*3.1. Materials*

BOF slag aggregates with a particle size less than four mesh original fine with a D50 at 0.75 mm obtained from CHC Resources Corporation were utilized. The chemical composition of fine BOF slags is depicted in Table 1. The key composition in BOF slag was CaO and Fe2O3; it also contains 4.7 wt% of free lime. Additionally, GGBS and FA were obtained from CHC Resources Corporation in Taiwan. GGBS possesses a particle size between 0.6–135.7 μm with D50 at 12.3 μm while those of FA range among 0.7–201.9 μm with a D50 at 22.2 μm. The chemical composition of GGBS powder and FA were also represented in Table 1. Alkali activator solutions with various SiO2/Na2O molar ratios were prepared by mixing sodium silicate solution (9.5 wt% Na2O, 29 wt% SiO2) and 6M sodium hydroxide. The molar ratio for SiO2/Al2O3 was kept at 50 and was controlled by sodium aluminate. The alkali solutions were prepared and ready to use the day before the experiment.

**Table 1.** Chemical composition of basic oxygen furnace (BOF) slag, granulated blast furnace slag (GGBS) and coal fly ash (FA).


## *3.2. Methods*

GGBS and FA were mixed at the designed weight ratio with an additional 3 wt% of wollastonite added during blending. Subsequent to pre-mixing for 3 min, the mixture was then activated with alkali solutions. After thorough mixing for 3 min, the fine BOF slags were then added into geopolymer paste for an additional 3 min blending to prepare geopolymer mortar with various geopolymer: BOF slag ratios. The geopolymer mortar was then cast into cylindrical molds of size ( Φ50 × 100 mm). Following 24 h of de-molding, the samples were then cured at room temperature until certain days for testing. In order to measure the volume stability, all the BOF slag-based geopolymer mortar samples were tested using the autoclave testing method according to ASTM C151 standard.

Before and after autoclave testing, the length, diameter, and volume change rate of the cylindrical specimen were calculated by dividing the circle portion of the cylindrical specimen into three sections with the center point of the circle. A Vernier caliper was used for measuring the length varieties of the cylindrical sample. For diameter measurements, an arbitrarily selected marked two points at the middle of the height of the cylindrical test body was measured and another two points were made at a position rotated by 90◦, which were also measured. The change in total volume is calculated from the above results.

#### **4. Results and Discussion**

#### *4.1. BOF Slags Expansion Behavior and Powdering Rate*

The powdering rate measurement was based on GB/T 24175-2009 (Test method for stability of steel slag). Before collecting an 800 g oversize part to put into an autoclave, the BOF slags were screened using a No. 4 sieve. The autoclave maintained the pressure at 20.8 ± 0.7 kgf/cm2, temperature 215.7 ± 1.7 ◦C for 3 h. Subsequently after drying and sieving again by employing the No. 4 sieve, the powdering rates were calculated as illustrated in Equation (3).

Table 2 demonstrates the results of powdering rates for various size ranges of BOF slags after autoclave tests. The powdering rate increased from 17.4% to 29.9% with decreasing particle size ranges from >3/4 inch–3/8 inch–No. 4 mesh. This implies that the finer particle size range of BOF slags contained more free lime.

> **Table 2.** Powdering rate for di fferent size range BOF slags after autoclave test.


$$\text{Powerering Rate} = \frac{\text{Weight of less than 4 mesh part}}{\text{Total Sample weight}} \times 100\% \tag{3}$$

In order to compare the performance of BOF slags (aggregate) in the Portland cement system and geopolymer system (binder), the binder/aggregate weight ratio was kept at 1:2.75 for the BOF slag expansion test. The expansion behavior of the BOF slag in the Portland cement system and geopolymer technology is presented in Table 3. Based on the experiment results, it is seen that in the case of the BOF slag in the Portland cement system, the sample totally collapsed after the autoclave test due to the steam, which accelerated the reaction of free lime, thus causing expansion. Nevertheless, in the case of the BOF slag in the geopolymer system, the sample still maintained its integrity after the autoclave test as displayed in Table 3. The average height, diameter, and volume changes recorded are 0.18%, 0.07%, and 0.35%, respectively. After the autoclave test, the samples were crushed, grinded and analyzed by XRD (X-ray Di ffraction) (Figure 2). The mineral phase of calcium silicate was found, indicating that the geopolymer technique can stabilize the untreated BOF slag fines. This is ascribed to

