Anomalies in Deformation Behaviour of TiAl Intermetallic

An explanation was proposed for a nonmonotonic temperature dependence of the yield stress σy(T) in TiAl having two extrema where σy(T) changes its temperature behaviour.

Збережено в:

Бібліографічні деталі
Дата:	2000
Автори:	Greenberg, B.A., Ivanov, M.A.
Формат:	Стаття
Мова:	English
Опубліковано:	Інститут металофізики ім. Г.В. Курдюмова НАН України 2000
Назва видання:	Успехи физики металлов
Онлайн доступ:	http://dspace.nbuv.gov.ua/handle/123456789/133373
Теги:	Додати тег Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:	Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:	Anomalies in Deformation Behaviour of TiAl Intermetallic / B.A. Greenberg, M.A. Ivanov // Успехи физики металлов. — 2000. — Т. 1, № 1. — С. 9-48. — Бібліогр.: 56 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine

id	irk-123456789-133373
record_format	dspace
spelling	irk-123456789-1333732018-05-25T03:03:27Z Anomalies in Deformation Behaviour of TiAl Intermetallic Greenberg, B.A. Ivanov, M.A. An explanation was proposed for a nonmonotonic temperature dependence of the yield stress σy(T) in TiAl having two extrema where σy(T) changes its temperature behaviour. Предложено объяснение немонотонной температурной зависимости предела текучести σy(T) в TiAl, где имеются две экстремальные точки, в которых происходит смена характера температурной зависимости σy(T). Запропоновано пояснення немонотонної температурної залежності межі плинності σy(T) в TiAl, де мають місце два екстремуми, в яких відбувається зміна характеру температурної залежності σy(T) . 2000 Article Anomalies in Deformation Behaviour of TiAl Intermetallic / B.A. Greenberg, M.A. Ivanov // Успехи физики металлов. — 2000. — Т. 1, № 1. — С. 9-48. — Бібліогр.: 56 назв. — англ. 1608-1021 PACS: 61.72.Bb, 61.72.Ff, 61.72.Lk, 62.20.Fe, 81.40.Lm DOI: https://doi.org/10.15407/ufm.01.01.009 http://dspace.nbuv.gov.ua/handle/123456789/133373 en Успехи физики металлов Інститут металофізики ім. Г.В. Курдюмова НАН України
institution	Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection	DSpace DC
language	English
description	An explanation was proposed for a nonmonotonic temperature dependence of the yield stress σy(T) in TiAl having two extrema where σy(T) changes its temperature behaviour.
format	Article
author	Greenberg, B.A. Ivanov, M.A.
spellingShingle	Greenberg, B.A. Ivanov, M.A. Anomalies in Deformation Behaviour of TiAl Intermetallic Успехи физики металлов
author_facet	Greenberg, B.A. Ivanov, M.A.
author_sort	Greenberg, B.A.
title	Anomalies in Deformation Behaviour of TiAl Intermetallic
title_short	Anomalies in Deformation Behaviour of TiAl Intermetallic
title_full	Anomalies in Deformation Behaviour of TiAl Intermetallic
title_fullStr	Anomalies in Deformation Behaviour of TiAl Intermetallic
title_full_unstemmed	Anomalies in Deformation Behaviour of TiAl Intermetallic
title_sort	anomalies in deformation behaviour of tial intermetallic
publisher	Інститут металофізики ім. Г.В. Курдюмова НАН України
publishDate	2000
url	http://dspace.nbuv.gov.ua/handle/123456789/133373
citation_txt	Anomalies in Deformation Behaviour of TiAl Intermetallic / B.A. Greenberg, M.A. Ivanov // Успехи физики металлов. — 2000. — Т. 1, № 1. — С. 9-48. — Бібліогр.: 56 назв. — англ.
series	Успехи физики металлов
work_keys_str_mv	AT greenbergba anomaliesindeformationbehaviouroftialintermetallic AT ivanovma anomaliesindeformationbehaviouroftialintermetallic
first_indexed	2025-07-09T18:52:43Z
last_indexed	2025-07-09T18:52:43Z
_version_	1837196549961023488
fulltext	9 МЕХАНИЧЕСКИЕ, ТЕПЛОВЫЕ И КИНЕТИЧЕСКИЕ СВОЙСТВА PACS numbers: 61.72.Bb, 61.72.Ff, 61.72.Lk, 62.20.Fe, 81.40.Lm Anomalies in Deformation Behaviour of TiAl Intermetallic B. A. Greenberg and M. A. Ivanov * Institute of Physics of Metals, Ural Division, Russian Academy of Sciences, 18 Sophia Kovalevskaya Str., 620219 Ekaterinburg, Russia * G. V. Kurdyumov Institute for Metal Physics, N.A.S. of the Ukraine, 36 Academician Vernadsky Blvd., UA -03680 Kyyiv-142, Ukraine An explanation was proposed for a nonmonotonic temperature dependence of the yield stress y(T) in TiAl having two extrema where y(T) changes its tem- perature behaviour. The comparison of y(T) curves for TiAl and typical curves for other materials (b.c.c. metals, semiconductors, Ni3Al-type intermetallics) allowed reconstructing the shape of the potential relief for dislocations in TiAl. The shape of the relief reflects existence of two types of dislocation traps (shallow and deep ones) and two types of potential barriers: low and high bar- riers for the capture of dislocations in shallow and deep traps respectively. The deformation behaviour of TiAl was described over the whole temperature inter- val allowing for the capture of dislocations in traps and their release therefrom. Expressions determining extremums of y(T) were derived. Possible depend- ences of the work hardening rate (T) in the region of the anomalous trend y(T) were analysed. Conditions of the anomalous behaviour of (T) were as- certained. It was proposed that the shape of the potential relief of a d islocation changed in the effective range of a microcrack. The capture of dislocations in deep traps, which is stimulated by concentration of stresses near a micro- crack, and the inhibited release of dislocations from the traps up to relatively high temperatures are viewed as a possible cause of TiAl brittleness. A model of the deformation behaviour of intermetallics after prestraining was proposed. This model enables one to describe two alternatives: observation or absence of the stress macrojump after prestraining. Предложено объяснение немонотонной температурной зависимости пре- дела текучести y(T) в TiAl, где имеются две экстремальные точки, в кото- рых происходит смена характера температурной зависимости y(T). Срав- нение кривых y(T) для TiAl и типичных кривых для других материалов (ОЦК металлов, интерметаллидов типа Ni3Al) позволяет реконструировать форму потенциального рельефа для дислокаций в TiAl. Форма полученного рельефа отражает существование двух типов дислокационных ловушек (мелких и глубоких) и двух типов потенциальных барьеров: низких и высо- Успехи физ. мет. / Usp. Fiz. Met. 2000, т. 1, сс. 9–48 Îòòèñêè äîñòóïíû íåïîñðåäñòâåííî îò èçäàòåëÿ Ôîòîêîïèðîâàíèå ðàçðåøåíî òîëüêî â ñîîòâåòñòâèè ñ ëèöåíçèåé 2000 ÈÌÔ (Èíñòèòóò ìåòàëëîôèçèêè èì. Ã. Â. Êóðäþìîâà ÍÀÍ Óêðàèíû) Íàïå÷àòàíî â Óêðàèíå. 10 B. A. Greenberg and M. A. Ivanov ких барьеров для захвата дислокаций в случаях мелких и глубоких лову- шек соответственно. С учетом захвата дислокаций в ловушки и выхода из них удалось описать характер деформационного поведения TiAl во всем интервале температур. Получены выражения, которые определяют экс- тремальные точки на кривой y(T). Проанализированы возможные виды зависимости коэффициента упрочнения (T) в области аномальной тем- пературной зависимости y(T). Установлены условия, при которых возни- кает аномальная температурная зависимость (T). Высказано предполо- жение, что форма потенциального рельефа для дислокаций существенно меняется в некоторой области вблизи микротрещины. Захват дислокаций в глубокие ловушки, который стимулируется концентрацией напряжения вблизи микротрещины, и невозможность им покинуть ловушки вплоть до относительно высоких температур рассматривается как возможная при- чина хрупкости TiAl. Предложена модель, которая описывает деформаци- онное поведение интерметаллидов после некоторой предварительной деформации. Эта модель позволяет описать две альтернативные воз- можности: наблюдение или отсутствие макроскачка напряжения после указанной деформации. Запропоновано пояснення немонотонної температурної залежності межі плинності y(T) в TiAl, де мають місце два екстремуми, в яких відбува- ється зміна характеру температурної залежності y(T). Порівняння кри- вих y(T) для TiAl та типових кривих для інших матеріалів (ОЦК металів, інтерметалідів типу Ni3Al) дозволяє реконструювати форму потенціаль- ного рельєфу для дислокацій в TiAl. Форма одержаного рельєфу від - творює існування двох типів дислокаційних пасток (мілких та глибоких) та двох типів потенціальних бар’єрів: низьких та високих для захоплення дислокацій відповідно у випадках мілких та глибоких пасток. З урахуван- ням захоплення дислокацій у пастки та виходу з них вдалося описати характер деформаційної поведінки TiAl в усьому інтервалі температур. Отримано вирази, які дозволяють знайти екстремуми на кривій y(T). Проаналізовано можливі типи залежності коефіцієнта зміцнення (T) в області аномальної температурної залежності y(T). Встановлено умови, за яких виникає аномальна температурна залежність (T). Висловлено припущення, що форма потенціального рельєфу для дислокацій суттєво змінюється в деякій області поблизу мікротріщини. Захоплення дислока- цій в глибокі пастки, яке стимулюється концентрацією напруги поблизу мікротріщини, та неможливість їм покинути ці пастки аж до відносно ви- соких температур розглядається як можлива причина крихкості TiAl. За- пропоновано модель, яка описує деформаційну поведінку інтерметалідів після деякої попередньої деформації. Ця модель дозволяє описати дві альтернативні можливості: виникнення або відсутність макрострибка на- пруги після вказаної деформації. Key words: TiAl, mechanical properties, yield stress, work hardening rate, flow stress anomaly, brittleness, microstructure, dislocation, Peierls relief, thermally activated blocking. (Received February 3, 2000) Anomalies in Deformation Behaviour of TiAl Intermetallic 11 CONTENTS 1. Introduction 2. Typical y(T) Curves for Various Materials and Transitions Between Dislocation States 3. Description of a Nonmonotonic Temperature Dependence y(T) for TiAl 3.1. Low-Temperature Drop of y(T) 3.2. Anomalous Temperature Trend of y(T) 3.2.1. Thermally activated blocking of dislocation sources 3.2.2. Temperature dependence of the work hardening rate 3.3. Drop of y(T) above the Temperature Peak 3.4. Evolution of the Dislocation Structure with Temperature 3.5. The Comparison of the Results for CuAu Ordered Alloy and TiAl 4. Possible Causes of TiAl Brittleness 5. Analysis of Transition from the High-Temperature Step to the Low-Temperature Step of Deformation 5.1. Reversibility and Irreversibility of y(T) 5.2. Transparency of the Dislocation Framework and Stress Macro- Jump 5.3. Mechanism Initiating Blocking of a Dislocation Source 5.4. Comparison of Two-Step Deformation of TiAl and Ni3Al 6. Conclusion 1. INTRODUCTION One generation of intermetallics change another, which is necessitated primarily by the need to have better aerospace materials. Today light- weight titanium aluminides replace intermetallics