Introduction
Powders of α-tricalcium phosphate (α-TCP; α-Ca₃(PO₄)₂) and β-tricalcium phosphate (β-TCP; β-Ca₃(PO₄)₂) are widely used as raw materials for hydraulic calcium phosphate cements (CPCs). These powders are typically obtained by grinding α-TCP and β-TCP solids produced by reactive sintering of a calcium-rich phase with a phosphate-rich phase, for example calcium carbonate (CC) with anhydrous dicalcium phosphate (DCP). Several recent studies have shown the importance of small variations of the CC/DCP ratio on the physico-chemical properties of β-TCP solids and powders. In this study, it was hypothesized that similar effects might occur with α-TCP. To test this assumption, several dozens of α-TCP powders were produced according to the same general procedure, but different CC/DCP ratios and different sintering temperatures. The powders were then thoroughly characterized. Some powders were also used to produce cements to determine the relationship between powder synthesis conditions and cement properties.
 
Materials and methods
The powders were produced by reactive sintering of CC and DCP, and subsequent grinding in a planetary mill. In a first series, 7 α-TCP powders with a nominal Ca/P ratio of 1.485, 1.490, 1.495, 1.500, 1.5025, 1.5050, and 1.5075 were produced. In a second series, the sintering temperature of powders produced with nominal ratios of 1.490, 1.495, 1.5025, 1.5050 was reduced from 1350°C to 1250°C, 1200°C, 1175°C and 1150°C. The powders were characterized by SEM, XRD, XRF, ICP-MS, laser diffraction, nitrogen adsorption (BET). Once mixed with water, their liquid limit and injectability were also determined. Cement pastes made with Na2HPO4 0.2M solution were produced and characterized for their setting time (indentation test) and reactivity (isothermal calorimetry). Finally, some of the set cements were characterized by diametral tensile testing and SEM (after impregnation in a resin and polishing).
 
Results and discussion
A >0.99 correlation was found between nominal and measured Ca/P ratios of the α-TCP powders, as well as between their phase purity measured by XRD, XRF and ICP-MS results. Up to 1.8% β-calcium pyrophosphate (β-CPP), 8% hydroxyapatite (HA), and 10% amorphous phase were measured depending on the nominal Ca/P molar ratio and the sintering temperature. The powders produced at 1350°C with a low Ca/P molar ratio contained small β-CPP impurities and presented numerous > 50 µm dense particles. Since these particles could not be found in similar powders produced at a lower sintering temperature, their presence was associated with the formation of a liquid CPP phase above 1280°C. These large dense particles reacted very slowly and were still found in cement samples after several days of reaction at 37°C. The setting time of the cement pastes dropped threefold when the nominal Ca/P molar ratio of the α-TCP powder increased from 1.485 to 1.5075. Powders containing HA impurities (Ca/P > 1.50) were in general much finer but contained a small fraction of > 100 µm large particles, resulting in dense agglomerates in the set cements.
 
Conclusion
This study demonstrates the importance of a precise control of the purity of α-TCP powders for CPC applications. The setting time may vary threefold with minor changes of composition.