Rational Analysis of Clays

A. E. Sabek.

Part I.

A Historical Study.

“Knowledge of the origin of a thing always gives us, from experience, the most fruitful points of view regarding its entire nature.”
Prof. Alb. Heim.

I. General course of the development of the idea of rational analysis of clays according to literary sources.

Applying the usual methods of quantitative analysis to the study of clays, as this was done not only up to 1876–1878, but also for a long time afterward, yielded a far from complete picture of how the individual mineralogical elements entering into the composition of clays were distributed, such as: the crystalline portion of silica, and also the comminuted portion of the rocks accompanying these clays. By the time of 1878–1879, chemical quantitative analysis played a secondary role in ceramic industries, as many technical authorities assert: Kerl, Seger, Bischof, Brongniart, Malagutti, etc., despite the efforts of technologists to use chemical analysis for the conscious conduct of ceramic production.

The first successful attempt to elucidate the mineralogical structure of clays in connection with their chemical composition belongs to the well-known Danish geologist Forchhammer and dates to 1835. The method employed by him consisted in dissolving the clayey substance (kaolinite, Thonsubstanz) in dilute sulfuric acid upon heating, while the remaining insoluble residues of rocks were then subjected both to microscopic and to chemical investigation. The dissolved clayey substance was likewise subjected to chemical investigation. In the form in which the method was proposed, it was distinguished by its complexity and required a great deal of labor and time for execution.

At first, Forchhammer’s interesting work escaped the proper appraisal of technologists. Only in 1876–1878 did Dr. Aron, a collaborator of the well-known expert in ceramic production Prof. Seger, give due attention to Forchhammer’s work and assign it its proper value.

If in Germany Forchhammer had no followers before 1876–1878, in France his followers were soon found. In their

In their studies of French clays, the scholars Brongniart and Malagutti¹ devote a series of articles to the investigation of the composition of clays by the rational method. They study in detail the dissolution in sulfuric acid of the constituent parts of clay. This investigation is of a purely scientific character and does not end with any proposals of a practical nature. The very name “rational analysis” belongs to Brongniart and Malagutti.

At first the hope of these scholars consisted in separating the clayey substance from the rocks accompanying it by elutriation, and then subjecting this purified clayey substance to chemical investigation. Having soon become convinced, however, that quantitative analysis of such elutriated clays cannot give a sufficiently clear explanation of the composition of clays, owing to their heterogeneity, they come to the conclusion that, for a satisfactory solution of the problem posed, it remains to resort to the so-called “rational analysis.” The first scientific critique of the method likewise belongs to them: they study feldspars and rocks similar to feldspars with regard to the action of acids and alkalis upon them, and conclude from their experiments that neither acids nor alkalis act upon the residues of rocks. The principal task of rational analysis—the separation of quartz from the soluble part of the silicates after digestion with sulfuric acid—is solved by them through the use of caustic alkali, in contrast to the soda used by Forchhammer. One may express regret that they did not leave a detailed description of the method: thus there is no information as to what concentrations they used, or how the separation of the soluble parts from the insoluble parts was carried out. These works date to 1855.

Ten years later Fresenius² studies Nassau clays by the same, but somewhat modified, method. He denies the rationality of using caustic alkalis for separating the amorphous and crystalline parts of silica, but unfortunately gives no data from which this follows.

Turning to a consideration of subsequent works, it must be noted that up to 1866 the development of the rational method proceeds in accordance with Forchhammer’s proposal, i.e. the removal of the clayey substance is carried out by dissolution with the aid of sulfuric acid. Müller’s work³ is distinguished precisely by this particular feature. Using phosphoric acid hydrate, he seeks to bring the clayey substance and feldspathic rocks into solution, while leaving quartz unchanged. Having fused the given clay with phosphoric acid, he dissolves this melt in boiling water and 1% sodium alkali. The defect of this method consists in the fact that between solubility in phosphoric acid

of the individual constituent parts there is not such a sharply expressed difference as is observed under the action of sulfuric acid. Möller’s method thus does not lead to any improvements in “rational analysis.” Subsequently Prof. Seger resumes similar attempts with sodium phosphonate, but soon becomes convinced of the method’s limited suitability and again returns to Forchhammer’s earlier proposal.

From 1876–1878 an intensive and fruitful development begins of the method proposed by Forchhammer. More than others in this respect was done by Aron and Prof. Seger. Both applied the new method in the most varied cases, not only in the scientific assessment of the quality of clays, but especially in application to the technology of ceramic industries. They showed that the new method of analysis can serve as a means for the technical control of a whole series of ceramic industries. They accompanied all their considerations with a number of examples taken from practice. It is therefore not surprising that technicians soon become acquainted with the method vigorously promoted by Prof. Seger and soon begin to apply it in production practice. At the present time every rationally organized ceramic production solves a whole series of essentially important problems by means of the above-mentioned method.

There was no lack in the literature of articles either completely denying the significance of the rational method in its application to practice, or regarding the new method with great distrust.

By persistent practical propaganda in factories and by a series of reports at meetings of specialists, as well as by a series of articles in special journals, such as, for example, “Notizblatt des Deutschen Vereins für Ziegel, Thonwaren, Kalk und Cementfabrication,” “Thonindustriezeitung” and “Sprechsaal,” the defenders of the new method achieved general recognition of the significance and importance of rational analysis for the technology of ceramic industries.

Seger’s appeal in the indicated articles for cooperative work on the application of the new method did not remain without response. One of the first to respond to the appeal was Lindhorst, introducing the first corrections to the method. In 1876, in the pages of “Thonindustriezeitung,” Lindhorst points out the advantage of using HFl in the decomposition of the insoluble residues of rocks in rational analysis instead of the fusion with alkalis proposed by Seger.

In 1879, with the publication of the 2nd edition of the well-known handbook by Bruno Kerl under the title “Handbuch der gesammten Thonwarenindustrie,” the method of rational analysis and its importance are mentioned for the first time; moreover, on page 53 of this work a brief description of the method is given. In the first edition of the same work, dating to 1871, rational analysis is not mentioned.

In 1890, in the 2nd edition of Jul. Post’s Chemisch-technische Analyse, there appeared a new description of the method of rational analysis, belonging to the engineer Olszewski. In this form the method is accepted by the Zurich professors Tetmajer and Schokke and is described in Mitteilungen der Anstalt zur Prüfung der Baumaterialien in Zürich, 1893.

Soon, in 1893, Prof. Seger gives a more detailed description of the rational method in Böckman’s chemisch-technische Untersuchungsmethoden, 3rd ed.

With the publication in 1895 of a new edition of Dammer’s Handbuch der Chemischen Technologie, in the section on ceramics we encounter a description of the method in the form in which it was practiced in the special laboratory attached to the journal Thonind.-Ztg. The description belongs to Prof. Hecht.

With the appearance of the first descriptions of rational analysis, a well-known expert on refractory clays, K. Bischof, comes forward as a zealous opponent of this method. However, having gradually become convinced of the expediency and suitability of the method, in 1899 the same Bischof appears in Lunge’s Chemisch-technische Untersuchungsmethoden with his own description of rational analysis. This description ends with the indication that the method has not yet been sufficiently developed to yield exact results.

Beginning with the appearance of the first descriptions of rational analysis by Cerl and ending with the last description by Bischof, i.e., over a period of more than 20 years, the method is gradually improved in its details. However, apart from separate descriptions in various manuals, no more significant articles of a critical character appear.

A certain exception in this respect is the appearance of an article by G. Vogt in Bericht des III Internationalen Congress für Angewandte Chemie, Bd. 3, 1898, under the title De la composition des argiles dite rationelle. The work of G. Vogt deserves great attention because it communicates a whole series of interesting observations. At the end of the article, however, it takes a new direction, leaving aside the detailed criticism and elaboration of the former method that interests us. At the same time, G. Vogt also leaves unfinished the new direction he proposed. No followers of this direction have yet been encountered in the literature.

All the articles cited above, with the exception of Seger’s, on the pages of Thonind.-Ztg., are exclusively descriptive in character. The authors of these descriptions do not enter into a critical evaluation of the method and often propose small modifications of it without adducing any decisive arguments to prove the correctness of their proposals.

On reviewing all the descriptions mentioned (see the appendix to Part I), what is striking is both the insufficient critical evaluation of the foundations of the method itself and the absence of detailed criticism in particular.

...the operations involved in carrying out its individual procedures. Thus, for example, the limits within which the method guarantees satisfactory results had not been determined. The accuracy of the method in general also seemed very doubtful; and in particular, given the variety of proposals, it was not clear on whose side the advantages lay.

To answer all the above and similar questions was the purpose of my article “Beiträge zur Kenntniss der Rationellen Analyse der Thone,” published in the journal Die Chemische Industrie in 1902. The present work on the rational analysis of clays is a development of the above-mentioned article, which appeared in print in the German language; the critique of the method is the subject of the second part of the work.

