Temperature Dependence Discontinuity in the Stability of Manganese

Subscriber access provided by Fudan University

Article

Temperature Dependence Discontinuity in the Stability of Manganese doped Ceria Nanocrystals Longjia Wu, Pratik P Dholabhai, Blas P. Uberuaga, and Ricardo H. R. Castro Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.6b01193 • Publication Date (Web): 05 Jan 2017 Downloaded from http://pubs.acs.org on January 17, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Crystal Growth & Design is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

Temperature Dependence Discontinuity in the Stability of Manganese doped Ceria Nanocrystals Longjia Wu1, Pratik P. Dholabhai2, Blas P. Uberuaga2 and Ricardo H. R. Castro1 1

Department of Materials Science & Engineering and NEAT ORU, University of California,

Davis, CA 95616; 2Materials Science and Technology Division, Los Alamos National Laboratory, Los Alamos, NM 87545 Abstract: CeO2 has strong potential for chemical-looping water splitting. It has been shown that manganese doping decreases interface energies of CeO2, allowing increased stability of high surface areas in this oxygen carrier oxide. The phenomenon is related to the segregation of Mn3+ at interfaces, which causes a measurable decrease in excess energy. In the present work, it is shown that, despite the stability of nanocrystals of manganese doped CeO2 with relation to undoped CeO2, the effect is strongly dependent on the oxidation state of manganese, i.e. on the temperature. At temperatures below 800 °C, Mn is in the 3+ valence state, and coarsening is hindered by the reduced interface energetics, showing smaller crystal sizes with increasing Mn content. At temperatures above 800 °C, Mn is reduced to its 2+ valence state, and coarsening is enhanced with increasing Mn content. Atomistic simulations show the segregation of Mn to grain boundaries is relatively insensitive to the charge state of the dopant. However, point defect modeling finds that the reduced state causes a decrease in cation vacancy concentration and an increase in cation interstitials, reducing drag forces for grain boundary mobility and increasing growth rates.

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

Introduction Cerium oxide is an important material for a variety of applications due to its characteristic catalytic and sensorial uses

1-3

. Since these applications exploit surface properties, inhibiting

coarsening is of critical importance to extend life cycles and maintain performance. Building on a parallel strategy of using surfactants to reduce droplet sizes in liquids, the usage of ionic dopants prone to segregating to interfaces (here defined as surface and grain boundaries) of nanocrystalline systems has been proposed to decrease particle sizes via a reduction in interfacial energies, with good success

4-7

. For instance, Mn3+ doping was observed to decrease the

crystallite size of CeO2 from 10.8 nm down to 7.3 nm when samples were synthesized and annealed at the same temperature 8. The stabilization was attributed to a concomitant effect of Mn3+ in reducing the grain boundary energy and the surface energy of the particles 9, which results from the dopant segregating to those locations. That is, segregation leads to decreased interfacial energies, which is responsible for a decrease in the driving force for coarsening, resulting in smaller crystallite sizes. From a thermodynamics perspective, segregation is linked to interfacial energy by the Gibbs adsorption isotherm expression for two components in a dilute system 10,11: dγ = −RTΓଶ,ଵ dln‫ݔ‬ଶ

(1)

Here, subscripts “1” and “2” represent solvent and solute, respectively;γis the interface energy; T is temperature; and R is the gas constant. Γ2,1 is the excess concentration of dopant at the interface, and x2 is the molar fraction of the dopant soluble in the particle (not segregated). This relation is valid for both surfaces and grain boundaries in a competitive way, such that when a dopant is introduced in a system, some fraction of the dopant will migrate to each type of interface. This fraction is governed by the energy of segregation, ∆Hseg, which is defined as the


Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


difference in energy when the system has the dopant in solid solution in the bulk as compared to the dopant segregated to either interface. Typically, every variant of these interfaces is characterized by a different value of ∆Hseg; however, here, we will consider ∆Hseg to be an average over all variants for a given type of interface (grain boundaries versus surfaces) to understand the overall impact on crystal stability. Krill et al. have derived a useful expression to show the relationship between the enthalpy of surface segregation and the interface energy change due to solute segregation 12: ߛ = ߛ଴ + Γ∆‫ܪ‬௦௘௚

(2)

