Factors Influencing the Properties of Extrusion-Based 3D-Printed Alkali-Activated Fly Ash-Slag Mortar

The mix proportioning of extrusion-based 3D-printed cementitious material should balance printability and hardened properties. This paper investigated the effects of three key mix proportion parameters of 3D-printed alkali-activated fly ash/slag (3D-AAFS) mortar, i.e., the sand to binder (s/b) ratio, fly ash/ground granulated blast-furnace slag (FA/GGBS) ratio, and silicate modulus (Ms) of the activator, on extrudability, buildability, interlayer strength, and drying shrinkage. The results showed that the loss of extrudability and the development of buildability were accelerated by increasing the s/b ratio, decreasing the FA/GGBS ratio, or using a lower Ms activator. A rise in the s/b ratio improved the interlayer strength and reduces the drying shrinkage. Although increasing the FA/GGBS mass ratio from 1 to 3 led to a reduction of 35% in the interlayer bond strength, it decreased the shrinkage strain by half. A larger silicate modulus was beneficial to the interlayer bond strength, but it made shrinkage more serious. Moreover, a simple centroid design method was developed for optimizing the mix proportion of 3D-AAFS mortar to simultaneously meet the requirements of printability and hardened properties.

Unlike the conventional building process, 3D-printed concrete is a from-work free construction technique. Hence, the first challenge of this technology is to prepare concrete materials that are compatible with printing technology. That is, the mixture needs good fluidity and self-supporting capacity to ensure that the continuous concrete can be extruded through the nozzle and bears the load generated by subsequent concrete layers [25,26]. The printability of concrete is mainly characterized by extrudability and buildability [17].
Materials 2022, 15, 1969 3 of 18 Trial and error or single factor variable methods are often used to optimize the mix proportion of 3D-printed concrete. Obviously, these design methods cannot take printability, interlayer bonding, and volume stability into consideration simultaneously. These three properties are the most important properties for 3D-printed geopolymers. In addition, the properties of geopolymers may vary greatly due to the regional nature of geopolymer raw materials. This brings much more difficulties to the design of 3D-printed geopolymers. The interlayer bond strength and volume stability, which are also extremely important for the engineering application of 3D printing materials, are usually not considered in mix proportioning. Therefore, it is necessary to develop a mix proportioning method of 3D printing geopolymer mixtures, which can take printability, interlayer bond strength, and volume stability into consideration at the same time. This will greatly help to promote the application of this eco-friendly material in 3D printing technology.
For a full understanding of the relationship between the composition of 3D-printed alkali-activated fly ash/slag (3D-AAFS) mortar and its properties, this study comprehensively investigated the influences of the sand-to-binder ratio, the relative proportion of FA-GGBS precursors, and the silicate modulus of the activator on the printability, interlayer bond strength, and volume stability of 3D-AAFS mortar for the first time. Moreover, a simple centroid design method was developed for mix proportioning of extrusionbased 3D-AAFS mortar to strike a balance among printability, interlayer bond strength, and volume stability. This study enriched the mix design concept for 3D-printed alkaliactivated materials.

