Introduction

The search for long-wave infrared (LWIR) transmitting materials with combined thermo-mechanical and optical performance has been the subject of intensive research efforts over the past 40 years. For applications as protective windows for optronic equipment (missile domes), high hardness and toughness are indicated to withstand harsh environments conditions (i.e. resistance to rain erosion, sand wind, salt fog, etc.) and require fine-grained optics. As a result, polycrystalline ceramics outperform their competitors that are single crystals and glasses in the development of transparent materials for applications as external structural windows and have become strategic materials particularly for the defence industry1,2,3,4.

Today, the chemical vapor deposited (CVD) zinc sulfide (ZnS) is the reference material for applications in the 3–5 μm and 8–14 μm atmospheric windows and has been used for decades1. However, its low hardness makes it poorly resistant to rain erosion. This vulnerability is primarily due to its fabrication process (CVD-Hot Isostatic Pressing or HIP), which produces coarse-grained microstructures, thus reducing both hardness and mechanical strength1. In the early 1980s, compositions of the CaLa2S4-La2S3 solid solution (hereby referred to as CLS) were developed as alternative window materials to CVD ZnS to meet the need for improved performance in both erosion resistance and long-wave IR transparency5,6,7,8,9,10,11. CLS compounds, that belong to the MLn2S4 (M = alkaline earth element, Ln = lanthanide element) family of cubic Th3P4-type structure, offer a broad transparency window (0.5–17 μm), high refractoriness (melting point >1800 °C) and hardness (HK = 570 kg.mm−2), low refractive indices (~ 2.5) and most importantly, substantially better resistance to rain erosion at all velocities compared to ZnS5,12,13. Despite these benefits, the reproducibility of the elaboration process has never been fully mastered and the material was never commercialized. Problems were thought to be a result of variability of the powder7 caused or increased by lengthy and complicated processes. The employed sintering techniques may also bring long-standing issues with phase impurity and point defects which distort the optical properties expected from the pristine material. For instance, sulfur loss - inherent to sulfides high-temperature treatment - is a major challenge in the processing of CLS compounds14,15,16.

The research presented here investigates barium lanthanum sulfide compounds of the MLn2S4 family for which there is very little literature data. The stoichiometric BaLa2S4 and compositions within the BaLa2S4-La2S3 solid solution (referred to as BLS) exhibit properties comparable to those of CLS compounds, with a potentially wider transmission window due to the presence of barium which is 3.5 times heavier than calcium. To the best of our knowledge, this compound has never been produced as dense ceramics. So far, only its lattice parameter and band gap have been reported17,18.

These initial attempts to develop transparent BLS ceramics were based on the work of C. Chlique19 on BLS powder synthesis and G. Durand12 on CLS ceramics processing. In the absence of data concerning the BaS-La2S3 binary system, we assumed, as with CaS-La2S3, the existence of a complete solid solution domain for La/Ba molar ratios >2, and a two-phase domain, i.e. BaS + BaLa2S4, for La/Ba < 2. A phase diagram, adapted from the calcium system, is proposed in Fig. 1. The composition of the BLS material developed in this work was arbitrarily set within the solid solution domain to a molar ratio La/Ba = 2.7 to prevent the formation of the BaS impurity phase, which would cause optical scattering. Previous work on CLS also showed that higher molar ratios (La/Ca = 5 or 10) resulted in the presence of La10S14O oxysulfide phase (also known as β-La2S3) as a secondary phase due to a less efficient stabilization of the CLS cubic phase by lower Ca contents12,14. The chemical formulation of the composition with a molar ratio La/Ba = 2.7 is Ba0.79La2.14S4, referred to as BLS2.7 in the following, while the stoichiometric compound is referred to as BaLa2S4.

Fig. 1
Fig. 1
Full size image

BaS/γ-La2S3 binary phase diagram, adapted from CaS/γ-La2S320.

