8/20/2019 Cyclic Oxidation

http://slidepdf.com/reader/full/cyclic-oxidation 1/5

Applied Surface Science 318 (2014) 275–279

Contents lists available at ScienceDirect

Applied Surface Science

 j ournal homepage: www.elsevier .com/ locate /apsusc

Cyclic oxidation kinetics and oxide scale morphologies developed

on alloy 617

Kh. A. Al-Hataba,∗, M.A. Al-Bukhaitia, U. Kruppb

a Mechanical EngineeringDepartment, Faculty of Engineering, Sana’a University, –Sana’a, Yemenb Faculty of Engineeringand computer Science, Universityof Applied Science, Osnabrück, Germany

a r t i c l e i n f o

 Article history:

Received 19 October 2013

Received in revised form 8 April 2014

Accepted 29 April 2014

Available online 10 May 2014

Keywords:

Alloy 617

High temperature

Cyclic oxidation

Two-stages

Oxide scale

Nodules

a b s t r a c t

In this paper, an attempt was done to investigate the cyclic oxidation behaviour of  alloy 617, cyclic

oxidation tests were carried out in laboratory air at 750, 850, and 950 ◦C up to 12 cycles (14 h/cycle). The

oxidation behaviour of alloy617 approximatelyfollowed the parabolic rate law and the average activation

energy is about 206kJ/mol. At lower temperatures, a two-stage oxidation kinetics were observed and

the transition time decreased as the oxidation temperature increased. SEM observations indicated that

continuous and relatively irregular oxide layers were formed that had a surface nodular-type structure

thickening with temperatures. XRD-patterns and SEM-EDS analysis revealed that the oxide scales were

mainly composed of Cr2O3 scale mixed with minor amounts of MnCr2O4. Other oxides were detected such

as NiO, TiO2  and MnTiO3 . Also, the geometrical-irregularities and Ni-metallic inclusions were detected

at the oxide/alloy interface. Moreover, Aluminium was internally oxidized to form A12O3  as elongated

particles, which were grown along grain boundaries via branch-like growth. The internal oxidation depth

was increased as the temperature increased.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

With regard to the broad high-temperature applications, a set

of solid solution hardening Ni-based superalloys are utilized for

steamgenerator (SG) and intermediate heat exchanger (IHXs) com-

ponents. As well known, SG would operate with reactor outlet

temperatures from 750 to 800 ◦C, while IHXs would operate up to

an outlet temperature of 950◦C [1]. Among solidsolution hardening

Ni-based superalloys, alloy 617 has been selected as a poten-

tial candidate alloy for SG and IHXs components because of its

excellent combination of high-temperature strength and good oxi-

dation resistance [2–4]. Currentlyavailablechromia-forming alloys

exhibit excellent oxidation resistance at temperatures lower than

1000 ◦C. This is due to their rely on the formation of continuous

and coherent protective Cr2O3   scale, which requires a sufficient

of Cr-content and higher Cr-diffusion supply for long service life.

Moreover, the service life of the chromia-forming alloys is consid-

erably reduced by the increase in Cr-depletion depth, presence of 

Cr-oxide volatilizationand underthermal-cycling conditions[5–7].

The good oxidation resistance of alloy 617 is derived from its

high Cr-content as well as Al is added primarily to improve the

∗ Corresponding author. Tel.: +967 771157027.

E-mail addresses: alhatab22@yahoo.com, dr.alhatab@suye.ac (Kh.A. Al-Hatab).

high-temperature oxidation resistance by partitioning into and

stabilizing the surface oxides [8,9]. The oxidation resistance and

the stability of the surface oxide layer depend on the interplay

between temperatures, alloycomposition, thermal cycling and oxi-

dizing environment. On the other hand, it is difficult to clarify the

oxidation mechanisms of alloy 617 due to its complex chemical

composition. Consequently, a brief review of the previous find-

ings on a similar model Ni–Cr–Al ternary system is necessary to

understand the oxidation behaviour of alloy 617. Ni–Cr–Al ternary

system had been extensively studied [10–20], which reported that

the oxide scale structures and oxidation mechanisms of Ni–Cr–Al

alloys can be classified into:

• Group-I has a low Cr < 5 wt.% and low Al< 2 wt.% contents that

leads to the formation of a less-protecting NiO scale and internal

oxidation of Al and Cr.•  Group-IIwith high Cr>15 wt.% andlow Al< 2 wt.% contents,forms

an external layer of Cr2O3   and the internal oxidation of Al takes

place to form Al2O3.• Group-III has a low Cr <15 wt.% and high Al> 7 wt.% contents,

which develops a continuous external layer of Al2O3.

