TH! CONTROL OF .SCALE AND
CORROSlON IN WATER SYSTEMS

Thesis for the Dogma of M. S.
MICHIGAN STATE COLLEGE

Robart Forresteilo Mc Cauhy.
1949

This is to certify that the

thesis entitled

m CONTROL 0? 80“ MD CORROSION Ill INTER SYSTHS

presented by

Robert I. McCauloy

has been accepted towards fulfillment
of the requirements for

A degree in Whaling

Major professor

Date 16 1949

 

Ls-.>".'o~

 

 

ITHE CONTROL OF SCALE AND
CORROSION IN WATER SYSTEMS

By
ROBERT FORRESTELLE g99AULEY

A THESIS

Submitted to the School of’Graduate Studies of niohigan
State College of Agriculture and Applied Science
in partial fulfillment of the requirements

for the degree of

EASTER OF SOIEHOE

‘Depertment of Civil Engineering

1949

This thesis attempts a general study of the problem of
controlling corrosion. scale and tuberculation in water
supply systems. It covers twelve term hours of study. exper-

imentation and preparation.

A survey of recent literature concerning this problem
is briefly summarized. with especial attention to the use
of sodium hexametaphosphate as a corrosion control in water
systems. The use of sodium hexametaphosphate in several
cities is discussed, together with results of studies report-
ed in the literature in connection with the use of this chem-
ical (often known as a glassy phosphate) under various

condition approximating those found in a public water system.

‘ A series of tests with sodium hexametaphosphate and
other chemicals used in corrosion control is reported and
results discussed. In general these experiments are tests
to determine the effect of dissolved oxygen. pH, and various
additions of Calgcn (commercial name for sodium hexametaphos-
phate) upon the rates at-shich steel bars corrode. These
tests are described in some detail. but mainly consisted of
preparing various waters for test. and rotating steel rods
in these waters in such a way as to prevent the entrance of

any additional supply of oxygen.

An unpublished series of tests By Professor Edward F.

Eldridge of lichigan State College Engineering Experiment

he.
or
"PJ
O:
to
JC

 

 

of. l. Iii... e in". to f .

Station. studying the value of sodium hexametaphosphate

as a means of inhibiting corrosion. has been used as a
basis for these tests. The tests reported. however. are
almost entirely original upon the part of the author. and
do not follow the tests made by Professor Eldridge. except

in a general way.

Finally. a series of tests to determine the effect of
sodium hexametaphosphate upon the softening and iron remov-
ing capacity of a zeolite softener are described and

discussed.

DEFINITION OF TERMS

The terminolOgy used in this paper will be that
adopted.by the Committee on liter Conditioning Methods
to Inhibit Corrosion which reported.to the American

water works Association of June 24, 1942.

"The term.CORROSION is restricted to mean the removal
of metal from.sn exposed.surface regardless of it's
subsequent fate. Ihere the metal removed becomes
encrusedd on the pipe as an insoluble oxide or other
compound. the terms ICE or INCEQSEAIION are
used. m refers to the presence of suspended
iron compounds in the water, and.may'be due to corrosion
or to the fact that the water itself contains iron.
EITTINQ refers to the localization of corrosion in

small. toll-defined areas of penetration.“

THE PROBLEM OF WATER PIPE CORROSION

All types of waters are corrosive to a greater or
lesser extent. Some waters are so aggressive due to the
presence of certain minerals and gasses that definite
measures must be taken to protect conduits carrying these
waters. else there will be extensive damage to the metal
of the conduits and great nuisance to the consumers of the
water. Iaters found in most water systems are only slight-
ly aggressive. yet even these are corrosive enough to slow-
ly and steadily damage pipe lines. storage tanks. heaters
and other parts of water systems. In addition. most water
systems tend to build up blankets of iron oxide and hy-
'droxide which will on occasion break loose and harass
housewives and other users with epidemics of 'red water“.
Pipe corrosion. thereibre. from this view alone is a se-
rious enough problem to warrant extensive investigation.
for unexpected surges of iron oxide deposit often causes
extensive damage to laundry. household fixtures and other
utilities. 'Red water” nuisance always causes complaint
and tends toward dissatisfaction with a water supply man-

zugement.

From an economic viewpoint an even more costly and
damaging result of corrosion is the restriction and even

stoppage of pipe lines by the tubercles of corrosion.
u

These tubercles are particles of the pipe conduit. & to

1 inch long. which have broken loose due to corrosion. In

time these tubercles tend to almost fill portions of water
pipes. increasing friction and reducing carrying capacity

to a fraction of the original value. Removal of these re-
strictions is expensive and difficult. and at best a tem-

porary measure. Pipe lines must sometimes be replaced if

tuberculation becomes excessive. and even partial restric—
tion of pipe areas causes great expense due to reduced

capacity and increased power costs.

The corrosive action of water is greatly increased by
increased temperatures. Storage tanks. piping and heaters
iccnnected with hot water systems are gubject to damage far
more than cold water pipes. and often are eaten out and
damaged beyond repair in a fraction of their otherwise use-
ful lives.

As yet. little progress has been made in reaching any
complete solution of the problem. The following methods
have in some cases inhibited corrosion:

l. Galvanised coatings. Zink galvanizing is some-
times successful. yet may be attacked by some waters as
readily as the metal coated. Coatings of metals other
than sink may prove resitant to the more aggressive waters.
and therefore useful for iron pipes. but little data of
this nature is at present available.

5

2. Coating of paint. cement and other water-proof
materials. Various coatings have been used. some with con-
siderable success. Some of the most successful materials
have proved to impart a very bad taste and are therefore
objectionable. The most common and successful coating ma-
terial is a cement-sand mixture. Any coatings thus far
developed have been difficult to place and maintain during
laying of the pipe. and are sometimes disturbed. A "local-
ised' corrosion results. lo coating material has proved

useful in hot water lines.

3. Treating to raise the pH of the water. laying ,
down of a calcium carbonate scale on the inside surface of
the pipe and removal of carbon dioxide from the water.
These three means of protecting pipe lines are secured by
treating to obtain a positive stability index. usually
with lime or some other hydroxide. This method of protect-
ing pipe lines is that in most common favor at the present
time. Although one of the best and most popular methods
of controlling corrosion. the calcium carbonate scale is
very soft and portions tend to break away. leaving the
pipe exposed. In some instances this method has been of
little or no help at all. In particular it is very unsat-

isfactory in hot water systems.
H. The use of polyphosphates and silicates to pro-

6

duos protective films on the interior of pipe lines. These
chemicals have in some instances proved very successful.
and in other instances have proved unsuccessful. Experi-
mental data is not yet available on proper dosage and other
factors necessary to produce satisfactory protective films.
Polyphosphates. in particular. are one of the most promis-
ing methods of attacking the corrosive problem. This paper
will discuss sodium hexametaphosphate ( a polyphosphate)
commercially known as Calgon. a chemical which is coming

into extensive use as an agent for inhibiting corrosion.

5. Removal of oxygen. Oxygen is the most active
agent in causing aggressive waters. lo completely success
ful method of removing oxygen in large quantities has been
developed. vacuum. followed by sodium sulphate. heat or

scrap metal has been used experimentally.

6. Cathodic protection. Cathodic protection con-
sists of creating a slightly negative potential in the
metal which is sufficient to prevent the formation of pos-
itive Fe ions. This method is used extensively in connec-
ticn with the protection of water tanks. lo successful.
means of cathodic protection for water pipe lines has been

developed.

7. Asbestos cement pipe and other conduits which

are not subject to corrosion. Asbestos cement pipe. brass

piping in homes and glass/tube hot water heaters are prom-
ising new developments which promise to help eliminate much
of the 'red water' problem of domestic and possibly some
commercial systems. Asbestos cement pipe is coming into
wide usage in the United States at the present time. but
has not yet won the full confidence of practicing engi-
neers. Asbestos cement pipe does not have the strength of

cast iron pipe. Glass tube hot water heaters are a very

recent development.

Theory of the Chemistry of Corrosion

The most common explanation of the corrosion of
the interior metal of pipe lines is the electro-chemical
reaction theory. The reaction takes place in small
areas due to slight differences of metal structure and
of the water in contact with the metal. Using iron. the
most common example. we have the following chemical

explanation:

320 = 8* + 03’ ( water dissociates)
Te (dissolving) = Fe++ + 2 electrons
and s- 2 electrons = 2H ( Atomic hydrogen)

The hydrogen combines with the dissolved oxygen in

the water as follows:

Iron ions form ferrous hydroxide. which is one form

of iron rust:
Fol” + 203’ = ”(03);
re(os)2 + 02 (dissolved oxygen) = Fe(OH)3

It is obvious that the presence of dissolved oxygen

in water is the most active agent for causing pipe corro-

9

sion and that corrosion is not noticeably present except

in the presence of dissolved oxygen.

Carbon dioxide is also an active agent of corrosion.
Carbon dioxide combines with water to form carbonic acid.
which in turn dissociates into hydrogen ions. This increase
in the number of hydrogen ions. in turn increasing the rate
at which iron is dissolved into the water. An inspection
of the chemical equations of corrosion indicates that

ionic hydrogen is necessary for a rapid rate of corrosion.

‘ Carbon dioxide also has the effect of dissolving pro- .p
tective coatings of rust and other scale. leaving the pipe
walls clean and easily accessible to the dissolved oxygen

in the water.

The rate at which pipe lines corrode is greatly de- v“
pendent upon the purity of the iron of the pipe. Pure
iron does not corrode. but cast iron pipe is never pure
iron. but rather made up of iron of varying degrees of
purity. The relation between corrosion and impurities of
the metal of cast iron pipes is well explained by Babbitt

and Doland as follows:

“The corrosion of the metal (of the pipe) is the re-
sult of setting up an electro-chemical action between two
dissimilar metals. or between minute impurities in a metal

that bear a different electric potential toward each other.

10

sion and that corrosion is not noticeably present except

in the presence of dissolved oxygen.

Carbon dioxide is also an active agent of corrosion.
Carbon dioxide combines with water to form carbonic acid.
which in turn dissociates into hydrogen ions. This increase
in the number of hydrogen ions. in turn increasing the rate
at which iron is dissolved into the water. An inspection
of the chemical equations of corrosion indicates that

ionic hydrogen is necessary for a rapid ratecf corrosion.

. Carbon dioxide also has the effect of dissolving pro- .,
tective coatings of rust and other scale. leaving the pipe
walls clean and easily accessible to the dissolved oxygen

in the water.

The rate at which pipe lines corrode is greatly de- ./
pendent upon the purity of the iron of the pipe. Pure
iron does not corrode. but cast iron pipe is never pure
iron. but rather made up of iron of varying degrees of
purity. The relation between corrosion and impurities of
the metal of cast iron pipes is well explained by Babbitt

and Doland as follows:

'The corrosion of the metal (of the pipe) is the re-
sult of setting up an electro-chemical action between two
dissimilar metals. or between minute impurities in a metal

that bear a different electric potential toward each other.

