Antagonistic Effects of Hydrated Lime and Calcium Formate on Early-Age Strength in High Volume Fly Ash Composites: Mechanisms and Engineering Implications

Abstract

The utilization of high-volume fly ash (HVFA, ≥50% cement replacement) in concrete is pivotal for sustainable construction but hindered by low early-age strength. This study investigates the individual and combined effects of hydrated lime (HL) and calcium formate (CF) on the strength development, hydration kinetics, and microstructure of HVFA pastes (60% and 70% FA). Individual additions of 11% HL (HVFA60) or 14% HL (HVFA70) raised 28-day compressive strength by 18% and 22%, respectively, and shortened final setting from 10.0 h to 3.8 h. Similarly, 3% CF increased 28-day strength by 15% (HVFA60) and 12% (HVFA70) while cutting final setting to 2.1 h and 3.3 h. In contrast, combining HL and CF suppressed strength by 15–22% despite accelerating final setting to less than 1 h. Isothermal calorimetry showed a 40% reduction in cumulative heat release at 44 h for the combined system. XRD, TGA and SEM confirmed 20–30% lower C-S-H content, 25% less CH, and a rise in porosity when HL and CF were used together. These findings demonstrate that HL and CF act as competing accelerators, where rapid heat release compromises microstructural integrity. For practical applications using HVFA materials, individual use of HL or CF is recommended to enhance early-age performance, while combined application should be avoided to prevent strength reduction.

1. Introduction

Concrete is one of the most widely used materials in construction. The major binding component of ordinary concrete is Portland cement (PC), whose production process results in 1.6 billion tonnes of greenhouse gas (GHG) emissions worldwide on an annual basis [1]. The cement industry accounts for nearly 8% of global GHG emissions [2]. To reduce the carbon footprint of concrete, the concrete industry is interested in using increasing volumes of other available pozzolanic materials such as fly ash [3].

Fly ash (FA), a type of supplementary cementitious material (SCMs), is a by-product of coal-fired power plants. Large volumes of FA are available in countries that rely on coal to produce electricity. There is a large volume of literature reporting the replacement of PC with FA. Concrete mixes with more than 50% of their PC content (by weight) substituted by FA are termed high-volume fly ash (HVFA) concrete [4]. The replacement of PC by FA improves the workability of concrete [5,6] especially beneficial for 3D printing building materials [7] and reduces the likelihood of thermal cracking [8,9]. The improvement of workability is mainly attributed to the round shape of FA contributing to the ball-bearing effect [10]. On the other hand, various studies have reported the replacement of PC with metakaolin, which significantly contributes to CO₂ emission reduction (by 20%) [11] while increasing the cost of concrete manufacturing by 17%, as reported by [12]. According to Lima et al. [12], incorporating construction and demolition waste as recycled aggregate and fly ash in cementitious materials can enhance the sustainability and cost-effectiveness of the resulting concrete. Due to reduced carbon footprint, some recent studies investigate the use of FA in construction material regarding different aspects like strength, durability and microstructure [13].

Despite these advantages, earlier research has indicated that substituting PC with FA reduces the early strength of HVFA concrete, with a higher proportion of replacement leading to a more significant decrease in the early strength [14]. The decreased early stage strength of HVFA concrete is mainly due to the lower rate of the pozzolanic reaction between FA and calcium hydroxide (CH) compounds [15]. Moreover, the decrease in the early strength of HVFA concrete is accompanied by an increase in the setting time [15], which affects the concrete construction process. Consequently, for concrete containing higher volumes of FA, it is necessary to increase its early strength and reduce its setting time, which are required for field applications.

The use of concrete admixtures is a major method to address the HVFA issue of low early age strength. Calcium formate (CF), a calcium salt of formic acid, is a prime example and is now primarily utilized as a preservative in fodder for all animal species in the European Union [16]. Initial investigations considering the effects of CF on the properties of concrete date back to the early 1980s, when Gebler [17] studied the effectiveness of CF as a chloride-free accelerating admixture. It was found that CF could effectively accelerate compressive strength development when the Calcium: Sodium (CA:SO) ratio was greater than four. A decade later, Gao et al. [18] found that CF could be mixed with ferric sulfate and crystal embryo to form a concrete hardening accelerator, which increased 3-days and 7-days strength of PC concrete by 47% and 39%, respectively. Ma et al. [19] studied the influence of adding 1.5% CF to concrete (w/cm ratio of 0.5) incorporated with 2–4% FA and found that the 1-day strength increased by 133.7% compared with that without CF. The increase in the early strength of PC concrete with the addition of CF could be attributed to the reduction in the pH of the liquid phase that promotes the hydration of C₃S [18]. Zhou et al. [20] investigated the effects of CF on the early age strength and microstructure of HVFA (60% and 70% replacement) pastes. They reported that a 3% addition of CF relative to the binder weight would be the optimal dosage for the HVFA mix. Adding CF to concrete can reduce waste disposal because CF is a by-product of the food industry. This may further decrease the use of PC to reduce CO₂ emissions. Previous research suggests that the incorporation of an appropriate amount of CF may improve the properties of concrete and contribute to the reduction of CO₂ emissions.

