THE ACTION OF SEA WATER ON CEMENT.
Questions are frequently raised in connection with sea-coast
works as to whether any deleterious effect will result from
using sea-water for mixing the concrete or from using sand and
shingle off the beach; and, further, whether the concrete,
after it is mixed, will withstand the action of the elements,
exposed, as it will be, to air and sea-water, rain, hot sun,
and frosts.
Some concrete structures have failed by decay of the material,
principally between high and low water mark, and in order to
ascertain the probable causes and to learn the precautions
which it is necessary to take, some elaborate experiments have
been carried out.
To appreciate the chemical actions which may occur, it will be
as well to examine analyses of sea-water and cement. The water
of the Irish Channel is composed of
Sodium chloride.................... 2.6439 per cent.
Magnesium chloride................. 0.3150 " "
Magnesium sulphate................. 0.2066 " "
Calcium sulphate................... 0.1331 " "
Potassium chloride................. 0.0746 " "
Magnesium bromide.................. 0.0070 " "
Calcium carbonate.................. 0.0047 " "
Iron carbonate..................... 0.0005 " "
Magnesium nitrate.................. 0.0002 " "
Lithium chloride................... Traces.
Ammonium chloride.................. Traces.
Silica chloride.................... Traces.
Water.............................. 96.6144
--------
100.0000
An average analysis of a Thames cement may be taken to be as
follows:--
Silica................................ 23.54 per cent.
Insoluble residue (sand, clay,
etc.)............................ 0.40 "
Alumina and ferric oxide............... 9.86 "
Lime.................................. 62.08 "
Magnesia............................... 1.20 "
Sulphuric anhydride.................... 1.08 "
Carbonic anhydride and water........... 1.34 "
Alkalies and loss on analysis.......... 0.50 "
-----
100.00
The following figures give the analysis of a sample of cement
expressed in terms of the complex compounds that are found:--
Sodium silicate (Na2SiO3)........ 3.43 per cent.
Calcium sulphate (CaSO4)......... 2.45 "
Dicalcium silicate (Ca2SiO4).... 61.89 "
Dicalcium aluminate (Ca2Al2O5).. 12.14 "
Dicalcium ferrate (Ca2Fe2O5)..... 4.35 "
Magnesium oxide (MgO)............ 0.97 "
Calcium oxide (CaO)............. 14.22 "
Loss on analysis, &c............. 0.55 "
-----
100.00
Dr. W. Michaelis, the German cement specialist, gave much
consideration to this matter in 1906, and formed the opinion
that the free lime in the Portland cement, or the lime freed in
hardening, combines with the sulphuric acid of the sea-water,
which causes the mortar or cement to expand, resulting in its
destruction. He proposed to neutralise this action by adding to
the mortar materials rich in silica, such as trass, which would
combine with the lime.
Mr. J. M. O'Hara, of the Southern Pacific Laboratory, San
Francisco, Cal., made a series of tests with sets of pats 4 in
diameter and 1/2 in thick at the centre, tapering to a thin
edge on the circumference, and also with briquettes for
ascertaining the tensile strength, all of which were placed
in water twenty-four hours after mixing. At first some of the
pats were immersed in a "five-strength solution" of sea-water
having a chemical analysis as follows:--
Sodium chloride.................... 11.5 per cent.
Magnesium chloride................. 1.4 " "
Magnesium sulphate................. 0.9 " "
Calcium sulphate................... 0.6 " "
Water.............................. 85.6 " "
100.0
This strong solution was employed in order that the probable
effect of immersing the cement in sea-water might be
ascertained very much quicker than could be done by observing
samples actually placed in ordinary sea-water, and it is worthy
of note that the various mixtures which failed in this
accelerated test also subsequently failed in ordinary sea-water
within a period of twelve months.
