EFFECT OF MULTI-PASS HOT ROLLING ON RECRYSTALLISATION BEHAVIOUR OF FERRITIC STAINLESS STEEL TYPE 409
EFFET DES PASSES MULTIPLES DE LAMINAGE A CHAUD SUR LE COMPORTEMENT DE LA RECRISTALLISATION DE L’ACIER
INOXYDABLE FERRITIQUE TYPE 409
R. Benchouieb1, D. Berdjane1,S. Achouri1, O.Ghelloudj1, F. Lemboub1
1 Research Center in Industrial Technologies CRTI, P. O. Box 64 Cheraga 16014 algiers, Algeria r.benchouieb@crti.dz
benchouieb.rachid@yahoo.fr
RESUME
Le but de cette recherche est d’étudier l’effet de la déformation interrompue sur le comportement de la recristallisation de l’acier inoxydable ferritique type 409. Le laminage à chaud a été exécuté sur un laminoir expérimental en utilisant des échantillons rectangulaire de cet acier inoxydable ferritique.
La recherche a été effectuée en utilisant l’acier inoxydable ferritique type 409 qui contient des précipités, utilisant initialement le laminage expérimental des échantillons et les techniques de la métallographie optique quantitative avec seulement une seule passe de déformation. Cet acier inoxydable ferritique à taille de grain fine montre un rythme plus lent dans la recristallisation et ceci est attribué à la présence du titane.
Les effets de la restauration statique et la recristallisation statique complète entre deux passes de déformation sur la recristallisation cinétique subséquente ont été examinés. Dans cet acier inoxydable ferritique lorsque la restauration est présente entre les passes de déformation, la recristallisation subséquente est retardée et cela en comparaison avec celle d’une seule passe avec la déformation combinée. Il parait que ce retardement dépend de la quantité de la restauration statique subi entre les passes de déformation et de la température initiale de laminage. Ainsi l’influence de la restauration précédente sur la recristallisation est sans doute plus prononcée après les plus grands temps de pause. la cinétique de la recristallisation de la déformation interrompue lorsque le matériau est complètement recristallisé entre les passes de déformation s’est accélérée comparativement avec la simple grande passe de déformation.
Mots clés: Acier inoxydable ferritique type 409, précipités, laminage à chaud, restauration dynamique, restauration statique, recristallisation statique.
ABSTRACT
The aim of this research is to study the effect of interrupted deformation on the recrystallisation behaviour of ferritic stainless steel type 409. Hot rolling schedules have been performed on an experimental rolling mill using rectangular slabs of this ferritic stainless steel.
The research has been carried out using material of ferritic stainless steel type 409, which contains precipitates, initially using experimental rolling of small slabs and quantitative optical metallographic techniques with one pass only. Stainless steel type 409 displays a slower rate of recrystallisation than other ferritic stainless steels and this is attributed to the presence of titanium.
The effects of static recovery and full static recrystallisation between two passes on subsequent recrystallisation kinetics have been investigated. In this stainless steel, when recovery occurs between the passes, subsequent recrystallisation is retarded compared with that for a single pass of the combined strain. Increasing the recovery time increases the retardation of recrystallisation even more. Also it has been found that when recrystallisation is complete between the two passes, the rate of subsequent recrystallisation after the second deformation was considerably accelerated.
Keywords: ferritic stainless steel type 409, precipitates, hot rolling, dynamic recovery, static recovery, static recrystallization.
1. INTRODUCTION
All commercial rolling operations involve a series of deformations given in separate passes which are separated by intervals of time as the stock is reversed through a single rolling stand, or travels from one stand to another. In each pass, the strain rate rises rapidly to a maximum and then falls to zero on exit from the rolls. These conditions are considerably more complex than in normal laboratory tests, but it has been found that changing the strain rate during deformation, as in rolling, results in only small changes in static recrystallisation kinetics from the kinetics occurring when tests are carried out at a constant strain rate equal to the mean value for rolling. Simulation of such interrupted deformation schedules has been used in laboratory tests to study the influence of rest periods on static recrystallisation kinetics and hence on structure. The resultant structure from such operations evolves from both the dynamic and static changes in structure. However, it has been found that when partial recrystallisation takes place between two passes subsequent recrystallisation is retarded compared with that for a single pass of the combined strain. With increasing fraction recrystallised after the first pass, there is greater retardation in subsequent recrystallisation time, and a tendency for the final recrystallised grain size to increase after the second pass. In addition when recovery occurs between passes, the recrystallisation rate after single deformation to the same total strain is greater than after the interrupted ones.
