Vpliv oligoelementov na nekatere lastnosti zelo čistega železa in jekla Influences of residuals on some properties of high purity iron and steel C. Goux, J. Rochette, P. Benaben, R. Tardy, J. Rogez, M. Foucalt-Villard, M. Ghannam, J. Le Coze, B. Saulnier, J. Paire, J. Y. Boos A. UVOD Do danes je zbrano zelo veliko informacij o vplivu elementov, ki nastopajo kot primesi v kovinah, in posebno v jeklih. To je predvsem posledica težav, ki jih povzročajo primesi pri izdelavi in uporabi kovinskih delov. Zaradi tega izrazi, kot npr. »oligoelemcnti« takoj vzbujajo v zavesti metalurgov misli o škodljivem vplivu, kar v splošnem ni nujno in je odvisno od tega, za katere lastnosti gre. Ugoden vpliv nekaterih primesi na določeno lastnost materiala običajno spregledamo. Številni taki primeri so bili odkriti ob pojavu težav pri izdelavi ali uporabi kovinskih delov, izdelanih iz zelo čistih surovin. Praktiki pa so le bili zmožni identificirati škodljive elemente in izdelati učinkovite ukrepe za kompenzacijo njihove prisotnosti. Pri tem so delavnice in obrati imeli vlogo laboratorijev, ker je delo potekalo v industriji. V poskusih, ki naj bi pojasnili vlogo določenega oligoelementa ali grupe elementov, običajno delamo z laboratorijskimi šaržami jekla iste osnovne sestave, ki mu potem dodajamo točno določene količine oligoelementov, oz. določenega oligoelementa. Pri takem načinu dela običajno ne upoštevamo prisotnosti običajnih primesi v osnovnem materialu, npr. jeklu, in pri tem predpostavljamo, da njihov vpliv lahko zanemarimo. V naslednjem enostavnem primeru bomo videli, da na ta način lahko pridemo do napačnih sklepov. Prvotno smo v raziskavah vpliva različnih nečistoč na mehanske lastnosti železa uporabljali kot osnovni material običajno elektrolitsko železo. Tako so bile določene prehodne temperature pri raziskavah vpliva žvepla in različnih ostalih nečistoč na udarno žilavost. Nepričakovano se je pokazalo, da je bila prehodna temperatura približno 250 °C, in sicer neodvisno od vsebnosti žvepla v območju 20—50 ppm. (Vse koncentracije se v nadaljevanju nanašajo na količino mase). Navidezno očividen zaključek je bil, da žveplo v koncentracijah do 50 ppm nima vpliva na žilavost čistega železa. Še bolj presenetljiva je bila ugotovitev, da je prehodna temperatura precej nižja, približno enaka sobni temperaturi, ko smo v istih poskusih kot osnovni material namesto običajnega elektrolitskega železa uporabili zelo čisto železo, ki smo ga izdelali po lastnem rafinacijskem postopku. Poleg tega smo ugotovili, da je prehodna temperatura precej občutljiva na vsebnost žvepla, saj A. INTRODUCTION Up to the present time, a vast amount of information has been gained about the influence of trace elements in metals, particularly in steels. This is mostly the result of troubles caused by such elements in the manufacture and in the current use of metal parts. Because of the way metallurgists have got acquainted with trace elements, the term »residual« quite eommonly imply adverse effects. However, depending on the properties under consideration this is not necessarily true because the influence of a given residual element is usually overiooked if it is favorable: a number of such cases have been discovered when the use of raw materials of high purity instead of common ones resulted in various kinds of difficulties either in the manufacture or in the subsequent use of the parts. Nevertheless, men in the practice have been able to identify in many cases harmful elements and to devise efficient measures in order to counterbalance their pre-sence: working on an industrial scale, the wonk-shop is their laboratory. Furthermore, in order to better inteipret the role of particular element or group of elements in a given steel, it is a common procedure to use laboratory scale heats whose base composition is defined by the steel, the contents of some elements being suitably altered. Doing so, the impor-tant fact is that people disregard the usual residuals of the base steel, assuming that their influence can be negleoted. A simple example wiil show how this can lead to erroneous conclusions.1 In order to test the influence of various impu-rities on the mechanical properties of iron, com-mercial electrolytic iron was used at first as the base material. The transition temperature of the impact strength was determined in the presence of various impurities and among others, of sulphur. Surprisingly it proved