the enormous quantities of free silicon present in the geopolymer matrix. This free silicon reacted with free lime or free-MgO on the BOF slag surface to form a stable compound. When the BOF slag-based geopolymer is subjected to an external force and cracks crop up, the moisture will enter into the BOF slags. The free silicon in the geopolymer matrix will be dissolved and brought into internal BOF slag to react with free-CaO or free-MgO to form stable calcium silicate or magnesium silicate. This reaction can effectively inhibit the expansion of the BOF slags.

**Table 3.** BOF slag expansion behavior in Portland cement system and geopolymer system after autoclave test.


**Figure 2.** Results of XRD (X-ray Diffraction) analysis of the BOF slags in the geopolymer system after autoclave test.

#### *4.2. E*ff*ect of SiO2*/*Na2O Molar Ratio on the Properties of BOF Slag-Based Geopolymer Mortar*

In order to understand the effect of the SiO2/Na2O molar ratio on the features depending on the BOF slag-based geopolymer mortar, the experimental parameters were adjusted, and the experimental results exhibit that the compressive strength depending on the SiO2/Na2O ratio of the BOF slag-based geopolymer mortar is augmented as the ratio of SiO2/Na2O increases, as portrayed in Figure 3. Both tend to enhance the compressive strength as the curing time increases. When the ratio of SiO2/Na2O was between 1.28 and 1.5, the compressive strength at the age of 56 days was about 32–38 MPa. Likewise, when the ratio of SiO2/Na2O was 1.6, the compressive strength achieved 53 MPa on the day 56. However, only the decrease of compressive strength after 56 days for SiO2/Na2O = 1.4 was found. The reason for this is large amounts of micropores generated in the structure of the geopolymer that hinder the development of compressive strength, and no such phenomenon is found in a ratio higher or lower than the SiO2/Na2O ratio of 1.4 [30]. Moreover, it is found that most of the specimens subjected to the autoclave expansion treatment according to the ASTM C151 standard possess a tendency to increase the compressive strength, and the strength on day 56 can obtain 40–55 MPa as illustrated in Figure 3b. A careful observation of the surface of the specimen after the autoclave test revealed that it still has a small amount of surface peeling, which may be due to the fact that the BOF slag particles on the surface of the specimen cannot be entirely covered by the geopolymeric slurry, or maybe because of the highly thin coating. Therefore, there is still a small part of the reaction expansion phenomenon. However, due to its substantial increase in compressive strength, it indicates that the overall performance of the test body subsequent to the autoclave expansion test is still very stable.

**Figure 3.** Effect of SiO2/Na2O molar ratio on the curing age and compressive strength of BOF slag-based geopolymer mortar before autoclave expansion test (**a**) and after autoclave expansion test (**b**).

The effect of the SiO2/Na2O molar ratio on the curing age of the BOF slag-based geopolymer mortar was subjected to an autoclave expansion test specimen, and its linear expansion, diameter expansion, and bulk expansion characteristics were analyzed. The results are shown in Figure 4. As illustrated in Figure 4, the volume expansion ratio of each of the samples was 0.4% or less, of which, the expansion ratio of SiO2/Na2O = 1.4 and 1.5 is the lowest and is only 0.1% or less. Furthermore, the linear expansion ratio and the diameter expansion ratio were analyzed ahead, and the linear expansion ratio was mostly less than 0.1%, and the diameter expansion ratio was also 0.15% or less. This shows that the untreated BOF slag has high stability under the geopolymer system. According to the results of this experiment, the subsequent selection of the BOF slag-based geopolymer mortar was carried out with a ratio of SiO2/Na2O of 1.5.

**Figure 4.** Effect of SiO2/Na2O molar ratio on the length, diameter, and volume changes after autoclave test, (**a**) volume expansion ratio (**b**) linear expansion ratio (**c**) diameter expansion ratio.

