CONFERENCE ON DIFFUSION IN METALS AND ALLOYS

On May 9–12, 1955, a conference on diffusion in metals and alloys was held in Kiev, convened by the Institute of Metal Physics of the Academy of Sciences of the Ukrainian SSR and the Technical Section of the Scientific Council under the President of the Academy of Sciences of the USSR.

At the conference, the results of investigations were heard and discussed on the theory of atomic diffusion in alloys, the study of diffusion in solid solutions, the relation of diffusion parameters to quantities characterizing interatomic bonding, the development of methods for measuring diffusion parameters and the elasticities of component vapors in metallic alloys, the distribution of components in alloys, diffusion during the sintering of metallic powders, and the use of diffusion phenomena in technological production processes.

Scientists from Moscow, Leningrad, Kiev, Kharkov, Dnepropetrovsk, Sverdlovsk, Stalino, and Tbilisi took part in the conference, as did representatives of the Academy of Sciences of the USSR, the academies of sciences of the Union republics, and various ministries of the USSR.

Altogether, 45 reports and communications were heard at the conference.

In his introductory address, G. V. Kurdyumov characterized in detail the state of the problem of diffusion in metallic systems. The significance of the study of diffusion processes is determined by the role played by the mobility of atoms in resistance to deformation at high temperatures. The direct characteristic reflecting resistance to deformation at high temperatures is the diffusion coefficient, which determines the rate of diffusional displacements of atoms. The diffusion coefficient depends not only on the activation energy (bond energy), but also on the bond energy and its derivatives through the temperature-dependent factor. Thus, the diffusion coefficient is determined by a complex of bond-strength characteristics and depends on the features of the structure and mechanism of diffusional displacements.

In the report by A. A. Smirnov, “Theory of Atomic Diffusion in Alloys” (based on the work of M. A. Krivoglaz and A. A. Smirnov), the results of a theoretical study of the temperature and concentration dependence of the diffusion coefficient and of the influence of phase transitions, as well as of introduced atoms, on the diffusion process were presented. The speaker noted that, contrary to the widespread view according to which the diffusion coefficient in alloys depends exponentially on the reciprocal temperature \(1/T\), the microscopic theory of diffusion leads to the conclusion that a more complex dependence exists, which can be detected when studying diffusion in alloys over a sufficiently wide temperature range. In this connection it proves necessary to characterize diffusion not by a single constant activation energy, but to introduce an effective activation energy for different temperature intervals.

Characteristic features of the curves of temperature dependence should appear upon transition to the ordered state. If ordering

if it is a second-order phase transition, then on the curve of the dependence of \(\ln D\) on \(\frac{1}{T}\) there should be a kink at the transition temperature \(T_0\). If, however, the indicated transition is a first-order phase transition, then at \(T = T_0\) not only the effective activation energy changes discontinuously, but also the diffusion coefficient itself. On the concentration curve of the diffusion coefficient, jumps and kinks should likewise be encountered at those compositions at which the alloy passes into an ordered state (in the case \(T = \mathrm{const}\)).

The theory also makes it possible to explain the sometimes observed strong influence of a small amount of impurity on the diffusion coefficient and on diffusion parameters. If the interaction energies of atoms in an alloy are known, then it proves possible to predict the character of the influence of an impurity on the diffusion coefficient.

The theory developed for the influence of introduced atoms on the self-diffusion of a metal makes it possible to explain, in particular, the influence of a small amount of carbon impurity in gamma iron on the decrease in activation energy and on the experimentally observed decrease in the parameters for the self-diffusion of iron, observed in the experiments of P. L. Gruzin, Yu. V. Kornev, and G. V. Kurdyumov.

The report by P. L. Gruzin was devoted to the question “Study of diffusion and self-diffusion in metals and alloys” and presented the results of work by a group of collaborators. The results of the study of self-diffusion in a number of metals (silver, iron, cobalt, chromium, titanium, and tantalum) made it possible to verify the relationship between the activation energy of self-diffusion and the heat of sublimation, the melting temperature, and the coefficient of linear expansion.