Here, Γ is the excess concentration of dopant at the given interface. This equation has been used to estimate the change in interfacial energies for a given dopant in a few systems, and has been the basis for the design of grain growth resistant materials in which a segregating dopant is chosen to cause a quasi-zero grain boundary energy 13. The concomitant role of both interfaces (grain boundary and surface) in the energetics of nanocrystas is critical to interpret coarsening behavior (the term nanocrystal refers here to a single grain that can alone be forming a nanoparticle or be combined with others in a polycrystalline nanoparticle). Although most works focus on surfaces, a prime consideration for describing coalescence, it has been shown that grain boundaries can create highly stable metastable configurations, leading to stable nanocrystals. That is, high surface energies are replaced by lower energy grain boundaries, dramatically decreasing the driving force for growth, and creating virtually stable nanocrystals 14. In a previous work, the energetic effects of Mn on both surfaces and grain boundaries have been reported by using a combination of water adsorption microcalorimetry and oxide melt solution calorimetry 9. These techniques offer state-of-the-art assessments of the interfacial



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

energies that can be later used to interpret stability trends

Page 4 of 33

8,15,16

. Reduction of both interface

energies with increasing Mn content was associated with a clear segregation profile. The grain boundaries had significantly larger concentrations of Mn as compared to surfaces, resulting in a more pronounced decrease in energy. Interestingly, the enthalpies of segregation for both interfaces were similar, and the difference in segregated quantities was primarily responsible for the observed difference in energetics 9. In this work, we evaluate the effect of the lowered energy in the overall coarsening behavior as a function of time and temperature. Since Mn exhibits a variable oxidation state that is sensitive to temperature

17-19

, the effects of defect charge compensation on diffusion were

evaluated. We observe that below 700°C, grain size is inversely proportional to the interface energy (smaller grain sizes for higher Mn contents, i.e. lower interfacial energies), showing a critical role of thermodynamics in the interface evolution, i.e. nanocrystalline stability. On the other hand, for higher temperatures (above 800 °C), grains grow and deviate from the trend established by the room temperature energetics. Reduction of manganese was observed (from Mn3+ to Mn2+) at these temperatures. By using atomistic and point defect modeling, it is shown that the change in the oxidation state of Mn changes the relative concentrations of cation defects, causing a reduction in drag forces for grain boundary mobility and increasing growth rates. At this point, even though the interface energies somehow thermodynamically stabilize the system, the massive mobility allows for interface elimination despite the low driving force.

Experimental Procedure (1) Sample Synthesis and Characterization Manganese-containing and pure cerium dioxide were synthesized by co-precipitation


Page 5 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


method. Cerium nitrate hexahydrate [Ce(NO3)3•6H2O, 99.99%, Alfa Aesar] and manganese carbonate [MnCO3, 99.9%, Alfa Aesar] were used as precursors, and an ammonia solution with 1.5M was used as the precipitant. Details on the synthesis process can be found in a previous work 8. After precipitation samples were all calcined at 600 °C for 8 h under oxygen flow. Coarsening was studied in those nanocrystals by annealing them at 700, 800 or 900 °C for different times (ranging from 30 min to 4h). A Bruker D8 Advance diffractometer, operating at 40 KV and 40 mA with Cu Kα radiation (λ=1.5406Å) and a spinning sample holder, was used to collect the X-ray powder diffraction (XRD) patterns. Data were acquired over a 2θ range of 20- 90°, with a 0.017° step size and 0.7s dwell time. WPF (Whole Pattern Fitting) refinement was performed using JADE software (version 6.11, 2002, Materials Data Inc., Livermore, CA), and used to calculate crystallite sizes (grain sizes) from diffraction peak broadening. Error bars from the grain sizes were obtained from the whole-profile fitting accuracy. The surface areas of the nanoparticles were measured using Micromeritics ASAP 2020 instrument. All samples were degassed under vacuum at 400 °C for 12 h and reoxidized under oxygen (P=700 mmHg) at 400°C for 12 h. Five-point adsorption isotherms of nitrogen were acquired at the relative pressure range from 0.05 to 0.30 at -196°C. Each sample was measured three times to obtain average values, which are reported. Thermal analysis was performed using a DSC/TG SETSYS 1600 system (Setaram Inc., France) to record thermogravimetry (TG) and differential scanning calorimetry (DSC) simultaneously. Around 40 mg of each sample was heated from room temperature to 1300 °C at a heating rate of 10°C/min in a 20 mL/min air flow. The baseline correction was made by recording a TG/DSC run with empty crucibles under the same conditions.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 33

As a result of coarsening, partial agglomeration (sintering) is inevitable given the high reactivity of nanocrystals. This is a critical parameter that needs quantification for reliable calorimetric assessments. While surface areas and crystallite sizes can be directly measured as described above, grain boundary area quantification requires geometrical assumptions. In order to calculate grain boundary areas (AGB), the following equations were used: 20,21 ଺଴଴଴

‫ܣ‬ூ = ఘ∙ீௌ

(3)

