Associate Professor M. I. Meshcheryakov

On the Methodology for Determining the Uniformity of Volume Changes (Constancy of Volume) of Portland Cement According to the All-Union Standard of 1927*)

The strength of artificial stone-like masses obtained by cementing loose ingredients (sand, gravel, crushed stone, stones) with the aid of some binding substance is based essentially on the strength of the cementing material used.

Portland cement, widely used in construction practice as a binding material, deserves special attention in this respect. This is explained by the fact that the transformations occurring in the loose components of a monolith, dependent on changes in temperature, humidity, and chemical processes, are extremely slow in them, whereas in Portland cement, which binds the ingredients of the monolith into a single whole, the processes of hydration and molecular movements occurring in it proceed immeasurably more intensively and for a long time, influencing the binding capacity of the cement not always in a positive sense, i.e., in the direction of increasing the strength of the whole.

The outward expression of the processes occurring in Portland cement is a change in the volume of the cement mass, which may be either uniformly general throughout the entire mass or local. The latter most often creates excessive local stresses exceeding the force of adhesion of the particles, and leads to a rupture of continuity.

An excessively large uniformly general change in volume, in the direction of increase, results in the loosening of the cement mass and likewise weakens the bonds that have formed.

It is natural, therefore, that in all countries the standards defining the technical conditions for accepting batches of Portland cement for work include a clause on the uniformity of volume change (“constancy of volume”), with an indication of those signs by which the soundness of the cement, its suitability and reliability as a binder for construction works, are established. It should be noted that the soundness of Portland cement with respect to volume change is determined by negative signs, by the absence of the deformations indicated by the standards, the presence of which testifies to non-uniformity of volume change.

The interest and attention that have been and are being devoted to questions of determining the uniformity of volume change of Portland cement at international congresses on the testing of construction materials, at congresses of engineers and manufacturers in the cement, concrete, and reinforced-concrete fields, are also understandable: a correct characterization of Portland cement in this respect guarantees the reliability and durability of the structure in the construction of which cement is assigned a very responsible role. In journal articles**), which have treated questions of pe—

*) All-Union Standard. Portland cement. “Technical Conditions” OST. 77. “Methods of Mechanical Testing” OST 78. G.; mandatory from May 1, 1927.

**) “Zodchii,” 1917, Nos. 7, 9, 10–13; 14–15; “Portland Cement,” Nos. 5, 9–10.

revision and amendment of the standard norms as a whole, the determination of the soundness of change in the volume of Portland cement constituted one of the most important points of discussion, as the most controversial one.

In all cases, at the center of attention was the question of a method of testing that would guarantee the obtaining of indications sufficiently reliable for characterizing the future behavior of Portland cement with respect to changes in its volume.

Of the proposed methods of testing—by a pat in water and under heating, by an apparatus and according to the Le Chatelier method, by calcining the ball, testing in boiling water, in steam, and others—not one has received universal recognition; and work on evaluating methods for testing Portland cement for the soundness of change in its volume is still being carried on at the present time (“Stroit. Promyshl.,” 1928, No. 1. Engineer Aleksandrov, “On the Soundness of Change in the Volume of Portland Cement”). However, the requirements of construction practice and of Portland-cement production made it necessary to adopt some one of the recommended methods for testing Portland cement with respect to the soundness of changes in its volume during the process of hardening, finding such a method, for one reason or another, sufficiently indicative for characterizing the material being tested, and to introduce it into the standard norms.

Different countries have legalized different methods of testing. The All-Union Standard of 1927 for the USSR established the former method of testing for soundness of change in volume, adopted in the acceptance standards for Portland cement in 1905 and 1911–1913: testing a pat in water (27 days) and in air by a hot test at 120°C for 2 hours.

The methodological part of the determinations of the soundness of change in the volume of Portland cement by one method or another, especially with respect to recognizing deformations characterizing defective Portland cement, was considered and established at congresses and in the literature with insufficient completeness.

In the All-Union Standard of 1927 of the USSR, the methodological instructions in the part concerning the determination of signs of defectiveness of Portland cement and the methods of recognizing them are given very concisely. The text reads: “Cement is recognized as defective if noticeable distortions or radial cracks are found on the pats.”

For all the apparent clarity of the text, the practice of testing indicates that the brief formulation of the standard norms leaves many unresolved, ambiguous cases and must, for the sake of definiteness, be developed in greater detail.

As regards the evidential value of the signs of unsoundness of change in the volume of Portland cement indicated in the standard—distortion of shape and radial cracks—no misunderstandings arise, nor do they arise with respect to the procedures for carrying out the tests: the number of specimens; preparation of the pats; temperature; curing in water. Difficulties and ambiguities arise in examining the specimens (pats) and recognizing deformations.