Since a critical reconsideration of the question of the best method of rational analysis of clays, in all its completeness, answered a pressing need, my above-mentioned article was received very favorably by both the foreign and the Russian literature. Since 1902, i.e. since the time when the above-mentioned article appeared, several of the most recent articles on the question of rational analysis of clays have appeared, of which the first in time is the dissertation of I. Koerner under the title “Beitrag zur Kenntnis der Elsasser Thone,” which gives a somewhat different direction, namely: Koerner proposes, after the usual treatment with sulfuric acid, to use \(1/2\%\) HF to accelerate and achieve distinctness in the separation of the clayey substance from the residue of rock and crystalline quartz; in this connection one may suppose that at the same time the finely divided rocks, as well as quartz, may partly dissolve. However, this latter dissolution is extremely slight. Koerner’s new method, notwithstanding its great interest and novelty, must nevertheless be regarded as insufficiently investigated critically.

Among the subsequent works, the compilatory article by Berdel stands out, on the pages of the journal Sprechsaal, beginning with No. 27 through No. 37 inclusive, for 1903. In addition to a historical survey of the development of the method, it contains a description of a small series of experiments on the action of alkalis on feldspars and quartz, and in its conclusions it fully agrees with the propositions accepted and demonstrated in my earlier article cited above.

The most recent article on the question that interests us is the very interesting monograph by Scheffler: “Beiträge zur Kenntniss der Westerwaldthone,” 1906. In this work, unlike the others, the question is treated in comparatively great detail, the descriptions are made in detail, and parallel experiments are presented by which one may judge the degree of accuracy of the method. Also of interest are comparisons of the results obtained both by Scheffler’s method and by Koerner’s method. As was to be expected, the results obtained by Koerner’s method are always smaller than by the other procedure. The drawback of the method adopted and described by Scheffler is its unusual duration, without any special gain,

moreover, in precision. The article contains no critically valuable indications concerning rational analysis.

Passing to a review of Russian works devoted to the study of clays, we must mention the monograph of Prof. Bogdanov: “Chemical Investigation of the Kiev Clays,” in the Transactions of the Kiev Society of Naturalists for 1883, vol. VII, issue 1. Here, among other things, the contradictory nature of the data in the literature regarding the rational analysis of clays is pointed out.

Especially worthy of attention is the very detailed monograph of Prof. P. A. Zemyatchensky of St. Petersburg University, entitled: “Kaolinite Formations of the South of Russia,” published in the Proceedings of the St. Petersburg Society of Naturalists, vol. XXI, issue 2 for 1896. In the chemical part of this extensive work there are many valuable indications on the question that interests us.

From that time onward, as far as I know, no articles or works have appeared in Russian literature treating clays in general, or in particular the rational analysis of clays.

True, there is also a very detailed work by the Master of Chemistry S. S. Tugutt—his dissertation: “On the Chemical Structure of Certain Alumosilicates,” published in the Scholarly Memoirs of Yuriev University for 1894. Unfortunately this work is more mineralogical than chemical in character, and touches little upon the question that interests us.

Among the articles “On the Rational Analysis of Clays,” one should also mention the article by the late Prof. of the Mining Institute K. Iv. Lisenko; to him belongs the idea of presenting, in brief form, the content of my article, printed earlier in German on the rational analysis of clays, in the pages of the Mining Journal for 1903, vol. I. The article ends with a critical comparison of my observations with certain observations and conclusions made by Prof. Zemyatchensky in the chemical part of his work.

Prof. Lisenko’s final conclusion—that an entire series of correct solutions concerning the rational analysis of clays appears possible, but that it will become possible to give preference to one or another only after a very thorough study, in a general formulation, of the action of acids and alkalis on the constituent parts of clay—led me to the thought of expanding my first task. Indeed, the general relations of clays to chemical agents have been so little studied that it is necessary to undertake a systematic series of experiments in this direction.

The solution of the above-stated problem, “on the chemical action of acids and alkalis on clays,” in its general form represents a very extensive task and may require a whole series of years before completion. Nevertheless, by three successful series of experiments carried out

together with three of my pupils—Engineer-Technologist P. N. Grebenshchikov and senior students I. I. Krasikov and G. S. Nagornov—I succeeded in clarifying, as we shall see below, a number of interesting questions concerning both certain properties of clays in general and the action of acids and alkalis on clays. The third part of my work is devoted to the exposition of these observations. Certain peculiarities of the properties of clays noted here seem to make it possible to outline a new method for solving the rational analysis of clays.

2. Forchhammer’s Investigation.

The German scholar Gmelin used, as a general method in certain cases, for the separation of complex mineralogical rocks into simpler constituent parts, the distinction of their solubility in water, hydrochloric acid, and sulfuric acid.

This method was apparently used by the Danish geologist Forchhammer when he was asked to evaluate the merits of domestic kaolins. At that time there was no generally accepted method for the comparative assessment of the quality of clays. At first it was proposed to place in parallel the study of kaolins from Germany, known for their applications, with unstudied domestic ones. Quantitative analysis served as the method of investigation. Soon, however, it became necessary to be convinced that quantitative analysis gave entirely incomparable data for very similar clays. It remained to seek another method, and Gmelin’s method was chosen as such. The method employed by Forchhammer consisted in boiling the clays with dilute sulfuric acid, whereby the clay substance (Thonsubstanz, kaolinite) dissolved, while the residue after dissolution, having been washed with weak alkali and water and examined under the microscope, indicated that this residue consisted of grains of quartz and feldspar. This observation is set forth in Poggendorfs Annalen, Bd. XXXV, S. 331, under the title: “Ueber die Zusammensetzung der Porzellanerde und ihre Entstehung aus dem Feldspath”.

I place below a detailed description of the course of work according to this method of Forchhammer, in view of its importance as the first example of the application of sulfuric acid to the analysis of clays, and also in view of the rarity of the publication in which the cited article appears.

Forchhammer, in studying the Zetlitz kaolin, proceeded in the following manner: 79.868 English grains of clay (1 English grain = 0.0648 gr.) were dried at 100° C. and ignited, after which they lost 7.458 grains of water; this clay was poured over with dilute sulfuric acid, transferred into a platinum vessel, and boiled until sulfuric acid vapors appeared. After this the resulting mass was boiled with water and washed on a filter with hot water until ...

it no longer reacts with litmus. The solution is boiled with an excess of ammonium carbonate; alumina is precipitated, which after washing and ignition gives a weight of 21.282 grains.

Placing the filter on a glass plate, they wash off the powder insoluble in sulfuric acid into a platinum crucible with wash water; the filter is dried, burned, and the ash is weighed. The correction is 0.451 grain. The powder in the crucible is boiled three times with sodium carbonate, the alkali is diluted with water and filtered in the presence of an excess of ammonium chloride, evaporated to dryness, dissolved again, and then the silica is washed with hot water on the filter. The silica obtained in this way is partly soluble in water; therefore the filtrate and wash waters are preserved, and the washing is continued until the spot obtained from a drop of water on a platinum plate, when moistened with a new drop, no longer dissolves, and in another place on the plate the drop ceases to leave a coating. The silica obtained by the above method weighed 23.76 grains.

Evaporating the filtrate and wash waters, igniting and dissolving again, there was obtained in all 0.776 SiO₂. The residue of clay, washed with sodium carbonate, is again washed off the filter, and the filter is burned, obtaining in the residue 0.122 grain. The washed precipitate is treated with boiling hydrochloric acid, precipitated with ammonium carbonate, washed from the filter, leached by boiling with sodium carbonate, and then washed. By this method there were obtained a further 0.176 grain of alumina and 0.537 grain of residue on the filter, 0.246 grain of silica and 24.506 grains of insoluble silica. The initially obtained 21.214 grains of alumina were dissolved in hydrochloric acid, leaving undissolved 0.343 (a) grain of silica. The solution, not containing even traces of sulfuric acid, was boiled with caustic potash containing a small amount of carbonic acid. The unwashed precipitate was dissolved with hydrochloric acid from the filter, averaged and precipitated with succinic ammonia, whereby after ignition 0.670 (b) grain of iron oxide was obtained. Ammonia precipitated from the remaining solution 0.500 grain (e) of a yellowish-white powder, consisting of a substance different from alumina; with a small amount of ammonium oxalate a small quantity of lime is precipitated, which after ignition amounts to 0.081 (d) of carbonic lime. Therefore the alumina amounts to 19.620 grains. The filtrate separated from the first precipitate of alumina was evaporated to dryness and the residue ignited; the residue was dissolved with a small amount of sulfuric acid and water, giving thereby 0.187 grain (I) of silica. The alumina, precipitated from this filtrate by ammonium carbonate, amounts to 0.136 grain (12), and in the residue there are very small traces of potassium and sodium.

