Elsevier

Cement and Concrete Research

Volume 34, Issue 9, September 2004, Pages 1697-1716
Cement and Concrete Research

The work of Powers and Brownyard revisited: Part 1

https://doi.org/10.1016/j.cemconres.2004.05.031Get rights and content

Abstract

Powers and Brownyard [Studies of the physical properties of hardened Portland cement paste. Bull. 22, Res. Lab. of Portland Cement Association, Skokie, IL, U.S. J. Am. Concr. Inst. (Proc.), 43 (1947) 101–132, 249–336, 469–505, 549–602, 669–712, 845–880, 933–992 (reprint)] were the first to systematically investigate the reaction of cement and water and the composition of cement paste. They introduced the concept of nonevaporable (water retained in P-dried state) and gel water (additional water retained upon saturation). Their specific volumes (νn and νg) are lower than that of free water, causing chemical shrinkage. The retained water was furthermore related to the content of the four most abundant clinker phases, viz alite, belite, aluminate and ferrite.

Their work is recapitulated here. Major aspects, such as the specific volume of nonevaporable and gel water, are addressed, as well as the issue of gel water being “compressed”. Subsequently, it will be demonstrated that their water retention data enable the study of the molar reactions of the calcium silicate phases and the reaction products C-S-H (C1.7SH3.2 when saturated) and CH, which represents a principal innovation. Using the molar reactions and the specific volumes of nonevaporable (νn) and gel water (νg), the density of saturated C-S-H and its porosity are derived.

Introduction

In a pioneering work, Powers and Brownyard [1] were the first to systematically investigate the reaction of cement and water and the formation of cement paste. In the late 1940s, they presented a model for the cement paste, in which unreacted water and cement, the reaction product and (gel and capillary) porosity were distinguished. Major paste properties were determined by extensive and carefully executed experiments, including the amount of retained water and the chemical shrinkage associated with hydration reaction. These properties were furthermore related to the content of the four most important clinker phases, viz alite, belite, aluminate and ferrite. Additionally, the composition of the cement paste was related to engineering properties, such as compressive strength, shrinkage, porosity, water permeability and freezing/thawing. Their model furthermore distinguished gel and capillary porosity. The impact of this standard work is paramount, and their concepts and results are used in contemporary cement and concrete science [2], [3]. For a historical overview of the development of Portland cement research until 1960, starting with the first publications and dissertation by Le Chatelier in 1887, the reader is referred to Steinour [4].

This paper addresses the work of Powers and Brownyard for four reasons. First, although their work is widely cited (although perhaps not widely read), a rigorous review of their theoretical model and experimental results is still lacking. Taylor [2], Neville [3], Czernin [5], Locher [6], Hansen [7] and Jensen and Hansen [8] summarise the most important features of the model, but the underlying experiments, model and equations deserve a more in-depth discussion. Furthermore, these authors make use of later work by Powers [9], [10], which refers to Cement 15754 to illustrate the implications of the model. This latter cement is, however, not representative of contemporary Portland cements (CEM I, corresponding to ASTM Type III cement), which contain more alite. Based on Cement 15754, the model predicts a minimum water cement ratio of 0.42 [7], [8] to 0.44 [2] for complete hydration. The present analysis will reveal that this value is actually much lower for CEM I (around 0.39), being in line with values accepted nowadays. This aspect is a second motivation for the present analysis of their work. Third, in later publications, the work of Powers and Brownyard [1] was corrected (e.g., their prediction of chemical shrinkage) or criticized (e.g., their drying and sorption experiments). The relevance of these comments will also be addressed in the present study. Finally, and probably most importantly, it will be demonstrated that their results enable the study of the reactions of the four clinker phases and quantification of their reaction products, which is a principal innovation. A short treatment of these reactions was presented recently [11].

In the past, the model predictions have been compared with pure C3S1 hydration by Locher [12] and Young and Hansen [13]. To this end, the water binding of cement as given by Powers [9] was used, i.e., the data pertaining to Cement 15754. This approach is permitted, as C3S is a major constituent of Portland cement. The model of Powers and Brownyard [1], however, contains specific information in regard to the reaction of each individual clinker phase, such as C3S. This aspect of their model, overlooked in the past, has been presented elsewhere recently by Brouwers [11], [14] and will be analysed in detail here.

Using their model and data yields some major advantages. First, in contrast to plain clinker hydration experiments, their experiments and model are based on a real cement–water system. In contrast to hydrating pure clinker minerals, a mix of clinker minerals that contains impurities and that hydrate simultaneously is more compatible with practice. Furthermore, their experiments concerned paste hydration, which is also closer to reality than are bottle hydration experiments that are frequently reported. Applying their work to the hydration of the calcium silicate phases yields the reaction stochiometry, as well as the amount, density, porosity, water content, etc. of the most abundant reaction product, C-S-H. Likewise, in the approach of Le Chatelier, a picture that is generally accepted nowadays, the calcium silicate phases are assumed to react independently from the aluminate phase [4]. In a future publication [15], it will be seen that the work of Powers and Brownyard [1] can also be applied to the reactions of aluminate, ferrite and sulphate phases. The type of products formed and their quantities can be derived from their water retention data. Their classical work contains much more information and will have a wider application than is generally appreciated.

