Journal of Thermal Spray Technology

, Volume 18, Issue 4, pp 694–701

Partial Evaporation of Strontium Zirconate During Atmospheric Plasma Spraying

Authors

    • Forschungszentrum Jülich GmbH, IEF-1
    • State Key Laboratory of Rare Earth Resources Utilization, Changchun Institute of Applied ChemistryChinese Academy of Sciences
    • Graduate School of Chinese Academy of Sciences
  • Daniel Emil Mack
    • Forschungszentrum Jülich GmbH, IEF-1
  • Maria Ophelia Jarligo
    • Forschungszentrum Jülich GmbH, IEF-1
  • Xueqiang Cao
    • State Key Laboratory of Rare Earth Resources Utilization, Changchun Institute of Applied ChemistryChinese Academy of Sciences
  • Robert Vaßen
    • Forschungszentrum Jülich GmbH, IEF-1
  • Detlev Stöver
    • Forschungszentrum Jülich GmbH, IEF-1
Peer Reviewed

DOI: 10.1007/s11666-009-9360-z

Cite this article as:
Zhang, Y., Mack, D.E., Jarligo, M.O. et al. J Therm Spray Tech (2009) 18: 694. doi:10.1007/s11666-009-9360-z

Abstract

Perovskite-type SrZrO3 has been investigated as a candidate material for thermal barrier coating application. During plasma spraying of SrZrO3, SrO volatilized more than ZrO2 and the coating composition deviates from initial stoichiometry. In this investigation, partial evaporation was investigated by spraying SrZrO3 powders into water. The influences of spraying current, distance and particle size of the powder on the partial evaporation were also investigated in a quantitative way. With optimized spraying parameters, stoichiometric SrZrO3 coating was produced by adding an excess amount of Sr in the precursors before plasma spraying to compensate for the volatilized component.

Keywords

partial evaporationplasma sprayingstrontium zirconatethermal barrier coating

Introduction

The use of thermal barrier coatings (TBCs) allows the increase of the turbine inlet temperature, hence an increase of the efficiency of turbine engines (Ref 1, 2). Plasma-sprayed TBCs typically consist of an oxidation-resistant metallic MCrAlY (M = Ni and/or Co) bond coat and a ceramic top coat. Present top coat materials are based on 7-8 wt.% Y2O3-stabilized ZrO2. However, at higher temperature, volume changes induced by phase transformation and sintering give rise to cracks, leading to coating failure. To overcome this shortcoming, the search for new TBCs materials has been intensified. Perovskite-type SrZrO3 has been investigated as candidate TBCs material. Details of SrZrO3 are published in literature (Ref 1, 2), and Table 1 summarizes its thermophysical properties.
Table 1

Thermophysical properties of SrZrO3 (Ref 1-3)

Crystal system

Orthorhombic

Lattice parameters, nm

 

    a

0.5816

    b

0.8225

    c

0.5813

Thermal expansion coefficients, K−1

 

    α

10.9 × 10−6

Melting point, K

 

    Tm

2883

Shear modulus, GPa

 

    G

98.5

Young’s modulus

 

    E

269

Debye temperature, K

 

    θD

591

Vickers hardness, GPa

 

    Hv

5.74

Thermal conductivity at 1000 °C, W·(m·K)−1

 

    λ

2.08

‘‘Steady state’’ sintering rate at 1200 °C, s−1

1.04 × 10−9

Besides electron beam physical vapor deposition, usually TBCs are deposited by atmospheric plasma spraying with typically very high plasma flame temperature. The rapid heating of the particles gives rise to melting of the particles, sublimation, vaporization and evaporation across a liquid-vapor interface. Vaporization is driven by vapor concentration gradients existing between the free stream and the particle surface at temperatures below the boiling point while evaporation happens as the surface temperature reaches the boiling point (Ref 4). Physical vaporization and evaporation of the powders has no component change, which is controlled by vapor diffusion and heat transfer (Ref 5). In addition, plasma spraying can be accompanied by selective evaporation of a component of multi-component powder (Ref 6). As the material is heated in the plasma flame, substantial different evaporation rates of the constituents will lead to stoichiometry changes (Ref 7). For example, partial evaporation of CuO from YBa2Cu3Ox (Ref 6), P2O5 from Ca5(PO4)3OH (Ref 8), La2O3 from La2Zr2O7 (Ref 7) and CeO2 from La2Ce2O7 (Ref 9) have been reported.

