Paper on alluminizing

J. Mater. Sci. Technol., Vol.25 No.4, 2009 433

Aluminizing Low Carbon Steel at Lower Temperatures

Xiao Si1)†, Bining Lu2) and Zhenbo Wang1)

1) Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences,Shenyang 110016, China

2) The High School Affiliated to Renmin University of China, Beijing 100080, China

[Manuscript received March 9, 2009, in revised form March 16, 2009]

This study reports the significantly enhanced aluminizing behaviors of a low carbon steel at temperaturesfar below the austenitizing temperature, with a nanostructured surface layer produced by surface mechanicalattrition treatment (SMAT). A much thicker iron aluminide compound layer with a much enhanced growthkinetics of η-Fe2Al5 in the SMAT sample has been observed relative to the coarse-grained steel sample.Compared to the coarse-grained sample, a weakened texture is formed in the aluminide layer in the SMATsample. The aluminizing kinetics is analyzed in terms of promoted diffusivity and nucleation frequency in thenanostructured surface layer.

KEY WORDS: Nanostructured materials; Surface mechanical attrition treatment; Aluminizing;

Diffusion; Nucleation

1. Introduction

Aluminizing is an effective surface modificationprocess for improving corrosion resistance of steels.Various aluminizing processes have been developedby enriching the surface layer of steels with a highconcentration of Al to form iron aluminide diffu-sion coatings, so that the ability to form impervi-ous and tenacious alumina scale is enhanced in cor-rosive media[1–3]. Nevertheless, as limited by the in-volved diffusion of Al and reaction kinetics betweenAl and Fe, effective aluminizing is normally performedat high temperatures with austenite phase. An ironaluminide coating can only be achieved on steels attemperatures above 900◦C with a duration of severalto dozens hours for the pack aluminizing process thatis most commonly used in industry. Holding at suchhigh temperatures might induce serious distortion ofworkpieces, carbide precipitation and grain coarsen-ing of the steel matrix, hence, degradation of mechan-ical properties. Apparently, lowering the aluminizingtemperatures of steels is of great significance for min-imizing these negative effects and widening the appli-cation of aluminizing techniques. More specifically,aluminizing steels in the ferrite state at a tempera-ture below 700◦C would be very much desired[4].

By means of a recently developed surfacenanocrystallization technique, surface mechanical at-trition treatment (SMAT)[5,6], lowering the aluminiz-ing temperature of steels becomes feasible. SMATenables to substantially refine grains in the surfacelayer of various steels into the nanometer scale via re-peated and multidirectional plastic deformation[5–10].Due to the significantly enhanced diffusion and chem-ical reactivity of the nanostructured surface layer pro-duced by SMAT, gaseous nitriding has been suc-cessfully carried out on a Fe plate at 300◦C[11],evidently lower than conventional gaseous nitrid-ing temperatures (∼550◦C). In addition, a muchthicker chromized surface layer has been obtainedon an SMAT low carbon steel sample than that onthe coarse-grained counterpart after the same pack

† Corresponding author. Senior Engineer; Tel.: +86 2423971882; E-mail address: [email protected] (X. Si).

chromizing treatment[12].In the present work, we demonstrate the possibil-

ity of lowering aluminizing temperature by SMAT ina commercial low carbon steel plate, which is amongthe most broadly used steels. The microstructure,hardness, chemical and phase compositions of the alu-minized SMAT surface layer were investigated in com-parison with those of the coarse-grained one after thesame treatment.

2. Experimental

A commercial low carbon steel, with compositions(wt pct) of Fe, 0.11C, 0.01Si, 0.39Mn, 0.024S (max),0.01P (max), was used in the present work. The sam-ple was annealed at 950◦C for 120 min in vacuum toeliminate the effect of mechanical deformation and toobtain homogeneous coarse grains. A plate sample(100 mm×100 mm×4.0 mm in size) was subjected toSMAT, of which the set-up and procedure have beendescribed previously[6,12]. In brief, a large number ofhardened steel balls (8 mm in diameter) were placedat the bottom of a cylinder-shaped vacuum chambervibrated for 60 min by a generator at a frequencyof 50 Hz at ambient temperature. The as-annealedsample was fixed at the upper side of the chamberand impacted by flying balls repeatedly and multi-directionally. Because the sample surface was plasti-cally deformed with high strains and high strain rates,grains in the surface layer were effectively refined.

The SMAT sample and the coarse-grained onewere aluminized under same conditions, i.e., at twotemperatures (500 and 600◦C, respectively) for 8 hin a packed powder mixture of 50Al, 2NH4Cl and48Al2O3 (in wt pct) in a double container designedby Meier et al.[13]. After aluminizing treatment, thesamples were wire-brushed and ultrasonically cleanedto remove adhering packed materials.