The introduction of carbon into iron and into iron—nickel, iron—chromium, and cobalt—nickel—chromium alloys leads to a substantial acceleration of the diffusion process and to a decrease in its parameters. The introduction of chromium into iron, nickel, and iron—nickel alloys, on the other hand, increases the diffusion parameters at definite concentrations. In some alloys it was shown that the presence of internal grain-boundary surfaces arising as a result of phase transformations can accelerate the diffusion process. This effect is manifested especially clearly in diffusion in steel and in iron alloys in which a martensitic-type phase transformation occurs.

The report noted that the processes of melting and recrystallization are characterized by definite levels of diffusional mobility.

The report by S. T. Kishkin was devoted to the study of the diffusion mechanism of the influence of active media on the properties of metals. In the high-temperature range, the diffusional mobility of atoms in alloys leads to “local” changes in chemical composition, which lead to disruption of the strength of alloys under prolonged action of a static load. The local character of fracture that arises develops in the second and third stages of creep.

The speaker then discussed in detail the mechanism of penetration of liquid metals into the interior of a solid specimen. With the aid of autoradiography it was found that, for example, tin diffuses intensively in nickel along grain boundaries and much more slowly within the grain volume. Boron substantially slows the diffusion of tin along grain boundaries in nickel: at \(700^\circ\mathrm{C}\) the diffusion coefficient decreases from \(0.46 \cdot 10^{-11}\) to \(0.52 \cdot 10^{-12}\ \text{cm}^2/\text{sec}\). It is characteristic that boron also slows the diffusion of tin within the grain volume. It may be supposed that boron substantially improves the structure of the grain boundaries. In accordance with this, the addition of boron is found to increase the high-temperature strength of nickel alloys both in air and in contact with liquid metal.

At \(700^\circ\mathrm{C}\) the diffusion coefficient of tin in nickel along grain boundaries is approximately 100 times greater than in the grains; at \(800^\circ\mathrm{C}\), 10 times greater; and at \(1000^\circ\mathrm{C}\), only 1.5 times greater. Apparently, with increasing temperature and diffusional mobility of atoms, the difference in the structure of the crystal lattice of the grains and of the boundary regions decreases.

S. Z. Bokshtein delivered a report on the distribution and diffusion of components in the volume and along the grain boundaries of metallic alloys. In studying, by the method of autoradiography, the distribution of elements and impurities in iron and nickel alloys, a considerable chemical inhomogeneity was found within a single phase and in individual structural elements. Thus, in nickel, molybdenum, niobium, zirconium, and impurities of tin, antimony, and cerium are found at grain boundaries and in interdendritic regions. Inside the grains are iron and tungsten.

During heat treatment, processes of redistribution of elements take place. The rate of these processes depends on the structure of the alloy, the diffusion mobility, and the initial state of the alloy. During prolonged annealing of an alloy, instead of the normal process of homogenization, heterogenization of the alloy is observed. The nonuniform distribution of elements in an alloy is connected with the peculiarities of the structure of polycrystalline bodies. Of substantial importance is the question of interfaces in metals, in particular the state of grain boundaries and the character of diffusion mobility in these regions. However, this question has been little studied. The author studied the displacement of tin into nickel, cobalt, copper, and iron, and also the diffusion of carbon into $\alpha$-iron. It turned out that diffusion proceeds predominantly along the grain boundaries, and this character of diffusion gradually disappears with increasing temperature. For the diffusion of tin into nickel $(700—1000^\circ)$ the following values of the activation energy were obtained: $52\,000$ cal/m for diffusion in the volume and $21\,000$ cal/m for diffusion along grain boundaries. The high rate of diffusion along grain boundaries is apparently explained by an increased concentration of vacancies.

A report by B. A. Movchan and V. N. Svechnikov, “On equilibrium chemical inhomogeneity in alloys at high temperatures,” was devoted to the same question. Experiments established that in nonferrous alloys prepared from technically pure components, and also in alloyed steels, high-temperature heating produces enrichment by certain elements of the boundary zones of grains and other regions of distorted crystalline structure. This phenomenon occurs in the solid phase below the solidus line ($\sim 200^\circ$ C) and, when the temperature is raised considerably, is accompanied by melting of these regions. Upon subsequent cooling of the alloys, the equilibrium chemical inhomogeneity present at high temperatures contributes to the development of phase inhomogeneity. Both of the indicated factors lead to brittle fracture of the alloy.

M. P. Zheldak reported on a study of the distribution of sulfur and phosphorus in cast tubes by the method of macroautoradiography.

