Gold is an interesting and unique catalyst that is relatively inert in the bulk state, but catalytically active when its particle size is reduced to the nanoscale regime.Reference Hutchings¹ Excellent catalytic activities are observed for highly dispersed gold nanoparticles on metal oxide supports, particularly for CO oxidation.Reference Schubert, Hackenberg, van Veen, Muhler, Plzak and Jurgen Behm² The catalytic activity of gold nanoparticles is strongly dependent on particle size, but the composition and structure of the supporting oxide also play a role.Reference Schubert, Hackenberg, van Veen, Muhler, Plzak and Jurgen Behm²

The exact nature of the catalytically active species within metal oxide support systems is still unclear.Reference Hutchings¹ Several studies have suggested that Au⁰ is the active species, but mechanisms involving cationic gold (Au³⁺, Au⁺, or both) have also been proposed.Reference Fierro-Gonzalez and Gates³ The low concentration and fine dispersion of gold particles on metal oxide supports hamper experimental investigations of the catalytic species.Reference Hutchings¹ Modeling the system is also a challenge as a wide array of interactions between the gold nanoparticles and oxide supports are possible.Reference Fierro-Gonzalez and Gates³

To investigate the importance of Au³⁺ as a catalytic species, Kurzman et al.Reference Kurzman, Ouyang, Bin Im, Li, Hu, Scott and Seshadri⁴ have studied CO oxidation in the presence of La₄LiAuO₈. Originally reported by Abbattista et al.,Reference Abbattista, Vallino and Mazza⁵ La₄LiAuO₈ crystallizes in an ordered variant of the Nd₂CuO₄ structure, containing layers of corner-connected LiO₄ and AuO₄ square planes separated by two fluorite-type layers of La and O. Each AuO₄ moiety is isolated, which makes La₄LiAuO₈ an interesting model compound to investigate heterogeneous catalysis by square planar Au³⁺ ions. Although not an effective catalyst for the oxidation of CO, the compound is remarkably stable toward thermal decomposition, which occurs near 935 °C.Reference Kurzman, Ouyang, Bin Im, Li, Hu, Scott and Seshadri⁴ The observed weight loss upon decomposition has been reported as 2.7%, corresponding to the formation of metallic Au and either 2 La₂O₃ + ½ Li₂O or 1.5 La₂O₃ + LiLaO₂.Reference Kurzman, Ouyang, Bin Im, Li, Hu, Scott and Seshadri⁴ LiLaO₂ is more likely to form than Li₂O, but the results are ambiguous as only one small diffraction peak could be indexed to LaLiO₂ in the decomposed material.

LiLaO₂ is the only binary compound in the La₂O₃–Li₂O system that forms between 750 and 1000 °C.Reference Abbattista and Vallino⁶ It is believed to be isomorphous to LiEuO₂, but the detailed crystal structures of LiEuO₂ and LiLaO₂ are not known. LiLaO₂ was observed as a decomposition product of Li₃La(C₂O₄)₃·9 H₂O,Reference Deb, Baruah and Dass⁷ and it has been reported as an effective catalyst for methane oxidation.Reference Kinnemann, Kieffer, Kaddouri, Poix and Rehspringer⁸

The thermal stability of La₄LiAuO₈, along with the paucity of thermochemical data for trivalent gold oxide compounds, led us to investigate the energetics of La₄LiAuO₈ and its decomposition product LiLaO₂. Here, we present the enthalpies of formation for La₄LiAuO₈ and LiLaO₂ obtained by high-temperature oxide melt solution calorimetry and investigate the energetics of the La₄LiAuO₈ decomposition.

The La₄LiAuO₈ powder was synthesized according to previously reported methods by Kurzman et al.Reference Kurzman, Ouyang, Bin Im, Li, Hu, Scott and Seshadri⁴ LiLaO₂ was synthesized following the procedure by Abbattista and Vallino.Reference Abbattista and Vallino⁶ LiOH·H₂O (Aldrich Chemical 131066-3, Aldrich, St. Louis, MO) was heated for 16 h at 250 °C to form LiOH. An excess of LiOH (2:1 ratio) was ground with La₂O₃ (Aldrich Chemical 131281-8), pelletized, and wrapped in silver foil. The sample was heated at 775 °C for 10 h.