Cross-sectional observations of the as-SMAT andthe aluminized samples were performed on a NovaNano-SEM 430 scanning electron microscope (SEM).Al distribution in the aluminized surface layer wasmonitored by using a fully quantitative (Oxford Pro-grams) X-ray energy dispersive spectroscope (EDS).

434 J. Mater. Sci. Technol., Vol.25 No.4, 2009

Fig. 1 (a) Cross-sectional SEM morphology and (b) a typical bright-field TEM image of the top surface layerof the SMAT low carbon steel sample. The insert in (b) shows the corresponding selected area electrondiffraction pattern

A protective layer of pure Ni of ∼50 µm in thick-ness was electrodeposited onto the sample surface forpreparing the cross-sectional samples. Microstructureof the top surface layer of the SMAT sample was alsoobserved by using a Philip EM-420 transmission elec-tron microscope (TEM). In addition, X-ray diffraction(XRD) analysis of the surface layer was carried outto identify the phase information in the aluminizedsurface layer, by using a Rigaku D/max 2400 X-raydiffractometer with Cu Kα radiation.

The microhardness variation along depth from thealuminized surface was measured on cross-sectionalsamples by using a Nano Indenter XPTM (Nano in-struments) fitted with a Berkovich indenter. Themaximum load for each measurement was 9 mN withduration of 5 s, and the distance between any twoneighboring indentations was at least 10 µm. Theload-displacement data obtained during the first un-loading were analyzed using the Oliver-Pharr methodto determine hardness[14].

3. Results and Discussion

3.1 Microstructures of the SMAT surface layer

Clear evidences of plastic deformation have beenobserved in the SMAT surface layer of ∼200 µm inthickness, as shown in the cross-sectional SEM mor-phologies in Fig. 1(a). Grains in the surface layerare significantly refined and the microstructure differsmarkedly from that in the coarse-grained matrix (seethe bottom part in Fig. 1(a)). The degree of deforma-tion increases with decreasing depth from the topmosttreated surface, so that it is difficult to distinguish themicrostructure in the top surface layer of ∼100 µmby SEM. TEM observations in the top surface layerof the SMAT sample (as shown in Fig. 1(b)) revealthat the microstructure is characterized by ultrafineequiaxed ferrite grains with random crystallographic

orientations, as indicated by the selected area electrondiffraction (SAED) pattern (inset in Fig. 1(b)). Themean grain size obtained from a number of TEM im-ages indicates that it has been refined from ∼50 µmto ∼9 nm in the top surface layer by SMAT. Detailedmicrostructural characterizations of the SMAT sur-face layer by XRD and TEM show that the grain sizeincreases with increasing depth and it reaches 100 nmat the depth of ∼18 µm[12].

Ferrite grains are refined via sequential formationof dislocation cells in original grains, transformationof cell walls into subboundaries, and evolution of sub-boundaries into highly misoriented grain boundaries(GBs) separating the nanocrystallites[7,9,12]. Whenferrite grains are refined to a critical size, plastic de-formation occurs in carbide phase. Carbides in thesteel are progressively refined into smaller particlesand/or dissolved into the ferrite phase with increas-ing strain and strain rate[9,12], so that no cementite isobserved in the top surface layer in Fig. 1(b).

3.2 Aluminizing kinetics of the SMAT sample

The cross-sectional SEM observations for theSMAT and the coarse-grained samples after the alu-minizing treatment at 600◦C for 8 h are shown inFig. 2(a) and (b), respectively. It is clear that a con-tinuous and dense aluminide surface layer (the darklayer) has been formed on both samples. MeasuredAl concentration profiles (see Fig. 2(c)) indicate thatthe atomic concentration of Al is about 70% in thealuminide surface layers on both samples. In compar-ison with the aluminide coating formed on the coarse-grained sample (∼16 µm in thickness), the coating onthe SMAT sample (∼52 µm) is much thicker after thesame aluminizing treatment. A similar difference hasalso been observed on the samples with and withoutSMAT after the aluminizing treatment at 500◦C for8 h, as listed in Table 1. The thickness of the alu-

J. Mater. Sci. Technol., Vol.25 No.4, 2009 435

Fig. 2 Cross-sectional SEM morphologies of the SMAT (a) and the coarse-grained (b) low carbon steel samplesafter the aluminizing treatment at 600◦C for 8 h. (c) and (d) show variations of Al concentration andhardness with the depth from the topmost surface, respectively