of the first generation like Ni3Al and heavy superalloys based on those intermetallics [1, 2]. Ti- tanium aluminides, in the first place TiAl, possess excellent properties [3– 5]. Unique complex of physical properties of TiAl includes a high strength/density ratio, heat resistance, high-temperature strength, resis- tance to oxidation, corrosion and creep, and good fatigue characteristics. But low plasticity of TiAl hampers realization of these properties. Methods for improving TiAl plasticity can be elaborated if we understand causes of deformation strengthening and brittleness of TiAl. This goal is pursued by many researchers, including the authors of this review. The series of tita- nium aluminides extends further. It includes orthorhombic titanium alu- minides based on Ti2(Al,Nb) [6–8]. They are inferior to TiAl with respect to the strength-density ratio. But their plasticity is better. Elongation about of 16% was achieved for Ti2(Al,Nb) at room temperature. New problems always arise when one generation of intermetallics replaces another. They include a loss of plasticity when passing from Ni3Al to TiAl or diffi- 12 B. A. Greenberg and M. A. Ivanov culties involved in production of a single-phase material in the case of orthorhombic aluminides, etc. An the interval from room temperature to 600C. An anomalous depend- ence anomalous temperature dependence of the yield stress y(T), which was inherited from intermetallics of the first generation, should be in- cluded in the aforementioned properties of TiAl. The yield stress of TiAl increases in y(T) is an excellent property: the higher the temperature, the larger the stress at which plastic flow begins. However, y(T) is non- monotonic for TiAl in the whole temperature interval. A nonmonotonic temperature dependence of the yield stress y(T) in single crystals of TiAl is observed at all orientations (Fig. 1): those fa- vourable for single dislocations and superdislocations [9–11]. The analysis of those data was focused mainly on the anomalous trend of y(T). How- ever, the y(T) curve has two extrema and, correspondingly, exhibits a normal temperature behaviour in two different temperature intervals (at low and high temperatures) and an anomalous behaviour at intermediate temperatures. Therefore, the nonmonotonic dependence y(T) need be interpreted for the whole temperature interval. Moreover, the work hard- ening rate (T) also has a nonmonotonic dependence, which does not duplicate the dependence y(T). Kawabata et al. [9] showed that the work hardening rate (T) may have an anomalous trend at some orientations and a normal trend at other orientations in the region of the anomalous behaviour of y(T). However, Unui et al. [11] noted that even if (T) has an anomalous trend, the temperature of the (T) peak is 100 to 200 lower than the temperature of the y(T) peak. One more significant fea- ture of the deformation behaviour of TiAl is a very small strain-rate sensi- Figure 1. Yield stress of TiAl single crystals in different temperature ranges [9, 10]. Anomalies in Deformation Behaviour of TiAl Intermetallic 13 tivity of the flow stress, which is observed [9, 11] in the region of the anomalous behaviour of y(T). In our opinion, the aforementioned features of the deformation behav- iour are due to transitions between possible states of dislocations, primar- ily to thermally activated transitions. The comparison of typical y(T) curves observed for various materials made it possible to reconstruct the shape of the potential barrier for dislocations in TiAl. Using those data, we tried to describe the temperature dependence of the deformation characteristics in the whole temperature interval. In addition, assuming that the potential relief for dislocations may change near a crack tip, we made an attempt to explain the brittle–ductile transition in TiAl. 2. TYPICAL y(T) CURVES FOR VARIOUS MATERIALS AND TRANSITIONS BETWEEN DISLOCATION STATES It was proposed [12–15] to describe plastic deformation as an evolution of a dislocation population, which is determined by multiplication of dislo- cations and their transformations. These processes occur on the back- ground of elastic stress fields created by the dislocation ensemble itself. The transformations actually represent transitions between possible states of dislocations. It is reasonable to think that dislocations are ‘born’ glissade. If the energy of a dislocation in the glissade (g) form is larger than in the sessile (s) form, gs transformations are possible. During these transformations a gain in the dislocation energy is achieved at the expense of a loss of mobility by dislocations. Reverse sg transforma- tions, which return dislocations to the glissade form, are usually thermally activated. Direct gs transformations may be either thermally activated or athermal. It is necessary to distinguish thermally activated processes among gs transformations. It is these processes that determine the temperature dependence y(T). Figure 2 presents typical y(T) curves for some materials. A possible potential relief, i.e. the potential energy of various dislocation states, is sketched schematically in a simplest form for the same materials in Fig. 3. Since the energy of the glissade form of dislocations is higher than the energy of the sessile form, the potential barrier separating these states is asymmetric: the potential well for the glissade form is shallower than the well for the sessile form. Metals having a b.c.c. lattice are characterized by a sharp drop of y(T) with growing temperature [16]. It is assumed to be an established fact that planar and non-planar configurations of the wide core of disloca- tions represent g- and s-states respectively in b.c.c. metals [17]. It is the thermally activated sg transition that determines the observed depend- ence y(T). But the gs transition is not a thermally activated one. Oth- erwise, y(T) would have an anomalous trend, which would change to the normal trend with growing temperature. 14 B. A. Greenberg and M. A. Ivanov A typical feature of semiconductors is that plastic flow is first observed at relatively high temperatures, and y(T) exhibits a normal trend [18]. In the given case gs transitions are due to immersion of a dislocation into a deep Peierls valley. s.c. b.c.c. Yield stress a d c b Figure 2. Typical y(T) curves for some materials: a—b.c.c. metal; b—semi- conductor; c—Ni3Al intermetallic; d—TiAl intermetallic. Ni3Al TiAl Anomalies in Deformation Behaviour of TiAl Intermetallic 15 In both cases (Fig. 3a, 3b), the potential well of the g-state is so shal- low that thermal activation is unnecessary to escape this potential well. Figure 3. Schematic picture of potential relief for dislocation in different materials; g—glissade state; s, s—sessile state. 16 B. A. Greenberg and M. A. Ivanov Here the potential well is conditional only. The potential well for the s- state (Fig. 3b) is much deeper than for the s-state (Fig. 3a). Therefore sg transitions are possible only at relatively high temperatures, a fact which determines the observed temperature dependence y(T) in semi- conductors. Such intermetallics as Ni3Al are characterized by the y(T) curve hav- ing the shape shown in Fig. 2c: a weak normal dependence at low tem- peratures (region I); an anomalous trend at intermediate temperatures (region II); a normal trend at high temperatures (region III) [1, 2]. A non- monotonic temperature dependence y(T) with two extrema, which is ob- served for TiAl, is shown schematically in Fig. 2d. A specific feature of the y(T) curve for TiAl is that the temperature drop at low temperatures is much larger than for Ni3Al. The observed temperature dependence y(T) can be explained if one assumes that the potential relief for dislocations is more complicated in intermetallics than in the aforementioned simple materials. What is meant here is that intermetallics contain, in addition to one type (g) of glissade dislocations, two types of sessile dislocations, which correspond to shal- low (s) and deep (s) potential wells. Dislocations move for some time on the relief with shallow wells, which allows some free path length , and then ‘fall’ into a deep well. To emphasize this feature of the relief, shallow wells would be repeated several times in the sketch diagram of the relief for Ni3Al (Fig. 3c) and TiAl (Fig. 3d). Wells of the s-type prove to be very shallow in Ni3Al. Barriers separating these wells are easily surmounted and the relief has little effect on the tem- perature dependence of the yield stress even at low temperatures. Deep wells of the s-type correspond to a blocked Kear–Wilsdorf configuration. As is known [19], such barriers appear due to resplitting of a superdislo- cation when a superpartial dislocation goes to the cube plane (Fig. 4a). As distinct from gs transitions, a relatively high potential barrier need be surmounted for gs transitions (see Fig. 3c), which are realized by a thermally activated mechanism. These transitions take a long waiting time, during which a dislocation undergoes a number of gs transitions. Note also a large asymmetry of the potential barrier separating g- and s-type states. This leads to a rigorous relation between activation ener- gies of forward (gs) and reverse (sg) transitions: Ugs'Us'g Conse- quently, gs transitions determine an anomalous trend of y(T) in the region II, and sg transitions are responsible for the normal trend of y(T) in the region III. Let us analyse the character of the presumed potential relief for a dis- location in TiAl (Fig. 3d). The main distinction from the situation dis- cussed above (Fig. 3c) is that potential wells of the s-type are not so shal- low in TiAl as in Ni3Al. Consequently, thermal activation is required to es- cape these potential wells. This determines a sharp drop of y(T) with temperature in the region I. When temperature is increased, dislocations may go to the s-state by a thermally activated mechanism. Similarly to Anomalies in Deformation Behaviour of TiAl Intermetallic 17 the case considered above, the transition takes a long waiting time, dur- ing which movement of a dislocation is connected with potential wells of the s-type. Consider first a single dislocation. It is assumed that gs transitions are connected with submersion of a dislocation in deep Peierls valleys [20]. One can expect a potential well of the s-type to be deep as in semi- conductors (Fig. 3b). Then a question arises: which main g-state should single dislocations in TiAl have so that their blocking in deep Peierls val- leys requires, unlike in semiconductors, thermal activation and causes an anomalous trend of