to be almost inidepen-dent of the sulphur content in the range which had been considered, that is 20 to 50 ppm: it amounted approximately to 250 °C. The apparently obvious conclusion was that underneath 50 ppm, sulphur does not influence the brittleness of pure iron. Then, the same measurements were made using the first batches of a just started production of refined iron. The stili great surprise was that je naraščala za približno 10 °C za vsak ppm žvepla. Nadalje se je pokazalo, da je vpliv žvepla odvisen od prisotnosti drugih nečistoč. Glavne rezultate lahko rezimiramo v naslednji obliki: 1. Prehodna temperatura čistega železa konti-nuirno narašča z vsebnostjo žvepla in doseže približno 600 °C pri 60 ppm S. 2. Ogljik, če je raztopljen v osnovi, močno zmanjša vpliv žvepla. Tako npr. pri 30 ppm ogljika narašča prehodna temperatura za vsak ppm žvepla samo za 3 °C. 3. Kaže, da oligoelementi, ki so prisotni v elektrolitskem železu, razen žvepla in ogljika, ne vplivajo na obnašanje železa. Dodatek žvepla do približno 20 25 ppm povečuje prehodno temperaturo. Pri večji vsebnosti žvepla se prehodna temperatura skoraj ne spreminja. Do sedaj ni bila možna jasna razlaga teh rezultatov, predvsem pa je nepojasnjena interakcija med žveplom in ostalimi oligoelementi. Omenjeni rezultati pa jasno kažejo, da nekontrolirana prisotnost primesd lahko povzroči težave. Če se povrnemo k industrijskim jeklom, kaže, da raziskave vpliva različnih elementov lahko privedejo do napačnih rezultatov, kar je odvisno od vsebnosti in vrste primesi, ki so prisotne v osnovnem materialu. To je eden izmed vzrokov, zaradi katerih se vedno znova pojavlja problem nečistoč v kovinah kljub številnim in zelo koristnim rezultatom, ki so že doseženi v metalurški industriji. Za boljše razumevanje teh problemov je bil že zdavnaj dan predlog, da se k tovrstnim raziskavam pristopa na naslednji način: Najprej je treba pripraviti osnovni material zahtevane nominalne sestave, in sicer z uporabo čim bolj čistih elementov. Vsebnost nečistoč mora biti tako majhna, da jo zanesljivo lahko zanemarimo. Osnovni sestavi potem dodajamo točno določene količine ustreznih oligoelementov. Navidez je to zelo enostavno, dejansko pa nastopa cela vrsta vprašanj, ki jih tukaj ne moremo obsežno obravnavati. Zadostuje ugotovitev, da vsebnost primesi nikoli ni dovolj majhna, da bi jo lahko »zanesljivo zanemarili«. Celo v ekstremno čistih materialih vedno ostaja odprto vprašanje vpliva prisotnih primesi. To pomeni, da priprava ustreznega jekla osnovne sestave zahteva uporabo ekstremno čistih materialov, ki jih praviloma ni na tržišču in jih moramo zato skrbno pripraviti s posdbnimi postopki. B. PRIPRAVA ZELO ČISTEGA ŽELEZA IN JEKLA i-« V skladu s temi izivajanji smo razvili takšne posebne postopke v našem laboratoriju. Najprej smo razvili postopek za pripravo zelo čistega železa. Potem smo se lotili priprave čistega kroma in niklja v zvezi z nerjavnimi jekli. Glede na to, da the transition temperature was much lower, near room temperature. Furthermore, it was quite sen-sitive to the sulphur content, increasing by some 10 °C per ppm of sulphur. However, subsequent experiment have shovvn that the influence of sulphur may be greatly modified in the presence of other impurities. On the whole, the main results could be summarized as follows: 1. In refined iron, the transition temperature increases continuously as the sulphur content goes up, reaching eventually some 600 °C vvith 60 ppm of sulphur. 2. Carbon, if dissolved in the matrix conside-rably moderates the influence of sulphur. For instance, vvith a carbon content of 30 ppm, the transition temperature increases only by 30C per ppm of sulphur. 3. Apart from sulphur and carbon, the residual elements vvhich are present in electrolytic iron do not seem to influence the behaviour of the metal. Novv, if sulphur is added, it increases the transition temperature as far as its content does not go beyond about 20/25 ppm; then, as mentioned previously, the transition temperature remains practically constant. It has not been possible to clearly explain ali these results, particularly the i.nteraction betvveen sulphur and residual elements. Nevertheless, they shovv hovv uncontrolled trace elements may prove troublesome. Turning back to industrial steels, it appears that, depending on the residuals vvhich are present in base steel, experiments on the influence of various elements can lead to erratic