Table 1 presents chemical composition of reject fly ash and fine ash used in the experiment. Composition of RFA and fine ash are pretty similar, and major components were SiO₂ and Al₂O₃. Ignition loss of RFA was more than double of fine ash, and reject fly ash has far higher unburned carbon and volatile substances than fine ash. Generally, it is known that unburned carbon of fly ash absorbs AE agent when it is blended in the concrete, thus hinders inclusion of air.^8,9) Because admixture such as AE was not used in the present study, it was judged that there would not be any problem in expressing physical properties of mortar.

Average particle size of reject fly ash by grinding time is presented in Table 2. As curding time was increased, though average particle size was decreased, reduction of particle size was gradually decreased as can be seen in Table 2. This means that as particle diameter is decreased, grinding would be difficult, which is in conformity with general grinding mechanism.

Figure 1 shows micro-structure of reject fly ash before and after grinding with a vibratory milling for one hour. Reject fly ash before grinding (b) displays a general similar sphere shape, but the sphericity was lower than the fine ash. Generally, spherical fine ash has a ball-bearing effect, which increases fluidity of the concrete and mortar.¹⁰⁾ However, it was found that reject fly ash (c) after grinding one hour was changed to broken sphere (angled micro-particles).

XRD patterns of reject fly ash before and after grinding are presented in Fig. 2. Generally, major crystals in the coal ash are mullite and quartz, and major peaks of these crystals are presented at an around 27°.¹¹⁾ In this study, above peaks were observed at an around 27°, and major peaks were decreased under the grinding condition. At this time, peak strength in the non-grinding condition was 1363 CPS (Count/Second), and this value became decreased in a range of 808 ~ 908 lower than that CPS after grinding. However, no big difference was observed by increasing the grinding time. The reason why XRD peak strength was decreased is due to increases in the amorphous contents. Therefore, from the above results, it could be assumed that crystalline property of the reject fly ash becomes deteriorated by grinding. However, reduction of peak strength by increased grinding did not show a clear trend. Rather, peak strength was increased in some of the samples, which needs further analysis.

Figure 3 shows SEM photos for the rejected fly ash before and after grinding. In Fig. 3 (a) and (b) atomic arrangement of particles could be observed. These are representative photo obtained from crystalline particles. Therefore, it was judged that (a) and (b) are parts of quartz particle analyzed in the XRD Pattern. Especially, in photo (b), dislocation which is a type of line defects was confirmed.¹²⁾ From the photo (d), trace of falling parts of particles and many holes were found from several places inside the particles as compared with photo (c). This phenomenon might be due to the strong impact from steel balls which are the grinding media inside of vibratory mill.

Above defect can increase energetically-modified degree for the particles by making particles unstable. After that, mixed water and alkali ion move at the interface of the particle, particle changed to stable phase. That is, defect portion presented in the particle is filled by hydration, and stability is achieved. In order to fill the defects, new materials have to be produced at the defective portion. In case of cement mortar and concrete, mixed water and alkali ion move, and new hydrate should be formed at the defective portion. Ying wang et al. has expressed this process as “nucleus production effect”.¹³⁾ Further, large specific area can act as a heterogeneous nucleus producing surface that promotes sedimentation of the hydrate. Therefore, crack or holes produced during grinding can play a vital role on the production of hydrates by attributing increases in the specific area.