Christ et al. [21,22] showed that there is no significant effect of 

Mo and Co on the oxidation behaviour of alloy 617 and hence, it

can be treated as the Ni–Cr–Al alloys group-II. The role of alloying

http://dx.doi.org/10.1016/j.apsusc.2014.04.199

0169-4332/© 2014 Elsevier B.V. All rights reserved.

8/20/2019 Cyclic Oxidation

http://slidepdf.com/reader/full/cyclic-oxidation 2/5

276 Kh.A. Al-Hatab et al. / Applied Surface Science 318 (2014) 275–279

 Table 1

Nominal chemical composition of alloy 617.

Ni Cr Si Al Fe Co Mo C Ti Mn

Bal. 22.1 0.1 0.96 0.95 11–14 8.5–10 0.08 0–0.2 0.03–0.3

elements such as Ti and Mn on oxidation behaviour of alloy 617

must be considered. Several considerable studies have been carried

out to evaluate the oxidation behaviour of alloy 617 in air, watervapor, and helium environments [23–42]. In general,Janget al. [35]

showed that compared to air environment, the oxidation rate of 

alloy 617 was not significantly affected in helium and steam envi-

ronments. On the other hand, the oxide morphology and structure

of alloy 617were strongly affected by the environments. They have

shown that the oxidation kinetics followed a parabolic oxidation

rate lawthat is related to thegrowth of external oxide scalesas well

as to internal oxidation for all of the above environments [21–42].

The oxide scales developed on alloy 617 specimens oxidizing in air

are mostly composed of Cr-oxide scales with NiO and/or NiCr2O4

additions [35,36,41,42]. While, the oxidizing of alloy 617 in helium

environment resulted in an external oxide scale that was mostly

composed of Cr2O3 mixed with less amounts of MnCr2O4 and TiO2

isolated grains [26–28,31,35]. In addition, isolated nodular MnTiO3oxides were formed on the MnCr2O4/Cr2O3   and sub-layer Cr2O3

oxide was formed in steam and steam + 20vol.% H2 [36,40].

Theevolution of cyclicoxidation kinetics andoxide scalesdevel-

oped on the alloy 617 at high temperatures has not been clearly

identified. Also, the effect of temperatures and the role of alloying

elements also has not been properly investigated. So, the present

work is a part of theattempt to understand theoxidation behaviour

of alloy 617 in air at750, 850, and 950 ◦C for upto 12cycles (168h).

2. Experimental procedure

Thealloy617was used inthe as-received conditions. Itsnominal

chemical composition is shown in Table 1.

Fig. 1a illustrates the as-received microstructure that wasobserved with an optical microscope after chemical etching. The

optical image of the as-received alloy 617showed a polycrystalline

microstructure and is typical austenite with many annealing twins

and precipitation of titanium nitrides and carbides preferentially

located at the grain boundaries and inside the matrix- grains [43],

and an average grain size of 55m was calculated using image-J as

shown in Fig. 1b.

The sheet of alloy 617 was cut into square coupons with dimen-

sions of 20×20mm2 and 1mm in thickness. The square coupons

were polished with SiC paper up to the 1200 polishing grade, then

ultrasonically cleaned and dried.

Cyclic oxidation tests were carried out in laboratory air up to

12 cycles. Each cycle was composed of heating (5◦C/min) to the

Fig. 1. (a) Surface microstructure optical image of as-received alloy 617 and (b)

Imaje-J statistical result.