Strong solutions were next made of the individual salts
contained in sea-water, and pats were immersed as before, when
it was found that the magnesium sulphate present in the water
acted upon the calcium hydrate in the cement, forming calcium
sulphate, and leaving the magnesium hydrate free. The calcium
sulphate combines with the alumina of the cement, forming
calcium sulpho-aluminate, which causes swelling and cracking of
the concrete, and in cements containing a high proportion of
alumina, leads to total destruction of all cohesion. The
magnesium hydrate has a tendency to fill the pores of the
concrete so as to make it more impervious to the destructive
action of the sea-water, and disintegration may be retarded or
checked. A high proportion of magnesia has been found in
samples of cement which have failed under the action of sea
water, but the disastrous result cannot be attributed to this
substance having been in excess in the original cement, as it
was probably due to the deposition of the magnesia salts from
the sea-water; although, if magnesia were present in the cement
in large quantities, it would cause it to expand and crack,
still with the small proportion in which it occurs in ordinary
cements it is probably inert. The setting of cement under the
action of water always frees a portion of the lime which was
combined, but over twice as much is freed when the cement sets
in sea-water as in fresh water. The setting qualities of cement
are due to the iron and alumina combined with calcium, so that
for sea-coast work it is desirable for the alumina to be
replaced by iron as far as possible. The final hardening and
strength of cement is due in a great degree to the tri-calcium
silicate (3CaO, SiO2) which is soluble by the sodium chloride
found in sea-water, so that the resultant effect of the action
of these two compounds is to enable the sea-water to gradually
penetrate the mortar and rot the concrete. The concrete is
softened, when there is an abnormal amount of sulphuric acid
present, as a result of the reaction of the sulphuric acid of
the salt dissolved by the water upon a part of the lime in the
cement. The ferric oxide of the cement is unaffected by sea-
water.
The neat cement briquette tests showed that those immersed in
sea-water attained a high degree of strength at a much quicker
rate than those immersed in fresh water, but the 1 to 3 cement
and sand briquette tests gave an opposite result. At the end of
twelve months, however, practically all the cements set in
fresh water showed greater strength than those set in sea-
water. When briquettes which have been immersed in fresh water
and have thoroughly hardened are broken, the cores are found to
be quite dry, and if briquettes immersed in sea-water show a
similar dryness there need be no hesitation in using the
cement; but if, on the other hand, the briquette shows that the
sea-water has permeated to the interior, the cement will lose
strength by rotting until it has no cohesion at all. It must be
remembered that it is only necessary for the water to penetrate
to a depth of 1/2 in on each side of a briquette to render it
damp all through, whereas in practical work, if the water only
penetrated to the same depth, very little ill-effect would be
experienced, although by successive removals of a skin 1/2 in
deep the structure might in time be imperilled.
The average strength in pounds per square inch of six different
well-known brands of cement tested by Mr. O'Hara was as
follows:--
TABLE No. 16.
EFFECT OF SEA WATER ON STRENGTH OF CEMENT.
Neat cement 1 cement to 3 sand
set in set in
Sea Water Fresh Water Sea Water Fresh Water
7 days 682 548 214 224
28 days 836 643 293 319
2 months 913 668 313 359
3 months 861 667 301 387
6 months 634 654 309 428
9 months 542 687 317 417
12 months 372 706 325 432
Some tests were also made by Messrs. Westinghouse, Church,
Kerr, and Co., of New York, to ascertain the effect of sea-
water on the tensile strength of cement mortar. Three sets of
briquettes were made, having a minimum section of one square
inch. The first were mixed with fresh water and kept in fresh
water; the second were mixed with fresh water, but kept
immersed in pans containing salt water; while the third were
mixed with sea-water and kept in sea-water. In the experiments
the proportion of cement and sand varied from 1 to 1 to 1 to 6.
The results of the tests on the stronger mixtures are shown in
Fig. 32.
The Scandinavian Portland cement manufacturers have in hand
tests on cubes of cement mortar and cement concrete, which were
started in 1896, and are to extend over a period of twenty
years. A report upon the tests of the first ten years was
submitted at the end of 1909 to the International Association
of Testing Materials at Copenhagen, and particulars of them are
published in "Cement and Sea-Water," by A. Poulsen (chairman of
the committee), J. Jorsen and Co., Copenhagen, 1909, price 3s.
[Illustration: FIG. 32.--Tests of the Tensile Strength of
Cement and Sand Briquettes, Showing the Effect of Sea Water.]