೉ೃವ

‫ீܣ‬஻ =

஺಺ ି஺ೄ

(4)

ଶ

Here, AI is the overall interface area (assuming all particles are isolated), ρ is the theoretical density of the sample, GSXRD is the grain size refined from the x-ray diffraction peaks (also from TEM) and AS is the surface area measured by the BET method. Equation (3) has the assumption that all the nanoparticles are spherical and have a narrow size distribution (a reasonable assumption for ceria powders prepared by coprecipitation according to TEM analyses8). For the unlikely case of nanoparticles composed by a single nanocrystal and with negligible agglomeration, AI and AS should be equal to each other. The division by 2 in Equation (4) is because the grain boundary is formed by two surfaces during agglomeration.

(2) Atomistic and Point Defect Simulations Our simulations involved three distinct aspects. We first examine the interaction of Mn2+ and Mn3+ dopants with three high symmetry grain boundaries. The methodology used here is the same as that reported before 9. To briefly summarize, using grain boundary structures developed previously

22,23

, Ce cations were, one-by-one, replaced by Mn cations, either in the 2+ or 3+

charge state (potentials were taken from references

24-26

and are reproduced in Table 1) and the

energy of the system minimized using LAMMPS 27, keeping the volume constant. This produced


Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


segregation energy profiles for each microstructural feature considered, from which the segregation energy was extracted. In the calculation of charged systems (CeO2 with an aliovalent dopant), we added a background charge-compensating charge. However, we note that, for the relative energies that are the focus of this work, the contribution from this background charge identically cancels. Finally, there was only one dopant introduced in each calculation. We also examined dopant-defect interactions. In this case, a bulk simulation cell of CeO2 containing either a Ce interstitial or a Ce vacancy was constructed. Then, as with the grain boundary calculations, we substituted one-by-one every Ce ion with the dopant and minimized the energy. At the temperatures considered in this study (around 800 °C), the fact that doped CeO2 exhibits high oxygen conductivity suggests that the defects responsible for conductivity – oxygen vacancies – are not bound to dopants and that any such complexes are dissociated

28

. Thus, in

this work, we have only considered purely substitutional dopant geometries and did not consider charge-compensating defects. At lower temperatures, we would expect dopant-defect complexes to be bound, altering the segregation behavior to grain boundaries 23. Finally, we adapted the point defect model recently presented by Beschnitt et al. for cation diffusion in CeO2 as a function of dopant concentrations 29. Their model predicts the equilibrium defect concentrations for doped CeO2 assuming that the mechanisms responsible for defects are Frenkel, anti-Frenkel, and Schottky disorder reactions, oxygen reduction reactions, and the impact of aliovalent doping. We essentially used exactly the same model, with the only difference being that we considered two different charge states for the dopant. This is reflected by Eq. 10 in Ref. [30], the charge neutrality condition, which originally was '''' [Acc'Ce ]+ 2[O''i ]+ 4[VCe ]+[Ce'Ce ] = 4[Ce....i ]+ 2[VO.. ]



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 33

(5) for a 3+ dopant. For a 2+ dopant, this becomes

2[Acc''Ce ]+ 2[O''i ]+ 4[VCe'''' ]+[Ce'Ce ] = 4[Ce....i ]+ 2[VO.. ]

(6)

Otherwise, the model is exactly as outlined in Ref. [30]. All thermodynamic parameters for the model – formation enthalpies and entropies – are as presented in Ref. [30]. Further, this model makes the same assumption used in our atomistic calculations: dopants and defects do not cluster. Again, this assumption is only valid at the higher temperatures of the experiments.

Results and Discussion Experimental Coarsening of Mn-CeO2 Figure 1 shows the X-ray diffraction patterns of pure ceria and Mn doped ceria samples calcined at 600 °C (as synthesized) and further annealed at 700, 800 and 900 °C for 4h (longest annealing time). Only the fluorite structure was observed in the patterns, regardless of the temperature, time, and dopant concentration, with the absence of any second phases. In general, there is a clear sharpening of the reflection peaks with increasing temperature of calcination. This is also true for increasing time at a particular annealing temperature, as demonstrated in the grain sizes refined from the diffraction peaks (GSXRD) and plotted in Figures 2a and 3, for samples calcined at 700, 800 and 900 °C. Figure 2a reveals an increase of the grain size for all of the samples with increasing calcination time, as expected for any thermally activated process, leveling out for longer annealing times. Furthermore, after four hours, although the grain sizes for all of the Mn doped ceria samples are very close to each other considering the error bars in the grain size results, there is a non-negligible reduction of the grain size from pure ceria to Mn doped ceria samples,


