GRAIN BOUNDARY SEGREGATION OF BORON. AN EXPERIMENTAL AND THEORETICAL STUDY

HAL Id: jpa-00225938

https://hal.archives-ouvertes.fr/jpa-00225938

Submitted on 1 Jan 1986

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

GRAIN BOUNDARY SEGREGATION OF BORON.

AN EXPERIMENTAL AND THEORETICAL STUDY

L. Karlsson, H. Norden

To cite this version:

L. Karlsson, H. Norden. GRAIN BOUNDARY SEGREGATION OF BORON. AN EXPERIMENTAL AND THEORETICAL STUDY. Journal de Physique Colloques, 1986, 47 (C7), pp.C7-257-C7-262.

�10.1051/jphyscol:1986744�. �jpa-00225938�

Colloque C7, suppl&ment au n o 11, Tome 47, Novembre 1986

GRAIN BOUNDARY SEGREGATION OF BORON. AN EXPERIMENTAL AND THEORETICAL STUDY

L. KARLSSON and H. NORDEN

D e p a r t m e n t o f P h y s i c s , C h a l m e r s U n i v e r s i t y o f T e c h n o l o g y , S - 4 1 2 96 G d t e b o r g , S w e d e n

w -

The size of the boron atom is intermediate between those of common interstitial and substitutional elements in austenitic stainless steels and nickel (1). Consequently, in these macerials bomn has a very low solid solubility (2.3). a saong tendency to segregate to grain boundaries (3-7) and interacts strodgly with lattice imperfections.

In general the enrichment of solute atoms to grain boundaries can be due either to equilibrium or non-equilibrium segregation. Equilibrium segregation occurs when a material is held at a temperature sufficiently high to permit appreciable diffusion of solute atoms. The interfacial free energy of the loosely packed interface regions is then reduced by the adsorption of solute atoms. The segregated atoms are localized to a few atomic layers at the boundary (8).

Non-equilibrium segregation is a dynamic process which takes place during cooling from high temperatures and results in a wider zone enriched in solute, the width depending on the heat treatment procedure. W e enrichment is an effect of mobile vacancy-solute complexes diffusing down vacancy gradients towards vacancy sinks. The change in equilibrium vacancy concentration during cooling creates the vacancy gradients necessary for segregation to occur (4.9.10).

Both nonequilibrium and equilibrium segregation can occur simultaneously. However, the shape of the resulting concentration profiles and the time and temperature dependences of the mechanisms are quite differen& which makes it possible to separate the effects. Earlier investigations have shown thatboron has a strong tendency to segregate to grain boundaries in austenitic stainless steels and nickel (3.4) and qualitative indications exist that non-equilibrium segregation is the dominating process in steel (4). However, no quantitative information about the segregation process in austenitic stainless steel exists and any direct observation of non-equilibrium segregation in nickel has not been r e p o d to date.

The aims of the present investigation have been (i) to quantitatively describe grain boundary segtegation of boron in austenitic stainless steel as a function of the heat treatment, (ii) to investigate the influence of the segregation on the precipitation behaviour and elemental distribution at grain boundaries a n d (iii) to determine which s&r&ation mechanisms operate in nickel.

Three austenitic stainless steels (316L with 4 0 ppm or <1 ppm B and "Mo-free 316L" with 23 ppm 8, see Table I) and one nickel-boron alloy (Specpure Ni with 60 ppm B) were studied.

A series of heat treatments of the steels were made in order to study the effects of cooling rate (0.25 to >600 'CIS) and scaning temperature (500 to 1400 'C) (see (5)). A preliminary study was also performed on the nickel-boron alloy cooled at 40 'Us from 1090 'C.

The elemental dismbutions and the precipitation at grain boundaries were studied both on a very fine and a c o w scale using a combination of microanalytical techniques: secondary ion mass specuomeuy (SIMS), transmission electron mi~roscopy (TFM), energy dispersive X-ray analysis (EDS), field-ion microscopy (FIM), atom-probe microanalysis (AP) and imaging atom-probe micmanalysis (IAP), and also by wmprter simulations.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1986744

C7-258 JOURNAL DE PHYSIQUE

Table I. Composition of the steels

Element B C Cr Fe Mn Mo N Ni Si

"Low bomn

AISI316L" wt% c 1 ppm 0.015 17.4 bal. 1.76 2.61 0.05 13.1 0.58

"High boron

AISI 316L" wt% 40ppm 0.016 17.5 bal. 1.76 2.62 0.05 13.2 0.55

"Mo-free

AISI 316L" wt% 23 ppm 0.016 17.4 bal. 1.80 0.01 0.04 13.2 0.62

Specimens for FQA, AF' and LAP analysis were prepared by the floating layer technique.

