CEMENTand CONCRETERESEARCH. Vol. 9, pp. 747-755, 197g. Printed in the U.S.A. 0008-8846/79/060747-09502.

00/0 Copyright (c) 1979 Pergamon Press, Ltd.

ELECTRON MICROPROBE ANALYSIS OF PORTLAND CEMENT CLINKERS A. Ghose and P. Barnes Department of Crystallography, Birkbeck College, University of London, Malet Street, London WCIE 7HX.

(Communicated by H.F.W. Taylor) (Received Sept. 5, 1979)

ABSTRACT Using modern electron microprobe techinques, we have undertaken a detailed composition analysis of all the major phases in portland cement. A preferential incorporation of certain elements into either alites or belites is confirmed, and compared to existing ideas on alite decomposition. In addition the composition data are discussed in terms of solid-solutions existing in the various phases; in particular there is evidence of a further solid solution in the interstitial phase between C3A and the alumino-ferrites. Nous avons entrepris, en utilisant des techniques modernes de microscopie electronique, une analyse d~taill~e de la composition de toutes les phases majeures du ciment de Portland. L'incorporation pr~f~rentielle de certains ~l~ments soit dans les alites, soit dans les belites, est confirm~e et compar~e aux idees actuelles au sujet de la d~composition des alites. On exam4ne aussi les donn~es de la composition en tant que solutions-solides existant dans les diff~rentes phases; en particulier on constate l'existence d'une solution solide suppl~mentaire dans la phase intersticielle entre C3A et les alumino-ferrites. Introduction The use of energy dispersive systems (EDS) as a means of fairly rapid chemical composition analysis of cement phases has already been established. (1,2,3,4,5). The techniques can however be further i-T.roved from both the points of view of quantitative accuracy and certainty of phase delineation. This progress is reported here as well as the application of these techniques to a statistical chemical study of all the major portland cement phases, together with a discussion of its implications on the chemistry of cement formation. Experimental Specimens were cut and polished using standard techniques (6), due care being taken regarding the planarlty of the surface under study. An additional preparation feature was the use of light ion-etching (described in

747

748 A. Ghose, P. Barnes

Vol. 9, No. 6

detail elsewhere (7)) to enhance grain deliniation in the Scanning Electron Microscope (SEM) image and to clean the surface without any effective loss of planarity. In this study a Cambridge Stereoscan SII SEM was used in conjunction with an EDAX 707A energy dispersive analyser. The spectral intensities obtained were compared with those of pure standards using a m o d i f i e d version of a single standard method (8) and converted to weight percentage results using a modified (9) version of the standard ZAF/FRAME correction routines (I0). Because of inherent difficulties in defining (ii) the exact effective X-ray take off angle with EDS-systems and also uncertainties regarding the corrections for X-ray absorption by the specimen, systematic errors of analyses (particular to a given composite material) can arise up to a few percecnt. In an attempt to quantify and therefore remove these systematic errors, a careful study was first performed on pure synthetic preparations of stoichiometric C2S, C3S, C~A, C~AF and C6A2F to cover the range of cement material. Table i summarises the results and indicates the relevant corrections applicable with our system; as it turns out the largest correction necessary is -2.7% for calcium levels in C3S phases. TABLE I

Analysis of Synthetic Phases No. of grains analysed, 115 174 20 Random Error O//N 0.07 0.02 0.09 0.03 0.09 0.06 0.14 0.14 0.23 0.12 0.09 0.15

Phase C3S C2S C3A

Element Ca Si Ca Si Ca A1 Ca A1 Fe Ca A1 Fe

Theor. wt.% 52.66 12.30 46.54 16.31 44.50 19.97 32.99 II.10 22.98 34.35 15.42 15.95

Expt. av. wt.% 55.35 13.16 47.43 16.99 45.88 20.00 34.89 11.25 23.20 36.23 15.73 16.21

~ 0.72 0.214 1.22 0.37 0.39 0.29 0.89 0.87 1.35 0.76 0.55 0.91

Systematic Deviation +2.69 +0.86 +0.89 +0.68 +1.38 +0.03 +1.90 +0.15 +0.22 +1.88 +0.31 +0.26