In addition to CF, hydrated lime (HL) or CH is another popular additive that can also be added to HVFA concrete to accelerate early age strength gain and reduce setting time. Bentz and Ferraris [20] reported that incorporating 5% HL into HVFA pastes (w/c = 0.3; 50:50 FA:PC by mass) significantly reduced setting times. For Class C FA, initial and final setting times decreased from 8.2 h to 5.3 h and from 8.8 h to 6.0 h, respectively. For Class F FA, the corresponding reductions were from 8.6 h to 5.2 h and from 10.2 h to 5.9 h. Nochaiya et al. [21] studied the influence of HL on concrete with 50% FA replacement and a w/b ratio of 0.5 and found that the addition of 5% HL not only reduced the setting time, but also increased the 7-day strength due to accelerated pozzolanic reactions at an early age. Similar results were also reported by George and Sofi [22], who investigated the effects of 5% HL supplementation in FA concrete with 30% PC replaced by FA. Gunasekara [1] used 13% to 18% HL (by mass of paste) and 1% to 3% nano SiO₂ (by mass of paste) for super HVFA concrete with up to 80% replacement, accomplishing comparable mechanical performance to pure Ordinary Portland cement (OPC) concrete, although it has limited practicality owing to its high cost.

Although studies on the effects of sole CF or HL on the early-stage strength development of HVFA concrete have been reported, there is a scarcity of information regarding the effects of combined CF and HL on HVFA concrete. Therefore, studies on the joint influence of CF and HL are necessary. HL is considered a low-cost material, and the addition of HL could further reduce the use of FA or PC, thus reducing the total cost of cement-based building materials. Ultimately, a reduction in the use of PC could further decrease carbon emissions. The addition of an appropriate amount of single CF or HL has positive effects on the early strength of HVFA concrete, and it is assumed that the addition of combined CF and HL may further promote strength development.

This study investigated the effects of admixtures consisting of CF and HL on the strength of HVFA pastes that contain 60% and 70% FA considering early age up to 28 days. The mixture of pastes (pure HL and the combination of HL and CF) was designed based on the calculation of the amount of HL required to fully react with FA and then the Pozzolanic Index (PI) was subsequently determined, which measures the reactivity of FA with calcium hydroxide (Ca(OH)₂) and water in concrete. X-ray diffraction analysis and scanning electron microscopy were used to characterize the microstructures of HVFA pastes containing CF and HL. Calorimetry and thermogravimetric analysis were used to characterize the reactivity in the HVFA paste incorporated with CF and HL.

2. Materials and Methodology

2.1. Materials and Mix Design

In this experimental program, Cement Australia (Queensland, Australia) supplied OPC, FA, CF, and HL. CF is an organic substance with the chemical formula C₂H₂O₄Ca. It was a white powder that was soluble in water. HL is an inorganic substance with the chemical formula Ca(OH)₂. It is a white powder that could not be completely dissolved in water.

Table 1. Chemical composition of PC and FA.

		PC	FA	HL
Chemical composition (%)	CaO	63.7	3.5	71.6
	SiO2	19.9	55	1.4
	Al2O3	4.6	27.2	0.7
	Fe2O3	2.57	9.2	0.3
	MgO	1.39	1.21	4.6
	K2O	0.69	0.8	0.05
	Na2O	0.59	1.28	0.4
	P2O5	0.04	0.55	-
	Mn2O3	0.06	0.12	-
	SrO	0.07	0.05	-
	TiO2	-	1.49	-
	SO3	2.7	0.11	1
	LOI	3.9	1.2	19.95
Mineral phases (%)	C3S	65.78	-	-
	C2S	18.04	-	-
	C3A	7.82	-	-
	C4AF	7.88	-	-

presents the chemical compositions of HL, PC, and FA, which were provided by Cement Australia and characterized by XRF (Bruker S2 PUMA). Figure 1 [20] shows the particle size distribution measured at the University of Melbourne using a Malvern Mastersizer 3000 (Malvern, UK).

The mix design for pastes with the addition of pure HL and the combination of HL and CF was developed based on the calculation of the amount of HL required to fully react with FA at the age of 28th day. This begins with the calculation of the Pozzolanic Index (PI). The PI describes the degree of FA reactivity with Ca (OH)₂ and water in concrete. The mix design was created following ASTM C 618 [23] and ASTM C 311 [24] in a water/binder ratio of 0.484, as presented in

Table 2. Mortar mix proportions (kg/m³).

Mix Notation	PC	FA	Sand	Water	28-Day Strength (MPa)
100PC	500	−	1375	242	Fc = 45.0
[80PC + 20FA]	400	100	1375	242	Fca = 39.9

Note: F_c = Strength of 100PC mortar; F_ca = Strength of [80PC + 20FA] mortar.

.

The calculation of the pozzolanic index (PI) follows the procedure described by Dunstan and Zayed [25] from Equations (1)–(8).

$$N=1.598\times C_v\times H_v=1.598\times \left({\displaystyle \frac{500}{3120}}\right)\times 0.676=0.173$$

(1)

$$D=\left(H_x{C}_v\right)+W_v=\left(0.676\times \left({\displaystyle \frac{500}{3120}}\right)\right)+0.242=0.350$$

(2)

$${\left({\displaystyle \frac{N}{D}}\right)}^3={\left({\displaystyle \frac{0.173}{0.350}}\right)}^3=0.121$$

(3)

$$S_F={\displaystyle \frac{F_c}{\left[2.143\times {\left(\frac{N}{D}\right)}^3\right]}}={\displaystyle \frac{6607.9}{2.143\times {\left(\frac{0.1731}{0.3503}\right)}^3}}=\mathrm{25,554.6}$$

(4)

$$K={\left[{\displaystyle \frac{F_c}{2.143\times S_F}}\right]}^{{\displaystyle \frac{1}{3}}}={\left({\displaystyle \frac{6607.9}{2.143\times 25554.6}}\right)}^{{\displaystyle \frac{1}{3}}}=0.4813$$