#### *4.3. E*ff*ect of GGBS*/*FA Ratio on the Properties of BOF Slag-Based Geopolymer Mortar*

The results of the compressive strength of the BOF slag-based geopolymer mortar, with varying proportions of the GGBS and FA, are shown in Figure 5. The compressive strength increased with increased GGBS content (Figure 5a). This could be due to the fact that the structure of geopolymer mortar is denser and more increased with curing age. The BOF slag-based geopolymer mortar of different powder ratios was subjected to a compressive strength test after an autoclave treatment, and the results are shown in Figure 5b. According to the experimental results, it can be found that although the surface of the test piece has a small part of peeling after the autoclave test, the compressive strength is still significantly increased. This means that the test specimen is still very stable after the autoclave expansion test. The higher the content of fly ash, the significantly higher the strength of the sample after the autoclave test. This may be due to the high silicon content of fly ash in the system to suppress the BOF slag expansion [31].

**Figure 5.** Effect of GGBS/FA ratio on compressive strength of BOF slag-based geopolymer mortar (**a**) before autoclave expansion test; (**b**) after autoclave expansion test.

The autoclave expansion test, ASTM C151, was carried out to understand the stability and volume expansion rate of BOF slag-based geopolymer mortar samples, with different proportions of GGBS:Fly ash, shown in Table 4. Table 4 highlights that in the BOF slag-based geopolymer mortar sample, the excessive GGBS or the excessively high FA content has a high expansion rate. Moreover, with a ratio of GGBS:FA = 6:4, the volume expansion rate is 0.53% after 56 days of curing, but the GGBS:FA = 3:7, after 56 days of curing, has an expansion rate as high as 0.71%. Only between the GGBS:FA = 5:5, an expansion rate of less than 0.1% exists. The linear expansion ratio and the diameter expansion ratio were less than 0.3%.

**Table 4.** The effect of GGBS/FA Ratio on the expansion ratio of BOF slag-based geopolymer mortar after autoclave test.


Similarly, in the ratio of GGBS:FA = 5:5, both the diameter and linear expansion rate were less than 0.03%. Although the FA in the system can provide more silicon to inhibit the expansion of the BOF slag, the powder itself has large shrinkage and insufficient strength, thus cannot add too much. On the other hand, in the sample with higher GGBS content, although the content of silicon—which can react with free lime—is reduced, it is still su fficient to inhibit the expansion of the BOF slag due to its own strength, and its volume expansion is still 0.53%. In the GGBS:FA = 5:5, the strength provided by the GGBS in the system and the FA that can react to inhibit free calcium reached the optimum amount, and consequently the expansion rate and age are the best, <0.1% for 7–56 days curing.

#### *4.4. Laboratory Horizontal Double Shaft Mixer Tests*

This phase of the test is mainly to simulate the large-scale test of the actual plant. GGBS and FA (5:5) are used as source materials. The alkaline liquid has NaOH concentration of 6M, the SiO2/Na2O molar ratio is 1.5, the SiO2/Al2O3, as well as molar ratio, is 50. The moisture content of the BOF slag is controlled at 10%, where the BOF slag is not pretreated before the tests. To simulate the actual factory test, this experiment prepared a BOF slag-based geopolymer mortar with a horizontal biaxial mixing machine and poured a test specimen of Φ10 cm×20 cm for the compression test and the autoclave expansion test. The mixture proportion and process are shown in Table 5. Table 6 shows the hardening time of the BOF slag-based geopolymer mortar. According to the hardening time test results, the initial setting time is about 3 h; the final setting time is about 9 h.

**Table 5.** Simulation the large-scale experiment for BOF slag-based geopolymer mortar.


**Table 6.** Hardening time result of simulation of large-scale experimental for BOF slag-based geopolymer mortar.


The compressive strength of the simulated large-scale experimental BOF slag-based geopolymer mortar is shown in Figure 6. As the curing age increases, the strength of the test body increases. At seven days of age, the strength reached about 27 MPa; at 28 days of age, the strength reached 40 MPa. It is said that the simulation of the large-scale experimental of BOF slag-based geopolymer mortar has excellent strength performance.

