Erosion-corrosion of thermal spray coatings

Surface and Coatings Technology, 43/44 (1990) 859—874

859

EROSION-CORROSION OF THERMAL SPRAY COATINGS BU QIAN WANG and GANG QIANG GENG Xian Institute of Highways, Xian, Shaanxi (People’s Republic of China) ALAN V. LEVY Lawrence Berkeley Laboratory, University of California, Berkeley, CA 94720 (U.S.A.)

Abstract Nine thermal spray coatings on mild steel, including hard tungsten carbide—cobalt and softer, more ductile metals that are supposed to provide wear protection, were eroded—corroded at elevated temperature using 250 ~im fluidized bed combustor bed material and quartz particles. It was found that all the coatings had lower material wastage rates than the base metal, 1018 steel. When the coatings were eroded by bed material particles, the ductile metal coatings had the same or lower material wastage rates than the harder tungsten carbide—cobalt. When quartz particles were used as the erodent, the ratio of the target hardness to the hardness of the erodent particles appeared to play a significant role in the erosion-corrosion behavior of the thermal spray coatings. The relationship between coating composition and morphology and the material wastage was determined.

1. Introduction Some thermal spray coatings are being used on structural steels in energy conversion and utilization systems to prevent surface degradation by corrosion, erosion or combinations of these mechanisms. The ability of the coatings to protect base materials against erosion-corrosion (EC) is related to their composition and processing, and especially to their morphology, i.e. the form and structure of the coating. The hardness of the coatings does not directly relate to their erosion resistance [1—3].However, some data in the literature [4, 5] suggested that the relative hardness values of the erodent particles and the target sample might play a significant role in the erosion behavior of brittle materials. Decreasing the ratio of the target material hardness to the hardness of impacting particle resulted in a dramatic increase in the erosion rates, mainly when the ratio was in the range of unity [5,61. In the present work a number of metallic and hard material coatings designed to protect metal surfaces against EC were applied to 1018 steel by 0257-8972/90/$3.50

© Elsevier Sequoia/Printed in The Netherlands

860

several thermal spray processes. The erosion behavior of the coatings was determined at the standard test conditions used at Lawrence Berkeley Laboratory (LBL) to determine the EC behavior of heat exchanger tubing steels. The test conditions approximately simulated the EC environment in the convection pass, superheater region of a circulating fluidized bed combustor (CFBC). Two different erodents were used. All the coatings were commerically available coatings applied by industrial thermal spray companies. The relationship between coating composition, morphology and the material wastage was determined, and also the role of the ratio of the target hardness to the hardness of the erodent particles in the EC behavior of the coatings was studied.

2. Test conditions 2.1. Materials The target materials for the EC tests were nine thermal spray coatings on mild steel. The characteristics of the coatings are listed in Table 1. The test surfaces of the coatings were polished to 600 grit prior to testing. The thicknesses of the coatings listed in Table 1 were measured after polishing. 2.2. Test description The tests were carried out in the LBL room temperature [7] and elevated temperature, blase nozzle [81 types of testers. Air was utilized as the carrier fluid for the particles, creating a generally oxidizing atmosphere at elevated temperature. Two different erodents were used; 250 ~m CFBC bed material number 2 and 250 ~.tmSi02 fused quartz. The CFBC bed material erodent had an angular shape and consisted primarily of Si02 and CaO, as determined by X-ray diffraction analysis. The detailed test conditions are listed in Tables 2 and 3. The material wastage of the specimens was determined by microscopic dimensional measurements to determine the thickness changes of the specimens in the central part of the erosion zone.

3. Results and discussion 3.1. Erosion—corrosion rates The material wastage of specimens eroded—corroded by the different erodent particles at 450 °C, a particle velocity of 20 m s’ and an impact angle of 30° are listed in Table 2. For comparison, the data [9] for the 1018 mild steel substrate are also listed in Table 2. From Table 2 it can be seen that all the coatings tested, using both erodents, had lower material wastage than bare 1018 steel. Therefore all the coatings tested were protective and can improve the EC behavior of 1018 mild steel tubing. Generally the Si02 fused quartz was more erosive than the

861

-4

Cl)

~

a

h

-4

CI)~

—

~ .~

;~

~

t~. U~

.-4’~OLL~

Q

a

C~o Qoc)

0

0

a

_4

,—l

—

CI)

ç

b

~ ~

~

~

-4

~-4

+

~Q00

0

~

~

L.~Q

C.)