(5)

$$A={\displaystyle \frac{K{P}_w}{P_d}}-{\displaystyle \frac{2.85P_w}{G_d}}=0.481\times {\displaystyle \frac{100}{2330}}-2.85\times {\displaystyle \frac{100}{2530}}=-0.092$$

(6)

$$B=\left(1.598H_x{C}_v\right)-\left(K{H}_x{C}_v\right)-K{W}_v=1.598\times 0.714\times \left({\displaystyle \frac{400}{3120}}\right)-0.481\times 0.714\times \left({\displaystyle \frac{400}{3120}}\right)-0.481\times 0.242=-0.0143$$

(7)

$$P_x={\displaystyle \frac{B}{A}}={\displaystyle \frac{-0.0143}{-0.092}}=0.155$$

(8)

The following calculations were used to further calculate the HL percentage required to theoretically fully react with the FA from the PI. The total binder content was set to 300 g, which was the amount of binder for one 50 mm × 50 mm × 50 mm cube. Thus, the PC contents were 120 g (300 g × 40%) and 90 g (300 g × 30%) for the HVFA60% and HVFA70% pastes, respectively. Assumptions are that complete cement hydration is at 28 days and the hydrated lime generated during hydration is 25% [25].

The weight of available free hydrated lime produced from the OPC cement is given by:

$${HL}_{free}=0.25\times P_c\times H_x$$

(9)

The necessary amount of hydrated lime to react with fly ash will be:

$${HL}_{FA}=1.85\times P_x\times P_p$$

(10)

where P_c is the weight of cement, P_p is the weight of fly ash $H_x \mathrm{i}\mathrm{s} \mathrm{t}\mathrm{h}\mathrm{e} \mathrm{h}\mathrm{y}\mathrm{d}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{t} \mathrm{c}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{b}\mathrm{y} (0.914 w/c)/(w/c+0.17).$

P_p can then be derived by letting HL _free = HL_FA. Solved by the excel solver function, it gives that P_p is 35.39 g and 21.94 g for HVFA 60% paste and HVFA 70% paste, respectively. The P_p solved here represents the weight of the FA that would be fully consumed by the HL generated from the OPC. The calculations in

Table 3. Calculation of hydrated lime percentage required to replace FA in HVFA60% and HVFA 70% pastes.

Description	Formula	HVFA60	HVFA70
Original FA content, FAOriginal	Total binder × FA%	180.00	210.00
Remaining FA not reacted; FAremained	Original FA content-Pp	144.61	188.06
FA content	${\mathrm{F}\mathrm{A}}_{remained}/(1+1.85\times 0.155)$	112.38	146.15
extra hydrated lime content, HLextra	${\mathrm{F}\mathrm{A}}_{remained}/(1+1.85\times 0.155)\times (1.85\times 0.155)$	32.23	41.91
lime percentage in binder	${\mathrm{H}\mathrm{L}}_{extra}/total binder$	11%	14%

were then followed to further calculate the HL percentage required to replace FA.

Therefore, Table 3 demonstrates that theoretically, all the FA could be consumed by reacting with Ca(OH)₂ if 11% and 14% of HL is added to replace FA in HVFA60% and HVFA 70% pastes, respectively. These two values help to select the mixes in the mix design. In the mix design for HVFA pastes with the addition of pure HL, the percentage of HL added to HVFA pastes was selected to be the content below, exactly at, and beyond 11% and 14% for HVFA 60% and HVFA 70%, respectively. For the HVFA pastes with the addition of combined CF and HL, the CF content was fixed at 3% for both HVFA60% and HVFA70% pastes, as 3% was the optimal CF dosage to achieve the highest compressive strength of HVFA mixes, as indicated in our early study [20].The dosages of HL in the HVFA pastes with combined CF and HL were then selected to vary from below the calculated threshold values (11% for HVFA 60% pastes and 14% for HVFA 70% pastes) to beyond. The mix design is summarized in

Table 4. Summary of mix design for HVFA pastes with the addition of pure HL and combined CF and HL.

Mix Notation	Cement (g)	FA (g)	CF (g)	HL (g)	W/b	Water (g)
PC100	300	0	0	0	0.25	75
FA60	120	180	0	0	0.25	75
FA60CF3	120	171	9	0	0.25	75
FA60PHL5	120	165	0	15	0.25	75
FA60PHL8	120	156	0	24	0.25	75
FA60PHL11	120	147	0	33	0.25	75
FA60PHL14	120	138	0	42	0.25	75
FA60PHL17	120	129	0	51	0.25	75
FA70	90	210	0	0	0.25	75
FA70CF3	90	201	9	0	0.25	75
FA70PHL8	90	186	0	24	0.25	75
FA70PHL11	90	177	0	33	0.25	75
FAPHL14	90	168	0	42	0.25	75
FA70PHL17	90	159	0	51	0.25	75
FA70PHL21	90	147	0	63	0.25	75
FA60HL5	120	156	9	15	0.25	75
FA60HL6	120	153	9	18	0.25	75
FA60HL8	120	147	9	24	0.25	75
FA60HL9	120	144	9	27	0.25	75
FA60HL12	120	135	9	36	0.25	75
FA60HL13	120	132	9	39	0.25	75
FA60HL15	120	126	9	45	0.25	75
FA60HL17	120	120	9	51	0.25	75
FA70HL5	90	186	9	15	0.25	75
FA70HL6	90	183	9	18	0.25	75
FA70HL8	90	177	9	24	0.25	75
FA70HL9	90	174	9	27	0.25	75
FA70HL12	90	165	9	36	0.25	75
FA70HL15	90	156	9	45	0.25	75
FA70HL17	90	150	9	51	0.25	75
FA70HL20	90	141	9	60	0.25	75
FA70HL25	90	126	9	75	0.25	75

. FAXX refers to the mix only contain FA. FAXX-PHLXX indicates that the paste contains FA and HL. FAXX-CFXX means that the paste includes FA and CF. FAXX-HLXX is a paste composed of FA, CF, and HL. The compressive strength tests, setting time tests, isothermal calorimetry and microstructural analysis were conducted, and the procedure of each test is described in Section 2.2, Section 2.3, Section 2.4, Section 2.5 and Section 2.6. The mix with the highest strength in each type of mix at the age of 28 days was selected for the setting time tests, isothermal calorimetry and microstructural analysis tests including TGA, XRD and SEM.

