Eutectoid steel strips are designed for the production of parts for intensive use such as clutches, seat slides, and springs as they exhibit<p>excellent strength levels and wear resistance. These properties arise from the unique morphology of lamellar pearlite which can be considered<p>as a self-laminated nanoscale composite. However, a spheroidization annealing step is nowadays necessary to improve the cold forming properties before further cold rolling steps.<p>This thesis is aimed at improving the tensile ductility of the hot rolled products of eutectoid composition in order to eliminate the intermediate<p>annealing step. Two strategies are proposed.<p>The first is to transpose the concept of controlled rolling developed for HSLA to<p>eutectoid steels. Through a strict adjustment of the austenite processing and of the cooling strategy, it is possible to improve the ductility<p>of the final lamellar microstructure. The way the processing parameters influence the hot deformation of austenite, the eutectoid transformation and of the subsequent spheroidization annealing is deeply<p>investigated. It is found that refinement and pancaking of austenite<p>is beneficial as it reduces the pearlite block size improving the total<p>tensile elongation. Accelerated cooling is of paramount importance to<p>achieve fine Interlamellar spacing (ILS), which lead to high strength<p>levels and accelerate spheroidization during subsequent annealing.<p>The second approach involves intercritical or warm deformation. Warm processing of eutectoid steels is first explored by torsion testing<p>and then up-scaled to a pilot rolling-line. The interactions between thermomechanical parameters, rolling forces generated and microstructural<p>evolution are carefully scrutinized. During concurrent hot deformation, spheroidization of cementite takes place almost instantaneously<p>in both torsion and rolling. The restoration processes occurring in the ferrite matrix depends on the strain path and the strain rates. Low strain rates (0,1 s−1) and simple shear promotes the formation of a recrystallized-like HABs network of about 3μm in size.<p>Plane strain compression and high strain rates (10 s−1) leads to the formation of a typical recovered dislocation substructure (LABs) of 1μm in size. During annealing, no recrystallization occurs and the LABs substructure remains stable. This substructure influences drammatically the mechanical properties: the strength is very high and the work-hardening behavior is poor due to high recovery rate in the region close to the LABs. However, due to the presence of spheroidized<p>cementite particles the ductility of warm rolled eutectoid steels is higher than that of ultra fine grained low carbon steels. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Abstract This research work focuses on how to improve the toughness of heat affected zones (HAZs) of low heat input welds in the case of high strength thermomechanically processed (TMCP) and recrystallization controlled rolled and accelerated cooled (RCR) plates with yield strengths of 355–500 MPa. Experimental work was aimed at the investigation of the intragranular nucleation of acicular ferrite or bainite in hot-rolled plates and the evaluation of the Charpy V and CTOD toughness of the most critical sub-zones of the weld HAZ using simulated specimens with a cooling time t8/5 = 5 s. The zones studied were the coarse grained HAZ (CGHAZ), the intercritically reheated coarse-grained HAZ (ICCGHAZ) and the intercritical HAZ (ICHAZ), the metallographical analyses consisted of microstructural investigations complemented with hardness measurements. Optical, scanning and transmission electron microscopy techniques were employed together with image and electron backscatter diffraction (EBSD) analysis. The test results showed that the toughness of the various sub-zones of the HAZ is improved by promoting intragranularly nucleated ferritic-bainitic (acicular) microstructure in both the CGHAZ and in the base plate. In this way, the sub-zones subjected to intercritical thermal cycles (the ICCGHAZ and the ICHAZ) develop evenly distributed M-A constituents between ferrite and bainite laths. These favourable microstructures can be achieved by using titanium killing or by avoiding niobium microalloying by using copper plus nickel alloying instead. In the laboratory experiments titanium killed steel, containing titanium-manganese oxide/manganese sulphide inclusions with a number density of 300–750 particles/mm2, develops a largely acicular ferritic microstructure in the base plate provided the austenite grain size is greater than about 120 μm and the cooling rate is in the range 6–11 °C/s down to 500 °C. Under plate mill conditions, no significant amount of acicular ferrite could be obtained, because it was not possible to achieve austenite grain sizes larger than about 70 μm after rolling. However, a significant fraction of acicular ferritic-bainitic microstructure was achieved in the CGHAZ, when the austenite grain size exceeded 90 μm. When achieved, a uniform distribution of M-A particles in an acicular ferritic-bainitic microstructure improves toughness. Cracks nucleate at numerous sites on M-A/ferrite boundaries or bainite packet interfaces, but they are initially arrested in the acicular matrix. Crack growth finally occurs by linking of the numerous arrested microcracks.