Raw Materials and Sample Preparation
In this study, Grade I fly ash (FA) powder in compliance with GB/T 1596-2017 [61] and Grade 95 ground granulated blast furnace slag powder (GGBS) in accordance with GB/T 18046-2017 [62] were used to prepare 3D-printed alkali-activated fly ash/slag (3D-AAFS) mortars. Their chemical compositions were determined by X-ray fluorescence (Axios mAX, PANalytical, Almelo, The Netherlands) and are given in Table 1. The density and Blaine specific surface area were 2350 kg/m 3 and 263 m 2 /kg for fly ash and 2860 kg/m 3 and 487 m 2 /kg for GGBS. The fine aggregate used was river sand with a fineness modulus of 2.47. The particle size distributions of powder materials and sand are shown in Figure 1. The alkaline activator was prepared using sodium hydroxide pellets (analytical grade, purity ≥ 98%), liquid sodium silicate with an original silicate modulus (Ms = SiO 2 /Na 2 O) of 3.1 (water content of 62%), and distilled water to achieve different silicate moduli (i.e., 0, 0.5, and 1). 3D-AAFS mortars with a water-to-binder (w/b) ratio of 0.35 and sand-to-binder (s/b) ratio varying from 0.8 to 1.2 were prepared. The water contained in liquid sodium silicate was considered in the total mixing water. The mix proportions of 3D-AAFS mortars are presented in Table 2. According to the designed silicate modulus shown in Table 2, the alkaline solution was prepared 2 h before the experiment and cooled to room temperature (25 ± 3 • C). The powder materials were mixed thoroughly in a Turbula shaker (WAB AG, Basel, Switzerland) for 24 h. Before preparing the 3D-AAFS mortars, the dry-mixed powder and sand were introduced into the mixer, and then the alkaline solution was added and stirred at 500 rpm for 4 min. 3D-AAFS mortars with a water-to-binder (w/b) ratio of 0.35 and sand-to-binder (s/b ratio varying from 0.8 to 1.2 were prepared. The water contained in liquid sodium silicat was considered in the total mixing water. The mix proportions of 3D-AAFS mortars ar presented in Table 2. According to the designed silicate modulus shown in Table 2, th alkaline solution was prepared 2 h before the experiment and cooled to room temperatur (25 ± 3 °C). The powder materials were mixed thoroughly in a Turbula shaker (WAB AG Basel, Switzerland) for 24 h. Before preparing the 3D-AAFS mortars, the dry-mixed pow der and sand were introduced into the mixer, and then the alkaline solution was added and stirred at 500 rpm for 4 min.

Rheological Tests for AAFS Mortar
Rheological tests were performed by a Rheolab QC rheometer (Anton Paar, Graz Austria) with a cylindrical geometry of 41.94 mm in inner diameter. The type of rotato was ST22-4 V-40-SN20452, and the height and width of each blade were 40.00 mm and 22.00 mm, respectively. During the test, the temperature was kept at 25 °C using a wate bath.

Dynamic Test
The procedure of the dynamic yield stress test consisted of preshearing at 100 s −1 fo

Rheological Tests for AAFS Mortar
Rheological tests were performed by a Rheolab QC rheometer (Anton Paar, Graz, Austria) with a cylindrical geometry of 41.94 mm in inner diameter. The type of rotator was ST22-4 V-40-SN20452, and the height and width of each blade were 40.00 mm and 22.00 mm, respectively. During the test, the temperature was kept at 25 • C using a water bath.

Dynamic Test
The procedure of the dynamic yield stress test consisted of preshearing at 100 s −1 for 60 s, resting for 15 s, ramping up from 0 to 100 s −1 in 60 s, stopping for 30 s, and then ramping down from 100 to 0 s −1 in 60 s. The Herschel-Bulkley model (H-B model, Equation (1)) was used to characterize the rheological behavior of the samples.
where τ is the shear stress (Pa), τ d,0 is the dynamic yield stress (Pa), K is the consistency coefficient (Pa·s), and n is the dimensionless fluidity index.

Static Test
Before the static yield shear test, the sample was presheared at 100 s −1 for 60 s to reach a consistent initial state [63]. The test was run at a constant shear rate of 0.02 s −1 for 60 s, the peak value of the measured shear stress was denoted as the static yield stress. The test was repeated every 5 min until the yield stress reached the upper limit value of the Materials 2022, 15, 1969 5 of 18 instrument. In this paper, the model (Equation (2)) proposed by Roussel [64] was used to fit the static yield stress data, which was defined as the measured peak shear stress.
where τ 0,0 and A thix are the static yield stress and the structural build-up rate, respectively. Table Test The flowability of the AAFS mortar was evaluated by using a flow table test on the basis of ASTM C230 [65]. All specimens were tested every 10 min after mixing. It has been proven that the spread diameter determined by the flow table test is related to its pumpability and extrudability [34,66,67]. In this work, it was found that the mixture with a spread diameter less than 200 mm cannot be continuously extruded from the 3D printer.

Buildability Test
A self-designed device was used to measure the buildability of 3D-AAFS mortar. The details of the device and operations information can be found in our previous work [33]. In this method, the demolded AAFS mortar specimen with a diameter of 50 mm was subjected to a load, which equaled 20 times the weight of the specimen. The deformation under this 20-layer load was recorded and used to evaluate the buildability. The buildability of mortar is acceptable when its deformation is smaller than 0.2% [33].