results. This is one of the reasons why the problem of trace elements in metals is permanent-ly reopened, in spite of the many results vvhich are so valuable in the metallurgical industries. As an aid tovvards a better understanding of these problems, it has long been proposed to go a vvay vvhich can be described as follovvs. Considering the constituents of a steel vvhose contents are explicitly specified, a base material of the same nominal composition is prepared using pure elements. The impurity contents must be so Iow that they are certainly negligible. Then, vvell controlled amounts of suitable elements can be added to this base steel. Apparently, this is a very simple procedure. Aetually, it raises a num-ber of questions which cannot be discussed at length here. Suffice it to say that it is never easy to assert that residual impurities are »certainly negligible«. Even in extremely pure materials, the influence of these elements remains alvvavs questionable. This means at least that the prepa-ration of suitable base steels requires materials of exceptionnally high purity. As a general rule, they are not available as commercial produots so they must be prepared by carefully controlled procedures. se lahko nabavi zelo čist silicij, se sedaj ukvarjamo s postopkom za pridobivanje zelo čistega mangana. Razvoj postopka za pripravo vsakega zelo čistega elementa temelji na dveh osnovnih zahtevah: 1. postopek mora omogočiti izdelavo elementa zahtevane čistoče, 2. postopek mora zagotoviti izdelavo zadosti velike količine materiala. Izhajajoč iz v literaturi opisanega postopka rafinacije, v vsakem konkretnem primeru najprej s preliminarnimi poskusi ugotovimo, koliko ta postopek omogoča izpolnitev omenjenih zahtev. Običajno so potrebne manjše prilagoditve ali celo precejšnje spremembe. Nato organiziramo manjše obrate za normalno proizvodnjo po tako razvitem postopku. Na si. 1 vidimo shemo postopka pridobivanja zelo čistega železa4. Izhodni material je običajno elektrolitsko železo, ki ga s pomočjo klora prevedemo pri 300—350 °C v FeCl3. Železov klorid potem raztopimo v konc. HC1. Najpomembnejši del rafi-nacijskega postopka je dvojna ekstrakcija z butil acetatom (C6Hl202). V prvi stopnji železov klorid prehaja v organsko fazo. Pri tem večina nečistoč ostane v kisli raztopini, ki jo nato dekantiramo in dolijemo čisto vodo. V drugi stopnji železov klorid prehaja nazaj v vodno raztopino. Tako dobimo zelo čisto vodno raztopino železovega klorida. Z amoniakom nato precipitiramo železo v obliki Fe-hidroksida, ki ga zelo skrbno približno 10 krat izpiramo s čisto vodo, da odstranimo amonijev klorid. Po sušenju dobljeni železov oksid reduciramo z vodikom v železovo gobo, ki je po vsebnosti vseh nečistoč, razen ogljika in kisika, zelo čista. Gobo lahko potem neposredno uporabimo kot zelo čisto železo v vseh primerih, v katerih so dopustne manjše primesi ogljika in kisika. Ta rafinacijski postopek smo uporabljali v našem laboratoriju pri obsežnih raziskavah vpliva različnih nečistoč na krhkosti železa1-2. B. PREPARATION OF HIGH PURITY IRON AND STEEL1—8 Working along these lines, such procedures have been developed in our laboratory. The initial step was the preparation of iron. Then, in view of studying stainless steels nickel and chromium have been dealt with. As high purity silicon can be purchased, the čase of manganese is now foeing considered. For each element, the procedure to be developed is selected in accordance vvith two general rules: 1. it must be able to yield a product of the required purity. 2. it must be applicable to fairly large quanti-ties of material. Talking in each particular čase a refining process which has been described in the littera-ture, preliminary tests shovv how far it can satisfy these requirements. Some simple adjustments or more important modifications are usually neces-sary. Then small shops are organized for a regular production according to a well defined procedure. Figure 1 shows the general principle of the process for iron4. The starting material is a com-mercial electrolytic iron. It is attacked by chlorine at a temperature of about 300/350 °C and trans-formed into iron chloride. Then FeCl3 is dissolved into a concentrated solution of hydrochloric acid. The most important part of the refining process is now a double liquid-liquid extraction vvith butylacetate (C6H1202). In a first step, the iron chloride goes into the organic phase vvhereas