Cobalt Ferrite has important, size-dependent magnetic properties. Consequently, an overview of particle size is important. Co-precipitation in air was the fabrication method used because it is comparatively simple and safe. The effects of three different reaction times including 1, 2, 3 hour(s) on particle size were compared. Also, the effectiveness of three different capping agents (Oleic Acid, Polyvinylpyrollidone (PVP), and Trisodium Citrate) in reducing aggregation and correspondingly particle size were examined. Using Welch’s analysis of variance (ANOVA) and the relevant post hoc tests, there was no significant difference (p=0.05) between reaction times of 1 hour and 2 hours, but there was a significant difference between reaction times of 2 hours and 3 hours. Potentially, because of increased coarsening for the 3 hour reaction time. PVP and Oleic Acid were shown to be effective in reducing aggregation; however, Citrate was not effective. Possibly, the synthesis procedure was inadequate.

The following investigations have been carried out in this thesis 1)Cordierite is already well known for its low thermal expansion behaviour. Chemical substitutions at various octahedral and tetrahedral sites have been done and their thermal expansion characteristics have been studied Synthesis of cordierite in more reactive environment provided by AlF3 used as sintering aid has been attempted 2) Diffuse ferroelectric phase transition of lead based perovskite materials leads to low expansion region. Solid solutions of lead iron niobate with lead titanate is investigated to increase the structural distortion and see it this low expansion region can be extended to wider temperature Preparation of materials with higher tetragonal distortion In PbTi03- BlFeO3 system is undertaken to study the thermal expansion anisotropy. 3) Composites between lead iron niobate(+(x) and lead titanate (-(x below Tc) has been undertaken to prepare low expansion hulk over a wide temperature range 4) Acoustic emission has been employed as a tool to detect the microcracking in solid solutions between PFN1-x, PTx, and PT1-x, ,BFx, It is hoped to understand relation between magnitude of lattice distortion transition temperature and microcracking in ceramics of the class of materials.

This thesis presents experimental and related theoretical studies of pyrochlore titanate oxides, boron nitride nanotubes, and multiferroic bismuth ferrite. We have investigated these systems at high pressures and at low temperatures using Raman spectroscopy. Below, we furnish a synoptic presentation of our work on these three systems. In Chapter 1, we introduce the systems studied in this thesis, viz. pyrochlores, boron nitride nanotubes, and multiferroic BiFeO3, with a review of the literature pertaining to their structural, electronic, vibrational, and mechanical properties. We also bring out our interests in these systems. Chapter 2 includes a brief description of the theory of Raman scattering and infrared absorption. This is followed by a short account of the experimental setups used for Raman and infrared measurements. We also present the technical details of high pressure technique including the alignment of diamond anvil cells, gasket preparation, calibration of the pressure, etc. Chapter 3 furnishes the results of our pressure-and temperature-dependent studies of pyrochlore oxides which has been divided into eight different parts. In recent years, magnetic and thermodynamic properties of pyrochlores have received a lot of attention. However, not much work has been reported to address the quasiparticle excitations, e.g., phonons and crystal-field excitations in these materials. A material that shows exotic magnetic behavior and high degree of degenerate ground states can be expected to have low-lying excitations with possible couplings with phonons, thereby, finger-printing various novel properties of the system. Raman and infrared absorption spectroscopies can, therefore, be used to comprehend the novel role of phonons and their role in various phenomena of frustrated magnetic pyrochlores. Recently, there have been reports on various novel properties of these systems; for example, Raman and absorption studies [Phys. Rev. B 77, 214310 (2008)] have revealed a loss of inversion symmetry in Tb2Ti2O7 at low temperatures which has been suggested as the key reason for this frustrated magnet to remain in spin-liquid state down to 70 mK. Powder neutron-diffraction experiments [Nature 420, 54 (2002)] have shown that an application of isostatic pressure of about 8.6 GPa in spin-liquid Tb2Ti2O7 induces a long-range magnetic order of the Tb3+ spins coexisting with the spin-liquid phase ascribing this transition to the breakdown of the delicate balance among the various fundamental interactions. Moreover, Raman and x-ray studies have shown that