Temperature Programmed Oxygen Desorption and Sorption Processes on Pr2-XLaxNiO4+δ Nickelates

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. Temperature programmed oxygen desorption and sorption processes on Pr2-XLaxNiO4+δ nickelates Alexandra Usenka, Vladimir Pankov, Vaibhav Vibhu, Aurelien Flura, Jean-Claude Grenier, Jean-Marc. Bassat

The present work is devoted to the investigation of temperature programmed oxygen desorption -sorption (TPD) processes performed on SOFC innovative cathode materials with Pr2-xLaxNiO composition. The experiments were performed using the coulometric titration technique involving an , over the temperature cycle 20 900 20 parameters were the time and temperature dependences of both the titration current and the oxygen over-stoichiometry. TPD spectra of Pr2-xLaxNiO (x 0) exhibited two sharp peaks (maxima) of titration current with excellent resolution at p(O2) = 50 Pa, induced by removal of oxygen from different sites of the crystal lattice. TPD spectrum of Pr2NiO recorded under the same conditions drastically differs from the spectra of Pr2-xLaxNiO as four sharp maxima of titration current are detected. The TPD spectra of Pr2-xLaxNiO then strongly depend on the nature of the rare-earth (La or Pr) in link with the oxygen over-stoichiometry amount.

Introduction
Due to their high oxygen ion diffusivity and electronic conductivity at high temperature, the mixed ionic-electronic conducting (MIEC) lanthanide nickelates Ln2NiO4+ with K2NiF4-type structure can be used as solid oxide fuel cell (SOFC) cathodes, separation membranes or oxygen sensors (1 6).
Oxygen exchange with atmosphere, i.e. desorption-sorption processes, obviously depends on the crystallographic structure as well as on the oxygen non-stoichiometry of the oxides. Temperature programmed desorption (TPD) followed by solid electrolyte coulometry/potentiometry gives valuable data on the oxygen exchange processes: temperature at which oxygen releases (desorption), stoichiometry changes and structure transition occur (7). Interestingly it also allows estimating the nature of desorpted oxygen species with respect to their mobility (8).
To investigate oxygen TPD and ionic transport in MIEC oxides being induced by a chemical potential gradient, it is convenient to use the multifunctional solid electrolyte (7 13). It acts both as independent and supplementary technique to thermo-gravimetry, gaseous chromatography, X-ray and neutron diffraction in a wide range of temperature and oxygen partial pressure (8). A devices application for investigation of various compounds has been reported by Vashook et al. (8).
described in the work of Teske (7), where properties of perovskite-like cuprates Y Ba Cu O were investigated. For example, the time dependence of oxygen exchange gave useful informations about oxygen diffusion in the materials. Spectra of temperature programmed desorption and sorption (TPD) detected the oxygen variation with temperature. According to the authors, TPD spectra of Y Ba Cu O could be associated with series of structural and conductivity transitions.
Refs (9, 10) described the operating modes of the OSEC device that are required to build p T diagrams and to investigate the thermodynamic properties of mixed nonstoichiometric oxides. For instance, a tentative analysis of oxygen non-stoichiometry and electrical conductivity of the binary strontium cobalt oxide SrCoOx has been performed in ref. (11). The authors demonstrated strong evidence of correlation for maxima in TPD spectra (within the temperature range 500 and the relationship of a -disorder transition of the cubic high-temperature phase. Regarding nickelates with K2NiF4 type structure, oxygen non-stoichiometry combined with electrical conductivity measurements were performed on La SrxNiOy and Pr SrxNiOy (3,12,13). TPD spectra allowed concluding about the existence of weakly bonded oxygen, capable to reversibly exchange with the gas phase at temperature as low The present work is devoted to the investigation of oxygen TPD in Pr2-xLaxNiO4+ powders by coulometric solid electrolyte technique, i.e.