2.2. Compressive Strength Test

Specimens for the compressive test were cast using cube molds, each measuring 50 mm by 50 mm by 50 mm. The molded samples were wrapped in damp burlap for the initial 24 h to reduce moisture loss. After being removed from the molds, the specimens were cured in a water tank at room temperature (around 25 °C) until the time of testing.

The compressive strengths of the samples were measured at 3, 7, and 28 days using a four-column automatic testing machine (Technotest Modena, Model KE 300, Modena, Italy). The tests were performed in accordance with ASTM C109 standards [26]. The loading rate is fixed at 0.9 MPa/s. Three samples for each mix were tested at each age to include the error bars in calculations.

2.3. Setting Time Test

An Automatic Vicat Apparatus (VICAMATIC-2) was utilized to measure the setting time, following the ASTM C191 standard procedure [27]. A 1 mm Vicat needle was inserted into the paste sample every 10 min. Both the initial and final setting times of each paste were recorded. The initial setting time is the interval from mixing until the needle penetrates the paste to a depth of 25 mm. The final setting time is the duration from when water first contacts the paste until the needle no longer produces any visible penetration.

2.4. Isothermal Calorimetry

The heat of hydration of the pastes was monitored according to ASTM C1702 [28] testing method. A TAM Air 8-Channel Standard Volume Calorimeter (TA Instruments, New Castle, DE, USA) was used, and the temperature was maintained at 23 °C. Prior to the paste mixing, all the raw materials were stored at room temperature of 23 °C for 24 h. Soon after each paste mixing is finished, about 30 g of the mixture was transferred into an empty ampoule, which was immediately sealed by a cap and then carefully placed into the calorimeter. Both the rate of heat development and the cumulative hydration heat were monitored from the start of hydration to 44 h. Kinetic differences between mixes were assessed from isothermal heat-flow parameters (time to main peak, peak magnitude, and acceleration slope) at 23 °C. Activation energies were not calculated, as this requires calorimetry at multiple temperatures. It is worth noting that all calorimetry tests were conducted under the isothermal condition. No adiabatic or semi-adiabatic temperature rise measurements were performed; therefore, the results represent intrinsic hydration heat at constant temperature, without thermal acceleration from self-heating. This approach was selected to isolate chemical effects of HL and CF. Environmental coupling between heat evolution and temperature rise could be addressed in future work using semi-adiabatic calorimetry.

2.5. Thermogravimetric Analysis (TGA)

For pastes subject to 3, 7, and 28-day curing, thermogravimetric analysis (TGA) was conducted using a PYRIS TGA machine (PerkinElmer, Waltham, MA, USA). The dynamic heating ramp was varied between 30 °C and 1000 °C under a N₂ atmosphere with a heating rate of 10 °C/min. The pastes were first crushed and milled into particles that were fine enough to pass through the sieve of 63 μm. On the testing day, the reaction was stopped before starting the TGA. For each paste, approximately 5–12 mg of crushed powder was placed on a filter paper above a vacuum pump; then, the powders were entirely immersed in acetone and dried using a vacuum pump. The reaction for each paste was stopped five times. The plotted TG curve reflects the mass of the sample as a percentage of that of the original sample during heating. The samples lose their mass due to decomposition during heating, and different compounds decompose in different temperature ranges. In this study, the decomposition of C-S-H, known as dehydration (Ldh), was recorded between 105 and 400 °C. It is worth noting that Calcium Aluminate trisubstituted (AFt) and Calcium Aluminate monosubstituted (AFm) also decomposed during this temperature range. The decomposition of Ca(OH)₂, referred to as dihydroxylation (Ldx), was recorded within a temperature range of 400–600 °C. The decomposition of CaCO₃, referred to as decarbonation (Ldc), occurs between 600 and 1000 °C. The first derivative of the TG curve (DTG) for each sample is also given and provides information about minor changes that can influence the actual temperature ranges to be considered. The calcium hydroxide (CH) content and chemically bound water (W_B) of each paste were calculated using Equations (11) and (12), respectively [29].

$$CH=4.11Ldx+1.68Ldc$$

(11)

$$W_{B }=Ldh+Ldx+0.41Ldc$$

(12)

Following the quantitative DTA protocol established by Pyzalski et al. [30], the heat-transfer characteristics of the PYRIS TGA were verified with calcium hydroxide decomposition (ΔH = 1.02 kJ g⁻¹ at 450–550 °C) as an internal chemical standard [31]. This calibration confirmed temperature accuracy within ±2 °C and enthalpy accuracy within ±3%, ensuring that the mass-loss assignments (C-S-H 105–400 °C, CH 400–600 °C and CaCO₃ 600–1000 °C) remain internally consistent for all HVFA pastes investigated.