However, the depth accessible for analysis is typically 100-500 nm whereas the grain size of the heat treated materials was larger than 100 p~ Therefore, in a 6nal specimen preparation step, material was removed in a controlled manner by stepwise elecmpoliihing and subsequent examination by TEM until a grain boundary was suitably positioned for analysis (Fig. I). The methods of specimen preparation and of mM, AP and IAP analysis of elemental disnibutions at grain boundaries have been

described in considerable detail elsewhere (611). Figure 1. TEM (brightfreld) micrographs @a FIM specimen with a grain bowuiary (arrows) (a) w o r e and (b) qftw AP analysis. "Mo-fee 316L." with 23 ppm B cooled at 31 "CIS from 1250 "C.

-

px~erimental re-

Boron segregation was found after all heat treatments, at all grain boundaries except at coherent twin boundaries. The general segregation behaviour was the same for all materials studied. Therefore, results for the 316L steel with 40 ppm B are used to exemplify the segregation characteristics.

AP and IAP analysis showed that on a very fine scale the segregation profiles after cooling from 1075 and 1250 "C were smooth, peaked at the boundary with a tail of deneasing concentration away fiom the boundary (Figs. 28-i and 3). The amount of segregated boron increased with increasing starting temperature and was largest at intermediate cooling rates (Figs. 2g-i and 4a). The width of the boron enriched zone increased with increasing starting temperature and was larger for lower cooling rates. Investigations on a coarser scale using SIMS showed that at the lowest cooling rates, the boron distribution changed from an enrichment localized to the boundary to a wider, mox broadly spread enrichment (Figs.

2 d 4 .

The segregation behaviour of the steels shows that non-equilibrium segregation is the dominating process for the hjgher starting temperatures (1075 and 1250 'C) and that the relative magnitude of enrichment increases with decreasing boron content A low bomn content is therefore no guarantee that appreciable bomn segregation will not occur. The measured segregation profiles in nickel indicate that bomn non-equilibrium segregation occurs also in nickel.

Equilibrium segregation dominated in the steels rapidly cooled from 8WC. The enrichment was in this case localized to within a few atomic layers at the boundary (Fig. 5). The energy at austenite boundaries was estimated at 0.65 M.04 eV by applying the McLean equation for equilibrium segregation (12). For further details on the experimental results see (5-7).

com~uter simulations

A computer model was developed to simulate non-equilibrium segregation. An iterativenumerical method of solving the diffusion equations, the explicit f i ~ t e difference formula method (13), was used to calculate segregation profiles one-dimensionally between a grain boundary and the centre of the grain. In the segregation model free boron atoms, free vacancies and vacancy-boron complexes were considered. It was assumed that local equilibrium between these three entities ismaintained concurrent with their behaviour as distinct entities diffusing down their own concentration gradients.

During cooling through a large temperature interval the me equilibrium concentration of vacancies, and thus complexes, cannot be maintained except at vacancy sinks (e.g. grain boundaries). Thus vacancy gradients are formed and a flow of vacancies and complexes, towards the sinks will occur. Vacancies, free or in complexes, are annihilated at the sink where bomn atoms, previously forming complexes, are deposited.

DISTANCE FROM G.B. (pm)

-50 -40 -30 -20 -70 0 10 20 DISTANCE FROM G. 6. (nm)

240

230 PI250 'C

- .-

*

0 7 ²¹⁰

m

200 0

0 2 5 5 0 75 -50 -40 -30 -20 -10 0 10 20

DISTANCE FROM G.B. (pm) DISTANCE FROM G. B. (nm)

200 ⁰

-

-50 -40 -30 -20 -10 0 10 20 DISTANCE FROM G. 6. (nm)

Figure 2. B distributions after cooling the 316L steel with 4OppmB ar 530,13 or 0.29 U s from 1250 'C.

(a-c) Computer simulations, (d-f) SIMS "B ion micrographs obtained using 02+ as primary ions, and (g-i) AP concentration profiles for B.

A sufficient condition for boundary enrichment to occur is that complexes are more mobile than free boron atoms. The boron concentration then has a maximum at the boundary, further out there is a depleted zone and the original concentration is maintained far from the boundary (Fig. 2a). The maximum enrichment at the boundary occurs at some intermediate cooling rate (Figs. 2b and 4b) and the enrichment becomes more broadly spread for lower cooling rates. The boundary concentration increases rapidly with increasing starting temperature (Fig. 4).