Preparation of 3D-AAFS Mortar Specimen
For mortar printing, a lab-scale 3D concrete printer with a round nozzle of 30 mm diameter, introduced by earlier works [31,63,64], was employed in this study. In this work, the printing speed was set at 30 mm/s, and the printing interval time was controlled at 5 min. All the samples for mechanical tests and drying shrinkage tests were cut from the printing specimen. The 300 × 75 × 75 mm 3 blocks were printed. Immediately after printing, both ends of the printed block were cut off, and the middle part of block (285 × 75 × 75 mm 3 ) was used for the drying shrinkage test, and the sample was cured at 25 ± 2 • C and 50 ± 5% relative humidity. For the interlayer bond strength test, the printed blocks were cured at 25 • C and 98% relative humidity for 2 days and then cut into 70 mm × 70 mm × 70 mm cubes. The cubes were cured at 25 • C and 98% relative humidity.

Interlayer Bond Strength Test
The interlayer bond strengths of the specimen at 3, 7, 28, and 90 days were measured by the slant shear bond strength, as shown in Figure 2. In the slant shear strength test, the load was directly applied to the mold, and the inclined angle of the specimen was set to 60 degrees [68,69]. The loading rate was 2.4 kN/s. The interlayer shear stress (τ) was calculated by Equation (3).
where τ (Pa) is the interfacial shear stress, P (N) is the critical load of interface sliding, and A is the bond area of the specimen.

Drying Shrinkage
The drying shrinkage test in this study was carried out according to ASTM C157 [70]. For the measurement of drying shrinkage strains, the 3D AAFS mortar specimen was placed in the shrinkage frame in the vertical position, and an electronic dial gauge was mounted on the top, as shown in Figure 3. The specimens were kept at a controlled temperature of 25 ± 2 • C and relative humidity of 50 ± 5%. The dial gauge reading was recorded at the age of 1, 2, 3,5,7,9,11,14,21,28,56, and 90 days. The shrinkage strain of the mix reported in the paper is the average of three specimens. where τ (Pa) is the interfacial shear stress, P (N) is the critical load of interface sliding, A is the bond area of the specimen.

Drying Shrinkage
The drying shrinkage test in this study was carried out according to ASTM C157 For the measurement of drying shrinkage strains, the 3D AAFS mortar specimen placed in the shrinkage frame in the vertical position, and an electronic dial gauge mounted on the top, as shown in Figure 3. The specimens were kept at a controlled perature of 25 ± 2 °C and relative humidity of 50 ± 5%. The dial gauge reading was orded at the age of 1, 2, 3, 5, 7, 9, 11, 14, 21, 28, 56, and 90 days. The shrinkage strain o mix reported in the paper is the average of three specimens.

Drying Shrinkage
The drying shrinkage test in this study was carried out according to ASTM C157 For the measurement of drying shrinkage strains, the 3D AAFS mortar specimen placed in the shrinkage frame in the vertical position, and an electronic dial gauge mounted on the top, as shown in Figure 3. The specimens were kept at a controlled perature of 25 ± 2 °C and relative humidity of 50 ± 5%. The dial gauge reading was orded at the age of 1, 2, 3, 5, 7, 9, 11, 14, 21, 28, 56, and 90 days. The shrinkage strain o mix reported in the paper is the average of three specimens.     [71]. Figure 4a reveals that the τ d,0 value of AAFS mortar is sensitive to the s/b ratio. The rise of the s/b ratio leads to a significantly higher τ d,0 . As seen, the τ d,0 of P1-S1.2-M0 is nearly twice the value of P1-S0.8-M0. The increase of s/b ratio results in a higher solid volume fraction and the increase of particle friction, which is mainly responsible for the greater dynamic yield stress [71]. Moreover, Figure 4b indicates that the FA/GGBS ratio also affects τ d,0 of the mortar. The value shows a slight decline when the FA/GGBS ratio increases from 1 to 3 due to the spherical geometry and smooth surface of FA particles [71][72][73][74]. Figure 4c shows that τ d,0 has a 40% reduction when the silicate modulus of the activator solution increases from 0 to 0.5. This suggests that the presence of sodium silicate in the activator solution greatly lowers τ d,0 compared to the AAFS mortar activated only by NaOH (P1-S1-M0).