The aim of this project is to determine the effect of interrupted deformation on the subsequent recrystallisation behaviour of ferritic stainless steel type 409.
Small slabs of a commercial plate stainless steel type 409 have been tested in hot rolling under isothermal deformation conditions. Several deformation schedules of up to two passes have been performed in order to determine the influence of a first
pass microstructure on the behaviour of the metal during and after a second pass.
Optical and electron microscopy have been systematically used to observe the microstructural features of representative specimens and the quantitative measurements related to the corresponding exprimental conditions. In the future work all this information facilitates a computational procedure to predict recrystallisation behaviour under all deformation conditions using a computer model.
2. MATERIALS AND METHODS Stainless steel type 409 was supplied in the form of hot rolled commercial plate, with thickness of 30mm and initial grain size of 16010µm. The commercial plate analysis is shown in table 1.
Element Weight(%)
C 0,016
S <0,001
P 0,014
Mo 0,03
Ni 0,2
Si 0,42
V 0,05
Cr 12,2
Mn 0,31
Nb <0,02
Ti 0,32
Co <0,02
Cu 0,04
N 161ppm
Table 1. Chemical composition of the ferritic stainless steel type 409 employed.
The plates were machined and ground to obtain specimens of dimensions of 1203015mm thick for the rolling experiments.
Hot rolling experiments for stainless steel type 409 were performed with the same rolling mill, temperature was recorded in the same way. After reheating at temperature of 900oC for 20 minutes, the slabs were hot rolled at a temperature of 870oC from a thickness of 15mm to 8.4mm
and 4.7mm thickness at a controled strain rate of 3.3s-1. The rolling schedules were given in a single pass or double pass sequence with different pauses between passes and then the slabs were water quenched after the last pass. To freeze the as-deformed structure, the time elapsed between finishing the last pass and the quenching operation was of the order of 4 seconds. a total equivalent time for recrystallisation was calculated by equation (1).
QrexRT
dt
QrexRT
teq
exp i exp (1) Where:teq is the equivalent time for annealing.
Ti is the temperature in a particular time interval dt.
T is the annealing temperature.
R is the universal gas constant (R=8.314Jmol-1K-1).
Qrex is the activation energy for static recrystallisation, 120kJ.mol-1 for this ferritic stainless steel type 409 [1].
The metallography for this ferritic stainless steel type 409 was carried out at the position of 0.21WC distance from the side surface (Face X in figure 1), as the temperature history at this point had been found to be close to the average of the cross section [2].
Figure 1. Schematic diagram of the samples taken for annealing. The faces marked X were the planes chosen for metallographic examination.
After sectioning and linishing, specimens were ground down through a series of successively finer silicon carbide papers, 1200 grade was the finest used. Polishing was then performed on diamond wheels in
the sequence 6, 1, 1/4µm grades. Various etches were tried electrolytically and chemically but most of these were unsatisfactory, since substructures in unrecrystallized grains were not revealed.
The chemical etch of Kalling’s reagent containing: 20ml HCl, 15ml H2O, 65ml CH3OH, 1g CuCl2, as well as Schaftmeister’s reagent containing: 50ml HCl, 5ml HNO3, 50ml H2O, were selected finally, since these reagent gave good grain boundary delineation and revealed substructures within the unrecrystallized grains, but it also produced some pitting using electro-etching. The specimens were also chemically etched with stirring the solution for a determined time until the grain boundary delineation was completely revealed. After taking the specimens out of the solution, they were rinsed with water followed by methyl alcohol, and dried in hot air.