—

~

—

+•~

ç)OC.)

~.

_$~.

~

~CJ}.~L 4

1.. 04

~

~

h

88

8

0.4.41.4 ~040. 1.4

~

888

~

C.) +4

~

~L)Z-4ZC~l -

04

~

.~

~+

~QC)

CI)Cl)

zz ~

oo~c~~co a)

~00

~ a)

~

862

I 03 CO C-’

C0~t.:c4i -4

‘.3 ‘4

P

8 C-

0

C.)

C—

C-- N ‘C CO CO CO C’)

—aaaaaaaa

—

C 0

8

N CO 40 001

40 CC ~CC

bL.—~

‘4

I.a) .0

8

.~

I a) ‘4

‘4

-. ~

.~

8 V .~

-4

010

40

-~.-~aaa

0

~ -i0~

81.4 a) 4.44.4

‘4.4

IICO

.o.~ 88

0

a)

~

+4

‘4

-

CIDCI)

ZZ C-N

CID

zz

0

+4)a 00)

C--

00C’l~00’)’0O~ V V — C) 01 + V V +4 —• +4~4 +4 +4 CC

+4

~0 E-~

(12(1)

~

Q~1.4

0)

863 TABLE 3 Material wastage of three coatings at different test conditions Designation

Thickness loss Qim) Eroded at room temperature 2

V~—1Oms1

V=20ms1

cz=30° ~=90°

a=30°

cz=90°

a=30° a=90°

s=30°

1.4

7.2

2.6

9.4

7.5

2.7

9.8

2.4 1.8

2.8 1.3

3.3 1.8

6.3 1.7

8.1

3.4 7.8

10.1 11.6

V—lOms

Metco 76FNS (number 1) SMI-28 Metco 468

Eroded at 450 °C

V=20ms’

Erodent particle, 250 ~im Si0 (V=20ms’) or 600g (V= lOms’). 2 t = 5 h (V

1) =

or 8 h (V

=

~90O

10 m s’); loading, 375 g

20 ms

CFBC bed material. When the CFBC bed material was the erodent, the hardness of the coatings had no direct relationship to material wastage. The ductile, softer coatings had the same or lower material wastage than the harder tungsten carbide-cobalt coatings. When the Si0 2 fused quartz was the erodent, the hardness of the coatings had an inverse relationship to material wastage. The ratio of the coating hardness to erodent hardness vs. the material wastage is plotted in Fig. 1. It can be seen that there are two ranges in the curve. Above a ratio of 0.56 the curve is smooth and shows a slight reduction in material wastage with an increase in the coating:erodent hardness ratio. Below a ratio of 0.33 the curve is a steep straight line showing a dramatic increase in material wastage with a decrease in the ratio. The former range corresponds to the hard material coatings and the latter range covers the softer, ductile metal coatings; see Table 2. References 5 and 6 reported that when the ratio of the target hardness to the hardness of the erodent particle approached unity, there was a sharp transition in erosion rates. In those investigations target samples were brittle materials and the relationship between the hardness ratio and the erosion rate resulted from tests in which the same target material was eroded using erodents with different hardnesses. In this investigation different coatings with different hardnesses were the target materials that were eroded using the same erodent. When the Si02 fused quartz was the only erodent used, there was a different correlation between the hardness ratio and the material wastage. A sharp transition occurred, but at a ratio of 0.4, not unity. As mentioned above, when CFBC bed material was the erodent, there was no correlation at all. The CFBC bed material contained CaO, a soft weak constituent that is prone to shattering, which results in covering the surface of the target with a layer of material. Therefore, when this kind of erodent

864 8.40

7.0 Test conditions: T=450C,t—5hr,a~30 ,V=20m/s 375gm loading, Air,250~imSi02

6.0 0

5.0 C

a a o

4-1

-.