10

An electric current is set up between the two dissimilar
metals. and it continues to flow until the electrolyte is
neutralized or the electrodes are changed into substances
of the same electric potential. Under normal conditions
in nature the electrolyte between the two metals is con-
tinuously renewed. so that the corrosion of the anode is
continuous unless other conditions set in to interrupt the
current. The rapidity of corrosion is dependent on the
difference of potential between the parts of the metal and
on the strength of the electrolyte to neutralise the elec-
tric charges delivered to it.

“The rusting of iron and steel can be explained by
this hypothesis. Iron rust is a hydrated oxide of iron
(Tea 03. nfizo). It will form only in the presence of iron.
oxygen and water. The absence of any one of these will
preclude the formation of rust. The oxygen dissolved in
the water and the water itself form a weak electrolyte so-
lution. low if 23;; iron is introduced into the water.
experiment has demonstrated that it will not rust. Ordi-
nary wrought iron and steel contain many impurities. The
less homogeneous the composition of the metal the greater
the difference of electric potential between spots on
the surface and. hence the more rapid.the formation of

rust or the corrosion of the metal.
“Among the principal factors to effect corrosion of

11

metals may be included: Electrode potential; overvoltage
of metal; protective-film formation by the metal itself;
homogeniety of the metal. both chemical and physical; the
kinds of metals involved; the electrolyte; the temperature;
protective-film; turbulence in the solution; oxygen concen-
tration in the solution; ionic concentration in the elec-
trolyte. particularly hydrogen-ion concentration; and elec-
trolysis. The relative importance of the various factors
causing corrosion varies with the environment involved;
but. in general. electrode potential. overvoltage. hydro-
genion concentration or pH. and oxygen concentration are

always important .

”Oxygen plays an important part in corrosion because
when metal goes into electrolytic solution nascent hydro-
gen is formed. which reacts with the oxygen. removing the
protective film of hydrogen. exposing more metalix surface
for the formation of hydrogen. with resultant corrosion of

the metal.‘

12

THE USE OF SODIUI HEMIETAPHOSPHATE
IR
CORROSION CONTROL

“The soluble leta- and pyrophisphates possess certain
properties which are unusual. The complex phosphate ions
produced by these substances are capable of taking up such
ions as calcium. iron and others which might be precipitat-
ed under conditions prevailing in the water. and of holding
them in the form of a soluble complex ion' - E.l. Moore.
Associate Professor of Sanitary Chemistry. Harvard Gradu-
ate School of Engineering and Chairman of the American
later Works Association Committee on Water Conditioning

Methods to Inhibit Corrosion.

The use of sodium hexametaphosphate in controlling
corrosion is a very recent development. not over ten years
old. Probably the first uses of this chemical as a corro-
sion inhibitor were in Columbus. Ohio. Dallas. Texas and
Hempstead Long Island.

For years the ability of sodium meta—and pyrophos-
phates to hold up ions of calcium and iron has been recog-
nised. Calgcn. a phosphate glass. was first used in water
works practice in Columbus. Ohio for the purpose of re-
ducing the amount of recarbonation necessary after lime-
soda softening. so that the pH of the water could be
raised for the purpose of inhibiting corrosion. The phos-
phate glass proved successful for this purpose. the pH was

successfully raised. and 'red water' conditions were much

13

improved by the treatment. particularly in hot water tanks.

Following the successful use of the chemical at Colum-
bus. its use was begun at Dallas. Texas. where a reduc-
tion of 'red water“ was obtained by the use of glassy phos-
phate alone. combined with a slow lowering instead of

raising of the pH.

At about the same time that too phosphate treatment
was begun in Dallas its use was also started at Hemetead.
Long Island in the expectation of improving I'red water' con-
ditions by raising pH. 'Red water difficulties were elim-
inated within a short period of time. but further tests
indicated that the high ph (9.2-9.u; was not necessary for
successful elimination of 'red water“ troubles. and equally

good results were obtained with pH of about 7.0.

Trials by the nearby cities of Garden City. Long
island and Fairhaven. Massachusetts. showed without ques-
tion. that satisfactory results could be obtained with
these waters with glassy phosphates at pH 6.2 and 5.“ re-
spectively and that the phosphate itself was capable of
greatly improving 'red water' conditions.

These successful uses of phosphate in water supply
systems encouraged further tests. both in the field and
laboratory. Experimentation indicated that glassy phos-

phates were capable of decreasing corrosion in iron and

In

other pipes by adsorbing the phosphates on the metallic
surfaces (probably as phosphates of the metals) and by
holding up ions of iron which would otherwise precipitates

as iron hydroxide (rust).

Although considerable laboratory data have been gath-
ered in connection with the use of metaphosphates in water
systems. the best results have been obtained in actual ap-
plication to water systems. The reason for this seems to be
that the metaphosphates require considerable volume and ve-
locity of flow. and that this volume and velocity (which
is often lacking in laboratory tests ) is as important to

good results as the concentration of the chemical.

One of the most interesting experiments carried on
with metaphosphates was that of E. Hood of the New Haven.
Connecticut later Company. Hood fed New Haven water (A
soft. mildy aggressive water with pH of 6.6-6.8 and iron
content of 0.05 ppm)through 45 foot lengths of 3/“ inch
black iron pipe connected to a constant level box. He
passed untreated water through all three pipes. Iron de-
terminations were made at various intervals of time and
showed that the iron content of the untreated water in-
creased 8 fold to an average of 0.u1 ppm. Treatment with
0.11 ppm of metaphoephate held the iron content to 0.11
ppm average. or-twice that of the raw supply. Treatment
with 0.33 ppm prevented\any apparent increase in the iron

15

 

 

 

content of the water after it entered the pipe. Soft marble
slabs were placed under the discharge of the treated water

for 72 hours without any signs of staining.

Mood also studied the effect of metaphosphate treat-
ment on loss of flow in 3/” inch black iron pipe. His ex-
perinents indicated that this treatment reduces. but does
not eliminate loss of flow due to tubercle formation. After
27 days test untreated water showed a reduction of flow of
u1.1$. 0.11 ppm treated.water showed a reduction of flow of
19.3% and 0.33 ppm treated water showed a reduction of

15.5%

Evans and 8huey.of the Cincinnati.0hic later Company.
experimented with tap water. allowing flow under constant
head an an initial rate of 3 gallons per minute through
two 20 foot lengths of Q-inch black iron pipe. h ppm of
mstaphosphate being applied to water passing through one
pipe. Results are tabluated on the following sheet.

is

Iron ppm

letaphosphats Untreated

Treated
Initial flow (gpm) 3 3
Decrease in flow (gpm) 0.31 1.5“
Decrease in flow ($) 10 51
Increase in loss of head (inches) 37.25 78.1}
flasen lilliams C at start 111 111
Essen lilliams c at end 65 27
R s f I w Te t n t t an Ie h s -
T e T It 0 Aft D B 0

Also of interest and deemed worth of reproduction
is a plotting by Ir. H.P. Stockwell. Jr.. Chemical En-
gineer for the Ottawa. Ontario Hater Purification Plant.
showing the effect of metaphosphate treatment on iron
content of water passing through new wrought-iron pipe.

 

 

.-.+.... r

I. .onT—l Js"...~. LI Il.‘,..t.l-mu'. . a ..'e:’-

lhile most data of field use of the glassy phosphates
are of a favorable nature. many conflicting results have
been reported which seem to indicate that the chemical is
not successful as a corrosion inhibitor in all cases. Some
very reliable engineers and chemists are of the belief that ,
the phosphates do not have any value as corrosion inhibi-
tors. For instance. Frank E. Hale. Director of Laboratories.
Department of later Supply. Gas and Electricity. New York
City. is of the opinion that hexametaphosphate is corrosive,
and that reported improvement of 'red water‘I conditions are
due to the ability of the phosphates to hold iron in solu-
tion rather than any ability to inhibit corrosion. Because
of his strong beliefs he has not permitted the use of the
phosphates in New York City. fearing damage might result to
the cities' buildings. A few other well recognized men in
the waterworks field seem to agree with Hale. However.the
general concensus of opinion seems contrary to that of Ir.
Hale. and much evidence has been gathered to substantiate
the opinion that the phosphates do have great value. Indeed.
many adverse reports seem to be based upon inconclusive
data or experiennts under conditions vhich do not measure

the phosphates' true value.

Already mentioned in this connection is the necessity
for a large volume of flow in pipe lines to obtain a good
coating of iron phosphate protection on the inside of the

18

 

I9rilu
I D

... s... ..

in! viral" a, .s

i .I D

t 4|-

 

water pipes. Tests indicate that two or three parts per
million concentration with continued passage of water
through pipe lines is of more value than a concentration
ten times as strong under stagnant or low flow conditions.
In fact recent studies indicate that hexametaphosphate is
of only slight value unless accompanied by a considerable
volume and velocity of flow. However. once a satisfactory
film is formed. stagnant or low flow conditions no longer
seem to affect the efficiencies of the phosphates.

For this reason it is usually recommended that an
initial feed of five or more parts per million be used for
a period of several weeks. followed by a slow reduction
to one or two parts per million. It is of course most im-
portant that the initial heavy concentrations of hexame-
taphosphate be accompanied by as high a volume and veloci-
ty of flow as possible.

Tests also indicate that the usual concentration of
one or two parts per million of glassy phosphate do not
inhibit corrosion of hot water systems. though they prob-
ably do reduce the rate of corrosion. It is well recog-
nised that hot water is several times as corrosive as
cold water and therefore is much more difficult to con;
trol. A commercial form of sodium hexametaphosphate

known as Ilicromet has proved very helpful in inhibititg

19

. .llbh huff.Erilf.-vllsr all: - Dis, "

 

 

corrosion in hot water lines. licromet is a glass which
dissolves at a rate of about 0.8% of its weight per day.
If placed in the cold water inlet line so as to supply
about 25 parts per million concentration to the water a
considerable reduction in the rate of corrosion is observ-
ed. Reports indicate that glassy phosphates in a strength
of 25 parts per million will completely inhibit corrosion

in many hot water lines.

Studies of corrosion show that the rate of corrosion
normally increases with an increased rate of flow. The
corrosion of iron results in the formation of a protective
film of hydrogen gas on the pipe interiors. but this film
is swept away by the movement of rapidly flowing water.
The large volume of water flowing provides an ample supply
of oxygen. and corrosion is very rapid. However. studies
by Owan Rice of Calgon. Inc.. and others have indicated
that high rates of flow decrease instead of increase the
rate of corrosion if as much as one part per million of

sodium hexametaphosphate is added to the water.

In this connection. studies by Rice with #0 foot
lengths of black iron pipe are of interest. Rice stud-
ied head losses in the pipes during 28 days of.time.re-
porting results obtained with pipes through which water «

with concentrations of zero to five parts per million of

20

/.

sum. .. {WWIgnl-uwlelsy5rtr.’ {Ilia-W

phosphate was passed. His results indicate that in this
particular test. five parts per million of phosphate re-
duced the head loss due to formation of tubercles to one

sixth of that obtained with raw water.