Starting from 21.214 grains of clay, the sum of the constituent parts obtained is 21.282 grains.

Omitting here a part of the discussion concerning certain peculiarities in the course of the analysis, I pass on to its conclusion: from the analysis given above it follows that the ratio between alumina, silica, and water is expressed by the formula:

\[ \mathrm{Al_2O_3\ 2SiO_2\ 2H_2O}. \]

This ratio for silica and water attains great precision; as regards alumina, however, the precision would be no less only if to it were added the ferric oxide and the unknown substance separated from the alumina. Besides the Zettlitz kaolin, the following were analyzed:

1. Porcelain clay from Schneeberg.
2. Clay from Halle.
3. Clay from St.-Irie near Limoges.
4. Porcelain clay from Bornholm.
5. Earthy leucite from Kalia.
6. Plastic clay from Gross-Almerode.

A large part of the clays listed have a composition expressed by the following percentage composition:

which corresponds to the formula:

\[ \mathrm{Al_2O_3\ 2SiO_2\ 2H_2O}. \]

In the subsequent exposition the author, by microscopic investigation, proves that the residue insoluble in acid is nothing other than feldspar.

Dwelling in detail on the chemical composition of feldspar, he subsequently establishes its most probable chemical formula:

\[ \mathrm{K_2O\ Al_2O_3\ 6SiO_2} \]

Forchhammer’s work, notwithstanding the insufficient development at that time of the methods of quantitative analysis, may, as regards the accuracy of its results, be considered not only not obsolete but, on the contrary, the significance of this work deserves all the greater attention because, despite the considerable improvement of the methods of quantitative analysis, the most recent works do not in the least refute Forchhammer’s conclusions, but only confirm their correctness anew.

Considering all the complicated operations in the description given, it is easy to conclude that this method is too complex for practical purposes and, in such a form, could find few adherents for use in practice. The very importance of the proposed separation of clay into mineralogical constituent parts, even if it was recognized, was recognized only by such eminent scholars as Brongniart and Malaguti, or others like them.

3. Works of Prof. Brongniart and Malaguti, Dr. Fresenius and Moller¹.

If the conclusions drawn by Forchhammer with the aid of the new method concerning the composition of kaolins did not find supporters among technicians engaged in the ceramic industry, they nevertheless aroused general scientific interest. Thus, for example, a series of articles by Prof. Brongniart soon appeared, written together with his regular collaborator Malaguti, who were interested in elucidating the composition of native clays used as raw material at the well-known Sèvres manufactory. Being placed in such close contact with the practice of production, they were interested in clarifying both the chemical and the mineralogical composition of the kaolins used in the work. At first they resorted to elutriation of the kaolins, hoping by this means to separate the clay substance from the rocks accompanying it, and they came to the conclusion that by this method separation into mineralogical elements is not attainable: together with the clayey substance there is also elutriation of the finest particles of rock. As a result, they were compelled to resort to a new method, in which the clayey substance could be separated by successive treatment, first in boiling sulfuric acid and then in alkalis. Accidental minerals, such as quartz, feldspar, and others, in these operations, in their opinion, do not dissolve. To verify what had been said, they subjected a whole series of feldspars, both unaltered and subjected to weathering, to the action of the same reagents.

In rational analysis two essentially different tasks are distinguished: 1)—the separation of the clayey substance from rocks by means of dissolution in sulfuric acid, and 2)—the separation of quartz from rocks. Of these tasks Brongniart and Malaguti study the first part with particular thoroughness.

The observations of Brongniart and Malaguti are very interesting, and it is therefore regrettable that they did not leave a detailed description of the course of the entire analysis—they indicate neither the concentrations of the reagents nor the other conditions under which the best results could be obtained. From a comparative consideration of the 33 analyses carried out by them,

it remains the general impression that they themselves considered the method insufficiently precise, or as requiring great caution in drawing their conclusions. They point out that, before treatment with sulfuric acid, they first treated the clays with caustic alkali, sp. gr. 1.075, whereby all the free soluble silica is removed, leaving the kaolin untouched. My further observations indicate that kaolin is soluble in alkalis, while, according to Prof. Lunge’s observations, the solubility of crystalline quartz is beyond doubt.

Besides a certain verification of the foundations of the method and confirmation of its scientific significance, they at the same time expressed the view that this method may have very great importance for practical purposes. They cite their experiments, in which, according to analysis of a porcelain body, a composition made up from chemically pure oxides, such as \(Al_2O_3\), \(SiO_2\), \(K_2O\), and \(CaO\), melts at a different temperature and has properties entirely different from those of the porcelain body. The same proposition was subsequently developed in an interesting Russian article, where it is confirmed by a considerable number of experiments. This work belongs to the engineer-technologist F. Ya. Yakovlev and is entitled: “A Brief Outline of the Development of Porcelain Manufacture in Russia,” Moscow, 1882. From the above-cited experiments, Brongniart and Malaguti come to the conclusion that only porcelain bodies composed on the basis of rational analysis could satisfy identical conditions of melting and of other qualities.

Brongniart and Malaguti reduce the results of their scientific investigations to the following propositions:

1. The so-called true kaolins, even after elutriation from the coarsest foreign impurities, are nothing other than a mixture of clayey substance and a residue of silicates of various bases, insoluble in acids and alkalis.

2. The clayey substance (kaolin) is freed from this residue by successive treatment with sulfuric acid and caustic alkali. This method we call rational analysis.

3. The clayey substance is a compound of silica, alumina, and water in ratios that are almost always definite and mutually equal. This ratio may be expressed by an invariable formula.

4. However, in many of these clays there is found an excess of silica in a free form, which may be dissolved by caustic potash according to definite rules; in this way it is possible to separate the silica from the hydrous alumina silicate which constitutes the true clayey substance (kaolin). The remaining, water-containing, alumina silicate may in general be expressed by the formula: \(Al_2O_3, 2SiO_2 + 2H_2O\).

1. The excess of silicic acid in kaolin clays is separated from kaolin by the means indicated above, and its formation may be ascribed to the gradual decomposition of feldspar under the action of electricity. Feldspar first passes into kaolin \(Al_3S_3\) (according to the old notation) and into the insoluble potassium silicate \(KSq\), and this latter then again breaks down into the insoluble potassium silicate \(KS_3\) and silicic acid, which, mixing with the clayey substance, remains in it.

2. Finally, deviations in the available quantity of this excess silicic acid in different clays may be ascribed to the subsequent action of aqueous masses, which carry away with them a greater or lesser quantity of this free and soluble acid.

Ten years later there appeared in print a very detailed work by Fresenius¹, devoted to the study of the Nassau clays. In most of his conclusions Fresenius adheres to the conclusions of Brongniart and Malaguti. Fresenius, among other things, notes the advantage of applying boiling dilute sulfuric acid for the decomposition of the clayey substance; he rejects the use of caustic alkali for removing soluble silicic acid, considering such treatment inexpedient because of the partial dissolution of crystalline silica. Preferring soda to caustic alkalis, Fresenius inclines more toward Forchhammer’s method.

Thus, without dwelling more closely on the study of the action of caustic and carbonated alkalis, scholars wavered, preferring to employ in the treatment of clays now caustic, now carbonated alkalis.

Subsequently, Prof. Lunge dwelt in detail, in a whole series of dissertations by his students, on the dissolution of crystalline and amorphous silica by the action of caustic and carbonated alkalis.

Möller’s article, published in 1866², differs substantially from what has been set forth above in this chapter. Möller, being a staunch opponent of Forchhammer’s proposal, made an attempt to direct rational analysis along a new path, namely: for separating crystalline quartz from rocks, he proposes using phosphoric acid hydrate. In his opinion, with a gradual increase in temperature, phosphoric acid first dissolves silicates little by little and only later, at a higher temperature, crystalline silica. The quantity of phosphoric acid required for dissolving the clayey substance and the rocks must exceed the quantity of the latter by approximately 30 times. The resulting melt is then boiled with 1% caustic sodium. The fusion of the remaining quartz is repeated with phosphoric acid once more and then, taking it up on

filter, washed with acids and alkalis and, finally, with water. The concentration and other details are described by Müller in his article. The solubility of the constituent parts of clay does not differ so sharply from one another that one might hope to obtain definite results. The data obtained fluctuate now in the direction of increasing the silica, now in the direction of decreasing it, depending on the amount of flux and the degree of heating. There is no control over when the fusion should be stopped. It is clear that this method offered little hope of possible improvement. And indeed, apart from Seger’s attempts to use sodium phosphonate salt for the same purpose, no other analogous proposals are encountered.