Tb2Ti2O7,Sm2Ti2O7,and Gd2Ti2O7 undergo a structural transition followed by an irreversible amorphization at very high pressures (~ 40 GPa or above) [Appl. Phys. Lett. 88, 031903 (2006)]. In this chapter, therefore, we present our temperature-and pressure-dependent Raman studies of A2Ti2O7 pyrochlores, where ‘A’ is a trivalent rare-earth element (A = Sm, Gd,Tb, Dy,Ho, Er,Yb, and Lu; and also Y). Since all the group theoretically predicted Raman modes of this cubic lattice are due to oxygen vibrations only, in Part (A), we revisit the phonon assignments of pyrochlore titanates by performing Raman measurements on the O16 /O18 − isotope based Dy2Ti2O7 and Lu2Ti2O7 and find that the vibrations with frequencies below 250 cm−1 do not involve oxygen atoms. Our results lead to a reassignment of the pyrochlore Raman phonons thus proposing that the mode with frequency ~ 200 cm−1, which has earlier been known as an F2g phonon due to oxygen vibration, is a vibration of Ti4+ ions. Moreover, we have performed lattice dynamical calculations using Shell model that help us to assign the Raman phonons. In Part (B), we have explored the temperature dependence of the Raman phonons of spin-ice Dy2Ti2O7 and compared with the results of two non-magnetic pyrochlores, Lu2Ti2O7 and Y2Ti2O7. Our results reveal anomalous red-shift of some of the phonons in both magnetic and non-magnetic pyrochlores as the temperature is lowered. The phonon anomalies can not be understood in terms of spin-phonon and crystal field transition-phonon couplings, thus attributing them to phonon-phonon anharmonic interactions. We also find that the anomaly of the disorder activated Ti4+ Raman vibration (~ 200 cm−1) is unusually high compared to other phonons due to the large vibrational amplitudes of Ti4+-ions rendered by the vacant Wyckoff sites in their neighborhood. Later, we have quantified the anharmonicity in Dy2Ti2O7. We have extended our studies on spin-ice compound Dy2Ti2O7 by performing simultaneous pressure-and temperature-dependent Raman measurements, presented in Part (C). We show that a new Raman mode appears at low temperatures below TC ~ 110 K, suggesting a structural transition, also supported by our x-ray measurements. There are reports [Phys. Rev. B 77, 214310 (2008), Phys.Rev.B 79, 214437 (2009)] in the literature where the new mode in Dy2Ti2O7 at low temperatures has been assigned to a crystal field transition. Here, we put forward evidences that suggest that the “new” mode is a phonon and not a crystal field transition. Moreover, the TC is found to depend on pressure with a positive coefficient. In Part (D), we have presented our results of temperature-and pressure-dependent Raman and x-ray measurements of spin-frustrated pyrochlores Gd2Ti2O7, Tb2Ti2O7,and Yb2Ti2O7. Here, we have estimated the quasiharmonic and anharmonic contributions to the anomalous change in phonon frequencies with temperature. Moreover, we find that Gd2Ti2O7 and Tb2Ti2O7 undergo a subtle structural transition at a pressure of ~ 9 GPa which is absent in Yb2Ti2O7. The implication of this structural transition in the context of a long-range magnetically ordered state coexisting with the spin-liquid phase in Tb2Ti2O7 at high pressure (8.6 GPa) and low temperature (1.5 K), observed by Mirebeau et al. [Nature 420, 54 (2002)], has been discussed. As we have established in the previous parts that the anomalous behavior of pyrochlore phonons is due to phonon-phonon anharmonic interactions, we have tuned the anharmonicity in the first pyrochlore of the A2Ti2O7 series, i.e., Sm2Ti2O7,by replacing Ti4+-ions with bigger Zr4+-ions, presented in Part (E). Our results suggest that the phonon anomalies have a very strong dependence on the ionic size and mass of the transition element (i.e., the B4+-ion in A2B2O7 pyrochlores). We have also observed signatures of coupling between a phonon and crystal-field transitions in Sm2Ti2O7. In Part (F), we have studied spin-ice compound Ho2Ti2O7 and compared the phonon anomalies with the stuffed spin-ice compounds, Ho2+xTi2−xO7−x/2 by stuffing Ho3+ ions into the sites of Ti4+ with appropriate oxygen stoichiometry. We find that as more and more Ho3+-ions are stuffed, there is an increase in the structural disorder of the pyrochlore lattice and the phonon anomalies gradually disappear with increasing Ho3+-ions. Moreover, a coupling between phonon and crystal field transition has also been observed. In Part (G), we have examined the temperature dependence of phonons of “dynamical spin-ice” compound Pr2Sn2O7 and compared with its non-pyrochlore (monoclinic) counterpart Pr2Ti2O7. Our results conclude that the anomalous behavior of phonons is an intrinsic property of pyrochlore structure having inherent vacant sites. We also find a coupling between phonon and crystal-field transitions in Pr2Sn2O7. In the last part of this chapter, Part (H), we