C~AF

39

C6A2F

39

Discussion In applying these results to real cements, the C2S/C3S corrections per Table I were applied to belites/alites respectively and linear interpolations applied between the C3A/C~AF/C6A2F corrections (according to the F-content) for the interstitial phases. N o c o r r e c t i o n s were applied for other (minor) elements. All grains studied were of dimensions > 5~m so as to minimise errors of sampling due to electron beam-spread (12) and secondary fluorescence. The results are discussed in two sections: Alites and Belites The results for the full compositional analyses are given in Table II (b) and (a), showing the obtained in various types of cement clinker. minor elements are incorporated preferentially of alite and belite grains range of weight percentages We have observed that certain into alites or belites and it

< o TABLE II(a) z o o~ Fe Nil 0.12 Nil 0.06 Nil 0.16 0.02 0.08 Nil 0.14 0.04 0.26 0.22 0.70 Nil 0.04 Nil 0.07 0.11 0.40 0.28 0.94 0.02 0.43 0.02 0.09 0.02 0.30 1.15 1.54 0.20 0.43 0.11 0.39 0.14 0.80 0.03 0.16 0.04 0.26 0.06 0.11 0.03 0.09 0.03 0.18 Nil 0.07 0.02 0.08 Nil 0.12 Nil 0.05 Nil 0.07 0.02 0.06 Nil 0.10 Nil 0.09 Nil 0.06 Nil 0.06 Nil 0.10 Nil 0.12 Nil 0.06 0.02 0.05 Nil 0.12 0.02 0.06 Nil 0.05 Nil 0.10 Nil 0.13 Nil 0.05 Nil 0.10 Nil 0.04 Nil 0.07 Nil 0.12 Nil 0.11 0.03 0.72 0.08 0.32 0.21 0.87 0.14 0.55 0.33 0.59 0.12 0.53 0.07 1.03 0.40 0.63 0.22 0.55 0.02 0.07 0.12 1.16 ..j
4~

Range of Elemental Composition of Belites in commercial clinkers (the two values denote the range of percentages by weight) (Cements of the same type from different manufacturers have been banded together in the following tables) Si 12.98 15.64 13.66 14.97 12.85 14.08 13.28 15.31 12.57 13.89 13.21 15.33 13.25 16.35 14.02 14.98 13.06 14.14 13.72 14.83 13.17 15.76 12.98 16.35 0.09 0.61 0.13 0.61 0.06 0.47 0.04 0.74 0.16 0.23 0.04 0.18 0.91 1.68 0.17 0.31 0.08 0.27 0.35 1.28 0.09 0.20 0.09 0.29 0.90 1.08 0.06 0.15 0.12 0.37 0.09 0.38 0.07 0.74 0.51 1.61 0.07 0.33 Nil 0.39 0.10 0.29 0.07 0.27 0.73 1.32 0.10 0.29 0.05 0.44 0.04 0.22 0.10 0.24 0.09 0.30 1.02 1.48 0.02 0.33 0.50 1.21 0.05 0.22 0.11 0.38 0.07 0.34 0.53 1.45 0.05 0.34 0.02 0.58 0.10 0.81 0.03 0.13 0.13 0.50 0.09 0.52 0.57 1.36 0.08 0.36 0.05 0.61 0.06 0.48 0.05 0.14 Nil 0.10 0.14 0.37 0.07 0.32 0.71 1.27 0.09 0.58 Nil 0.10 0.29 0.44 0.04 0.15 0.02 0.06 0.12 0.60 0.07 0.42 0.37 1.69 0.03 0.44 Nil 0.14 0.10 1.15 0.02 0.23 Nil 0.10 Nil 0.18 Na Mg A1 P S K Ti V Cr Mn

Cement Type/Process

Ca

OPC/Wet Process

43.25 47.17

nr c~ -D O m z -< tn L~
w r-

OPC/Wet Process Low Flux

44.24 45.54

OPC/Wet Process High FerriticPhase

44.81 46.47

OPC/Dry Process

44.56 46.91

OPC/Semi-Dry Process

45.04 47.09

c~

OPC/Semi-Dry Process, High Flux

45.03 46.51

m ~o tn ov

OPC/3 West German clinkers

43.23 45.88

OPC/Libyan clinker

42.96 44.05

OPC/Wet Process PFA and shale Flux added White Cement

45.19 46.51

o vco z t~

45.53 46.86

Sulphateresisting Cement

44.71 47.83

Overall

42.96 47.83

0.28 1.69

0.02 0.58

Nil 1.21

0.04 1.54

0.02 0.26

Nil 0.12

Nil 0.18

Nil 0.12

0.02 1.16

T A B L E II(b)