During plasma spraying of SrZrO3, SrO was found to volatilize faster than ZrO2 with the result that the coating composition deviated from stoichiometric composition (Ref 2). The existence of excess ZrO2 in the plasma-sprayed SrZrO3 coating might influence the coating properties. To deposit a stoichiometric SrZrO3 coating, an excess amount of Sr must be added in the precursors before plasma spraying to compensate the loss of SrO by partial evaporation.

Although partial evaporation during plasma spraying of multi-oxides is common, detailed investigation of this phenomenon is very limited. Accurate analysis of evaporation mechanisms involved during plasma spraying of multi-oxides is necessary. This paper presents the quantitative investigation of the partial evaporation of SrZrO3 by spraying powders into water.

Experimental

Plasma Spraying

Commercial SrCO3 (Aldrich, >98%, Munich, Germany) and ZrO2 (Aldrich, 99%, Munich, Germany) powders were mixed by ball-milling with a Sr/Zr atomic ratio of 1.136 and calcined at 1400 °C for 24 h with a heating rate of 5 K/min. This means more than 10% of Sr was added in the precursors. The powders were spray-dried in a spray-dryer (Mobile MinorTM‘2000’ Ex Model H) with N2 as the drying medium, then heat treated at 1300 °C for 5 h to remove the organic binding compounds.

All plasma-sprayed coatings have been produced with a three cathode gun Triplex II in a Multicoat facility produced by Sulzer Metco, Wohlen, Switzerland. The gun was mounted on a 6-axis robot, which allowed well-defined coating sequences. Sand blasted austenitic steel was used as substrate for the coatings. The dimension of the standard steel substrates was a square with a side length of 50 mm and a thickness of 2 mm. Substrate temperatures were measured using the 4 M8 pyrometer system (Land instruments GmbH, Germany). These coatings were sprayed with argon and helium as plasma gases with flow rates of 45 and 6 standard liters per minute (slpm), respectively. The feeding gas was 2 slpm of Ar. The Robert speed was 500 mm/s. Particle speed and temperature monitoring were performed by an Accura spray-g3 diagnostic system (Tecnar, St. Bruno, Canada).

With a Triplex I gun (Sulzer Metco, Wohlen, Switzerland), some SrZrO3 powders were also sprayed into water to investigate the evaporation of oxide components. The schematic illustration is shown in Fig. 1. The argon and helium plasma gas flow rates during spraying were 20 and 13 slpm, respectively. The feeding gas was 1.5 slpm of Ar. The spraying distance from the exit of the torch and the water was 320 mm and the spraying angle was 45°. A 4L boiler filled with ordinary water was used to collect plasma-sprayed powders. After plasma spraying, the solution was evaporated until totally dry. The collected powders were characterized.
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig1_HTML.gif
Fig. 1

Schematic illustration of plasma spraying SrZrO3 into water

Characterization

The phase composition of the powders was analyzed by x-ray diffraction (XRD) with a D500 diffractometer (Siemens AG, Germany) equipped with diffracted-beam monochromator for Cu-Kα. Particle size distribution was measured by Fraunhofer diffraction with 633 nm laser light using the Analysette 22 (Fritsch GmbH, Idar Oberstein, Germany). Chemical analysis of samples was performed using optical emission spectroscopy (OES 4.1, TJAIRIS-INTREPID) after evaporation in inductively coupled plasma (ICP). After plasma treatment, the SrZrO3 powders were vacuum impregnated with epoxy. After hardening of the resin, the sample was ground mechanically with SiC abrasive paper and polished on soft disks with diamond suspension. A cross section of the sample was investigated by a scanning electron microscope (SEM) (ModelJXA 840, JEOL, Tokyo, Japan) with an energy dispersive x-ray spectrometer (EDX) (Model Inca, Oxford Instruments, Oxfordshire, UK).