Fig. 2. Kinetic curvesof alloy 617 sampesoxidized in air up to 12 cycles: (a) Linear

plots, (b) Log-log plots, (c) sequare of weight gain vs time in seconeds, (d) and (e)

two-stageoxidation at 750 and850 ◦C, and (f) kp  vs T−1 .

desired temperature, then 14h holding at the desired tempera-

ture, followed by furnace cooling (5◦C/min) with a subsequent

weight measurements using electronic balance with a measur-

able sensitivity of 0.1mg. The characteristics of the oxide scales

were investigated by X-ray diffraction (XRD), scanning electron

microscopy in combination with energy-dispersive X-ray spec-

troscopy (SEM-EDS) techniques. XRD was performed using Cu K(= 1.5418 A) a t 2◦ scale ranging from 16 to 90◦ and scanning rate

of 5 s/step (step 0.03◦).

3. Oxidation results

The details on the cyclic oxidation kinetics and oxide scale

microstructural observations are discussed in the following sec-tions.

 3.1. Cyclic oxidation kinetic results

The cyclic oxidation kinetics curves of alloy 617 have been

illustrated in Fig. 2. Fig. 2a represents the kinetics curves of weight

gains (w/A) in mgcm−2 vs the number of cycles. It shows that a

longertime wasneeded to detecta significant change in theweight

gain at 750 ◦C, but a quite identical behaviour with significant

increase in weight gain occurred at 850◦C. At 950 ◦C, oxidation

was accelerated compared to lower temperatures. As well known,

external and internal oxidation of Cr, Ni, Al, Mn and Ti could be the

main reason for the weight gains increased as time and tempera-

ture increased [9,44]. Asshown in Fig.2b, theslopes in log–log scale

8/20/2019 Cyclic Oxidation

http://slidepdf.com/reader/full/cyclic-oxidation 3/5

Kh.A. Al-Hatab et al. / Applied Surface Science 318 (2014) 275–279 277

Fig. 3. X-ray diffraction patterns of alloy 617 oxidized in air up to 12 cycles at: (a)

850 ◦C and (b) 950 ◦C.

are approximately about 0.5 which indicates that the oxidation

kinetics followa parabolicrate lawin theentire temperature range.

To calculate the parabolic rates, the weight gains were squared

and differentiated with respect to time as:

d(W )2

dt   = kpt  (1)

As shown in Fig. 2c, the parabolic rate constants, kp, were cal-

culated from the plot of (w/A)2 in mg2 cm−4 vs time in seconds

and the results are summarized in Table 2. Moreover, as indicated

in Figs. 2d and e, alloy 617 exhibited a two-stages oxidation kinet-

ics at 750 and 850 ◦C and the transition time between these stages

was decreased as the oxidation temperature increased. While, only

a one-stage oxidation kinetics was obtained at 950 ◦C, where a

very short and rapid initial stage with a higher weight gains were

occurred. Also, the two stages oxidation kinetics showed that the

slope of straight lines for each stage represents the parabolic rate

constants and denoted as k-I and k-II  that are also listed in Table 2.

Fig.2f shows theArrhenius plot of theparabolicrateconstantsvsthe inverse of absolute temperatures. The activation energy, Q , and

Q -I   for stage-I and Q -II   for stage-II are determined by considering

the temperature dependence of the parabolic rate constants that

obey the following Arrhenius-type expressions:

k p|Stage-I = 19338.06 ∗ e(−29075.56/T )

k p|Stage-II = 4.392 ∗ e(−18919.45/T )

k p|General = 586.43 ∗ e(−24767.09/T )

(2)

The activation energy, Q , Q -I   and Q -II   are calculated from the

slope of Arrhenius curves shown in Fig. 2f and are listed in Table 2.

 3.2. X-ray diffraction results

The X-ray diffraction analyses carried out on the alloy 617 sam-

ples oxidized at 850 and 950 ◦C up to 12 cycles are displayed in

Fig. 3. The X-ray patterns indicate that the diffraction peaks for

the underlying metal substrate were clearly noticed for all sam-

ples. This means that X-rays were completely penetrated the oxide

layer thickness and all the oxide phases will be detected. However,

the external oxide scales were strongly composed of Cr2O3, other

oxides such as MnCr2O4, TiO2   and MnTiO3   are weakly detected.

Hence, the Cr-oxide scale was mixed with less amounts of these

phases. Strong Cr2O3   peaks were detected at the 850 ◦C, then at

950 ◦C, MnCr2O4, TiO2  and MnTiO3   peaks gradually increased. In

contrast to the previous results in air [3], NiO and NiCr2O4  peaks

were not detected.