y(T). Moreover, the choice of the main state deter- mines the reason why an anomalous trend of y(T) is replaced by the normal trend at low temperatures. A single dislocation in the main g-state is supposed not to be con- stricted but to have a wide planar core [21]. In this case transitions of two types, gs and gs, are possible from the main state. Put another way, a potential well corresponding to the g-state in TiAl has walls of different height: a low wall for the gs transition to an non-planar wide core and a high wall for the gs transition to a constricted dislocation localized in a deep Peierls valley. Only gs transitions are thermally activated, be- cause they require formation of a constriction as a necessary step. There- fore blocking of a single dislocation in deep Peierls valleys occurs through thermal activation rather than spontaneously. Thus, at orientations favourable for single dislocations, a non- monotonic dependence y(T) is determined by the following thermally ac- (a) (b) (c) Figure 4. Schematic picture of Kear–Wilsdorf barrier: a—Ni3Al, b, c—TiAl; double line—complex stacking fault, bold line—superstructural stacking fault, dotted line—{010} APB. 18 B. A. Greenberg and M. A. Ivanov tivated transformations: the normal trend at low temperatures (region I) depends on sg transformations of a wide non-planar core to a planar core; an anomalous trend at intermediate temperatures (region II) is due to gs transformations of a wide core to a constricted form localized in a deep Peierls valley; the normal trend at high temperatures (region III) is determined by sg processes of the dislocation release from a deep Peierls valley. Similarity of the curves describing a nonmonotonic behaviour of y(T) in TiAl at different orientations is observed. The analysis of numerous TEM data suggests some feature in common for a single dislocation and a superdislocation or, more precisely, the constituent 1/2101] superpar- tial dislocation: no band of a complex stacking fault is observed in both cases [22]. It seems reasonable to assume that the superpartial disloca- tion has, like a single dislocation, a wide core. But the core of the 1/2101] superpartial dislocation differs structurally from the core of a sin- gle dislocation. This is because an APB band borders on the superpartial dislocation and couples it to an adjacent partial dislocation. Thus, a 101]-superdislocation in the main g-state has a glissade con- figuration, which includes a superpartial dislocation with a wide planar core. Similarly to a single dislocation, transitions of two types, gs and gs, are possible for a superdislocation in the main state. The potential well corresponding to the g-state has walls of different height (Fig. 3d): a low wall and a high wall for gs and gs transitions respectively. The gs transition describes rearrangement of the wide core of a superpar- tial dislocation from a planar to a non-planar form, while the gs transi- tion gives rearrangement of a glissade superdislocation caused by its re- splitting (Figs. 4b, c). The analogy readily transpires to the gs trans- formation of a glissade superdislocation to a blocked Kear–Wilsdorf con- figuration in Ni3Al discussed earlier. Referring to gs transitions in TiAl, they are largely similar to rearrangement of the wide core of a superpar- tial dislocation in Pt3Al [23]. Consequently, at orientations favourable for superdislocations, a non- monotonic dependence y(T) is determined by the following thermally ac- tivated transformations: the normal trend at low temperatures (region I) depends on sg transformations of the wide core of a superpartial dislo- cation from an non-planar to a planar form; an anomalous trend at inter- mediate temperatures (region II) is due to gs transformations of a su- perdislocation to a blocked form as a result of resplitting; the normal trend at high temperatures (region III) is determined by sg transformations of a superdislocation from a blocked to a glissade form. Although this se- quence of transformations describes rather fully, as shown later, the ob- served dependence y(T), the actual picture is more complicated, be- cause two types of superdislocations exist: some are not observed in the region II and others in the region III. Causes of their vanishing will be dis- cussed in section 3.4 below. The form of the potential relief described above is rough. In actual fact, Anomalies in Deformation Behaviour of TiAl Intermetallic 19 the potential relief exists in a multidimensional configuration space. One can say only about plane sections of the potential relief depending on which co-ordinates characterizing the state of a dislocation are fixed. Moreover, an energy barrier is characterized by its shape in addition to the wall height. It is essential whether the barrier walls are steep or not. However, the proposed approach can be used as the first approximation. 3. DESCRIPTION OF A NONMONOTONIC TEMPERATURE DEPENDENCE y(T) Let us derive expressions that determine the yield stress in each of the three temperature intervals for the presumed form of the potential relief in TiAl (Fig. 2d). 3.1. Low-Temperature Drop of y(T ) The plastic deformation equation (allowing also for the elastic part of de- formation) for a population, which includes dislocations in g- and s-states, can be written as [13–15]: 1 1 1 1 2 2 2 2           d d d d c s          , (1) where dimensionless quantities c and  s denote characteristic deforma- tions. They have the form   c gsf b  2 3 1 and   s sg  / . (2) Here gs is the free path length connected with gs transformations; sg—the frequency of sg transformations; b—the Burgers vector magni- tude; —the shear modulus; f—the Schmid factor for the given slip sys- tem;  denotes some numerical coefficient. The second term in brackets describes the contribution from the elastic part of deformation. It tends to unity in the region of elastic deformation and to zero in the region of plas- tic deformation. The notation (1) assumes that sources of dislocations operate rapidly and adjust the applied stress to the density of dislocations  in accor- dance with the known relation    and f/b . (3) In addition, the initial density of dislocations is taken to be small. We shall not consider the case of slow operation of sources and, correspondingly, a finite time of adjustment between  and , when it is necessary to intro- duce an equation for the population growth by multiplication of disloca- 20 B. A. Greenberg and M. A. Ivanov tions [13–15]. The relative contribution of the terms in equation (1) is determined by the relation between the lifetime of the s-state  1 sg and the observation time  /  . In the limiting case, when sg 1   /  , i.e. the s-state is a short-lived one, the relation d d sg          /  1 , (4) is fulfilled. Therefore, in accordance with (1), the work hardening rate in the re- gion I is small and the yield stress  I y is determined by the expression    y I s c . (5) It is assumed that only sg transformations are thermally activated in the region I (see section 2). Therefore  sg sg sgU kT        ~ exp  (6) where Usgis the activation energy of the sg transformation. Taking into account (2) and (6), the expression (5) readily yields  y I y I sgU kT        ~ exp 2 . (7) Here ~ I y changes with  , because s is proportional, in accordance with (2), to  . However, expression (7) holds only if one assumes that quanti- ties c and s are independent of the stress. Otherwise, expression (5) should be considered as an equation for determination of  I y. If the stress dependence of the activation energy Usg is taken into account,  I y exhibits a weaker dependence on temperature and strain rate [13, 15]. The nor- mal temperature trend of  I y(T) is preserved. This situation will not change until thermally activated gs transformations, which cause blocking of dislocation sources, become significant. 3.2. Anomalous Temperature Trend of y(T) In many cases specific features of the deformation behaviour of TiAl stem from one and the same reason: thermally activated blocking of dislocation sources. A direct consequence is an anomalous temperature depend- ence y(T). One more obvious consequence is the aforementioned van- ishing of dislocations of definite type in some temperature intervals as discussed below. Less obvious but nonetheless significant are conse- Anomalies in Deformation Behaviour of TiAl Intermetallic 21 quences of dislocation source blocking, which is initiated by some exter- nal effect. Such unexpected consequences include, in the final analysis, brittleness of TiAl (see section 4) and stress macrojump (see section 5). 3.2.1. Thermally activated blocking of dislocation sources A dislocation source can operate if a dislocation segment passes the critical configuration before it is blocked. As a result, the free path length,  F, of a dislocation belonging to a source serves as the maximum linear dimension of the source [24–26]. A segment, whose length exceeds  F, becomes blocked rather than acquires the critical configuration. In other words, the source will not operate. For plastic flow to begin, the applied stress should be higher than the characteristic minimum value of the switch-on stress Fof a dislocation source, which is connected to the length  F, by the relation F f  b/F. (8) Here and henceforth the index F refers to quantities associated with dis- location sources. The introduction of F as a threshold value determining the behaviour of a dislocation population does not mean that plastic flow is prohibited in principle at stresses lower than F. One should allow for smearing of the switch-on stress F. This is due to both the statistical character of thermally activated processes and fluctuating character of the internal stress field. Similarly, another threshold stress—the stress determined by (1) and needed to overcome the elastic counteraction of the dislocation framework—is smeared too. Smearing of these threshold stresses can be taken into account by the method proposed in Refs. [14, 15]. In the case of the thermally activated gs transformation F (T)can be written as   F F gs FU kT  ~ exp / ( ) , (9) where U gs F ' is the blocking activation energy of a dislocation belonging to a source. From (8) and (9) it follows that  F F gs F T U kT ( ) ~ exp         , (10) Assuming that the yield stress y II (T) is determined by the switch-on stress of a dislocation source F(T), from (10) we immediately obtain an anomalous temperature dependence of the stress y II (T). Let it be empha- sized that expressions (8) and (10) proceed from the assumption that in 22 B. A. Greenberg and M. A. Ivanov the region II the s-state of a dislocation belonging to a source is long- lived and activation energies are interrelated as U gs F ' U s g F ' . From (7) and (9) it is seen that y II T) is independent of the strain rate  . As stated ear- lier, the strain-rate sensitivity of the flow stress, which is observed in the region II, is weak indeed. 