These experiments by Rice show that water of low pH
corrode pipes in a uniform manner with little tuberculation.
As the pH increases corrosion becomes more localized and
the result is formation of tubercles rather than a uniform
corrosion. If metaphosphate is added. this same condition
holds. except that tubercles are smaller in number and

size. darker in color and more dense.

when a test similar to that Just described with to
foot lengths of black iron pipe was performed to study the
effect of pH on head loss due to tuberculation. results in-
dicated that treated and untreated water showed a head loss
rise over ten times as high at pH 7.5 as at pH 5.0. A con-
centration of 5 parts per million of metaphosphate helped
reduce rise in head loss in pipes exposed to water with
both high and low pH. The test shows that high pH. al-
though greatly reducing rate of corrosion. does increase

head losses in pipes due to formation of tubercles.

It would seem that metaphosphate treated waters give
best results with low pH because of the ability of phos-

phates to give protection against corrosion. a protection

21

 

 

,I.

. e
a
» '
I.
.~
4o
e.‘
a.
h
..
§
,;
. V
.r.
‘
3
u
e

_. J tdlnE. T... at. first. ff, ..

E‘si‘
\

 

that heretofore couli be obtained only by raising pH. These
lower ph's reduce tuberculation. and therefore reduce pipe
head loss. This procedure of reducing pH is contrary to
standard water works practice. but is the method recommend-
ed by operators who have used hexametaphosphate with suc-
cess.

The matter of reducing the pH of water from standard
values of 8.5 or 9.0 to around 6.0 or 6.5 is not without
problems in itself. .A quick reduction in pH will cause an
epidemic of 'red water“ with accompanying dissatisfaction
and complaints. Since metaphosphate has the effect of low-
ering pH. its use requires a careful regulation of pH dur-
ing the first few weeks of use. pH is most successfully
lowered by very small incrments to the desired value. Even
those small increments of reduction are often accompanied
with a breaking loose of rust crusts and.a very rusty water.
If an adequate plan of flushing accompanies the reduction
of pH. little if any difficulties will result.

Although the reduction of corrosion and.tuberculatioh
are of great importance. the prevention of precipitation
of calcium carbonate scale and the prevention of 'red water"
due to precipitation of dissolved iron from well water
(after oxidation) are equally vital to good water works

OPOrat 1011 e

Good'custcmer relations” and maintenance of resonable

22

pipe head losses are usually dependent upon preventing
scales of rust and calcium carbonate from forming on the
inside of pipe lines. Such precipitations cause roughness
on the inside of the pipes. reduce the area available for
carrying water and often break loose at intervals to cause

'red water“.

While there may be some questions as to the ability
of the metaphosphates to inhibit corrosion. there is no
question as to the ability of the chemical to hold cal-
cium and iron in a complex ion. In the past. most water
supply people have preferred to have a certain amount of
precipitation of calcium carbonate upon the interiors of
pipe lines to prevent corrosion. But if such protection
can be obtained with a phosphate-iron protective scale on
the pipe interior these calcium carbonate scales are unde-
sirable since they increase head loss. decrease carrying
capacity. and tend to clog water meters and valves. Cal-
cium carbonate scale in hot water coils is most undesir-
able since such scale reduces the ability of the coil to
conduct heat.

later treated by the lime-soda method tends to form
heavy deposits of calcium carbonate scale because the
treatment precipitates calcium carbonate and recarbona-
ticn often fails to change all suspended carbonates into

bicarbonates. Many such plants are now using metaphos—

23

phates to hold the calcium carbonate in suspension. and

to avoid calcium carbonate growth upon the grains of filter
sand. For instance. the Lansing. Michigan lime-soda treat-
ment plant adds 0.25 parts per million of lalco-5l9 ( a met-
aphosphate manufactured by the Rational Alumbnate Corpor—
ation) just before filtration and 0.25 parts per million of
lalco-5l9 after filtration.

lore interesting yet is the ability of small amounts
of metaphosphates to prevent the precipitation of dissolv-
ed iron frOm well water. lell water often contains appre-
ciable quantities of dissolved iron and usually is devoid
of oxygen. lhen such water is aerated in reservoirs and
storage tanks oxygen is absorbed. causing the ferrous iron

to be oxidized to a ferric iron. which is a form or rust.

Rust so formed often precipitates in water mains and
causes 'rsd water' after a rise of fall in pH or other
change in the normal chemical condition of the water. Any
water with over 0.3 parts per million of iron is generally
considered to be of questionable value as a public water
supply. Iater far below this value in iron content can form
rust scale upon the interiors of pipe mains and seriously
tharaes the water supply management with 'red water“ prob-

lems.

The addition of one part per million of sodium hexa-
2k

 

..I\Il....

.21 “Ji'il ’Illltl‘

 

metaphosphate for each part per million of iron (as Fe). if
added before chlorination and before the water has an oppor-
tunity to absorb oxygen will prevent the precipitation of
dissolved iron. Once the iron is held in a complex ion by
the phosphate the iron will be carried completely through the
distribution system.

After prolonged contact of the water with air the iron
will oxidize. but 2 parts per million of iron is not notice-
able when oxidized and an iron content below 5 parts per mil-
lion is not particularly objectionable. though there may be
some taste. If as much as four times as much metaphosphate
as iron is present in the water. the iron will not be oxi-
dized. even if exposed to a very high oxygen content in the

water after the metaphosphate is added.

The ability of metaphosphates to hold iron in solur
ticn has been strikingly demonstrated to this writer by ob-
servation of the results of tests with iron rods. describ-
ed in this paper. In all cases. it was observed that
tests which badly discolored untreated waters did not cause
any particular discoloration of waters containing twelve
or more parts per million of Calgon. Since in many of
these tests the loss of iron by the rods exposed to Calgon
treated waters was as high as the loss in untreated waters

it must be concluded that heavy concentrations of iron

25

 

 

were held in suspension in'the treated waters by the action

of phosphate ions.

lhen metaphosphates are dissolved in water the phos-
phorous pentoxide P2 05 combines with water to form meta-

phosphoric acid or pyrophosphcrio acid thus:
P2 05 + 820 8 2 HPO; (letaphosphorio Acid)
p205 .4 2 H20 .-. 31.9207 (Pyrcphcspherie Acid)

Both metaphosphcric acid and pyrophosphcrio acid are
capable of holding iron and calcium in suspension. Tm two

acids are in almost every way similar.

letaphcrphcric acid and pyrophosphcrio acid both slow-
ly hydrolise to crthophosphoric acid. which has no value.

as an inhibitor of corrosion.

letaphosphates are usually fed by dissolving the phos-
phate glass in solution. Because of the tendency to hydro-
lize to an crthophosphate it is well to make up a fresh so-
lution every ten days. lost metaphosphates are shipped in
broken glass in 100 pound waterproof bags. Not over three
months supply should be on hand at one time because the
glass tends to become covered with a film of reverted phos-
phate if stored under moist conditions. The addition of
soda ash in small quantities helps keep strong solutions
in a metaphosphate form for a longer time than is other-

wise possible.
26

 

nirivil-l|x|1|l
..r. I
. . . 3 ., Inljt.

letaphcspates are a true glass and have no lindting
solubility. However it is recommended that a maximum feed
solution of one pound per gallon be used. because more con-
centrated solutions become too viscous to handle. The
usual feed solution used varies from one part per hundred
to one part per thousand concentration. The solution is
usually fed by a force feeding arrangement of the type
usually used for adding chemical solutions to water supplies.

High concentrations of metaphosphate are highly cor-
rosive to metal. 'ood or concrete tanks are best for hold-

ing the concentrated feed solution.

An exhaustive study of the physiology of sodium hema-
metaphospahte by K.K. Jones. Assistant Professor of Physi-
ology and Pharmacology. lorthwestern Medical School. has
led to the conclusion that the metaphosphates are rela-
tively non-toxic when taken by mouth. and can cause no
toxic or other undesirable effects in concentrations which
might be used in water supply systems. Orthcphosphates are
considerably more toxic than pyrophosphates. However. it
is estimated that a concentration of crthophosphate fifty
times higher than would ever be used in a water distribu-
tion system would be required to produce an even slightly
toxic effect. Feeding very heavy doses of metaphosphate

to rats and rabbits over long periods of time produced no

27'

 

K

 

unpleasant results other than temporary cases of diarrhea.

letaphosphates have no effect upon sofening or iron
removing capacity of seclite softeners. A series of exper-
iments reported in this paper show no difference in results
with treated and untreated water. It is interesting to note
that hexametaphosphate is a sequestering or softening agent
in itself and capable of softening water. It is often
used for this purpose in households and laundries. However.
the very small additions of phosphates used for scale and
corrosion control are too minute to have any sofetning ef-

foot.

The cost of adding sodium.hexametaphcsphate to a water
supply varies from 31.00 to $1.50 per million gallons for
each part per million added. Thus. if two parts per mil-
lion is added at a cost of about $2.50 per million gallons
treated the annual cost would be about 8900 per year for
a water system supplying one million gallons per day or
about 3.09 per capita per year. This cost is not exces-
sive if corrosion is inhibited. 'red water' conditions are
corrected and pipe tuberculation is reduced. The cost of
sodium hexametaphosphate is from ten to fifteen cents per

Pound.

At present over 300 water supplies throughout the
country are using metaphosphates. Results reported by a

28

 

.s-. .I‘Ir ii.‘tuzhh‘ eh’kfi—
. .

.r’l ... h.

 

few of these water supplies are noted below:

ngghaven. Massachusetts

Some complaint was first reported after use of Calgon
was begun due to initial disturbance of old iron oxide de-
posits. but after pH was lowered to 5.5. 'red water“ con-
ditions disappeared and such conditions are now unknown.

pH was lowered over a period of several months.

winghegtgz. gasgachusetts
An original feed of 10 ppm of metaphosphate was used

for one week. then dropped to 2 ppm. Results were satis-
factory so a feed of 2 ppm.was applied to the enttre city
system. Within one month all iron pickup in pipes had dis-
appeared. The improvement of conditions was greatly ac-
celerated by thorough flushing of mains to increase rate of

flow and remove loose deposits of iron oxide.

Sumter. Sguth Cazglip;
Sumter. South Carolina suffered periodically from

“wild water“. This was caused by the necessity of turning
into the system a well with iron content of 3.8 ppm during
periods of high demand. The water was 'tamed' tn the well
water with high iron content. Results were so satisfact-
ory that the well with high iron content was kept in con-

stant use.

29

 

, » 1....n!»

It}

a}.

Cgptgn.'0hig

Results were similar to Sumter. South Carolina in that
a well with 1.5 ppm iron had to be turned into the distribu-
tion system during periods of high demand. Addition of 3 ppm

of metaphosphate corrected all 'red water“ troubles.

ngwind. West Vigginia
Berwind water supply is treated by the lime-soda method

 

at the rate of #00.000 gallons daily. Considerable “after
precipitation“ of carbonate deposit developed. both on fil-
tor sand and in lines and valves of the filtration plant.

2 ppm of 'Halco Ho. 18 Ball“ (a 66% phosphate. 15% sodium
carbonate glass) eliminated all difficulties.

Hitgg, Vegt Vigginia
The city of Dunbar. West Virginia is supplied with

water by a seven mile long pipe line from Nitrc. The line
was cleaned mechanically. resulting in widespread Fred wa-

ter' complaints. Iron content of the water ran as high as

3.4 ppm.

Calgon was applied in a concentration of 10 ppm for
two days. then reduced to 5 ppm for two days. then applied
in a concentration of 1 ppm for six months. Application

was at the Nitro end of the seven mils line.