2.6. X-Ray Diffraction (XRD)

XRD was used to qualitatively analyze the mineralogical composition of the different pastes. The device used was A Bruker D8 Advance X-ray diffractometer with a Ni-filtered CuKα radiation source (Bruker, Karlsruhe, Germany) (wavelength = 1.54 Å). The diffraction data were collected between 5–85° 2θ, and the step size and scan rate were 0.02° and 1.0 s per step, respectively. An anti-scatter blade was added to minimize the intensity of the diffracted background at low angles. The incident beam had a divergence of 1.0°, and a 2.5° Soller slit was employed in the diffracted beam. The spinning rate was set at 15 rpm. Pastes with a curing period of 28 days were selected for the XRD test, and powders for this test were prepared in the same way as described for TGA.

2.7. Scanning Electron Microscopy (SEM)

To investigate the microstructure of the pastes at the age of 28 days, a scanning electron microscopy (SEM) FlexSEM 1000 name/brand machine equipped with a backscatter electron detector (Hitachi, Tokyo, Japan) was used. The operating voltage was 15 kV. Elemental analysis of the pastes was performed using an energy-dispersive X-ray spectroscope (EDS) attached to the SEM apparatus. For sample preparation, a small piece of paste was placed into a cylindrical mold (diameter of 25 mm and height of 10 mm) before the mold was filled with resin, making the paste embedded in the solidified resin. The sample was then polished with the help of a grinding paper, which provided a smooth surface of the sample, exposed for SEM examination.

3. Results and Discussion

3.1. Compressive Strength Development

The compressive strength results for OPC, pure HVFA pastes, HVFA pastes with HL, CF, and combined HL and CF are shown in Figure 2, Figure 3, Figure 4 and Figure 5. Comparing OPC, pure FA60, and pure FA70 pastes, the compressive strength decreased with an increase in FA dosage because of the postponed pozzolanic reaction between Ca(OH)₂ and FA. As the time progressed from 3 to 28 days, increasing rates of compressive strength for pure FA60 and FA70 pastes were observed. For instance, the increasing rates of the compressive strength of FA60 pastes are 34.2% from 3 to 7 days and 72.3% from 7 to 28 days. This is because of the pozzolanic time-dependent reaction when sufficient Ca(OH)₂ accumulated from OPC hydration.

Figure 2 and Figure 3 demonstrate that the FA60 and FA70 pastes reached their highest compressive strengths at HL dosages of 11% and 14%, respectively, at the age of 28 days. This matches the calculations from the PI index of the FA in Section 2.1. Figure 4 and Figure 5 compare the effects of the addition of different admixtures to HVFA60 and HVFA70 pastes. It can be seen that the compressive strengths of both HVFA60% and HVFA70% pastes were improved at all the ages of 3, 7 and 28 days by the adding of HL or CF separately. At the very early ages of 3 and 7 days, it seems that the adding of CF was more effective in improving the compressive strengths of HVFA60% and HVFA70% pastes compared to the adding of HL. In contrast, at the age of 28 days, adding optimal dosages of HL seemed to improve the compressive strength of both HVFA60% and HVFA70% pastes more compared to adding optimal dosages of CF. However, with the addition of both HL and CF to HVFA pastes, the compressive strength either remained unchanged or slightly declined compared to the corresponding reference mixes.

With the addition of both HL and CF, it was observed that the HVFA mixes became quite hot almost immediately after mixing, indicating that heat was rapidly released after mixing. The heat of hydration released too quickly is not beneficial for the microstructural development of pastes [10]. This could be the reason why the HVFA pastes with both HL and CF do not have good strengths. Although rapid heat evolution is usually linked to higher strength in ordinary concrete, the present HVFA mixes lost 15–22% of their 28-day strength. The mechanisms underlying this apparent contradiction, revealed by isothermal calorimetry and microstructural analyses (presented in Section 3.3, Section 3.4, Section 3.5 and Section 3.6), are discussed subsequently.

3.2. Setting Time

The test results for the initial and final setting times of the pastes are listed in

Table 5. Setting times of cement pastes with different mixtures.

Mix ID	Setting Time (h:min)
	Initial	Final
FA60CF3	0:43	2:04
FA70CF3	1:19	3:17
FA60HL5	0:20	0:55
FA70HL5	0:25	1:07
FA60HL15	0:17	0:32
FA70HL15	0:24	0:59
FA60PHL11	2:9	3:46
FA70PHL14	2:1	3:43
FA60	7:10	10:01
FA70	9:14	11:39
PC100	2:25	6:31

. As expected, compared to OPC paste, FA60 and FA70 pastes had longer initial and final setting times. FA70 paste has a longer initial and final setting time than FA60 paste. It is commonly known that FA replacement of OPC delays the hardening process.

Adding HL to the FA60 and FA70 pastes shortened both their initial and final setting times compared to the pure FA60 and FA70 pastes. However, FA70-PHL5 could have a longer setting time than FA60-PHL5 due to the higher amount of FA in the paste.

Incorporating CF into the FA60 and FA70 pastes also decreased their initial and final setting times relative to the pure FA60 and FA70 pastes. Additionally, it was observed that the pastes containing CF had a shorter setting time than the pastes with HL. Thus, the addition of CF has a stronger effect on reducing the setting time of the pastes. Previous research indicates that the addition of a single HL or CF can reduce the setting time of FA pastes [32,33]. These findings are consistent with those of the experiments. This is mainly due to the fact that both the HL and CF could accelerate the hydration of PC. This is mainly associated with the hydration of C₃A and C₃S [32,33]. This is reflected in the heat of the hydration curves. This hydration process is further investigated in detail through isothermal calorimetry tests in the next section.