3.2.2. Temperature dependence of the work hardening rate From (9) it is seen that the source switch-on stress F(T)drops with de- creasing temperature. At sufficiently low temperatures F(T)becomes smaller than y(T), which is determined by (5), and can be neglected. However, F(T) should outgrow  I y(T) with increasing temperature, be- cause these quantities have different temperature trends. The region II of an anomalous temperature dependence of the yield stress starts at tem- peratures when F(T) >  I y(T). Thus, the temperature of the minimum, Tmin, in the y(T) curve for TiAl is determined by the equation    F s c T( )min  . (11) Write the equation of plastic deformation allowing for gs transforma- tions in the form analogous to (1):  1 1 2 2 2 2       d d c s  ' ' , (12) where   c gsf b' '  2 3 1 and   s s g' '  / . (13) If both gs transformations are thermally activated, we have   gs gs gsU kT    ~ exp ,            s g s g s gU kT ~ exp . (14) In equation (12) the first term in the right-hand side describes blocking of dislocations owing to gs transformations, while the second term gives unblocking of dislocations as a result of sg transformations. As distinct from (1), the first term in (12) does not include a factor allowing for elastic deformation. Equation (12) holds only when  F(T) and elastic part of deformation, for which , extends up to  values equal to F  F(T)/ and it is omitted from this equation. Thus, initial values for equation (12) are  F(T)and  F where F(T)is determined by expression (10). At the beginning of deformation, when the flow stress is still close to F(T) Anomalies in Deformation Behaviour of TiAl Intermetallic 23 equation (12) can be written in the form: 1 2 2            d d c F F s c             . (15) Let us consider the case when blocked configurations of the s-type are practically indestructible and the second term on the right of (15) can be neglected. This case corresponds to moderate temperatures. From (15) we have d d c F           2 . (16) Hence, considering (10), (13) and (14), we obtain    ~ exp - U U kT gs gs F        . (17) From (17) it is seen that (T) is independent of temperature if activation energies are nearly equal, i.e. Ugs´  U F gs´. The fact that the temperature dependence of the work-hardening rate (T) is indeed observed is an evidence that blocking of dislocations at the stage of multiplication and during subsequent deformation has different activation energies U F gs and Ugs respectively. From (17) it follows that the relation U F gs  Ugs should be fulfilled for the coefficient (T) to have an anomalous temperature de- pendence. This relation seems to be reasonable, because a dislocation belonging to a source can be blocked easier than in the bulk thanks to the presence of pinning points. If Ugs  2U F gs a coefficient (T) exhibits a stronger temperature dependence than y(T). If U F gs  Ugs  2U F gs  the temperature dependence (T) is weaker than that of y(T). The observed types of the temperature dependence T) [9] are illustrated in Fig. 5. 3.3. Drop of (T) Above the Temperature Peak When the temperature rises, sg transformations become significant and can determine the onset of plastic flow. In this case, y(T) can be presented as y III  which is given by an expression analogous to (5):   y I I I s c  ' ' . (18) Taking into account (13) and (14), from (18) we have  y I I I y I I I U U kT s g gs         ~ exp 2 . (19) 24 B. A. Greenberg and M. A. Ivanov If the stress dependence of the activation energy is neglected, it readily follows that y III depends, unlike F on  , namely    and exhibits a normal temperature behaviour, because the obvious condition Ugs  Usg is fulfilled (see section 2). Thus, growth of the yield stress with increasing temperature should be followed by its drop described by expression (19). In this case the temperature Tmax of the maximum in (T) is found from the equation F(Tmax)  y III (Tmax), (20) which can be written as   F s c   . (21) From (15) and (20) it is seen that the work hardening rate (T) de- creases near the temperature Tmax and turns to zero at   Tmax. This means that (T) passes its maximum at a lower temperature than y(T) does. Put another way, the (T) peak is shifted relative the y(T) maxi- mum towards lower temperatures. As stated earlier [11], this behaviour of the deformation characteristics indeed is observed both for TiAl and other intermetallics. Figure 5. Temperature and orientation dependence of work hardening rate [9]. Anomalies in Deformation Behaviour of TiAl Intermetallic 25 When T  Tmax, the upper and lower yield stress exists at high tempera- tures in TiAl [9,11] as in semiconductors. Expression (19) just determines the lower yield stress  L y. If the stress dependence of the activation en- ergy is taken into account, weaker temperature and rate dependencies of  L y can be obtained as for  L y considered above. The presence of the up- per yield stress  U y indicates [13–15] that the characteristic time  of dislo- cation multiplication, i.e. the switch-on time of dislocation sources and, correspondingly, the time required for adjusting  and , is relatively long. The analysis of the equation for the dislocation population growth during slow operation of dislocation sources showed that the upper yield stress is proportional to  . Eventually, the deformation behaviour of TiAl is different in regions I and III, because potential wells corresponding to s- and s-states have different depth (see Fig. 3d). Therefore two types of traps exist: shallow traps of the s-type and deep traps of the s-type. Release of dislocations from traps leads to a normal temperature dependence y(T). It is quite reasonable that the drop of y(T) with growing temperature in regions I and III is due to release of dislocations from shallow and deep traps re- spectively. However, the presence of two types of traps is insufficient for the ob- served trend of y(T). Dislocations should first be captured in these traps. Assume that a potential well corresponding to the g-state has walls of the same height as for transition to s- and s-type states. In this case TiAl would exhibit a different deformation behaviour. Indeed, s-type traps would be unobservable, because gs transitions to the long-lived state always dominate over gs transitions to the short-lived state at compa- rable free path lengths of dislocations [27]. The density of blocked s-type dislocations actually is much higher than the density of blocked s-type dislocations, because s-type dislocations accumulate during the whole observation time equal to /  , while s-type dislocations accumulate only during their lifetime  1 sg  /  . But, then, a strong drop of y(T) with growing temperature in region I would be unobservable. The more so that s-type traps would not have any effect on the deformation behaviour of TiAl if the barrier for gs transitions is higher than the barrier for gs transitions. Only when the barrier for gs transitions is lower than for gs transitions, both types of traps show up themselves: shallow traps in region I, and deep traps in regions II and III. The origin of deep traps is different for dislocations of various types: they result from immersion of single dislocations to deep Peierls valleys, whose presence is due to co- valent-like interatomic bonds, and resplitting of superdislocations, which is followed by formation of low-energy surface defects. 3.4. Evolution of the Dislocation Structure with Temperature TEM investigations of the dislocation structure of TiAl helped ascertaining 26 B. A. Greenberg and M. A. Ivanov types of dislocations observed in various temperature intervals and forms (glissade or blocked) of dislocations [28–34]. Vanishing of superdisloca- tions 1/2112] at intermediate temperatures (region II) and superdisloca- tions 101] at high temperatures (region III) was noted. We relate this ef- fect to thermally activated blocking of dislocation sources. If a population comprises dislocations of different types, plastic flow will be realized by dislocations whose sources are switched on at a lower stress F. Sources, which can be easily blocked (the activation energy U F gs is small), will not be switched on and dislocations of the correspond- ing type will be unobservable. It may be thought that easy recombination of a superpartial dislocation causes blocking of 1/2112] superdisloca- tions. Indeed, calculations [32]. suggest that the wide core of a superpar- tial dislocation will be constricted due to the stress field of an adjacent partial dislocation. This stress field will favour blocking of sources of 1/2112] superdislocations. Consequently, 1/2112] superdislocations are unobservable in the region II. This example is also significant as direct evidence that the stress field may stimulate blocking of dislocation sources. However, superdislocations 1/2112] can be made observable in re- gion II by the following method [32]: deformation at 196C and weak de- formation in situ at 300C. Superdislocations 1/2112] are injected during low-temperature deformation, when sources are not blocked, and be- come blocked after deformation in situ. A blocked superdislocation 1/2112] is shown in Fig. 6. Moreover, the following situation may arise: sg transformations are already realized for dislocations of some type, while blocked configura- tions remain indestructible for dislocations of another type. The switch-on stress F of the last type of dislocations will continue growing with in- creasing temperature in accordance with (10). Plastic flow will start at a smaller stress, which is connected with gs transformations and decreases with temperature in accordance with (19). One may think that this situation is realized in region III where 101] su- perdislocations are unobservable [29, 32]. In the final analysis, this is ex- plained by indestructibility of blocked superdislocations of the Kear– Wilsdorf type in TiAl. These blocked superdislocations cannot be easily transformed to glissade configurations, because it is difficult to constrict the band of a superstructural stacking fault. The 101] superdislocations may be observed at high temperature as product of dislocation reactions only. Therefore the drop of y(T) in region III is due to unblocking of sin- gle dislocations or 1/2112] superdislocations. Figure 7 shows a dislocation