The room temperature oxygen over-stoichiometry of the synthesized oxides was determined either by iodometric titration (6) or by thermo-gravimetry analysis (TGA) measurements performed under Ar/5%H2 flow with a SETARAM setup (MTB 10-8 balance). Both values (table I) are in very good agreement. Oxygen desorption-sorption processes with atmosphere were studied on Pr2-xLaxNiO4+ 900 required value of oxygen partial pressure set by coulometry dosing was 50 Pa. Argon was used as a carrier gas and air was used as a comparison gas. The considered parameters were the time and temperature dependences of both the titration current (I2 in Fig. 1) and / or the oxygen over-stoichiometry. To simplify further discussions, the dependence will be called TPD spectra (7), the spectra data concerning not only desorption but also oxygen sorption. coulometric regime. The measurement system consists of two identical solid electrolyte cells, each one using a tube made of 8 mol. % Y2O3 stabilized ZrO2 (8YSZ). Each cell has a pair of oxygen pumping (Ii, pump zone) and potential measurement (Ui, gauge zone) electrodes. Porous Pt partly covering areas of inner and outer walls of YSZ tube ensures the electrical contacts and serves as ele trode components. The cells allow stabilizing a steady-state gas flow with a given oxygen concentration by presetting the electrode voltage (Ui) at a known temperature. The voltage is controlled by feedback adjustment of the coulometric titration current (Ii) between the oxygenpumping electrodes.
The experimental reactor and the cells are connected by stainless steel capillaries. If the reactor with a sample is placed between two such cells, the oxygen partial pressure can vary in a wide range (8). The temperature of the reactor can be selected independently of the cells temperature.
Any oxygen partial pressure p(O2) in the in-flowing gas (Ar) can be modified in cell 1 by coulometric dosing with O2 to reach the required value p(O2)'. Cell 2 keeps the required value of oxygen partial pressure p(O2)'' after the flowing gas has just left the reactor by pumping out or pumping in O2. If no oxygen exchange between the flowing gas and the sample takes place in the reactor, no change in oxygen partial pressure (p(O2)''= p(O2) ') is observed and titration current I2 keeps constant value. If oxygen desorption occurs, oxygen partial pressure rises in the gas flow of the reactor (p(O2) ''> p(O2) ') and cell 2 pumps out the O2 excess to the value p(O2 2 2 -out procedure is detected through the decrease of I2 (titration current) if U2 is kept constant. If the sample absorbs O2, oxygen partial pressure decreases in the gas flow of the reactor (p(O2) ''< p(O2) ') and cell 2 pumps in O2 till reaching the value p(O2)' = p(O2) ''. The pumping-in procedure is detected through the increase of I2 if U2 is kept constant.
All oxygen exchange in the reactor should be accompanied by a deviation of the coulometric titration current (I2,t) from the (I2, base) value, if U2 is kept constant. The oxygen exchanged mass m (O2)) may be calculated according to the law: (1) Knowing the mass and chemical composition of the investigated sample, the TPD spectra can be readily built.
Prior to the coulometric measurements, powders of Pr2-xLaxNiO4+ were flowed at following temperature cycle was then applied in order to build the TPD spectra: i) heating ii) isothermal plateau iii) cooling fterwards, each heating-cooling cycle performed for coulometric measurements was carried out in Ar atmosphere with p(O2) = 50 Pa at flow rate 48 52 ml/min. The duration times will be detailed in the following when necessary.
TGA measurements were used to compare oxygen over-stoichiometry defined with   Prior to further detailed consideration of the spectra, one should pay attention to the titration curves of the samples recorded under cooling conditions. At the beginning of the cooling, and from maximum temperature down to 290