According to the text of the standard, deformations “noticeable to the eye” are taken into account. When deformations of one kind or another are sharply expressed, ascertaining them presents no difficulty upon simple inspection. But in testing practice one often encounters deformations only just beginning to appear within the short testing period established by the standard (29,

days), with cracks on the pats that are very fine, sometimes visible only under a magnifying glass, and sometimes detectable only by a special method, based on hairline cracking and described below, during testing with water.

Given the insufficiency of the method of examination and recognition of defects established by the standard, the variety of techniques used in testing above all leads to the obtaining of incomparable data: with the naked eye one may often fail to discern a fine crack and regard the cement as changing its volume uniformly. Upon more careful investigation and the application of other techniques these cracks are discovered. Should the cement in that case be recognized as having failed the test? The experimenter remains at a loss and at the same time may be reproached that his methods are arbitrary and that he overestimates the indications. The incomparability of test data and the apparent arbitrariness of the experimenter in examining specimens can be eliminated only by the direct establishment in the standard of a method for recognizing fine deformations. Another aspect of the matter is that the standard lacks a numerical characterization of the observed deformations. In fact, how great must distortions of shape be in order for the cement to be recognized as unsound? Prof. Zhitkevich, in his work Concrete and Concrete Works (ed. 1912, p. 39), gives the following numerical ratio: “a bulging of the pat or its concavity not exceeding ¼ inch with a pat diameter of 3 inches may be accepted as the limit for sound cement.” Without disputing Prof. Zhitkevich’s indication, I shall note only its concrete meaning: the normal thickness of the pat is 10 mm; ¼ inch is equal to approximately 6 mm; i.e., the permissible curvature, according to Prof. Zhitkevich, may reach 60%. In the absence of direct indications in the standard, one may doubt that experimenters take Prof. Zhitkevich’s numerical ratio as a guide, even if it is known to them. It is difficult to find a scientific basis for the indicated numerical ratio; meanwhile one may also have in mind another criterion for the significance of deformations: this is the old rule of the German specifications compiled by the Union of German Manufacturers: “the cement pat must give absolutely no distortions and cracks along the edges” (C. Naske. Production of Portland Cement. Russ. trans. Ed. 1913, p. 303).

According to the All-Union Standard of 1927, the test of a batch of cement for uniformity of change of its volume is performed on 4 specimens (pats): 2 by heating, 2 in water for 27 days. Hardly anyone will guarantee that of the 2 pats both will necessarily be defective or both will pass the test. In practice cases are not rare when one pat is defective and the other is not. What conclusion about the cement should the experimenter or the acceptance commission give in the presence of such contradictory indications? In the absence in the standard of any instructions for such a case, the matter is left to the discretion of whoever gives the conclusion about the tested batch of cement.

Thus, in order to resolve the uncertainties that arise in the practice of testing cement for uniformity of change of its volume, in the methodological part of the standard it would be necessary—having established the types of deformations characterizing nonuniformity of change of the volume of cement—to give instructions:

1. By what method should deformations be recognized?
2. What is the limiting magnitude of deformations (absolute or relative), expressed by a number, by which one could firmly

be guided by in determining the significance of deformations for the defective quality of the cement?

1. At what number of specimens from a batch does the discovery of deformations of a definite number and magnitude give the right to regard the cement as falling outside the range of normal cements (in simple terms—to reject it)?

In addition to these instructions, it would be desirable to add to the standard such a detail, at first glance unimportant, as the establishment of a method for placing the pats in the drying cabinet during the hot test, in order to avoid one-sided overheating. Such one-sided overheating, for example, is possible when the pats are placed in the cabinet on little stools, with the specimens set on edge vertically. A similar method was prescribed by the Austrian standards up to 1913 (C. Naske, Manufacture of Portland Cement. Russian trans. Ed. 1913, p. 312).

An instruction for carrying out the test for uniformity of change of volume of Portland cement, including the desirable additions listed above, presupposes the existence of methods for recognizing the features established by the standard, guaranteeing their complete detection, the possibility of taking them into account, and their quantitative assessment. The experience of testing laboratories and of construction practice with cements is sufficiently extensive that, from the methods in use, it should be possible to select the most reliable and sufficiently substantiated ones. Thus, for example, with respect to distortion of the shape of the pat (swelling, concavity), one such method was indicated by the American standards*): “the curvature of the specimen may be determined by means of a straightedge.” Unfortunately, the American standards stop at this point, without specifying exactly whether every distortion of the specimen’s shape—slight or considerable—characterizes the tested cement as defective. In the sense of carrying out measurements of deformations—convexity—concavity—this presents no particular difficulties. More essential is the question of the magnitude of the maximum permissible deformation.

