Rheology and structuring in organo-ceramic composites

https://doi.org/10.1016/S1359-835X(00)00179-2Get rights and content

Abstract

The processing–structure–property relationships for calcium aluminate based organo-ceramic composites are reviewed. Processing behavior was measured with a torque-outfitted Banbury mixer, structuring behavior was analyzed with electron microscopy, and mechanical properties were determined using flexural strength measurements. These results are discussed for composites composed of either a polyvinyl alcohol or phenol formaldehyde resin organic phase. The two materials have different processing behaviors, driven by different paste formation mechanisms, but are observed to have similar microstructure and mechanical properties. The rheology of model, non-calcium aluminate based composites was studied using a lubricated squeezing flow rheometer. The relaxation modulus following a step strain was measured for reactive systems that stiffened due to a crosslinking polymer phase or a polymerizing polymer phase and compared to that measured for non-reactive polydimethyl siloxane (PDMS) systems with filler fractions from 0 to 65% by volume and 65% filled PDMS of differing molecular weights. Comparison of the normalized relaxation moduli of the model reactive and non-reactive materials suggests similarities between the effect of filler amount and crosslinking amount and between the combined effects of molecular weight and filler amount and the degree of polymerization.

Introduction

Reactive organo-ceramic composites consisting of calcium aluminate particles embedded in a dense polymer matrix represent an important new class of high-strength materials. These systems have the advantage that they can be fabricated at or near room temperature as extrudable pastes, prior to low temperature heat treating, to produce hardened materials with properties similar to fired ceramics. The paste state is generated by chemical and/or physical interactions that take place between the organic and ceramic phases during high-shear processing, leading to rheological properties similar to those of highly-filled polymer melts. A unique but essential feature of these pastes is that due to the progression of chemical and/or physical interactions taking place, the processing behavior of the paste changes with time, i.e. the paste becomes stiffer and its viscosity increases as it is mixed or processed. In order to take advantage of the continuous processing possibilities of these pastes, one needs to understand how the paste rheology evolves with the processing time.

These composites were first developed by Birchall et al. [1] as a means to improve the flexural strength of hydraulic cement. By combining a small amount of a polyvinyl alcohol (PVA) polymer with a calcium aluminate cement (CAC) in a high-shear mixing environment, they produced hardened materials with flexural strengths in excess of 200 MPa. By comparison, a cast and dried cement paste has a typical flexural strength of 3–10 MPa. Since then, many combinations of cement and polymer have been reported in the literature, but the CAC–PVA composite has the highest flexural strength and has been the most extensively studied. Due to problems with the durability of the PVA composite, Pushpalal et al. [2] replaced the PVA with a hydrophobic phenol formaldehyde resin (PR) and were successful in making a more durable, but still high-strength composite.

Previous studies on CAC–PVA and CAC–PR composites have shown that quantitative information on the mixing chemistry and resultant properties of the polymer-particle network can be generated using Banbury mixing. This type of processing protocol, in conjunction with mechanical testing and microscopy studies of hardened composite, has been used as part of a comprehensive study in the authors' laboratory on the processing–property–structure relationship in these composites [3], [4], [5], [6], [7]. In an attempt to generalize the work on these two specific systems and to understand better the evolving paste rheology of these reactive composites, the authors have studied two model reactive systems whose mechanism of paste stiffening is known [8].

The objective of this study is to review the processing–property–structure relationship for the CAC based composites and to present some new information on the rheological and processing characterization of model highly-filled reactive composites.

Section snippets

Organo-ceramic composites

Two different composite systems, both based on CAC, have been extensively studied. Their formulation and strength in the hardened state are compared to that of the ordinary portland cement in Table 1. The pastes are prepared by pre-blending the components in a planetary mixer and then feeding them into a high-shear Banbury mixer. In the mixer the material is transformed from a damp powder into a viscoplastic paste within about 30 s. The torque required to maintain a constant mixing rate is

Rheological studies of model highly-filled systems

From the work with the CAC–PVA and CAC–PR systems, it becomes clear that each candidate combination of polymer and ceramic is likely to require its own Edisonian investigation in order to understand and optimize its processing and rheological behavior. In an attempt to avoid this, and to build a general understanding of the processing behavior of the entire class of materials, the authors have worked with four model systems. Each model system is a combination of ceramic particles and an organic

Conclusions

Organo-ceramic composites based on CACs have mechanical properties in the range of fired ceramics but are able to be processed at or near room temperature. The relationship among processing behavior, property development, and structure development has been studied with Banbury mixing, mechanical testing, and electron microscopy. Composites based on both PVA and PR are seen to have very different torque-mixing time behavior in the Banbury mixer indicating differing mechanisms of paste formation.

References (10)

There are more references available in the full text version of this article.

Cited by (6)

View full text