S. S. Nosyreva and A. M. Polyakova described the use of the radioactive isotope sulfur-35 for studying the causes of the occurrence of stone-like fracture in structural steels. The use of isotopes made it possible to study the behavior of sulfides and their distribution as a function of heat treatment and of the cooling rate after overheating. The latter factor strongly affects the formation of stone-like fracture in structural steel.

A number of reports were devoted to methods for measuring the coefficients of diffusion and the elasticity of vapor of components in alloys.

In the report by L. I. Ivanov, I. S. Kulikov, and M. P. Matveeva, a new method developed by them for studying diffusion was described, based on studying the rate of isotopic exchange between two specimens of an alloy of identical composition, taking place in a closed high-vacuum chamber. In the first period of time the reaction rate is determined by evaporation, and then self-diffusion of the labeled component in the radioactive specimen begins to affect the rate of isotopic exchange. From the kinetic curve of accumulation of activity on the inactive specimen one can calculate the vapor pressure and the diffusion coefficient.

Thus, the temperature dependence of vapor elasticity and of the diffusion coefficient for pure iron was studied. The data obtained are very close to those in the literature.

Yu. V. Kornev reported on a method for measuring the vapor elasticity of metals, based on measuring the rate of outflow of saturated vapor through a small opening into a high vacuum. For this purpose radioactive isotopes were used as indicators. In this way the heats of sublimation for a number of metals were determined; the results agree with the literature data.

S. N. Kryukov and A. A. Zhukhovitskii, in their report, described a method for the simultaneous measurement of the diffusion and thermodynamic characteristics of alloys with the use of radioactive isotopes. The method is based on studying the process of exchange between two solid solutions in vacuum. The theory leads to the dependence of the integral activity of the acceptor on the annealing time. At short times this dependence is proportional to time, and at long times to the square root of time. The method was tested on the example of silver and gave good results.

M. E. Yanitskaya and A. A. Zhukhovitskii also reported on a method for determining the thermodynamic characteristics of solutions, in which the measurement of the diffusion flux from the sample is determined by a quantity proportional to the vapor elasticity. In this way the heat of evaporation for silver and the thermodynamic activity for silver–gold alloys were determined. The data agree well with those in the literature.

As is known, the existing methods for determining diffusion coefficients with the aid of radioactive indicators are divided into two groups: a) methods involving a violation of the integrity of the specimen under study (removal of layers, autoradiography), and b) methods not involving destruction of the specimen.

A number of reports at the conference were devoted to these methods. In the report by M. A. Studnits, a new technique was presented for measuring diffusion coefficients, based on the autoradiographic method and making it possible to determine separately the diffusion coefficient in the volume and along grain boundaries, without applying a layer to the surface of the specimen. The error in determining the diffusion coefficients is 20–30%.

V. A. Geodakyan and A. A. Zhukhovitskii described a new “imprint” method for measuring small diffusion coefficients with the use of radioactive isotopes. The method is based on measuring the concentration on the surface of the specimen by means of an “imprint.” The time of diffusion annealing is reduced by a factor of 1000 in comparison with the “layer-removal” technique.

B. T. Borisov, V. I. Golikov, and B. Ya. Lyubov, in the report “On Certain Adsorption Methods for Studying Diffusion in Metals,” dealt with the problem of studying diffusion along grain boundaries. A new scheme for solving the problem was developed, differing from Fisher’s scheme in that it requires knowledge of the thickness of the boundary layer between grains.

In the report by P. L. Gruzin, “On the Question of Methods for Studying Diffusion in Solids Based on the Use of Isotopes,” techniques belonging to the two groups mentioned above were considered, and on the basis of their analysis it was concluded that the most reliable results are obtained by direct measurement of the integral activity of a specimen when thin layers are removed from it. A method was also proposed based on determining the ratio of the radiation intensities of two components, one of which plays the role of a “witness.”

The report by M. A. Krishtal was devoted to a metallographic method for determining the diffusion coefficient of carbon and the self-diffusion of iron in iron phases. The method is based on studying the kinetics of decarburization of white cast irons in a hydrogen atmosphere at high temperature. The thickness of the austenite layer was monitored metallographically. The diffusion coefficients of carbon in the gamma phase were determined. The deviation from the data of Wells and Mehl was 10–12%. The influence of alloying ...

additions on the diffusion coefficient of carbon in the gamma phase. For diffusion of carbon in ferrite the heat of activation proved to be equal to 32,000 cal/m. The coefficients of self-diffusion of iron in the gamma phase turned out to be 3–4 orders of magnitude greater than those obtained by the method of radioactive isotopes.