Rheological Behavior
An increase in the silicate modulus in the NaOH + sodium silicate activator solution can further reduce τ d,0 of the AAFS mortar. It is attributed to the stronger repulsive electrostatic force of particles caused by the increase of silicate modulus of activator solution [72,73]. over, Figure 4b indicates that the FA/GGBS ratio also affects τd,0 of the mortar. The v shows a slight decline when the FA/GGBS ratio increases from 1 to 3 due to the sphe geometry and smooth surface of FA particles [71][72][73][74]. Figure 4c shows that τd,0 has a reduction when the silicate modulus of the activator solution increases from 0 to 0.5. suggests that the presence of sodium silicate in the activator solution greatly lower compared to the AAFS mortar activated only by NaOH (P1-S1-M0). An increase in silicate modulus in the NaOH + sodium silicate activator solution can further reduc of the AAFS mortar. It is attributed to the stronger repulsive electrostatic force of part caused by the increase of silicate modulus of activator solution [72,73].   (2)) was employed to fit these d and the results are given in Table 3. The results show that mortar with a larger s/b has a greater Athix, indicating that a higher s/b ratio is beneficial for the structural bui of AAFS mortar. By comparing the Athix values of P1-S0.8-M0, P2-S0.8-M0, and P3-M0, it can be found that the increase in FA proportion dramatically slows down the st tural buildup rate of mortar. The Athix value of the mortar with an FA/GGBS mass rat 3 is only one quarter of the value for the mortar with a ratio of 1 due to the lower diss tion rate of the glassy structure of FA as compared to GGBS [74]. In addition, the struct buildup rates of those activated by NaOH+sodium silicate are much slower than t activated by NaOH. Increasing the silicate modulus of the activator from 0.5 to 1 slig decreased Athix. This means that the increase in silicate modulus is unfavorable for structure formation of AAFS mortar, which is owing to the retardation effect of the la Ms on the formation of reaction products [73,74].   (2)) was employed to fit these data, and the results are given in Table 3. The results show that mortar with a larger s/b ratio has a greater A thix , indicating that a higher s/b ratio is beneficial for the structural buildup of AAFS mortar. By comparing the A thix values of P1-S0.8-M0, P2-S0.8-M0, and P3-S0.8-M0, it can be found that the increase in FA proportion dramatically slows down the structural buildup rate of mortar. The A thix value of the mortar with an FA/GGBS mass ratio of 3 is only one quarter of the value for the mortar with a ratio of 1 due to the lower dissolution rate of the glassy structure of FA as compared to GGBS [74]. In addition, the structural buildup rates of those activated by NaOH+sodium silicate are much slower than those activated by NaOH. Increasing the silicate modulus of the activator from 0.5 to 1 slightly decreased A thix . This means that the increase in silicate modulus is unfavorable for the structure formation of AAFS mortar, which is owing to the retardation effect of the larger Ms on the formation of reaction products [73,74]. Figure 6a,b exhibit the changes in fluidity and buildability of AAFS mortar with time, respectively. The spread diameter measured by the flow table has been proven to be useful for predicting the extrudability of cementitious material [66,67]. A larger spread diameter corresponds to a better extrudability [66]. It can be seen from Figure 6a that the initial spread diameters of all mortars (at 5 min) are larger than 200 mm, and all of them can be continuously extruded from the nozzle. The initial spread diameter of the AAFS mortar increases with decreasing s/b ratio, increasing FA/GGBS ratio, and increasing silicate modulus of the alkaline solution. As seen in Figure 7, the spread diameter of mortar shows a good negative correlation with its dynamic yield stress. Therefore, the effects of these Materials 2022, 15, 1969 8 of 18 three factors on the dynamic yield stress of AAFS mortar should be responsible for their influences on the initial extrudability. Moreover, the spread diameter of mortar with a larger s/b ratio, a lower FA proportion, or a smaller silicate modulus of activator solution reduces faster to be smaller than 200 mm (the critical value corresponding to the acceptable extrudability) earlier.   Figure 6a,b exhibit the changes in fluidity and buildability of AAFS mortar with t respectively. The spread diameter measured by the flow table has been proven to be ful for predicting the extrudability of cementitious material [66,67]. A larger spread di eter corresponds to a better extrudability [66]. It can be seen from Figure 6a that the in spread diameters of all mortars (at 5 min) are larger than 200 mm, and all of them ca continuously extruded from the nozzle. The initial spread diameter of the AAFS mo increases with decreasing s/b ratio, increasing FA/GGBS ratio, and increasing sili modulus of the alkaline solution. As seen in Figure 7, the spread diameter of mortar sh a good negative correlation with its dynamic yield stress. Therefore, the effects of t three factors on the dynamic yield stress of AAFS mortar should be responsible for t influences on the initial extrudability. Moreover, the spread diameter of mortar wi larger s/b ratio, a lower FA proportion, or a smaller silicate modulus of activator solu reduces faster to be smaller than 200 mm (the critical value corresponding to the acce ble extrudability) earlier.   Figure 6b shows that the initial deformations of all AAFS mortars are greater than 0.2%, which indicates that the buildability levels of all specimens are not acceptable. The deformation value decreases with time, reflecting the improvement of buildability. The increase in the s/b ratio has a positive effect on the buildability of mortar, and mortar with a high sand content reaches an acceptable buildability quicker. The buildability shows an insignificant difference as the FA/GGBS mass ratio increases from 1 to 2 during the whole measurement period. However, the buildability of mortar with an FA/GGBS ratio of 3 is slightly poorer than that for a lower FA/GGBS ratio. The increase in the silicate modulus of the activator solution greatly harms the initial value and the development rate of buildability. Figure 8 plots the static yield strengths of AAFS mortars vs. their deformation. The result shows that the buildability of mortar correlates well to its static yield strength. A higher static yield strength responds to a greater capacity to resist the weight of subsequently extruded layers. The effects of the s/b ratio, FA/GGBS ratio, and silicate modulus on the buildability should be attributed to their effects on the static yield stress. According to the result shown in Figure 8, the buildability of AAFS mortar is acceptable when the static yield stress is larger than~3.7 kPa. The higher structuration rate (A thix ) resulted from an increase in the s/b ratio, a decrease in the FA/GGBS ratio, or a decrease in the silicate modulus, which accelerated the development of AAFS mortar buildability to meet the requirements of 3D printing manufacturing.  Figure 6b shows that the initial deformations of all AAFS mortars are greater than 0.2%, which indicates that the buildability levels of all specimens are not acceptable. The  Figure 6b shows that the initial deformations of all AAFS mortars are greater than 0.2%, which indicates that the buildability levels of all specimens are not acceptable. The on the buildability should be attributed to their effects on the static yield stress. According to the result shown in Figure 8, the buildability of AAFS mortar is acceptable when the static yield stress is larger than ~3.7 kPa. The higher structuration rate (Athix) resulted from an increase in the s/b ratio, a decrease in the FA/GGBS ratio, or a decrease in the silicate modulus, which accelerated the development of AAFS mortar buildability to meet the requirements of 3D printing manufacturing. In this study, tE was defined as the duration time AAFS mortar maintains extrudabil ity, which can be estimated by the time the spread diameter becomes smaller than 200 mm. The tB was defined as the needed time for obtaining good buildability, corresponding to the time at which the deformation under a 20-layer load decreased to less than 0.2% Table 4 lists the tE and tB of AAFS mortars estimated from Figure 6a,b. Both tE and tB de termine the open time for the 3D printing process. The total operation time of 3D printing should be shorter than the tE owing to the extrudability requirement. Moreover, to ensure that the bottom layers, especially the first layer, can endure the weight of subsequent lay ers being deposited on the top without excessive distortion and failures, the start time o 3D printing is suggested to be later than tB. Table 4 shows that an increase in the s/b ratio not only shortens tE but also reduces tB. Increasing the FA/GGBS ratio of the binder or the silicate modulus of the activator extends the tE and prolongs the tB. Increasing the FA/GGBS ratio of the binder or the silicate modulus of the activator retards the reaction products formation, which reduces the structural buildup of mortar and thus extends the tE and prolongs the tB, as seen in Table 4. In this study, t E was defined as the duration time AAFS mortar maintains extrudability, which can be estimated by the time the spread diameter becomes smaller than 200 mm. The t B was defined as the needed time for obtaining good buildability, corresponding to the time at which the deformation under a 20-layer load decreased to less than 0.2%. Table 4 lists the t E and t B of AAFS mortars estimated from Figure 6a,b. Both t E and t B determine the open time for the 3D printing process. The total operation time of 3D printing should be shorter than the t E owing to the extrudability requirement. Moreover, to ensure that the bottom layers, especially the first layer, can endure the weight of subsequent layers being deposited on the top without excessive distortion and failures, the start time of 3D printing is suggested to be later than t B . Table 4 shows that an increase in the s/b ratio not only shortens t E but also reduces t B. Increasing the FA/GGBS ratio of the binder or the silicate modulus of the activator extends the t E and prolongs the t B . Increasing the FA/GGBS ratio of the binder or the silicate modulus of the activator retards the reaction products formation, which reduces the structural buildup of mortar and thus extends the t E and prolongs the t B , as seen in Table 4.