0,,, -~

S~._00

3.0 a

‘—

o

C

~~~~——0--—

1.) -~

2,0

4-a

1.0 0 0.2

0.4 0.6 0.8 1.0 1.2 ratio of hardness (hardness of coating/hardness of erodent)

Fig. 1. The relationship between the ratio of the coating hardness to the erodent hardness and the material wastage.

particle impacted the surface of the target, the damage pattern was no longer based on the indentation fracture mechanics of the target material [10, 11]. From Table 2 it can also be seen that the coatings had different relative material thickness losses with each of the two erodents used. The WC— (Cr,Ni,Co) coating, SMI-28, had the lowest metal wastage and the WC—Co Metco 76F-NS coating had the highest material wastage, when CFBC bed material was the erodent. When Si02 fused quartz was the erodent, the Metco 76F-NS coating had the lowest material wastage. In ref. 1 it was reported that the straight erosion behaviors of coatings tested at room and elevated temperature were very dependent on the shapes and properties of the erodent particles. Different erosion behaviors occurred for the same material or coating when it was impacted by different particles and the ranking of materials tested varied when different erodent particles were used. In this work the data in Table 2 are from specimens tested at 450 °C. X-ray diffraction analysis determined that essentially no oxidation of the coatings tested occurred, except for the 420SS coatings. Therefore the surface was only subjected to elevated temperature erosion without oxidation. The tests in ref. 1 were also erosion only so that both test series had similar results.

U

865

3.2. Effect of test variables on material wastage Three thermal spray coatings, the two best-performing coatings and the worst-performing coating listed in Table 2, were eroded with 250 ~tm Si02 fused quartz at room temperature and 450 °Cat impact angles of 30°and 90° and particle velocities of 10 m s’ and 20 m s~.The material wastages of the three coatings at different test conditions are listed in Table 3. 3.2.1. Temperature It can he seen that when the test temperature was increased from room temperature to 450 °Cthe material wastage of the two best-performing coatings, Metco 76F-NS (number 1) and SMI-28, increased only a small amount. However, the worst-performing coating at 450 °C,Metco 468, had the lowest material wastage of the three at room temperature. Increasing the test temperature from 25 °Cto 450 °Cgreatly increased the material wastage of the Metco 468 coating at both impact angles of 30° and 90°, and at both 1. particle velocities of 10 m s~ and 20 ms 3.2.2. Impact angle From Table 3 it can be seen that the material wastages of the Metco 468 coating at room temperature for all of the test conditions were nearly the same. However, the other two coatings tested had erosion behavior that was typical of brittle materials at both test temperatures and both particle velocities, i.e. greater material loss at an impact angle cx of 90° than at cx = 30°. Even the Metco 468 metallic coating at 450 °C had the erosion behavior of a brittle material. The data listed in ref. 12 indicated that both the elongation and the strength of some Ni—Cr—Al alloys dramatically decreased at certain elevated temperatures. The Metco 468 coating may be one of these alloys. It is postulated that the Metco 468 coating, which is a Ni—Cr—Al alloy, had the best ductility at room temperature, which accounted for its low material wastage. However, on increasing test temperature to 450 ~C, the ductility of this coating was dramatically reduced, resulting in the higher material wastage and the erosion behavior of a brittle material. 3.2.3. Particle velocity In Table 3 it can be seen that when the particle velocity was increased from 10 m s’ to 20 ms’ at both test temperatures the material losses increased for the three coatings even though the erodent loading was reduced from 600 g to 375 g and the test time was reduced from 8 to 5 h. The effect of particle velocity on changing the thickness loss was less at room temperature for the Metco 468 coating than for the others. 3.3. Metallographic analysis The erosion material wastage rates correlated well with the composition and morphology of the coatings. Figure 2 shows that the flame spray 420SS eroded by the cracking and chipping of a thin, intermittent iron + chromium oxide scale layer, which was corroborated by X-ray diffrac-