Ho residual of metaphosphate was found in the Dunbar

end of the line after six months time. .Iron content rose

30

 

2.4 .IIAI'IIIIVIILEDI: JWs

 

and fell during the period of application. Just as good
results were obtained after use of metaphosphate was dis-
continued as during the time of use. Sodium hexametaphos-
phate did not demonstrate any capacity to inhibit corrosion

or reduce 'red water."

Siggx Falls Army Air Basea Sguth Dakgta
125 failures of hot water systems per year was reduced

 

to zero failures per year by adding 5 ppm of sodium septa-

phosphate.

‘Las Vegas Army Air Bags. How lexicc

numerous complaints of “red water“ were received each

 

day. 2 ppm of Calgon reduced 'red water“ to almost zero.
Four months flushing of mains. coupled with application of
Calgon in the above concentration cleared mains. showed a
great reduction in the rate of hot water tank corrosion
and greatly improved the entire system with respect to

head losses.

CONCLUSIONS

1. letaphosphates hold iron and calcium in complex
ion combinations if added before water is chlorinated or
aerated. This action greatly reduces the effects of
I'red water'. due to iron pickup in the pipe lines and to
natural iron content of the water. Calcium carbonate
scale is greatly reduced by the addition of metaphosphate
to water distribution systems and calcium carbonate
growth is prevented upon sand grains of sand filters by

small concentrations of the chemical.

2. letaphosphates in many cases have proved very
helpful in the reduction and even total elimination of
corrosion in pipe distribution systems. However the
chemical has not been successful in all cases and the
ability of metaphosphates to inhibit corrosion in all 2
cases is in question. Many conflicting reports. both

pro and con. are reported in the literature.

3. Metaphosphates are non toxic in concentrations
employed for inhibiting corrosion and reducing scale and

'red water'.

h. Volume and velocity of water passing through
pipe mains appear to be as important as quantity of phos-

32

phate used. Large pipes require lower concentrations for

good results than small pipes.

5. Application of metaphosphates at the rate of 0.2
ppm to 10 ppm is recommended. depending upon results de-
sired. pipe velocity. volume of flow and length of time

over which the metaphosphate is used.

6. Use of metaphosphate is not unduly expensive if

corrosion is inhibdtsd or “red water“ corrected.

7. Inhibiting of corrosion in hot water systems is
more difficult than in cold water systems and a concen-
tration of 25 ppm or more of metaphosphate is often re-
quired for satisfactory results. This is most easily
provided by a slow dissolving phosphate glass such as

Micromet.

8. Metaphosphates have no effect upon the soften-
ing and iron removing capacity of a zeolite softener.
Their use previous to lime-soda softening will interfere
with the softening unless a slight excess of alum. iron
sulfate. or other coagulant is added to adsorb the meta-

phosphate.

33

TESTS TO STUDY THE CORROSION OF STEEL BARS
UNIER VARIOUS CONDITIONS WHICH MIGHT EXIST
IN MUNICIPAL AND OTHER WATER SUPPLY SYSTEMS

GENERAL DISCUSSION OF TESTS WITH STEEL BARS TO STUDY
THE VALUE OF SODIUM HEXAMETAPHOSPHATE IN CONTROLLING
CORROSION IN IRON AND STEEL PIPES OF RATER SYSTEMS

Shortly after the end of Warld War 11 Professor
Edward F. Eldridge of Michigan state College Engineer
Experiment Stat ion began a series of emperiments with
Sodium Hexametaphosphate. more commonly known by the
trade name of Calgon. His purpose was to study the
value of this chemical in controlling corrosion of
iron and steel pipes. Professor Eldridge served as
an employee of the U. 3. Army during the war and in
his work used Calgon extensively in combating corrosion
of pipes and plumbing fixtures in army carnps. It is
his cpinicn that several million dollars was saved

during this period by using Calgon in corrosion control.

The author has studied the results of these exper-
iments. and has used them as a basis for the studies
described in this paper.. Some changes have seemed
wise in view of results obtained from the early exper-
iments” For instance, Professor Eldridge originally
began his experiments with the idea of using the iron
content of the water from bottles (after tests) as an
index of corrosion. Loss of weight of bars, however.

proved a matter and more easily studied index of corr -

34

osion, and during his later experiments he used this
method of measuring corrosion of the bars. This latter

method is that used in this series of tests.

In performing these experiments the author has
rebuilt the apparatus used by Professor Eldridge, and
has found it a satisfactory method for making a general
study of the value of Sodium Hexametaphosphate under

various conditions such as might exist in water systems.

35

DESCRIPTION OF APPARATUS AND PROCEDURE USED IN TESTS

WITH STEEL BARS AND VARIOUS STRENGTHS OF CALGON SOLUTION

APPARATUS

1.
2.
3.

Mechanism to revolve rods in quart milk bottles.
12 one quart milk bottles.

12 quarter inch cold-rolled steel bars 5% inches
long.

12 rubber steppers fitted with short pieces of
glass tubing to act as bushings for stirring rods.
12 six-inch lengths of stirring rods to connect
stirring mechanism to steel rods.

Short pieces of rubber tubing to fasten steel rods

being tested to glass stirring rods.

PRELIMINARY £1300an

1.
8-.

b.

Clean steel rods thoroughly.

wash and.remcve dirt and rust with steel wool.
wash with carbon tetrachloride.

wash with alcohol, dry; and place in desiccator.
so not touch with hands after alcohol wash.

When dry weigh to nearest 0.0001 gram.

EQEE; This procedure is used after each test run except

that steel wool is used to clean bars before first test

only. and is omitted thereafter. Bars are marked with

file marks for identification.

MAKE UP WhTER

I.
2.
3.

NOTE:

 

Draw off five gallons of tap water.

Hake dissolved oxygen test on water.

Siphon the water to each bottle, filling com-
pletely. Add variables and.make pH tests.

Attach weighed bars to glass stirring rods with
short pieces of rubber tubing.

Insert stepper to fill the glass tube in the cork

and remove all air bubbles.

Each bottle is calibrated for volume of liquid

it contains during actual test, that is with glass

stirring rods. steel rods. and rubber tubing in position

for operation. Calgon is made up in 1000 parts per

million strength and added to test milk bottles by means

of pippettes.

PROCEIURE

1.

Allow test bars to rotate at specified.3peed
(usually 42 rpm) for 24 or 48 hours.

Step stirring apparatus. Remove test bars

and wipe dry. Clean with hydrochloric acid
solution, wamh with distilled water, rinse
with carbon tetrachloride and alcohol,place in
ddsiccator and weigh.

Insert specially'prepared stapper for with-
drawing water after tests. snake bottle and
‘Vithdraw samples for dissolved oxygen and

pH tests.

57

4. Make dissolved oxygen and pH tests.

5. Record results and plot applicable curves.

38

 

 

 

 

 

 

STIRRING APPARATUS USED IN TESTS WITH STEEL RODS

‘ due to corrosion

Grams loss in weight per rod

TEST NUMBER ONE

Curve‘ indicating loss in weight of steel bars
with various additions of Calgon (in ppm)

..ozeo
..o24o
-02..
,0...
..0180
-.0160
..0140
..0120
. .6156
. .ooao
..ooeo

. .0040

0

————-—— 48 hour test
24 hour test

———

 

   
  

G)

X

Dnnn- 2hrm. 4srn. ennui Bhnn lonhn

Parts Per llillion Calgon added

.39

TABULATION OF RESULTS FROM TEST # l.

MICHIGAN

STATE COLLEGE TAP WATER WETH CALGON SOLUTION

Bottle Capacity Calgon wt of rod It of rod Loss in
(ml) added before after weight
(ppm) test test (grams)
(grams) (grams)
1 928 0 35.1941 35.1737 .0203
2 920 2 35.2914 35.2670 .0244
3 924 4 35.0860 35.0662 .0198
4 916 6 35.2376 35.2203 .0173
5 920 8 35.2448 35.2269 .0189
6 919 10 35.3440 35.3353 .0093
7 923 0 34.8910 34.8661 .0249
8 921 2 35.3459 35.3238 .0221
9 928 4 35 . 2442 33 . 2211 . O 231
10 932 6 35.0930 35.0702 .0228
11 919 8 35.1606 35.1400 .0206
12 917 . 10 34.8900 34.8719 .0181
Remarks:

Dissolved oxygen bottle 1 before test 8.6. After 28 hours

2.4. Dissolved oxygen bottle 7 before test 8.6. After
48 hours 1.7. Beds in bottles 1-6 exposed 24 hours.

bottles 7-12 exposed 48 hours

40

Discussion of Test Number One

Test number one was intended as a 9:481 test in
measuring the value of Sodium Hexametaphosphate (Calgon)
as a means of deterring the corrosion of iron and steel
in water pipes. Rocedure consisted of adding various
amounts of Calgon to test bottles. filling with water,
inserting weighed steel rods. making a dissolved oxygen
test of the water used. and revolving the rods for periods
of 24 and 48 hours. Following these periods of emosure
the rods were washed together in dilute solutions of
hydrochloric acid. carefully weighed, and loss of weight
determined. Loss in weight was used as an index of
corrosion, and loss of weight has been plotted against

parts per million of Calgon used on the accompanying sheet.

Of particular interest is the tendency of loss of
weight to decrease with increased amounts of Calgon.
The sharpest break in the curve is noted to be between
8 and 10 parts per million strength. It must be noted
however that small increases in the amounts of Calgon
used does not always result in similar decreases in loss
of weight. 0n the contrary increased amounts of Calgon
seem to cause noticeable changes in weight losses at

certain points in the curve only.

Also of interest is the flatness of the 48 hour

41

curve as compared to the 24 hour curve. This is
obviously caused by the almost complete chemical com-
bination of most of the dissolved oxygen with iron
during the first 24 hours of exposure. The rate of
corrosion is sharply reduced during the second 24 hour
period, particularly with respect to the rods exposed

to water containing little or no Calgon. Dissolved
oxygen at the beginning of the test was 8.5 ppm, after
24 hours 2.4 ppm and after 48 hours 1.6 ppm, thus indi-
cating that the greater part of the combination of

iron with oxygen to form oxides and.hydroxides took
place during the first 24 hours. The curve showing

loss of weight with various strength Calgon solutions defi-
nitely shows that oxygen dissolved in water is the

most important agent of corrosion, and.that the rate

of corrosion was several times higher during the first
24 hours when an ample supply of oxygen was present than
.during the second 24 hour period when oxygen was largely

absent.

Since the rods were all washed together in dilute
solutions of hydrochloric acid. loss of weight of the
rods seems to be an excellent method of measuring corro-
sion. After washing with hydrochloric acid solution
all rods were a bright gray color. indicating that all
ferric oxide and hydroxide were removed. An even stronger

argument for believing that this procedure is a true

42

measure of corrosion of the steel bars is the appearance
of the test bottles after the 24 hour test. Number one
bottle was very red in color (a typical I'red water“) and
contained a very heavy concentration of red,precipitate.
No Calgon was added to bottle number one. Number six
bottle, containing 10 parts per million Calgon was still
reasonably clear after a 24 hour period. slightly red.
and with a very light precipitate. The other four 24éhour
test bottles varied in color between these exmremes. very
much in line with the losses of weight noted. No such
comparison could be made after 48 hours of exposure
because all bars had become quite corroded after that
time.

