To remove the coarse particles by sifting will reduce the specific gravity of a sample of Portland cement, as the un-ground particles are from the harder burned and denser portion of the clinker, and to remove these denser particles will, of course, decrease the average density of the sample. This is not always the case with natural cements, as is shown by the following tests: —
As received . . .
Pass 50 ....
Ret. on 50 . . .
Pass 100 ... .
Ret. on 50 ...
Pass 50 ... .
Ret. on 50 . . .
The apparent density or weight per cubic foot of Portland will be reduced more than the specific gravity by the removal of the coarse particles; because not only will the true density be decreased, but the packing, which is facilitated by a wide range in the sizes of the particles, will be less perfect than when the coarse particles are present. In § 89 a table is given show-ing the changes in specific gravity and weight per bushel occasioned by removing the coarse particles by sifting.
Table 11 gives the results of a number of tests on Portland and natural cements to determine the relative time of setting of samples from which the coarse particles had been removed by the No. 200 sieve, while Table 12 gives results obtained with a sample of natural cement of varying fineness.
In Table 11, 30 per cent, of water was used for all Portland cements, and 36 per cent, for all naturals, but the consistency varied as stated in the table. It is seen that in nearly every case the setting was hastened by removing the coarse particles, though this may have been due in part to the fact that with the same percentage of water the finer cement gave a stiffer paste.
For the tests in Table 12, the attempt was made to make all of the mortars of the same consistency by varying the percentage of water. As would be expected, the coarse particles are very slow setting. In fact, what hardness they attained was probably due largely to the fine dust that adhered to the grains. These coarse particles may be considered as practically inert, and their presence in a sample would naturally make it slow setting. To show this by actual test, however, is very difficult, as the amount of water required to bring the mortars to the same consistency varies with the amount of coarse particles present, and as there is no very satisfactory method of testing the consistency, the tests for time of setting have in them this indetermination.
Cement Passing No. 20
Cement Passing No. 200
Time to bear
1/4 lb. wire.
Time to bear
1/4 lb. wire.
. . .
. . .
. . .
Note : — 30 per cent, water used for all Portlands.
36 per cent, water used for all natural cements.
water Used as Per Cent, of Cement.
Time to Bear 1/4 lb. Wire.
Time to Bear 1 lb. Wire.
Pass No. 20 sieve
" 50 "
" 100 "
Retained on 50,
reground to pass 100
Pass No. 50, retained
on No. 100
A cement having a certain quantity of coarse particles will frequently give a higher tensile strength when tested neat than a cement from which the coarse particles have been removed by screening. The reason for this may be found in the fact that a wide range in the sizes of grain of the powder facilitates packing, both when dry and when mixed with water to form a paste. Another reason is that the unground particles are stronger than the hardened mortar, and, considering the broken section of a briquet, the break does not take place through these particles, but they are pulled out of their bed; this virtually increases the area of section. Were the same sample of cement reground, so that a certain proportion of the coarse particles was rendered active, it might then give a higher strength, neat, than at first. If so, the reason would be found in the fact that the coarse particles, being the hardest burned, were really from the best part of the cement clinker, and rendering these particles active by fine grinding increased the cohesive properties of the cement so much as to overcome the physical effect of the coarse particles, which, when judged by neat tests, appear to be beneficial. The above serves to illustrate the difference between sifting and fine grinding which are so frequently confused in treating this subject.
83. Among the many tests that have been made to show the effect of sifting on the cohesive and adhesive strength of cements, a few may be given as follows: —
Mr. Maclay1 gives a few experiments to show that the presence of coarse particles increases the cohesive strength, neat, seven days.
Lieut. W. Innes2 gives two tables of results obtained by experimenting on very coarse cements. The tables show that removing the particles that would not pass through sieves of 1,296 meshes and 2,500 meshes per square inch, decreased the strength when tested neat at the ages of three months and six months; but increased the strength when sand mortars were used. The differences at six months were relatively somewhat less than at three months. By separating a sample of cement into two parts, that passing a sieve having 2,500 meshes per square inch and that retained on the same sieve, and then remixing the screenings with the fine portion, he found that the highest strength, neat, six months, was given by the mixture containing the largest amount tried (70 per cent.) of screenings.
1 Trans. Am. Soc. C. E., Vol. vi.
2 Minutes Proc. Inst. C. E., Vol. xxv.
84. In the tests of cement for the Cairo Bridge1 a series of experiments was made to determine the effect of coarse particles on the value of both Portland and natural cements. The cement was separated into two parts, by a sieve having 10,000 meshes per square inch. Briquets were made both neat and with sand, the cement used being made of 100, 90, 80, 70 and 60 volumes of sifted cement to 0, 10, 20, 30, and 40 volumes, respectively, of cement screenings. The briquets were broken when six months old.