It was also noted that incorporating both CF and HL together shortened the initial and final setting times more noticeably than pastes containing only CF or only HL. For example, the FA60HL15 paste had a final setting time of just 32 min, meaning it fully hardened in about half an hour. However, pastes that set too rapidly are difficult to compact properly. This may explain why HVFA mixtures with both CF and HL added exhibit lower compressive strengths compared to HVFA pastes containing only CF or only HL.

3.3. Heat of Hydration Development

The effects of HL, CF, and combined HL and CF on the hydration reaction of the FA pastes during the first 44 h are illustrated in Figure 6 and Figure 7. The first peak in the heat flow curves is mostly related to the hydration of C₃A and C₄AF, as the aluminate phases hydrate quickly after contact with water. The second peak corresponded to the hydration of C₃S to form C-S-H as the major compound to take the strength. The third peak was caused by the renewed hydration of the aluminate phases when it reacted with gypsum. The exact times for the occurrence of each major peak for each mix are listed in

Table 6. Time for occurrence of each peak in the heat flow curve of each mix.

Mixes	The Time for the Occurrence of Peaks (hour)
	1st Peak	2nd Peak	3rd Peak
FA60	0.03	9.98	15.38
FA60PHL11	0.04	8.13	11.8
FA60CF3	0.15	5.90	-
FA60HL5	0.14	18.00	-
FA70	0.05	9.30	14.96
FA70PHL14	0.05	7.80	10.74
FA70CF3	0.07	4.74	-
FA70HL5	0.45	14.73	-

for comparison.

For the HVFA60% mixes with the addition of different admixtures, as shown in Figure 6b, the addition of 3% pure CF slightly increased the height of the first peak compared to the pure FA60 mix, while the addition of pure HL did not significantly affect the intensity of the first peak. The combined HL + CF addition slightly lowered the first-peak intensity and delayed its onset from 0.03 h to 0.14 h. compared to the pure FA60 mixes. For the HVFA70% mixes in Figure 7b, the addition of 14%HL to FA70 mixes (FA70PHL14) or the addition of 3%CF to FA70 mixes (FA70CF3) increased the intensity of the first peak. In the FA70HL5 mix (5% HL + 3% CF), the first peak was delayed from 0.05 h to 0.45 h and its intensity was reduced compared with the pure FA70 mix. Thus, for both FA60 and FA70 mixes, the addition of single HL or CF generally promoted the hydration rate of aluminate phases. The improved hydration rate of aluminate phases by the adding of CF has also been reported in a previous study [34]. Nevertheless, the combined addition of CF and HL delayed and reduced the hydration of aluminate phases in both HVFA60% and HVFA70% pastes.

The second peaks of the heat of hydration curves for both the FA60 and FA70 mixes show a similar trend with the addition of admixtures. The FA60 and FA70 pastes with 11% and 14% pure HL accelerated the occurrence of the second peak compared to the pure FA60 and FA70 pastes, respectively (from 9.98 h to 8.13 h in HVFA60% mixes and from 9.30 h to 7.80 h in HVFA70% mixes). The addition of single CF or HL accelerated the occurrence of the second peaks as well as increasing the intensities. The adding of CF has a larger extent of accelerating the occurrence of the second peak compared to the adding of HL (the second peak advanced to 4.74 for HVFA60% mix and 5.90 h for HVFA70% mix). However, the addition of both HL and CF significantly decreased the intensity of the second peak compared to that of the FA60 and FA70 pastes. These phenomena means that the adding of single HL or CF could accelerate the hydration of C₃S with CF having a larger effect, while the adding of combined HL and CF decreased the hydration of C₃S significantly.

The third peaks of the heat of hydration curves for the HVFA pastes with the addition of optimal dosages of pure HL (11% for HVFA60 paste and 14% for HVFA70 paste) are remarkably brought forward and heightened compared to FA60 and FA70 pastes (from 15.38 h to 11.80 h in HVFA 60% mixes and from 14.96 h to 10.74 h in HVFA 70% mixes). However, the third peaks of the HVFA pastes with the addition of CF, combined CF and HL were inconspicuous. Thus, the addition of pure HL accelerated the reaction between the aluminate phases and gypsum, whereas the addition of the single CF or combined CF and HL weakened the renewed reaction of the aluminate phases.

The lower first peak with adding of combined HL and CF is attributed to elevated Ca²⁺/OH⁻ levels, which alter sulfate–aluminate equilibrium and slow the initial C₃A reaction. When CF and HL are added together, their rapid, competing consumption of Ca²⁺ and SO₄²⁺ ions suppresses C₃S nucleation and growth, leading to a markedly lower and delayed second peak. The reduced third peak with combined HL and CF reflects rapid early sulfate consumption, limiting later AFm formation and thus diminishing late aluminate hydration. Although not measured directly, adsorption of calcium-formate complexes onto C-S-H surfaces and concomitant Al³⁺/SO₄²⁻ complexation are consistent with the delayed C₃S peak and the suppressed AFt-renewal peak observed in HL + CF mixes. These reactions would lower the effective supersaturation needed for continued hydration and provide an additional pathway for the observed antagonism

To sum up the trend of heat flow by the adding of different admixtures, the adding of pure HL slightly improved the hydration of aluminate phases at the beginning of mixing, accelerated the occurrence of C₃S and intensified the renewed hydration reaction of aluminate phases. The addition of pure CF enhanced the hydration of the aluminate phases at the beginning, promoted the hydration of C₃S, and weakened the renewed hydration of the aluminate phases. The addition of combined CF and HL significantly reduced the hydration of C₃A, C₃S as well as the renewed hydration of C₃A.