Precursor powders were synthesized using a combustion method and subsequently sulfurized in pure H2S atmosphere. The densification of these powders was then investigated through both hot-pressing and natural (pressureless) sintering to produce transparent ceramics. The study examined the influences of these techniques on the purity, microstructure and optical properties of the ceramics, discussing the conditions necessary for achieving LWIR transparency free of impurity absorption. Crystal structure refinements performed on the processed ceramics revealed a deviation between the BLS theoretical lattice parameters, as computed from Vegard’s law in the literature, and the experimental values, leading to inconsistencies with the compositional and optical properties of the ceramics. The reinvestigation of the lattice parameter of the stoichiometric BaLa2S4 is presented first, allowing for the computation of the theoretical lattice parameter and theoretical density for any composition of the solid solution.

Results

Determination of stoichiometric BaLa2S4 lattice parameter

Considering the Ca-La-S system, CLS compositions are formed by the addition of Ca2+ bivalent cations into the defective γ-La2S3 with the cubic Th3P4-type structure (La8/3□1/3S4 where □ represents a lanthanum vacancy)17,21. The addition occurs by progressively filling 1/3 of the cation vacancies and substituting 2/3 of La3+ ions, with a statistical distribution of Ca2+ ions. As a consequence, for molar ratios La/Ca >2, there exists a complete solid solution domain for the ternary sulfides defined as \(\:{Ca}_{x}{La}_{\frac{8-2x}{3}}\:{\Box}_{\frac{1-x}{3}}\:{S}_{4}\) (0 < \(\:x\) < 1) with a cubic lattice parameter that evolves linearly according to a Vegard’s law20. However, regarding the Ba-La-S system, Flahaut et al. reported in 1965 a significant deviation between the experimental and theoretical lattice parameters for the BaLa2S4-La2S3 solid solution. Unlike other ternary sulfide systems that comply with Vegard’s law, the results for the barium’s system were never presented or further discussed. Additionally, regarding the lattice parameter of the stoichiometric compound BaLa2S4, there is no consensus in the literature which reports at least two different values: 8.917 Å17 and 8.9090 Å22,23. To address this, we determined the lattice parameter of BaLa2S4 to compute a corrected Vegard’s law and verify its applicability to the BLS system. To achieve this, we studied a composition within the La/Ba < 2 two-phase domain of the binary diagram to ensure the synthesis of the stoichiometric compound BaLa2S4 and compare its experimental lattice parameter to those reported in the literature. Since Vegard’s law computations for the CLS solid solution were proven to be accurate12, it was assumed that the γ-La2S3 lattice parameter was indeed correct (\(\:a\) = 8.731 Å24) and that only BaLa2S4 needed verification.

A starting La/Ba molar ratio of 1.9 (referred to as BLS1.9) was chosen to work with a composition within the two-phase BaS + BaLa2S4 domain of the BaS-La2S3 phase diagram. After sulfurization, the powders were characterized by XRD followed by Rietveld refinement. The final Rietveld refinement patterns are available as Supplementary Figs. S1 and S2. Figure 2 shows the XRD pattern of the BLS1.9 powder before and after natural sintering. The powder XRD pattern indicates the formation of a BaLa2S4 phase with the expected presence of residual BaS. The Rietveld refinement conducted on the powder gives a lattice parameter of 8.9283(1) Å with approximately 1.6 wt.% of BaS which is in perfect agreement with the calculated amount of residual BaS expected from the BLS1.9 composition (i.e. 1.6 wt%). Let us note, however, that, due to the detection limit of the equipment (about 1–2 wt%), the XRD pattern of the ceramic obtained from natural sintering does not show any noticeable secondary BaS phase but only the BaLa2S4 compound.

Fig. 2
Fig. 2
Full size image

X-ray diffraction patterns of pure BLS1.9 powder and BLS1.9 ceramic sintered naturally for 12 h at 1250 °C under pure H2S.

The refined cubic lattice parameter of the ceramic is 8.9324(1) Å (compared to 8.909 Å or 8.917 Å in the literature17,22,23. If the system does obey a Vegard’s law, a theoretical lattice parameter of 8.8905 Å (Fig. 3) and a theoretical density of 5.0484 g.cm−3 are determined for the BLS2.7 composition. The cubic lattice parameter of BaLa₂S₄ (8.9324 Å) is larger than that of CaLa₂S₄ (8.687 Å21), consistent with the difference in ionic radii between Ba²⁺ and Ca²⁺ ions in 8-fold coordination25. Additionally, the experimental densities of BLS align well with those of CLS (ρ = 5.06 g/cm³ for BaLa₂S₄ (our study) and ρ = 4.52 g/cm³ for CaLa₂S₄21).