ul o

Range of Elemental C o m p o s i t i o n of Alites in commercial clinkers (the two values denote the range of p e r c e n t a g e s by weight) Si 9.76 11.95 10.67 11.22 10.25 11.65 10,37 11.92 0.12 0.48 0.21 0.83 0.62 1.10 0.09 0.26 Nil 0.12 0.02 0.09 0.02 0.10 Nil 0.10 0.12 0.42 0.15 0.50 0.48 0.95 0.11 0.29 Nil 0.11 0.03 0.13 0.03 0.15 Nil 0.11 0.17 0.34 0.17 0.67 0.55 1.43 0.09 0.44 Nil 0.04 0.03 0.09 0.02 0.16 Nil 0.09 Nil 0.08 Nil 0.10 Nil 0.08 0.12 0.40 0.09 0.85 0.38 2.96 0.03 0.62 Nil 0.09 0.03 0.14 0.02 0.24 Nil 0.09 Nil 0.12 Nil 0.09 Nil 0.12 Nil 0.05 Nil 0.11 Na Mg A1 P S K Ti V Cr Mn Fe 0.02 1.16 0.08

Cement Type/Process

Ca

OPC/Wet Process

47.18 52.63

OPC/Wet Process Low Flux

47.97 50.79

0.44
0.08 0.66 0.04 0.30 O Ln

OPC/Wet Process High F e r r i t i c P h a s e

49.97 52.42

OPC/Dry Process

49.91 52.09

OPC/Semi-Dry
10.07 11.15 8.31 11.92 10.76 11.77 10.99 11.78 9.62 10.42 10.60 11.94 9.96 12.12 8.31 12.12 0.07 0.48 0.08 0.98 0.13 0,35 0.08 0.61 0.36 1.25 0.34 2.96 0.16 0.47 0.08 0.22 0.61 1.10 0.22 0.32 0.30 0.37 0.69 1.42 0.18 0.37 0.07 0.27 0.03 0.34 0.03 0.62 0.14 0.26 0.64 0.94 0.64 0.86 0.08 0.13 0.01 0.04 Nil 0.04 Nil 0.06 Nil 0.08 Nil 0.12 0.07 0.46 0.10 0.98 0.34 0.88 0.04 0.32 Nil 0.012 0.04 0.39 0.08 0.37 0.03 0.12 0.04 0.10 0.03 0.14 0.02 0.39 0.12 0.32 0.19 0.58 0.53 2.26 0.07 0.33 Nil 0.09 0.03 0.09 0.08 0.44 0.58 0.89 0.66 1.07 0.07 0.27 0.02 0.15 0.03 0.09

Process

47.76 51.34

Nil 0.18 0.04 0.10 0.02 0.16 0.07 0.12 0.06 0.09 0.02 0.09 0.02 0.12 0.02 0.24

Nil 0.I0 0.02 0.08 Nil 0.07 Nil 0.08 Nil 0.08 0.03 0.08 Nil 0.11 Nil 0.11

Nil 0.11 Nil 0.08 Nil 0.14 Nil 0.09 Nil 0.07 0.02 0.07 Nil 0.12 Nil 0.14

Nil 0.09 Nil 0.06 Nil 0.09 Nil 0.06 Nil 0.06 Nil 0.07 Nil 0.09 Nil 0.12

0.14 0.44 0.03 1.85 0.04 0.40 0.17 0.39 0.20 0.79 0.04 0.06 0.12 1.73 0.02 1.85

OPC/Semi-Dry Process, High Flux

50.32 52.94

OPC/3 West German clinkers

48.01 51.83

OPC/Libyan clinker

46.40 49.48

OPC~Jet Process PFA and shale added

49.01 51.38

White Cement

49.17 52.51

Sulphateresisting Cement

47.66 52.04

O ~D Z O O~

Overall

46.40 52.94

< o ._J TABLE II(c) 2 o

Range of Elemental Composition in the Interstitial Phase in commercial clinkers (the two values denote the range of p e r c e n t a g e by weight) A1 Na
Mg P S K Ti V