Results and Discussion

Composition of the Synthesized Powders

An excess amount of SrO was added in before calcination and Sr/Zr ratio of this powder is 1.136. The XRD result of the synthesized powders is shown in Fig. 2. The main phase composition was found to be SrZrO3 (PDF#01-070-0283, Orthorhombic) and the secondary phase was Sr4Zr3O10 (JCPDS#23-1416, Orthorhombic). This result is in accordance with ZrO2-SrO phase diagram (Fig. 3). No other phases were detected.
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig2_HTML.gif
Fig. 2

XRD of SrZrO3 powders after calcination (1400 °C for 24 h) and before spray-drying

https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig3_HTML.gif
Fig. 3

ZrO2-SrO phase diagram (Ref 10)

Powders Plasma-Sprayed into Water

After spray-drying and before plasma spraying, powders were sintered at 1300 °C for 5 h. Furnace heat treatment of the spray-dried powders has many advantages. Heat treatment will remove the organic binding compounds, make the particles dense and increase the mechanical stability of the powders. Figure 4(a) indicates the morphology of spray-dried powders after being sintered at 1300 °C for 5 h. Most of the spray-dried powders are spherical and some are hollow. During spray-drying the large-sized pores inside the particles result mainly from the evaporation of liquid and the small-sized pores result from the incomplete sintering of the submicron precursors. The surface of the particles is very porous (Fig. 4b). Such a way of obtaining hollow particles can be useful to produce powders resulting in coatings of controlled properties, such as thermal conductivity or elastic modulus (Ref 11). Compared with irregular shape powders, spherical particles have better flowability. During thermal spraying, these spherical and free-flowing powders can enter the center of the plasma flame easily.
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig4_HTML.gif
Fig. 4

(a) Spray-dried powders after sintering at 1300 °C; (b) surface of the powders after sintering at 1300 °C; (c) powders plasma-sprayed into water; (d) surface of the powders sprayed into water; (e) cross section of the powders sprayed into water; (f) spherical particles and newly formed needle-like structure; (g) EDX of spherical particles; and (h) EDX of newly formed structure

Although the processes of plasma spraying of coatings and plasma treatment of powders are very similar, their final products are totally different. The plasma treatment of powders is usually carried out to improve some of their properties, such as flowability, density, internal or external morphology, etc. (Ref 11). As shown in Fig. 4(c), most of the spherical structures were preserved after plasma treatment. However, some new needle-like structures were formed after plasma treatment. As shown in Fig. 4(d), the surface of the powders became much denser after plasma treatment. Cross section of the powders (Fig. 4e) indicates that most of the spray-dried powders were densified in the plasma flame. Some spheres have a porous center with a dense outer shell. Big holes in the center of powders can be found after being plasma-sprayed into water. As the SrZrO3 powders were prepared by the spray-drying method, particles were initially spherical but very porous. During plasma treatment a large temperature gradient exists between the surface and the core of a porous particle. SrZrO3 has low thermal conductivity and it is possible to have the surface melting, evaporating and the core still remain porous (Ref 12). Therefore, heating, melting and evaporation may proceed at the same time in such a particle (Ref 13).

The particle size distributions of the spray-dried powders before and after being plasma-sprayed into water are shown in Fig. 5(a) and (b), respectively. Both frequency histograms and cumulative frequency curves are indicated with logarithmic abscissa. The maximum frequency decreased from 13 to 8% after plasma-sprayed into water. The histogram also became wider and the proportion of lower size particles became higher. These prove distribution of plasma-treated powder size is shifted in the direction of smaller sizes. After plasma processing, usually the distribution of powder sizes shifts in the direction of smaller sizes (Ref 8, 11). The particle size decrease of the powders during plasma treatment is caused by densification and shrinkage because powders are molten totally or partially in the plasma flame. The smallest particles might be blown away or totally evaporated in the plasma flame and only larger particles were preserved.
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig5_HTML.gif
Fig. 5

Particle size distributions of spray-dried powders before (a) and after (b) being plasma-sprayed into water

As shown in Fig. 4(g), after plasma treatment, the composition of the spherical powders is still SrZrO3. However, main composition of the newly formed needle-like structure is Sr and no Zr can be detected in the new structure (Fig. 4h). In spectrum 2 (Fig. 4h), the Ca, Mg, Na, S and Cl are probably from ordinary water used for collecting powders during plasma spraying.

