【正文】 e absorption of the acetone into the pores of paste can stop the hydration process by eliminating the remaining water. After removing the free water, the samples were dried in vacuum at room temperature. The morphologies of steel slag and its hydration products were characterized using a FEI Quanta200F scanning electron microscope under a high vacuum condition (SEM). EDX was used to identify the element distributions of the mineral phases and hydration products. The steel slag and its hydration products were mineralogically determined by Xray diffraction. XRD measurements were conducted with a TTR III diffractometer using nickelfiltered Cu K 1 radiation (=? ), 50 kV voltage and 200mA current. The nonevaporable water content of paste was obtained as the difference in mass between the sample heated at 65 ?C and 1000 ?C normalized by the mass after heating 65 ?C, and correcting for the loss on ignition of unhydrated samples. Thermogravimetric analysis (TGDTG) was carried out using a Setaram thermoanalyser at a heating rate of 10 ?C/min up to 900 ?C. Apart from the original steel slag, two other kinds of steel slag were obtained by sieving the original steel slag into two portions. Steel slagA represented the portion of particles smaller than 61 m, which accounts for 81% of the total mass of the original steel slag. Steel slagB represented the other portion. Mortar bars of 40mm40mm160mm were prepared. The watertobinder ratio of all mortars was . The sandtobinder ratio of all mortars was . Mortars made by binders posed of 100% cement, 80% cement and 20% original steel slag, 80% cement and 20% fly ash, 80% cement and 20% steel slagA, 60% cement and 40% original steel slag, 60% cement and 40% fly ash, and 60% cement and 40% steel slagA were denoted by CM, SM1, FM1, SSM1, SM2, FM2, SSM2, respectively. Mortars were cured first at 20177。1 ?C and 95% relative humidity for the first day, and then cured in water of 20177。1?C for the remaining ages. At the age of 3, 28 and 90 days, their pressive strengths were tested according to Chinese National Standards GB/T176711999.2. Results and discussion. Characteristics of the mineral positions of steel slagFig. 1 shows the particle size distributions of steel slag and cement. The steel slag contains % (volume fraction) particles with diameters smaller than 10 m and % particles with diameters larger than 60 m. The cement contains % and % of particles with diameters smaller than 10 m and larger than 60 m, respectively. Steel slag has a little higher proportion of small particles (10 m) and much higher proportion of large particles (60 m) than Portland cement, however, the proportion of moderate particles (10–60 m) in steel slag is far less than that in Portland cement, so steel slag indicates a relative poor continuity in particle size distribution. The poor continuity is also reflected in Fig. 2, a morphological micrograph of steel slag grains at 1000amplification.In XRD spectrum of steel slag, characteristic peaks of mineral phases concentrate at 30–45? [1,16–18]. Fig. 3 pares the XRD spectrums of the original steel slag, steel slagA, and steel slagB. It is derived from Fig. 3 that steel slagB is lower at C3S, C2S, C2Al2Si3O12 and C12A7 but richer at RO phase than the original steel slag. Steel slagA is richer at C3S, C2S, C2Al2Si3O12 and C12A7 but lower at RO phase than the original steel slag. Fig. 4 and Table 2 show the morphology and EDX results of the small particles in the steel slag, respectively. It is derived from EDX that these small particles are posed mainly by silicate and aluminate, and besides, a small amount of RO phase and Fe3O4.abFig. 4. SEM morphologies of small steel slag particles.. Contribution of mineral to cementitious performance of steel slag The results of Ref. [16] showed that most silicate and aluminate in steel slag have hydrated after 90 days of reaction, producing C–S–H gel, C–S–Al–H gel and Ca(OH)2。 but RO phase, Fe3O4 and C2F barely participate in the hydration due to their low activity. XRD result of the hydration products of steel slag at 360 days is given in Fig. 5, in which gels are not reflected because of their amorphous structures. Fig. 5 indicates that RO phase, Fe3O4 and C2F still show ultra low reaction extent even after having undergone hydration for 360 of the hydration products of steel slag at 360 days is given in Fig. 6. The large unhydrated particles inlaid in the gel are diagnosed as RO phase by EDX analysis. RO phase not only contributes nothing to the cementitious properties of steel slag, but also forms a thin layer between the RO particle and the surrounding gel due to the relatively large particle size and smooth surface, so RO phase plays a negative role in the cementitious properties of steel slag to a certain extent.Table 2Chemical position of particles in Fig. 4 (at.%).Fig. 5 alsoshowsthat part of silicate and aluminate are still unhydrated even at 360 days. Generally speaking, hydration activity of silicate and aluminate is related with the particle size: the larger the size, the lower the activity. Some silicate and aluminate particles of steel slag are too large in size that only slight hydration occurs at the surfaces of these particles at 360 days and a very small amount of hydration product is generated (Fig. 7). To a specific cementitious material, the amount of its hydration products is proportional to the nonevaporable water content of the products. The major cementitious ponents of both steel slag and Portland cement are C3S and C2S, and both with C–S–H gel and Ca(OH)2 as major hydration products, so it is available to conduct a horizontal parison to the nonevaporable water contents of the hydration products of steel slag and Portland cement. The nonevaporable water contents of the hydratio