Cement Type/Process 7.97 15.38 11.69 13.88 0.18 0.43 0.11 1.18 0.21 0.49 0.17 0.42 0.14 0.70 0.14 1.56 0.23 1.17 0.21 0.44 1 .06 2.74 0.69 1.38 0.49 1.00 0.64 0.14 0.65 0.11 1.62 0.61 7.73 0.55 5.55 0.24 14.65 0.58 14.65 Nil 0.20 Nil 0.10 0.03 0.12 0.09 0.26 Nil 0.14 Nil 0.30 0.58 5.40 Nil 0.11 Nil 0.09 Nil 0.14 Nil 0.07 Nil 0.07 Nil 0. I0 Nil 0.75 Nil 1.62 1.45 2.50 Nil 0.07 0.03 1.62 0.24 I .94 Nil 0.25 Nil 0.42 0.03 0.74 0.05 0.77 0.04 0.34 0.04 2.59 0.19 1.57 0.04 0.18 0.08 0.47 0.03 1.54 Nil 2.59 1.07 3.20 Nil 0.12 Nil 0.07 0.04 0.38 0.07 0.73 0.12 0.81 0.13 0.49 0.11 0.37 0.07 0.75 0.12 0.43 0.16 0.48 0.03 0112 0.05 0.37 0.03 0.90 0.66 I .35 Nil 0.22 Nil 0.04 0.18 0.57 0.26 0.40 7.37 13.33 10.29 16.29 10.50 13.70 12.59 14.34 9.31 15.26 9.00 12.30 10.69 13.52 13.35 16.46 6.19 0.30 0.69 9.36 21.07 0.30 21.07 0.61 3.17 2.43 4.82 7.05 12.25 1.61 3.29 7.66 15.19 1.84 2.77 2.26 13.51 1.51 4.82 5.67 7.29 1.72 2.67 6.22 10.31 1.69 2.65 3.75 14.58 1.45 2.73 6.44 17.11 1.52 2.34 4.45 5.52 1.83 2.93 1.56 15.15 1.06 7.73 0.11 1.62 0.30 6.92 Nil 0.30 Nil 0.09 Nil 2.33 0.04 0.90 Nil 0.11 Nil 0.09 Nil 0.10 0.03 0.06 Nil 0.08 0.03 0.08 0.02 0.10 Nil 0.10 Nil 0.07 0.02 0.08 Nil 0.08 Nil 0. II

Ca

Fe

Si

Cr Nil 0.07 0.02 0.04 Nil 0.06 Nil 0.08 Nil 0.08 Nil 0.06 Nil 0.07 Nil 0.08 Nil 0.06 Nil 0.05 Nil 0.08 Nil 0.08

Mn Nil 0.30 0.04 0.21 Nil 0.66 Nil 0.80 Nil 0.18 0.03 0.10 Nil 0.48 Nil 0.18 0.12 0.22 0.03 0.08 Nil 0.40 Nil 0.80 ~4 tn .-a C~ ;O O -O m

OPC/Wet Process

29.69 43.18

OPC/Wet Process Low Flux

35.33 38.01

OPC/Wet Process High F e r r l t l c P h a s e

31.33 36.10

OPC/Dry Process

32.03 40.40

Z F" L~ t~ C~
r-

OPC/Semi-Dry Process

31.84 34.84

OPC/Semi-Dry Process, High Flux

34.74 39.55

OPC/3 W e s t German clinkers

28.82 38.68

m O~ t~ O V-L~ O V-

OPC/Libyan clinker

29.95 33.47

35.68 40.87

OPC/Wet Process PFA and shale added White Cement

38.84 43.05

0.11

Z L~

Sulphateresisting Cement

28.68 39.14

11.10
6.19 16.46

Overall

28.68 43.18

752 A. Ghose, P. Barnes

Vol. 9, No. 6

is interesting to discuss these preferences from the point of view of the decomposition kinetics of alite into belite and CaO. K. Potassium is found to be far more abundant in belites (0.04 to 1.5%, average 0.42%) than alites (0.02 to 0.4%, average 0.07%) in keeping with the proposition ([3) that K promotes the dissociation of alite into belite and CaO. S. Sulphur in molten form in the eutectic K2SO ~ - CaSO~ is thought (13) to be the main controlling factor in the dissociation of alite. Although general sulphur levels detected in the clinkers in our study are low, some higher levels were detected in belites (up to 1.2%) from certain cements unlike alites (up to 0.12%), again supporting the above ideas on alite dissociation. Fe,P. Iron (13) and phosphorus (14,15) (in amounts > i%) are to accelerate the dissociation of alites, although our results preference for their incorporation into belites, though in the phosphorus this could be because the critical I% level had not also show case been believed no clear of reached.