present our Raman studies of Er2Ti2O7. Here, we show that in addition to the anomalous phonons, there are modes that originate from photoluminescence transitions and some of these luminescence lines show anomalous temperature dependence which have been understood using the theory of optical dephasing in crystals, developed by Hsu and Skinner [J. Chem. Phys. 81, 1604 (1984)]. Temperature dependence of a few Raman modes and photoluminescence bands suggest a phase transition at 130 K. In Chapter 4, we furnish our pressure-dependent Raman studies of boron nitride multi-walled nanotubes (BNNT) and hexagonal boron nitride (h-BN) and compare the results with those of their carbon counterparts. Using Raman spectroscopy, we show that BNNT undergo an irreversible transition at ~ 12 GPa while the carbon counterpart, multi-walled carbon nanotubes, show a similar transition at a much higher pressure of ~ 51 GPa. In sharp contrast, the layered form of both the systems (i.e. h-BN and graphite) undergo a hexagonal to wurtzite phase at nearly similar pressure (~ 13 GPa of h-BN and ~ 15 GPa for graphite). A molecular dynamical simulation on boron nitride single-walled nanotubes has also been undertaken that suggests that the polar nature of the B−N bonds may be responsible for the irreversibility of the pressure-induced transformations. It is interesting to see that in hexagonal phase both the systems have almost similar mechanical property, but once they are rolled up to make nanotubes, the property becomes quite different. Chapter 5 presents the temperature dependence of the Raman modes of multiferroic thin films of BiFeO3 and Bi0.7Tb0.2La0.1O3. Though there have been several Raman investigations of BiFeO3 in literature, here we emphasize the observation of unusually intense second order Raman phonons. Our results have motivated Waghmare et al. to suggest a theoretical model to explain the anomalously large second order Raman tensor of BiFeO3 in terms of an incipient metal-insulator transition. In Chapter 6, we summarize our findings on the three different systems, namely, pyrochlores, boron nitride nanotubes, and BiFeO3 and highlight a few possible experiments that may be undertaken in future to have a better understanding of these systems.

A utilização do eletrodo tubular no processo de soldagem a arco submerso leva a um aumento na produtividade com um custo relativamente baixo, pois não requer investimentos complementares em equipamentos. A mudança do eletrodo sólido para o tubular acarreta variações microestruturais no metal de solda que influenciam as propriedades mecânicas, tais como dureza, tração e tenacidade ao impacto. Este trabalho tem como objetivo principal estudar a tenacidade à fratura dinâmica do metal de solda, mostrando que o uso do eletrodo tubular diminui a propagação de trincas ao impacto, além de melhorar as propriedades de dureza e tração. Para a confecção da junta soldada foi utilizado o aço ASTM-A516 com espessura de 37,5 mm; para o metal de solda, o eletrodo sólido AWS EM12K, com diâmetro de 4,0 mm, e fluxo AWS F6A4 e o eletrodo tubular AWS E71T-5, com diâmetro de 4,0 mm, fabricado em caráter experimental, untamente com os fluxos AWS F6A4 e AWS F7A8 (Fluxo Neutro). Os ensaios de impacto, dureza e tração foram realizados em corpos de prova com e sem alívio de tensão, e o ensaio Charpy instrumentado em corpos de prova entalhados e em corpos de prova entalhados e com trinca por fadiga, nas temperaturas de 25, 200, 400 e 600OC. O metal de solda utilizando eletrodo sólido foi o que mostrou menor percentual de ferrita acicular (54%), enquanto que o utilizando eletrodo tubular e fluxo AWS F6A4 foi o que mostrou maior percentual de ferrita acicular (89%). Os ensaios de dureza mostraram uma influência do microconstituinte ferrita acicular: maior dureza para o maior percentual de ferrita acicular. Os limites de resistência e de escoamento também sofreram o efeito deste microconstituinte, apresentando maiores valores quando os percentuais de ferrita acicular foram maiores. Para a análise da tenacidade para o início de propagação da trinca foram utilizados os métodos da variação da taxa da flexibilidade elástica, o método da energia da carga máxima revisada e o método da energia da carga máxima. Foram calculados os valores de JId e de KJd (K derivado de J). Os únicos valores validados, de acordo com a norma, foram os de JId calculados pelo método da variação da taxa da flexibilidade elástica. Os resultados encontrados, utilizando este método, mostraram, à 600OC, a influência da ferrita acicular, sendo que o metal de solda com maior percentual deste microconstituinte forneceu maior valor de tenacidade. Os valores de JId utilizando o método da variação da taxa da flexibilidade elástica para os corpos de prova com trinca por fadiga ficaram próximos aos resultados encontrados quando o ensaio foi realizado