3.4. Thermal Gravimetric Analysis (TGA)

The TG and DTG patterns of the samples hydrated for 28 days are presented in Figure 8 and Figure 9. The TG curves describe the decrease in the weight of the samples with increasing temperature. The mass of the sample at specific temperature points and the corresponding calculated Ldh, Ldx and Ldc values are listed in

Table 7. Ldh, Ldx, and Ldc of the pastes.

Mix ID	m_105	m_400	m_600	m_1000	Ldh	Ldx	Ldc
FA60	98.4	89.9	88.5	83.5	8.5	1.4	5.0
FA60CF3	97.8	87.1	85.1	78.8	10.7	2.0	6.3
FA60PHL11	97.3	87.6	84.0	79.0	9.7	3.6	5.0
FA60HL5	95.9	94.3	92.2	90.9	1.7	2.1	1.3
FA70	98.3	90.3	89.2	84.5	8.0	1.1	4.7
FA70CF3	98.0	87.3	85.5	79.2	10.7	1.8	6.3
FA70PHL14	97.0	88.0	84.5	78.0	9.0	3.5	6.5
FA70HL5	93.6	91.7	89.8	87.8	1.9	1.9	1.9

.

Comparing the Ldh values of pastes with HL and CF and the combination of HL and CF, the pastes with the combination of HL and CF had smaller Ldh values compared to the pastes with pure HL and CF. As the Ldh value represents chemically bonded water from CSH, the pastes with a combination of HL and CF had a lower content of CSH than the pastes containing HL and CF. As CSH is the major hydration product responsible for the concrete strength, this finding agrees well with the compressive strength results that the pastes with both CF and HL have lower strength compared to the pastes with HL or CF.

The Ldc value for the paste with both CF and HL was the lowest compared to that of the pastes with a single CF or HL. Ldc represents the CaCO₃ content of the paste. As the TGA testing was conducted in a N₂ atmosphere, CaCO₃ could only be obtained from the carbonation of Ca(OH)₂ during hydration. Thus, the pastes with both CF and HL had the lowest degree of carbonation compared to the pastes with HL or CF separately.

The Ca(OH)₂ contents of the different mixes are illustrated in Figure 10. The pastes with a single CF increased the Ca(OH)₂ content compared to the pure HVFA mixes. The pastes with single HL or single CF increased Ca(OH)₂ slightly more than the pure HVFA mixes. It is possible that CF stimulated the hydration of OPC to form more Ca(OH)₂. HL can also react with FA to produce more hydration products. The addition of both CF and HL decreased the Ca(OH)₂ content compared with the other mixes.

The chemically bound water (W_b) of the different mixes is shown in Figure 11. The W_b of pastes with a single CF or HL was higher than that of pure HVFA mixes. The pastes with HL had a higher W_b compared to the pastes with CF. The pastes with both HL and CF considerably decreased the W_b compared to the other mixes. The W_b values of all mixes followed the same trend as the compressive strength development of these mixes.

3.5. XRD Patterns

The XRD patterns of the HVFA mixtures incorporating CF, HL, and a combination of CF and HL are shown in Figure 12 and Figure 13 for comparison purposes. The peak intensities in the XRD patterns of the HVFA60% and HVFA70% mixtures displayed a similar tendency. Quartz, as the main component of FA, was detected in all HVFA mixtures.

When comparing the peaks related to CH content in different mixtures, the addition of optimal dosages of CF or HL increased the CH content in both the HVFA60% and HVFA70% mixtures. This is in line with the conclusion that adding of HL can enhance the pozzolanic reaction of FA cement paste [35] and CF can accelerate the OPC hydration [20] as well as pozzolanic reaction of FA [36]. Notably, mixtures with HL had a higher CH content than those with CF at the age of 28 days. Conversely, the addition of a combination of HL and CF led to a reduction in CH content compared to the pure HVFA mixtures. Both TGA and XRD indicated consistent relative trends in CH content across mixes, although absolute values differed. Such discrepancies are common and arise from the fundamentally different measurement bases: TGA estimates CH from the mass loss on dehydroxylation, which can overlap with other bound water losses, whereas XRD relies on peak intensities from crystalline CH and does not account for amorphous or poorly crystalline material. Given these methodological differences, we focus our interpretation on the agreement in relative changes in CH between mixes, which is consistent between both techniques.

Regarding the ettringite (AFt) content in the HVFA mixtures, the HVFA60% and HVFA70% mixs showed the same trend. the HVFA mixtures with optimal CF had a higher ettringite content than the pure HVFA pastes. It was also found in previous studies that CF could increase the AFt content in cement [33,36]. In contrast, the HVFA pastes with optimal HL had less ettringite than the pure HVFA pastes. Moreover, the mixture with combined HL and CF exhibited near-complete AFt absence, demonstrating weakened aluminate rehydration. This is also reflected in the isothermal calorimetry that the 3rd peaks of heat flow curves of HVFA mixes with both HL and CF were missing.

Concerning the CaCO₃ content of the mixtures, the same trend could also be observed in HVFA 60% and HVFA 70% pastes. The mixture with only HL had the highest CaCO₃ content. The pure HVFA pastes had the second-highest amount of CaCO₃. The mixture with CF had less CaCO₃, and the mixture with combined HL and CF had the least CaCO₃. This was consistent with the TGA results.

