Grain boundary strengthening in austenitic nitrogen steels

https://doi.org/10.1016/S0921-5093(99)00272-5Get rights and content

Abstract

The effect of nitrogen and carbon on the strengthening of the austenitic steel Cr18Ni16Mn10 by grain boundaries is studied. It is established in accordance with previous results that, in contrast to carbon, nitrogen markedly increases the coefficient k in the Hall-Petch equation. Because of a pronounced planar slip induced by nitrogen and the absence of any noticeable segregation of nitrogen atoms at the grain boundaries, nitrogen austenite presents an excellent object for testing different existing models of grain boundary strengthening (pile-up of dislocations, grain boundary dislocation sources, work hardening). Based on the analysis of available data and measurements of interaction between nitrogen or carbon atoms and dislocations it is shown that the nitrogen effect can be attributed to a strong blocking of dislocation sources in grains adjacent to those where the slip started.

Introduction

A contribution to the yield strength from grain boundaries is described by the Hall-Petch equationσ0.20+kyd−1/2

where σ0 is a friction stress, ky is a coefficient characterizing the transfer of slip through the grain boundaries and d is the grain size. There exist three main groups of theories of the Hall-Petch equation: the pile-up models [1], [2], [3], [4], [5], those based on work hardening [6], [7], [8], [9] and the grain boundary source theories [10], [11], [12].

The first one was proposed by Hall [1] and Petch [2] who postulated the formation of dislocation pile-ups against the grain boundaries, causing, in accordance with results [3], [4], a stress concentration at the grain boundary or within it. This pile-up model was later modified by Cottrell [5] who determined a critical stress needed to unpin dislocation sources in the adjacent grain. The main objection against the pile-up model is that planar slip is not a general case. In particular, pile-ups are not observed in metals and alloys with a bcc crystal lattice, but the Hall-Petch equation is nevertheless obeyed.

According to the work hardening model, the dislocation density is increased with a decreasing grain size as d−1 and the term kd−1/2 in the Hall-Petch equation is due to work hardening. This model requires a stress-independent σ0, a parabolic stress-strain relation σ∼√ε and a linear relation between the Hall-Petch slope and the square root of strain, k∼√ε. Such a behaviour was found in niobium [8], but it was not observed in Fe-Co alloys [13]. Moreover, the Hall-Petch slope decreases with strain in copper [14], silver [15], aluminium [16] and iron [17].

The work hardening model has been strongly supported by the theory proposed by Ashby [18] who was the first to introduce a definition of geometrically necessary dislocations stored during deformation of plastically non-homogeneous materials and showed that an increase in their density near the grain boundary with strain causes a grain-size-dependent contribution to the work hardening of polycrystals.

In the grain boundary source theory the Hall-Petch equation is derived from the capacity of grain boundaries to emit dislocations under loading, which does not require a stress concentration created by a pile-up, and in contrast to the work hardening model, the Hall-Petch slope k does not depend on strain. Experimental observations of dislocation loops emitted by grain boundaries [19], of slip lines at grain boundaries [20] and of grain boundary ledges which are supposed to be the dislocation donors [21] are consistent with this theory. An important prediction which follows from the grain boundary source theory is that the capacity of grain boundaries to emit dislocations depends on their structure and composition. The segregation of atoms onto grain boundaries has to affect the Hall-Petch slope k through the stabilization of the ledge structure and an increase in their density. It was shown [22], [23], that ageing of iron increases k, consistent with the grain boundary source theory.

The development of materials with ultrafine-grain size has caused a new wave of studies of the Hall-Petch relation [24], [25], [26]. The pile-up approach was used for the interpretation of experimental data obtained for ultra-fine grain sizes and a remarkable decrease in the Hall-Petch slope was established [24], [25]. The Hall-Petch strengthening was also found to be insignificant for nanophase materials [26].

The grain boundary strengthening of austenitic steels is not very effective. It is half of that in ferrite [27]. In this relation alloying of austenitic steels with nitrogen is of particular interest for two reasons: (i) nitrogen is known to increase the Hall-Petch slope k in the iron austenite [28], [29], [30], [31], [32], [33], [34], and (ii) nitrogen produces some unusual dislocation and twin structures and affects the mechanical properties of austenitic steels markedly.

Norström [28] was the first to show that nitrogen in AISI 316 L type austenitic steels causes, beside a thermal solid solution strengthening, a significant grain size dependent increase in the yield strength which does not depend on the temperature up to 600°C. He has also supposed that a nitrogen-induced increase of the twin density can give rise to an increased Hall-Petch coefficient for a normal grain size.

The data of Norström were confirmed by Degallaix et al. [29] studying austenitic steels of higher nitrogen content. The ‘twin’-hypothesis was also used as an explanation for the nitrogen effect. Werner [30] has studied CrNi and CrMn austenitic steels alloyed with nitrogen up to 0.6 w/o and demonstrated a large increase in the Hall-Petch slope due to nitrogen. Combining this effect with the nitrogen-induced work hardening, Uggowitzer and Speidel [31] have succeeded in obtaining a yield strength of ∼3000 MPa in the austenitic steel Cr18Mn18N0.6.

