Calcium aluminosilicate glass (4.5SiO₂-3Al₂O₃-1.5P₂O₅-CaO-2CaF₂) was prepared in the laboratory by an established melt quenching route [26] . Poly-acrylic acid solution (PAA) (C₃H₄O₂) was provided by BASF plc, Cheshire, UK. The analytical grade Hydroxyapatite (HA-Ca₁₀(PO₄)₆(OH)₂) of particle sizes 15 μm and 30 μm, was purchased from the chemical company Teknimed, France. Three types of glass ionomer cements GIC, 15HAGIC and 30HAGIC were prepared using the ratio of powder and liquid as Glass:PAA = 2:1 and Glass:HA:PAA = 2:0.5:1. The powder and the polymer liquid were mixed by a UHMWPE spatula on a PMMA mixing plate within 60 seconds and then cast into Teflon molds onto a steel plate. The steel mold was clamped using a G clamp and the GICs pastes were allowed to set for 60 minutes in an air oven at 37˚C. The molds were then unclamped and all samples were placed into deionized water in glass beakers, which then were stored in a water bath (Thermo Fisher Scientific Inc., Waltham, United States) at 37˚C. The samples were then kept for different aging periods and tested.

Three different types of cement namely LG26, 15HA, and 30HA were mixed homogeneously with KBr and were ground by a mortar and pestle to further reduce the particle size because the scattering of the infrared beam by larger particle size creates the sloping baseline of the spectrum. Evaporating the KBr, a standard pellet of all the samples was prepared for FTIR. The Nicolet FTIR spectrometer (Magna-IR860) coupled with a deuterated triglycine sulphate (DTGS) detector, was used for FTIR spectroscopy studies. The spectra were collected by transmission infrared source with evolved packet system mode with a speed of 0.6329 ms⁻¹. The frequency ranges 400 - 4000 cm⁻¹ and a resolution of 0.4821 cm⁻¹ was set up with 100 scans over 15 minutes. All the spectrum was filtered i.e., the background noise was eliminated by subtraction.

X-ray powder diffraction study was done for LG26, GIC, 15HAGIC, and 30HAGIC by Equinox XRD machine (model number 3000) with Cu-Kα radiation. The LG26 powder was taken after sieving to investigate whether there was any crystallization occurring during quenching. The 2-theta range was 10 - 90 degrees with a resolution of 0.02983 degrees, with 35 kV - 25 mA throughout the experiments. The total scan time for the whole 2-theta range was 7200 seconds. The JCPDS database was used as a reference to determine the phases. The peak position was identified by Match software.

The particle size and distribution of raw powder of hydroxyapatite and glass and the fracture surface after flexural test of different GICs (GIC, 15HAGIC, 30HAGIC) were studied by electron microscope. The SEM machine used here was a high-resolution JEOL 6060 (Philips Co. Japan) operated at 10 keV operating voltage. The cylindrical-shaped samples were used to study the fracture surface of GICs.

Several mechanical properties like compressive strength (CS), diametral tensile strength (DTS), and flexural strength (FS) of different glass ionomer cement were investigated by Instron mechanical testing machine (Model-1195) and compared with British Standard (BS) according to the requirement of American Dental Association (ADA) specification [27] [28] . Figure 1 shows the Teflon split mold, prepared cylindrical samples of (a) dimension 4 ± 0.1 mm in diameter and 6 ± 0.1 mm thickness for CS (b) dimension 4 ± 0.1 mm of diameter and 2 ± 0.1 mm in thickness for DTS and (c) dimension 25 ± 0.1 mm in length and 2 ± 0.1 mm in width and FS measurements. CS was assessed by 5 kN load with a cross-head speed of 1 mm/min using Equation (1), DTS was measured according to diametral compression of 1kN load, and a crosshead speed of 1mm/min, using Equation (2) [27] . FS was studied by applying 500 N load with a cross-head speed of 1 mm/min, using formula (3) [27] [28] .