3. RESULTS
This research was carried out using ferritic stainless steel type 409 which was received in the form of hot rolled commercial plates with a uniform structure composed of equiaxed grains of size of 160m after a complete recrystallisation. To obtain different starting grain size, slabs of this material were reheated at temperatures of 900oC, 1100oC and 1200oC from 900 seconds up to 2400 seconds and immediately water quenched. The results of this treatment were given in table 2.
Specimen Nbr
Rh.T (°C)
Rh.t (S)
do (µm)
1 900 900 160
2 900 1200 160.5
3 900 1800 162
4 900 2400 161
1 1100 900 170
2 1100 1500 170.8
3 1100 2100 173
1 1200 900 174
2 1200 1500 175
3 1200 2100 176
Rh.T is the reheating temperature (°C).
Rh.t is the reheating time (S).
do is the initial grain size (µm).
Table 2. Effect of reheating temperature and reheating time on initial grain size of the this ferritic stainless steel type 409.
The starting grain size prior to rolling operations was of 160m, which was found to be independent of reheating temperature.
The activation energy for static recrystallisation was found to be 185kJ.mol-1 which was derived from the effect of annealing temperature on recrystallisation kinetics.
In order to determine the isothermal recrystallisation kinetics after deformation to different effective strains at a strain rate of 3.3s-1 where dynamic recovery is the only restoration process during deformation, slabs of this material were deformed at a temperature of 870oC and effective strains of 0.58 and 1.16.
Experimental results are presented in figure 2 as volume fraction recrystallised XV versus Log effective time.
Figure 2. Effect of strain on the recrystallisation kinetics of this material, deformed at a temperature of 870oC, do=160µm.
3.3. Interrupted deformation tests
3.3.1. Effect of recovery on subsequent recrystallisation kinetics
Slabs of ferritic stainless steel type 409 with a starting grain size of 160m were hot rolled in a two rolling pass schedule, to a total effective strain of 1.16 separated by short interpass effective times at a starting rolling temperature of 870oC to give no interpass recrystallisation at testing temperature, to determine the effect of recovery on subsequent recrystallisation kinetics. For this purpose slabs were given a single deformation to an effective strain of 0.58, from which effective times were selected when the material undergoes only static recovery (Figure 3).
Figure 3. Effect of annealing temperature on the recrystallisation kinetics of this material, deformed at a temperature of 870oC, do=160µm.
The volume fraction recrystallised XV as a function of annealing effective time from the last deformation, is illustrated in figure 4. These were hot rolled in two passes to a total effective strain of 1.16 separated by effective time intervals of 40 seconds and 118 seconds.
Figure 4. Effect of recovery on the recrystallisation kinetics of this material, deformed at a starting rolling temperature of 870oC, do=160µm.
From figure 4 where the single deformation and interrupted ones were plotted in the same logarithmic scale, it can be seen that the recrystallisation kinetics of interrupted deformation has slowed down compared with a single deformation performed to the same total effective strain. These retardations seem do not depend on the amount of static recovery undergone between passes.
Avrami-type plots of the recrystallisation data of this material are illustrated in figure 5 as Log[ln(1/(1-x))] versus Log effective time. The time exponent K in the Avrami equation was found to be 0.6 in the range of effective time investigated (40-118 seconds).
Figure 5. Avrami-type plots of the recrystallisation data of this material,
deformed at a starting rolling temperature of 870oC, do=160µm.
In order to see the relationship between the effective time of 50% volume fraction recrystallised and the effective pause time, values of t50% were taken from figure 4 and plotted on a linear scale as function of effective pause time, as shown in figure 6.
Figure 6. Effect of effective pause time (recovery) on the time for 50% static recrystallisation of this material, deformed at a starting rolling temperature of 870oC, do=160µm.
It can be seen from figure 6 that the effective time for 50% volume fraction to be recrystallised after the second pass does slightly vary with the effective pause time after the first pass in the range investigated.
The dependence of average recrystallising grain size on fraction recrystallised under the deformation conditions mentioned above is presented in figure 7.
Figure 7. Dependence of the average recrystallising grain size on the fraction
recrystallised of this material, deformed at a starting rolling temperature of 870oC, do=160µm.
From this figure values of the final extrapolated recrystallised grain size drex were taken and plotted on a linear scale as a function of effective pause time (static recovery), as shown in figure 8.