S. D. Gertsriken and V. G. Lozovik delivered a report on the topic “On Certain Problems of Diffusion.” Certain cases of diffusion were considered, for example that of a finite layer in a semi-infinite specimen, and some methods of processing experimental data were proposed. Tables were compiled that make it possible to read the values of the diffusion coefficient from experimentally found concentrations, or amounts of diffusing substance.

I. L. Mirkin spoke with the report “On the Mechanism of Diffusion in Metals in the Solid State.” The theory of the diffusion mechanism by means of vacancies must take into account additionally that: 1) the energy of formation of a defect in the lattice also includes the change of state in a certain volume surrounding the vacancy, and 2) displacement in the lattice is the result of the cooperative motion of several nearest atoms. A group of atoms with an altered mutual arrangement and reduced density may be likened to a region of “liquid” inside a solid body. This group may be called a “relaxation.” It may be shown that the ratio of the diffusion coefficients in the solid and liquid states at a given temperature is the fraction of the total number of atoms participating in relaxations at this temperature. It turned out that the change in volume for creating a relaxation, calculated from self-diffusion data, is close to the change in volume on melting of the metal. Moreover, as the hydrostatic pressure is increased to 12,000 atmospheres, both quantities decrease in exactly the same way. The general conclusion is that the diffusion process is a consequence of the cooperative displacement of a group of atoms that successively pass into a state analogous to the liquid state. This idea should promote further, more profound study of the nature of a solid metal and its properties.

At the meeting, reports were heard in which the results of general investigations of diffusion in alloys and chemical compounds were presented.

S. D. Gertsriken and I. Ya. Dekhtyar reported on the results of studying cobalt diffusion in some of its alloys. Alloys of the systems cobalt—aluminum, cobalt—iron, and cobalt—nickel—manganese were studied. The report presented data showing that the greater the energy of interatomic bonding, the greater the activation energy of cobalt diffusion in cobalt–aluminum alloys. In this system the activation energy increases up to about 50 at.% Al, and then falls sharply, which is explained by the presence of a large excess number of structural vacancies.

For the alloy Co + Fe (50%) a sharp increase in the diffusion coefficient is observed in the transition from the γ- to the α-region. The activation energy for this alloy proved to be less than the activation energy of self-diffusion of the pure components. This is explained by a decrease, for this alloy, in the filling coefficient of \(d\)-vacancies, which causes a decrease in the bond energy.

I. N. Frantsevich and D. F. Kalinovich spoke about the transfer of carbon in gamma iron under the influence of a constant current, using radioactive isotopes. It was experimentally proved, contrary to Drakin’s opinion, that all the carbon dissolved in γ-iron participates in electrotransfer, which indicates the absence in the alloy of non-ionized carbon atoms.

Participation in the transfer of electricity to the anode by iron ions was shown indirectly.

A report by B. M. Noskov was devoted to the question of the influence of alloy components on diffusion constants. Proceeding from his own and literature data on the study of diffusion in alloys based on iron, cobalt, and nickel, the speaker came to the conclusion that the activation energy is located

in functional relation to the bond energy of the crystal lattice. The data obtained indicate that manganese and chromium increase the energy of interatomic bonds in the iron lattice, whereas carbon decreases it. In some cases the action of alloy components may be considered additive. Thus, the strengthening effect of chromium on interatomic bonds in iron may to some extent be compensated by the weakening effect of carbon.

The available experimental data appear insufficient for a quantitative theory of the interaction of atoms in an alloy. Qualitatively, it may be supposed that the atoms of some electron-donor element give up part of their electrons to another atom of a transition element, and the donor atom may interact with many atoms of its nearest environment. The change in activation energy with increasing concentration of the alloying element is considered in connection with the change in the ratio of heteronymous and homonymous bonds.

The speaker considers it expedient to study the concentration dependence of self-diffusion in various systems over a wide range of concentrations in the absence of a concentration gradient.