$CS=\frac{4\times p}{\pi \times d^2}$ (1)

$DTS=\frac{2\times p}{\pi \times d\times T}$ (2)

$FS=\frac{3\times F\times l}{2\times b\times h^2}$ (3)

Where p is the maximum applied load at the specimen, d is the diameter of the cylindrical sample. For FS, F is the maximum load at fracture of the sample, l is the distance between two supports, b is the width of the specimen, and the thickness of the specimen is h.

The FTIR study of every single element and different GICs (GIC, 15HAGIC, 30HAGIC) like PAA, LG 26, 15HA, 30HA, 15HAPAA, and 30HAPAA was done and the major peaks in FTIR spectra are shown in Figure 2 .

In the FTIR spectrum for 40% PAA solution, as shown in Figure 2(a) , the main absorption peak at 1697 cm⁻¹ was due to carbonyl (C = O) stretching vibrations [29] - [31] . The inter- and intra-layer stretching vibrations of H-bonded O-H are represented by the absorption band at around 3359 cm⁻¹ [32] . The monomer and PAA’s C-H scissor vibrations peak appeared at 1447 cm⁻¹ and the stretching vibration of C-O was observed at 1257 cm^{−1 [29]} . The FTIR spectroscopy analysis of ion-leachable glass (LG 26) powder is shown in Figure 2(b) . The peak at 450 cm⁻¹ is associated with stretching vibrations of Si-O-Si [33] . P-O bending and Si-O-Al stretching vibrations are associated with the peak at 552 cm⁻¹ [34] . The stretching vibrations of Al-O bond is represented by the peak at 713 cm⁻¹ [35] . The peaks at 966 cm⁻¹ and 1070 cm⁻¹ are associated with Si-O stretching and P-O bonds [33] [36] . The FTIR spectrum of hydroxyapatite of particle sizes 15 μm and 30 μm is similar (not included here) but compared in Figure 2(e) and (f) blue curve. The peak at 3571 cm⁻¹ is due to absorbed water and hydroxyl vibrations. Vibrations of PO₄ are associated with the peak at 1041 cm⁻¹. The peaks at 400 and 472 were phosphate ν₂ vibrations. The peak at 1400 cm⁻¹ is due to carbonate contamination [37] . Figure 2(c) shows the FTIR spectrum of 15HAPAA (15HA:PAA = 0.5:1.0) liquid (yellow) is compared with PAA (blue) where the two main peaks at around 1600 and 1700 cm⁻¹ are still similar in the blended liquid. The only noticeable peak in 15HAPAA blend liquid different from PAA peaks is at 1550 cm⁻¹ which is due to calcium carboxylate with calcium in a bridging state [38] . Like Figure 2(c) , Figure 2(d) show a similar result for 30HAPAA liquid (15HA:PAA = 0.5:1.0). Figure 2 (e) shows the FTIR spectra of 15HA powder and 15HAGIC cement after 4 weeks of setting. Figure 2(f) shows the FTIR spectra of 30HA powder and 30HAGIC cement after 4 weeks of setting. In 15HAGIC and 30HAGIC the peaks are shifted slightly to the right. From 800 cm⁻¹ to 1300 cm⁻¹ there is a broad peak with a major peak at 1045 cm⁻¹ in 15HAGIC and at 1050 cm⁻¹ in 30HAGIC because of Si-O stretching vibrations. The peaks at 570 cm⁻¹ in 15HAGIC and 566 cm⁻¹ in 15HA are congruent peaks ascribed to phosphate ν₄ vibrations. Similarly, the peak at 568 cm⁻¹ in 30HAGIC and 563 cm⁻¹ in 30HA are due to phosphate ν4 vibrations [37] . There might be some unreacted PAA in the cement and thus peaks around 1460 cm⁻¹ in both 15HA and 30HA are due to C-H scissoring of PAA and monomers [29] . The peaks at 1553 ( Figure 2(e) ) and 1570 ( Figure 2(f) ) are due to calcium carboxylate with calcium in bridging mode [38] .