4. Aron’s article propagating Forchhammer’s method.

Before 1874 it is difficult to cite examples from which it would be evident that Forchhammer’s method was applied in the investigation of clays; and only in that year did Dr. Aron, having set himself the goal of finding the method of clay investigation most satisfactory for practice, make use of it. Dr. Aron dwelt with special attention on the above-mentioned article by Forchhammer and gave it a fair assessment. Dr. Aron’s discussion was published in the long out-of-print edition: “Notizblatt des deutschen Vereins für Ziegel,- Thonwaren,- Kalk- und Cement-Fabrikation.” Berlin, 1874, under the title: “Eine Bemerkung über Untersuchung von Thonen” von D-r Julius Aron.

In this article the author sets forth the significance of rational analysis, both for practice and for the general understanding of the composition of clays, with such detail and persuasiveness that I shall allow myself to cite several excerpts, so as not to repeat, in substance, the very same thing in my own words.

“Both the test itself and the judgment of the technical suitability of any clay on the basis of the results of its test present entirely peculiar difficulties. It is not enough to establish the elementary composition of the clay, i.e., the relative quantities of silicic acid, alumina, and other constituent parts, in order to be able to judge those properties of clay materials which are very essential in practical respects.”

“Every clay is a mixture of substances of different chemical composition, possessing different physical properties; therefore every clay manufacture strives, as its first aim, by means of various technical procedures, to bring this mixture of substances to a greater or lesser homogeneity.”

“The interest of the manufacturer of clay products in a given material ceases precisely when chemical homogeneity is established, i.e., fusion or general sintering of the mass of which the body consists.”

This is the essential difference between the manufacture of clay products and that of glass.”

“In both branches, objects of the most varied forms are produced. In the glass industry an object is given its form when the whole mass has first been brought to complete formlessness. In clay manufacture, however, objects are given, while still in the unfired state, the form which they must retain after firing; i.e., in this case they do not reach the limit at which the whole mass is brought to complete homogeneity.”

“This is precisely the reason why tests of clays intended for technical purposes must give indications both of the nature and of the properties of the individual substances composing the clay, and why it is in no way correct to regard clay as something chemically homogeneous.”

“If one breaks and examines a mettlach tile fired at a high temperature, it is nevertheless easy to distinguish in it the mineral constituent parts, despite the firing, which in its manufacture has been carried as far as sintering (Sinterung). Porcelain likewise does not constitute an exception in this case.”

“And so, in all clay manufactures one observes the absence of complete chemical homogeneity. What right, then, have we afterwards to unite all the component parts of a clay into a homogeneous mass, as is involuntarily done by chemical analysis?”

Simultaneously with the appearance of Aron’s articles, which pointed out the inadequacy of a single quantitative analysis for the technical evaluation of clays owing to the heterogeneity of the latter, there appeared in print a well-known article by K. Bischof, “On the Determination of the Degree of Refractoriness of Clays from the Results of Quantitative Analysis of Clays.” The determination of refractoriness is reduced to the calculation of the so-called “coefficient of refractoriness,” on the assumption of complete homogeneity of clays.

Thus, two currents became established in the literature: one inclined to regard clays as chemically homogeneous substances, the other, quite the opposite, as chemically heterogeneous.

Aron appears as a vigorous opponent of Bischof; he constructs his objections as follows: if, for example, we imagine two bricks, made up of a definite quantity of one and the same easily fusible clay, mixed in one case with coarse quartz grains, and in the other with grains of feldspar of equal size, then both bricks, at the same low firing temperature in both cases, would acquire completely identical qualities and strength, despite the difference in their elementary composition.

An easily fusible clay sinters at a temperature at which neither quartz nor feldspar yet takes any part in the melt—

Rational Analysis of Clays

... melting. In this case the clay cements the mineral grains, the chemical composition of which plays no role whatever.

From these considerations, Aron declares, it cannot be an advocate of Dr. Bischof’s method, which, without making any distinction between brick clays and the more homogeneous refractory clays, determines by one and the same procedure the coefficient of refractoriness or fusibility.

Aron’s considerations were undoubtedly correct, as is evident from K. Bischof’s article in No. 4 of the Notizblatt des deutschen Vereins für Ziegel-, Thonwaren-, Kalk- und Cement-Fabrikation for 1873, in which Bischof himself, attempting to find numerical expressions for the fusibility of two brick clays on the basis of quantitative analysis, by the calculation set forth in detail in the cited article, arrives at the conclusion that his calculations are of little applicability to the given case. Further, trying to find the cause of such a phenomenon, he gives the following explanation: both clays melt at a temperature insufficient for the formation of acid silicates, and therefore both the added silica and the silica present in excess do not come into a state of fusion.

The discrepancy between the expected results and reality is nothing other than a consequence of the incorrect assumption of the complete homogeneity of clays.

The coefficient of refractoriness is the closer to the truth, the closer the melting temperature of all the constituent parts of the material is, i.e. the more precisely the conditions of its complete homogeneity are observed.

Only the rational analysis of clays¹ can answer the unexpected results cited above. The latter also makes it possible to establish the origin of the clayey substance from feldspar or from rocks related to it.

Dissolution in the rational analysis of clays is carried out with sulfuric acid; the soluble silica separated in this process is removed with the aid of sodium carbonate. This alternating action of sulfuric acid and sodium carbonate is continued until the latter finally ceases to dissolve silicic acid. All the clay passes into solution, while the undissolved elements of the rock remain in the residue, such as feldspar, quartz, etc. This procedure, which until now has been used for solving purely scientific questions, on the basis of the points of view stated above, appears to be the only correct method. With its help, while simultaneously elucidating the mechanical analysis and the elemental composition of the total mass of clay, one can obtain proper indications concerning the technical suitability of clayey material in all respects.

The importance of mechanical analysis for the technical evaluation of the quality of clays has been noted by an entire series of articles by Aron and Prof. Seger. The chief role of mechanical analysis comes down to clarifying the relation of the given clayey material to the operations of preparation, namely, molding and drying.

As regards firing, however, i.e., where mechanical analysis cannot provide proper indications, there rational analysis comes into play.

As to the extent to which the latter, as an analytical method, guarantees scientific accuracy for the strict separation of clay from the mineral constituent parts, Aron does not venture to state, since this was only the first attempt to make use of the method, while the degree of accuracy of the method had as yet been little studied. By his article Aron wishes only to arouse interest among technical specialists and to provoke both written and oral exchanges of opinions in meetings of societies on this question. To simplify the complex problem of general analysis, Aron proposes that the total mass of clay be mechanically divided, and that part of it whose particle diameters are \(=0.01\) mm, at a water-flow velocity of 0.18 mm per second, be called the clayey substance and that only it be subjected to rational analysis. As may be seen, this proposal fully coincides with the proposal of Brongniart and Malagutti.

Aron further describes the drying of clays practiced in his laboratory at \(130^\circ\) C and treatment with hydrochloric acid to remove salts of lime, iron-oxide hydrate, etc.

Concluding his interesting article, which opened a new path for the technical evaluation of clays, Aron indicates that it is precisely along this path that he expects to obtain a series of technically applicable analyses, while for the present he sets as his principal aim the announcement of the method he has found in the literature, in order to call upon specialists for an exchange of opinions and a critique of the method he proposes.

As is evident from the preceding exposition, the entire modification of Forchhammer’s method, according to Aron’s proposal, consists in the use of \(\mathrm{Na_2CO_3}\) for leaching out dissolved silica instead of \((\mathrm{NH_4})_2\mathrm{CO_3}\); the remaining treatment stays as before.

In its details the procedure has not yet been worked out, but it is clear that it must, to a considerable degree, be simplified for practical purposes.

As regards the limits of applicability of rational analysis, Aron indicates in general terms that the more homogeneous and uniform a clay is in composition, the greater the importance quantitative analysis has in deciding practical questions. But as soon as homogeneity is disturbed, there arises an urgent necessity to resort to the method of rational analysis or to methods of mechanical elutriation.

5. Brief Review of Methods of Mechanical Analysis.

From the preceding we have seen that both Brongniart and Malagutti, as well as Aron, placed great hopes on the mechanical levigation of clays and advised that, before rational analysis, one should resort to levigation and subject only the levigated portion to more detailed study. This path seems quite natural because, in most manufactures, clay as a raw material first undergoes the operation of levigation and only in this form is introduced into the composition of masses, which must be specified in accordance with the chemical composition of the raw material. Thus it is clear that, in practice, it is more important to know the composition not of the raw clay, just brought from the quarry, but of the levigated clay that goes directly into the work.

In the great majority of cases in the various ceramic industries, where the question is not melting but firing clay articles, i.e., where the plastic part of the materials envelops the non-plastic part, the chief interest of the investigation must fall on the study of the active mass of the binder, or the so-called clay substance (Thonsubstanz). On the quality of this latter depend both the degree of plasticity and the binding capacity of the clay. Both these qualities are of very great importance in forming, in drying, and also in firing the articles.

