Experimental evidence based on the peak area measurements of the endothermic reaction of kaolinite and also (kaolinite): (inert material) mixtures was produced to show that the use of an optimum amount of material (200 milligrams for the experimental set-up used) gives the nearest approach to the quantitative reproduction of peak areas in relation to the mass of active material. It was also shown that such an optimum amount of active material increases to a maximum the ability of the method to semi-quantitatively resolve reactions with the same sign which occur at closely separated temperatures. The experimental evidence refers to the resolution of the two low temperature moisture peaks of calcium montmorillonite between 100-200°; and also the pyridine-montmorillonite system.

Montmorillonite is not uncommon in the china clay of Devon and Cornwall but occurs in such small proportions that few commercial samples will give X-ray diagrams showing a line at 15-18 Å, in the oriented aggregate technique of Nagelschmidt (1940) and MacEwan (1946). Particle size fractionation is an essential starting point for the study of these clays, montmorillonite concentrating at the fine end. The sensitivity of the X-ray method may be greatly increased by a preliminary concentration in this manner. However, there are many clays with properties which suggest that a trace of montmorillonite is present, but in which it cannot even be identified by X-rays. If the fractionated clays are converted to the H-clay and the cation exchange capacity of the fractions measured by titration with alkali the presence of montmorillonite is shown by the much higher exchange capacity of the finest fractions. In this way montmorillonite may be detected, and roughly estimated, when the proportion present is as low as 0.2 per cent. of the total clay.

In responding to the kind invitation of the Clay Minerals Group, may I say that my notes are intended to express appreciation by a geologist of the illuminating work of the Group, as well as a plea for many favours to come?

History repeats itself. Forty years ago, when the coarser detrital and authigenic constituents of sediments were first being systematically examined, it became evident that their significance could not be assessed until we knew, in broad outline at least, the characters of the assemblages in all the major stratigraphical divisions. The more finely-grained materials were at that time in the main intractable: their investigation has had to await the development of the now familiar techniques utilized by the Group.

For some time research workers have directed their attention to the study of ferriferous minerals. Researches on iron ores have led us to take up this problem. We shall give here the results of our observations on some minerals considered to be pure, belonging to the collection of the Museum National d'Histoire Naturelle, and on samples extracted by sedimentation from various ores.

This shows that the different samples belong to three types, with 10Å, 14Å and 7Å spacing. We will consider only the last two.

1. 14Å type (chlorite).—This includes the Chamoson chamosite, the bavalite from Bas-Vallon, metachlorite from Radauthal, and the chlorites extracted from ores from Soumont (Calvados), Auboue, Angevilliers, Andun-le-Tiche, Godbrańge (Lorraine).

One of the most outstanding contributions to petrogenesis and a continuing basis for petrologic argument is the classic reaction relation first propounded by Bowen (1922, 1928). This pregnant summary of petrogenetic process, included in a reaction series, has received affirmation anew from the more recent approach of crystal chemistry and atomic structure of the minerals emphasizing the concept of continuous and discontinuous reaction (Brammall, 1936). The evidence upon which this reaction relation was initially proposed is essentially twofold, firstly a thorough acquaintance with the regular association of certain minerals, particularly the essential and varietal minerals, in igneous rocks and this stemmed mainly from microscopic study, whereas the second line of evidence comprises the guidance received from laboratory synthesis of mineral aggregates (simplified rocks) from prepared melts.

During investigations on clays and minerals, of which only small amounts were readily available, it became necessary to find an accurate micro-method for determining cation-exchange capacities. Usual methods, which are far too numerous to detail here (see, e.g., Kelley, 1948) required considerably more than the 10-50 mg. of material that could be spared. Three possible methods were tested, viz., (a) an adaptation of the manganese method of Bower and Truog (1940 a); (b) a method employing saturation with strontium and utilising spectrographic determination of the sorbed strontium; (c) an adaptation of the usual ammonium acetate method (Schollenberger and Dreibelbis, 1930; Jackson and Truog, 1939).