Page 9 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


representing an increase in the overall stability of nanocrystals caused by the dopant. The dependence of the microstructural evolution on dopant content is very consistent with the energetics of the interfaces found in the nanocrystals. Table 2 shows surface and grain boundary energies experimentally measured in this system in our previous work 9. With increasing Mn content, there is a decrease in surface energy from 1.08 J.m-2 to 0.95 J.m-2, while the grain boundary energy decreases even more remarkably from 0.87 J.m-2 to 0.30 J.m-2 with the incremental addition of manganese. From the viewpoint of interface area evolution induced during calcination, in contrast to grain size evolution, one can better analyze the energetic effects in the microstructure development. The data in Figure 2b show both surface and grain boundary area evolution in the samples during calcination at 700 °C. A very distinct difference between the surface areas of the doped and undoped samples is observed. While all surface areas decrease with annealing time, the doped samples are shifted to significantly lower values. While this may sound like a contradiction to the thermodynamics of the system, since the surface energy of the doped samples is systematically smaller, one may note that the total interface energetics must include the grain boundary term. The grain boundary areas for the doped samples are much higher than that for the pure samples for all compositions, a fact consistent with the lower energies of those interfaces with respect to the surface energy term that causes an overall smaller energy for the doped nanoparticles. That is, while the energies of both types of interfaces is reduced by doping, that of grain boundaries is reduced by a much greater extent, and thus the system can further minimize its energy by increasing the area associated with grain boundaries, at the expense of surfaces. Interestingly, Figure 2b shows that while the surface area decreases for all samples with time, the grain boundary area slightly increases and levels for all compositions (more for the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 33

highly doped samples). For the doped samples, this leads to a crossover between majority surface area to majority grain boundary area, such that after annealing they have comparatively high amounts of grain boundaries. Figure 3 shows the grain size of pure ceria and the Mn doped ceria samples annealed at 800 °C and 900 °C, respectively, for different times. For the samples annealed at 800 °C, the results show almost overlapping trends, but a close inspection reveals that when the calcination time is no longer than 30 min, the Mn doped ceria samples show a slightly smaller grain size as compared to pure ceria. For annealing times larger than 1h, the grain sizes of the doped ceria samples surpass those of pure ceria, particularly for 5 and 10 mol% Mn doped ceria. For the samples calcined at 900°C, doped samples have systematically larger grain sizes regardless of the annealing times, but no particular trends are observed within the doped ones. Figure 4 shows a comparative plot of both surface and grain boundary area for the sample doped with 10 mol% Mn and the undoped CeO2 to highlight the behavior change as a function of temperature. While Mn is stabilizing grain boundaries at 700 °C with respect to CeO2, allowing smaller crystallite sizes (Figure 2a), the dopant causes both surface and grain boundary instabilities at higher temperatures. This reverse in coarsening behavior at high temperatures can be potentially related to a change in the Mn oxidation state. Figure 5 shows thermogravimetric analysis of 10 mol% Mn doped ceria, as well as pure ceria nanoparticles for reference. From the curves, one can see that there is a significant mass loss before 600 °C for both of the samples, which is due to the evolution of physically and chemically adsorbed water

21

. For pure ceria,

there is no obvious mass loss after 600°C, corresponding well to other literature 30 reporting ceria reduction occurs only above 1200 °C in air. On the other hand, for 10 mol% Mn doped ceria, mass loss starting from 820°C and ending at 1000°C is observed. Previous studies have


Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


attributed this phenomenon to a change in the oxidation state from Mn3+ to Mn2+ in this temperature range 17,19. The relationship between the manganese oxidation state and observed coarsening enhancement at high temperatures can be hypothetically attributed to: (1) the change in valence of manganese is causing a change in the segregation tendency. That is, the interface energy decrease shown in Table 2 for the doped samples with respect to undoped CeO2 has been related to the segregation of the dopant, as described by equation 2. Since one of the effects of manganese reduction is the decreased radius mismatch between Mn ions and Ce ions (Mn3+ radius is 0.58 Å, Mn2+ is 0.83 Å, Ce4+ is 1.01 Å), reduction may cause a different segregation tendency and should be investigated in terms of the consequences to the interface energies of the system. Unfortunately, it is extremely difficult to measure the grain boundary energy and surface energy of the Mn doped ceria samples annealed at high temperatures due to the very small interface areas (see Figure 4). Calorimetric signals for interface energies are directly dependent on interface areas, and reliable values cannot be attained once the areas are

Temperature Dependence Discontinuity in the Stability of Manganese

Recommend Documents