Results and discussion
, an oxygen sorption is observed. The heated samples are oxygen deficient while they exhibit an oxygen overstoichiometry at room temperature before heating. Due to presence of oxygen in the carriergas, they start to absorb O2 when cooling the powder, which can be seen from values of the titration current which are higher than the base current. Oxygen sorption stops at 290 , resulting in a quick current dropping to the base current value. Such a thermal behavior is the same for all the samples and will not be more discussed below.

TPD of La2NiO4.16
Two main desorption peaks are visible on the TPD spectrum of La2NiO4.16, located around 240 is chosen to refer on it); one can conclude that under the considered experimental conditions, there are two regions of very intensive and continuous . The second maximum of the titration current cannot be accurately determined, because it is stopped when heating at 900 (12) also observed two desorption maxima in the TPD spectrum of La2NiO4.15 powder in argon flow with lower oxygen partial pressure (10 Pa) in the temperature cycle 20 1050 with those of ref. (12) leads to conclude that the real location of the second maximum takes place at a temperature highe The oxygen desorption occurs with an apparent constant rate during heating La2NiO4.16 in the range 400 i.e. between the two previous peaks.
The absorption of O2 during cooling represents a diffuse maximum, i.e. a broad peak in the TPD spectrum (Fig. 2) Table   II, one can see a difference between the initial and final over-stoichiometry of the powders. Because the material is prepared under air, but here cycled under lower pO2, the amount of desorbed oxygen while heating is larger than the amount of absorbed oxygen while cooling, as expected: for instance the initial oxygen over-stoichiometry of La2NiO is 0.16, the final one being 0.124. The influence of the oxygen content in the gas flow can be confirmed by the data reported in ref. (12) for the same material: after cycling under p(O2) = 10 Pa, the final oxygen over-stoichiometry is 0.06.  Fig. 2, two sharper maxima of oxygen desorption (compared to those of LNO) are visible in the TPD spectra of Pr2-xLaxNiO (x 0). More especially the second one becomes sharper and sharper when the Pr amount increases. In addition, at increasing Pr content, the start of the first oxygen desorption as well as the first maximum itself (which corresponds to the maximum speed of oxygen release) are shifted to lower temperatures (Table III) compared to La2NiO4. 16. The same conclusion on the temperature range of the second oxygen desorption maximum can be drawn. It should be also noticed that the released oxygen amount during heating increases with the increase of Pr content in Pr2-xLaxNiO . This is clearly visible from the value of the over-stoichiometry and after heating (see and in Table II) and the maxima values of oxygen desorption (Fig. 2). The amount of released oxygen depends on the initial overstoichiometry of the oxides, which in its turn depends on the Pr content, provided the preparation conditions of Pr2-xLaxNiO are the same. It is also well-known that the oxygen over--earth cation: the smaller the radius of the rare-earth cation, the higher the oxygen content is. For La, the value is reported to be 0.16 0 0.29 (14 19). In the Pr2-xLaxNiO increases, as expected. Very interestingly, the TPD spectrum of Pr2NiO4.26 drastically differs from the previous spectra (x < 2). Four sharp maxima (peaks) of oxygen desorption are observed at the (labeled 1, 2, 3, 4 in Fig. 3). Two additional prominent peaks (compared to the spectra of PLNO and LNO) are located at 370 (labeled 2 and 3, Fig. 3).
Besides the four mentioned maxima in the TPD spectrum of Pr2NiO4.26, our data (labeled 5 and 6, the existence and resolution were confirmed in experimental series with l TPD Spectra Comparison Our TPD spectrum is in rather good agreement with that reported for Pr2NiO by Sadykov et al. (20). Due to different oxygen partial pressure and inert atmosphere used in our experiments, the observed maxima logically shift to lower temperatures compared to ref. 20 According to the available data in literature (6,20), the peaks observed in the temperature range 210 located in Pr2O2 layers (NaCl-type structure) but without any link with the possible lowering of the mean praseodymium charge (which would be between 3+ and 4+), as supposed by Sadykov et al. (20). Indeed, the charge is only 3+ in Pr2NiO whatever the oxygen over-stoichiometry in air at T = 25 by XANES studies (21). Combination of experimental and modelling data concerning interstitialcy / interstitial diffusion mechanisms from (20, 22 25) allows the following suggestions: the peaks 1 3, 5, 6, depicted in Fig. 3 could be likely attributed to mobile i) and reflect different dynamic positions of Oi in crystal lattice under heating at T= ~ 400 , corresponding to the sharp narrow peak (2 in Fig. 3), is especially noteworthy. In our experiments, the TPD spectra constantly preserved this peak as compared to 1, 3 6 ones. This peak can be assigned to the orthorhombic-tetragonal phase transition (22, 25) in Pr2NiO . According to single crystal neutron diffraction investigations performed by M. Ceretti et al. (23) on Pr2NiO at 2O2 rock salt layer from an embedding matrix of NiO2 layers takes place in this temperature range. Thus, this peak induced by crystal lattice transformation can be found whatever the initial oxygen stoichiometry. In refs. (16,22) it has been confirmed that the initial non-stoichiometry value, typically exceeding 4.20, does not significantly affect the presence of this phase transition.
Th 4, Fig. 3) is supposed to result from oxygen removal from the direct environment of the nickel cations.