The shape of the simulated concentration profiles, their variation with starting temperature, cooling rate and boron content were in very good agreement with the experimental results. Diffusion coefficients for boron and vacancy-boron complexes in austenite were estimated from comparisons of the cooling rate dependence of the boundary concentration as measured and as calculated. The diffusion coefficients; ~~=2l0-~ex~(-l.l5(e~)/k?") and

~ ~ ~ = 2 ~ 1 0 ~ e x ~ ( - l . l 5 ( e ~ ) l k ~ ) (m2/s) respectively, suggest that the mobility of boron m austenite has a stronger temperature dependence than previously reported A detailed description of the model and the results can be found elsewhere (10).

AISI 316L

8,600 'CIS

z

0

Y

0

-50 -40 -30 -20 -10 0 10 20 DISTANCE FROM G. 6. (nm)

Figure 5 . AP concennation profrle for B at a grain boundary in 316L steel with 4 0 p p m B rapidly cooled (>600 CIS) from 800 'C.

C7-260 JOURNAL DE PHYSIQUE

Figure 3. FIM and lAP micrographs of a grain bouminry in the 316L steel wirh 40 ppmB cooled at 27 "CIS from 1075 'C. The position of the boundary during analysis is indicated with arrows.

Grain boundm urecioitation AlSl 316L .

The same precipitate morphology - plate-like with a

5

maximum dimension 5 to 10 times larger than the

2

thickness - was observed in the 316L steel with 40 ^t

ppm B and in the "Mo-free 316L" steel with 23 ppm B. w

=

Mo-containing steel tetragonal M3B2, MgB3 and M2B z o

were found (Fig. 6 andTable XI) whereas m the W

"Mo-free steel" onhorhombic %B was found (Table 1 I , ^1.1,

11). Furthermore, although the b&on content of the _1.20

Mo-containing steel was almost twice that of the ^I

"Mo-free steel" the rate of precipitatenucleation and a -T0=1250 "C.

growth was much higher for the "Mo-free steel". In the

?

3 16L steel with 40 ppm B also a thin (<2 nm), ⁰ 4

LL ‘ - - T o = 900 "C

continuous layer was found at the boundary after cooling at low and intermediate cooling rates (Fig. 7 ) .

The layer was rich in B, Fe, Mo and Cr and had a .... ...

_

stoichiometry of typically MgB. It was not possible to

$

determine whether the layer was crystalline or not but it a

5

appeared to be amorphous when extracted on carbon - - - - _ _ _ _ _ _ films. No phase with a composition close to M9B has 1.00, ,' , ---__

previously been reported in austenitic steels, neither has lo-' 100 l o 1 102 103 104 1 0 5 the formation of boron rich boundary layers.

COOLING R A T E (OCls) The accumulation of boron at grain boundaries by

non-equilibrium segregation can affect the precipitation

behaviour by changing the boundary composition into Figure 4. (a) Experimental and (b) calculated em'chment other phase fields than those expected from the total factors (the average boron concenhation within boron content For example, in the 316L steel with 50 and 375 nm respectively, frqm the boundary 40 ppm B three different types of borides precipitated. divided by the bulk concentration) for the 316L Judging from the phase diagram (2) two of these would steel with 40 ppm B.

not be expected to form for this boron content. The

variation of the size distribution with cooling rate can also, for both types of steel (Table n), be explained in terms of non-equilibrium segregation. For intermediate cooling rates, which gave the highest boron content of the boundary region, the precipitates differed little in size. However, at the lower cooling rates large variations of precipitate sizes and number densities were found. Thus, factors such as grain boundary orientation could be of greater importance when less boron is available during the precipitate nucleation and growth.

. .

mmsltton of the erain boundarv reeion

In all steels and for all heat Ueatments, also when no precipitation occurred, a boundary zone was found with a composition significantly different from the bulk composition. The width of the zone increased with decreasing cooling rate and increasing starting temperature. However, the differences were most marked at intermediate cooling rates.

Different groups of elements showed markedly different concentration profiles. In the 316L steels Cr and Mo were depleted adjacent to the boundary, Fe, Ni and Mn were enriched adjacent to the boundary, N and Si showed no significant concentration variations with distance and B (and C) was enriched at and adjacent to the boundary (Figs. 2 , 3 and 8).

/.

.-

temp. ("C) ('CIS) (nm)

0.29 M3B2 - tetragonal 30-400

" M9B -layer (2)*'

3 - z

800 2600 M3B2 - tetragonal

" MI3B -layer g 10: 6-1075 'C

. 8-27 " C I S IT

t- 2 . t

$ 5 -

2 .

