The absence of the -30 ºC peak on the 3 d and 7 d exposures for the 0.5 w/c specimen raises questions regarding microstructure formation under UWVE. Although DOH data exhibit only a minor effect due to the exposure, the DSC data suggest the microstructure might be considerably different. The absence of the freezing peak at -30 ºC indicates that either capillary pores remained percolated or the capillary pores were enclosed by inter-LD C-S-H, but no water remained within the capillary pores. Regardless, given evidence that pastes hydrated at lower relative humidities exhibited higher levels of silicate polymerization [34], that the observed -45 ºC peak height decreased, and that the amount of bound water is proportional to the amount of free water [26,33], it is quite likely that the final microstructure of the 0.50 w/c 3 d and 7 d specimens at 28 d are different from the corresponding saturated specimens. Evidence to this effect for 0.4 w/c mortars can be found in the Bager experiment [7]. Using the diffusion coefficient of 14 d curing as a reference, Bager found that the diffusion coefficient of specimens cured 2 d and 7 d were a factor of 3 and 2 greater, respectively.
Moreover, simply re-wetting a paste that has been exposed prior to the formation of the -30 ºC peak might not negate any detrimental effects. Based on re-wetting experiments on 0.40 w/c and 0.60 w/c pastes [27], water loss in both the -30 ºC and the -45 ºC peaks is not recovered upon re-wetting, despite the fact that the total initial evaporable water content was nearly fully recovered upon re-wetting.