The samples were characterized by powder x-ray diffraction using a Bruker D8 Advance diffractometer (Cu Kα radiation; Bruker, Madison, WI) equipped with a LynxEye solid state detector. Rietveld refinements were performed using the XND code; the GSAS/EXPGUI suite was used for simultaneous Lebail fitting and Rietveld refinement; and the JADE 8.0 software was used for particle size estimation from Scherrer broadening. Thermogravimetric analysis (TGA)/differential scanning calorimetry (DSC) analysis was performed on a LABSYS Evo (Setaram Instruments, Caluire, France) from 30 to 1200 °C in air with heating rate of 10 °C/min.

High-temperature oxide melt solution calorimetry was performed using a specially built twin calorimeter of the Tian–Calvet type operating at 800 °C that has been previously described.Reference Navrotsky⁹ Pelletized samples (∼5 mg) were dropped into a platinum crucible containing 30 g of a lead borate, 2PbO·B₂O₃, solvent. The drop solution experiments were performed with air bubbling through the melt to sweep out evolved gases. The calorimeter was calibrated using ∼5 mg pellets of α-Al₂O₃ (99.997 wt% metal basis; Alfa Aesar, Ward Hill, MA). The drop solution enthalpy value for Li₂O is necessary component of the thermochemical cycle. Lithium oxide absorbs water to form LiOH·H₂O, and it is difficult to accurately measure the ΔH _dsol directly. However, the ΔH _dsol has been calculated from the ΔH _dsol of the more stable Li₂CO₃ compound and related thermodynamic parameters.

The powder x-ray diffraction pattern of La₄LiAuO₈ confirmed that the sample was nearly phase-pure with small amounts of La₂O₃ and metallic gold. Quantitative analysis by multiphase Rietveld refinement indicated the sample contained 0.5 wt% face-centered cubic (fcc) Au and 2.8 wt% La₂O₃. The x-ray diffraction pattern of the LaLiO₂ sample agreed with that obtained by Abbatista and VallinoReference Abbattista and Vallino⁶ and also indicated a small amount of La₂O₃. Because phase fractions cannot be estimated from diffraction data of a multiphase mixture when one (or more) of the phases lacks a structural model, quantification of the impurity amounts was determined by thermogravimetry, described in the following section. Preparing pure LaLiO₂ is problematic because it is a very reactive material; the absence of any high quality structural work in the literature reflects the difficulty in obtaining the phase. Despite numerous attempts, it was not possible to prepare better samples of LiLaO₂. The major impurity phases were lithium hydroxide and lanthanum oxide and have been corrected for in the thermochemical calculations.

The TGA of La₄LiAuO₈ indicates that the material persists up to 931 °C [Fig. 1(a)]. The DSC curve contains two endothermic peaks, one accounting for the decomposition of La₄LiAuO₈ and the second corresponding to the melting of gold (1064 °C). Kurzman et al.Reference Kurzman, Ouyang, Bin Im, Li, Hu, Scott and Seshadri⁴ previously reported the decomposition temperature for La₄LiAuO₈ in air as 935 °C, in good agreement with our results. The weight loss after decomposition of La₄LiAuO₈ is 2.6 wt%, which correlates with the final products: Au + 2 La₂O₃+ 1/2 Li₂O or Au + 1.5 La₂O₃ + LiLaO₂. The diffraction pattern of the decomposition products contains peaks of metallic gold and lanthanum oxide. An additional broad peak at 29° 2θ may be associated with a major peak of LiLaO₂, though this assignment is unclear given the absence of intensity at 35° 2θ, which is claimed to be the strongest reflection associated with LiLaO₂.Reference Abbattista and Vallino⁶

FIG. 1.

The thermogravimetric analysis/differential scanning calorimetry (TGA/DSC) curves for (a) La₄LiAuO₈ indicate that the compound is stable until 931 °C. The weight changes observed in (b) LaLiO₂ indicates that a small amount of the initial starting materials are still present in the final sample.