0 - 0

m 0. ⁸

-

206 ppm 6

I

316L" 1250 31 MZB - orthorhombic 50

* Maximum dimension **Maximum thickness 1

LbFigure 6. TEM (bright$eld) micrograph ofMjlBZ andfaulted

tetragonal M $ l extracted from a grcun boundary in the 316L steel with 40ppm B cooled at 13 "Cls from 1250 'C.

(b) Mo. ( c ) Ni and (d) Fe from the 316L steel cooled at 27 'CIS from 1075 "C (cf. Fig. 3). The full lines show the proji1es as derivedfrom TEM, FIM, h P andIAP data.

Figure 7. FIM micrograph of a brightly imaging grain boundary layer in the 316L steel with 4Oppm B.

The 2 nm thick. continuous layer was formed during cooling at 0.29 'CIS from 1250 "C.

JOURNAL DE PHYSIQUE

The concentration profiles for the "Mo-free steel" were different from the profiles for the Mo-containing steels in that no grain boundary layers were found and in that the Cr-depletion was more pronounced at the higher cooling rates. A zone at the boundary was depleted of Cr andMn, and a depletion of Fe localized to within a few atomic layers at the boundary was also found. The Fe depletion was accompanied by a slight enrichment in the adjacent zone. Some Mo was found at the boundary and indications of a slight enrichment of Ni were also seen. B (and C) was enriched at and adjacent to the boundary. For a more complete description of the precipitation behaviour and elemental distributions see (14).

CONCLUSIONS

Grain boundary segregation of boron and its influence on the precipitation behaviour and elemental distributions in the boundary region have been studied in austenitic stainless steels and nickel using a combination of microanalytical methods (SIMS, TEM, EDS, FIM, AP and IAP) and computer simulations.

The segregation was mainly of the non-equilibrium type in the steels for the higher starting tempera- tures (>800 'C). The preliminary results on nickel suggest that non-equilibrium segregation of boron occurs also in nickel.

Equilibrium segregation dominated in the steels after rapid cooling from 800 'C. The binding energy of boron at austenite boundaries was estimated at 0.65 m.04 eV.

A computer model was developed to simulate non-equilibrium segregation. Good agreement between experimental results and simulations was achieved and diffusion coefficients for boron and vacancy-boron complexes were estimated at :

~ ~1 0 - ~ e x ~ ( - = 2l . l ~ ( e ~ ) / k ~ ) and ~ ~ ~ = 2 . 1 0 - ~ e x ~ ( - l . 15(eV)/kT) (m2/s) respectively.

Tetragonal M2B, M3B2 and M5B3 precipitated at grain boundaries in the 316L steel with 40 ppm B whereas orthorhombic M2B formed at grain boundaries in the "Mo-free steel". In the 316L steel also a thin (12 nm), continuous boundary layer, containing mainly B, Fe, Cr and Mo and with stoichio- metry M9B, formed during cooling at intermediate and low cooling rates.

In all steels and for all heat treatments a boundary zone was found with a composition significantly different from the bulk composition. The differences were most marked at intermediate cooling rates.

Non-equilibrium segregation of boron can affect the precipitation behaviour by changing the boundary composition into phase fields other than those expected from the total boron content.

Boron and molybdenum have a synergistic effect on the precipitation behaviour and elemental distribution at grain boundaries.

-

This work was financially supported by the Swedish Natural Science Research Council (NFR) and the Swedish Board for Technical Development (STU).

REFERENCF3

1. Smithells C.J., "Metals reference book" 5th ed., London, Butterworth, (1976) 98.

2. Goldschmidt H.J., J. Iron Steel Inst., 289 (1971) 900.

3. Suto H. and Sato S., Trans. JIM, 21 (1980) 83.

4. Williams T.M., St0nehamA.M. and Harries D.R., Met. Sci., 10 (1976) 14.

5. Karlsson L., Nord6n H. and Odelius H., Acta Metall., submitted for publication.

6. Karlsson L. and Norden H., Acta metall., submitted for publication.

7. Karlsson L. andNord6n H., Roc. 4th Jap. Inst. of Met. Int. Symp., Minakami Spa (nov-85), accepted for publ.

8. Hondros E.D. and Seah M.P., Int. Met. Rev., 222 (1977) 262.

9. Faulber R.G., J. Mat. Sci., 16 (1981) 373.

10. Karlsson L., Acta Metall., submitted for publication.

11. Karlsson L. and Nord6n H., J. de Phys., 45, C9 (1984) 391.

12. McLean D., "Grain Boundaries in Metals", Oxford, Clarendon Press, (1957) 116 13. Crank J., "Mathematics of Diffusion", Oxford, Univ. Press, (1975).

14. Karlsson L. and Nord6n H., Acta Metall, submitted for publication.