Dissolved oxygen tests were made on bottles con-
taining untreated water (no Calgon) only. No pH measure-
ments were attemptdd.on this first test. Tests in whidh
particular attention has been paid to these factors.
together with tests with heavier concentrations of Calgon

are shown later.

43

. due to corrosion

Grams lose in weight per rod'

TEST NUMBER TWO

Curve indicating loss in weight of steel bars
with various additions of Calgon (in ppm)
added to Michigan State College tap water
from which dissolved oxygen has been removed

.0140
10130
.0120
.0110
.0100
.0090
.0080
.0070
.0000
.0050
.0040
.0030
.0020

.0010

X ‘ter not treated

0 psi

4 ppm

flippn-

 

.12 ppm 16 ppm .20 mm

Parts Per Million Calgon Added

TABULATION OF RESULTS FROM TEST # 2. MICHIGAN
STATE COLLEGE TIP WATER WITH ZERO DISSOLVED OXYGEN

VARIOUS STRENGTH CALGON SOLUTIONS ADDED TO WATER

Bottle Dissolved Calgon wt of rod.ln of rod.Loss in
oxygen added before after weight
(pm) (ppm) (creme) (grams) (crime)
1 . 8. 2 0 35.1737 35.1599 .0138
2. 0.2 20 35.2700 35.2685 .0015
3. 0 2 35.0662 35.0645 .0017
4. 0 4 35.2205 35.2176 .0029
5. 0 6 35.2269 35.2234 .0033
6. 0 8 35.3357 35.3324 .0033
7 . 0 10 34. 8661 34. 8648 . 0013
8. 0 12 35.3238 35.3219 .0019
9. 0 14 35.2211 35.2189 .0022
10. O 16 35.0702 35.0685 .0017
11. O 18 35.1400 35.1391 .0009
12. 0 20 34.8719 34.8704 .0015
Benefits:
Bods were exposed to solutions for 24 hours. They were

stirred for 12 hours. then left in solution without

stirring for 12 hours.

45

Discussion of Test Number Two

Test number two was performed to observe the
relationship of dissolved oxygen to the corrosion of
steel bars, and to find out whether Calgon is an
effective means of combatting corrosion in waters

containing very low quantities of dissolved oxygen.

Test bottle number one was used as a control,

that is no Calgon was added. and the dissolved oxygen
content was permitted to remain the same as Michigan
State College tap water. Test bottle number two also
contained a normal amount of dissolved oxygen. (8.2 ppm)
but 20 parts per million of Calgen was added to this
bottle. All dissolved oxygen was removed from the other
ten test bottles. and various amounts of Calgon were
added. varying from 2 parts per million for test bottle
three ts 20 parts per million for tstt bottle twelve.

Results from this test are interesting. Test
bettle one, the control. lost .0138 grams of weight
due to corresion. The bettle contained a typical "Red
‘ter' after 24 hours of testing. with a light sediment
of red percipitate in the bettom. The rod was covered
with a light film of rust. which stained a cloth used
for wiping the rods. This loss of weight ef .0138 grams

46

v.41llulull'l.s.hfi.lniflalw

compares with .0203 grams of weight lest by the rod
exposed to similar conditions in test one. This rar-

iance in loss of weight is probably explained by the

slight variation in procedure, that is the rods were
retated a full 24 hour period during test one, while

test two consisted of a 12 hour period of stirring followed
by 12 hours of exposure to the solutions without stirring.
This would seem to indicate that corrosion proceeds at

a higher rate when water is in motion than when it is
still, though of course this single test can be used as

an indicat ion only.

All rods in solutions from which dissolved oxygen
had been removed were still of a bright gray color after
24 hours of eaqaosure,indicating a very slight corrosion.
A comparison between test bottles one (which contained
8.2 ppm of dissolved oxygen and suffered a weight loss
of .0138 grams) and three (which contained no dissolved
oxygen and suffered a weight loss of only .0015 grams)
when clearly that iron and steel corrosion in water
pipes is caused by the dissolved oxygen content of the
water. It is true that bottle three contained 2 parts
per mill ion of Ca1gon,but this amount has only a slight 0‘

influence on the rate of corrosion.

A comparison of weight lessee in bottles containing

various strengths of Calgen shows that it has helped only

4'7

slightly in reducing corrosion in these oxygen-free

waters.

All bottles from which dissolved oxygen had been
removed contained almost completely unstained waters
after the test. A comparison between the colors of the
waters after the test is in many ways as good an indi-
cation of corrosion as the loss in weight, and.as indi-
cated above, in this case showed a typical red water in
the bottle containing 8.2 parts per million of dissolved
oxygen and no Calgon, and almost no change in celor

in bottles containing no dissolved oxygen.

Dissolved oxygen was removed from bottles three
through twelve by adding 150 parts per million of sodhmm
sulphite to each bottle. This chemical combines with
dissolved oxygen to form.sodium.su1phate. and is some-
times used for removing small quantities of dissolved oxy-
gen from a water of a municipal water system. It must
be used.with great care, however. since it is poisonous,
and can be used in only very small quantities. As the
figure of 150 parts per million indicates. a consider-
able quantity of this chemical is necessary to remove all

of the dissolved oxygen from a water.

Bettle two. containing 8.2.parts per*million of
dissolved oxygen and 20 parts per million of Calgon

demonstrates that this strength solution of Sodium Hexa-

48

metaphosphate is very effective in reducing corrosive
effect of the dissolved oxygen ininichigan State College
tap water. The red removed from this bottle after 24
hours of exposure was almost as bright and shiney as
when put in, and only .0015 grams of weight loss was
noted. Almost no staining of the water containing

20 parts per million Calgon was noted.

49

Grams loss in weight per rod'

-.0200

due to corrosion

TEST NUMBER THEE

Curve indicating loss in weight of steel bars
exposed for 24 hours to Michigan State College
tap water which was adjusted for various pHs
by the addition of sodium hydroxide and calcium
hydnoxide

'.0260

'50240

.0200

.0180

.0160

.0140

.0120

.0100

.0080

.0060

 

.0040

.0020
0 ”.8 ‘8.4 0.8 9.6 10.2 ‘1008
pH of Nor

50

MICHIGAN
STATE COLLEGE TAP WATER WITH VARYING pH

TABULATION or RESULTS FROM rear #3.

Bottle pH before pH after It of rod Wt of rod .Loss in
test test before after weight
(ppm) (ppm) (grams) (gnome) (grams)
1 7.4 7.2 35.1536 35.1303 .0233
2 7.5 7.5 35.2630 35.2449 .0181
3 7-9 8.0 35-0577 35-0375 -0202
1+ 8.2 8.0 35.2176 35.1929 .0247
5. 8.5 8.1 35.2234 35.2024 .0210
6 8.7 8.0 35.3324 35.3134 .0190
7 9.2 8.5 34.8648 34.8475 .0173
8 10.0 9.8 35.3219 35.306“ .0155
9 10.3 10.2 35.2189 35.2067 .0122
10 10.8 10.6 35.0685 35.0618 .0067
11 11.1 10.8 35.1391 35.1335 .0056
12 11.4 11.1 34.8704 34.8676 .0028
Remarks:

7.“ represents tap pH of Michigan State College water.

pH of bottles 2-6 raised by addition of lime Ca(OH).

pH of bottles 6-12 raised by addition of sodium hydroxide

solution la(0H).

51

No chemical added to bottle #1.

Discussion of Test Number Three

Test number three is a study of the effect of pH
on rate of corrosion of steel bars“ Performance of
this test consisted of making up a saturated solution
of lime water and decanting the clear solution. This
clear lime water was added to bottles two to six in-
clusive in increments of five milileters - that is.
five milileters to bottle two. ten milileters to bottle
three, fifteen milileters to bottle three etc. Bottle
one was left untreated, and used as a control for
comparison purposes. Because of the reaction of lime
with the bicarbonate hardness of the water. a percipi-
tate was noted after thirty milileters had been added
to bottle sevon. For this reason, bottles seven throng
twelve were treated with sodium hydroxide solution to
raise the pH. Explaination of the formation of the
percipitate lies in the combination of lime water with
dissolved carbon dioxide in the water until all carbon
dioxide was removed. Following this removal of carbon
dioxide the lime water was free to react with the bicar-
bonate hardness of the water. )3 of the water was raised
by the combination of lime water with carbon dioxide to
form Calcium Bicarbonate, but as soon as this reaction
was completed further additions of lime water caused a
fall in pH due to removal of the bicarbonate alkalinity

52

when lime combined with calcium bicarbonate to form

non-soluble cal cium carbonate .

Rods were inserted in bottles as usual, and revol-
ved for periods of twenty four hours in the test sol-
utions, following which they were removed. inspected.

cleaned with hydrochloric acid and weighed.

Bods one through six showed a strong tendency to
corrode. It should be noted that none of the solutions
in these six bottles showed a pH higher than 8.1 after
twenty four hours of exposure. This pH is sligitly above
that of llichigan State College tap water. but not enough
so to have much effect on the rate of corrosion. It is
noted that the rate of corrosion for these waters with pH
up to 8.1 was only slightly less than in bottle one, which
was not treated at all. Beginning with red seven, which
was exposed to water of pH 8.5 to 9.2. a definite decrease
in the rate of corrosion was noted. men pH reached the
value of 10.6 to 10.8 (for red ten) a very sharp decrease
in the rate of corrosion was noted. Indeed it can be
seen that a pH above 10.0 is very effective in reducing
corrosion. while values below this have only a tendency

to reduce the rate of corrosion.

As in previous tests. inspection of test bottles

and of rods after testing showed results much in line

53

with results obtained by studying loss of weight of

the steel rods. Bottles one through six contained
typical red water after 24 hours of exposure. a1thou@
the color was lighter in bottles with higher pHs. Rods
one through six showed considerable corrosion. Bottles
nine through twelve, all with pH higher than 10.2 showed
no discoloration of the water, but rods were slightly
corroded. Bottles eleven and twelve with pH higher than
10.8 were still very clear. and these rods were still of

a bright gray color after twenty four hours of testing.

It should be observed that this short test does
not truly measure the value of high pHs as a means of
combating corrosion, since one of the principal values
of high p88 are their tendency to cause a protect ivo
coating of calcium carbonate to be laid down. This is
of course a long time process. and no such coating
could be laid down during this short test, and under
the methods of testing employed.

i .hiupuli.