P. L. Gruzin and G. B. Fedorov reported on work studying the diffusion of chromium in solid solutions of nickel. It was found that the addition of 20% chromium to nickel substantially slows the diffusion of chromium in the temperature region below 900°, while the activation energy increases by 20%. The addition of 2.5% titanium to nichrome leads to a further increase in the activation energy of chromium diffusion. The data obtained are in agreement with the results of studying interatomic interaction in alloys by G. V. Kurdyumov and N. T. Travina.

The presentation by M. G. Lozinskii was devoted to the influence of compressive stresses on the rate of diffusion of carbon in alpha- and gamma-iron. The experiments performed made it possible to detect a decrease in the width of the zone of diffusion of carbon from steel (1.48% C) into Armco iron as the compressive stresses increased (0.15, 1.5, and 15 kg/mm²).

B. M. Noskov, E. V. Kuznetsov, and G. V. Shcherbedinskii presented a report on the influence of internal boundaries of a separation on the coefficient of self-diffusion of iron in iron–nickel–carbon alloys. The investigation was carried out with specimens of an alloy Fe + 25% Ni + 0.9% C. Preliminary experiments showed that the boundaries of martensitic crystals are retained at temperatures above the point of transformation of martensite into austenite. These boundaries may be diffusional pathways of accelerated diffusion. The data obtained indicate that at 900–1000° the diffusion coefficient of Fe increases by approximately 2–3 times.

I. Ya. Dekhtyar gave a report on the topic “The influence of the concentration of a third element on the process of diffusion of manganese in nickel–manganese alloys.” Nickel–manganese alloys with various contents of copper, silicon, titanium, and tin were investigated, using the vacuum-evaporation method for studying the diffusion of manganese. It was shown that the assumption of a monotonic dependence of the activation energy on small concentrations of alloying elements is not always valid. At 8 at.% Si, 4 at.% Sn, and 2 at.% Ti in the corresponding systems, maxima were found on the curves of the dependence of the activation energy on the concentration of the third element. For Ni–Mn–Si alloys, the maximum of the activation energy of manganese diffusion corresponds to a minimum on the curve of the change in the lattice parameter as a function of silicon concentration. For Ni–Mn–Sn alloys, the maximum of the activation energy corresponds to a minimum in the change of the galvanomagnetic effect as a function of tin content. The assumption is advanced that the change in the activation energy, as well as in other characteristics, is caused by a change in the structural state of the alloy, associated with the formation of certain regions of short-range order.

S. D. Gertsriken gave a report on the topic “Determination of the parameters of hole formation.” The energy of hole formation for pure metals

determined from the curves of the dependence of the coefficient of expansion or of the electrical resistance on temperature in a wide interval. For Al, Ag, Sn, and Cu, the energy of hole formation is approximately one third of the activation energy of self-diffusion.

A. G. Lesnik and N. P. Plotnikova presented a report on the influence of heat treatment on the diffusion of chromium in iron–chromium alloys.

Diffusion in alloys with chromium contents (up to 24%) was studied by the method of vacuum evaporation in the temperature range 980–1100° C. It was shown that the activation energy of chromium diffusion increases with the time of preliminary annealing of the alloys at 1200° C, reaching a certain maximum value. This effect is explained by the possibility that a process of short-range ordering occurs at 1200° C.

A. Ya. Shinyaev, in his presentation, spoke about the diffusion of iron in iron–nickel and iron–molybdenum alloys. The activation energy of diffusion for a solid solution changes continuously with concentration. In the presence of a chemical compound (for Fe—Mo alloys), the activation energy attains a maximum value.

F. A. Santalov spoke about pore formation during the distillation of a volatile component from solid solutions. Alloys Ag—Zn and Ag—Cd with additions of 0.2% tin or lead were studied. Zinc or cadmium was distilled off at 650° C and a pressure of \(10^{-4}\) mm Hg. The study showed that tin lowers the rate of distillation of zinc from a silver–zinc alloy, as well as the total porosity. In the opposite direction acts an addition of tin in silver–cadmium alloys. Lead lowers the rate of distillation of zinc and cadmium from alloys based on silver and reduces the total porosity of the specimen. In connection with porosity and shrinkage of the specimen, it was pointed out that these phenomena cannot be ignored in studying diffusion by the method of evaporation in vacuum.