most of the impurities remain in the acid solution. The latter one is decanted and replaced by pure water. In the second step of the process, the iron chloride goes bacik into the aqueous phase. In this way, one gets a solution of very pure iron chloride. Then, iron hydroxide is precipitated by means of ammonia. It is carefully washed vvith very pure vvater: about 10 operations are necessary to remo-ve most of the ammonium chloride. After drying the iron oxide is reduced to iron sponge by means of hydrogen. Aipart from the carbon and oxygen contents, the sponge iron is very pure. It can be used in cases vvhen some carbon and oxygen can be tole-rated. This is vvhat has been done in our labora-tory in an extensive investigation of the influence of various elements upon the brittleness of iron1-2. Novv, if carbon and oxygen must be avoided, or if the metal has to be obtained in a solid form, the iron sponge vvill be melted in a silver vvater-cooled boat by induction heating. In the present state of the technique we currently produce bars up to about 1200 g vvhich are 50 cm in length. With a nevv equipment vvhich has just been tested, bars of 4 kg have been melted; it should be possible to go up to 10 kg. Elektrolitsko Fe © Elec FeCI3 _ trolvtic :hio,- Extraction Cisti Pure FeClj.HjO Kloriranje Glavne izločene nečistoče Ekstrakcija Principal impurities separated off AI,Bi,Capr,Co,Cu.Mo.Mg,NiJPb, Si,Zn, Čisto Fe Slika 1 Shema postopka pridobivanja čistega železa Fig. 1 General principle of the refining procedure for iron. Če se zahteva zelo velika čistoča tudi glede vsebnosti ogljika in kisika ali če moramo imeti kompakten kos zelo čistega železa, železovo gobo pretalimo v vodno hlajeni srebrni ladjici v indukcijski peči. Na sedanji stopnji razvoja tega postopka uspešno izdelujemo palice, dolžine 50 cm, s težo približno 1200 g. Nova oprema, ki smo jo pravkar preizkusili, omogoča izdelavo 10-kilogram-skih palic, vendar smo do sedaj dosegli izdelavo 4-kilogramskih palic. Čeprav je naš postopek rafinacije podoben conski rafinaciji, je v bistvu povsem drugačen, saj odstranitev ogljika in kisika dosežemo z interakcijo staljenega metala z atmosfero v peči. Nadalje, indukcijska tuljava se premika vzdolž železne palice z razmeroma veliko hitrostjo 12 cm/min v obe smeri. S tako consko homogenizacijo poboljšamo homogenost izdelka. Ogljik in kisik iz železa reagirata v atmosferi argona. Tako odstranimo ogljik v obliki CO, preostali kisik odstranimo s čistim vodikom. Standardni postopek sestoji iz 60 prehodov, po 30 v vsako smer. Da bi lahko ugotavljali učinkovitost rafinacij-skega postopka, smo namenili posebno pozornost razvoju ustreznih analitskih metod za določanje vsebnosti najbolj kritičnih nečistoč5'8. V sodelovanju z osmimi francoskimi laboratoriji6 smo potem izvedli obsežen analitsko-razisko-valni program raziskav. Tako smo določili vsebnost 76 elementov v vzorcih, ki so bili izdelani iz 6 palic zelo čistega železa. Dobljene rezultate smo analizirali glede na: 1. reproduktivnost rafinacijskega postopka s primerjavo vsebnosti primesi v posameznih palicah, 2. homogenost posamezne palice in 3. kvaliteto različnih analitskih metod. Vsebnost nečistoč v izhodnem elektrolitskem železu se je spreminjala v precej širokem območju. Tako smo za posamezne elemente dobili naslednje območje, oz. vrednosti (vppm mase): Ti = Značilna analiza pretaljenega čistega železa Typical analysis of the remelted pure iron Vsebnost nečistoč v ut. ppm lmpurity content in wt. ppm Hf Ta W. Re Os lr\ Pt 'Au Hg, 77 Pb Bi, Po At Rn hniiiiii 'lili.ii.ii iml iiiiiiiiiiiiiiiiiriiiTiiiiiffliiiiiiii......."i Ra Ac Bi m lin m Nd II Pm E! i"n £u lini lllll Tb n m »1 llllll111L11 Z lllll III mi Th lili Pa M n Slika 2 Mejne vsebnosti nečistoč v čistem železu Fig. 2 Limits of impurity contents in refined iron. Although similar to a zone-refining process, our treatment is in fact quite different. Actually, the removal of carbon and oxygen is due solely to interactions of the metal with the surrounding atmosphere. Furthermore, the heating coil moves along the bar at a relatively high speed of 12 cm/min in either direction. This is a zone-levell-ing procedure vvhich improves the homogenity of the product. Carbon is eliminated by combina-tion with the internal oxygen in an atmosphere of argon. Then, oxygen is removed in an atmosphere of pure hydrogen. In the standard procedure, 60 passes are performed, 30 in each direction. Clearly, the validity of a refining technique is never better than the efficiency of the correspond-ing analytical testing methods. Therefore, a great effort has been made in our laboratory in order to develop suitable methods whereby the contents of the most critical impurities could be determi-ned5-8. Then, an extensive program of analysis has been carried out in cooperation vvith eight other French laboratories6. The contents of 76 elements have been determined in specimens taken out of six different bars. The results have been interpreted in terms: 1. of reproducibility of the production process from one bar to the other. 