Chemical compositions of the SrZrO3 powders before and after plasma treatment are listed in Table 2. Chemical analysis indicates that before plasma treatment, the Sr/Zr atomic ratio of the powders was 1.136. However, after plasma treatment, the ratio changed into 1.049, which means more Sr than Zr evaporated during plasma treatment of the powders.
Table 2

Chemical compositions of the SrZrO3 powders before and after plasma treatment

Powders

Sr, wt.%

Zr, wt.%

Sr/Zr, atomic ratio

Before plasma treatment

40.7

37.3

1.136

After plasma treatment

39.0

38.7

1.049

During plasma treatment of the powders, the following stages take place: heating of the solid phase, melting of the solid phase, heating of the liquid phase, vaporization, partial or total evaporation and even re-solidification depending on the surrounding temperature and rate of heat transfer (Ref 5, 14). Ceramic particles, especially the ones with low thermal conductivities, start to evaporate after a few millimeters of trajectory, and for ceramics this vaporization starts long before the particle core has begun to melt (Ref 15). During plasma spraying, the temperature of the plasma flame is much higher than the melting point of SrZrO3 and evaporation is inevitable. SrZrO3 decomposes easily at the melting temperature. It can be anticipated that both SrO and ZrO2 were evaporated during the plasma spray process, the evaporation rate of strontium oxide and zirconia might be quite different by virtue of the differences in their thermophysical properties (like melting point). Table 3 shows a comparison of some thermophysical properties of SrO and ZrO2, which indicate that the latter is more thermally stable than the former. The evaporation rate is driven by the intrinsic vapor pressure of the material and limited by diffusion of vapor through the boundary layer surrounding the particle surface (Ref 5). Figure 6 indicates vapor pressure of SrO and ZrO2 at different temperatures. SrO has a higher vapor pressure than ZrO2, which explains the faster loss of SrO compared to ZrO2 during plasma spraying. More evaporation of SrO leads to stoichiometry changes. Partial evaporation of SrO also happened when SrZrO3 was synthesized by floating zone technique with radiation heating (Ref 18). An excess of SrCO3 was also added to the feed rod material to compensate the evaporation losses (Ref 18).
Table 3

Physical properties of SrO and ZrO2 (Ref 16, 17)

Properties

SrO

ZrO2

Melting temperature, K

2804

2963 (β-ZrO2)

Boiling temperature, K

~3273

4573

Heat of sublimation ΔH × 10−6, J/kg·mole

530.635 ± 12.142

741.064 ± 25.121

Crystal lattice energy, 10−6 J/kg·mole

3311.759

11195.503

https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig6_HTML.gif
Fig. 6

Oxide vapor pressure of SrO and ZrO2 (Ref 17)

The XRD result of spray-dried SrZrO3 powders after plasma treatment is shown in Fig. 7. The main phase composition is still SrZrO3. Traces of SrCO3 (PDF#01-071-2393, Orthorhombic) can be detected.
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig7_HTML.gif
Fig. 7

XRD of spray-dried SrZrO3 powders after plasma treatment

Influence of Plasma Spraying Current, Distance and Particle Size on Sr Evaporation

It is well known that spraying parameters can have an important effect on the composition of as-sprayed coatings. Partial evaporation of Sr was found during plasma spraying of SrZrO3 coating. In order to investigate the influence of spraying parameters on the Sr evaporation, parameters including spraying current, distance and particle size were varied. The result is shown in Table 4.
Table 4

Influences of plasma spraying current, distance and particle size on Sr evaporation

Coating no.