em corpos de prova somente entalhados. Os resultados mostraram que a substituição do eletrodo sólido pelo tubular levou à maior tenacidade e a propriedades mecânicas superiores. / The cored wire application in the submerged arc welding process leads to a increase productivity with relative low cost, because it doesn’t require complementary investments in equipments. Changing from solid to cored wire promotes microstructural modification in the weld metal that enhances mechanical properties, such as hardness, stretching and impact toughness. The principal aim of this work is to study the dynamic fracture toughness of the weld metal. The use of cored wire reduces the impact crack propagation, as well as hardness and strength properties. The welded plate was composed of ASTM-A516 steel with 37.5 mm thickness and for making the weld metal was utilized AWS EM12K wire, with 4.0 mm diameter, and AWS F6A4 flux, and AWS E71T-5 cored wire, with 4.0 mm diameter, making in experimental mode, with AWS F6A4 and AWS F7A8 (Neutral Flux) fluxes. Impact, hardness and tensile tests were carried out in specimens with and without stress relief. The instrumented Charpy tests were carried out at notch and notch plus fatigue crack specimens, at 25, 200, 400 and 600OC temperatures. The solid wire weld metal produced 53.9% of acicular ferrite, while the cored wire weld metal and AWS F6A4 flux produced 88.8% of acicular ferrite. The hardness values were influenced by acicular ferrite and showed higher hardness for acicular ferrite higher values. Tensile strength and yield stress data suffered the same effect of this micro constituent also, they had bigger amount when the acicular ferrite percentiles were bigger. For toughness analysis to crack propagation start were utilized the compliance changing rate method, the energy revised method, and the maximum load energy method. JId and KJd (K derived of J) values were calculated according to the standards. The results met for this method had acicular ferrite influence at 600OC, where the weld metal with the biggest micro constituent had bigger toughness value. The JId values utilized the compliance changing rate method for the precracked specimen got values near to the notched specimen. The change solid wire by the cored wire had better toughness, and it gets better the other mechanical properties.

The constantly tightening emission regulations demand the engines to be moreefficient, to get more power out of smaller engines. Higher engine loads andcomponent temperatures are causing more stresses to engine components. Therefore,a company that produces engines wanted to study if it would be possible to increasethe capabilities of the components by optimising the used material. In this final project work, a cylinder head will be studied. The cylinder heads for theengines are made of ductile cast iron. The limits of that material are near safety limits,and therefore a better material is needed. In this work are some previous studiesanalysed and tried to find how to optimise the used material. The optimised materialshould have better thermal conductivity properties combined with sufficient strengthproperties. Previous studies were analysed to gather knowledge of the elements that affect thematerial parameters. Those studies showed that copper, silicon, pearlite fraction, andthe use of chills are the elements to be optimised. Silicon and pearlite fraction waschosen as optimisation parameters because of their effect on the thermal conductivityand strength properties. Copper was chosen as an optimisation variable due to its effecton the pearlite formation. Chills were used to affect the cooling rate and thereby thepearlite formation. The work was made using MAGMASOFT™ simulation software to simulate cylinderhead casting. The simulated cylinder head was divided into 4 parts for the simulations.For those sections were then set targets for pearlite fraction according to previousstudies. The silicon content was kept constant in the simulation, based on the studiespresented in this work. Copper content was simulated with variations from 0 to 0.7weight-%, and chill heights were simulated from 20 to 60 mm and without chills. After simulating the different variables, the results were analysed. Then the selectedcasting simulation result was mapped to finite element simulation mesh to include thelocal material parameters to finite element simulation. With the finite elementsimulation, the estimated lifetime of the component was simulated. By analysing the casting simulation results, an optimal combination was found. Theoptimal material parameters for a cylinder head casting would be copper 0.5weight-%, silicon 1.9 weight-% and chills thicker than 40 mm on the flame plate. Theoptimised material gives more possibilities to develop engines even further when thecomponent demands are growing.