The texts of foreign standards that adopt the pat method for determining the constancy of volume of cement do not provide numerical relationships and, in respect to the definiteness of their instructions, with the exception of the German standards with their categorical requirement, do not differ from the All-Union Standard of the USSR of 1927. Only in countries that have legalized in their standards the Le Chatelier method (England) is deformation regulated by a definite number. This latter circumstance indicates that the magnitude of the permissible deformation can be established, and that it is desirable also to work it out for the method of testing cement by the pat, which will provide an objective criterion for evaluating the indications observed during testing and will give the experimental material the desired comparability.

Following the instruction of the All-Union Standard of 1927, and determining deformations “noticeable to the eye,” the experimenter, when inspecting specimens with the simple, unaided eye, introduces too much that is subjective: first of all, the acuity of the observer’s vision plays an enormous role. It may be assumed with a high degree of probability that, not relying on eyesight, in many cases inspection is also practiced with the aid of a magnifying glass. In the author’s opinion, when inspecting specimens even with a magnifying glass, one cannot in some cases be sure that the pattern of cracking in the specimens will be established completely, not to mention here the number of cracks and their magnitude (length).

*) C. Naske. Manufacture of Portland Cement, 1913.

The simplest and most reliable method for detecting cracking in specimens is a technique based on wetting with water, with colored solutions, of the hairline vessels that the cracks in the pats represent. By this method all cracks are revealed, whatever their size, both in number and in magnitude.

The testing of cement pats with water or another liquid can be carried out in various ways. The simplest procedure is as follows: the cement pat, held in the hand or suspended on crossed wires, is brought into contact with water on the side on which it lay on the glass, as far as possible without wetting the edges and the upper dry side. The latter is not essential, but it is important for the clarity of the picture. When the pat is wetted, the water fills a through crack faster than it penetrates into the pores throughout the whole mass of the pat, moistens the cement mass along the entire length of the crack, and the latter appears on the dry surface of the specimen with complete definiteness. Instead of water, colored liquids may be used (ink, bluing), phenolphthalein; the latter after the test with water, so that a solution of caustic lime is evident in the crack. Contact of the specimen with the liquid by the indicated method may be replaced by wetting it with a brush from the reverse side.

The method described reveals the complete picture of cracking in the pats: the number of cracks, from considerable to negligible, their length, tortuosity, etc. The cracks found may be measured, at least in length, if this is necessary.

With time, the liquid that has filled a crack is absorbed by the porous material of the walls, and the appearance of the crack becomes blurred; in any case, however, it is possible to mark the end of the crack with sufficient accuracy.

This test, as is self-evident, is directly applicable to the testing of pats after the hot test. Here water may serve as the liquid, and only for recording both the general picture and the dimensions of the cracks may a colored liquid be used, as well as in the case when it is desired to follow the development of cracking in specimens during further storage.

The testing by the described method of specimens that have been stored in water for 28 days is somewhat more complicated. Having just been taken out of the water and only slightly dried, they are saturated with water, which also fills their cracks, if any are present. Testing with water will yield no result. The water must be removed from the cracks. Air drying for 1–2–3 days, or heating at a low temperature (30–40°C), will already give sufficient dryness of the specimens to make it possible to test them with water. But these drying procedures introduce into the determination a new factor—the drying out of the cement specimen—which, according to the author’s experience, affects the increase in cracking and leaves it debatable whether the cracks formed during storage in water or during lying in air. For the irreproachability of the test, the freeing of the cracks from the water contained in them must be carried out by a procedure that does not cause new or additional cracking. For this purpose the saturated state of the specimen should, as far as possible, not change. With some success, removal of water from cracks is achieved by keeping the specimens in alcohol and then treating them, for the purpose of the quickest possible drying, with ether, after which testing with water should be carried out as stated above.

More contrasting results on specimens taken out of water are obtained by a method based on the diffusion of staining liquids or an alcoholic solution of phenolphthalein. The colored liquid, penetrating into the crack by diffusion, stains it both from the back side and from the upper, dried side. It should be noted that diffusion proceeds comparatively slowly, especially in very fine cracks and at the tips of cracks. The pat must be kept moistened with the colored liquid or phenolphthalein for 15 minutes or more. The dye solution must not be too concentrated. Diffusion proceeds more successfully if the pat has first been treated to remove water from the cracks. A trial colored solution is applied with a brush to the back side of the pat. If the edges of the specimen prove sufficiently porous and, under prolonged action, become saturated with the colored liquid, producing a stained field, then the crack, if there is one, stands out against the general stained background with a deeper coloration.