Particle size, μm

Current, A

Spraying distance, mm

Sr/Zr decreased, %

A

45-100

530

200

32.3

B

45-100

440

200

20.2

C

45-100

440

100

19.8

D

45-100

370

200

6.2

E

>100

440

100

13.3

F

>100

370

200

3.8

Speed and temperature monitoring of the particles were performed by the particle diagnostic system Accura Spray. Accura Spray is a sensor specifically designed for industrial online monitoring of thermal spray processes. It allows control of coating properties by providing real-time measurements of the average particle temperature, velocity and flow rate as well as the vertical position and profile of the plume itself (Ref 19). Particle temperatures are determined by two-color pyrometry and particle velocities are obtained from cross-correlation of signals which are recorded at two closely spaced locations (Ref 20). Figure 8 shows speed and temperature of the SrZrO3 particles measured with Accura Spray (plasma current I = 530 A, spraying distance d = 100 mm). The measurement time was 80 s. For a current of 530 A, the mean particle speed and particle temperature were found to be 283 m/s and 2251 °C, respectively. However, when the current was 370 A, the mean particle speed and particle temperature were reduced to 192 m/s and 1667 °C, respectively. This proves that an increase of spraying current will lead to higher particle temperature and velocity. The heat transfer from the jet to the particles is sufficient to melt more powders and as a result a higher amount of SrO is volatilized. For coatings D, B and A, higher spraying current leads to more Sr evaporation.
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig8_HTML.gif
Fig. 8

Speed and temperature of the SrZrO3 particles measured with Accura Spray (plasma current I = 530 A, spraying distance d = 100 mm)

The particle size of the powders used for plasma spraying also has an influence on the final Sr/Zr ratio of the coating. The evaporation of SrO can be reduced effectively by increasing the particle size of the powder used for plasma spraying. Comparing with coatings C-D, coatings E-F used larger size powders and showed less Sr evaporation. The rate of heat transfer through the boundary layer is proportional to the surface area of the particle (Ref 14). The evaporation time of a particle is proportional to its surface area and smaller particles have larger evaporation rates (Ref 5, 21). Therefore, the partial evaporation is proportional to the particle diameter. A similar result has also been reported by Cao (Ref 9). During plasma spraying of La2Ce2O7, the evaporation of CeO2 can be reduced effectively by increasing the particle size of the powder for plasma spraying (Ref 9).

For coatings B and C, longer spraying distance leads to a little higher degree of SrO evaporation. With longer spraying distance, the dwell time of the particle in the plume is longer, leading to higher total SrO evaporation.

Stoichiometric SrZrO3 Coating Received

Figure 9 indicates x-ray diffraction patterns of plasma-sprayed coatings. The main phase for all coatings is SrZrO3 while the secondary phases are different. The peaks at 2θ = 25.2°, 28.2°, 29.8° can be assigned to SrCO3, ZrO2 (m) and ZrO2 (t), respectively. Even if coatings were sprayed with different parameters and the resultant Sr/Zr of the as-sprayed coating are different, the phase compositions of all of the as-sprayed coatings indicated by XRD are still the same. The secondary phases of the as-sprayed coatings are SrCO3 and unstable tetragonal ZrO2 (Fig. 9a). The metastable tetragonal ZrO2 is typical of thermally sprayed zirconia coatings, which is due to the rapid solidification at a cooling rate >100 K s−1 of the melted liquid droplets. Under these conditions, the equilibrium phase was not retained and metastable crystalline or non-crystalline products may be formed (Ref 8). In the near future, thermal cycling test of SrZrO3 coating will be carried out with a surface temperature of 1350 °C. For comparison, these coatings were also sintered at 1350 °C. Thermal treatment can partly regenerate the phase of perovskite. During sintering, all of the newly formed SrCO3 reacted with ZrO2 (t) to form SrZrO3 again. That is why both of these two peaks disappeared after sintering (Fig. 9b-c). The Sr/Zr ratio has an influence on phase composition of the sintered coating. If Sr/Zr is a little larger than 1, after sintering the phase composition of the coating is pure SrZrO3 (Fig. 9c). Within limits of XRD, no secondary phases were detected. This means that after sintering, pure SrZrO3 coating can be recovered only when Sr/Zr ratio of the coating is a little larger than 1 (Sr/Zr = 1.038 for this coating). If the Sr/Zr ratio is smaller than 1, m-ZrO2 can be detected in the coating after sintered at 1350 °C for 72 h (Fig. 9b). During sintering, nonequilibrium t-ZrO2 transformed into m-ZrO2 upon cooling according to the following phase transformation:
$${t{\text{-ZrO}}}_{2}{\mathop{\rightleftarrows}\limits^{950\,{^\circ\text{C}}}_{1185\,{^\circ\text{C}}}}{m{\text{-ZrO}}}_{2}$$
https://static-content.springer.com/image/art%3A10.1007%2Fs11666-009-9360-z/MediaObjects/11666_2009_9360_Fig9_HTML.gif
Fig. 9