The ‘twin’ hypothesis was tested by Varin and Kurzydlowski [32] who varied grain size and density of coherent twin boundaries by a thermo-mechanical treatment of type 316 austenitic steels. They have shown that an increasing twin density has no measurable effect on the 0.2% yield strength. Based on the data of Briant [33] concerning the nitrogen segregation at the surface of intergranularily fractured austenitic steels, the authors [32] interpreted their results in terms of grain boundary segregation, i.e. in accordance with the grain boundary source theory. However, according to other data, nitrogen does not seem to have a detectable affinity to grain boundaries either in austenitic [34] or in ferritic steels [35].

An increasing planar slip is considered in Refs. [36] and [37] as a reason for a more effective grain boundary strengthening in nitrogen austenitic steels.

The aim of this paper is to test the available interpretations of the Hall-Petch equation for the yield strength using experimental data obtained on nitrogen austenitic steels. The following pre-conditions suggest this task is feasible: (i) nitrogen promotes planar slip in austenitic steels, (ii) as a rule, it increases work hardening in austenite, (iii) in contrast to carbon, nitrogen does not form a notable grain boundary segregation.

Section snippets

Experimental

The austenitic steel Cr18Ni16Mn10 containing (wt.%) 0.07C, 0.06N, 18.48Cr, 16.13Ni, 9.64Mn, 0.45Si, 0.004S, 0.008P (steel N1 in the following) was chosen for the study. A reason for this choice was its high stability to phase transformations, as was shown in Ref. [38].

Nitrogen was introduced by plasma arc remelting with various pressures of nitrogen in the nitrogen/argon gas mixture (Table 1). During remelting steel N1 in a conventional induction furnace under argon 0.25 wt.% of carbon was

Mechanical tests

The influence of interstitial elements on the yield strength of the studied steels is shown in Fig. 1 in Hall-Petch co-ordinates. Nitrogen increases effectively the grain boundary strengthening, which is consistent with the available data (the data of Norström [28] are presented in the same figure). It is noteworthy that carbon increases the Hall-Petch slope k less effectively than nitrogen.

Cold worked samples were studied in order to evaluate the effect of nitrogen on work hardening. The yield

Discussion

Predictions of different models of the grain boundary strengthening can be evaluated based on the above mentioned experimental data.

First of all a decisive role of planar slip in controlling the Hall-Petch slope is not confirmed. TEM data show that there is no nitrogen-enhanced planar arrays of dislocations in the solution treated steels (Fig. 4). However, when deformation starts, planar slip occurs in nitrogen austenitic steels because of nitrogen-induced short range atomic ordering. It is

Conclusions

Nitrogen in the austenitic steel Cr18Ni16Mn10 increases the coefficient k in the Hall-Petch equation, in accordance with the available data of nitrogen-enhanced grain boundary strengthening. Carbon is shown to cause a smaller increase in the Hall-Petch slope than nitrogen. An unusual effect of nitrogen on the substructure and mechanical properties of austenitic steels provides an opportunity to test different theories of grain boundary strengthening.

The nitrogen-enhanced planar slip during

Acknowledgements

The authors wish to thank the BMBF for financial support (ref.: 03 N 900020).

References (61)

  • H Conrad

    Acta Metall.

    (1963)
  • H Conrad et al.

    Mater. Sci. Eng.

    (1967)
  • R.P Carreker et al.

    Acta Metall.

    (1953)
  • W.E Carrington et al.

    Acta Metall.

    (1965)
  • D.J Dingley et al.

    Acta Metall.

    (1967)
  • E Werner

    Mater. Sci. Eng.

    (1988)
  • R.A Varin et al.

    Mater. Sci. Eng.

    (1988)
  • C.L Briant

    Scripta Metall.

    (1987)
  • Yu.N Petrov et al.

    Scripta Mater.

    (1999)
  • V.G Gavriljuk et al.

    Acta Mater.

    (1998)
  • G Thomas

    Acta Metall.

    (1963)
  • P Müllner et al.

    Mater. Sci. Eng.

    (1993)
  • B.P Kashyap et al.

    Acta Metall. Mater.

    (1995)
  • Yu.N Petrov

    Scripta Metall. Mater.

    (1993)
  • E.O Hall

    Proc. Phys. Soc. London

    (1951)
  • N.J Petch

    J. Iron Steel Inst.

    (1953)
  • R.W Armstrong et al.

    Phil. Mag.

    (1962)
  • J.S Koehler

    Phys. Rev.

    (1952)
  • A.H Cottrell

    Trans. TMS-AIME

    (1958)
  • H Conrad et al.

    Trans. Jpn. Inst. Metals

    (1968)
  • A.A Johnson

    Phil. Mag.

    (1962)
  • J.C.M Li

    Trans. TMS-AIME

    (1963)
  • J.C.M Li et al.

    Metall. Trans.

    (1970)
  • C Crussard

    J. Austr. Inst. Metals

    (1963)
  • M.J Marcinkovski et al.

    Trans. TMS-AIME

    (1965)
  • R.P Carreker

    Trans. TMS-AIME

    (1957)
  • R.P Carreker et al.

    Trans. TMS-AIME

    (1957)
  • Y Ohba

    Trans. Nat. Res. Inst. Metals

    (1967)
  • M Ashby

    Phil. Mag.

    (1970)
  • E Hornbogen

    Trans. ASM

    (1963)
  • Cited by (89)

    View all citing articles on Scopus
    View full text