Fig. 4. Surface oxide scale morphologies developed on alloy 617 oxidized in air

up to 12 cycles : (a) SE M at 750 ◦C, (b) SEM at 850 ◦C, (c ) SEM at 850 ◦C (higher

magnifications), (d) EDX element analysis at positions marked in c, (e) BSE SEM at

950 ◦C,and(f)SEM at950◦C (higher magnifications).

 3.3. Surface and cross-sectional oxide scale structures

Fig. 4 shows the typical surface oxides morphologies formed on

alloy 617 oxidized for 12 cycles. At 750◦C, Fig. 4a indicates that

the oxide scales were only formed on partially localized surfaceareas and the alloy initial surface still visible. At 850 ◦C, in addition

to the formation of fine nodular-type structure, large nodules and

platelets were formed as shown in Figs. 4b and c. The EDS spec-

tra taken from the surface oxide scales are shown in Fig. 4d and

confirm that the large nodules oxides were mainly composed of 

Cr and are rich in Mn and Ti as compared with the fine nodular

type structure. The EDS analysis reveals that a less amount of Ti-

oxide was associated with Cr-oxide. At 950◦C, the alloy surface is

mostly covered with a dense nodular-shape structure and a par-

tial thicker grain boundary ridges were formed as shown in Fig. 4e

and f. The enhanced outward growth of nodules can be evident by

their end up joining together over the entire surface of the alloy.

Fig. 4a–c are sequentially show the continuous growth of external

oxide scale and grain boundary ridges with temperatures. As tem-perature raised, the denser Cr2O3   scale was formed mixed with

MnCr2O4 platelets and MnTiO3-nodules.

Fig.5a and b showsthe SEM backscattered cross-sectional views

of the oxidized samples at 850, and 950 ◦C respectively. Fig. 5b is

a cross-sectional view of alloy 617 specimen oxidized at 850 ◦C

up to 12 cycles; it indicates that a semi-continuous and very thin

oxide layer was formed. But at 950 ◦C, a continuous and relatively

thick external scale layer was formed with an about 8-m aver-

age thickness. This structure appeared compact and uniform. The

oxide layer wascharacterized by its largely wavy and chromia “fin-

gers” protruded into the alloy base. Ni-metallic inclusions (white

phase) were participated within oxide scale and at the scale/alloy

interface. The total oxide layer thickness wasapproximately 8-m.

EDS elemental mapping of cross-sectional oxide scales developed

8/20/2019 Cyclic Oxidation

http://slidepdf.com/reader/full/cyclic-oxidation 4/5

278 Kh.A. Al-Hatab et al. / Applied Surface Science 318 (2014) 275–279

 Table 2

Summary of theoxidation kinetics parameters of alloy 617 oxidizingin air at temperatures750, 850, and 950◦C.

Temp.◦C W mg cm−2 n k-Img2 cm−4 s−1 k-IImg2 cm−4 s−1 kpmg2 cm−4 s−1

750 0.13 0.6 1.1×10−08 5.5×10−08 2.2×10−08

850 0.27 0.6 6.7×10−08 1.05×10−079.9×10−08

1.5×10−07 a 3.2×10−08 a

950 0.79 0.4 1.2×10−06 1.2×10−06 1.2×10−06

5.5×10−06 a 7.8×10−07 a 8.8×10−07 b

Activation energy kJ/mol Q -I   Q -II   Q 

241.73±10 157.30±10 205.91±10a Ref. [36].b Ref. [3,35].

on alloy 617 at 950 ◦C after 168h isshownin Fig. 5c. It confirm that

theexternaloxidescale is mostlycomposed of Cr-rich oxide,which

was partitioning with little amounts of MnCr2O4 and MnTiO3. Fur-

thermore, a thin and semi-continuous layer consisting of Al-oxide

was formed just beneath the Cr2O3   layer. Also, Al was internally

oxidized to form Al-oxide fingers extends along the alloy grain

boundaries via branch-like growth and the internal oxidation takes

place evident by dark-grain boundary area as shown in Fig. 5b.