U

osrf.oefuneroded)

j

surfoce(eroded)

lOirn

Orn

j~ ~,

url°

~JJVL~

(~npositiooofsorface (eroded)

_

As

Compost tion of

cross

cross section (1)

section 20cm

I

Composition of cross section (2f

___________

Compost tfon of

cross section (3)

XBB 899-7859 405 sire bonding and 420 stire thermal spray Nooeie Te.4ter T450’C Erosion a30’ 250pm CFBC bed material No.2.V..20m/s 375cm 1~’~~— C—tAr. Air

Fig. 2. Surface and cross-section of flame-sprayed 420S5 eroded by CFBC bed material.

.4

.4

867

tion analysis. The scale layer had bed material, i.e. calcium and silicon, embedded in it. The coating was quite dense with only a few small pores present in the area shown. Contrast this type of porosity with that shown in Fig. 3 for the hypersonically sprayed 420SS. The presence of more and larger pores resulted in a 25% higher material wastage when CFBC bed material was the erodent, 4.0 jim compared with 3.2 jim, as seen in Table 2. When Si02 was the erodent, material loss was by direct erosion of the metal by the platelet mechanism of erosion, as can be seen in Fig. 4. There was no embedment of erodent particles to form a layer as occurred when CFBC bed material was the erodent. The material wastage was greater when Si02 was the erodent, indicating that the scale layer of bed material and iron oxide that formed when the erodent was CFBC bed material provided some degree of protection to the coating. The bed material erodent caused the formation of the protective layer, insulating the substrate from the direct erosion process. Hence, the carbides and metal alloys had about the same material wastage. When Si02 was used as the erodent and the substrate was being directly eroded, the high WC content coating had lower thickness losses than the softer metals. The WC + Co coatings eroded by a cracking and chipping mechanism, as can be seen in Fig. 5. The excess loss of the Metco 76F-NS (number 1) coating eroded by CFBC bed material was due to the defective coating—base metal bond. Some area of this coating broke away after the test. The reason for the low metal wastage of the SMI-28 coating eroded by FBC bed material can be seen in Fig. 6. The coating was dense, with relatively few, small pores. Compare this coating with the more porous hypersonic spray JK112 coating whose hardness level was similar but whose material loss was twice as great. Figure 7 shows the cracked and chipped surface of the Metco 76F-NS (number 2) coating eroded by Si02. The smooth character of the eroded surface indicated that only very small chips could be eroded, accounting for the low, 2.5 jim, material thickness loss. There was no embedment of Si02 in the surface of the coating. The FeCrA1Mo coating (Metco 465) that had high metal wastage in the 5 h test using 5i02 erodent is shown in Fig. 8. The NiCrA1Co Metco 468 coating had the same morphology as this figure. It lost material by the direct erosion of the metal with no protection from any kind of layer on its surface.

4. Conclusions (1) All the coatings tested had lower material wastage than the base material, 1018 steel. (2) Generally, the Si02 fused quartz was more erosive than the CFBC bed material, except eroding coatings Metco 76NS, No. 1 and JK112.

868

:~.i

___

____

—i---

Composftion of ssrface

(uneroded area)

~onpos1 tion

:

L-~

.~

of oar face (eroded areaf

Fe

Conposition of cross

section (1)

H Cross

Ii

sersion

Nezz1r~Ct005t

sprayed 420 stables:

250pm CFBC bed material No.2,V20n/n 375gm loadbng.t—5he. Air

ConpostlooofrrossseotiOn (2) 1418

899—7961

Fig. 3. Surface and cross-section of hypersonically sprayed 420SS eroded by CFBC bed material.

869

Lfssr~~ 1i

___ _______________

I

~ CompositIon

j_~-

~cJ of cross section

(1)

~LJJF Componitlnn

of

Crnss section

14141 405 * 420 coatIng Noecle Tester T450’C Erosion t—30’ 250,m sand V20m/n 375gm loadieg,t5hr. Air