Section snippets

Theory

In this section, the main features of the model by Powers and Brownyard [1] are recapitulated. The terms and symbols used are, as much as possible, consistent with theirs, as well as those used in later work by Powers [9], [10], [16], [17], [18].

Powers and Brownyard [1] distinguished three phases in the cement paste viz capillary water (=unreacted water), unreacted cement and cement gel (Fig. 1). The cement gel consists of solid hydrated cement and water-filled gel pores; this water is referred

Water retention and shrinkage

As said, the model as discussed in the previous section requires quantification of nonevaporable and gel water, as well as their specific volumes. To this end, Powers and Brownyard [1] executed and reported numerous experiments with cements of different compositions, with neat cement and with mortars, and at various water/cement ratios (w0/c0) and various hardening times. Most mixtures were cured at 21.1 °C (70 °F), but a few mixtures were steam cured in an autoclave at 215.5 °C (420 °F). Here,

Reactions of calcium silicate phases

Powers and Brownyard [1] presented a literature review of the reactions products and were aware that the products of the clinker phases C3S and C2S were “microcrystalline” CH and a “colloidal gel”, also named “colloidal hydrous silicate” and “calcium silicate hydrate” (pp. 106–132, 260 and 488). In later work, this product was called “tobermorite gel” [23], [32], [33], [34], [35], and nowadays, it is generally called C-S-H [2]. It is known to be a poorly crystalline to almost amorphous material.

Conclusions

Powers and Brownyard [1] presented a model that accounts for unreacted cement, free water, the hydration product (which is porous in itself, i.e., gel porosity) and chemical shrinkage (Fig. 1). Careful execution of experiments resulted in quantity and specific volume of both nonevaporable water and gel water. The water retention in P-dried and saturated states was furthermore related to the mineralogical composition of the cement.

The most important features of the model are discussed in detail

List of symbols

    Roman

    A

    Al2O3

    B

    =B′+1 = (wg+wn)/wn

    B′

    wg/wn

    c

    Mass of reacted cement [g]

    c0

    Initial mass of unreacted cement [g]

    C

    CaO

    C3A

    Aluminate or pure C3A

    CH

    Portlandite

    C3S

    Alite or pure C3S

    C2S

    Belite or pure C2S

    C-S-H

    Calcium silicate hydrate

    CS̄

    Anhydrite

    CS̄H0.5

    Hemihydrate

    CS̄H2

    Gypsum

    C4AF

    Ferrite or pure C4AF

    F

    Fe2O3

    H

    H2O

    k

    Number of absorbed layers of gel water

    M

    Mass of 1 mol of substance [g/mol]

    m

    Mass of reacted phase or formed product [g]

    m

    Maturity factor (degree of hydration), c/c0

    n

    Number of moles

    RH

    Relative humidity

    RO

    Remaining oxides

Acknowledgements

The author is grateful for the advice given by the late Dr. H.F.W. Taylor, Emeritus Professor of Inorganic Chemistry, University of Aberdeen, U.K. (to whom this commemorative issue of Cement and Concrete Research is dedicated). Dr. E. Spohr from the Forschungszentrum Jülich, Jülich, Germany, is acknowledged for discussing MD simulations on water adsorption. The author furthermore wishes to thank the following persons and institutions for providing copies of references: Mr. W.J. Burns from the

References (40)

  • T.C. Hansen

    Physical structure of hardened cement paste, a classical approach

    Mat. Struct.

    (1986)
  • T.C. Powers

    Physical properties of cement paste, Bull. 154, Res. Lab. of Portland Cement Association, Skokie, IL, U.S.

  • T.C. Powers
  • H.J.H. Brouwers, The work of Powers and Brownyard revisited: composition of Portland cement paste, CE & M Research...
  • F.W. Locher

    Stoichiometry of tricalcium silicate hydration

  • J.F. Young et al.

    Volume relationships for C-S-H formation based on hydration stoichiometries

    Mater. Res. Soc. Symp. Proc.

    (1987)
  • H.J.H. Brouwers

    Chemical reactions in hydrated ordinary portland cement based on the work by Powers and Brownyard

  • H.J.H. Brouwers

    The Work of Powers and Brownyard Revisited: Part 2

    (2004)
  • T.C. Powers

    The physical structure and engineering properties of concrete, Bull. 90, Res. Lab. of Portland Cement Association, Skokie, IL, U.S.

    ((March 1956) 1958)
  • T.C. Powers

    Structure and physical properties of hardened Portland cement paste, Bull. 94, Res. Lab. of Portland Cement Association, Skokie, IL, U.S.

    J. Am. Ceram. Soc.

    (1958)
  • Cited by (189)

    View all citing articles on Scopus
    View full text