Speaking on this question, I. Ya. Dekhtyar noted that the question of pore formation has been studied in a number of extensive investigations. In particular, Buffi’s work gives data showing the connection of pore formation with the mechanism of diffusion by means of vacancies, and that the diffusion coefficients measured by different methods for the same objects vary within the limits of 10–20%. It has been convincingly shown that pore formation is a phenomenon accompanying the diffusion process under the most diverse methods of investigation, and is not characteristic only of the method of evaporation in vacuum.

F. N. Tavadze and E. S. Kartozia, in their presentation on the question of the diffusion of magnesium in iron and its alloys, noted the importance of studying the diffusion of elements considered insoluble in iron, with the aim of obtaining alloys with new properties (magnesium cast iron).

A number of reports were devoted to the study of diffusion processes in the sintering of metals.

I. M. Fedorchenko gave a report on the topic: “On the question of surface diffusion during the heating of metallic powders.” The sintering of contacting particles is considered in connection with diffusion processes occurring in the surface layers. The change in the specific surface area of a powder is directly related to the coefficient of surface diffusion. Processing the experimental data for iron powder with the aid of the formula obtained showed that the activation energy of the process of restructuring the surface layer of the particles is equal to 16,000 cal/gram-atom, which is several times less than the activation energy of volume diffusion.

In the report of Ya. E. Geguzin, “On the sintering of mixtures of metallic powders (Cu—Ni),” formulas were proposed describing the concentration dependence of the linear shrinkage of compacts. The temperature and time dependences of linear shrinkage due to the diffusion process are discussed.

In another communication, Ya. E. Geguzin spoke about the study of the early stage of diffusion creep in connection with the kinetics of the removal of distortions.

A connection was established between the kinetics of elongation of threads made of lead-tin alloys subjected to various degrees of deformation and the kinetics of stress relaxation.

R. I. Garber, V. M. Mikhailovskii, and L. M. Polyakov reported on a study of the strength of the bond between metal blocks by means of adhesion. It was shown that the determining condition for the formation of a strong bond is the pressure on the contact surface.

The report by Ya. E. Geguzin and M. N. Ovcharenko was devoted to the study of processes occurring on the surface and in the near-surface layer of a metal at high temperature in vacuum. The process by which roughness arises on the surface of a polished section was described, and a proposal was made concerning the nucleation mechanism of its formation.

I. M. Choporov spoke about the results of investigations of the diffusion of tungsten in titanium carbide in a mixture of tungsten and titanium carbide powders. The parameters of tungsten diffusion in titanium carbide in a mixture of powders were determined (by the method of S. D. Gertsriken and M. A. Faingold).

A number of reports were devoted to the question of using radioactive isotopes to study technological production processes, as well as to the use of diffusion processes in industry.

G. A. Blokh, V. Ya. Demidionova, G. I. Miklukhin, I. I. Kukhtenko, A. F. Rekashova, R. V. Nikulina, M. P. Przhebyl’skii, S. B. Kalik, G. I. Laevskaya, E. Tsypenyuk, and E. Reznichenko reported on the study of diffusion processes in the iron industry by means of radioactive isotopes. T. A. Potapova reported on the diffusion of Ca-45 into nepheline concentrate and on the effect on it of small additions of alkaline and phosphorus-containing compounds. R. I. Mishkevich spoke about low-temperature nitrogenizing of structural alloy steel. Ya. P. Aleshin gave a report, “New Developments in the Technology of Siliciding Steel.” M. V. Lavrov and A. A. Ladygina reported on thermodiffusion chromium plating of certain grades of chromium-nickel-tungsten steel and high-strength complex-alloyed ferrite. A. I. Chizhikov spoke about the rate of equalization of the concentration of certain elements in the bath of an open-hearth furnace.

In the resolution adopted by the Conference it is noted that, in the field of the study of diffusion and interatomic interaction, a number of successes have been achieved, but many questions have still not found their proper place in research. These include, first of all, the study of diffusion in refractory systems and in liquid metals, clarification of the mechanism of atomic displacements in metals, the use of autoradiography and X-ray spectrography methods for investigating chemical microinhomogeneity in alloys, and problems of diffusion theory.

Noting the fruitfulness of the direction of work of the various participants, the Conference drew attention to the insufficient application of the results of diffusion studies for the rationalization of a number of technological processes in production, and for concrete problems of metal physics.

I. Ya. Dekhtyar