In determining the most essential part of clay—the clay substance—one encounters, however, not a few difficulties; these difficulties are so great that Brongniart and Malagutti, as well as Seger and Aron, who had placed great hopes on the separation of the clay substance from the accompanying rocks by means of mechanical analysis, in the end considered it only as a preliminary preparation for rational analysis. Perhaps these methods, after half a century, have advanced so far that one may now count on obtaining more reliable results? To answer this question, I give below a review of the existing methods of mechanical analysis in a brief exposition, with an assessment of their degree of accuracy and suitability both in technology and for scientific investigations.

The following principal methods of mechanical analysis exist¹:

A. Method of levigation based on the application of the velocity of fall of particles in water.

The simplest methods of levigation consist in applying the force of gravity, whereby particles of different size and different weight settle to the bottom at different velocities, since water exerts, in this case—

this is a resistance of varying degree, slowing the settling of the particles to the bottom.

a) Method of decantation in a closable glass cylinder with a siphon for removing the suspension.

By this method, in Prof. Van Bemmelen’s opinion, sufficiently accurate results can be obtained, but its inconvenience lies in the difficulty of separating the coarser particles; moreover, the duration of the work required by this method is very considerable.

b) Prof. Williams’s method according to the idea proposed by Prof. A. A. Fadeev.

The essence of this method is likewise reduced to the action of the force of gravity on the particles and to the different rate of settling. The separation of the particles proceeds according to the time of settling. The capacity of the vessels is considerably increased and is better adapted for decanting the liquid from the sediment. With the simultaneous use of Dr. Schene’s apparatus, this method makes it possible to make a larger number of gradations according to grain size. The essence of the method is the same decantation, but, owing to the successful arrangement of the corresponding apparatus, the work is considerably simplified and requires no such large expenditure of time.

c) Apparatus for washing out in the form of vessels with stopcocks, or in the form of flasks with divisions.

This includes the apparatus of Kühn, Knop, Beningsen, Clausen, etc. In general these instruments are far from perfect and are used exclusively for the study of soil for agronomic purposes; for technical purposes, in the analysis of clays, they are too coarse.

B. Apparatus with a vertical jet of water:

a) Nöbel’s apparatus with four vessels, the capacities of which relate to one another as $1^3 : 2^3 : 3^3 : 4^3 = 1 : 8 : 27 : 64$.

The chief drawback of the apparatus consists in the fact that in the 40 minutes for which it is designed, the clay does not always have time to separate into 4 completely homogeneous parts; also, owing to the poorly chosen acute angle in the lower part of the vessels, currents of different velocities may form inside the vessel, giving unequal results.

b) Dr. Schene’s method with one or two elutriating vessels and a piezometer for measuring the corresponding pressures of the water column under which the washing-out is carried out. Dr. Schene proposed an exact theory of the action of the instrument. With continuous and uniform use, this instrument gives very accurate results. A detailed separation of the finest particles, however, cannot be obtained—

...be done with the aid of this apparatus. If there is a desire to obtain such a separation as well, one resorts to the methods developed by Prof. Williams and Fadeyev.

c) Finally, Hilgard’s cylindrical apparatus with a winged mechanical stirrer can give very satisfactory results, but the apparatus is bulky and inconvenient, and requires a driving force at a speed of 500–600 revolutions per minute. All these inconveniences force one to avoid the use of this apparatus; its use is especially widespread in England and America.

This exhausts the methods of mechanical analysis of clays; among all the methods, special attention is drawn to Dr. Schöne’s method¹, the most fully developed theoretically, and also to Prof. Williams’s method, set forth in his dissertation².

Soon after the publication of Dr. Schöne’s dissertation in 1872, those interested in developing the question of a proper technical assessment of the quality of clays—Prof. Seger and Dr. Aron—placed great hopes on the mechanical analysis of clays. Thus, in the article “Die Mechanische Analyse des Thones”³, Prof. Seger dwells in detail on the significance of mechanical analysis in the investigation of clays. At the end of the article, however, he concludes with words of dissatisfaction: “the difference in qualities among clays is determined by the size of the grains, their kind, the properties of their surface, and the mutual relationship of these quantities; however, for judging in practice the qualities of clay, prolonged experience and the systematization of observations are necessary, and until this exists, mechanical analysis can give few practical indications.”

Among Russian works, apart from the well-known dissertation of Prof. Schöne, “On the Mechanical Analysis of Clays,” attention is merited by the dissertation of Prof. V. R. Williams of the Petrovsky Agricultural Academy (1892), entitled: “An Experimental Study in the Field of Mechanical Analysis”⁴. In the latter work a review is given of the most important preceding data on mechanical analysis, together with a whole series of the author’s own observations, both on soils and on clays. The examples indicate that the method has attained a considerable degree of perfection, and it is only regrettable that there are no investigations in which, on the basis of data from mechanical analysis, other physical and chemical properties of clays would be compared. In the work of Prof. P. A. Zemyatchensky: “Kaolinite Formations of Southern Russia”⁵, there is a series of interesting data and individual comparisons on the same question. However, mechani-

mechanical analysis must in any case, in this interesting work, play the secondary role of a means, and it is not applied everywhere.

Summing up everything known about the mechanical analysis of clays, we observe that elutriation in all cases is accompanied by a whole series of difficulties. The constituent parts of clay are separated in mechanical analysis according to the size of the component particles entering into it. Such separation gives, for particles of different specific gravity, different magnitudes at one and the same velocity of the water current. This consideration alone indicates the impossibility of expecting the separation of the grains of the clayey part from the other component parts and, indeed, a whole series of observations has fully confirmed these expectations, so that Prof. Seger, who at first expected to obtain important conclusions, comes in the end to the conclusion that the methods of mechanical analysis are of little suitability for clarifying the composition of clays. Mechanical analysis retains only a certain significance as an indicator of the relative distribution of particles of different size in the general mass of each clay; in this connection it is customary to count as clayey substance all particles whose diameters are less than a certain known magnitude. This boundary, being conventional, could be shifted with equal right in either direction, and still there is no guarantee of the complete separation of the clayey substance from the accompanying rocks.

In general, therefore, mechanical analysis has not led to the expected results. The chief difficulty in obtaining proper results consists in the fact that the separation of grains by size is not accompanied by separation into mineralogically homogeneous parts, whereas in most industrial processes at high temperatures, when chemical interaction of the component parts begins, precisely such separation is desirable.

Despite the fact that the methods of mechanical analysis, especially since the publication of the dissertations of Dr. Schene and Prof. Williams, enjoy recognition and are widespread in Russia and abroad, and also despite the fact that in various monographs and scientific treatises a considerable number of elutriations of clays have been presented, nevertheless it may still be said that there is not one of them in which the data would not be more or less fragmentary, and not one work in which all these data would be systematized and correlated with other properties of clays.

This is why, despite the elaboration of the method itself, it has not yet received proper application either in practice or in natural-historical descriptions of clays.

Agronomy constitutes a certain exception, where a connection has been established between the fineness of soil particles and its fertility, and where particular applications lead to known kinds of norms by which one may judge the greater or lesser fertility of the soil.

The general conclusion concerning mechanical analysis will be this: by this route there is as yet little hope of separating the clayey substance from the accompanying rocks, and thus there remains the only path—the decomposition of clays into their mineralogical constituent parts. This is the path of rational analysis.

6. First verification of the foundations of the method of rational analysis.

Experiments of Prof. Seger and the indications of Dr. Lindhorst.

For a long time, in response to Dr. Aron’s call to test the accuracy of the proposed method, no volunteers were to be found. Seger and Aron themselves were compelled to undertake the solution of the task posed.

The first, after Brongniart and Malaguti’s attempts, was Seger, who, although not in a large number of experiments, undertook to verify the suitability of the method.

On page 40 of Seger’s collected works¹ we find the following: “until 1876 rational analysis was applied chiefly for the purpose of determining the composition of the clayey substance, while the mineral constituent parts, which play an essential role in a whole series of productions, received no proper attention. Thus, for example, in the book Bruno Kerl analyses of kaolins from various deposits are presented in the form of complete analyses, with the addition of an analysis of the part soluble in hydrochloric acid; the rest is designated as a residue insoluble in acids.”

Bischoff and Richter, in their analyses, distinguished two kinds of silicic acid: crystalline and amorphous, i.e. chemically combined; after dissolution of the amorphous acid, the residue is taken to be grains of sand or crystalline silica. In fact, clays are very rarely encountered in which the residue insoluble in acids would consist of quartz alone exclusively; this residue always contains alumina and alkalis, which indicates an admixture of alkaline silicates.