Mg. By comparison, magnesium is believed (13) to have a stabilising effect on alites and this agrees with our observations of a small preference for Mg in alites (0.08 to 0.9%, average 0.42%) compared to belites (0.04 to 0.74%, average 0.18%). AI, Na. Aluminium and sodium are not thought capable of markedly altering the alite:belite ratio in clinkers (a slight retardation of alite dissociation in the case of Na (13)), and our results also show no clear preference. Lattice substitutions into alites and belites

The proposed substitution schemes found in the literature for alites are varied: Mg 2+ is believed to replace Ca 2+ while both Fe 3+ and Al--~-~are capable of substituting for either Ca 2+ or Si ~+, (16,17,18). Another point of view is that of coupled substitution (19) of A13+ and Mg 2+ for Si ~+. In the case of belites, the general view (20,21,22,23) is that of Na +, Mg 2+ and K + replacing --T+ and A13+ and Fe 3+ replacing Si ~+ whereas ~ divalent Fe 2+ can replace Ca 2+ in a reducing atmosphere (20). Group substitutions of (SiO~) ~- by (PO~) 3- and (SO~) 2- have also been proposed (23,24). An investigation of proposed substitution schemes was initiatied using the composition data described above. A computer analysis was made of all the alite/belite results, recasting the data (-into I00 C3S/C2S gram atom proportions) and comparing the fits obtained using various substitution schemes with the stoichiometric value. For example, a deviation of 0.5 gram atom of Ca and substituting ions for the stoichiometric value of 28.57 in i00 gram atom pure C2S would be considered a reasonable fit. Bearing in mind the above discussion of the literature and the accepted rules concerning ionic radii and charge, the general substitution scheme (i) Na, Mg, K substituting for Ca (ii) AI, P, S, Fe substituting for Si was written into the analysis, tl~ugh of course in practice some interstitial substitution would also take place to minimise the charge imbalabce of the oxygen network. The results showed a generally acceptable fit (at least 70 to 80% within 0.5 gram atoms of the stoichiometric value) with Na, K, Mg substituting for Ca though in a number of cases the fit could be improved by alternatively substituting Mg for Si. AI, P, S, Fe substituting for Si also yielded acceptable fits though occasionally with an improvement

Vol. 9, No. 6 MICROPROBEANALYSIS, CLINKERS, SOLID SOLUTIONS

753

&

b A

FIG. 1 Phase diagrams determined for the interstitial matrix (points) using two different substitution schemes (1 (a) and (b) - see text above) and also without any substitution a s s ~ e d (I (c)) . The open circles indicate the position of particular synthetic preparations as indicated in (c).

(particularly with alites) by dividing the A1 substitution between Si and Ca. We feel, however, that in spite of the extensive precautions taken in our own composition analyses (described earlier), we cannot seriously attempt to draw any further definite conclusions than those stated above using present SEM microanalysis methods. Interstitial ~L1trix The composition data for the interstitial phases have been plotted (in proportions of C, A and F) on phase diagrams as in Figure I. Figure I(c) shows the density of phases obtained using just the correction levels obtained from a comparison with the synthetic preparations (Table I). Phase diagrams 1 (a) and i (b) have been obtained with the following substitution schemes: (1) (2) (3) (4) (5) as (i) above P, S substituting for A1 Ti, ~h~ substituting for Fe (a) Si substituting for A1 (refer to Figure l(a)) (b) Si substituting for Fe (refer to Figure I (b))

Comparisons of figures I (a) and i (b) suggest that the 2-3% Si consistently found preferentially substitutes for AI, although even then a complete match