Cements from representative firms in different countries were
obtained for use in making the blocks, which had coloured glass
beads and coloured crushed glass incorporated to facilitate
identification. Each block of concrete was provided with a
number plate and a lifting bolt, and was kept moist for one
month before being placed in position. The sand and gravel were
obtained from the beach on the west coast of Jutland. The
mortar blocks were mixed in the proportion of 1 to 1, 1 to 2,
and 1 to 3, and were placed in various positions, some between
high and low water, so as to be exposed twice in every twenty-
four hours, and others below low water, so as to be always
submerged. The blocks were also deposited under these
conditions in various localities, the mortar ones being placed
at Esbjerb at the south of Denmark, at Vardo in the Arctic
Ocean, and at Degerhamm on the Baltic, where the water is only
one-seventh as salt as the North Sea, while the concrete blocks
were built up in the form of a breakwater or groyne at Thyboron
on the west coast of Jutland. At intervals of three, six, and
twelve months, and two, four, six, ten, and twenty years, some
of the blocks have, or will be, taken up and subjected to
chemical tests, the material being also examined to ascertain
the effect of exposure upon them. The blocks tested at
intervals of less than one year after being placed in position
gave very variable results, and the tests were not of much
value.
The mortar blocks between high and low water mark of the Arctic
Ocean at Vardo suffered the worst, and only those made with the
strongest mixture of cement, 1 to 1, withstood the severe frost
experienced. The best results were obtained when the mortar was
made compact, as such a mixture only allowed diffusion to take
place so slowly that its effect was negligible; but when, on
the other hand, the mortar was loose, the salts rapidly
penetrated to the interior of the mass, where chemical changes
took place, and caused it to disintegrate. The concrete blocks
made with 1 to 3 mortar disintegrated in nearly every case,
while the stronger ones remained in fairly good condition. The
best results were given by concrete containing an excess of
very fine sand. Mixing very finely-ground silica, or trass,
with the cement proved an advantage where a weak mixture was
employed, but in the other cases no benefit was observed.
The Association of German Portland Cement Manufacturers carried
out a series of tests, extending over ten years, at their
testing station at Gross Lichterfeld, near Berlin, the results
of which were tabulated by Mr. C. Schneider and Professor Gary.
In these tests the mortar blocks were made 3 in cube and the
concrete blocks l2 in cube; they were deposited in two tanks,
one containing fresh water and the other sea-water, so that the
effect under both conditions might be noted. In addition,
concrete blocks were made, allowed to remain in moist sand for
three months, and were then placed in the form of a groyne in
the sea between high and low-water mark. Some of the blocks
were allowed to harden for twelve months in sand before being
placed, and these gave better results than the others. Two
brands of German Portland cement were used in these tests, one,
from which the best results were obtained, containing 65.9 per
cent. of lime, and the other 62.0 per cent. of lime, together
with a high percentage of alumina. In this case, also, the
addition of finely-ground silica, or trass, improved the
resisting power of blocks made with poor mortars, but did not
have any appreciable effect on the stronger mixtures.
Professor M. Möller, of Brunswick, Germany, reported to the
International Association for Testing Materials, at the
Copenhagen Congress previously referred to, the result of his
tests on a small hollow, trapezium shape, reinforced concrete
structure, which was erected in the North Sea, the interior
being filled with sandy mud, which would be easily removable by
flowing water. The sides were 7 cm. thick, formed of cement
concrete 1:2 1/2:2, moulded elsewhere, and placed in the
structure forty days after they were made, while the top and
bottom were 5 cm. thick, and consisted of concrete 1:3:3,
moulded _in situ_ and covered by the tide within twenty-four
hours of being laid. The concrete moulded _in situ_ hardened a
little at first, and then became soft when damp, and friable
when dry, and white efflorescence appeared on the surface. In a
short time the waves broke this concrete away, and exposed the
reinforcement, which rusted and disappeared, with the result
that in less than four years holes were made right through the
concrete. The sides, which were formed of slabs allowed to
harden before being placed in the structure, were unaffected
except for a slight roughening of the surface after being
exposed alternately to the sea and air for a period, of
thirteen years. Professor Möller referred also to several cases
which had come under his notice where cement mortar or concrete
became soft and showed white efflorescence when it had been
brought into contact with sea-water shortly after being made.
In experiments in Atlantic City samples of dry cement in powder
form were put with sea-water in a vessel which was rapidly
rotated for a short time, after which the cement and the sea-
water were analysed, and it was found that the sea-water had
taken up the lime from the cement, and the cement had absorbed
the magnesia salts from the sea-water.
Some tests were carried out in 1908-9 at the Navy Yard,
Charlestown, Mass., by the Aberthaw Construction Company of
Boston, in conjunction with the Navy Department. The cement
concrete was placed so that the lower portions of the surfaces
of the specimens were always below water, the upper portions
were always exposed to the air, and the middle portions were
alternately exposed to each. Although the specimens were
exposed to several months of winter frost as well as to the
heat of the summer, no change was visible in any part of the
concrete at the end of six months.