structure with blocked dislocations after deformation at 196C. Microphotographs of superdislocations blocked at intermediate temperatures are given in Fig. 8. A typical microphotograph [31] shows a configuration, which we re- ferred to as ‘a tree’, was observed more than once [33]. The trunk of the Anomalies in Deformation Behaviour of TiAl Intermetallic 27 tree is a blocked superdislocation in the form of a Kear–Wilsdorf barrier and branches are represented by single dislocations. Single dislocations blocked at intermediate temperatures are seen in Fig. 9. The fact that blocked dislocations are observed both at very low [10, 34] and relatively high temperatures confirms the above supposition on two types of traps. Micro- structure of TiAl after high temperature deformation is shown in Fig. 10. Typical microphotograph of blocked single dislocations is presented in Figs. 9a, 9b. A clearly pronounced rectilinearity of long segments of sin- gle dislocations is remarkable [30, 31]. Figure 9c depicts two families of blocked single dislocations having mutually perpendicular Burgers vec- tors. However, single dislocations having a number of pinning points with bent segments between those points were observed too [11, 35, 36]. But mostly they extended along their Burgers vector (Fig. 9d). A paradoxical situation arises: on the one hand, y(T) and (T) exhibit an anomalous trend and, on the other hand, TEM data suggest possible unblocking of single dislocations at intermediate temperatures. The microstructure in Fig. 9d can be interpreted as follows. Single dis- locations emitted by a source are localized in deep valleys. Subsequently double kinks can be formed. Screw segments of these double kinks Figure 6. Dark-field electron micrographs [32] taken from a sample of Ti–50Al alloy deformed at 196C (a) and after two-stage treatment (b–d): beam direction BD near [110]; Burgers vectors of dislocations A and B are 1/2[1 1 2] and 1/2[ 1 10] respectively. 28 B. A. Greenberg and M. A. Ivanov again are localized in deep valleys. When Schmid factors of a single dis- location are equal in both octahedral planes, double kinks can be emitted in these planes. A node in a single dislocation, where such kinks meet one another, serve as a pinning point. Therefore observation of pinning points is only the sign accompanying above—mentioned transitions of a single dislocation. The situation is fully described by expression (10) and Eq. (12). Ex- pression (10) gives an anomalous trend of F(T) and, correspondingly, y(T). As the degree of deformation increases and the flow stress ex- ceeds F, the second term on the right of Eq. (12), which describes sg transformations (or unblocking of dislocations), grows. However, as long as the condition sc  2 F is fulfilled, an anomalous trend of (T) will con- tinue. Thus, the cause of an apparent discrepancy between TEM data and observed anomalies is that the characteristic time of dislocation multipli- cation is much shorter than /  .  Therefore sg transformations, which were not realized during dislo- cation multiplication, become possible upon subsequent plastic deforma- Figure 7. Microstructure of TiAl at 196C: a, b—dark-field images of dislocations in Ti–56Al [10]; c, d—bright-field images of dislocations in Ti–50Al (courtesy O. V. Antonova and N. V. Kazantseva). Anomalies in Deformation Behaviour of TiAl Intermetallic 29 tion. Elementary events of unblocking are observed in an electron micro- scope. Figure 8. Bright-field images of the dislocation configuration (‘tree’) in a sample of Ti–50Al alloy after deformation at 400C [31]. Figure 9. TEM images of single dislocations at intermediate temperature: a, b— microstructure of Ti–50Al alloy deformed at 400C [31]; c—dark-field electron micrographs [32] taken from a Ti–50Al sample deformed at 540C; two families of blocked screw dislocations with Burgers vectors: 1—1/2[110 ], 2—1/2[110], re- corded with g = [020]; d—single dislocation in a sample of -Ti47Al51Mn2 alloy de- formed at 500C, g = [ 220 ] [35]. 30 B. A. Greenberg and M. A. Ivanov 3.5. The Comparison of the Results for CuAu Ordered Alloy and TiAl To gain a deeper insight into the nature of TiAl, the deformation behav- iours of TiAl and CuAu alloys were compared. The anomaly of the yield stress was detected [37] for the polycrystal of ordered CuAu alloy without the lamellar structure. It is known that the critical temperature Tc, for ordering of the CuAu alloy is 410C and, besides, CuAuII is the equilibrium phase within the interval from 385C to 410C. The regimes of thermomechanical treat- ment were selected using appropriate TTT curves [38]. As a result a polycrystal of CuAuI had relatively coarse grains (5 to 10 m) and was almost free of the lamellar structure. Mechanical tests were performed within the temperature interval from 196C to 350C. Foils for TEM analysis were prepared using a standard technique from sample de- formed to 3%. Figure 11 shows the yield stress  as a function of the test tempera- ture. From Fig. 12 it is seen that (T) exhibits a nonmonotonous tem- perature trend with a maximum near Tmax 300C. The  value at Tmax is nearly twice as high as the value at room temperature. A detailed Figure 10. Microstructure of Ti–50Al after deformation at 800C: a–c—dark-field images of node cascade recorded with g  [1 1 1]; (d) schematic picture. Disloca- tions with Burgers vector: 1—1/2[110], 2—1/2[1 1 0], 3—1/2[ 1 12], 4—[011], 5— [10 1 ], 6—[010], 7—1/2[11 2 ] (see [32]). Anomalies in Deformation Behaviour of TiAl Intermetallic 31 analysis of the microstructure of the CuAu alloy deformed at various tem- peratures can be found elsewhere. Both single dislocations and superdis- locations with the Burgers vectors 101] and 1/2112] were observed. Moreover, numerous microtwins were seen. APB bands in the cube plane were also detected. Figures 12 and 13 illustrate typical fragments of the microstructure of the CuAu alloy deformed at different temperatures. Blocked 101] superdislocations were observed in the region of the anomalous temperature trend. The long rectilinear dislocation in Fig. 13b was identified as a screw superdislocation. In addition, a characteristic configuration was observed: the so-called ‘tree’. The trunk had a frag- mented structure, which included segments of superdislocations having both 101] and 1/2112] Burgers vectors. This configuration was ob- served in TiAl (Fig. 8). In CuAu, the configurations of the ‘tree’ type ap- peared even at room temperature (Fig. 13a) and persisted to the tem- perature Tmax. Blocked dislocations were not detected among a host of Figure 11. Temperature dependence of the yield stress 0.2(T) of the CuAu or- dered alloy [37]. Temperature, C Yield stress, 0.2 200 200 0 120 160 200 400 32 B. A. Greenberg and M. A. Ivanov single dislocations clearly seen in Fig. 13. It was surprising to observe blocked dislocations at 196C (Fig. 12). Their type has not been identi- fied so far. Referring to the very fact of their observation, one can draw an analogy with observation of blocked dislocations in TiAl at 196C (Fig. 7). The nonmonotonous temperature dependence of (T), which is ob- tained for the CuAu polycrystal (Fig. 11), is similar to the dependence ob- served for the TiAl single crystal (Fig. 1). For the sake of comparison the (T)curve was obtained for the CuAu alloy with the grain size (0.5 m) much smaller than in the alloy under study. It was found that the nonmonotonous temperature trend is replaced by a plateau (270 MPa). A similar dependence was observed for TiAl polycrystals with fine grains [39, 40]. It is worth noting that the mean grain size, at which the anoma- lous (T) vanishes, is much smaller in CuAu than in TiAl. In CuAu alloy, unlike in TiAl alloy, blocking of single dislocations was not detected in the region of the anomalous temperature behaviour of (T). This fact is an evidence that in TiAl single dislocations are blocked through a special mechanism, which is related to features of the TiAl in- termetallic itself rather than those of the L10 superstructure. The observa- tion of the anomalous (T) in CuAu is due to the fact that compatibility of deformation for a polycrystal always requires multiple slip. 4. ON THE POSSIBILITY OF CHANGES IN THE POTENTIAL RELIEF NEAR A MICROCRACK If the yield stress has a nonmonotonic temperature dependence, the temperature dependence of plasticity (T) for TiAl includes low plasticity up to the temperature Tmax of the y(T) peak, after which (T) increases with elevating temperature. A combination of a small yield stress and low plasticity at room temperature is unusual. Figure 12. Microstructure of the CuAu ordered alloy deformed by 3% at 196C, bright-field image of dislocations (courtesy O. V. Antonova and N. V. Kazantseva). Anomalies in Deformation Behaviour of TiAl Intermetallic 33 Room temperature is at the boundary between regions I and II. As stated earlier, a low y value corresponds to gs transitions, i.e. capture of dislocations in shallow traps and their release from these traps. Deep traps do not show up themselves yet. The potential relief in region I actu- ally acquires a shape similar to the one typical of b.c.c. metals (Fig. 3a). However, the shape of the potential relief can change in the vicinity of a microcrack. Indeed, it is known that a stress field alters the structure of a dislocation core. An example relating to a 1/2112] superdislocation was considered in section 3.4. Consider a configuration consisting of a crack and its associated screening dislocations. Assume that the characteristic crack stress– Figure 13. Microstructure of the CuAu ordered alloy deformed by 3%: a—at room temperature, dark-field image of dislocations in the reflection g [020], BD [10 1 ]; b—at 250C, bright-field image of dislocations [37]. 