The pattern of cracking in the specimen is obtained with sufficient relief, but less distinctly than in the test with water on a dried pat.

In any case, the diffusion method, applied to specimens just taken out of water, while excluding any side effects on the physical state of the specimen, establishes the cracks that appeared during storage of the test piece in water; therefore, the indications of the experiment are indisputable.

Thus, both with respect to the warping of specimens and with respect to their cracking, experimental material for taking defects into account can be obtained in sufficiently complete form.

The further task remains to develop a numerical evaluation of the indicativeness of the experimental observations for the characterization of the cement, determining, if admissible, the limiting magnitude of deformations as the norm by which the experimenter should be guided when judging the significance, for the characterization of the cement, of the indications of its tests.

The number of batches of cement that give defective test pieces when tested for uniformity of change of volume in water increases noticeably, according to the author’s observations, if, during testing, the technique described above—the water test—is used to reveal the complete pattern of cracking. Thus, stresses leading to rupture of continuity in the body of the cement specimen develop under water more energetically than during the artificial curing of cement in the hot test, which is also consistent with observations obtained earlier. These destructive tensile stresses manifest themselves most intensively in test pieces kept in water for 27 days, when they subsequently dry in air. Already after 1–3–4 days of storage in air of test pieces of cement taken from the water—regardless of whether they had passed or failed the normal test*)—the appearance of new cracks on the specimens could be established. With further storage in air and repeated water tests of the pats after various periods—a week, two weeks, a month, etc.—the following changes regularly had to be recorded:

1. The number of pats with cracks increases.
2. The number of cracks on the same specimens increases.

*) The tested batches of cement all passed the normal test in water. Additional tests were carried out on long-stored cement that did not exhibit constancy of volume.

1. The cracks themselves increase; it is especially easy to note their lengthening. The length of the cracks was observed to be up to 5 cm, i.e., half the diameter of the tile.

Some of the batches tested in this way, according to observations, after only 2 months after gauging gave up to 80% of cracked pats (16 out of 18; batches of 3000 barrels each). The dried pats are easily etched with water, thereby revealing the full picture of the developing cracking.

On pats tested by heating—defective ones and those that passed the test—no development of cracking was observed.

Observations were conducted on samples of 12 batches of cement tested over the course of a year. Unfortunately, the author has at hand cement from only one plant. To verify by observation on cements from other plants how general the noted phenomenon is lies beyond the author’s possibilities; therefore, for the present, one must speak of the observation as of a particular case.

The study of the noted phenomenon in cements from various plants is of interest in that, if results analogous to those described were obtained for a number of cements from different plants, it would be possible to raise the question whether Portland cement’s tendency, after setting under water and during service in air, to produce rupture of continuity is not a property of Portland cement, and to seek a genetic connection between this property and the cracking of concretes that so troubles practitioners of concrete and reinforced-concrete construction*). In fact, in the production of concrete structures intended for service in air, the concrete placement is kept for some time in a moist state (an analogy with curing cement specimens in water) in order to give the cement the opportunity to set under normal conditions. After this begins the process of drying of the structure in the open air—precisely the process which, in testing cement specimens for uniformity of change of volume, occurs when they are stored in air. The results of the latter are described above, and they consist in the development of cracking in cement tiles during drying.

Similarly, construction practice notes a change in the volume of concrete during drying, with the appearance of cracks on the surface.

It is beyond dispute that cracking of concretes is a phenomenon dependent not on the properties of the cement alone. A significant role is played here by tension and compression during wetting and drying and by changes in temperature, apparently of the inert material—the aggregate. The presence of the latter in concretes determines the complete nonuniformity of the mass, under which it is difficult to expect agreement in the movement of the material under one or another change in physical state. But if rupture of continuity, as a result of the processes occurring during drying in air, is inevitably produced by that material whose purpose is to bind into a monolith all components introduced into the concrete mass, then the question of cracking of concretes during service in air receives new illumination.

Hence, determinations of the uniformity of change of volume of cements in air, under various conditions of preliminary storage of specimens, are of essential importance; this also must be taken into account and regulated by the All-Union Standard of the USSR.

) Portland Cement, 1918, Nos. 1–3. Prof. Kunitsky, “On the question of the durability of reinforced-concrete structures.” Perkhun. Zeitschrift für Bauwesen*, 1916, H. 1–3.