Compressive strengths of mortar into which reject fly ash was substituted or used at 10% by grinding time on the third day, 7 days, and 28 days are presented in Fig. 4. As grinding time became extended to 60 min., though compressive strengths on the 3^rd, 7^th, and 28^th day were increased, this trend was followed by no great changes. Maximum compressive strength was 35.7MPa on the 3^rd day, 48.0MPa on the 7^th day, and 54.6MPa on the 28^th day during grinding for 60 min. These values were higher by 0.4MPa as compared with compressive strength of reject fly ash mortar without grinding on the 28^th day, and by 0.5MPa as compared with OPC mortar (Fig. 5). Fig. 6 shows separate expression of compressive strength on the 28^th day by grinding time. Compressive strength was increased until grinding for 60 min., after which showed similar values. Besides, analysis results for specific surface area of reject fly ash by grinding time showed that as grinding time was increased, increases rate of the specific surface area was decreased. (Fig. 7) Therefore, after grinding for 60 min., since changes in the specific surface area was meager, compressive strength value might have been similar. These results suggested that grinding time for 60 min. was the most effective for performance and economic feasibility of reject fly ash. Compressive strength of the mortar by mixing ratio of ERFA as compared with OPC mortar on which ERFA was not mixed on the 28^th day are presented in Fig. 8. When 10% of ERFA was substituted, mortar showed greater compressive strength than OPC mortar. The compressive strength of mortar was reduced as substitution rate was increased. This is a phenomenon occurring while producing Ca(OH)₂ which is a hydrate of cement as mixing ratio of cement was decreased. The major components of rejected fly ash SiO₂ and Al₂O₃ are major pozzolans and reacts with Ca(OH)₂, to produce calcium silicate hydrate (C-S-H gel).¹⁴⁾

Above hydrates make hydrated structure more compact, attributing improvement of physical property of mortar. That is, while Ca(OH)₂ which is necessary for pozzolan reaction of reject fly ash becomes lacking, thus strength becomes impaired. Further, as cement contents were decreased, decreases in the C3A and gypsum also might have affected on decreases in the strength. Therefore, it was judged that RFA substitution at 20% which exhibited similar strength with OPC mortar would be adequate in order to secure optimum physical property of mortar wherein pulverized fuel ash was used.

Contents of mixed water were fixed at 40% as against ERFA in order to analyze and evaluate hydration characteristics of cement, and then left inside thermo-hygrostat at temperature 23°C-relative humidity at 90% for one day. After that, it was cured under water. XRD analysis was conducted for the hydrates which was cured for three days and for 28 days. Fig. 9 shows XRD patterns of hydrates cured for three days and 28 days. While, Table 3 presents peak strength of Ca(OH)₂ in terms of value.

In all the XRD Pattern, peaks of Ca(OH)_2, which is one of the major hydrates were found. Crystal peak of Ca(OH)₂ was presented at near 18.2˚. Apart from this, ettringite peak at near 9˚ was also observed. Above peaks are the typical hydrates which are observed in OPC hydrates. Therefore, it could be assumed that energetically-modified mixed cements prepared in this study would express similar physical properties with OPC.¹⁵⁾

Ca(OH)₂ peak strength of hydrates prepared by using rejected fly ash was increased from 754 to 1016 as curing duration was increased (3 days →28 days). In case of energetically-modified reject fly ash, peak strength was increased with a less degree from 646 to 700. That means, Ca(OH)₂ crystalline properties of both specimens were gradually increased as curing days were extended. However, the increment of energetically-modified reject fly ash was relatively less. This phenomenon might be due to difference in the activity between reject fly ash and ERFA. Since coal ash reacts with Ca(OH)₂ which is a hydrate of cement, increase and decrease of Ca(OH)₂ peak strength are affected by amount of Ca(OH)_2, that is involved in the pozzolan reaction. That is, the lower strength increases of hydrate of energetically-modified hydrates means the more amount of Ca(OH)₂ was involved in hydration than hydrates of reject fly ash, which indicates that activation of reject fly ash by grinding is beneficial for promoting hydration.

Micro-structure of cured hydrate for seven days using reject fly ash and ERFA are presented in Fig. 10. The ash particles presented in the hydrates of reject fly ash have not been sufficiently hydrated. Some of the ash particles which were not ground were observed in the hydrates of energetically-modified reject fly ash. However, hydration was progressed to some extent in the ground particles, thereby hydrates were linked with matrix. Especially, at the fracture surface of the particles, hydration was well advanced, thus particle shape could not be observed properly. Such hydration characteristics would be important factors attributing in improving compressive strength.