. F...'v

due to corrosion

Grams loss in weight per rod',

TEST NUMBER FOUR

Curve indicating loss of weight of steel bars
exposed for 24 hours to Michigan State College
tap water treated to 23 parts per mill ion
hardness with zeolite sand, and adjusted to
various pHs by the addition of sodium hydroxide

..0200

'e 01 80 x\ X
'e 01 60 ‘ \»-\

X
..0140 ‘\\\

“0120 x \\

-0100 ‘\\\X

..0080 \\\
..0060 \\\\K
..0040

..0020

.0 7.4 8.0 8.6 .9.2 9.8 10.4

pH of water

55

11.0

TABULATION 0F RESULT FR0u.TEST #4. MICHIGAN
STATE. COLLEGE TAP WATER, TREATED To 26 PARTS
PER MILLION HARDNESS WITH SODIUM ZEOLITE, AND
70TH VARYING pHS

Bottle Mean pH Dissolved Rod's loss
Oxygen in weight
remove d ( grams )

(ppm)

1 7.4 5.6 .0189
2 8.0 7.6 .0150
5 8.3 4.8 .0176
4 8.6 4.8 .0122
5 8.8 5.2 .0188
6 8.8 5.2 .0165
7 9.0 4.2 .0154
8 9.5 3.6 .0136
9 9.8 5.0 .0135
10 10.1 2.4 .0095
11 10.8 o .0053
12 10.8 o .0053
Remarks:

7.4 represents tap pH of Michigan State College water.
pH of bottles 2 through 12 raised by adding sodium
hydroxide solution Na(0H)

56

 

Discussion of Test Number Four

Test number four was performed to observe the rela-
tion of pH to the rate of corrosion of steel bars exposed

to zeolite softened waters.

This test was performed in.much the same manner as
test three, that is. waters of various ple were:made
up by adding sodium hydroxide solution to water softened
to twenty five parts per million hardness with zeolite
sand. Bottle one was treated for hardness only; that is
no sodium hydroxide was added. and this bottle was used
as a control. Note: Michigan state College water nor-
mally'has a.hardness of about three hundred fifty'parts

per million.

A careful note of dissolved oxygen in each bottle
before and after the usual 24 hours test was made, and
these values are shown in tabulating the results for

this test.

Although a general relationship between oxygen
removed and.loss in weight of the bars can be noted.
no definite relationship between these two quantities
can be determined. It is of'interest that no change

in the amount of dissolved oxygen in bottles eleven and

twelve was observed during the period of testing, and
loss of weight for rods exposed to waters in these two
bottles is small when compared to that of other rods

used in the test.

curves obtained from this test compare very much
with those obtained from a similar test with nichigan
State College tap water from which no hardness had

been remove d.

No considerable change in loss of weight is to be
seen until pH becomes 10 or higher. At this point a
very radical change in the rate of corrosion is obser-
ved. Waters with pH of 10 or higher were unstained
after 24 hours of testing. while waters of pH. below
this value were stained to varying degrees. with waters

below pH 9.6 showing a typical 'red water" color.

All waters treated with sodium hydroxide were
observed to have covered the rods with a slimy coating.
which no doubt provided considerable protection against
the corrosive effect of the dissolved oxygen in the
water. Rods nine and eleven, pH 9.8 and 10.1 were
noted to have been badly corroded in a few spots only,
these particular points no doubt being points where a

slight varieties of purity of the metal secured.

58

TAEULATION OF RESULTS FROM TEST #5. MICHIGAN
STATE COLLEGE TAP WATER, TREATED TO 23 PARTS

PER MILLION HARDNESS WITH SODIUM ZEOLITE, AND
WITH VARIOUS AMOUNTS OF CALGON SOLUTION ADDED

Bottle Calgon Rod's loss
added in weight
(Pp!!!) (grams)

1 0 .0169

2 2 .0155

3 4 .0165

4 6 .0167

5 8 .0154

6 10 .0088

7 12 .0131

8 14 .0056

9 16 .0168

10 10 .0138

11 20 .0155

12 22 .0156

Remarks:

Steel rods used were the same as those used in previous
four tests. { inch in diameter, and 6 inches long.
flight of all bars is about 35 grams. Calgon solution

used was two weeks old.

59

Discussion of Test Number Five

Test number five was performed to observe the
effect of various additions of Calgon to zeolite softened

water on the rate of corrosion of steel bars.

Results of the test would seem to indicate that
Calgon does not inhibit corrosion of the bars in the
zeolite softened water used. No appreciable difference
between losses of weight with and without additions of

Calgon is to be noted.

Other investigators, particularly G. B. Hatch and
Owen rice, both of Hall Labratories. Pittsburgh. Penn-
sylvania, have observed that glassy phosphates seem to
have little value when used in zeolite softened.waters
unless some calcium is present. Zeolite softening changes
calcium salts to sodium.salts. EXperiments seem to
indicate that calcium.must be present with metaphosphates
in a ratio of at least 0.2 parts calcium to 1.0 part of
metaphosphate in order to obtain satisfactory results

when glassy phosphates are used to inhibit corrosion.

A further test, test ten, will repeat test five

using a water of very high oxygen content to note if

results similar to those from.this test are obtained.

due to corrosion

Grams lose in weight per rod'

-..0120 x \

TEST NUMBER SIX

Curves indicating loss of weight of steel bars
exposed to Michigan State College Tap water and
to Michigan state College water softened to 40
parts per million hardness with a zeolite water
softener. Various amounts of Calgon added to

each water. All rods polished.with steel wool
before test.

'°°14° ——~—— Softened

‘ _,_ _ te
~ah\\\g* Tap we r

..0100 ‘\\\\\
x

. \\
\ x
X X\
,,0080 ‘\‘\\\\
+ \\\\
-.0000 \>
..0020
.o o 4 8 12 16 so

Parts Per Million Calgon Added

61

TABULATION OF RESULTS FROM TEST # 6. MICHIGAN
STATE COLLEGE TAP EATER COMPAREIDFOR CORROSIVENESS
WITH MICHIGAN STATE meER SOFTENEI)TO 4O PARTS
PER MILLION HARDNESS WITH A ZEOL ITE SOFTENER.
DISSOLVED OXYGEN VARIED, RODS POLISHED BEFORE TEST

Bottle Calgon Rod's loss Remarks
added in wei ht
(PPm) (era-ms?
1 0 .0122 softened
2 4 .0092 Softened
3 8 .0115 Softened
4 12, .0116 Softened
6 16 .0096 Softened
6 20 .0074 Softened
7 0 .0074 Tap
8 4 .0079 Tap
9 8 .0074 Tap
10 12 .0089 Tap
11 16 .0026 Tap
12 20 .0018 Tap
Remarks:

Dissolved.oxygen of tap water 3.2. of softened water 6.0

62

Discussion Of Test Number Six

Because of poor results obtained with the use of
sodium hexametaphosphate in reducing the corrosiveness
Of zeolite treated water in the previous test (number
five) the test was rerun with certain alterations which
permitted results with zeolite water to be compared

with those obtained with tap water.

Rods were first polished with steel wool in order
to remove any incrustation or other protective film
which remained from previous tests, and then weighed to
four places as usual. Six bottles were then filled with
tap water with a dissolved oxygen content of 3.2 parts
per million and six were filled with zeolite softened
water (40 parts per million hardness remaining) with
a dissolved oxygen content of 6.0 Addition of Calgon
solution were then made to give concentrations of
4 to 20 parts per million for both softened and tap

water.

Bocause of the differences of dissolved oxygen content
a comparison cannot be made between corrosion of bars in
softened and in tap water. Rather, the test indicates
that Calgon is effective in inhibiting corrosion of pol-

ished steel bars in both zeolite softened water find in

63

tap water.

Poliahing the bars removed a rather heavy outside
film.0r skin which in the previous test (number five) was
no doubt quite effective in inhibiting corrosion of all
bars and no doubt, in part at least, explains the unsatis-

factory results Obtained in that test.

Results from test six, while showing that polished
steel bars are protected to a considerable extent by
glassy phosphate, does raise the question as to whether
the chemical is so successful in inhibiting corrosion
of ugpolished steel bars. this question is studied in

test number nine.

As usual, bottles containing stronger concentrations
of glassy phosphate showed no ”red water" while bottles
with weak concentrations of phosphate or no phosphate

at all were badly stained with iron rust.

64

due to corrosion

Grams lose in weight per rod'z

..0080

..0070

..0060

..0050

~.OO4O

- .0030

TEST NUMBER SEVEN

Curve indicating loss of weight of steel bars
exposed for 24 hours to Michigan State College
tap water with a dissolved oxygen content of
about 6.2 and with various amounts or calgon
solution added

..0110
wOlOO X

..0090

.0020

. .0010

. O O ,4 e is 16 20

Parts per million Of Calgon

65

 

24

TAHULATION OF RESULTS FROM TEST #7. MICHIGAN
STATE COLLEGE WATER WITH VARIOUS ADDITIONS
or CALGON SOLUTION AND LOW OXYGEN CONTENT

Bottle Calgon Rod's loss

added in wei t

(PM) (gram
1 0 .0101
2 2 .0098
3 4 .0108
4 6 .0081
5 8 .0108
6 10 .0086
7 12 .006?
8 14 .0060
9 16 .0048
10 18 .0040
11 20 .0026
12 22 .0034
Emmarks:

All bottles contained waters with dissolved oxygen con-
tent of about 6.2 at beginning of test. This compares
with dissolved oxygen contents of 7.6 to 8.2.for other

tests with calgon.

Discussion of Test Number Seven

Test number seven varied from previous tests with
Calgon and.Michigan State College tap water in that
larger amounts of Calgon were applied to the water than
previously, the test was begun with a lower dissolved
oxygen content 1.6.2 as compared with about 8.0) and the
stirring apparatus was run for 16 hours, then stepped
for 8 hours during the 24 hour test. In other respects
the standard prodedure outlined in previous tests was

followed.

A plotting of loss of weight against amount of
Calgon used shows a very considerable reduction in rate
of corrosion with increasing amounts of Calgon. A com-
parison of bottles one through twelve after the test
showed a typical “red water" in bottles one and two.
with an increasing clearness and decreasing amount of
red color as the amount of Calgon added increased.
Bottles eleven and twelve, with 20 and 22 parts per
million Calgon reSpectively. were almost completely

clear after 24 hours of testing.

No doubt the erratic but steady decrease in rate
of corrosion is explained in the varying purity of the

metal of the bars, which in some cases overcomes the

67

protective effect of the Calgon. Bars with slight

impurities corrode much rapidly than the pure metal.

This test indicates that increased additions of
Calgon doubtless leads to greater corrosion control.
but amounts above 22 parts per million are not practical

for water systems. and therefore are not of great inter-

set to this particular study.

68

due to corrosion

Grams loss in weight per rod'

TE

T NUMBJR EIGHT

Curve indicating loss of weight of steel bars
exposed for six days to Michigan state College
tap water Changed daily. and with various
amounts of calgon solution added. Dissolved
oxygen content average of about 7.0

e08

-.O7

..O6

..04

..03

.02

s 12 1e 20 ,

Parts Per Million of Calgon

69

24

TABULATION OF RESULTS FROM TEST # 8. MICHIGAN
STATE COLLEGE TAP WATER WITH VARIOUS AIEETIONS

0F CALGON SOLUTION, 7 DAY TEST WITH WATER CHANGED
EVERY 24 HOURS, NEW CALGON SOLUTION ADDED

Bottle Calgon Rod’s less

added in wei t

(PPR) (Creme
1 0 .0696
2 2 . 07 24
3 4 .0741
4 6 .0718
5 8 .0733
6 10 .0642
7 12 .0724
8 14 4 .0748
9 16 .0694
10 18 .0660
11 20 .0648
12 22 .0667
Remarks:

Tap water in bottles was changed every 24 hours,Ca1gon
added each time water was changed. Dissolved oxygen ~

content of water added varied from 6.8 to 7.2.

