Full Text
STUDY OF THE ATOMIC STRUCTURE OF AGING ALLOYS
Understanding the mechanism of aging of alloys (i.e., the spontaneous, or artificially induced by annealing, improvement of their properties) is of great importance for practical and theoretical physical metallurgy. Work on the study of aging alloys is being carried out widely both in the Soviet Union and abroad, and a large number of articles is devoted to this question. Research-
study of the mechanism, i.e., of changes in the atomic structure during aging, is carried out most effectively by methods of X-ray structural analysis, in particular by methods modified for solving the given problem. In a recently published work by Geisler and Hill*) a brief survey is given of the methods and results of a large number of X-ray studies of the aging of aluminum alloys, and new experimental data are also presented concerning the alloys Al—Ag and Al—Mg—Si. The authors draw a number of conclusions which, in their opinion, are of a general character.
As is known, the process of aging of an aluminum alloy is regarded as a process of decomposition of a supersaturated solid solution of the alloying component in aluminum; the cause of the decomposition is the decrease in the solubility of the component when the temperature is lowered. If at some temperature (usually 450–500°C) an alloy with a definite percentage content of the alloying component is an equilibrium solid solution of substitution, then after quenching from this temperature to a lower one (20°—−200°C) the solid solution will no longer be a stable phase. During annealing, certain structural changes will occur in it, as a result of which the alloy will possess an atomic structure that is equilibrium or metastable for the given temperature. (This process of structural change is accompanied by a change in the mechanical properties of the alloy, for the improvement of which new alloy compositions are usually created purely empirically.)
The course of the structural change during isothermal annealing is apparently the same for the entire range of annealing temperatures, and differs only in the rate of occurrence. At room temperature the process is so slow that it practically stops at the initial stage, which at higher temperatures (200–300°C) may pass almost unnoticed. The later stages of aging are more favorable for X-ray study and for this reason have been investigated more fully.
The initial stage of aging appears in the form of effects of two-dimensional diffraction in photographs obtained from single crystals of alloys. These diffraction patterns may be interpreted in such a way that very thin (several atomic layers) flat plates of unknown structure are formed in the crystal of the solid solution, the so-called “Guinier–Preston zones” (abbreviated G.-P. zones), closely connected with the crystal lattice of the original solid solution.
In all its complexity the course of the aging process may be described by the following scheme¹: matrix (supersaturated solid solution) → matrix + G.-P. zones → depleted matrix + metastable precipitate phase → → equilibrium matrix + stable precipitate phase.
As has been shown for a number of alloys²—⁴, the metastable phase is characterized by being structurally organically connected with the matrix (“coherent” with it), i.e. the matrix passes into a particle of this phase continuously, which is possible by virtue of the identity of the adjoining atomic planes of the two lattices. The stresses in the matrix and in the particles, caused by such a forced matching of lattices, are precisely what are responsible for the increased hardness of the aged alloy. As a result of prolonged annealing or at a still higher temperature, the particles of the metastable phase grow and “break away” from the matrix, turning into particles of the stable phase of the same composition, but with their own structure; in this process the stresses are relieved and the mechanical properties deteriorate.
Most fully in this respect the systems Al—Cu and Al—Ag have been studied ⁵˒²˒³˒⁶. In the first case the stable precipitate phase is the compound CuAl₂ with a tetragonal lattice (the so-called θ phase); its metastable
) A. H. Geisler and J. K. Hill, Acta Crystallographica 1*, 238 (1948).
FROM CURRENT LITERATURE
the form (phase $\theta$) has a fluorite lattice. The coincidence of atomic planes is achieved on the $(100)$ planes of the matrix, owing to which the precipitates of the phase have the form of plates parallel to these planes. In the Al—Ag alloy, pure silver is precipitated first in the form of a coherent
Fig. 1. Part of the reciprocal lattice for a crystal of an aged Al—Ag alloy. Double strokes are rods corresponding to the planes of the G.-P. zones; open circles are points of the reciprocal lattice of the $\gamma'$ phase.
metastable phase (phase $\gamma'$) with a hexagonal lattice and then of the $\gamma$ phase, having the normal cubic structure of silver. Here the “fitting” proceeds along atomic planes of the type $(111)$. The transformation into the stable form does not disrupt the crystallographic orientation of the precipitate plates relative to the matrix, which is detected in microsections in the form of Widmanstätten structures.