**Figure 6.** The compressive strength of simulation large-scale experimental BOF slag-based geopolymer mortar.

The autoclave expansion test of the simulated large-scale experimental BOF slag-based geopolymer mortar is shown in Table 7. After the autoclave expansion test, the BOF slag-based geopolymer mortar sample is complete, and only a slight surface is peeling. The compressive strength test before and after autoclave found that the surface peeling did not affect the stability of the test specimen.


**Table 7.** Autoclave expansion test (seven days) results of simulating the large-scale of the experimental BOF slag-based geopolymer mortar.

#### *4.5. BOF Slag-Based Geopolymer Mortar Tests in Ready-Mixed Plant*

4.5.1. Ready-Mixed Plant Small Scale Tests

The mixed proportion for Ready-mixed plant small scale test is shown in Table 8. The main difference between the two tests is the ratio of binder and aggregate, which is the additional amount of the BOF slags.


**Table 8.** Mixture proportion of ready-mixed plant small-scale test.

The fresh properties of the ready-mixed plant small-scale test mixture are shown in Table 9. According to the results, it is found that the BOF slags of the high-water content of BOF slag-based geopolymer mortar showed excellent workability. The initial slump flow of Test-1 is 540\*580 mm, and it is allowed to stand for 45 min after, and the slump flow is reduced to only 520\*520 mm. Test-2s initial slump flow is 470\*480 mm, and after allowing it to stand for 45 min, its slump flow is reduced to 440\*440 mm. The results of the two tests were found to have low slump flow loss performance.

**Table 9.** Fresh properties mixture after ready-mixed plant small scale test.

The compressive strength of the ready-mixed plant small scale tests is shown in Table 10 and Figure 7. As the curing age increases, the strength of the test body increases. The strength of the late Test-2 is higher than that of Test-1 because the strength source of the geopolymer system is from the geopolymer slurry rather than an aggregate. Therefore, in Test-2, where the BOF fine aggregate is relatively low, the compressive strength will be higher than Test-1. However, after the autoclave test, the compressive strength decreases in Test-2. The reason for this is due to water release and formed cracks.

**Figure 7.** Compressive strength of the ready-mixed plant small scale test.


**Table 10.** Compressive strength of concerning plant small scale test.

#### 4.5.2. Ready-Mixed Plant Pilot-Scale Tests

Ready-mixed plant pilot-scale test parameters are shown in Table 11. The ratios of FA and GGBS are 5:5 and 6:4, the ratios of binder and aggregate are 1:2.936 and 1:3.575. The total test volume is 1.5 cubic meters, and the total weight in each test is approximately 3.5–3.6 tons.

**Table 11.** Mixture proportion of ready-mixed plant pilot-scale test.


The fresh properties of the ready-mixed plant pilot-scale tests are shown in Table 12. The test number of Test-3 and Test-4 have a slump flow of 380\*390 mm and 510\*490 mm, respectively. The reason for this di fference is the increase in the use of FA in Test-4 samples, which is spherical in shape and contributes to its fluidity [32,33].

**Table 12.** Fresh properties of ready mixed plant pilot-scale tests.

The compressive strength and their autoclave test of ready-mixed plant pilot-scale experiments are shown in Table 13 and Figure 8. As the curing age increases, the strength of the test body increases. The compressive strength of Test-3 is slightly higher than that of Test-4. The main reason is the adjustment of the ratio of FA to GGBS. Test-4 is higher in the amount of FA, and the reactivity of FA itself is poorer than that of GGBS powder which causes its intensity to be slightly lower than Test-3. According to the results of the autoclave expansion test, as the curing age increases, the structure of the test body is more complete, and the test specimen has no break point after the autoclave expansion test. The expansion changes after the autoclave test for the ready-mixed plant pilot-scale tests are shown in Table 14. All the expansion test results can be controlled around −0.41% for diameter and liner changes.


**Table 13.** Compressive strength and autoclave test of ready-mixed plant pilot-scale tests.

**Table 14.** Expansion changes after autoclave test for ready-mixed plant pilot-scale tests (curing time 28 days).


**Figure 8.** Compressive strength of ready-mixed plant pilot-scale test.