Interlayer Bond Strength
The developments of the interlayer bond strength of the 3D-AAFS mortars are shown in Figure 9. A rapid increase in the interlayer bond strength of 3D-AAFS mortar can be seen at early stages. Most of the mortars reach more than 60% of the strength of 90 days within 7 days. The early interlayer bond strength is enhanced with an increase in the s/b ratio, and an increase in the s/b ratio from 0.8 to 1.2 leads to an increase in the strength of 1 MPa at 7 days. However, the influence of the s/b ratio on the interlayer strength of 3D-AAFS mortar at 90 days is very small. Conversely, the FA/GGBS mass ratio shows a significant impact on the interlayer bond strength of the printed specimen. The strength of the specimen has a 50% reduction at 7 days and a 40% reduction at 90 days due to the increase in the FA/GGBS ratio from 1 to 3. Figure 9c reveals that the printed mortar with a greater silicate modulus has a much higher interlayer bond strength after 7 days. As the silicate modulus increases from 0 to 1, the strength of the printed specimen is finally increased by~2 MPa. seen at early stages. Most of the mortars reach more than 60% of the strength of 90 days within 7 days. The early interlayer bond strength is enhanced with an increase in the s/b ratio, and an increase in the s/b ratio from 0.8 to 1.2 leads to an increase in the strength of 1 MPa at 7 days. However, the influence of the s/b ratio on the interlayer strength of 3D-AAFS mortar at 90 days is very small. Conversely, the FA/GGBS mass ratio shows a significant impact on the interlayer bond strength of the printed specimen. The strength of the specimen has a 50% reduction at 7 days and a 40% reduction at 90 days due to the increase in the FA/GGBS ratio from 1 to 3. Figure 9c reveals that the printed mortar with a greater silicate modulus has a much higher interlayer bond strength after 7 days. As the silicate modulus increases from 0 to 1, the strength of the printed specimen is finally increased by ~2 MPa. In summary, the increase in the s/b ratio only improves the interlayer bond strength at the early stage, but it insignificantly influences the final strength. Increasing the FA/GGBS mass ratio negatively affects the interlayer bond strength, while enlarging the silicate modulus is positive to improve the strength.