70

Discussion of Test Number Eight

Tests one through seven have consisted of adding
Calgon solution or other chemicals to water, then insert-
ing weighed steel rods and stirring rods for 24 or 48
hour periods in these solutions in such a manner as to
exclude oxygen other than that dissolved in the water
at the beginning of the test. These tests. while having
great value in the study of corrosion under the various
conditions involved do not approximate the exposure
of water pipes to the corrosive effect of water. Water
pipes are almost always exposed to a continuous supply
of fresh water which usually caries a high amount of

dissolved oxygen.

Test number eight was performed to observe the
results which might occur in steel pipe which is used
in large water lines. The steel rods were weighed as
usual, then placed in the bottles containing various
amounts of Calgon solution. At the end of 24 hours,the
water was changed. and new additions of Calgon solution
were made in the same strength as previously. After a
second 24 hour period the water was changed as before,
new additions of Calgon solution were made, and the rods
were stirred in the stirring device for three 24 hour
periods, with the water being changed story 24 hours.

fresh amounts of Calgon being added each time. Finally

71

the water was changed. Calgon solution added, and the
rods allowed to sit for 24 hours in the bottles without
stirring, Rods were then weighed as usual and loss of
weight of rods computed. The entire test took about one
weeks time. Dissolved oxygen of the tap water used was

about 7.0

Results of the test indicate than an exposure of
this type causes a continued corrosion of the bars during
the entire time of exposure. A loss in weight of .07
grams average, as compared with a loss of weight of .02
grams or less in 24 hours shows that as long as a fresh
supply of dissolved oxygen is present, corrosion will
continue, although at a slower rate than during the first

24 hour period.

Observation of the bars demonstrated that corrosion
occurs very slowly in quiet water. but very rapidly when
the water is in motion. The bars were exposed to the
water in the bottles for the first 48 hours without
stirring, and during this period almost no rusting of
the bars was observed. However, as soon as the stirring
was begun a very rapid mating of all the bars was noted.
This was because mot ion of the bars removed the protective
coating of hydrOgen gas which had formed on the metal, per-

mitting corrosion to progress at a very rapid rate.

72

Very slight protection was provided the bars by
the sodium.hexametaphosphate in this test. Bars in
solutions with 20 and 22 parts per million of Calgon
showed a slightly lower loss of weight than any others.,
but the difference is slight. This may be because
the seven day period was not a long enough time to permit
a heavy film of protective phOSphate to form on the metal.
The test shows that a wide variation in results is
often obtained with phosphates, and that differences in
these results are hard to explain. In this case no
adequate explainatien of the failure of the phosphate
to provide protection can be offered. since previous
tests have indicated.that glassy phosphate usually

gives considerable protection.

Observation of color of the water in bottles con-
sistently'ehowed.that bottles having 6 parts per million
or more of Calgon were clear, and contained no "red water?
while water with less Calgon were badly stained.

Thus. though there may be some question as to the ability
of Calgon to provide protection in all cases, there is
little question.that the chemical has great value in
combating 'red.water" by holding up iron ion, and
preventing oxidation of ferrous oxide to the hydroxide,

which is the objectionable red color of rusty water.

73

list. a . thelrulciw

 

 

Grams loss in weight per ro’d due to corrosion

TEST NUMBER NINE

Curves indicating loss of weight of steel bars
exposed to Michigan State college Tap water and
to Michigan State College water softened to 40
parts per million hardness with a zeolite water
softener. Various amounts of Calgon added to
eaCh water

 

-.0080 ————- Tap water
softened
x
V“""’ 1*
.0070 k‘\.
x \

.0050 * \

..OO4O

..0030

-.0020

-.0010

-0 OPPm 4PPm 8ppm 12mm léppm 20:91)!!!
Parts Per Million Calgon Added

74

TAEULATION OF RESULTS PROM TEST # 9. MICHIGAN
STATE COLLEGE TAP WATER.COMPARED EOR CORROSIVE
EFFECT WTTH‘MICHIGAN STATE COLLEGE WATER SOETEEED
TO 40 PARTS PER MILLION HARDNESS WITH A ZEOLITE
WATER SOFTENER. SAME AMOUNTS OF CALGON SOLUTION

ADE?) TO BOTH WATERS

Bottle Calgon Rod‘s loss Remarks
added in weight
(PPM) (grams)
1 0 .0072 Tap
2 4 .0067 Tap
3 8 .0066 Tap
4 12 .0051 Tap
5 16 .0058 Tap
6 20 .0044 Tap
7 0 .0073 Softened
8 4 .0073 Softened
9 8 .0076 softened
10 12 .0066 Softened
ll 16 .0049 Softened
12 20 .0046 softened
Remarks:

Dissolved oxygen at start of test was 3.6 in tap water

and 3.0 in softened water. Test was run for 24 hours

75

Discussion of Test Number Nine

In test number six a comparison was made of loss
of weight of polished steel bars exposed to tap water
and to zeolite softened water. In that particular test
rods were polished with steel wool to remove film which
might provide protection during time of exposure. Because
of the good results obtained with glassy phosphates in
that test it was felt that a similar test with unpolished

bars would be of interest.

In test number nine bars which had been exposed
for a period of one weeks time in test eight were washed
with dilute hydrochloric acid as usual and weighed.
Inspect ion of the bars indicated that a film of rather

dark color remained on the bars after the acid wash.

Test was run as usual with tap water and with zeol-
ite water softened to 40 parts per mill ion hardness.
Calgon was added in concentrations of 4 to 20 parts per
million to both tap and softened water. Dissolved oxygen
was 3.6 at the beginning of the test for tap water, as

compared with 3.0 for the softened water.

A study of loss of weight of the bes shows that a

slightly higher loss of weight was experienced by the

76

bars exposed to 2801 its softened water than to those
exposed to tap water, even though the dissolved oxygen
content of the tap water was slightly higher than that
of the softened water. In addition, bars exposed to
softened water were those removed from bottles with
high concentrations of glassy phoSphate in the previous
test, giving these bars whatever protect ion might remain
in the way of a phOSphate-iron film or skin deveIOped

in the seven day exposure of test number eight.

A considerable reduction in rate of corrosion was
noted in bars exposed to waters carrying higher concen-
trations of Calgon, both in the case of tap and softened
water. The shape of "loss of weight curves“ is generally

the same for both waters.

This test would seem to indicate that softened waters
are slightly more corrosive than unsoftened waters. and
that glassy phosphate is helpful in inhibiting corrosion
in both type waters. Results, as usual, were very erratic.
Curves shown with this test indicate a 2931 only toward
a lower rate of corrosion when larger additions of Calgon
are made. Several points on the curves vary considerably
from this general trend. and no emlainat ion can be

offered for these variations.

7'7

TABULATION OF RESULTS FROM TEST # 10. MICHIGAN
STATE COLLEGE TAP WATER SOFTEZTED T0 ZERO HARNESS
WITH SODIUM ZEOLITE, AND WITH VARIOUS AMOUNTS 0F

CALGON SOLUTION ADDED.

Bottle Calgon Loss in

Added. Weight

(PP!) (grams)
1 0 .0227
2 2 .0230
3 4 .0230
4 6 .0247
5 8 .0248
6 10 .0217
7 12 .0260
8 14 .0230
9 16 .0243
10 18 ..0249
11 20 .0242
12 22 .0201
Remarks:

water softened to zero hardness, and aorated.to

dissolved oxygen content of 9.6.

78

Discussion of Test Number Ten

Test ten is a repeat of test five in which the effect
of various additions of Calgon to zeolite softened water
was studied with reapect to the rate of corrosion of
steel bars. Test ten is altered slightly however, in
that a water of zero hardens was used instead of one
of 23 parts per million.hardness. and in that the water
used was aerated to obtain a very high oxygen content.
later used in test ten had a dissolved oxygen content of

9.6 parts per'million.

Because of the high oxygen content a very rapid
rate of corrosion was obtained. weight losses being
higher than in any previous tests. However no appre-
ciable difference in weight losses was observed between
bars subjected to water with and water without additions
of Calgon. Untreated water caused a weight loss of
.0227 grams in bar one, while water treated with 20
parts per million Calgon caused a weight loss of .0201
grams in bar twelve. The slight difference between these
weight losses is so small that we must conclude that
phosphate glasses have slightly inhibited corrosion in
a concentration of 20 parts per million, and are of

almost no value at all in lesser strengths.

79

The value of 23 parts per million.hardness of test
five was obtained by the use of standard soap solution.
A recheck of the soap solution indicates that this value
is high, and.that the true value is nearer zero. This
would show a very close comparison between tests five
and ten, both tests showing Calgon to be of little or no
value for inhibiting corrosion in zeolite softened waters
with very low hardness. However, test six, with a zeolite
softened water of 40 parts per million hardness (probably
a nearly true value) showed.that 20 parts per million
Calgon inhibited corrosion to about 55% of the untreated

value,

These three tests would indicate that phosphate
glasses are of‘value in inhibiting corrosion ihen enough
calcium is present to give the water a slight hardness.
but are of no value in zeolite softened waters of zero

hardness or thereabouts.

80

 

a . .

y. v.1: ass.
.s‘. .F

Illsll\

Grams loss in weight per ro'd due to corrosion

.0150
.0140

.0130
.0120
.0110
.0100
.0090
.0080
.0070
.0060
.0050
.0040

.0030

.0020

.0010

TEST NUMBER ELEVEN

Curves indicating loss in weight of steel bars

 

xwith various additions of Calgon (in pm)

002 added to tap water

\ .___.___ N0 CO2 .dded

.0000 .0 ppm 4 ppm 3 ppm 12 ppm
Parts Per Million Calgon added

81

1.6 ppm 2.0 moIII

TABULATION OF RESULTS FROM TEST # 11. MICHIGAN
STATE COLLEGE TAP EMTER.IHTH HIGH OXYGEN CONTENT.
002 CONTENT OF SIX BOTTLES RAISEIJABOVE NORMAL
VALUE FOR.PURPOSE OF COMPARISON. VARIOUS AMOUNTS
OF CALGON SOLUTION ADDED TO BOTTLES.

Bottle Calgon Rod‘s loss Remarks

added in wei ht

(ppm) (grams?
1 0 .0142 Aerated only
2 4 .0141 Aerated only
3 8 .0144 Aerated only
4 12 .0101 Aerated only
5 16 .0128 Aerated only
6 20 .0125 Aerated only
7 0 .0189 Aerated and 002 added
8 4 .0146 Aerated and 002 added
9 8 .0127 Aerated and 002 added
10 12 .0148 Aerated and 002 added
11 16 .0096 Aerated and 002 added
12 20 .0108 Aerated and 002 added
Remrks:

All bottles contained water with oxygen content of 9.2
No 00g added to bottles 1 through 6, 002 content of bottles

7 through 12 increased from normal 26‘ ppm to 3'? ppm

82

Discussion of Test Number Eleven

Test number eleven is a study to determine the
effect of increased 002 content upon the aggressiveness

of Michigan state College tap water.