The results obtained by the first two methods were usually somewhat high—not surprising, as polyvalent ions would have difficulty in bridging exchange spots any distance apart and might consequently be sorbed as M(OH)+ rather than as M++ (Bower and Truog, 1940 b). Method (c) gave results in very good agreement with those obtained with the macro-method normally employed and was the one finally adopted. A complete description and assessment of the method is at the moment in the press (Mackenzie, 1951) and only a brief indication need be given here.

Amongst hitherto described clay minerals, there are a number which have not yet been placed in clearly defined categories. In the course of an investigation of antigorites (Caillere, 1936), one of us has already shown that the thermal curves of bowlingites from Blackburn, Linlithgowshire, and Catkin Hills, Lanarkshire, of saponite from Catkin Hills and of diabantite from Bergen Hill, New Jersey, U.S,A., are very similar to the thermal curves of montmorillonites. We have therefore taken up again the study of these minerals with a view to determining more precisely their nature. We may recall that prior to the work already mentioned (Caillere, 1936), bowlingite and saponite were classed with antigorites, while the mineral from Bergen Hill, regarded as a diabantite, was classed with the chlorites.

In a series of test tubes, increasing amounts of pure cetyltrimethyl ammonium bromide (Lissolamine A) are added to a dilute clay suspension containing a known amount of clay. Visual examination of the contents of the tubes shows a stability minimum to exist at a certain concentration of the bromide. It may be assumed that in this stability minimum the clay micelles are neutralized by quantitative exchange of the exchangeable ions of the clay by an equivalent amount of strongly adsorbed organic cations of the bromide. Consequently the b.e.c, of the clay may be calculated from the concentration of the cation-active compound at the stability minimum. Since, however, by visual examination the stability minimum cannot be distinguished sharply enough, a better criterion had to be developed. As there are reasons to assume that the neutralized micelle will act as an emulsifier for oil-water systems, an emulsification test was expected to offer a way out. It was indeed found that on shaking the contents of the tubes with gas-oil, a rather stable emulsion was formed at a certain critical concentration of the cation active compound. Taking this concentration as an indication of quantitative exchange, the b.e.c, of the micelle may be easily calculated. Results of preliminary experiments are in good agreement with those obtained by application of conventional methods.

The structures of vermiculite and montmorillonite are very similar in that they consist of talc-like or pyrophyllite-like layers interleaved with layers of water molecules, the latter carrying exchangeable cations. In vermiculite, the exchange positions are normally filled by Mg ions, and the water layers consist of double sheets of water molecules with the exchangeable Mg ions presumably centrally placed. The water layers in natural Mg-saturated vermiculite are 5.1 Å high and it seems probable that this distance is fixed by the size of the hydration shells surrounding the exchangeable Mg ions as one of us (G.F.W.) has previously suggested.

Although vermiculite is found in particles of clay size, it also occurs as large flakes produced by the hydrothermal alteration of mica. This seemed a good reason for making a Fourier curve of the mineral in the c-direction with the object of examining the organization of the water layers. It was hoped that greater detail might be obtained from single crystals of vermiculite than would be possible in the case of montmorillonite, where oriented aggregates consisting of large numbers of very small crystals must of necessity be used.

The kieselguhrs or diatomaceous earths are composed essentially of more or less well crystallised silica, as well as a small quantity of alumina, iron oxide, and traces of other oxides. These earths are largely used as filtration aids; they can serve as supports for certain catalysts, and have also excellent calorifugic properties.

A study of different sorts of kieselguhrs used as supports for nickel and cobalt catalysts in the Fischer-Tropfsch synthesis has recently been carried out by Anderson et al. (1947). These authors find that a previous treatment of the raw kieselguhr by hot hydrochloric and nitric acids gives more active catalysts than those obtained using the same support untreated, and especially than those prepared from calcined kieselguhrs.

An exact knowledge of the chemical constitution of the surfaces of clay minerals is necessary to understand their properties. For this purpose as many methods as possible should be applied. One of these methods is the preparation and characterization of organic derivatives of clay minerals. Thus for example it ought to be possible to determine surface Si—OH groups.