Figure 8. Effect of effective pause time (recovery) on the recrystallised grain size of this material, deformed at a starting rolling temperature of 870oC, do=160µm.
It can be observed from this figure that the recrystallised grain size after the second pass is relatively insensitive to the pause time after an initial increase. The effect of effective pause time on the relation between recrystallizing grain size and effective time of holding after the last deformation has been determined in figure 9.
Figure 9. Effect of effective time of holding on the recrystallizing grain size of
this material, deformed at a starting rolling temperature of 870oC, do=160µm.
When partial recovery has occurred the grains grow quicker than when recrystallisation has occurred between passes.
3.3.2. Effect of full recrystallisation on subsequent recrystallisation kinetics The effect of full recrystallisation during hot rolling on subsequent recrystallisation kinetics was also investigated. Slabs of this material were given an effective strain of 0.58 at a starting rolling temperature of 870oC, quenched and cut into small specimens and annealed to obtain recrystallisation curves (Figure 2). In addition, slabs of this material were hot rolled in a two rolling pass schedule, to a total effective strain of 1.16 separated by an interpass effective time to give 100%
recrystallisation. The volume fraction recrystallised XV as a function of effective annealing time from the last deformation is illustrated in figure 4 and compared with the results where recovery only occurs. It is evident from figure 4 that the recrystallisation kinetics of interrupted deformation, and fully recrystallised between passes, has speeded up compared with single and interrupted ones performed to the same total effective strain. The time exponent K in the Avrami equation was found to be equal to 0.6.
It can also be seen in figure 9 that the recrystallising grain size after the second pass increases with increasing in time of holding. The growth rate of the recrystallizing grain size is slower compared with those of single and double deformation for a fixed time of annealing.
It is also noticed from figure 9 that the recrystallised grain size when the material is fully recrystallised between passes is smaller than the ones after a single pass and the ones after double passes separated by various recovery times.
3.4. Metallography of reheated, deformed and annealed structure
3.4.1. Single deformation tests
Examination by optical microscopy of specimens quenched from the reheating temperature taken as the starting grain size prior to rolling operations for ferritic stainless steel type 409 are shown in figure 10(a-b).
(a) Reheating temperature 1100oC, reheating time 900 seconds, d0=170µm, X50.
(b) Reheating temperature 900oC, reheating time 900 seconds, d0=160µm, X50.
Figure 10. Optical micrographs of specimens of this material, showing no effect of reheating temperature on the initial grain size.
It can be observed that the structure is composed of uniform equiaxed recrystallised grains. It was also noticed that the recrystallised grain sizes are independent of reheating temperature.
The evolution of the structure as the volume fraction recrystallised changes was examined by looking at the shape and location of the recrystallised grains.
Examination of specimens quenched immediately after the first pass showed deformed grains and no recrystallisation took place following a short period of time of 4 seconds. Subgrain structure within deformed grains at a temperature of 870oC is very poorly developed and not resolvable with the presence of undissolved coarse precipitates with different shapes (Figure 11(a-c)).
(a) X300.
(b) X375.
(c) X300.
Figure 11. Optical micrographs of specimens of this material, deformed at an effective strain of 0.58 and at a temperature of 870oC, showing a poorly developped substructure as well as coarse precipitates with different shapes.
Optical metallography of annealed specimens deformed at a temperature of 870oC indicated that inhomogenieties of recrystallisation in the longitudinal sections occurred through the thickness (Figure 12(a-b)) and that the grain boundary regions were preferred sites for nucleation (Figure 13).
(a) 25% thickness below the surface, annealing effective time 28599 seconds,
X75.
(b) Centre, annealing effective time 28599 seconds, X75.
Figure 12. Optical micrographs of specimens of this material, deformed at an effective strain of 0.58, showing variations in the volume fraction recrystallised through the thickness.
Figure 13. Optical micrographs of specimens of this material, deformed at an effective strain of 1.16 and at a temperature of 870oC, showing grain boundary regions as sites for nucleation.