Before proceeding to a series of analyses using the rational method, Aron and Seger consider it necessary to carry out a verification of the method. For this purpose Seger subjects a series of clays to the action of sulfuric acid and then observes what change the residue undergoes upon further treatment with sulfuric acid, and establishes the moment when these changes may be considered complete, i.e. when the residue from the action of sulfuric acid remains constant.

These experiments were carried out on two sorts of clay: a very plastic one from Grünstadt and a very lean brick clay from Schwarzehütte near Osterode on the Harz.

For each clay, 3 portions were weighed out, each about 2 grams; it was boiled with water in a platinum crucible until the particles were completely separated, then 10 cc of concentrated sulfuric acid were added, and it was then heated first on a water bath and afterward over an open flame until the sulfuric-acid vapors had completely volatilized, regulating the flame so that boiling with sulfuric acid at \(300^\circ\) C continued. In one portion the experiment ended at this point; to another sulfuric acid was added and, after the paste had been kneaded, it was again subjected to a 5–6-hour distillation of sulfuric acid at \(300^\circ\) C; in the third portion a similar operation was repeated once more, so that the action of sulfuric acid in the different portions lasted from 5 to 6, from 10 to 12, and from 15 to 18 hours; the residue was then leached twice with water, with dilute caustic soda, and finally with hydrochloric acid, in order to remove the separated soluble silica and the basic sulfuric-acid salts. The residue retained on the filter was washed, ignited, and weighed. The results obtained by this method are given below.

From this it is easy to conclude that even a double distillation, or, what is the same thing, the action of sulfuric acid for 10–12 hours, brings the decomposition to completion, and that a single distillation of sulfuric acid for 5–6 hours is insufficient.

However, in order to be fully certain, the clay was subjected to the action of sulfuric acid for 15–18 hours. In other words, the operation of distilling off sulfuric acid was repeated three times.

In order to determine whether, in the given case, mineral rocks were present in the residue, the residue was treated with dilute sulfuric and hydrofluoric acid and the silicon fluoride was distilled off, while the residue, consisting of sulfuric-acid salts of alumina and alkalis, was ignited and weighed. The residue amounted to:

Clay from Kronstadt . . . . . . . . . . . . . \(0,26\%\)
Brick clay from Schwarzenhütte . . \(8,54\%\)

From this it follows clearly that one of the clays did indeed contain almost exclusively grains of quartz in the residue, whereas the other contained a considerable precipitate of basic substances in the form of mineral residues of rocks not decomposed by sulfuric acid. These constituent parts during fusion must, of course, behave otherwise at high melting temperatures than quartz.

Further experiments were carried out on the resistance of feldspar to boiling sulfuric acid. For this purpose, powder of feldspar that had been subjected to weathering was treated with sulfuric acid and alkalis and, after weighing the washed and dried precipitate obtained, yielded:

$$ 97.76\% $$

of the feldspar taken previously. This slight difference of \(2.24\%\) indicates that feldspar may be regarded as almost undecomposable by acids, especially since even this small difference may arise from a scarcely noticeable decomposed portion of the feldspar or, as I have personally had occasion to observe, from a decrease in the weight of feldspar upon ignition, if such ignition was performed. This loss, resulting from the loss of moisture or other volatile constituents, may reach \(2\text{--}3\%\), as is evident from Part II of this work. In conclusion, it must not be left unsaid that the number of experiments performed is too small for a final solution of the problem posed above.

After dissolving the clayey substance in sulfuric acid, it is still necessary to separate quartz from the silicates. For this purpose attempts were made, as we saw earlier, to apply various kinds of solvents; thus, for example, Müller proposes phosphoric acid hydrate, and others sodium phosphonate and boric acid. The last two agents were again tested by Seger; however, owing to a certain solubility of silica in these reagents, and also owing to the incomplete dissolution of feldspars in them, the results obtained were negative. Having thus convinced himself of the futility of direct separation of the mineral constituents of clay, Seger proceeds further to the following method: the residue from treatment with sulfuric acid was mixed with a fourfold quantity of sodium-potassium carbonate; the mass was fused without any grinding, so that during grinding no parts of the substance would dissolve; such a fusion operation is readily accomplished when the content of \(SiO_2\) is high. After this melt had then been dissolved in water, it was treated with hydrochloric acid and, by repeated heating on a water bath, the silica was brought into an insoluble state; alumina and iron were precipitated from the filtrate with ammonia, and from these data the amount of alkalis was calculated.

If there is certainty that the clay contained quartz and feldspar, then by means of calculation it is possible, with sufficient accuracy, to determine both the content of quartz and that of feldspar. In those cases where the minerals do not belong to the feldspars, their amount may be calculated by multiplying the weight of alumina by the coefficient characteristic of the other mineral entering into the composition of the clay. In those cases where the mineral constituent has not been studied sufficiently, it is taken to be feldspar.

That with the aid of rational analysis it is possible to obtain a series of precise results, and that the path indicated by this method is very correct, is proved by a whole series of analyses carried out by Seger¹. According to these determinations, the clayey substance consists of the following relative quantities of silica, alumina, and water:

From these data it is evident that the ratio for kaolins and plastic clays in reality corresponds almost exactly to the formula:

\[ 2\mathrm{SiO_2},\ \mathrm{Al_2O_3},\ 2\mathrm{H_2O}, \]

as had earlier been shown by Forchhammer and again confirmed by Brongniart and Malaguti.

As for potter’s and brick clays, the ratio of $\mathrm{SiO_2}$ to $\mathrm{Al_2O_3}$ does not correspond to the above formula and may deviate from it more or less considerably, in either direction.

In addition to these principal results, Seger obtained many other interesting results with the aid of rational analysis; however, their consideration lies somewhat aside from our main aim, and therefore we shall proceed further to the remark of Dr. Lindhorst, published in the Thonindustriezeitung for the year 1878.

Dr. Lindhorst, referring to Prof. Seger’s proposal to use a mixture of potassium and sodium carbonates for fusing the residue from boiling with sulfuric acid, for the purpose of the quantitative determination of feldspar and silica—

Rational Analysis of Clays

...silica, indicates that it is not always immaterial which alkalis enter into the mineral constituent. In Seger’s method the alkalis are determined from the difference in weight, in the form of the sum of the alkalis. Moreover, in fusion with alkalis, errors are possible in the determination of silica. Therefore Dr. Lindhorst, in order to avoid these errors, proposes decomposing the residue from boiling with sulfuric acid by means of hydrofluoric acid. He further proposes determining alumina and alkalis, while taking silica as the remainder, calculating it by difference. Knowing the alkalis and alumina, it is then easy to convert these data to feldspar; the remaining part of the silica, not included in the composition of the feldspar, is usually counted as quartz.

Dr. Lindhorst’s proposal to use hydrofluoric acid for decomposing the mineral residues and to determine alumina, as the heaviest constituent of feldspar, for calculating the amount of the latter, subsequently became generally accepted. To the present day this proposal remains without any changes.

In this form, rational analysis, in the version proposed by Prof. Seger with Dr. Lindhorst’s correction, is accepted by almost all technical literature.

Although some objections are encountered, they are found chiefly in foreign literature, as, for example, in S. Bogdanov’s article on the Kiev clays, and also in Bourry in his Traité des industries céramiques, 1897, p. 87.

Comparing the data in the literature, S. Bogdanov, in the above-mentioned article, on the one hand points to their inconsistency and, in general, arrives at the following discouraging conclusion: “Thus, at the present time not a single method is known that makes it possible fully and precisely to solve the chief problem of the chemical investigation of clays—to determine their immediate constituent parts.”

It may be that, in general form, such a statement with respect to all clays in general is fair; but there is no doubt that, as is evident from Seger’s observations, rational analysis gives quite suitable and sufficiently accurate results for kaolins and plastic clays.

Here, however, one cannot fail to point out that the very basis of the method was subjected by Seger to insufficiently thorough critical evaluation; therefore inconsistency is found in the literature, and at times even complete distrust of this method. Among those who formerly treated the method with distrust was the well-known earlier authority on clays, Dr. Bischof. However, having later become convinced of the great suitability of the method, he even became the author of its description in Chemisch-technische Untersuchungsmethoden by G. Lunge, 1899 (see No. 5 of the appendix to Chapter I).

7. Study of Prof. P. A. Zemyatchensky, 1896

With the appearance of P. A. Zemyatchensky’s interesting work, “Kaolinite Formations of Southern Russia,” we have a complete collection of references in the scientific literature concerning kaolins and clays in general, and Russian clays in particular.

Certain sections of this work are closely connected with the task I have set myself and touch upon these questions in a quite distinctive way. In the general exposition of such chapters as that on mechanical analysis, I have been obliged to reduce the exposition to a minimum, bearing in mind that those wishing to acquaint themselves with the literature of the subject can do so from the corresponding chapter of P. A. Zemyatchensky’s work¹, which treats “the physical and chemical properties of hydrous silicic alumina in kaolinite formations.”