34 B. A. Greenberg and M. A. Ivanov concentration around the crack tip would shrink a wide core of the dislo- cation. Consequently, the dislocation would be captured in deep traps more easily. For a single dislocation this would mean that the dislocation drops more readily to deep Peierls valleys. However, one and the same elastic field cannot simultaneously com- press and expand a dislocation core. Therefore, if blocking of a single dislocation is facilitated near a crack, its release from a deep valley will not be easier than in the bulk of the material. This assumption can be formulated as follows: the ‘hill’ separating g- and s-states is truncated due to the stress concentration, but the poten- tial well, which corresponds to the s-state, remains deep. As a result, the potential relief approximates in shape the relief typical of semiconductors (Fig. 3b). The presumed change of the potential relief near a microcrack should considerably affect operation of dislocation sources in this area. From analysis of shielding of a crack by a dislocation, it is conjectured that the ubiquitous presence of dislocation sources in metals assures generation of copious dislocations. The one dislocation with shielding Burgers vector will be repelled from the crack, and the other with antishielding Burgers vector will be attracted to the crack and cause a crack-opening displace- a Figure 14. A crack and dislocations in plastic zone: a—dislocation sources (S) operate external to a crack and cause a crack-opening displacement [41]; b—a fully brittle crack in Si [41]; c—sessile configuration of 1/2<011> screw disloca- tions after nucleation of a crack in -sulphur; a crack is located at bottom right of micrograph [42]. Anomalies in Deformation Behaviour of TiAl Intermetallic 35 ment (Fig. 14a) [41]. Sharp cracks, near which dislocations are not seen, i.e. dislocation sources are inoperative, are observed in exceptional (non-metal) cases only. Such a crack in Si is shown in Fig. 14b. Taking into account similar- ity of the temperature dependences (T) observed for semiconductors [18] and TiAl, one can assume that a similar situation is realized in TiAl, despite different temperature dependences y(T) of these materials (see Fig. 2b, d). Let us consider switch-on of dislocation sources in the bulk of the ma- terial and in the vicinity of a crack. A low y value corresponds to easy switch-on of sources in the bulk at room temperature. However, the stress concentration near the crack facilitates gs transitions and, thus, decreases the free path length F Consequently, the stress F which is required to switch on a dislocation source, ‘automatically’ increases in accordance with (8). Therefore, instead of facilitating an intensive opera- tion of sources, as it usually does in materials with a low yield stress, the elastic field of the crack causes blocking of sources in TiAl. Moreover, if a dislocation from the bulk enters the plastic zone of a crack, it becomes blocked too. In this sense, the plastic zone of a crack itself acts as a large trap for dislocations and causes their blocking. Up to now, blocking of single dislocations in deep Peierls valleys was considered as an example. However, low plasticity is also observed at orientations, which have not single dislocations. One may think that trans- formation of the wide core of a superpartial dislocation to the constricted form is rendered easier near a crack. Therefore, transformation of a 101] superdislocation to a blocked Kear–Wilsdorf configuration or a nucleus of this configuration with a minimum width in the cube plane is facilitated. In either case, a band of a superstructural stacking fault appears. This im- pedes, as stated earlier, the reverse sg transformation of the 101] su- perdislocation. Di Persio and Escaig [42] obtained TEM data providing evidence that high stress near a crack favours the glissade–sessile transformation of a dislocation. Dislocations in the sessile form are clearly seen near a forth- coming crack in Fig. 14c. Although these results were obtained for - sulphur and not for TiAl, they confirm that a crack field can induce block- ing of dislocations. Thus, a combination of a small yield stress and low plasticity, which is observed in TiAl at room temperature, is explained by two types of dislo- cation traps: shallow traps are responsible for small y, while deep traps are the cause of low . As a result, TiAl has an unusual structure: the bulk of the material, where dislocation sources can be easily switched on, and inserts of the semiconductor type (plastic zones of cracks), where dislocation sources are blocked. In our opinion, capture of dislocations in deep traps, which is stimu- lated by concentration of stresses near a crack, and stay of dislocations in these traps up to the temperature Tmax lead to low plasticity in regions I 36 B. A. Greenberg and M. A. Ivanov and II at any orientation of a single crystal. The fact that in TiAl (T) starts growing near Tmax, when dislocations are unblocked intensively, confirms the supposition that brittleness of TiAl is connected with blocking of dislo- cation sources. The comparison of (T) trends for TiAl and Ni3Al [2] suggests that the barrier for the gs transition in Ni3Al is much higher than in TiAl. There- fore the decrease of the barrier near a microcrack in Ni3Al is not so dra- matic as in TiAl. Correspondingly, Ni3Al alloy looses less in plasticity. Plasticity decreases only on approaching the temperature Tmax. This fact also attests to a relationship between the ductile–brittle transition and blocking of dislocation sources. Thus, an efficient method for improving plasticity of TiAl may be alloy- ing with certain impurities that would favour growth of the potential barrier for the gs transition of a dislocation to a deep trap. 5. ANALYSIS OF TRANSITION FROM THE HIGH-TEMPERATURE STEP TO THE LOW-TEMPERATURE STEP OF DEFORMATION 5.1. Reversibility and Irreversibility of y(T) One of schemes of two-step deformation is as follows: the first step in- cludes small deformation (1–3%) at a high temperature T1, which is how- ever lower than the temperature Tmax of the y(T) peak; the second step is realized at a temperature T2, which normally is equal to room tempera- ture. Experiments performed by this scheme on alloys type Ni3Al [43–45] showed that on transition from the high-temperature (HT) step to the low- temperature (LT) step the stress sharply decreased (stress macrojump) almost to the yield stress 0(T2) observed in the case of one-step LT de- formation, which in what follows is taken as the reference value. Those results were treated as an evidence of reversibility of the yield stress, in line with the known experiments due to Cottrell–Stokes [46]. Some experiments [11] revealed a stress macrojump during two-step deformation of TiAl, as in the case of Ni3Al, while in other experiments [47, 48] the stress achieved by the end of the HT step remained almost unchanged on transition to the LT step: no reversibility of y(T) was ob- served. Figure 15 shows () curves for two-step deformation of TiAl. A stress macrojump is observed at the orientation [ 2 51], while no stress macrojump occurs at the orientation [010]. A contradictory picture arises: On the one hand, an anomalous de- pendence y(T) should cause a drop of the stress with decreasing tem- perature. On the other hand, the stress should be preserved if the dislo- cation structure is not recovered on transition to the LT step. A model [49] was proposed for description of two-step deformation. This model is based on an additive formula for the flow stress  written as Anomalies in Deformation Behaviour of TiAl Intermetallic 37  F (T) WH(T, ,  ) (22) Here WH(T, ,  ) is the stress increment caused by deformation strengthening (  will henceforth be omitted since the deformation rate   const under dynamic loading). It is assumed that only WH(T, ) de- pends on the dislocation density . Then, in the case of two-step defor- mation the switch-on stress of a dislocation source F(T) changes re- versibly on transition to the LT step and becomes F(T2), which coincides with the reference yield stress 0(T2) measured without preliminary de- formation. It is thought simultaneously that the quantity WH(T, ) preserves its value WH(T1) achieved by the end of the HT step, because density of Figure 15. Two-step deformation of TiAl: a—thermal reversibility of flow stress for [ 2 51]-oriented crystal, in which 101] superlattice slip is operative [11]; b—stress- strain curves for [010]-oriented crystal; lower curve, compression at room tem- perature; upper curves, two-step straining [47]. 38 B. A. Greenberg and M. A. Ivanov the blocked dislocation structure remains unchanged with decreasing temperature. Thus, in terms of the theoretical model in question, the plas- tic flow at the LT step starts at the stress y(T2)  0(T2) WH(T1). (23) From (23) it follows that the observed deviation from reversibility of y(T) (in other words, deviation of the yield stress at the LT step from the reference value) is determined by WH(T1), which describes an increase in the stress at the expense of strengthening at the HT step. In some in- stances this deviation indeed is nearly equal to WH(T1) (see, for example, Fig. 15a). In a known experiment [43], however, the yield stress at the LT step differed little from the reference value, while WH(T1) was high. When ex- perimental data obtained for various orientations [45] were processed [50] by the said scheme using relation (23), the stress WH(T1) proved to be too high or too low with respect to the observed deviation from re- versibility of y(T). Finally, the model [49] does not allow for possible conservation of the flow stress on transition to the LT step. Indeed, if (22) is true, the stress should drop nearly by {0(T1)  0(T2)} on transition from the HT step to the LT one. But no macrojump was observed in the aforementioned ex- periments [47, 48]. 5.2. Transparency of the Dislocation Framework and Stress Macrojump Since interpretation of experimental results in terms of the model [49] en- counters certain difficulties, we have proposed an alternative approach [26, 51, 52] and abandoned the additive formula (22) for the flow stress. Earlier we introduced two threshold stresses (3) and (10), which de- termine the onset of plastic flow in different cases. In a general case, in the presence of a dislocation structure formed beforehand, the condition of plastic flow should act as a ‘double key’, which ensures both operation of sources and the possibility to overcome the elastic counteraction of the dislocation structure. This condition can be written in the form:     y FT T( ) max ( ), . (24) Validity of expression (24) was convincingly proved by Ezz and Hirsch [49] who obtained the temperature dependence of the yield stress yp(T) for Ni3(Al,Hf)B intermetallic, which was subject to a strong cold deforma- tion beforehand. Figure 16 shows schematically yp(T) plotted [49] using a set of experimental data. The trend of the yield stress in the absence of cold deformation is Anomalies in Deformation Behaviour of TiAl Intermetallic 39 shown for comparison. The curve yp(T) has a plateau, which testifies to conservation of the stress caused by preliminary deformation. In other words, the aforementioned reversibility does not take place in the tem- perature interval of the plateau. Subsequently the plateau changes to an anomalous temperature trend at a certain temperature. So, the observed curve yp(T) indeed has a shape, which can be described by expression (24) and cannot be described by expression (23). Using (24), we made an effort to construct a model of deformation be- haviour of the intermetallic that would combine two seemingly incompati- ble options: observation or absence of a macrojump. Let us first consider superdislocations 101. This is the only type of dislocations in Ni3Al, but TiAl can contain other types of dislocations as will be discussed below. During the HT step, when T1  Tmax, a microstruc- ture with a high density of dislocations 1 appears. Since superdisloca- tions are blocked, this microstructure represents a rigid framework com- prising Kear–Wilsdorf barriers. These barriers are indestructible in the interval of an anomalous trend of y(T) and therefore they can hardly transform to glissade dislocations on transition to the LT step of deforma- tion. Consequently, the framework is inherited in the absence of recovery and a high density of dislocations is preserved at the LT step. In accor- Figure 16. Schematic of yield stress temperature dependence for Ni3(Al,Hf)B af- ter preliminary cold deformation [49].   