While using iron castings behind temperature of 500 C namely at cyclic heat straining degradation of structure and breaching iron castings. For this purpose introduce cast iron with content of Si and Mo. Focusing on improvement of SiMo51 for increased high-temmperature corrosion-and fatigue life in exhaust-gas temperatures up to 800 C. Thesis set up chemical microheterogeneity, properties, high temperature corosion and microhardness

In s industrial production is more and more requierement to a saving costs. Findig savings has been removed even to the areas using safety parts, f.e. parts of car chassis. This projekct agglomerate posibilities of technology controled cooling proces of die forgings from micro alloy steel, which is right example how we can spare production expenditure. Micro alloy steel 30MNVS6 was studied within the standard production conditions. A lot of tests were made. There where change a cooling rates of proceses without delay after forging proces. One experiment was made for two temperature of forging.proces. Two series of mechanical tests was made and then was observed phenomenon of metalurgical structure changed and mechanical properties.

Super duplex stainless steel (SDSS) has found a wide use in demanding applications such as offshore, chemical and petrochemical industries thanks to its excellent combination of mechanical properties and corrosion resistance. Welding of SDSS, however, is associated with the risk of precipitation of secondary phases and formation of excessive amounts of ferrite in the weld metal and heat affected zone. The present study was therefore aimed at gaining knowledge about the effect of multiple welding thermal cycles on the microstructure and possible sensitization to corrosion of welds in SDSS.Controlled and repeatable thermal cycles were produced by robotic welding. Oneto four autogenous TIG-remelting passes were applied on 2507 type SDSS plates using low or high heat inputs with pure argon as shielding gas. Thermal cycles were recorded using several thermocouples attached to the plates. Thermodynamic calculations and temperature field modelling were performed in order to understand the microstructural development and to predict the pitting corrosion resistance. Etching revealed the formation of different zones with characteristic microstructures: the fused weld zone (WZ) and the heat affected zone composed of the fusion boundary zone (FBZ), next to the fusion boundary, and further out Zone 1 (Z1) and Zone 2 (Z2). The WZ had a high content of ferrite and often nitrides which increased with increasing number of passes and decreasing heati nput. Nitrogen content of the WZ decreased from 0.28 wt.% to 0.17 wt.% after four passes of low heat input and to 0.10 wt.% after four passes of high heatinput. The FBZ was reheated to high peak temperatures (near melting point) and contained equiaxed ferrite grains with austenite and nitrides. Zone 1 was free from precipitates and the ferrite content was similar to that of the unaffected base material. Sigma phase precipitated only in zone 2, which was heated to peak temperatures in the range of approximately 828°C to 1028°C. The content of sigma phase increased with the number of passes and increasing heat input.  All locations, except Z1, were susceptible to local corrosion after multiplere heating. Thermodynamic calculations predicted that a post weld heat treatment could restore the corrosion resistance of the FBZ and Z2. However, the pitting resistance of the WZ cannot be improved significantly due to the nitrogen loss. Steady state and linear fitting approaches were therefore employed to predict nitrogen loss in autogenous TIG welding with argon as shielding gas. Two practical formulas were derived giving nitrogen loss as functions of initial nitrogen content and arc energy both predicting a larger loss for higher heat input and higher base material nitrogen content. A practical recommendation based on the present study is that it is beneficial to perform welding with a minimum number of passes even if this results in a higherheat input as multiple reheating strongly promotes formation of deleterious phases.