Tap water was aerated to raise the oxygen content
to a high figure so that the water would be aggressive.
Then, to half the bottles CO2 gas was added from.a gen-
erating device so that the normal 002 content of 26 parts
per'million of Michigan State College tap water was

raised to 37 parts per'million.

Various quantities of Calgon were added to both
series of six test bottles. concentrations ranging from
zero to 20 parts per million in each series. Test rods

were then inserted and rotated fer a period of 24 hours.

A study of results indicates that the untreated
water of bottles seven through twelve with higher 002
content was.more aggressive than the untreated water
e£>bcttles one through six which contained a normal
00% content. Calgon seemed to have a greater inhibatory
effect upon the water with higher 002 content however,
so that only a slight variation between rates of corrosion

is noted in bottles with a Calgon concentration of 8 or

more parts per’million.

Results of the test indicate that 002 increases
the rate of corrosion of untreated water. and that
water lath high dissolved oxygen content is by nature
so aggressive that even the addition of considerable
quantities of Calgon had only slight inhibitory effect.
Also worthy of note is a comparison between the results
of tests ten and eleven, ihich are in many ways quite
comparable with the exception that test ten used softened
water while test eleven used tap water. A.study of
results indicates that the softened water caused far
greater losses of weight,and therefore must be considered

considerably more aggressive.

84

TESTS TO STUDY THE EFFECT OF HEWTAPHOSPHATE
on THE sorrmmrc AND IRON REMOVING CAPACITY or
,A znours son-rm

 

Mom. ZEOLITE SOFTENER USED IN TESTS

Because of the ability of glassy phospuates to hold
Calcium and iron ions in solution some question.has been
raised as to the effect which adding the chemical to
water supplies might have on the ability of zeolite

water softeners to remove iron and soften water.

Aastudy of the question.seemed to be of value,
since the use of phosphates would be most undesirable
in localities ihere zeolite water softeners are used
in homes or business places if their use in any inter-

feres with the preper eperatien ef the sefteners.

A.small water softener was constructed by filling
a glass bottle (size 8 inches by 12 inches) with about
3 inches of { tot inch gravel and six.inches of zeolite
sand. The volume of zeolite sand in the softener was
about 0.2 cubic feet. A siphon arrangement for feeding
water to the softener was‘used so that a constant rate
of flow'might be obtained. Since water pressure varied
greatly in the tap line over each 24 hours period an
overflow was added so that water in the siphon feed
bottle would be of constant elevation. This arrangement
permitted computation of the amount of water passed
through the softener and the number of grains of hardness

removed between rechargings.

85

Two series of tests were run in this study. The
first series was run over a period of about one and a
half months to determine whether the addition of various
quantities of Calgon (1 part per million to 50 parts per
million) would have any apparent effect upon the operation
of the zeolite softener. Two methods of adding Calgon
were used in this test. A very low quantity of Calgon
was added by the addition of 100 grams of Micgomet ( a
soluble glassy phosphate which dissolves at the rate of
0.8% per day) to the feed bottle so that about 1 gram
would go into solution per day, giving a concentration of
about 1 part per million. Higher concentrations were
obtained by permitting a 1 to 100 solution of Calgon to
drip into the siphon bottle at a rate which added the
desired concentration (10 to 80 parts per million) of
the chemical to water passing through the softener.

In this first series of tests no attempt was made to
determine the iron removal in the softener on a quant i-
tat ive basis. Azrecord was kept of time which the soft-
cner was capable of operating with an effluent of low
hardness. both with and without the addition of Calgon.
The study showed that no apparent difference can be noted
between results with treated and untreated water in so
far as softening capacity is concerned. the softener

being capable of softening water for a period of about

86

2} hours without recharging, the hardness of the water
at the end of this time being about 20 parts per million.
liter 2% hours time the hardness of the water begins to
rise quite rapidly. Results seem to indicate that the
addition of glassy phosphates does not interfere with the

softening action of a zeolite softener.

Removal of iron by the softener was studied in these
first tests by permitting the softener to run for a
period of about one week without recharging, and then
reversing flow of the water to wash. Michigan State
college tap water contains between 0.1 and 0.2 parts
per million of iron. The study was made with untreated
water, and with waters containing various amounts of
glassy phosphate ranging from 1 part per million to 50
parts per million. No apparent difference in results
was noted. That is. one weeks operation without washing
resulted in the formation of a scum of red iron rust
on the surface of the softener. and several minutes
washing was required to remove the rust. This scum of
iron rust formed with treated and untreated water,regard-

less of concentration of Calgon used.

These first tests indicated that glassy phosphates
do not seem to have any effect on the softening and iron
removing capacity of a zeolite softener. It was then

felt that a reporting of the actual grains per gallon of

87

hardness removed by the softener with treated and un-
treated water. and a series of quantitative tests for
iron in the influent and effluent of the softener would

be of interest.

For the second series of tests all old zeolite sand
was removed and new sand added. (Synthetic sand was
used in all tests) The softener was then run for several
days, no record being kept, and recharging being done

whenever the water became quite hard. A test was then

 

run to determine the number of grains of hardness re-
moved during one cycle of operation after recharging.

Results are as- followszf

Time in Effluent Hardness
Operat io n Hardne s s Remove 6.
(ppl) (pp!)
0 hr 10 min 2 348
0 hr 50 min ' e 344
1 hr 30 min 12 338
2 hr 5 min 1.8 332
2 hr 50 min 24 a 326
3 hr 10 min 69 281
4 hr 30 min 129 221
5 hr 0 min 144 206

Note: Michigan State tap water has about 350 parts
per million hardness (as 08.003)

88

Next, 100 grams of Micromet was added to the siphon
feed bottle, which dissolved at a rate such as to provide
one part per million of phosphate for the water passing

through the softener. Results are as follows:

Time in Effluent Hardness
0pc rat is n Har dne s s Remove d
(ppm) (ppm)
0 hr 10 min 4 345
1 hr 0 min 10 340
1 hr 50 min 14 536
2 hr 30 min 25 330
2 hr 50 min 25 325
3 hr 10 min 30 320
4 hr 30 min 120 230

Finally,20 grams of Calgon was dissolved in 2 liters
of water. making a solution of 1 part per 100 Calgon.
This solution was fed into the siphon feed bottle with
a glass tube at the rate of 30 drops per minute.providing
a concentration of about 50 parts per million to the

water passing through the softener. Results were as

fell on:
T ime in Effluent Ear dne s s
Operat io n Hardno s s Remove d

(ppm) (pm)
0 hr 10 min 5 245

I)
I)

T ime in Effluent Hardness

Operat ion Hardne s s Remove d
(ppm) (ppm)
0 hr 30 min 3 847
1 hr 5 min 8 342
1 hr 30 min 10 340
1 hr 55 min 15 335
2 hr 30 min 20 330
3 hr 30 min 80 270
4 hr 25 min 160 190

On the following sheet a plotting of the results
from these three tests shows a very close comparison,
indicating clearly that Calgon has no effect upon the
softening capacity of a zeolite softener when used in
such quantities as might be added to a public water
system. Since the area under the curves represents
total number of grains of hardness removed, the capacity
of the softener is about 850 grains(or slightly more
if the softener is permitted to become completely dis-
charged). Sinoo the volume is about 0.2 cubic feet of
sand. the sand has a capacity of about 4200 grains per
cubic feet between rechargings,which is about normal
for synthetic zeolite sands.

A quantitative test for iron was run for all of the

above tests. Results indicated that all of the 0.2 parts

90

 

14.134

CURVES SHOWING A PLOTT ING 0F PARTS PER MILLION HARDN'ESS
REMOVED BY ZEOLITE sommn AGAINST TIME OF OPERATION

Micromet treated
50 ppm Calgon added
M.S.C. tap water

 

20.0 pp!

1 50 ppm

 

100 ppm

IO ppm

0.ppm 0 hr 1_ hr 2hr 8hr Ihr 5hr

Remark":
Curves show results obtained with tap water. water treated
with about one part per million of Micromet and water

treated to about fifty parts per million with Calgon

91

 

per million of iron in Michigan State College tap water
was removed by the zeolite softener. That is, an effluent
of zero iron content was obtained with raw water and also
with all water treated with Calgon. Samples of waters
treated with Calgon before softening were boiled to

remove any effect of the hexametaphosphate before iron
tests were made, and since a result of zero iron content
was obtained in all cases it is apparent that the icon
was removed by softener, and was not held in solution

by the hexamephosphate.
As already stated, these tests seem.to indicate that

glassy phosphate has no effect upon the softening or iron

removing capacity of a zeolite water softener.

92

 

'BIBLIOGRAPHY

l. The Value of Sodium Heggmetgphgsphate in the

0gptzgl 9f Diffggglties Dug 3g Cgrggsion in Watgr
Systems. -Committee Report. June 2h. l9h2, American

Water Works Association.

2. Recon} Develgpgepts in the Use 9f HezfiEetg-

phggphgtg in Watgr Treatmgnt - Owen Rice. Chemist for
Calgon. Inc. (A publication of Calgon Inc.)

 

3. Th; Phygiglggz pf Sodipg nggmgtaphgsphgtg
K.K. Jones. Assistant Professor of Physiology.

lorthwestern University Medical School. Journal of
American later Works Association. Vol. 32. NO. 9.

Sept.. l9u0.

h. The peg 9f Sgdigg Hggggetgphgsphate at Nitro.
lest Virtinia - P.L. McLaughlin. Sanitary Engineer.

Proceedings of the 15th Annual Conference on Water

Purification.

5. Private Communioatigp fggm Jgpg P, chbor.
Calgon. In.

6. ngrgsigp Cgpprpl with Ehreshgld Tgeggmgp .
G.B.Hatoh and Owen Rice. Industrial Engineering

Chemistry. Vol. 37. Page 752. August lghfi

7. c He hanism f C r s f W or P es.-
Thomas n. Riddick. Consulting Engineer. low York City.
Water Works and Sewerage. R - 149, Reference and Data.
19u7.

6. The Cgptrgl 9f Cgrrgsion. Robert S. Weston -
Water Works and Sewerage. September. 1943.

 

9. Scale ggd Cgrrooign antrol in Potable Water
Supplies.- R.T. Hanlon and A. J. Steffen together with

G.A. Rohlioh and L.H. Kessler Industrial and Engineer-

 

ing Chemistry. Vol. 37. Page 72%, August 19H5.

10. Nate; Tregtmgpt in Etgbicgke.- Franklin
McArthur. Township Engineer. later and Sewage. April. 1946.

ll. Causes gpd Effects of lgtor Pipe Corrgoion.

A paper by James A. Parks. Senior Chemist. Detroit. Michigan.

12. Scale gag nggggggp Preventigp by Chemiogl

Cgpdgtigping gf Wateg. R.T.Hanlon and D. Newton. Public
Works. April. 19U6.

13. lgtg; Supply Epgineering. Babitt and Doland.
Third Edition. Pages 332-335.

94

 

v...—

 

 

‘ ' . I . III I I. IIII 'Ii‘ II -I' ‘l I... 5"! I 'l‘l vl.l .Il.||| C H -1. . . I ~|| III I I.

, v '0.
I. Jet o ..I .
U s

.

 

 

 

 

 

 

 

mu“(I(Immunity)(WWW