The TGA/DSC curve of LiLaO₂ is more complex, with weight changes occurring at several temperatures [Fig. 1(b)]. Initially, a weight gain occurs followed immediately by a weight loss from 50 to 110 °C, indicating the presence of a small amount of the initial LiOH starting material. Additional evidence for the presence of LiOH includes the sharp endothermic peak in the DSC curve at 425 °C that corresponds to the melting temperature of LiOH. Decomposition of LiOH to Li₂O is also evident with the weight loss and broad exothermic peak from 450 to 825 °C. From the weight loss in the TGA curve, the amount of LiOH in the sample was estimated to be 1.8 wt%. The weight loss at 300 °C corresponds to the decomposition of La(OH)₃ to La₂O₃. Lanthanum oxide readily transforms to a hydroxide phase in the presence of a base (LiOH). The weight loss associated with the decomposition of La(OH)₃ indicates that La₂O₃ is present as 12 wt% in the sample.

La₄LiAuO₈ and LiLaO₂ dissolved completely in the lead borate melt, which was evidenced by a stable baseline during the calorimetric experiments, consistent measured enthalpies, and visual inspection of the solidified melt. The heat effects observed for both samples were endothermic. The average ΔH _dsol (in kJ/mol) for the La₄LiAuO₈, LiLaO₂, and Li₂CO₃ and two standard deviations of the mean were calculated from a total of eight to ten measurements for each and are summarized in Table I.

TABLE I.

Enthalpy of drop solution ΔH _dsol (kJ/mol) for Li₂CO₃, La₄LiAuO₈, and LiLaO₂ in lead borate at 800 °C.

^a Error is two standard deviations of the mean.

The thermochemical cycle used to calculate the enthalpy of drop solution for Li₂O is given in Table II. The formation enthalpy of Li₂CO₃ was calculated from Li₂O and CO₂ gasReference Chase, Davies, Downey, Furrip, McDonald and Syverud¹⁰ and the heat capacity of CO₂ gas from 25 to 800 °C.Reference Robbie and Hemingway¹² Using the enthalpy of drop solution for Li₂CO₃, the ΔH _drp−sol for Li₂O was calculated to be −23.83 ± 5.06 kJ/mol.

TABLE II.

Thermochemical cycles used to determine ΔH _f−ox for La₄LiAuO₈.

The enthalpy of drop solution for La₄LiAuO₈ was corrected for the 7.16 mol% La₂O₃ and 2.1 mol% Au resulting in ΔH _dsol−corr of 160.16 ± 1.84 kJ/mol. The thermochemical cycles used in the calculations for the heats of formation from the oxides for La₄LiAuO₈ are summarized in Table II. The ΔH _corr−dsol for La₄LiAuO₈ (ΔH ₅), the ΔH _dsol for LaO_1.5 (ΔH ₆),Reference Robbie and Hemingway¹² ΔH _dsol of Li₂O (ΔH ₄) and the integrated heat capacity from 298 to 800 °C for Au metal (ΔH ₇) and O₂ gas (ΔH ₈)Reference Robbie and Hemingway¹² were used to calculate the enthalpy of formation from the oxides for La₄LiAuO₈, −187.66 ± 5.84 kJ/mol.

The residue from the thermal analysis experiment was analyzed by powder x-ray diffraction. The diffraction peaks matched those of fcc Au and La₂O₃. Li₂O was not observed in the x-ray diffraction pattern. The three peaks associated with Au metal were fit, and the average particle size was calculated based on the ScherrerReference Scherrer¹¹ equation to be 39 ± 2 nm.

The effect of the Au particle size on the heat of formation of La₄LiAuO₈ was estimated using the following relationship

where SA is the surface area of the particle in m²/mol and γ is the surface energy (J/m²). The surface energy of gold nanoparticles is in the range of 1.2–1.5 J/m².Reference Arcidiacono, Bieri, Poulikakos and Grigoropoulos¹⁴ The value used for these calculations, 1.46 J/m², was taken from size-dependent evaporation data reported by Nanda.Reference Nanda¹⁵ Using this value, the difference in the ΔH _dsol−nano from the ΔH _dsol−bulk is 1.12 kJ/mol. Therefore, the ΔH ₇ value decreases from 20.97 to 19.85 kJ/mol. This change results in a ΔH _f−corr value for La₄LiAuO₈ of −188.77 ± 5.1 kJ/mol. The correction, however, is within the experimental error of the enthalpy of formation.