4. Discussion

The dependence of weight gains on the oxidation temperatures

was shown in Fig. 2. The kinetics results of alloy 617 such as the

parabolic oxidation rates, maximum weight gains, and the activa-

tion energy are summarized in Table 2. At 750 and 850 ◦C, it was

found that the two-stages oxidation process is occurring which

was previously investigated [36,45]. It can be seen that the tran-

sition time decreases as the oxidation temperature increases. But

at 950 ◦C, a single-stage linear relationship was prevailing that

had a very short and fast transient stage. Moreover, the gained

weight was increased as temperature increased and the mass

change is the sum of the mass gain by oxidation of alloying ele-

ments. Therefore, the gained weight should mainly result from

the external and internal oxides of Cr, Mn, Ti, and Al that formed

Fig. 5. Cross-sectional images of alloy 617 samples oxidized in air up to 12 cycles:

(a) BSI at 850◦

C,(b)BSI at950◦

C,and (c) EDX element analysis mapa at 950◦

C.

and still adherent to the alloy 617 [9,44]. However, the kinetic

results indicated that the oxidation of alloy 617 at 750 and 850 ◦C

was not so severe, but the gained weight was tripled at 950 ◦C.

Therefore, there is a critical temperature (850◦C) above it the

oxidation process was particularly severe. The oxidation kinetics

approximately followed the parabolic rate law that was increased

with temperature. The parabolic rates varied by two orders of 

magnitudes and are very close to the previously reported ones

[3,35,36].

Moreover, based on the kp   values obtained from the slope of 

lines in Fig. 2c, the activation energy, Q , was calculated to around205.91 ±10kJ/mol, which can be considered as the average value

of Q -I  and Q -II reported earlier. It has been suggested in the litera-

ture [36,40,46] that the high-temperature parabolic oxidation rate

of the alloy 617 is determined by predominant transport of Cr-ions

through a dense Cr2O3 layer where thereactionwith oxygenoccurs

at the scale/gas interface. Whereas, the activation energy for the

first stage is about 241.73kJ/mol which is very close to reported

activation energy for Cr-ions in Cr2O3   ranges from 240kJ/mol to

280kJ/mol [36,45–51]. The activation energy for the second stage

is about 157.30kJ/mol, suggesting that the oxidation kinetics is

affected by theformation of MnTiO3 andMnCr2O4 tomeettheoxide

scale boundary [36,45]. The difference in the activation energy

may be evident by the significance scale growth and thickening

occurring during the second stage [51–53]. The transition timebetween the two stages depends on the oxidation temperature

where a longer time was required at lower oxidation tempera-

ture.

Mn and Ti oxides were formed and incorporated in the Cr2O3

scale layer. Thus, Mn-Cr spinel was formed at the outer part of  

the oxide layer due to the faster diffusion of Mn than Ni, Cr,

and Fe in the Cr2O3   scale [3,27,36,40,49–53] . Unlike the oxides

formed under isothermal oxidation in air, the oxide scales iden-

tified through XRD and SEM/EDS analyses consisted of a large

nodular structure of MnTiO3   and isolated platelets of MnCr2O4

growth on the top of Cr2O3   instead of Ni oxide. Cr2O3   had fine

grain size comparedto MnCr2O4 and MnTiO3 oxides. Previous stud-

ies [3,35,41] reported that NiO/NiCr2O4/Cr2O3  multi-oxide layers

were formed on alloy 617 oxidized in air at 900◦

C after 1000h,but MnTiO3/MnCr2O4/Cr2O were formed in steam environment.

According to these studies, oxide scale growth proceeds with

inward diffusion of oxygen as well as the outward diffusion of Cr,

Mn, and Ti cations [35–40]. However, the requirement fora contin-

uous supply of Mn and Ti may not be able to be satisfied because

of their limited concentrations in the alloy 617 [3,35,41,44]. On the

other hand, Ni-metallic inclusions (white phase) were participated

within oxide scale and at the scale/alloy interface. The total oxide

layer thickness was approximately 8-m.

As reported in literature, about 1.1 wt.% Al is added to the alloy

617 increase the strength by precipitation hardening to improve

the oxidation resistance, stabilizing the surface oxides, and can

formdiscreteinternal andinter-granularAl2O3 oxidesin thematrix

underneath the external oxide layer [5,22]. However, Al was not

8/20/2019 Cyclic Oxidation

http://slidepdf.com/reader/full/cyclic-oxidation 5/5