X150
Observation by transmission electron microscopy was conducted on the specimens deformed at a low temperature of 870oC. Substructure with well-defined elongated subgrains is revealed in figure 14.
Figure 14. Transmission electron micrograph of specimen of this material, water quenching after deformation, showing well defined elongated subgrains deformed at an effective strain of 0.58.
X10500
Network dislocations as well as straight and parallel dislocations were observed in the subgrain matrix (Figure 15).
Figure 15. Transmission electron micrograph of specimen of this material, showing network dislocations as well as straight and parallel dislocations. X6350 It is also apparent that low temperatures of
deformation produce ragged
subboundaries, the subgrain interior has a higher density of dislocation. Figures 14 and 15 show well defined subboundaries surrounding the subgrains, some of them show traces of dissolved subboundaries in their interior.
Electron microscopy showed that static recovery at a temperature of 870oC after a strain of 0.58 caused a smooth appearance of substructure and this seems to be brought about by subgrain coarsening in some areas of the substructure (Figure 16(a-b)).
(a) X4900.
(b) X4900.
Figure 16. Transmission electron micrographs of specimens of this material, deformed at an effective strain of 0.58 and at a temperature of 870oC, annealed at the same temperature (870oC) for 380 seconds.
It is also apparent that some of the subgrains became equiaxed, since although forest dislocations as well as straight and parallel dislocation with some of them appeared to be just on the stage of subboundary formation or annihilation of dislocations are shown during the progress of static recovery and recrystallisation (Figure 16(b)).
Figure 17(a-c) shows an interesting feature about the morphology of the subboundaries, some of them show traces of dissolved subboundaries in their interior. This particular feature was found to be very common in this specimen and therefore suggests subgrain coalescence through the dissolution of subboundaries as a very likely mechanism for the change in size of the subgrains.
(a) X4900.
(b) X10500.
(c) X10500.
Figure 17. Transmission electron micrographs of specimens of this material, showing the morphology of the subboundaries.
Figure 18 shows the presence of very small and dispersed inclusions at subboundaries and inside subgrain matrix, surrounded by dislocations.
Figure 18. Transmission electron micrograph of specimen of this material, showing inclusions at subboundaries and inside subgrains. X18500
Figure 19 shows increased dislocation density associated with a stringer of inclusions.
Figure 19. Transmission electron micrograph of specimen of this material, showing a high dislocation density near a straight line formed of inclusions. X10500 3.4.2. Interrupted deformation tests As mentioned previously the structure which develops as a result of interrupted deformation depends on the amount of recovery and recrystallisation which have taken place between passes.
Figure 20(a-c) shows the structures resulting after a second hot rolling pass following partial recovery and full recrystallisation after the first pass, and a structure resulting from one pass deformation to the same total effective strain of 1.16.
(a) ε=1.16, pause effective time=0 seconds
(b) pause effective time=118 seconds
(c) pause effective time=46599 seconds Figure 20. Optical micrographs of deformed structures of specimens of this material resulting from two rolling passes separated by different time intervals.
ε1=0.58, ε2=0.58, RT1=870oC, do=160µm.
X75
It is clearly seen that the deformed grains in material which has been fully recrystallised after the first pass are smaller and less elongated than the ones which have undergone partial recovery after the first pass.
The accumulated strain of the two passes results in a marked reduction in grain size during recrystallisation after the second pass, and uniform equiaxed grains have been obtained (Figure 21(a-c)).
(a) ε=1.16, pause effective time=0 seconds
(b) pause effective time=118 seconds
(c) pause effective time=46599 seconds Figure 21. Optical micrographs of fully recrystallised structures of specimens of this material resulting from two rolling passes separated by different time intervals. ε1=0.58, ε2=0.58, RT1=870oC, do=160µm. X150
4. DISCUSSION 4.4. Hot deformation
4.4.1. Structural observations
Examination of specimens quenched immediately after a single deformation pass under rolling temperature of 870oC for this material, showed deformed grains and no recrystallisation took place following a short period of time of 4 seconds, indicating that this steel also behaved as typical dynamic recovery type of metals. In stainless steel type 409, the sample deformed at a temperature of 870oC shows that the original grain boundaries are no longer clearly visible (Figure 20), and that some larger regions contain only faint subgrains with evidence for the presence of indissolved coarse
particles within deformed grains and at grain boundaries.