Drying Shrinkage
Figures 10a-c display the changes in the drying shrinkage strain of 3D-AAFS mortars with time. The results demonstrate that the drying shrinkage of the printed specimen greatly depends on the s/b ratio and FA/GGBS mass ratio. The effect of the silicate modulus on the drying shrinkage is relatively slighter than that of the other factors. The strains caused by drying shrinkage for P1-S0.8-M0 and P1-S1-M0 were similar within 90 days. In summary, the increase in the s/b ratio only improves the interlayer bond strength at the early stage, but it insignificantly influences the final strength. Increasing the FA/GGBS mass ratio negatively affects the interlayer bond strength, while enlarging the silicate modulus is positive to improve the strength.

Drying Shrinkage
Figure 10a-c display the changes in the drying shrinkage strain of 3D-AAFS mortars with time. The results demonstrate that the drying shrinkage of the printed specimen greatly depends on the s/b ratio and FA/GGBS mass ratio. The effect of the silicate modulus on the drying shrinkage is relatively slighter than that of the other factors. The strains caused by drying shrinkage for P1-S0.8-M0 and P1-S1-M0 were similar within 90 days. This means that increasing the s/b ratio from 0.8 to 1 does not influence the drying shrinkage degree of mortar. However, further increasing the s/b ratio to 1.2 effectively reduces the drying shrinkage after 14 days. At 90 d, the drying shrinkage strain decreased from 0.12% to 0.08%. The 3D-AAFS mortar with a larger FA content shows a clear advantage in reducing drying shrinkage. Figure 10b shows that the dry shrinkage of the printed specimen with an FA/GGBS ratio of 1 is nearly 1.5 times larger than that for an FA/GGBS ratio of 2 and three times greater than that for an FA/GGBS ratio of 3 after 21 days. The variation in silicate modulus barely changes the drying shrinkage level of 3D-AAFS mortar before 28 days. After 28 days, the drying shrinkage increases slightly with increasing silicate modulus.