Fig. 4. Surface and cross-section of flame-sprayed 42088 eroded by sand.

(2)

899-7860

870

surface (sneroded)

LJ

nurface(eroded)

S

~

Co -

Composjtion of

surf ace(soerodet1)

Composition of

Ca

L

~

ssrface (eroded area)

S

Composition of cross

section

cross nection~)

[j 2lim

JK112 hypersonic sprayed WC+ 12% Co powder Nozele Tester T450’C Erosion 2SOurn CPBC bed material p30 Nn.2,V—20m/s 375 gm loading,t=5hr, Air XBB 099-7963

Fig. 5. Surface and cross-section of hypersonically sprayed JK112 WC—12%Co eroded by CFBC bed material.

871

Surface

ssrfaoe (sneroded area)

-

~~ i~ j~ I

C~ CeComposition of

(eroded area)

N

composition of surface (eroded area)

surface (uneroded area)

Cr

—

Ni

__

—

Composition of

crass sectionU)

As

C

ti

i 250pm 375gm

I

~

i

t

T4SOC

BFBC bed material No.2 loadiog.t5hr.Air.V—20m/s

C

—

________________

as .

Compositiam of crass

seotOns( 1) 04141 899-7868

Fig. 6. Surface and cross.section of SMI-28 coating eroded by CFBC bed material.

872

•

I

C mP~ff

lO 4m

_____________________________________________

I-

Composi Sian of

C ross

section

(A position)

cross secnian 41Cm

76 f—IS flame sprayed Cf. * I if C pcwdrr Necir Troirr f4)55f, (No.2) Ercsio 2Ni. n sand, V2is/s .3) 5gn I cad ug,t Nh r Air XBB 899—7867

Fig. 7. Surface and cross.section of Metco 76F-NS coating (number 2) eroded by sand.

(3) When the CFBC bed material was the erodent, the hardness of the coatings had no direct relationship to material wastage. The ductile, softer coatings had the same or lower material wastage than the harder, tungsten carbide-cobalt coatings. (4) When Si02 fused quartz was the erodent, the hardness of the coatings had an inverse relationship to material wastage. (5) The erosion material wastage correlated well with the composition and morphology of the coatings. Dense structure, the absence of pores and the build-up of a protective layer when CFBC particles were the erodent resulted in lower material wastage.

873

-

-

CtLJ

ii

-

Fe

‘

______

2iin

ii--- ~ Composition of

Metco 465(Fe—Cr—Ai—Mo) bade Tester 1451’C Erosion i=30’ 250Cm sand V20m/s 375gm loadlng,r—Shr, Air

-—

C russ section

(2)

0841

899—7871

Fig. 8. Surface and cross-section of FeCrA1Mo coating (Metco 465NS) eroded by sand.

874

Acknowledgment This research was sponsored by the Department of Energy under DOE! FEAA 10 10 0, Advanced Research and Technical Development, Fossil Energy Materials Program Work Breakdown Structure Element LBL-3, under Contract DE-AC035F00098.

References 1 A. V. Levy and B. Q. Wang, Erosion of hard material coating systems, Wear, 121 (3) (1988) 325—346. 2 A. V. Levy, The erosion—corrosion behavior of protective coatings, Surf. Coat. Technol., 36 (1988) 387—406. 3 Z. R. Shui, B. Q. Wang and A. V. Levy, Surf. Coat. Technol., 43/44 (1990) 875. 4 R. 0. Scattergood and S. Srinivasan, Erosion of brittle materials, Proc. Con!. on Tribological Mechanisms and Wear Problems in Materals, ASM’s Materials Week ‘87, Cincinnati, OH, October 1987, American Society for Metals, Metals Park, OH, pp. 1—10. 5 S. Srinivasan and R. 0. Scattergood, Effect of erodent hardness on erosion of brittle materials, Wear, 128 (1988) 139—152. 6 5. Wada and N. Watanabe, Toyota, Research and Development Laboratory, 1987, to be published. 7 A. V. Levy, The solid particle erosion behavior of steel as a function of microstructure, Wear, 68 (3) (1981) 269—287. 8 A. Levy, J. Yan and J. Patterson, Elevated temperature erosion of steels, Wear, 108(1) (1986) 43—60. 9 A. V. Levy, B. Q. Wang and G. Q. Geng, Characteristics of erosion particles in fluidized bed combustors. Proc. ASME 1~hFBC Con!., San Francisco, CA, May 1989, ASME, New York, 1989, pp. 945—951. 10 B. R. Lawn and R. Wilshaw, Indentation fracture: principles and application, J. Mater. Sci., 10 (1975) 1049. 11 A. G. Evans and R. Wilshaw, Quasi-static solid particle damage in brittle solids — I. Observations, analysis and implications, Acta Metall., 24 (1976) 939. 12 E. A. Brandes, Smithells Metals Reference Book, Butterworths, London, 6th edn., 1983, Chapter 22, pp. 71—72.