4.5. Static recrystallisation 4.5.1. Structural changes
In stainless steel type 409, for all the deformation conditions investigated static recrystallisation is nucleated preferentially at the original grain boundary and to a lesser extent at grain edges (Figure 13(a)).
Furthermore with increasing time of holding the nucleation takes place at deformation bands (Figure 13(b)). Barbosa and Santos [1] in material of ferritic stainless steel type 430, after cold rolling by 20 and 40% showed that nucleation takes place at grain boundaries only.
Observations in stainless steel type 409, seem to show that grain boundary migration is one of the mechanisms for nucleation. This is in good agreement with the investigation of Uematsu and Yamazaki [3] using material of ferritic stainless steel type 409 after hot rolling by 40% reduction. They found that the nucleation takes place at grain boundaries at temperatures higher than 1150oC by grain boundary migration and takes place at deformation bands at temperature lower than 1150oC. This finding is in line with earlier work by Richardson et al [4] where low dislocation densities in nickel, deformed in creep, lead to a boundary migration type of recrystallisation, whereas larger strains and dislocation densities lead to coalescence as the mechanism of recrystallisation.
4.6. Effect of particles
In general from the published works, solute additions lower the stacking fault energy, which makes dynamic recovery more difficult. This is observed in brass and copper, which does not form a well- defined subgrain structure under deformation conditions of high temperature and strain rate. Similarly, in Zr-Sn alloys, the addition of up to 5% tin reduces the stacking fault energy from 240mJ/m2 to 60mJ/m2 [5]. According to many investigators, in situations where
existing particles are not cut by the dislocations and are relatively stable, they serve as obstacles which help initially to build up the substructure and finally to stabilize it. This effect has been observed at strains not exceeding two in many systems: e.g. Al with Al2O3 particles, Ni with ThO2 particles, Ni-base superalloy with γ´ precipitate, ferrite with carbide precipitates and austenite with NbC particles or other carbides [6,7,8,9]. Under these conditions, the subgrain diameter may be reduced to the order of the interparticle spacing, which leads to an increase in flow stress. The presence of precipitates can prevent the formation of a subgrain structure altogether if the spacing is small enough, or if, as a result of their shearing, they induce the formation of matrix superdislocations, which can not leave their slip planes and also preventing grain boundary migration [10,11].
The results attributed to the influence of microalloying elements in ferritic stainless steel type 409 during static recrystallisation compared with the results obtained with ferritic stainless steel type 430 (Ti free grade) are presented in figure 22.
Figure 22. Isothermal recrystallisation kinetics of ferritic stainless steel type 430 and 409, deformed at an effective strain of 0,58 and at temperature of 870°C.
In our case, the type 409 alloy (0.016%
carbon and 0.32% titanium) in the temperature of 870oC contains titanium giving rise to yellow coarse, angular non- metallic particles, observed by optical microscopy and probably correspond to Ti(CxNy). These particles appear to be
responsible for the retardation of recrystallisation. As is known, second phase particles have a strong retarding effect on recrystallisation when they are small and finely dispersed (which is not the case since small finely dispersed particles were not revealed by transmission electron microscopy. The most important practical application of this effect is in controlled rolling of microalloyed steels in which strain induced precipitation of carbonitrides take place rapidly at high temperature, and when strain induced precipitation starts earlier than the expected starting of recrystallisation, recrystallisation is severely retarded.
However, coarse particles may create local lattice distortion in the matrix around the particles, if a critical strain rate and temperature condition is exceeded [12,13].
These distorted regions cause high local dislocation densities, which may result in particle stimulated nucleation of subsequent static recrystallisation, with important effects on both the kinetics of recrystallisation and the texture developed [14]. These yellow coarse particles Ti(CxNy) seem to affect recrystallisation indirectly through their influence on the preceding recovery processes and on moving grain boundaries since nucleation sites were observed only at grain boundaries (Figure 13(a)). Whatever the cause, it is clear that the retardation of static recrystallisation in this stainless steel is attributable, in a large measure, to the retardation of recovery caused by the titanium addition. The addition of titanium leads to a fully ferritic structure at all temperatures, since carbon and nitrogen, which are strong gamma stabilizers, are tied up in the form of titanium carbonitrides. However, titanium is known to retard recrystallisation kinetics, whether in solid solution or precipitate form [15].