(a) SrZrO3 coating in the as-sprayed case; (b) SrZrO3 coating (Sr/Zr = 0.950) after sintered at 1350 °C for 72 h; and (c) coating (Sr/Zr = 1.038) sintered at 1350 °C for 72 h

Proposed Forming Mechanisms of Needle-Like Structure and Strontium Carbonate

So far, the detailed forming mechanisms of needle-like structure in powders sprayed into water and SrCO3 in both coatings and powders sprayed into water are still not clear. One mechanism is proposed here.

After spray-drying and before plasma spraying, the powders were annealed at 1300 °C for 5 h to remove organic binder and carbon source. Therefore, the newly formed SrCO3 must have formed during plasma spraying or later. During plasma spraying, molten powders are in contact with the surrounding atmosphere. They may decompose or have chemical reactions with the surrounding atmosphere. For example, oxides might form on the particle surface during plasma spraying of metals. Volatile species might form through thermal decomposition or reaction with the surrounding gas (Ref 5). Like perovskite-structured LaMnO3, LaCoO3 and YBa2Cu3Ox (Ref 6, 22), SrZrO3 also decomposes easily at the melting temperature. The needle-like structure is probably formed from completely molten phase originating from the readily melted ultra-fine Sr-rich SrZrO3 particles in flight. The new structure contained only Sr without Zr. Composition of the needle-like structure may be SrO at the beginning. However, Sr is strong carbonate former like Ba (Ref 23). Formation of SrCO3 from SrO and atmospheric CO2 for the SrZrO3 coatings or plasma-treated powders might be inevitable.

Conclusions

Partial evaporation of SrO happened during plasma spraying of SrZrO3, which is confirmed by plasma spraying of SrZrO3 powders into water. During plasma spraying SrZrO3 into water, some new needle-like structures formed and it contained only Sr. Once the initial composition of the powders is fixed, the final Sr/Zr ratio of the sprayed coating is determined by spraying current, distance and particle size of the powders used. By varying the spraying parameters and adding an excess amount of SrO before plasma spraying, stoichiometric strontium zirconate coatings can be produced. Traces of SrCO3 and metastable tetragonal ZrO2 can be detected in the as-sprayed coating. During sintering SrCO3 reacts with part of ZrO2 to form SrZrO3 again and the remaining metastable tetragonal ZrO2 transformed into ZrO2 (m). After sintering, pure SrZrO3 coating can be received only when the Sr/Zr ratio of the coating is a little larger than 1. In the near future, microstructure optimization and thermal cycling behavior of the received stoichiometric strontium zirconate coating will be investigated.

Acknowledgments

The authors thank Ms. M. Andreas for spray-drying of the powders, Mr. K.-H. Rauwald for the manufacturing of the plasma-sprayed coatings, Dr. G. Mauer for the particle diagnostic system operation, Mrs. S. Schwartz-Lückge for particle size measurements, Mr. M. Kappertz and Dr. D. Sebold for the sample microstructure characterizations, Dr. W. Fischer and Mr. M. Ziegner for XRD measurements. We would also like to thank Ms. N. Merki at the ZCH, Forschungszentrum Jülich, Germany, who performed the chemical analysis. The experiments described in this paper were carried out in Forschungszentrum Jülich GmbH and the first author was financially supported by NSFC-50825204 and CAS-DAAD program.

Copyright information

© ASM International 2009