CS as shown in the bar diagram, Figure 3(a) , increased with time and the maximum value of average compressive strength was found to be about 86 MPa for 15HAGIC after 4 weeks from mixing. Finally, for 30HAGIC, the CS increased up to two weeks after mixing but decreased after that. It is worth noticing, that the cements containing HA with larger particle sizes showed higher CS in the first two weeks but lower CS after that, compared with the cements containing smaller particle sizes in HA. In this test, the alpha value was taken 0.05, where p > 0.05 has no significant difference between the two groups of cements but when p < 0.05 has a significant difference between the two groups of cements. From this t-test result, it is seen that the two sets of comparison groups GICs to 30HAGIC and 15HAGIC to 30HAGIC have significant differences as both groups had p-values of less than 0.05. The CS of GIC and 15HAGIC are not significantly different as the p-value was greater than 0.05.

DTS was assessed, as shown in Figure 3(b) , by an Instron machine by applying compressive forces along the diameter adopting the British Standard Institution rule [27] . From the Figure 3(b) , it is clear that the DTS of conventional GIC reached the highest value of about 15 MPa in 168 hours of setting. After 168 hours of aging, the DTS was decreased. The DTS of 15HAGIC continued to increase with time and for 30HAGIC, DTS was increased up to 3 weeks of aging, after that it was decreased. The results showed that the DTS was notably lower than the CS as the GICs are brittle materials, and crack propagates easily due to tensile loading than compression loading. In this test, shear stress arose between the specimen and the flat metal disk of the testing machine. As the shear failure was followed by tensile failure, the cements broke with a significantly small amount of load. The confidence level of the t-test was 95%. The t-test result of diametral tensile strength shows that there was no significant difference among different groups of samples as the p values were all clearly greater than 0.05.

FS of all cements increased with aging time except the conventional GIC. The FS of GICs dropped in the 3rd week and increased again in the 4th week after mixing. 15HAGIC showed an increasing pattern of FS trend for all aging times. During the aging period of four weeks, the flexural strength for 15GIC was the highest except in 168 hours. 30HAGIC also showed an increasing trend throughout aging time. Therefore, the t-test shows that the flexural strength of the two cements had no significant difference (since p > 0.05). It is noticeable that the general trend is increasing flexural strength for all individual cements.

The EDX spectra and SEM image(inset) of the HA (grain size 15 - 30 µm), and GICs are shown in Figure 4(a) . The EDX shows that the major ingredients of hydroxyapatite are calcium and phosphorus. The EDX spectra of pure hydroxyapatite confirmed the calcium-to-phosphorus ratio as 1.67. In Figure 4(b) , the SEM image(inset) of a cylindrical GIC sample shows that the outer surface has a lot of cracks due to dehydration. In the case of SEM imaging, vacuum water evaporates, and cracks are formed, therefore SEM is not a standard technique for microstructural characterization of GIC. The microscopic anatomy of GICs revealed that the fractured surface of cement specimens consists of both large and small glass particles which can readily be distinguished from the polymer matrix. Figure 4 (c) shows the SEM image (inset) and EDX spectrum of LG 26 with their composition. Figure 4 (d) EDX spectrum and the SEM of pure GICs. Similarly, Figure 4(e) shows the SEM image and EDX spectrum of 30HAGIC. Figure 5 shows the fracture surface of flexural samples of pure GIC, 15HAGIC, and 30HAGIC. From, these figures it is clear that the glass particles were pulled out of the fracture surface under the applied bending force.

Figure 6 depicts the X-ray diffraction pattern of 15 μm hydroxyapatite, 30 μm hydroxyapatite, pure GIC, 15HAGIC, and 30HAGIC respectively. This pattern indicates that the hydroxyapatite has a crystalline single phase. Its structure is hexagonal. The sharp peaks are between 30 and 40 degrees, in the 2-theta range and these peaks correspond to JCPDS 9-0432 file. These peaks are also the same as observed in enamel powder. From these spectra, it is evident that the pure set cement was amorphous as there is no prominent peak. After adding hydroxyapatite filler, some peaks are observed in 15HAGIC and 30HAGIC cements associated with the presence of hydroxyapatite. The peaks in HAGIC are shifted to the right compared to the pure HA. The peak in the 2-theta range of 20 - 35 degrees is ascribed to crystalline apatite structure.