T ó(T) óð(T) 40 B. A. Greenberg and M. A. Ivanov dance with (3), the stress achieved by the end of the HT step is 1(T1)   1 F(T1). (25) If the same slip system(s) continues operating at the LT step, plastic flow at a low temperature T2 of the second step should start at a stress     2 2 2 1( ) max ( ),T TF . (26) Since the stress F(T) exhibits an anomalous trend, then   F FT T T( ) ( ) ( ) 1 1 1 . (27) Consequently, at first glance, the following equality should be fulfilled:    2 2 1 1 1( ) ( )T T   (28) Therefore the stress cannot drop lower than 1(T1) with decreasing tem- perature. In our opinion, the situation can change if a new slip system operates at the LT step. In this case the framework inherited from the HT step does not contain dislocations of the new slip system. Let us simplify the task and assume that only one slip system I with the Burgers vector bI operates at the HT step. The key question is: can the dislocation frame- work turn transparent on transition to the LT step and, simultaneously, preserve its structure and density? Let the Burgers vector bII for a new slip system II be perpendicular to bI so that vectors bI and bII are parallel to diagonals of one and the same cube face. In this case, any of possible slip planes {111} is parallel to one of these vectors (Fig. 17). As a result, the framework axes do not lie in possible slip planes II but cross them. The framework acts as a ‘forest’ with respect to dislocations of the slip system II. Without going into detail of local interaction between nonparallel dislocations we shall note only that in this case the depend- ence of the type of 1/r on the distance r between dislocations, which is characteristic of the force of elastic interaction between parallel disloca- tions, vanishes [53]. But it is this dependence that underlies the relation (3). Taking into account the framework rigidity and mutual perpendicularity of vectors bI and bII, one may expect the elastic counteraction of the framework to dislocations II to be weak. Since the system II is devoid of its own framework, from (24) we readily have 2(T2) F(T2)  1(T1). (29) So, it is with this assumption that a macrojump of the stress can be ob- served on transition to the LT step of deformation. Anomalies in Deformation Behaviour of TiAl Intermetallic 41 Two points, which are significant on transition from the HT to the LT step, can be emphasized. Firstly, it is an anomalous trend of the yield stress and, correspondingly, a drop of the stress F(T) with decreasing temperature. Secondly, a slip system II, whose dislocations experience little elastic counteraction from the framework, can come into play. This brings about a question why system II is not involved at the first step even if, as it often happens in prestraining experiments, Schmid fac- tors of systems I and II are nearly equal. 5.3. Mechanism Initiating Blocking of a Dislocation Source Taking into account what has been said above, it is reasonable to think of I D A C A B AA (a) C A (b) D A B A A II Figure 17. Slip planes {111} whose intersection lines are parallel to different di- agonals of one and the same cube face. 42 B. A. Greenberg and M. A. Ivanov a mechanism that would prevent the switch-on of system II at the HT step and allow its switch-on at the LT step. Also, this mechanism should not hinder the switch-on of slip systems where Burgers vectors are parallel to different cube faces. Such slip systems indeed were observed at the HT step [43] and also during one-step deformation in the interval of an anomalous trend of y(T) (see, for example, [54]). So, the mechanism in question should be connected with thermally ac- tivated blocking, which takes place in one cube face whose diagonals are vectors bI and bII. Transformation of a glissade superdislocation to a Kear–Wilsdorf barrier may be thought of as thermally activated blocking. Obviously, such barriers are formed through slip to one and the same cube face for systems I and II only. It is this circumstance that distin- guishes systems I and II among the whole set of octahedral slip system. Let us consider collision of a glissade superdislocation II and a blocked superdislocation I. Consecutive blocking stages of a superdislocation II are shown in Fig. 18. This process includes formation of a constriction in the superpartial dislocation, recombination and bending of the recombined segment in the cube plane [55, 56]. Barrier I stimulates the cross-slip of a screw su- perdislocation II to the cube plane, because part of the APB band, which belongs to barrier I, disappears in the cube plane. For this reason, the activation energy for a superpartial dislocation II to slip under the action of barrier I to the cube plane is nearly equal to the recombination energy of this dislocation. The latter energy is lower than the activation energy of barrier I formation. This is also true for the cross-slip through generation and propagation of kinks. (b) {001} {111} (a) Figure 18. Cross-slip of superpartial dislocation followed by vanishing of the APB band in the cube plane; (a) and (b)—different stages of recombined segment bending [55]. Anomalies in Deformation Behaviour of TiAl Intermetallic 43 Assume that the slip system I is switched on first at the HT step and produces a framework with a high density of dislocations. Superdisloca- tions II, which have already been present in the initial state, are stopped by this framework (Fig. 18a). Segments of a superdislocation II, which are located between ‘forest trees’, represent potential dislocation sources. However, each ‘tree’ includes the aforementioned APB band. A blocked configuration II nucleates (see Fig. 18b). Blocking may develop along the superdislocation. Clearly, only a screw segment is blocked. However, if the framework stops an edge superdislocation II, a screw segment, which appears upon bending of this superdislocation, also ex- tends along the aforementioned APB band and slips to the cube plane. Thus, potential sources represented by superdislocations II and stopped by the framework are blocked. The mechanism of contact interaction, which we have just discussed, represents the mechanism in question re- sponsible for blocking of sources of dislocations having similar Schmid factors. If one of the systems is switched on first, the said mechanism suppresses the switch-on of the other system. This applies only to slip systems with mutually perpendicular Burgers vectors. As follows from the above discussion, ‘the second diagonal inhibit’ is ineffective with respect to diagonals of different faces. It remains to clarify why systems I and II cannot be switched on simul- taneously at the HT step. This fact may be attributed to a deviation from precise orientation. Even a slight deviation is sufficient for Schmid factors to be dissimilar. If it is assumed that dislocations multiply in an avalanche manner, one system immediately leads the other. Later, sources of the system II cannot be switched on at the HT step because of mechanism initiating the blocking. However, they can be switched on at the LT step if recombination of a superpartial dislocation is impeded and the said mechanism does not operate. 5.4. Comparison of Two-Step Deformation of TiAl and Ni3Al Maximum Schmid factors for possible types of dislocations are given in Table 1 for orientations at which TiAl underwent two-step deformation (references are given in the heading). A stress macrojump was detected in TiAl only at the orientation [ 2 51] among all orientations studied (see Fig. 15a). A similar orientation, namely [ 1 25], was examined in Ni3(Al,Ti) where a stress macrojump was revealed too [45]. From Table it is seen that the Schmid factor is a maxi- mum at the orientation [ 2 51] of the superdislocation 101] among all other types of dislocations possible in TiAl. Only one superdislocation slip system with the Burgers vector [011] has the Schmid factor equal to 0.490. The superdislocation slip system with the perpendicular Burgers vector [0 1 1] has a smaller Schmid factor equal to 0.435. Only the first slip system is observed at the HT step [11]. Other types of dislocations 44 B. A. Greenberg and M. A. Ivanov were not observed at this orientation. Considering the model proposed above, we suppose that the stress macrojump is due to the switch-on of a slip system II with the Burgers vector [0 1 1] at the LT step. Taking into account that slip systems I and II have the Schmid factor ratio of 1.14 and considering the data in Fig. 15a, we immediately have that the resolved stress for the system II, which corresponds to the beginning of the LT step, almost coincides with the reference stress, i.e. the resolved stress for the system I in the case of one-step LT deformation. Slip systems of single dislocations can operate at orientations [010] and [011], at which no stress macrojump was observed in TiAl [47, 48]. From Table it is seen that their maximum Schmid factors (f  0.405) coin- cide with the corresponding factors of superdislocation systems. Conse- quently, in addition to superdislocation slip systems, two slip systems of single dislocations corresponding to diagonals of the (001) face can be switched on at the HT step. Note that they have equal Schmid factors at the given orientations. The ‘second diagonal inhibit’ does not apply to these single dislocations, because their blocking is not associated with formation of Kear–Wilsdorf barriers. Both slip systems of single disloca- tions with mutually