Chapter II of the same work, entitled “The Relation of Clays to Various Reagents,” has a greater bearing on the question of the rational analysis of clays, and therefore I permit myself briefly to dwell on the most important points that introduce something new into the formulation of the matter.

It is noteworthy that, of all the methods of mechanical analysis, the author fixes his attention on Osborne’s method as the best. Unfortunately, this opinion is not substantiated. The absence of a justification is especially to be regretted because other authors, such as Van Schaffé, known for his activity in the field of mechanical analysis, regard this method unfavorably.

As for the method of rational analysis using sulfuric acid, the author treats it very negatively, explaining this as follows: “if one takes into account what is known concerning the action of the named acid on the most varied minerals: almost all of them are decomposed by it, and, moreover, to a very considerable degree, while some are even decomposed completely, especially if the decomposition is conducted at an elevated temperature and for a prolonged time, as is done in the analysis of clays.”

The influence of other reagents on kaolinites and mineral impurities has been studied very little, and, moreover, the available indications contain many contradictions. Thus Zemyatchensky states that, concerning the action of caustic potash and hydrochloric acid, the most varied statements exist in the literature, some of which assert that these reagents do not act at all on kaolinite, while others assert the opposite.

P. A. Zemyatchensky tests the action of hydrochloric acid on kaolinites taken from different deposits; he comes

in conclusion, that hydrochloric acid in general acts strongly, but this action is especially noticeable if the kaolin has previously been ignited at the temperature of dull-red heat.

The first of these propositions is confirmed by the following observations:

Treatment with 10% hydrochloric acid was carried out under identical conditions, in closed beakers on a boiling water bath for 10 hours, and the following results were obtained:

Further study of the properties of kaolinite leads P. A. Zemiatčensky to a new method of “rational analysis” of clays, based on the action of hydrochloric acid on ignited kaolins. On this occasion a series of examples is cited, which the author concludes with the following words: “the easy decomposability of kaolinites in hydrochloric acid after their preliminary weak ignition makes it possible easily and quickly to carry out the decomposition of any kaolinitic rock. For this we propose to use 10% hydrochloric acid, heating with it, weighed out, the kaolinitic rock in a closed beaker or conical flask on a water bath for 10 hours. The kaolinitic rock must be preliminarily ignited to weak dull-red heat, for which, as our experiments have shown, even one hour is sufficient.”

“When concentrated hydrochloric acid is used, decomposition will, of course, take place much more rapidly, but its use is dangerous, since it may exert a noticeable influence on silicates that may occur in clays, e.g. on feldspars, mica, hornblende, etc. The residue from the filtrate must be treated with a 10% solution of soda while boiling for half an hour, so as not to attack particles of quartz and other minerals that have not decomposed in 10% hydrochloric acid. This method makes it possible to determine rather accurately the quantity of undecomposed minerals present in kaolinitic rocks, chiefly quartz, feldspar, mica, hornblende, garnet, tourmaline, and others. Secondly, in combination with mechanical analysis, it can more closely elucidate the character of the ‘sand’ and ‘dust’ in the composition

which may include, in considerable quantity, particles of hydrous aluminous silicic acid (kaolinite).”

Without rejecting the method proposed by P. A. Zemyatchensky without further verification, I nevertheless cannot agree with his assertion that, by his method, the mineral rocks are affected less than by others, nor with the assertion that more concentrated acids act more strongly on kaolinites. The latter is refuted in Part III of this work. As for the first assertion, that the mineral rocks are little affected after ignition, the classic investigations of Vicat give rise to serious doubt; as early as 1828 he studied the effect of certain reagents on ignited minerals, or so-called pozzolans. The results of his investigation were published by him in the work “Résumé sur les mortiers et ciments calcaires” par L. J. Vicat. 1828, p. 31, in Chapter VII: “L’action des acides sur les puzzolanes naturelles et artificielles,” and in Chapter VIII, “Fabrication des puzzolanes artificielles”¹.

In the latter of these two chapters he states that, at the temperature of dark-red heat, clays that do not in any way react with lime become active toward the latter; at the same time, their resistance to acids decreases to a considerable degree. Consequently, P. A. Zemyatchensky’s conclusion completely coincides with Vicat’s statement, but the whole difference comes down to the fact that P. A. Zemyatchensky asserts, without relying on any experiments, that the changes undergone by the mineral rocks contained in clay are in this case very slight; meanwhile, the conclusion of Vicat is not such: from numerous experiments he arrives at the conclusion that many mineral rocks, upon ignition, pass into the so-called pozzolans and that, in doing so, their decomposability by acids increases to a very considerable degree; that is, relying on a series of experiments, Vicat asserts precisely the opposite of what P. A. Zemyatchensky accepts as an indisputable proposition.

Moreover, confirmation of the same conclusions of Vicat may be found in Fuchs. Fuchs, who investigated many silicates, found that both feldspar and many other silicates, after ignition, acquire pozzolanic properties to a greater or lesser degree; that is, they become readily decomposable by hydrochloric and other acids. From these assertions of eminent authorities it is possible to come to the conclusion that there are more data in favor of the view that the method of rational analysis proposed by P. A. Zemyatchensky leads to greater errors than each of those mentioned above. In any case, the opposite proposition—that the errors in P. A. Zemyatchensky’s method are small—must be regarded as unproven and very doubtful.

8. Study of Vogt from Sèvres¹ and Dr. Kaul from Nuremberg².

In 1898 an interesting study by Vogt appeared in print, under the title: “De la composition des argiles et de l’analyse des argiles dite rationelle,” in which the author thoroughly reviews the material on the chemical properties of clays, relying, however, chiefly on the French literature, and therefore illuminates the problem of rational analysis somewhat differently than had been done by the German school of Seger.

In general, his report falls into 2 parts: the first contains material that may serve as a check on the basic assumptions adopted by the German school for the rational analysis of clays. The question he raises is investigated by him in a far from exhaustive manner; nevertheless, the data he cites deserve great attention. The second part of Vogt’s study concerns the establishment of chemical formulas not for kaolins, but for clays of exceptional composition, such as, for example, certain clays of Chinese and Japanese origin, as well as exceptional clays of French origin, such as, for example, the marly clays from Argenteuil and Aube; for us this part is of less interest.

Below I give, in brief, the content of the part of his report that is of greater interest to us.

At the beginning of his exposition Vogt reports that the original idea of rational analysis belongs to Gmelin, who, in studying the chemical composition of rocks, applied treatment of these rocks first with water, then with hydrochloric acid, and, finally, with sulfuric acid. As applied to clays that do not contain soluble and calcareous salts, treatment with the first two reagents may be omitted without detriment. Hence the origin of Forchhammer’s method becomes clear.

To verify the basic assumptions on which the rational analysis of clays rests, it was necessary to subject crystalline quartz, feldspar, and mica—these simple rocks entering into the composition of the majority of more complex rocks—to the action of sulfuric acid and other reagents used in rational analysis.

Action of acids. Sulfuric acid, heated until excess evaporates, dissolves kaolinite completely. Nitric acid of sp. gr. 1.2 acts more weakly, but when heated to 210° in a sealed tube it dissolves kaolinite within 2 hours. Hydrochloric acid at atmospheric pressure acts considerably more weakly; however, if heat—

... heated to \(70^\circ\)—\(90^\circ\) C for 6 days, then kaolinite, of which 97.4% is soluble in sulfuric acid, dissolves only in the amount of 93.4%. Treating the same kaolin after calcination at \(1600^\circ\) C, they observed a considerably weaker action of the acids; thus, for example, hydrochloric acid dissolves no more than 35.5% of the clay. Different clays behave in the same way toward the action of the reagents cited. Further, Forch points out that clay from Dreux, after ignition at \(1000^\circ\) C, proved, like kaolin, to be less decomposable by hydrochloric acid, notwithstanding Brogniart’s assertion. Before ignition this clay dissolved up to 75.16%, and after ignition only 33.7%.

Feldspar (orthoclase) is almost entirely not decomposed by the action of acids.

On quartz acid likewise has absolutely no action.

White mica, or muscovite, \(K_2O\,Al_2O_3\,6SiO_2\,2H_2O\), is completely decomposed by sulfuric acid heated until vapors are evolved, contrary to what is asserted by most manuals of mineralogy. Nitric acid in sealed tubes destroys mica at \(120^\circ\) C over the course of 2 hours. Hydrochloric acid at atmospheric pressure decomposes up to 85.1% of the mica. After calcination at \(1000^\circ\) C it is less subject to the action of hydrochloric acid. In all cases the mica is taken finely pulverized; in this form it resists the action of acids, like kaolins, very little.

Action of a solution of caustic soda.