Fig. 3
Fig. 3
Full size image

Lattice parameter evolution within the BaLa2S4-La2S3 solid solution computed according to Vegard’s law from literature data22,23 and refined pure BaLa2S4 from this study. The x value represents the incorporated amount (by filling and substitution) of Ba2+ ions into defective γ-La2S3.

BLS2.7 powder characterizations

XRD patterns of both as-combusted and post-sulfurized BLS2.7 powders are presented in Fig. 4. The as-combusted precursor powder is composed of multiple phases with close or overlapping diffraction peaks. Readily identified phases are lanthanum oxysulfide and oxysulfate, lanthanum hydroxide, barium sulfate, barium carbonate, lanthanum oxide and barium oxide. For clarity, only the main phases are distinguished in Fig. 4. The 4 h-sulfurization treatment effectively converts the precursor entirely into pure BLS (BaLa2S4 ICCD card n° 04-002-2427), resulting in the formation of a yellowish powder.

Fig. 4
Fig. 4
Full size image

X-ray diffraction patterns of as-combusted and pure (post-sulfurized) BLS2.7 powders.

In order to further assess the purity of the post-sulfurized powder, FTIR spectroscopy analysis was carried out and corroborates XRD results presented above (Fig. 5). The as-combusted powder shows a broad absorption band around 3300–3610 cm−1 that is characteristic of H2O elongation modes26,27. Carbonate groups, identified by two absorption bands around 1350–1580 cm−1 and 800–900 cm−1, are associated to the presence of a barium carbonate phase identified through XRD28,29. Absorption bands at 1400 cm−1, 1000–1250 cm−1 and 560–700 cm−1 are attributed to the elongation vibrations of sulfate groups30,31,32,33,34. Finally, absorption bands centered around 400–550 cm−1 are characteristic of the vibrational modes of La-O and La-S bonds in La2O2S35. These absorption bands are completely eliminated after a 4 h-sulfurization treatment in H2S, indicating the high purity of the BLS2.7 powder.

Fig. 5
Fig. 5
Full size image

FTIR analysis of as-combusted and pure (post-sulfurized) BLS2.7 powders.

Figure 6 shows the SEM micrographs of the as-combusted and post-sulfurized powders. The as-combusted powders exhibit heterogeneous morphology and particle size distribution in agreement with the multiple phases highlighted by XRD and FTIR analysis. The powder mainly consists of agglomerates of up to a few tens of micron size with a few primary particles of nanometric size (~ 110 nm). After sulfurization, the pure BLS2.7 powder forms agglomerates composed of particles that are partially sintered as a result of the sulfurization heat treatment at 1000 °C. The specific surface areas are consistent with the observed morphologies. The 6 m²/g value measured for the as-combusted powder is reduced to 1 m²/g after sulfurization.

Fig. 6
Fig. 6
Full size image

SEM images of (a) as-combusted and (b) pure BLS2.7 powders.

It is worth noting that the conditions ensuring powder purity (heat treatment in H₂S for 4 h) are identical to those applied for CLS powders. This treatment leads to the formation of agglomerates in both compounds, consisting of partially sintered primary particles, and results in a decrease in SSA to 1 m²/g36.

Powders densification study

The ceramics were obtained after a 2 h-hot-pressing (120 MPa) at 1250 °C. Figure 7 presents the XRD patterns of BLS2.7 sample before and after hot-pressing. No drastic change is observed in the ceramic pattern compared to the initial powder except for the appearance of a small amount of La10S14O (β-La2S3, ICCD card n° 01–080–1900) as a secondary phase. Hot-pressed ceramics are black, thus opaque to visible light, revealing some reduction phenomenon (Fig. 8 (a)).

Fig. 7
Fig. 7
Full size image

X-ray diffraction patterns of pure (post-sulfurized) BLS2.7 powder, BLS2.7 ceramic hot-pressed for 2 h at 1250 °C and post-annealed for 12 h at 1250 °C under pure H2S.

Fig. 8
Fig. 8
Full size image

Photographs of a hot-pressed BLS2.7 ceramic (a) before and (b) after a 12 h-annealing under pure H2S.