The nature of G.-P. zones is clear to a far lesser degree. The subject of persistent discussion up to the present time remains the question whether they are nuclei of a metastable phase with its own lattice, or represent merely regions of redistribution of the atoms of the alloying element over the sites of the matrix. A number of American authors 7,3, including the authors of the paper under review, support the first interpretation, basing themselves chiefly on the fact that in Al—Cu and Al—Ag alloys G.-P. zones appear along the same atomic planes as the subsequent precipitates of the phases $\delta$ and $\gamma'$ in amounts already sufficient to produce three-dimensional diffraction.
In the present paper the authors consider that, on the basis of comparison and interpretation of diffraction patterns, they have shown sufficiently convincingly that Guinier–Preston zones are interlayers of a metastable phase (with its own crystal lattice), having only a small thickness insufficient to produce their own three-dimensional diffraction. This is proved by the fact that the reciprocal-lattice points of the phase $\gamma'$ in an Al—Ag alloy lie exactly on rods in reciprocal space from two-dimensional G.-P. zones.*) During the aging process the zones grow into thick interlayers of the phase $\gamma'$, and the rods in reciprocal space then contract into points (Fig. 1). The fact that, in reciprocal space for Al—Ag, as also for Al—Cu, the rods pass through the reciprocal-lattice nodes for the matrix is, in the authors’ opinion, not evidence of a direct relation of the zones to the matrix, but is explained only by the structural similarity of the two lattices; this is confirmed by the absence of such a correlation in the case of an Al—Mg—Si alloy. This conclusion, however, does not explain the fact that in the case of an Al—Cu alloy only those interferences from two-dimensional formations (zones) are observed which are permitted by the structure of the matrix itself, and that no superstructure interferences 8 are observed, which should have been present if the pattern were produced by thin interlayers of a metastable phase $\theta'$, having the fluorite lattice.
The second substantial result of the authors’ work is their discovery of an even earlier stage of aging preceding the stage of appearance of G.-P. zones. If the G.-P. zones appear in the form of two-dimensional diffraction effects, then this earlier stage produces patterns of one-dimensional diffraction, which may be explained by the fact that, before the formation of plate-like zones in the alloy, precipitate particles first form which in only one dimension have a magnitude sufficient to produce distinct diffraction.
The authors believe that at this stage the metastable phase $\gamma'$ precipitates in the form of needles with only one sufficiently large dimension ($>10a_{\mathrm{Al}}$). Such “one-dimensional” formations correspond in reciprocal space not to rods, but to considerable portions of a plane, which, in photographs taken with monochromatic radiation, should be revealed by the appearance of streaks in those places where, at a later stage, spots from two-dimensional diffraction by G.-P. zones appear. The authors prove this by a whole series of photographs, from which they calculate the direction of the needles in the direct lattice. For the Al—Ag alloy they give 6 directions of the type $[011]$.
The stage of one-dimensional formations is also found by the authors in the Al—Mg—Si alloy (1.4% Mg$_2$Si). They also trace the transition from this stage
*) The authors analyze in detail the geometry of the transition from X-ray diffraction patterns to reciprocal (diffraction) space and from reciprocal space to the direct space of the crystal lattice of the matrix. A number of formulas are given and a simplified method is proposed for constructing the reciprocal-space pattern from X-ray photographs.
to the zone stage. In the photographs reproduced by us (Fig. 2) it is evident how the streaks near the 111 spots (marked by an arrow) turn into separate points upon transition to a higher aging temperature. The direction of the planes in reciprocal space is defined as (001), which in the direct lattice corresponds to the direction of the needles along the three directions of the type \([00\bar{1}]\). After aging at a higher temperature (300° C) these needles turn into plates (G.-P. zones), located on matrix planes of type (001), (011), and, possibly, also (012) and (013), or (014). The authors indicate that, as for Al—Ag, these plates are nuclei of a metastable coherent phase with the matrix, the structure of which they have determined, but is not given in the present work.