Table 1 Comparisons of the average thicknesses (inµm) of aluminide surface layers on theSMAT and the coarse-grained (CG) lowcarbon steel samples after the aluminizingtreatments at 500 and 600◦C for 8 h, re-spectively. m is the ratio of k (see Eq. (1))on the SMAT sample to that on the CGsample

Temp./◦C SMAT sample CG sample m500 10.9±1.8 2.4±0.5 20.6600 52.5±9.3 16.3±4.3 10.4

minide coating on the SMAT sample aluminized at500◦C is comparable with that on the coarse-grainedsample aluminized at 600◦C. Formation of an obviousaluminide coating is difficult at temperatures below500◦C, due to the limited deposition rate of Al ontothe sample surface.

Because a large content (50 wt pct) of Al is con-tained in the pack powder mixture, a constant Al con-centration in the source might be expected during thealuminizing procedure at a fixed temperature and thegrowth kinetics of the aluminide layer can be repre-sented by a parabolic rate equation of the form[3,15],

y2 = kt (1)

where y is the thickness of the aluminide layer aftertreating duration of t and k is the mean growth rate.Therefore, the ratios (m) of k on the SMAT sample tothe one on the coarse-grained sample are derived forthe aluminizing treatments at 500 and 600◦C, respec-tively, as shown in Table 1. It is indicated that thegrowth kinetics of the aluminide layer on the SMATsteel is about 10 times higher than that on the coarse-grained sample at 600◦C, and the m value is doubledat 500◦C.

The much enhanced aluminizing kinetics in theSMAT low carbon steel is expected to result from the

formation of a nanostructured surface layer, in whicha considerable volume fraction of GBs (∼30 vol. pctfor an average grain size of 10 nm[16]) act as numer-ous fast diffusion “channels” for Al. In addition, ahigher energy state of GBs induced by SMAT rela-tive to the conventional GBs is expected to furtherincrease the diffusivity of Al in the nanostructuredsurface layer[17,18]. The lower m value at 600◦C thanat 500◦C in Table 1 might be induced by a faster graingrowth at the higher temperature. This is because thefraction and the excess energy of GBs may decreaseand result in a reduction of growth kinetics of the alu-minide coating on SMAT sample, accompanying thegrain growth at temperatures above 500◦C[12].

A previous work[19] revealed that the growth ki-netics of aluminide diffusion coating on an alloyedsteel was enhanced by shot peening at temperaturesbelow 667◦C and the enhancement effect progressivelydiminished as temperature increased. This work alsosuggests a positive effect of microstructure refinementon the aluminizing kinetics at lower temperatures.

The variation of hardness along depth in theSMAT sample aluminized at 600◦C was comparedwith that in the aluminized coarse-grained counter-part in Fig. 2(d). The hardness values of both sur-face layers are ∼15 GPa after the aluminizing treat-ment, while the matrix is about 3 GPa. However,the hardened surface layer on the SMAT sample ismuch thicker than that on the coarse-grained sampleafter the same aluminizing treatment. This differenceis consistent with the measured thicknesses of alu-minide surface layers on the samples. It is clear thatsurface hardness of the aluminized samples has beenpromoted by the formation of iron aluminide coatings.

3.3 Phase evolution in the aluminide layer

XRD patterns (as shown in Fig. 3) demonstrate

436 J. Mater. Sci. Technol., Vol.25 No.4, 2009

20 30 40 50 60

-Fe2Al5(c)(2

40)

(331

)(400

)

(112

)(1

30)

(002

)

(020

)

(200

)

(310

)

(a)

(b)

Inte

nsity

/ a.

u.

2 / deg.

Fig. 3 XRD patterns of the SMAT (a) and the coarse-grained (b) samples after the aluminizing treat-ment at 600◦C for 8 h. (c) shows an XRD pat-tern obtained from the reported powder diffrac-tion data (JCPD card No. 29-0043)

that aluminide coatings formed on the SMAT andcoarse-grained samples aluminized at 600◦C for 8 hconsist almost exclusively of η-Fe2Al5 phase. Com-paring with the XRD pattern for a standard powderspecimen of η-Fe2Al5 (Fig. 3(c), JCPD card No. 29-0043), where (002) and (130) Brag diffraction peaksshow the highest diffraction intensities, it is appar-ent that the iron aluminide coating on the aluminizedcoarse-grained sample is strongly textured because amuch higher intensity of (002) peak is detected, whileno obvious texture forms in the aluminized SMATsurface layer. According to the results of pole figureanalyses in literature [2], where a similar XRD pat-tern was obtained, a fibrous texture is expected in thecoarse-grained sample.