This reduction phenomenon had previously been observed with CLS compounds12,14,15,16. Two hypotheses were tentatively suggested, at the time, to explain this phenomenon: (1) A direct contact between the graphite from the die and the sample according to : CaLa2S4(s) + εC (s) \(\:\underrightarrow{kT}\) CaLa2S4−2ε(s) + εCS2 (g)37; (2) A gaseous-phase reduction caused by the formation of carbon monoxide (CO)12. G. Durand showed that CaLa2S4 blackening indeed did occur without any direct contact between the graphite die and the sample. The formation of CO would result from the possible presence of oxygen traces in the hot-press and/or the insulating graphite felts38,39. This process is highly likely to proceed here as XRD analysis identified some La10S14O phase as impurity phase.

As a result, the reduction phenomenon leads to the production of sulfur-deficient phases. To maintain electrical neutrality, metallic bonds are formed, which explain their black colour and opacity12,40. The slight deviations in stoichiometry in the resulting compositions leads to an increase in the lattice parameter as evidenced by the observed asymmetry of the diffraction peak bases at low 2θ after sintering (Fig. 7). To restore the sulfur lost during hot-pressing, the samples were annealed for 12 h (6 h dynamic flow + 6 h static atmosphere) under pure H2S at 1250 °C. As expected, the heat-treated samples changed from black to brownish after annealing (Fig. 8 (b)). While the diffraction peaks associated to the La10S14O oxysulfide phase are still observed, the asymmetry of the X-ray diffraction peaks of the BLS2.7 phase disappeared after annealing under H2S (Fig. 7). The Rietveld-refined lattice parameter is estimated to be: \(\:a\) = 8.8901(4) Å (Fig. S3). If we consider a Vegard’s law between the stoichiometric compounds BaLa2S4 (\(\:a\) = 8.9324 Å) and γ-La2S3 (\(\:a\) = 8.731 Å24) (Fig. 3), this value is extremely close to the theoretical one expected for the compound BLS2.7,i.e. 8.8905 Å. A densification rate of 99% is estimated considering a theoretical density of 5.0484 g.cm−3 for BLS2.7.

The elemental composition of the ceramic was determined using EDS analysis (Table 1). Despite the limited accuracy of EDS measurements (± 1 at.%), the results indicate that the observed La/Ba atomic ratio is close to the theoretical La/Ba ratio introduced during synthesis. The annealed ceramics show a lighter colour, indicating that the decrease of sulfur vacancy concentration after annealing in pure H2S reduces the optical absorption.

Table 1 Elemental composition of theoretical BLS2.7 and EDS analysis of hot-pressed BLS2.7 ceramics post-annealed at 1250 °C under H2S for 12 h and BLS2.7 ceramics sintered naturally at 1250 °C under H2S for 12 h.
Full size table

The SEM micrographs (Fig. 9) reveal a dense microstructure with some residual porosity. The average grain size is approximately 60 μm, with d10 and d90 percentiles around 25 μm and 105 μm, respectively. The micrometric intragranular porosity, along with the significant deviation between the statistical d10 and d90 grain size values, suggest abnormal grain growth, that is likely due to prolonged annealing conditions (12 h at 1250 °C).

Fig. 9
Fig. 9
Full size image

SEM images of thermally etched surface of BLS2.7 ceramic hot-pressed for 2 h at 1250 °C and annealed at 1250 °C for 12 h under H2S (a and b).

The hot-pressed ceramic is transparent in the infra-red region with a maximum transmission of 20% at 16.5 μm (Fig. 10). Absorption bands, that are visible throughout the material’s transparency window, are a consequence of oxygen impurities stabilizing the impurity phase La10S14O (or β-La2S3) that has been identified in the hot-pressed ceramics before and after annealing under H2S. During sintering and post-annealing, the latter phase would possibly transform into γ-La2S3 (only partially in this study) liberating oxygen, forming S-O bonds. The bands at 9 μm, 10 μm and 16 μm are associated with elongation modes of SO42− groups8,13,20,41 while those at 10.5 μm and 15 μm are associated with SO32− groups13,42. The bands in the range 12–14 μm could not be identified but have also been reported for the compound CaLa2S412,13.