754

Vol. 9, No. 6

A. Ghose, P. Barnes
with the s~/nthetic series is not obtained. ~ e t h e r this is a real effect or just represents the limitations of our correction proceedure is not clear. Another alternative explanation concerns the p o s s i b i l i t I, of intergrowth of very fine grains of C2S with the C3A and alumino-ferrites (25) thouqh the consistency of the results argue against this. Regardless of the precise reason the well defined range of experimental results parallel those for the synthetic alumino-ferrite series, and so it w o u l d appear reasonable to discuss the solid-solution ranges obtained in actual cements. In white cements the interstitial matrix is easily identifiable w i t h the C3A terminus of the results range, and a l t e r n a t i v e l y for sulphate-resisting clinkers w i t h the "C6A2F-C6AF2" part of the range. In the other cements there appears to be evidence for a further solid solution between C3A and the alumino-ferrites. A caution should be added here that the intergrowth effects m e n t i o n e d above and that of finite electron beam spread (12,25) in the SEM could detract from this viewpoint. However our attention to larger (> 5~m) grains and experience in phase delineation by ion-etching (7) w o u l d suggest this further solid solution to be real, and there is other evidence of such solid solutions in synthetic p r e p a r a t i o n s (26) and industrial clinkers (5). Conclusions The p r e s e n t study shows that with special care the accuracy of SEMm i c r o a n a l y s i s can be made sufficiently good to merit a study of the amounts of minor elements in alites and belites with a view to making comparisons between various possible lattice substitution schemes on the basis of divergence from stoicniometry. The same study can be made also with the interstitial m a t r i x but the problems of i n h o m o g e n e i t y and crystal intergrowth make the results a little less certain here. A__cknowledgements AG thanks the Cement Makers F e d e r a t i o n of the United K i n g d o m for financial assistance. We also thank Dr I. Smith and Dr J. Bye of Associated Portland Cement Manufacturers Ltd., Mr H. W e r m o u t h of Tunnel Cement Ltd. and Dr D. F. Barlow of Rugby P o r t l a n d Cement Company Ltd., for the samples and Professor J. W. Jeffery for helpful suggestions. References i. 2. S. Diamond, Cement and Concrete Res., 2, 617 J. Skalny, J. E. Mander and M. H. Meyerhoff, 5, 119 (1975). (1972). Cement and Concrete Res.,

3.

G. Yamaguchi and S. Takagi, Proc. 6th Int. Symp. Chem. Cement, Moscow, September (1974). P. Barnes, J. W. Jeffery and S. L. Sarkar, Cement and Concrete Res., 8, 559 (1978). M. Kristmann, Cement and Concrete Res., 8, 93 P. Barnes, N. T. Moore and S. L. Sarkar, 51, 251 (1977). (1978).

4.

5. 6.

Indian Concrete Journal,

7.

P. Barnes, G. M. Fonseka, A. Ghose and N. T. Moore; To be p u b l i s h e d in J. Matl. Sci.

Vol. 9, No. 6 MICROPROBE ANALYSIS, CLINKERS, SOLID SOLUTIONS

8. N. T. Moore, S. L. Sarkar and J. W. Jeffery, World Cement Technol., 8(6), 240, 245-7 (1977). J. C. Russ, EDAX Editor, ~, I (1975). K. F. J. Heinrich and H. Yakowitz, D. R. Beaman and L. F. Solosky, Microch. Acta, 182, 905 (1972).

755

9. I0. 11. 12.

Analytical Chem., 44,

1598 (1972).

G. Yamaguchi and S. T~kagi, Proc. 5th Int. Symp. Chem. Cement, Vol. I, p.181. Tokyo (1968). K. Mohan and F. P. Glasser, 379-384 (1977). Cement and Concrete Res., ~, 1-8, 269-276,

13.

14.

W. Gutt and M. A. Smith, Proc. 6th Int. Symp. Chem. Cement, Moscow, September (1974). R. W. Nurse, J. App. Chem., E. Woermann, 2, 708 (1952).

15. 16. 17.

T. Hahn and W. Eysel, Zement Kalk Gips, 16, 370 (1963). Proc. 5th Int. Symp. Chem. Cements,

T. Hahn, W. Eysel and E. Woermann, Tokyo (1968). H. G. ~ d g l e y J. W. Jeffery, and K. E. Fletcher,

18. 19. 20. 21. 22.

Trans. Brit. Cer. Soc. 62, 917 (1963). (1952).

Proc. 3rd Int. Symp. Chem. Cements p.30.

K. E. Fletcher, Mag. Concrete Res., 20, 167 (1968). H. G. Midgley and M. Bennett, M. Regourd and A. Guimier, September (1974). J. Forest, Silicates R. W. Nurse, Proc. Cement and Concrete Res., I, 413 (1971).

Proc. 6th Int. Symp. Chem. Cements, Moscow,

23. 24. 25. 26.

Industries, 32, 373 (1967). (1952).

3rd Int. Symp. Chem. Cements, p56. EDAX Editor, 5, 21 (1975). (1965).

T. Knudsen and N. Thaulow, P. Tarte, Nature,

207, No. 5000, 973