Computer Modeling of the Replacement of Coarse Cement Particles by Inert Fillers in Low W/C Ratio Concretes: Hydration and Strength

The model predictions for the two systems containing filler with w/s=0.25 are contrasted against those for the system with no filler in Figs. 2 and 3. In both cases, it can be seen that while the systems with the inert filler replacing the coarser cement particles actually hydrate more completely (due to their higher effective w/c ratio), their compressive strength development is less than that of the control (no filler) system. These strength differences are highlighted in Fig. 4, which shows the projected strength reduction vs. time for the two replacements considered in this study. It is observed that the compressive strength reduction is maximal after about 12 d and then decays towards zero, as the "filled" systems continue to hydrate due to their overall higher w/c ratio. For the 20.5 % replacement, the maximum projected reduction is only 6 MPa (or about 7 %), which may be quite acceptable from a performance standpoint. After 28 days, the standard age for acceptance testing, the projected reductions are about 5 MPa (6 %) and 11 MPa (13 %) for the 20.5 % and 30.8 % replacement levels, respectively.

**Figure 2:** Predicted degree of hydration and compressive strength development for base w/c=0.25 system and system with coarsest 20.5 % of cement particles replaced by inert filler. Solid line is original system and dotted line is system with 20.5 % replacement.
$\begin{figure}\special{psfile=compwc025f20.ps hoffset=0 voffset=35 vscale=65 hscale=65 angle=-90} \vspace{9.5 cm} \vspace{0.10in}\end{figure}$

**Figure 3:** Predicted degree of hydration and compressive strength development for base w/c=0.25 system and system with coarsest 30.8 % of cement particles replaced by inert filler. Solid line is original system and dotted line is system with 30.8 % replacement.
$\begin{figure}\special{psfile=compwc025f30.ps hoffset=0 voffset=35 vscale=65 hscale=65 angle=-90} \vspace{9.5 cm} \vspace{0.10in}\end{figure}$

**Figure 4:** Predicted reduction in compressive strength due to coarse cement particle replacement for Cement 135 with w/s=0.25. Solid line is reduction with 20.5 % replacement level and dotted line is reduction with replacement level of 30.8 %.
$\begin{figure}\special{psfile=diffwc025.ps hoffset=20 voffset=35 vscale=55 hscale=55 angle=-90} \vspace{9.5 cm}\vspace{0.10in}\end{figure}$

Similar results were obtained for the w/s=0.30 systems. Thus, only the strength differential plot is provided here, in Fig. 5. Here, the maximum difference in compressive strength is observed to occur at a slightly later age, 18 d to 40 d, since with a higher w/c ratio, the control system is able to maintain a significant hydration rate for a longer period of time. After 28 days, the projected reductions in compressive strength are about 6 MPa (7 %) and 11 MPa (13 %) for the 14.5 % and 22.3 % replacement levels, respectively. Therefore, in this case, only a 15 % replacement level of the coarsest cement particles may be acceptable. In principle, the lower the "initial" w/c (w/s) ratio, the greater the possible replacement level for the coarse particles without detrimentally reducing the compressive strength. The preliminary results presented here would suggest maximum replacements of 15 % and 20 % for w/c ratios of 0.30 and 0.25, respectively. Thus, the optimal replacement level is seen to be a strong function of the ultimate concrete mixture proportions. This implies that a given replacement level would only provide "optimal" performance in a limited range of concrete mixture proportions.

**Figure 5:** Predicted reduction in compressive strength due to coarse cement particle replacement for Cement 135 with w/s=0.30. Solid line is reduction with 14.5 % replacement level and dotted line is reduction with replacement level of 22.3 %.
$\begin{figure}\special{psfile=diffwc030.ps hoffset=20 voffset=35 vscale=55 hscale=55 angle=-90} \vspace{9.5 cm}\vspace{0.10in}\end{figure}$

Regarding the appropriate filler material to use, coarse limestone particles may offer one possibility. Studies [8,9] have indicated that limestone particles are only mildly reactive within the cement system, producing small amounts of a monocarboaluminate Afm-type reaction product at longer hydration times. With relatively coarse limestone particles in these lower w/c ratio concretes, these long-term reactions may be further limited by the same space limitations that prohibit all of the coarser cement particles from hydrating completely. The incorporation of limestone-cement reactions into the NIST CEMHYD3D computer model is the subject of current research, so that this reactivity will be examined quantitatively by both experiments and computer modeling in the future.

3. Results and Discussion