Fig. 10
Fig. 10
Full size image

IR transmission spectrum of hot-pressed BLS2.7 ceramic annealed for 12 h at 1250 °C under pure H2S (thickness: 1.4 mm).

Considering the observed reduction associated with the hot-pressing process, the powders were subsequently densified through natural sintering for 12 h at 1250 °C in an H2S atmosphere with the intent of preventing both any oxygen uptake and sulfur loss. The as-sintered ceramics exhibit a light brown colour (Fig. 11 (a)).

Fig. 11
Fig. 11
Full size image

(a) Photograph of BLS2.7 ceramic naturally sintered for 12 h at 1250 °C under pure H2S and (b) X-ray diffraction patterns of pure BLS2.7 powder and BLS2.7 ceramic sintered naturally.

XRD characterization identifies all diffraction peaks as belonging to BLS (BaLa2S4 ICCD card n° 04-002-2427) (Fig. 11 (b)) with no trace of La10S14O detected. An experimental lattice parameter of 8.8886(1) Å was determined by Rietveld refinement which is very close to the theoretical value of 8.8905(1) Å, considering our new Vegard’s law (Fig. S4). Additionally, EDS analysis confirms that the Ba, La, and S atomic percentages align with the theoretical values expected for the BLS2.7 compound (Table 1).

The SEM micrographs (Fig. 12) show a dense microstructure with large residual porosity. The grain size is approximately 20 μm, with d10 and d90 percentiles in the order of 10 μm and 30 μm, respectively. These grain sizes are relatively large but were expected given the high temperature and extended sintering time. The significant porosity can be attributed to the agglomeration of the powder, resulting from the partial sintering of the particles, as described in previous work, which promotes differential sintering within the agglomerates43,44. The processed ceramic exhibits a densification rate of 97% (according to the new Vegard’s law) which is insufficient to achieve transparency. However, with no open porosity, improved densification might be achievable through subsequent containerless Hot Isostatic Pressing.

Fig. 12
Fig. 12
Full size image

SEM images of thermally etched surface of BLS2.7 ceramic sintered naturally at 1250 °C for 12 h under H2S (a and b).

Discussion

Preliminary work in this study consisted in the determination of the BaLa2S4 lattice parameter to define the Vegard’s law of the supposedly solid solution formed between BaLa2S4 and the binary compound γ-La2S3. A lattice parameter of 8.8905 Å and a density of 5.0484 g.cm−3 were determined for the BLS2.7 composition. Refined lattice parameters of both hot-pressed and naturally sintered ceramics converged on the theoretical value computed from the Vegard’s law, hence on a La/Ba ratio of 2.7. Elemental composition of all ceramics was also confirmed thanks to EDS analysis, perfectly matching the expected BLS2.7 composition. Moreover, densification rates of the hot-pressed and the naturally sintered ceramics, estimated to be of 99% and 97%, respectively, are in accordance with their respective optical IR transparency and opacity properties. These results corroborate the validity of the new Vegard’s law computed from the BaLa2S4 lattice parameter determined in this work. Hence, this study is refuting, at least for BaS-rich compositions, Flahaut’s observation postulating for BaLa2S4-La2S3 systems, contrarily to CaLa2S4-La2S3, a clear deviation from Vegard’s law.

If only a few literature is available on barium lanthanum sulfides, their calcium analogues have been extensively studied. Saunders et al. defined a ternary solid solution domain between the stoichiometric ternary composition CaLa2S4 and the binary lanthanum sulfides, La2S3 and La3S45. These authors concluded that the processing of a transparent material required single-phased and fully sulfurized materials, respectively, to avoid any scattering phenomenon and limit absorption phenomenon5. Fully sulfurized compositions lie along the CaLa2S4-La2S3 join, where, as the anion network is completely filled, any sulfur deficiency in the compound results in an excess of cations (Ca2+ or La3+). This leads to the creation of charge carriers, brought by a lanthanum excess, and consequently the loss of the material’s optical properties12. Figure 13 highlights the CaLa2S4-La2S3-La3S4 ternary solid solution domain described above.

Fig. 13
Fig. 13
Full size image

(adapted from)12,45.

Ternary Ca-La-S phase diagram with the relevant phases. The CaLa2S4-La2S3-La3S4 ternary solid solution domain is highlighted in yellow colour. [] represents a cation vacancy.