The preparation of organic derivatives of clay minerals is not astonishing at all. Already in the last century Ebelman in 1846, Friedel and Crafts in 1863 and others prepared molecules containing carbon and silicon. Most outstanding is the work of Kipping accomplished between 1899 and 1944.

A series of samples of “metabentonites,” here referred to as K-bentonite, and related limestones were collected from the Upper Black River and Lower Trenton formations of the Ordovician System in Central Pennsylvania. The K-bentonite occurs in yellow and gray beds which range from half an inch to ten inches in width.

X-ray diffraction patterns of the 22 K-bentonite samples are similar to those for illite and have the (060) reflection at 1·50Å indicating that the clay mineral is of the dioctahedral variety. However, the value of the (001) reflection ranges from 10·2 to 10·6Å, suggesting that the material is a mixed layer clay composed of expanded and non-expanded layers of hydrous mica (montmorillonite and illite). This possibility was tested by saturating the clay with H, Ca, Na, K, NH4, and ethylene glycol and noting the effect of the treatment on the position of the (00l) spacings in the X-ray diffraction patterns. Using the curves published by Brown and MacEwan (1950) for the analysis of mixed layer complexes, it was found that approximately 20% of the layers are expanded and 80% are not. This ratio is supported by chemical analyses which show that of the exchangeable cations 20% are calcium, sodium, and magnesium and the remaining 80% is potassium.

The supposed analogy between hydrogen montmorillonite and a weak organic acid has encouraged several investigators, notably Berger and Deuel, to attempt to esterify the exchange positions using standard organo-chemical methods.

In our opinion the experimental evidence so far presented to establish the course and extent of the esterification reaction is inconclusive. Thus Berger determined the methoxyl content of the product obtained by treating dry hydrogen montmorillonite with diazomethane in ether without showing that sorbed methyl alcohol was absent. This omission is particularly relevant when Berger states that he dried his starting material at 150°C. Data on the dehydration of hydrogen montmorillonite has shown that under these conditions Berger's sample probably contained more than 2% sorbed water. Taking into account the difficulty of washing montmorillonite free of methyl alcohol, this amount of water would be sufficient to hydrolyse enough diazomethane to account for the observed methoxyl contents.

A talc or talc-like unit forms part of the structure of mica, illite and hydro-mica, vermiculite, montmorillonite and chlorite as well as tale and pyrophyllite. Transitions should be possible between these structures by chemical and/or thermal methods. Many such transitions have been observed in nature but comparatively few successful transitions have been carried out in the laboratory.

One aspect that has recently been under investigation at Leeds University Physics Department is the transition chlorite→mica. Treatment of a penninite by various acid media either with or without heat treatment failed to remove the brucite layer from the chlorite without also attacking the talc layers. No mica or talc-type of mineral was observed as an intermediate product, and a decrease of only 0.4 kX occurred in the spacing of the layers during acid attack.

Although it has been known empirically for some considerable time that the area under the peak of a thermogram (i.e. the area BCD on Figure 1) is approximately proportional to the amount of reactive constituent in the sample, only some five years ago was a simplified theory accounting for such a relationship proposed. This theory is, however, at best an approximation, and recently Arens has attempted a more rigorous mathematical treatment of some of the problems involved. Some incorrect assumptions and derivations in his treatment unfortunately rather vitiate the final result, and, in view of the wide interest likely to be exhibited in this theory, it is perhaps not out of place to draw attention to the inaccuracies here.

During an extensive survey of brickmaking clays by the differential thermal analysis method (carried out from 1947 onwards) it was observed that the clays of the Keuper Marl gave a series of very similar thermal curves which could not be explained by the presence of known clay minerals. Dr D. M. C. MacEwan of Rothamsted Experimental Station kindly carried out an X-ray examination of one of the samples for us, and he has reported the results of this examination, and the examination of a Keuper Marl sample from another source, to the last meeting of this Group. He postulated the existence of a “hydrated chloritic material”—sometimes exhibiting an expanding lattice stabilised by glycerol.

Since then the Keuper Marls have been further studied at the Building Research Station both by thermal analysis and X-ray methods and this paper is an interim report of this work. MacEwan's findings have been confirmed and some additional information is presented.