In conclusion, the findings from the XRD patterns were in agreement with the TGA and isothermal calorimetry results. Generally, mixtures with combined HL and CF had fewer hydration products compared to mixtures with the separate addition of HL and CF. Thus, the reduced formation of hydration products in mixtures with combined HL and CF could account for their lower compressive strengths, which had been previously confirmed. The antagonistic HL + CF effect could be attributed to calcium ion competition and aluminate phase destabilization. Specifically, calcium ion competition refers to CF preferentially accelerating OPC hydration, sequestering soluble Ca²⁺ ions and thereby starving the pozzolanic reaction between HL and fly ash (FA). Aluminate phase destabilization is that the rapid initial ettringite (AFt) formation, induced by CF, depletes sulfate ions (SO₄²⁻), compromising the stability of secondary ettringite and suppressing renewed aluminate hydration.

3.6. Morphological Analysis

Scanning electron microscopy (SEM) micrographs of the pastes are shown in Figure 14. The round particles were fly ash, and the black dots represent the pores in the paste. In the control samples HVFA60% and HVFA70% pastes, it can be seen that there are abundant unreacted FA particles (spherical morphology) and microcracks. Discontinuous C-S-H matrix with high porosity could be noticed. Compared to the pure HVFA60 and HVFA70 pastes, the pastes with single CF or HL had fewer pores, fewer FA particles and denser C-S-H morphology. As indicated by the previous compressive strength results, the compressive strength of the HVFA pastes with single CF or HL was higher than that of the pure HVFA pastes. It is confirmed in SEM images that the addition of CF or HL stimulates the hydration of FA and reduces the number of pores.

With the addition of both CF and HL in HVFA pastes, there were fewer FA particles in the samples than in the HVFA pastes with single CF or HL, indicating more hydration of FA. However, more pores and cracks were observed in the HVFA pastes with both CF and HL. Isothermal calorimetry results showed that the HVFA pastes with both CF and HL released a large amount of heat within a short time after mixing. A large amount of heat released quickly after mixing is not suitable for the microstructural development, inhibiting pore-filling C-S-H growth [10]. The quick release of heat was also associated with flash setting, as revealed in the setting time test. Thus, paste setting within a short time can be difficult to compact. All these factors could lead to the formation of pores and cracks in pastes with both HL and CF, weakening the compressive strength.

These microstructural features are also consistent with the macroscopic failure behaviour observed under compression. All specimens failed in a brittle manner, characterised by inclined shear cracks propagating at approximately 45° to the loading axis, which ultimately led to sudden loss of load-bearing capacity. Mixes with optimal HL additions (11% for HVFA60 and 14% for HVFA70) exhibited comparatively finer, more distributed crack patterns, consistent with their denser hydration products. By contrast, mixes containing combined HL and CF developed fewer but wider shear cracks, in line with the more porous microstructure identified by SEM.

4. Conclusions

This study examined the individual and combined effects of hydrated lime (HL) and calcium formate (CF) on the early-age strength of HVFA pastes containing 60% and 70% fly ash. Major findings are summarized below.

The addition of CF or HL alone could enhance the compressive strength of the HVFA paste. For HVFA mixes with CF, when 3% CF is added, the HVFA60% and HVFA70% mixes reached the highest strength, respectively [20]. For HVFA mixes with HL, the compressive strength testing results confirmed the calculation results that the addition of 11% HL and 14%HL into the HVFA60% and HVFA70% mixes reached the highest strengths. While the PI-guided HL levels (11% for HVFA60 and 14% for HVFA70) coincide with the strength maxima observed in this study, additional data across a wider range of HL contents, curing ages and fly ash sources are required to validate the PI as a general predictor. Such validation is planned as part of our ongoing research.

When HL and CF were added together, strength dropped by 15–22% even though setting was further accelerated. Isothermal calorimetry, TGA, XRD and SEM revealed that the co-addition suppressed hydration of C₃A and C₃S, reduced C-S-H and portlandite contents, and increased porosity and cracks.

The antagonistic interaction between calcium formate (CF) and hydrated lime (HL) in HVFA systems mainly arises from the three main mechanisms, which are calcium ion competition, aluminate phase destabilization and flash setting introducing pore and crack formation.

In the practical application of HVFA concrete, it is recommended to use hydrated lime (HL) or calcium formate (CF) individually to enhance early-age strength. Specifically, if priority is given to ensuring the 28-day strength, HL can be selected with dosages of 11% and 14% for HVFA 60% and HVFA 70% pastes, respectively. For scenarios requiring rapid development of early-age strength (such as emergency repair projects), 3% CF for HVFA60% or HVFA 70% pastes is recommended. The combined use of HL and CF for HVFA pastes is not suggested, as their antagonistic effects will lead to a reduction in strength and construction difficulties (e.g., excessively rapid setting which makes pouring unmanageable). While drying and autogenous shrinkage were not measured in this work, it is noted that the rapid setting observed in certain HL + CF combinations could potentially accelerate early-age shrinkage. Rapid stiffening limits the period during which internal stress relaxation can occur and may increase capillary tension, leading to higher autogenous shrinkage. Similarly, earlier strength gain could lead to faster restraint of volumetric changes, affecting drying shrinkage. These effects warrant dedicated study in future work. The present study focused on early-age strength; subsequent work should quantify carbonation depth, chloride migration and freeze–thaw resistance of HVFA mixes cured under variable humidity (50–90% RH) and temperature (10–40 °C) to confirm suitability for precast elements exposed to outdoor storage or aggressive environments. As deformability and post-peak ductility are critical for field performance, future work could therefore quantify stress–strain curves and fracture energy under displacement-controlled loading to explore how the microstructural variations identified here affect the overall mechanical response of HVFA systems.