The same observation can be made for our barium hot-pressed ceramics as they appear completely black and opaque after sintering. The annealing under H2S atmosphere enabled to correct the sulfur deficiency, restoring the ceramic colour and the IR transparency. As highlighted by the EDS analysis (Table 1), the annealing however did not allow to regenerate completely the BLS2.7 theoretical composition. The remaining small deviations in the stoichiometry (sulfur deficiency and lanthanum excess) still induce absorption phenomenon, explaining why only 20% transparency was reached in the IR region. In contrast, atomic compositions of naturally sintered ceramics converge onto the theoretical expected percentages, testifying that a perfectly sulfurized composition is obtained thanks to the thermal treatment under H2S. Unfortunately, the obtained microstructure, characterized by an important porosity rate (of approximately 3%), is not optimal for any infrared transparency. These porosities are challenging to eliminate and detrimental to the optical properties. Further research could focus on powder treatment (e.g. grinding) or new powder shaping approaches (wet techniques) to improve the homogeneity of the particles’ arrangement while maintaining a fine-grained powder.

In summary, pure barium lanthanum sulfide (BLS) powders were successfully synthesized by a combustion method at 700 °C followed by high-temperature treatment (1000 °C) under pure H2S atmosphere. Natural sintering and hot-pressing were employed to sinter BLS powders of composition Ba0.79La2.14S4 (BLS2.7) at 1250 °C into dense ceramics so as to achieve IR transparency. Hot-pressing (P = 120 MPa) densified the powders to over 99% in just 2 h at 1250 °C. Natural sintering under H2S also achieved a rather high densification rate (97%) at the same temperature after a 12 h-treatment. A reduction phenomenon is though observed for the hot-pressed ceramics (loss of sulfur, black colour) in the reducing environment of the hot-press. Reduction in sulfur content results in a tendency towards metallic character with a consequent increase in free carrier absorption effects. Sulfur loss and ceramics colour can be restored under thermal treatment in H2S. Our experimental conditions led to achieve a peak IR transmission of 20% at 16.5 μm for 1.4 mm thick ceramics. Contrary to our prediction, BLS does not offer a wider transmission window compared to CLS. The transparency of BLS ceramics, especially in the 6–18 μm range, makes them uniquely suited for use in harsh environments to detect atmospheric components such as O3, H2O or CO2 (search for exoplanets).

The synthesis of a composition belonging to the two-phase domain of the BaS-γ-La2S3 binary phase diagram allowed us to determine the lattice parameter of the stoichiometric BaLa2S4 compound. This value, different from the ones reported in the literature, led to define a Vegard’s law which allowed to determine the lattice parameter and density of the BLS2.7 studied compositions. The results are in complete agreement with the compositional and optical properties of the ceramics and indicate that, for BaS-rich compositions, the BaLa2S4-La2S3 system does obey a Vegard’s law.

Finally, this preliminary study also highlighted, for further processing studies, to concentrate on the elimination of residual La10S14O oxysulfide phase or “β” phase, generating “S-O” type absorption bands after sintering that significantly impact the IR transmission of the ceramics.

Materials and methods

Powders synthesis

Precursor powders were prepared by a solution combustion method described in previous work14,43,46,47. Starting materials, La(NO3)3, 6H2O (Alfa Aesar, 99.99%), Ba(NO3)2 (ACROS organic 99 +%) and thioacetamide (TAA) CH3CSNH2 (ACROS organic, 99 +%) were dissolved in distilled water at ~ 70 °C under magnetic stirring. Lanthanum nitrate and barium nitrate were introduced respecting La/Ba molar ratios of 2.7 and 1.9 allowing the synthesis of, respectively, the compound of composition Ba0.79La2.14S4 (BLS2.7) and the stoichiometric compound BaLa2S4 (BLS1.9). A fuel (TAA)/oxidant (nitrate) molar ratio of 1 was used in all preparations. After complete dissolution, the solution was introduced into a muffle furnace (Nabertherm GmbH L40/11/B410) pre-heated to 700 °C. The as-obtained white foamy product is subsequently ground and post-sulfurized in a tubular furnace (Naberthem GmbH R 50/250/13) in pure H2S flow at 1000 °C for 4 h.