1. Kaolin, subjected to the action of caustic soda of sp. gr. 1.08 for 2 minutes at boiling, is strongly affected by it. The solution of caustic soda contains silica and alumina. The decomposition of kaolin under the action of caustic soda proceeds with some approximation proportionally to time. By repeated action of caustic soda it is possible to attain complete decomposition of kaolin.

2. Orthoclase feldspar, treated with an alkaline solution, behaves like muscovite, i.e. it is eaten away very little.

3. Quartz (in the form of powder) and white sand from Fontainebleau, when treated with caustic soda of sp. gr. 1.08, dissolve after 15 minutes’ boiling. No more than 1% \(SiO_2\) passes into the solution. Further dissolution proceeds as though proportionally to time.

4. White mica is changed very little by such treatment; thus, after one boiling of the mica with caustic soda of sp. gr. 1.08, the filtered alkaline liquid contains only barely noticeable quantities of silica and traces of alumina; further repeated treatment with caustic soda has no effect on mica, or muscovite.

5. Magnesian mica—biotite is decomposed by alkali under the same conditions.

With regard to the chemical formula of the kaolinites, Forch made an observation contradicting the text of the memoir of Brogniart and Malagutti (Académie des sciences, Octobre, 1841).

Thus, under conditions similar to those described in the memoirs, it was observed that, when clays were dissolved, the ratio of alumina to silica was always preserved as

\[ \frac{\mathrm{Al_2O_3}}{\mathrm{SiO_2}} = \frac{1}{2}. \]

Brongniart and Malaguti, however, give their formula for kaolinite in the form \(\mathrm{Al_2O_3\ 3SiO_2\ 4H_2O}\), on the grounds that, when clay is boiled with a solution of caustic soda for one minute or no more than a minute and a half, a certain quantity of silica dissolves without traces of alumina.

A new confirmation of the correctness of the ratio in clays of alumina to silica as \(1 : 2\) when dissolved by sulfuric acid may be found in the article by J. M. van Bemmelen¹, Zeitschrift für anorganische Chemie, Bd 22, S. 269, 1904.

Summarizing all that has been set forth, Fogg arrives at the following principal conclusions:

1) Clays and micas (phlogopite, meroxene, as well as muscovite) are decomposed by boiling sulfuric acid; feldspar and quartz resist this action.

2) Clays, kaolinite, and magnesian mica are attacked by a boiling solution of caustic soda of specific gravity 1.08; quartz is attacked slightly; white mica and feldspar resist this action almost completely.

Consequence. Therefore, under the action of sulfuric acid on a clay containing mica, both the clay and the mica are decomposed, and in the alkalis separated from the undecomposed portion, after the separation of silica and alumina, there will be found quantities of potassium, sodium, magnesium, and lime proportional to the quantity of mica contained in the given clay.

On the other hand, by subjecting a mixture of clay and white mica to the action of a boiling solution of caustic potash, it was found that only the clay is attacked by the alkali, while the ratio of \(\mathrm{SiO_2}\) to \(\mathrm{Al_2O_3}\) remains the same as it was in the clays.

Some of the observations presented here are of great interest, although these observations are not numerous enough to regard them as definitively accepted. There are, moreover, grounds for cautious acceptance of some of these results. Almost simultaneously with Fogg’s observations, a very detailed and interesting dissertation by Dr. Kaul appeared in print in 1900, entitled: “Geologisch-chemische Studien über die Thon und Lehmvorkommen um Nürnberg,” in which he describes a series of brick and pottery—

clays; in the mineral residues of these clays he could not, using the microscope, find feldspars, but found only plates of mica. Therefore, when carrying out the rational analysis of clays according to Seger, he makes all his calculations by determining \(Al_2O_3\) in the mineral residues and then recalculating \(Al_2O_3\) as normal muscovite \(H_4K_2Al_6Si_6O_{24}\), or potash mica.

The investigations of Vogt and Cauer are in complete contradiction with one another in that Vogt points to the easy decomposition of various kinds of mica by acids; at the same time Cauer has grounds to suppose that micas remain unchanged under such treatment. Otherwise he would not have been entitled to recalculate the \(Al_2O_3\) found in the residue as muscovite.

As is evident from these comparisons, the question of applying rational analysis to clays containing varieties of mica among their mineral residues remains open.

9. Brief review and comparison of all the enumerated results.

The inadequacy of a single quantitative analysis for solving the problem of the composition of clays was recognized as early as the 1840s of the past century. Aron was the first to introduce into the literature a whole series of practical considerations, the importance of which is beyond doubt, pointing at the same time to the helplessness of chemists in giving a conscious answer to questions about the mineralogical composition of clays by means of a single quantitative analysis. Hence—the impossibility of consciously introducing into practice a whole series of ceramic manufactures. Aron further indicates the path outlined by Gmelin and successfully applied by Forchhammer to the solution of the theoretical problem of the chemical composition of clayey substance.

Professor Seger, who was sympathetic to Aron’s proposal, verifies, on the one hand, the propositions on which the method of rational analysis is based, and subsequently energetically propagates in the literature and himself applies the proposed method in a whole series of articles having both purely theoretical and narrowly practical significance. It must, however, be noted that Seger’s verification was of a somewhat hasty character and, as a consequence, that the critique of the foundations of the method was incomplete.

In the Notizblatt for 1875, Bischoff comes forward as a sharp opponent of rational analysis, calling this method too crude and associated with a whole series of inevitable errors. Subsequently, however, as is evident from Prof. Lunge’s compendium on technical analysis, Chemisch-technische Untersuchungsmethoden, 1899, he changes his opinion and decides even to undertake the editorship of the section on clay analysis

...in the above-mentioned book, where he describes the method of rational analysis, although still with some reservations at the end.

The method is set forth by him (see No. 5 of the appendix) without substantial changes in the form in which it had been proposed by Aron, supplemented and developed by Seger, with corrections introduced by Lindhorst.

If attention is paid to the critical evaluation of the method and to the elaboration of the details of the work performed, it will appear that, among the majority of specialists, the details are worked out in separate laboratories without discussion of these methods in the literature. In confirmation of what has been said, it is enough to enumerate the various descriptions of the method, differing from one another in details): Dr. Bruno Kerl. Die Thonwaarenindustrie, 1879, s. 57; D-r Post. Chemisch-technische Analyse, 1890 and 1897.—D-r Seger. Beckmans „Chemisch-technische Untersuchungsmethoden“, III Auflage, 1893.—D-r Hecht. Chemische Technologie von Dammer, 1895.—D-r Bischoff.* Chemisch-technische Untersuchungsmethoden von D-r G. Lunge, 1884 and 1899, and some other descriptions (see the appendices to Part I).

All these descriptions are distinguished by brevity, without proper indications of those special features which must be observed in order to obtain correct results. Descriptions of particular special cases, as, for example, in the case of carbonaceous clays, are not given at all. The absence, even in many dissertations, of duplicate determinations (controls) suggests that the criticism of the fundamental propositions has still been insufficiently developed for it to be possible to work by this method with complete confidence in its correctness.

The most recent works of Prof. Zemyatchensky indicate a considerable inaccuracy of the method (although, incidentally, without confirmation of this by experiment) in the event that sulfuric acid is used. His proposal to use hydrochloric acid with preliminary ignition of the clays contradicts the notion, long established thanks to the works of Vicat, of the considerable decomposability of mineral rocks by acids after ignition of kaolinite.

Foth asserts that micas are decomposable by sulfuric acid. This assertion is in complete contradiction with the assertion of Kayul.

Even in special laboratories the rational method is practiced not as an exact one, but as an approximate one, which at times gives an error that in extreme cases reaches 4%.

All that has been said above creates a picture on the basis of which it is difficult to give oneself an account of whether these errors are conditioned by the foundations of the method itself, or whether the procedures for carrying out the analysis have not yet been sufficiently developed in detail to always lead to the same invariable results.

*) At the end of the article, in the form of appendices, the best-known texts of descriptions of rational analysis are given.

In the second part of my work, I set forth my attempts to make sense of many of the contradictions cited above. First I present my attempts to verify the foundations of the method, as well as the degree of accuracy and the limits of its applicability. I then study the conditions under which considerable accuracy of the method is achieved; for this purpose it is necessary to work out the details of the practical implementation of the same method, while at the same time it will be possible to note certain particular features in individual cases.

1. J. M. Bemmelen. Beiträge zur Kenntniss der Verwitterungsprodukte der Silicate in Thon,-vulkanischen und Lateritböden. ↩↩↩↩↩↩↩↩↩↩↩↩

2. Dr Herm. Kaul. Inauguraldissertation, 1900: Geologisch-chemische Studien über die Thon-und Lehmvorkommen um Nürnberg. ↩↩↩↩

3. Segers gesammelte Werke, p. 33. ↩↩

4. See above. ↩

5. Proceedings of the St. Petersburg Society of Naturalists, vol. XXI, part 2, 1896. ↩