perpendicular Burgers vectors were actually observed in [32] (see Fig. 9c) and also in [11]. In this case, as can be seen from Fig. 17, any of possible octahedral slip planes includes blocked disloca- tions of the framework. Consequently, the framework cannot turn trans- parent to any new slip system on transition to the LT step. No stress macrojump is observed either at the orientation [001], al- though single dislocations are absent. However, superdislocations with three possible Burgers vectors type 1/2112] have a maximum Schmid factor. It is not improbable that these slip systems can be switched on. In this case the framework structure will be more complicated than the one considered above. Such framework can hardly turn transparent on transi- tion to the LT step. A feature in common of orientations [010] and [001], which are not TABLE 1. Maximum Schmid factors for possible types of dislocations. [11] [47] [48] [48] [ 2 51] [010] [011] [001] 1/2110] 0.381 0.405 0.405 0 101] 0.490 0.405 0.405 0.405 1/2112] 0.424 0.230 0.230 0.470 Anomalies in Deformation Behaviour of TiAl Intermetallic 45 equivalent in TiAl, is that slip systems of single dislocations or superdislo- cations 1/2112] can be switched on in addition to slip systems of su- perdislocations 101]. Different types of dislocations present in TiAl com- plicate the framework structure and make the picture of transition from the HT to the LT step smeared. A stress macrojump is always observed in 001 single crystals of Ni3Al, unlike in TiAl, during two-step deforma- tion [43, 45]. These two intermetallics need be analysed at same orienta- tions to determine causes responsible for their different or similar behav- iour during two-step deformation. 6. CONCLUSION This review offers a wider concept of the multivalley Peierls relief in TiAl. An earlier model [20] included a blocking mechanism of single disloca- tions by their immersion in deep Peierls valleys. That mechanism ex- plained observation of single dislocations in glissade and sessile forms. But it failed to describe the whole curve y(T). Using TEM data on evolu- tion of the dislocation structure and comparing deformation curves for TiAl with typical curves for other materials, we made an effort to recon- struct the shape of the potential relief for a dislocation in TiAl. In our opin- ion, a complicated potential relief, which is schematically shown in Fig. 3d, corresponds to the deformation behaviour of TiAl. This relief includes (i) potential wells of different depth corresponding to two types (shallow and deep) traps, and (ii) potential barriers of different height: low and high barriers for capture of dislocations in shallow and deep traps respectively. These representations were used to describe a nonmonotonic tempera- ture dependence of the yield stress. It was conjectured that the shape of the potential relief for a dislocation changes in the plastic zone of a mi- crocrack. The supposition on the presence of two types of dislocation traps is confirmed, firstly, by the fact that the yield stress drops first in the low- temperature interval and then in the high-temperature interval, and, sec- ondly, by observation of blocked dislocations both at extremely low tem- peratures and in the region of an anomalous trend of y(T). One may think that pulsation of the wide core of a dislocation, which is due to thermal fluctuations, leads to local narrowing of the core. Such random changes of the core structure are fixed thanks to the capture of a recom- bined segment in deep valleys in the case of a single dislocation or cross- slip to the cube plane in the case of a superpartial dislocation. Ultimately, a dislocation is captured in a deep trap. Numerous unusual features of the deformation behaviour of TiAl, which apparently are independent of one another, actually stem from a single effect. This effect is thermally activated blocking of dislocation sources. It is the temperature dependence of the dislocation source switch-on stress F(T) that determines a temperature anomaly of y(T). 46 B. A. Greenberg and M. A. Ivanov The fact that the stress F is independent of the strain rate explains a weak strain-rate sensitivity of the flow stress observed in the interval of the anomalous trend. All other factor being equal, a system with a lowest F operates out of other slip systems. This consideration can account for disappearance of certain types of dislocations in certain temperature in- tervals. Concentration of stresses causes blocking of dislocation sources near a microcrack, possibly leading to TiAl embrittlement. The stress F fits naturally the condition (26), which determines the onset of plastic flow in description of prestraining experiments. We assume that when a stress macrojump is observed, the low-temperature step begins at the switch-on stress of a new slip system, for which a dislocation framework inherited from HT step is transparent. ACKNOWLEDGEMENTS The authors would like to thank INTAS-97-31994 and the Russian Fun- damental Research Foundation (Grant No. 98-02-17278) for financial support. REFERENCES 1. D. P. Pope and S. S. Ezz, Int. Metall. Rev., 29: 136 (1984). 2. T. Suzuki, Y. Mishima, and S. Miura, ISIJ International, 29: 1 (1989). 3. H. A. Lipsitt, D. Schechtman, and R. E. Schafric, Metall. Trans. A, 6: 1991 (1975). 4. D. M. Dimiduk, Gamma Titanium Aluminides (Eds. Y.-W. Kim et al.) (Warrendale, PA, USA: TMS: 1995), p. 3. 5. Y. Umakoshi, H. J. Yasuda, and T. Nakano, Intermetallics, 4 Supplement: p. S65 (1996). 6. D. Banerjee, A. K. Gogia, T. K. Nandi, and V. A. Joshi, Acta Met., 36: 871 (1988). 7. D. Banerjee, Phil. Mag. A, 72: 1559 (1995). 8. P. K. Sagar, D. Banerjee, K. Muraleedharan, and Y. U. R. K. Prasad, Met. Trans. A, 27: 2593 (1996). 9. T. Kawabata, T. Kanai, and O. Izumi, Acta Met., 33: 1355 (1985). 10. T. Kawabata, T. Abumiya, T. Kanai, and O. Izumi, Acta Met. Mater., 38: 1381 (1990). 11. H. Inui, M. Matsumuro, D.-H. Wu, and M. Yamaguchi, Phil. Mag. A, 75: 395 (1997). 12. B. A. Greenberg and M. A. Ivanov, Mat. Sci. Eng. A, 153: 356 (1992). 13. M. A. Ivanov, B. A. Greenberg, and T. O. Barabash, Phys. Met. Metallogr., 86: 240 (1998). 14. M. A. Ivanov, B. A. Greenberg, and N. A. Kruglikov, Gamma Titanium Aluminides (Eds. Y.-W. Kim et al.) (Warrendale, PA, USA: TMS: 1999), p. 256. 15. M. A. Ivanov, B. A. Greenberg, and N. A. Kruglikov, Phys. Met. Metallogr., 89: in press (2000). 16. V. I. Trefilov, Yu. V. Mil’man, and S. A. Firstov, Physics of Refractory-Metals’ Strength (Kiev: Naukova Dumka: 1975) (in Russian). Anomalies in Deformation Behaviour of TiAl Intermetallic 47 17. V. Vitek and M. Yamaguchi, Interatomic Potentials and Crystalline Defects (Ed. J. K. Lee) (Warrendale, PA, USA: TMS-AIME Publ: 1981), p. 223. 18. H. Alexander, Dislocations in Solids (Ed. F. R. N. Nabarro) (Amsterdam: Elsevier Sci. Publ: 1986), Vol. 7, p. 113. 19. P. Veyssiere, ISIJ International, 31: 1028 (1991). 20. B. A. Greenberg, V. I. Anisimov, Yu. N. Gornostirev, and G. G. Taluts, Scripta metall., 22: 859 (1988). 21. S. Rao, C. Woodward, J. Simmons, and D. M. Dimiduk, High Temperature Or- dered Intermetallic Alloys VI: MRS Symp. Proc. (1995), Vol. 364, p. 129. 22. G. Hug, A. Loiseau, and A. Lasalmonie, Phil. Mag. A, 54: 47 (1986). 23. D. M. Wee, D. P. Pope, and V. Vitek, Acta metall., 32: 829 (1984). 24. B. A Greenberg and M. A. Ivanov, Gamma Titanium Aluminides (Eds. Y.-W. Kim et al.) (Warrendale, PA, USA: TMS: 1995), p. 299. 25. B. A Greenberg and M. A. Ivanov, Met. Phys. Adv. Tech., 16: 1159 (1997). 26. B. A Greenberg and M. A. Ivanov, Mat. Sci. Eng. A, 239: 813 (1997). 27. B. A Greenberg and M. A. Ivanov, Phys. Met. Metallogr., 78: 247 (1994). 28. G. Hug, A. Loiseau, and P. Veyssiere, Phil. Mag. A, 57: 499 (1988). 29. G. Hug, A. Loiseau, and P. Veyssiere, Revue Phys. Appl., 23: 673 (1988). 30. S. A. Court, V. K. Vasudevan, and H. L. Fraser, Phil. Mag. A, 61: 14 (1990). 31. B. A. Greenberg, O. V. Antonova, V. N. Indenbaum, L. E. Karkina, A. B. Notkin, M. V. Ponomarev, and L. V. Smirnov, Acta Metall. Mater., 39: 233; ibid., 243 (1991). 32. B. A. Greenberg, O. V. Antonova, L. E. Karkina, A. B. Notkin, and M. V. Ponoma- rev, Acta Metall. Mater., 40: 815; ibid., 823 (1992). 33. M. A. Morris, Phil. Mag. A, 68: 237; ibid., 259 (1993). 34. M. A. Morris, Phil. Mag. A, 69: 129 (1994). 35. B. Viguier, K. J. Hemker, J. Bonneville, F. Louchet, and J. L. Martin, Phil. Mag. A, 71: 1295 (1995). 36. S. Sriram, D. M. Dimiduk, P. M. Hazzledine, and V. K. Vasudevan, Phil. Mag. A, 76: 965 (1997). 37. B. A. Greenberg, O. V. Antonova, and A. Yu. Volkov, Intermetallics, 7: 1219 (1999). 38. B. A. Greenberg, G. Hug, O. V. Antonova, T. S. Boyarshinova, Z. M. Pesina, I. N. Sachanskaya, and A. Yu. Volkov, Intermetallics, 5: 297 (1997). 39. V. K. Vasudevan, S. A. Court, P. Kurath, and H. L. Fraser, Scripta Met., 23: 467 (1989). 40. S.-C. Huang and E. L. Hall, Metall. Trans. A, 22: 427 (1991). 41. R.M. Thomson, Phys. Metallurgy (Eds. R.W. Cahn and P. Haasen) (Elsevier Sci. Publ. BV: 1983), p. 1487. 42. J. Di Persio and B. Escaig, Dislocations 1984 (Eds. P. Veyssiere et al.) (Paris, France: CNRS: 1984), p. 267. 43. M. Dimiduk and T.A. Parthasarathy, Phil. Mag. Lett., 71: 21 (1995). 44. X. Shi, G. Saada, and P. Veyssiere, Phil. Mag. A, 73: 1419 (1996). 45. N. D. Bakhteyeva, B. A. Greenberg, A. V. Nemchenko, and Yu. N. Akshentsev, Phys. Met. Metallogr., 85: 481 (1998). 46. S. J. Basinski and Z. S. Basinski, Dislocations in Solids: IV Dislocations in Metal- lurgy (Ed. F. R. N. Nabarro) (Amsterdam, North-Holland: 1979), Vol. 4, p. 261. 47. M. A. Stucke, D. M. Dimiduk, and P. M. Hazzledine, High Temperature Ordered Intermetallic Alloys V: MRS Symp. Proc. (1993), Vol. 228, p. 471. 48. R. Mahapatra, Y. T. Chou, and D. P. Pope, Mat. Sci. Eng. A, 239: 456 (1997). 48 B. A. Greenberg and M. A. Ivanov 49. S. Ezz and P. B. Hirsch, Phil. Mag. A, 73: 1969 (1996). 50. N. D. Bakhteyeva, B. A. Greenberg, A. V. Nemchenko, Yu. N. Akshentsev, M. A. Ivanov, and Eu. N. Khlystov, Izvestiya Akad. Sci., 63: 42 (1999). 51. B. A. Greenberg, M. A. Ivanov, T. O. Barabash, and A. G. Blokhin, Phys. Met. Metallogr., 81: 374; ibid., 381 (1996). 52. B. A. Greenberg and M. A. Ivanov, Met. Phys. Adv. Tech., 17: 973 (1999). 53. B. A. Greenberg and M. A. Ivanov, Phys. Met. Metallogr., 79: 1373 (1995). 54. A. E. Staton-Bevan and R. D. Rawlings, Phys. status solidi (a), 29: 613 (1975). 55. B. A. Greenberg and M. A. Ivanov, Metallofiz. Noveishie Tekhnol., 21, No. 7: 55 (1999). 56. B. A. Greenberg and M. A. Ivanov, Metallofiz. Noveishie Tekhnol., 21, No. 8: 3 (1999).

Anomalies in Deformation Behaviour of TiAl Intermetallic

Репозитарії

Схожі ресурси