Few studies on the energetics of trivalent gold oxides have been reported. The enthalpy of formation for Au₂O₃ has been determined using several different techniques, including DSC.Reference Nanda¹⁵ The experimental values for the ΔH _f of Au₂O₃ range from −3.5 to −13 kJ/mol.Reference Ashcroft and Schwarzmann¹⁶ DFT calculationsReference Shi, Asahi and Stampfl¹⁷ estimate ΔH _f of Au₂O₃ to be −50.1 kJ/mol. Thus, La₄LiAuO₈ is energetically more stable than Au₂O₃ when compared on a per mol basis. The heats of formation for other trivalent gold oxides have not been reported.

The average ΔH _dsol for LiLaO₂ was based upon 10 drops, resulting in a value of 45.38 ± 1.40 kJ/mol (Fig. 1). The value was corrected for the presence of 12.6 mol% LiOH and 6.2 mol% La₂O₃. The drop solution value for LiOH has not been measured, but a thermochemical cycle can be written from ΔH _drp−sol (ΔH ₁) and ΔH _f−ox (ΔH ₈) of Li₂CO_3,Reference Chase, Davies, Downey, Furrip, McDonald and Syverud¹⁰ the integrated heat capacity of H₂O (ΔH ₁₀) and CO₂ (ΔH ₃) gas from 800 to 25 °C,Reference Robbie and Hemingway¹² and the formation enthalpy (ΔH ₁₁) of LiOH (Table III).Reference Chase, Davies, Downey, Furrip, McDonald and Syverud¹⁰ The corrected enthalpy of drop solution (ΔH ₁₃ = ΔH _dsol−corr) for LaLiO₂ is 15.62 ± 8.18 kJ/mol.

TABLE III.

Thermochemical cycles used to determine ΔH _f-ox for LiLaO₂.

The enthalpy of formation for LaLiO₂ was calculated from the ΔH _dsol−corr for LaLiO₂ (ΔH ₁₃), the measured enthalpy of drop solution for La₂O₃ (ΔH ₆), and the calculated drop solution value for Li₂O (ΔH ₄), resulting in a value of −41.4 ± 9.6 kJ/mol. The large error associated with the ΔH _f−ox for LaLiO₂ is due to the presence of LiOH and the error of the ΔH _f−ox of LiOH.

The thermal stability of La₄LiAuO₈ is interesting given the oxophobic nature of gold. Kurzman et al.Reference Kurzman, Ouyang, Bin Im, Li, Hu, Scott and Seshadri⁴ suggest that the stability can be attributed to the presence of covalent Au–O bonding due to the presence of highly electropositive counter cations. To provide insight into the decomposition of La₄LiAuO₈, the decomposition temperature was calculated from the experimentally derived thermodynamic data for La₄LiAuO₈ and LiLaO₂.

The enthalpy of decomposition based upon the formation of LaLiO₂ can be determined from the following thermochemical cycle

The Ellingham diagram for oxidation reactions involving solid phases plots ΔG⁰ versus temperature.Reference Ellingham¹⁸ As ΔG⁰ = ΔH⁰ − TΔS⁰, the slope of the line is equal to −ΔS⁰ and the lines are generally parallel to one another for most oxides. For the decomposition of La₄LiAuO₈, the oxygen is released rather than consumed so the entropy change is positive. The entropy value used for these calculations, per mole of O₂ gas, (145 J/mol K) was taken from the average value for metal oxides.Reference Gaskell¹⁹

At a given temperature and pO₂

so

The calculated decomposition temperature is within error of the experimental value.

Considering the formation of Li₂O as a decomposition product of La₄LiAuO₈ involves the following reaction

Inserting this ΔH _decomp value into the equation for ΔG _decomp results in a decomposition temperature of 1331 ± 43 °C, which is 396 °C higher than the experimental value. Therefore, the thermodynamics of the decomposition reaction of La₄AuLiO₈ indicate that LaLiO₂ is the likely product.

The results from the oxide melt solution calorimetry of La₄LiAuO₈ indicate that the compound is thermodynamically stable with respect to the constituent oxides and is more stable than binary trivalent gold oxide, Au₂O₃. The enthalpy of formation for LaLiO₂ with respect to the oxides is small and exothermic. The decomposition of La₄LiAuO₈ occurs at 931 °C and results in metallic gold, lanthanum oxide, and lanthanum lithium oxide. The formation of LaLiO₂ as a decomposition product for La₄LiAuO₈ was confirmed by calculating the decomposition temperature from measured thermodynamic parameters. The calculated value (979 ± 95 °C) agrees well with the experimental results.