Single-Point Imaging of Partially Dried, Hydrated White Portland Cement

Moisture content in concrete has a major influence on the concrete samples with spin-echo imaging methods (15, 16) . The spin–spin relaxation times of water in these rehydration durability and mechanical properties of the material during its service life (1, 2) . Resistance to corrosion, ability to experiments is only a few milliseconds so the quality of traditional spin-echo images is disappointing. Drying differs withstand repeated freeze/ thaw cycles, and resistance to explosive spalling at high temperatures all depend on the moisfrom the rehydration phenomena on a molecular scale and the transverse-relaxation times for the drying samples are ture content and distribution of moisture within the concrete (3–5) . Knowledge of the moisture distribution in drying significantly less (vide infra) . We are aware of no other NMR imaging studies of drying in portland cement-based concrete is thus of paramount importance given the ubiquitous use of concrete as a construction material. In order to materials. Concrete appears to be a natural subject for NMR imaging provide a quantitative basis for understanding water movement during drying as a result of capillary flow or molecular because the technique is inherently three dimensional and the cement paste in concrete is approximately 50% water by diffusion, it is essential to have a good experimental method for measuring local water content. The ideal method should mass. After full hydration, approximately 40% of this water is free to leave the material during drying (17) . We have have submillimeter resolution, be nondestructive, be rapid, and be easy to interpret. It should also be capable of extenbegun our study of concrete materials by concentrating on cement, one constituent in the composite material concrete. sion to three dimensions. One of the most promising solid-state NMR imaging In common with many other porous media, however, the cement system has transverse-relaxation times which are methods is the constant-time imaging (CTI) approach proposed by Emid and Creyghton (6) . Some variations of CTI considerably shorter than standard echo times. In a moist cured white cement, 1 the spin–spin relaxation times T2 and exist and have been discussed in the literature (7–9) , including the single-point imaging method (SPI) . In this CommuT* 2 were 320 and 170 ms, respectively, when measured at nication, we demonstrate the application of the SPI method 100 MHz. The spin–lattice relaxation time, T1 , measured to noninvasive examination of moisture content level, and its under the same conditions is also short, 1.2 ms. In a standard changes, during drying of hydrated white portland cement. H spin-echo imaging sequence, an echo time (TE) of 2 ms Gravimetric sampling (10) and g-ray attenuation (11) are is attainable; it is possible to shorten this time to approxitraditional methods for moisture content analysis in concrete mately 1 ms by reducing resolution. Eddy-current effects on materials. The first method, however, has very coarse resoluthe image with echo times this short are often pronounced tion and requires many samples because it is a destructive due to the finite settling time of the read gradient. Shorttest method. The g-ray attenuation method has the advantage echo-time, one-dimensional, spin-echo profiles of moist of being nondestructive but moisture determination is indicured white-cement paste cylinders displayed poor signalrect. Application of this method to a heterogeneous material to-noise and geometric distortions from the ideal profile gewith a variety of absorption cross sections tends to cloud ometry. experimental interpretation. A schematic description of the one-dimensional SPI seNMR has been used by a number of groups to study the chemical dynamics of cement hydration (12–14) . Several 1 White portland cement, water-to-cement ratio 0.5 by mass, was cast groups have attempted to image water invasion of cured into open, round bottom, polyethylene bottles (5 cm diameter, 5 cm length) . Samples were moist cured for 28 days at 23{ 27C, relative humidity greater than 95%. After curing, samples were dried at ambient temperature and † Permanent address: Institute for Isotopic and Molecular Technology, P.O. Box 700, 3400 Cluj-Napoca, Romania. humidity. Drying occurred by evaporation from the exposed surface of the cylinder. White portland cement was chosen for its low iron content. ‡ To whom correspondence should be addressed.