From thermodynamic considerations, θ-FeAl3phase possesses the lowest free energy of formationand it is expected to form preferentially in Fe-Al sys-tem. However, it is the η-Fe2Al5 phase that formsin most cases due to the higher growth rate andfavored crystallographic orientation (c axis)[2,3,15].For example, the growth rates of Fe2Al5 and FeAl3at 715◦C were reported to be 220 and 21 µm2/s,respectively[15]. The (00l) planes of Fe2Al5 phase arethought to be the most densely packed and smooth,giving the lowest surface energy[2]. In the coarse-grained sample, few Fe2Al5 particles are expected tonucleate at the GBs at the early stage of aluminiz-ing process and grow up with c axis aligned alongthe direction perpendicular to the interface betweencoating and matrix to minimize the surface energy.Therefore, a strong fibrous texture is developed afterthe aluminizing process. While a plenty of GBs in thenanostructured surface layer produced by SMAT sig-nificantly increase the nucleation frequency of Fe2Al5,the texture is more difficult to be developed because

grains might catch each other and stop growing at anearlier stage. It was discussed that the nucleation fre-quency might be increased by an order of about 106

with a reduction of grain size from 40 µm to 40 nm[12].

4. Summary

In conclusion, it has been demonstrated that alu-minizing low carbon steels at a temperature far belowthe austenitizing temperature is possible by the for-mation of a nanostructured surface layer by SMAT. Asurface layer consisted of η-Fe2Al5 phase, of ∼52 µmin thickness, is formed on the SMAT sample after apack aluminizing treatment at 600◦C for 8 h, morethan 3 times thicker than that on the aluminizedcoarse-grained counterpart. And the enhancement ef-fect is doubled at 500◦C. The enhanced aluminizingkinetics is expected to result from a much increasedGB diffusivity in the nanostructured surface layer. Inaddition, no obvious texture is detected in the Fe2Al5surface layer on the aluminized SMAT sample, due tothe significantly increased nucleation frequency of theFe2Al5 phase in the nanostructured surface layer.

AcknowledgementsThis work was financially supported by the Na-

tional Science Foundation of China (Nos. 50701044 and50890171), and the Ministry of Science and Technology ofChina (No. 2005CB623604).

REFERENCES

[1 ] R. Mevrel, C. Duret and R. Pichoir: Mater. Sci. Tech-nol., 1986, 2, 201.

[2 ] K. Murakami, N. Nishida, K. Osamura, Y. Tomotaand T. Suzuki: Acta Mater., 2004, 52, 2173.

[3 ] R.W. Richards, R.D. Jones, P.D. Clements and H.Clarke: Int. Mater. Rev., 1994, 39, 191.

[4 ] Z.D. Xiang and P.K. Datta: Acta Mater., 2006, 54,4453.

[5 ] K. Lu and J. Lu: J. Mater. Sci. Technol., 1999, 15,193.

[6 ] K. Lu and J. Lu: Mater. Sci. Eng. A, 2004, 375-377,38.

[7 ] N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu andK. Lu: Acta Mater., 2002, 50, 4603.

[8 ] H.W. Zhang, Z.K. Hei, G. Liu, J. Lu and K. Lu: ActaMater., 2003, 51, 1871.

[9 ] L. Zhou, G. Liu, X.L. Ma and K. Lu: Acta Mater.,2008, 56, 78.

[10] W.L. Li, N.R. Tao and K. Lu: Scripta Mater., 2008,59, 546.

[11] W.P. Tong, N.R. Tao, Z.B. Wang, J. Lu and K. Lu:Science, 2003, 299, 686.

[12] Z.B. Wang, J. Lu and K. Lu: Acta Mater., 2005, 53,2081.

[13] G.H. Meier, C. Cheng, R.A. Perkins and W. Bakker:Surf. Coat. Technol., 1989, 39-40, 53.

[14] W.C. Oliver and G.M. Pharr: J. Mater. Res., 1992,7, 1564.

[15] P.N. Bindumadhavan, S. Makesh, N. Gowrishankar,H.K. Wah and O. Prabhakar: Surf. Coat. Technol.,2000, 127, 252.

[16] C. Suryanarayana: Int. Mater. Rev., 1995, 40, 41.[17] Z.B. Wang, N.R. Tao, W.P. Tong, J. Lu and K. Lu:

Acta Mater., 2003, 51, 4319.[18] Z.B. Wang, K. Lu, G. Wilde and S. Divinski: Appl.

Phys. Lett., 2008, 93, 131904-1.[19] Z.D. Xiang and P.K. Datta: Scripta Mater., 2006, 55,

1151.