Mixture Design of 3D-AAFS Mortar Using the Simplex Centroid Design Method
Generally, the sand-to-binder ratio, the mass ratio of FA/GGBS, and the silicate modulus of the activator are the three most important parameters considered in the mix design process of alkali-activated fly ash/slag materials. These three parameters have been proven by many studies [75][76][77][78] to determine the workability, mechanical properties, and durability of mixtures for traditional engineering applications. 0.12% to 0.08%. The 3D-AAFS mortar with a larger FA content shows a clear advantage in reducing drying shrinkage. Figure 10b shows that the dry shrinkage of the printed specimen with an FA/GGBS ratio of 1 is nearly 1.5 times larger than that for an FA/GGBS ratio of 2 and three times greater than that for an FA/GGBS ratio of 3 after 21 days. The variation in silicate modulus barely changes the drying shrinkage level of 3D-AAFS mortar before 28 days. After 28 days, the drying shrinkage increases slightly with increasing silicate modulus.

Mixture Design of 3D-AAFS Mortar Using the Simplex Centroid Design Method
Generally, the sand-to-binder ratio, the mass ratio of FA/GGBS, and the silicate modulus of the activator are the three most important parameters considered in the mix design process of alkali-activated fly ash/slag materials. These three parameters have been proven by many studies [75][76][77][78] to determine the workability, mechanical properties, and durability of mixtures for traditional engineering applications.
The mix design of alkali-activated fly ash/slag materials for 3D printing manufacturing is more complex than that for conventional applications. Printable alkali-activated fly ash/slag materials should meet extra fresh-state and hardened-state requirements, including pumpability, extrudability, buildability, and interlayer strength. Moreover, the drying shrinkage of alkali-activated materials is more serious than that of Portland cement-based materials [78]. The cracks caused by drying shrinkage will threaten the appearance, quality, and even safety of 3D-printed alkali-activated material construction. Therefore, drying shrinkage should be a crucial hardened-state requirement if alkali-activated materials are used for 3D printing.
According to Sections 3.2-3.4, the mix proportion parameters, including the s/b ratio, FA/GGBS ratio, and silicate modulus, exert great influences on the printability, interlayer bond strength, and drying shrinkage of the 3D-printed alkali-activated fly ash/slag mortar. An increase in the s/b ratio shortens tE and tB, improves the interlayer bond strength, and reduces the drying shrinkage of the 3D-AAFS mortar. The rise of the FA/GGBS mass ratio of the binder leads to longer tE and tB, a weaker interlayer bond strength, and a smaller drying shrinkage. In addition, the increase in silicate modulus of the activator The mix design of alkali-activated fly ash/slag materials for 3D printing manufacturing is more complex than that for conventional applications. Printable alkali-activated fly ash/slag materials should meet extra fresh-state and hardened-state requirements, including pumpability, extrudability, buildability, and interlayer strength. Moreover, the drying shrinkage of alkali-activated materials is more serious than that of Portland cement-based materials [78]. The cracks caused by drying shrinkage will threaten the appearance, quality, and even safety of 3D-printed alkali-activated material construction. Therefore, drying shrinkage should be a crucial hardened-state requirement if alkali-activated materials are used for 3D printing.
According to Sections 3.2-3.4, the mix proportion parameters, including the s/b ratio, FA/GGBS ratio, and silicate modulus, exert great influences on the printability, interlayer bond strength, and drying shrinkage of the 3D-printed alkali-activated fly ash/slag mortar. An increase in the s/b ratio shortens t E and t B , improves the interlayer bond strength, and reduces the drying shrinkage of the 3D-AAFS mortar. The rise of the FA/GGBS mass ratio of the binder leads to longer t E and t B , a weaker interlayer bond strength, and a smaller drying shrinkage. In addition, the increase in silicate modulus of the activator extends the t E , prolongs the t B , enhances the interlayer bond strength, and enlarges the drying shrinkage.
Given that all the properties of 3D-AAFS mortar are closely related to the three mix proportion parameters, it is difficult to optimize the mix proportion to satisfy all the requirements. The simple-centroid design method [79] allows us to investigate the properties of mixtures simultaneously controlled by three factors.
In the simple-centroid design method, if there are n variables, 2n−1 groups of tests are needed to obtain the corresponding contour map. For 3D-AAFS mortar, the factors s/b ratio, FA/GGBS ratio, and silicate modulus are independent of each other. Therefore, a linear substitution calculation method is proposed, which transforms the actual value of variables into the equivalent value with the sum of 100%. The corresponding calculation formulas and values are shown in Table 5 and Equations (4)- (6). Figure 11 describes the seven test points of the simplex centroid design method with three variables, namely, the vertex, midpoint, and center of the triangle.
where x i , z i and ∆z i are the equivalent value, the actual value of the variable, and the maximum difference between actual values, respectively. Table 5. The actual value and linear substitution of three factors. Figure  seven test points of the simplex centroid design method with three variables, namely, the vertex, midpoint, and center of the triangle. where , and ∆ are the equivalent value, the actual value of the variable, and the max imum difference between actual values, respectively.  Figure   Based on the experimental results of the properties shown in Table 6, the contours of t E , t B , and 90-day interlayer bond strength and 90-day drying shrinkage of 3D-AAFS mortar related to the s/b ratio, FA/GGBS mass ratio, and silicate modulus are drawn and shown in Figure 12a-d. These contours can not only be used to predict the performances of 3D-AAFS mortar through its mix proportion, but can also be used to optimize the mix proportion to achieve the designed properties. For example, the interlayer bond strength of 3D printing material was suggested to be higher than~6 MPa, and the dry shrinkage of alkaline-activated materials should be lower than~0.09% [43,54,56,60,[80][81][82][83]. In practice, t E is always required to be larger than 1 h [29,80]. From the contours of each property, three critical lines of each property could be acquired to meet these required values, as shown in Figure 13. The overlapping area of three different areas in Figure 13 is regarded as the optimal mix proportion of AAFS mortar for 3D printing manufacturing, i.e., the optimum s/b ratio, FA/GGBS mass ratio, and silicate modulus are 0.8~1.0, 2.2~2.6, and 0.5~1.0, respectively. tE is always required to be larger than 1 h [29,80]. From the contours of each property, three critical lines of each property could be acquired to meet these required values, as shown in Figure 13. The overlapping area of three different areas in Figure 13 is regarded as the optimal mix proportion of AAFS mortar for 3D printing manufacturing, i.e., the optimum s/b ratio, FA/GGBS mass ratio, and silicate modulus are 0.8~1.0, 2.2~2.6, and 0.5~1.0, respectively. P2-S1-M0 55 20 6.7 ± 0.28 0.09 ± 0.006 6 P1-S1-M0.5 110 40 9.7 ± 0.14 0.12 ± 0.011 7 P1.7-S1.1-M0.