Powder sintering by hot-pressing

Hot-pressing was carried out under dynamic vacuum (about 0.3–0.4 mbar) using VAS S.A. (France) equipment. A typical quantity of 2 g of BLS2.7 powder was introduced in a 13 mm diameter graphite in order to obtain 3 mm-thick ceramics after sintering. The die was coated beforehand with boron nitride (BN, Alfa Aesar 99.5%) to prevent carbon diffusion and facilitate unmolding. Powders were sintered at 1250 °C with a heating rate of 10 °C/min. A load of 120 MPa (maximum pressure allowed on graphite die) was applied from 800 °C till the end of the 2 h-dwell time at 1250 °C. At the end of the dwell time, the pressure was released and samples were cooled to 1000 °C at a rate of 4 °C/min, then naturally cooled down to room temperature.

Powder sintering by natural sintering

Green bodies were shaped beforehand by uniaxial cold pressing. Typical quantities of 2 g of BLS2.7 powder and 1 g of BLS1.9 powder, introduced in 20 mm and 13 mm diameter stainless steel matrices, respectively, produced 2.5 mm to 2.8 mm-thick ceramics after sintering. BLS green bodies were placed in an alumina boat and sintered for 12 h in pure H2S (6 h dynamic flow + 6 h static atmosphere) at 1250 °C in a tubular furnace (Naberthem GmbH R 50/250/13).

Characterization of powders and ceramics

X-ray diffraction (XRD) patterns were recorded at room temperature in the 5–90° 2θ range with a step size of 0.0261° and an effective scan time per step of 40 s using a PANalytical X’Pert Pro diffractometer (Cu-L2, L3 radiation, λ = 1.5418 Å, 40 kV, 40 mA, PIXcel 1D detector). Data Collector and HighScore Plus softwares were used respectively to record and analyze the diffractograms. The XRD patterns for Rietveld refinements were collected at room temperature in the 5–120° 2θ range with a step size of 0.0131° and an effective scan time per step of 200 s. All calculations were carried out with Fullprof and WinPLOTR programs48,49. The pseudo-Voigt (NPr = 5) profile function was used and the background was approximated by linear interpolation between a set of background points. The I\(\:\stackrel{-}{4}\)3d (#220) space group and lattice parameters: a = b = c = 8.9324 Å and = = = 90 ° were used as input data. Thermal parameters and site occupancies were not refined (La: 12a (3/8, 0, 1/4), Ba: 12a (3/8, 0, 1/4), S: 16c (1/12, 1/12, 1/12)). The theoretical lattice parameter of the BLS2.7 compound was calculated using a Vegard’s law, interpolating between stoichiometric BaLa2S4 and γ-La2S3 (\(\:a\) = 8.731 Å24).

A Gemini VII 2390 Micromeritics analyzer was used to determine the specific surface area of the powders by the Brunauer-Emmett-Teller method. Before measurement, the samples were outgassed overnight in vacuum at 250 °C.

Scanning Electron Microscopy (SEM) for powders morphology and ceramics microstructure analyses was performed using a JEOL JSM 7100 F and JEOL IT-300 LA, respectively. Prior to SEM observations, powder and ceramic samples were metallized with gold. Particle size of the powders was estimated from SEM images via the ImageJ 1.53o software. Grain boundaries of the ceramics were revealed by thermal etching at 1010 °C for 30 min in H2S/N2 flow. SEM observations of the thermally etched samples were then used to determine the ceramic grain size from about 100 to 400 grains via ImageJ.

Elemental composition of the ceramics samples was determined through Energy Dispersive Spectroscopy (EDS), attached to the JEOL IT-300 LA SEM, under low vacuum conditions (100 Pa) without metallization.

Fourier Transform Infra-Red (FTIR) patterns of powders samples dissolved in KBr pellets and transmission spectra of the ceramics were recorded at room temperature with a Bruker Vector 22 FT-IR spectrometer, between 400 and 4000 cm−1 and 400 and 8000 cm−1, respectively. Prior to transmission measurements, the ceramics were optically polished using 0.1 μm alumina powder.

Densities and open porosity rates of the ceramics samples were determined, with an accuracy of ± 1%, using the Archimedes method, respectively, in absolute ethanol and after impregnation in distilled water.