Comparison of TPD Spectra
Summarizing the above-mentioned features concerned with non-monotonous change of oxygen desorption when heating at constant rate, one can correlate the nature of released oxygen with titration current peaks (maxima) in TPD spectra: 1) phase transformation in a material, 2) un-equivalent crystallographic positions of oxygen atoms in the crystal lattice of the compounds. In the case of the phase transformations, an abrupt change in bonding energy of oxygen with the lattice takes place and, as a consequence, an anomaly in the oxygen desorption rate can occur. In the second case, with the increase of temperature, distinct consecutive removal and/or migration of oxygen from interstitial and regular positions can proceed in accordance with the increase in bonding energy of oxygen in the crystal lattice (12). The sharp narrow peak matching the phase transition is observed only for PNO and not detected for PLNO phases. A hypothesis is that the amount of oxygen removed during the phase transition for PLNO phases is significantly lower than for PNO (cf. Table 2 and Fig. 2, 4) due to their lower ionic mobility in the attendance of lanthanum cations. Figure 4. Oxygen non-stoichiometry value vs. temperature (converted TPD spectra) of Pr2-xLaxNiO4+ nickelates in argon flow (pO2 = 50 Pa) and heating rate 6 /min.
The second reason for the absence of the narrow peak in TPD spectra for Pr2-xLaxNiO4+ (x 0) could be the vicinity of temperature ranges for weakly-bounded oxygen desorption (first peak) and oxygen removal linked to the phase transition (25). Indeed, from Fig. 2 it can be seen that the location of the first peak shifts towards higher temperature with decreasing x (for lowest Pr contents, which again supposes a lower mobility of such oxygen in PLNO), and can then overlap the narrow peak of oxygen desorption occurring during the phase transition.
To confirm the existence of the TDP peaks and their resolution for Pr2NiO and La2NiO , experiments were conducted As shown in the TPD of PNO (Fig. 5) all oxygen desorption maxima were either present in the temperature range under consideration for intensity and shape slightly changed (Fig. 5). Prolonged thermal treatment (depending on the heating rate) results in both cases in the shift of the peaks location to lower temperatures, as well as in a change of the shape of the last maximum ( Fig.5(a)).
Most likely, the two sharp maxima in TPD for Pr2-xLaxNiO (x 0) might be associated to removal of oxygen from different crystallographic positions in these nickelates. Over-stoichiometric oxygen in La2NiO structure is located in interstitial position with the coordinates ( , , close to layers formed by NiO6 octahedra. These interstitial oxygen ions, apparently, are the most weakly bonded in the nickelate lattice (26). Our suggestion is that during the heating of the nickelates interstitial oxygen ions leave the compound lattice at lower temperatures and cause the occurrence of the first oxygen desorption peak in the range 240 stoichiometry content, a further departure of oxygen apparently occurs accounting for the removing of oxygen from its normal (regular) lattice sites. Regular oxygen causes the occurrence as shown in Fig. 2.
When the second oxygen desorption peak occurs in Pr2-xLaxNiO (x 0), the over-0.090 (Fig.  4, Fig. 6 (a) (c)). This over-stoichiometric content of oxygen is possibly so strongly linked to the nickelate lattice that it does not result in desorption of oxygen at heating the powders up to temperatures 650 1, Fig. 6). In this temperature range, oxygen is more easily removed from the regular lattice sites. (c) Figure 6. TPD spectra for Pr2-xLaxNiO4+ (x = 0.5, 1.0, 1.5) nickelates, including in the temperature cycle 20 900 in Ar flow (pO2 = 50 Pa) and heating-/min.

Conclusions
TPD spectra of Pr2-xLaxNiO (x 0) exhibit two sharp peaks (maxima) of titration current with excellent resolution at p(O2) = 50 Pa that are connected with removal of oxygen from different sites of the crystal lattice. The maxima observed during heating of Pr2-xLaxNiO powders at temperatures 200 predominantly result from weakly bonded over-stoichiometric oxygen, i.e. interstitial oxygen in the crystal lattice. The statement might be supported by the fact that, when increasing the Pr amount, whose cationic radius is smaller than that of La, the oxygen over-stoichiometry of the nickelates (synthesized in the same conditions) increases and the oxygen desorption shifts to lower temperature compared to La2NiO . TPD spectrum of Pr2NiO recorded under the same conditions drastically differs from the spectra of Pr2-xLaxNiO with lower Pr contents: three sharp and two local maxima of titration current are located in the range of temperature 210 prominent first maximum in a relatively low temperature range is supposed to have the same nature as for PLNO and might be associated with the removal of interstitial sites with the coordinates , close to layers formed by NiO6 octahedra. The second (narrow) maximum of the TPD at 370 is most probably attributed to the orthorhombic-tetragonal phase transition in Pr2NiO . The peaks observed in the temperature range 400 65 different (nonequilibrium) migration positions of mobile oxygen from crystal lattice under heating.
The maxima of TPD spectra observed at temperature above 650 considered nickelate are apparently connected with the removal of oxygen from NiO6 octahedra building perovskite layers. It has been shown that the removal of oxygen from perovskite layers occurs when the lattice maintains some amount of oxygen over-0.09 for Pr2-xLaxNiO 0.14 for Pr2NiO ).