Mons. R. Feret, Chief of the Laboratory of Bridges and Roads,
Boulogne-sur-Mer, France, has given expression to the following
opinions:--
1. No cement or other hydraulic product has yet been found
which presents absolute security against the decomposing action
of sea-water.
2. The most injurious compound of sea-water is the acid of the
dissolved sulphates, sulphuric acid being the principal agent
in the decomposition of cement.
3. Portland cement for sea-water should be low in aluminium and
as low as possible in lime.
4. Puzzolanic material is a valuable addition to cement for
sea-water construction,
5. As little gypsum as possible should be added for regulating
the time of setting to cements which are to be used in sea-
water.
6. Sand containing a large proportion of fine grains must never
be used in concrete or mortar for sea-water construction.
7. The proportions of the cement and aggregate for sea-water
construction must be such as will produce a dense and
impervious concrete.
On the whole, sea-water has very little chemical effect on good
Portland cements, such as are now easily obtainable, and,
provided the proportion of aluminates is not too high, the
varying composition of the several well-known commercial
cements is of little moment. For this reason tests on blocks
immersed in still salt water are of very little use in
determining the probable behaviour of concrete when exposed to
damage by physical and mechanical means, such as occurs in
practical work.
The destruction of concrete works on the sea coast is due to
the alternate exposure to air and water, frost, and heat, and
takes the form of cracking or scaling, the latter being the
most usual when severe frosts are experienced. When concrete
blocks are employed in the construction of works, they should
be made as long as possible before they are required to be
built in the structure, and allowed to harden in moist sand,
or, if this is impracticable, the blocks should be kept in the
air and thoroughly wetted each day. On placing cement or
concrete blocks in sea water a white precipitate is formed on
their surfaces, which shows that there is some slight chemical
action, but if the mixture is dense this action is restricted
to the outside, and does not harm the block.
Cement mixed with sea water takes longer to harden than if
mixed with fresh water, the time varying in proportion to the
amount of salinity in the water. Sand and gravel from the
beach, even though dry, have their surfaces covered with saline
matters, which retard the setting of the cement, even when
fresh water is used, as they become mixed with such water, and
thus permeate the whole mass. If sea water and aggregate from
the shore are used, care must be taken to see that no decaying
seaweed or other organic matter is mixed with it, as every such
piece will cause a weak place in the concrete. If loam, clay,
or other earthy matters from the cliffs have fallen down on to
the beach, the shingle must be washed before it is used in
concrete.
Exposure to damp air, such as is unavoidable on the coast,
considerably retards the setting of cement, so that it is
desirable that it should not be further retarded by the
addition of gypsum, or calcium sulphate, especially if it is to
be used with sea water or sea-washed sand and gravel. The
percentage of gypsum found in cement is, however, generally
considerably below the maximum allowed by the British Standard
Specification, viz., 2 per cent., and is so small that, for
practical purposes, it makes very little difference in sea
coast work, although of course, within reasonable limits, the
quicker the cement sets the better. When cement is used to
joint stoneware pipe sewers near the coast, allowance must be
made for this retardation of the setting, and any internal
water tests which may be specified to be applied must not be
made until a longer period has elapsed after the laying of the
pipes than would otherwise be necessary. A high proportion of
aluminates tends to cause disintegration when exposed to sea
water. The most appreciable change which takes place in a good
sound cement after exposure to the sea is an increase in the
chlorides, while a slight increase in the magnesia and the
sulphates also takes place, so that the proportion of sulphates
and magnesia in the cement should be kept fairly low. Hydraulic
lime exposed to the sea rapidly loses the lime and takes up
magnesia and sulphates.
To summarise the information upon this point, it appears that
it is better to use fresh water for all purposes, but if, for
the sake of economy, saline matters are introduced into the
concrete, either by using sea water for mixing or by using sand
and shingle from the beach, the principal effect will be to
delay the time of setting to some extent, but the ultimate
strength of the concrete will probably not be seriously
affected. When the concrete is placed in position the portion
most liable to be destroyed is that between high and low water
mark, which is alternately exposed to the action of the sea and
the air, but if the concrete has a well-graded aggregate, is
densely mixed, and contains not more than two parts of sand to
one part of cement, no ill-effect need be anticipated.