In general, recrystallisation is controlled by precipitation of carbonitrides and clearly variables such as reheating temperature, which determines the solution rates of carbonitrides, amount of deformation/pass,
which affects the amount and mode of precipitation of carbonitrides, and finish rolling temperature, which alters the amount of precipitation on further cooling, must be carefully controlled.
4.7. Interrupted deformation tests
4.7.1. Effect of recovery on subsequent recrystallisation kinetics
In this stainless steel type 409, the results of interrupted tests when only static recovery has taken place have shown that increasing the pause time has a similar effect as stainless steel type 430 (Figure 4).
The rate of subsequent recrystallisation after the second deformation was also considerably retarded compared with that for a single deformation performed to the same total strain. As shown in figure 36, the effect of recovery on the recrystallisation kinetics (as described by the Avrami equation ) was to delay the recrystallisation by displacing the lines in plots of Log[ln(1/(1-x))] versus Log(t) to longer times. Once again it was found that the prior recovery delays the recrystallisation whilst the parameter K remains constant compared with the K value of single deformation for this fine grained material. The influence of recovery on recrystallisation would explain the increasing value of t50% of samples deformed under identical conditions as the effective pause time increases compared with samples deformed in a single pass deformation (Figure 6).
Simulation of such interrupted deformation schedules has been used in laboratory tests to study the influence of rest periods on ductility and structure [16-19], Rossard and Blain [19] showed that, in simulating the rolling conditions of reversing and continuous rolling mills, the schedule of the last few passes and the cooling rate after deformation determined the final structure. Thus, it should be noted that when the effective pause time increases the increase in the final grain size after subsequent recrystallisation was to be expected because of the reduced driving
force for recrystallisation. However further increase in the effective pause time had no effect on recrystallisation.
4.7.2. Effect of full recrystallisation on subsequent recrystallisation kinetics In this ferritic stainless steel type 409, the results of interrupted tests when full recrystallisation has taken place have shown that when the material is recrystallised between two passes, the rate of subsequent recrystallisation after the second deformation was also considerably accelerated compared with that of a single deformation performed to the same total effective strain (Figure 4). Because the grain size is virtually the same after the first pass as the initial grain size, this would not account for the accelerated recrystallisation. There must be some other reason. It is probable that during the very long anneal between passes that many carbides present would have an opportunity to dissolve and carbon to diffuse through the grains and precipitate onto the preexisting large nitride particles.
This would then remove any pinning or drag effect on the grain boundaries as they recrystallise.
By observing the microstructure shown in figure 21(c), it can be seen that the accumulated strain of the two passes, as well as the decrease in rolling temperature from 870oC to 840oC results in a reduction of approximately one third in the final grain size after the second pass, and that the recrystallised structure is constituted of small equiaxed grains. It can thus be seen that static recrystallisation plays a vital role in determining the structure of rolled products. In order to obtain uniform structures in the product it is necessary either to have complete recrystallisation between all passes so that uniform equiaxed grains are obtained, or alternatively, a sufficient number of passes should be given after recrystallisation has ceased so that all grains are heavily deformed and will give uniform structures
either on transformation during cooling or on subsequent treatment.
5. CONCLUSION 5.4. Hot working
The dynamic restoration process which operates during hot working at a constant strain rate of 3.3s-1 and temperature of 870oC is dynamic recovery.
Transmission electron microscopy (TEM) showed that lower temperatures of deformation (870oC) produce ragged subboundaries with subgrain interior more densely populated by dislocations.
In this ferritic stainless steel, at lower temperature of deformation (870oC), the original grain boundaries are no longer visible and some larger regions contain only faint subgrains.
There is evidence of the presence of coarse titanium carbonitrides Ti(CN) within deformed grains and at grain boundaries.