Conclusions
The following conclusions can be summarized from the presented findings: (1) The composition of alkali-activated fly ash/slag (AAFS) mortar exerts a tremendous Figure 13. Optimization of the mix proportion of AAFS mortar for 3D printing manufacture.

Conclusions
The following conclusions can be summarized from the presented findings: (1) The composition of alkali-activated fly ash/slag (AAFS) mortar exerts a tremendous influence on the printability of AAFS mortar, which relates closely to the rheological properties. The printability of AAFS mortar relates closely to the rheological properties. The increase in the s/b ratio enlarges the dynamic yield stress and accelerates the structural buildup of mortar, resulting in a faster loss rate of extrudability and a quicker growth rate of buildability. Conversely, increasing the FA/GGBS mass ratio or the silicate modulus reduces both the dynamic yield stress and structuration rate, which extends the duration time of extrudability and slows down the development of buildability. The printability of AAFS mortar is most sensitive to the silicate modulus of the activator. (2) The hardened-state properties of 3D-printed AAFS mortar also depend on its mix proportion. Increasing the s/b ratio is conducive to improving the interlayer bond strength and diminishing the drying shrinkage. The rise in the FA/GGBS mass ratio weakens the interlayer bond strength and reduces the drying shrinkage. The use of an activator with a larger silicate modulus is beneficial to the interlayer bond strength, but it causes a slightly larger drying shrinkage. (3) A simple centroid design method was developed for mix proportioning of extrusionbased 3D-printed AAFS mortar for the first time, which took printability, interlayer bond strength, and drying shrinkage into consideration at the same time. By restricting the fresh-state and hardened-state requirements, the optimum mix proportion of 3D-AAFS mortar can be obtained using this method. Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.