With further annealing, the metastable phase changes into the stable phase \(Mg_2Si\) with a fluorite-type lattice, the particles of which are arranged with respect to the matrix according to the following law:
plane (100) \(Mg_2Si \parallel\) (100) Al;
direction [110] \(Mg_2Si \parallel\) [100] Al.
From the data obtained, the authors estimate the order of magnitude of the particles precipitated in both initial stages. For the needle-like stage they give a particle size of \(10 \times 20 \times\) more than 100 Å, and in the following plate-like, two-dimensional stage: \(15 \times\) more than \(100 \times\) more than 100 Å; a schematic drawing explaining the development of plates of different directions from needles of only one crystallographic direction is given in Fig. 3.
It should be noted that, although at the beginning of the paper the authors discuss in detail possible methods for X-ray observation of aging effects, they themselves do not use the most effective method. Using the method of a stationary single crystal of the alloy, they work not with purely monochromatic radiation, but with mixed radiation, which complicates the calculation and somewhat hampers the interpretation of the photographs. Therefore, in the original part of the work they have to devote a special chapter to an analysis of possible diffraction patterns in the case of the presence of a whole spectrum in the radiation used. For these same reasons, one must be cautious about the authors’ interpretation of certain new diffraction effects observed by them. It seems to us that the explanation of these effects may be recognized as satisfac-
Fig. 2. Photographs of a single crystal of the alloy Al·Mg·Si under two aging conditions:
a) 4 hours at 220° C — streaks corresponding to the needle-like stage;
b) 0.5 hour at 300° C — points in place of the streaks, corresponding to the plate-like stage (G.-P. zones).
...only after obtaining them in a purer form, with monochromatized radiation*).
At the end of the article the authors draw several general conclusions concerning the aging mechanism of aluminum alloys. They believe that all particles in all systems grow from nuclei of submicroscopic dimensions, and that the stage of appearance of needle-like (one-dimensional) formations and the stage of G.-P. zones (two-dimensional formations) are only intermediate stages in the sequential growth of the nuclei of a coherent metastable phase up to the stage at which particles of this phase are already capable of producing three-dimensional diffraction patterns. The idea of such a sequence in particle growth is theoretically substantiated by them on the basis of ordinary diffusion concepts, similarly to the predominant growth of needles in the usual precipitation process. In the case under consideration, such a sequence should be carried out even more rigorously because of the necessary coherence of the precipitates with respect to the matrix. It is possible that the difference in diffraction patterns obtained with monochromatic radiation in the Al--Cu--Mg alloy after natural aging and aging at \(218^\circ\)C may be explained by the presence, in the first case, of only needle-like precipitates, and in the second also of plate-like precipitates; this is expressed in the appearance of sharp spots in the photograph from the crystal annealed at \(218^\circ\)C, against a background of blurred streaks characteristic of the naturally aged crystal\(^8\).
Fig. 3. Interpretation of the structure of an aged Al--Mg--Si alloy.
On the left, the arrangement of precipitate particles in the alloy crystal; on the right, images in the reciprocal lattice corresponding to the particles.
Needle-like stage
Plate-like stage
The needles grow into plates parallel to the planes \((100)\) and \((120)\) of the matrix.
In any case, the idea advanced by the authors appears very fruitful in constructing a general theory of the aging mechanism, although it must still be tested and proved on a large number of alloys.
Yu. Bagaryatskii
Cited Literature
- N. K. Hardy, Light Metals, Nos. 7, 8 (1944).
- G. D. Preston, Proc. Roy. Soc. A 167, 526 (1938).
- Barrett, Geisler, Mehl, Transac. A. I. M. E. 143, 134 (1941).
- Geisler, Barrett, Mehl, Transac. A. I. M. E. 152, 201 (1943).
- Wassermann and Weerts, Metallwirtschaft, 14, 605 (1930).
- Geisler, Barrett, Mehl, Transac. A. I. M. E. 152, 182 (1943).
- Collection Age Hardening of Metals, 1940, Oxford; articles by Fink and Smith, Mehl and Jetter.
- Yu. A. Bagaryatskii, Journal of Technical Physics 18, 827 (1948).
*) The method of investigating crystals of aging alloys with purely monochromatic radiation is used by Soviet researchers (Research Institute of Physics, Moscow State University), as well as by the French Guinier school.