5.5. Static recrystallisation
In this ferritic stainless steel, nucleation of recrystallising grains appears to be associated with triple points and grain boundaries and no evidence of nucleation at particles of titanium carbonitrides was observed.
The recrystallisation kinetics follow Avrami behaviour.
Ferritic stainless steel type 409 displays a slower rate of recrystallisation than stainless steel type 430 and this is attributed to the presence of titanium.
The recrystallised grain size seems to be independent of reheating temperature.
5.6. Interrupted deformation tests
In this ferritic stainless steel when recovery occurs between two passes, subsequent recrystallisation is retarded compared with that for a single pass of the combined strain. Increasing the recovery time further retardation of recrystallisation was observed after the second pass.
With increasing recovery time between two passes there is a tendency for the final
recrystallised grain size after the second pass to increase.
When the recrystallisation is complete between two passes, the rate of subsequent recrystallisation after the second deformation was considerably accelerated compared with that for a single deformation performed to the same total strain.
REFERENCES
[1]. A.J.Craven, K.He, L.A.J.Garvie and T.N.Baker: Acta Mat. Vol48 (2000).
PP.3857-3868.
[2]. A.N.Belyakov and R.O.Kaibyshev.
‘’Deformation mechanisms in a high- chromium ferritic steel’’ The Physics of Metals and Metallography, 79, PP.212- 228, 1995.
[3]. Y.Uematsu and K.Yamazaki: Stainless Steels '87 [ PROC.CONF], York, UK, 14-16 Sept 1987; The institute of Metals, P.386-393.
[4]. M.Militzer, "Modeling of the Interaction of Precipitation and Grain Growth," in Recrystallization and Grain Growth, Eds. G. Gottstein and D.A.
Molodov, (SpringerVerlag Berlin 2001) 361-366.
[5]. K.U.Snowden, P.A.Stathers and D.S.Hughes, Journal of Nuclear Materials, Volume 87, Issues 2–3, 1 December 1979, Pages 297-302.
[6]. S.Kobayashi, C.Zambaldi and D.Raabe: Acta Mater., 2010, vol. 58, pp.
6672–84.
[7]. S.Kobayashi, S.Zaefferer, A.Schneider, D.Raabe and G.Frommeyer:
Intermetallics, 2005, vol. 13, pp. 1296–
1303.
[8]. O.Engler, J.Hirsch and K.Lu¨cke: Acta Metall. Mater., 1995, vol. 43, pp. 121–38.
[9]. D.Raabe and W.Mao: Phil. Mag. A, 1995, vol. 71, pp. 805– 13.
[10]. F.J.Humphreys and M.Hatherly:
Recrystallization and Related Annealing Phenomena, 2nd ed., Elsevier, Oxford, 2004, p. 285.
[11]. R.E.Smallman and R.J.Bishop:
Modern Physical Metallurgy and Materials
Engineering, 6th ed., Butterworth- Heinemann, Oxford, 1999, p. 265.
[12]. P.N.Kalu and F.J.Humphreys., Annealing Processes-Recovery, Recrystallisation and Grain Growth, Proc.7th Riso Int.Symp.
Ed.N.Hansen et al., Riso National Lab.
Roskilde, Denmark, 1986, P.385-390.
[13]. F.J.Humphreys and P.N.Kalu., Acta Met., 1987, 35, P.2815-2829.
[14]. S.Kobayashi, C.Zambaldi and D.Raabe: Acta Mater., 2010, vol. 58, pp.
6672–84.
[15]. B.Feng, D.Dunne and T.Chandra:
The Minerals, Metals and Materials Society, 1990, P.319-324.
[16]. M.M.Farag, C.M.Sellars and W.J.McG.Tegart, Ref.5, P.60.
[17]. T.B.Vangham, Ref.5, P.68.
[18]. F.E.White and C.Rossard, ibid., P.14.
[19]. R.Moses, A.Zarei, Hanzaki, S.Yue and J.J.Jonas, ‘’Precipitation Behavior of IF409 Ferritic Stainless Steel,’’ 35th MWSP Conf. Proc., ISS-AIME, Vol.XXXI, 1994, PP.435-442.