2. of homogeneity of each bar. 3. of validity of the various analytical methods. In the starting electrolytic iron, the impurity contents exhibit vvide variations. We have obtained for instance the follovving results (contents in ppm: Ti = 10—50; Cr = 0.5—12; Ni = 6—50, vvith: S ~ 35; Mn ^ 12; Co ^ 9. In the iron sponge, the carbon and oxygen contents may be as high as 15 and 500 ppm respectively. For the other elements the composition is the same as in the subsequently remelted metal. The table of figure 2 summarizes the results concerning 76 elements vvhich have been ana-lysed in 6 remelted bars6. Comparing vvith the values indicated for the sponge iron vve see, in particular, that the contents of Ti, Cr, Mn, Co, Ni, are now well belovv 1 ppm. Only the contents of C, N, O, and Si exceed 1 ppm. Hovvever, the indicated limits also reflect the sensitivity limits of the analytičal methods. For instance in the čase of C and O, subsequent analysis by activation methods have disclosed contents near or under-neath 1 ppm. Furthermore, since the time vvhen the table has been established, the quality of our production has been improved. This concerns particularly sili-con vvhose content vvas relatively high in the slx bars because some silica had entered the iron hydroxide during the vvashing proces. Using vvater free of silica and teflon containers has allovved to avoid contamination the oxide. Similar procedures have been devised for nickel and chromium. For nickel, the procedure = 10—50; Cr = 0,5—12; Ni = 6—50; S^ 35; Mn^ ^12; Co ^9. V železovi gobi vsebnosti ogljika in kisika lahko dosežeta 15, oz. 500 ppm. Vsebnost ostalih elementov je bila skoraj enaka kot v pre-tailjenih palicah. Tabela na si. 2 kaže rezultate6 analize šestih palic glede na 76 elementov. Primerjava z analizo železove gobe kaže, da so vsebnosti Ti, Cr, Mn, Co in Ni sedaj precej pod 1 ppm. Samo vsebnosti C, N, O, in Si so večje od 1 ppm. Vendar je treba pri tem upoštevati, da nastopajo omejitve, ki jih narekujejo meje občutljivosti uporabljenih analitskih metod. Tako so npr. aktivacijske metode analize pokazale, da je vsebnost ogljika in kisika približno 1 ppm ali pod 1 ppm. Medtem smo že izboljšali kvaliteto naše proizvodnje zelo čistega železa, posebno glede na vsebnost silicija. Vsebnost silicija je bila v omenjenih šestih palicah razmeroma velika zaradi onesnaženja Fe-hidro-ksida s Si02 pri izpiranju z vodo. Izpiranje z vodo, ki ni vsebovala Si02 in uporaba teflonskih posod sta pripomogla, da smo se izognili temu onesnaženju. Podobna postopka smo razvili za pridobivanje Ni in Cr. Postopek za pridobivanje čistega niklja Slika 3 Shema postopka pridobivanja čistega niklja Fig. 3 General prineiple of the refining procedure for nickel. Značilna analiza pretaljenega ■ čistega niklja Typical analysis of the remelted pure nickel Vsebnost nečistoč v ut. ppm lmpurity content in wt. ppm lllll Ce lllll Pr mil Nd III Pm\Sm Himni iillll lllll Tb lllll R UMI m um II z um m mu □ mu Th Pa Slika 4 Mejne vsebnosti nečistoč v čistem niklju Fig. 4 Limits of impurity contents in refined nickel. Chromium (VI) Kromov (VI) oxyde oksid Elektroliza \ Pure Cr I kg/mesec Pretaljevanje Čisti Cr Slika 5 Shema postopka pridobivanja čistega kroma Fig. 5 General prineiple of the refining procedure for chromium. has reached a state of development comparable to that for iron. The only difference is that no extensive program of analysis has not been yet carried out. As suggested in figure 3, commercial carbonyl nickel is first dissolved electrolytically in a dilut^ solution of hydrochloric acid. Then, the solution is purified by means of iron exchange with suita-ble resins, namely: — resin DOWEX 50 W X 8 Form H+ 100—200 mesh — resin DOWEX 1 x 8 Form Cl- 200—400 mesh. The pure nickel chloride is then eleetrolyzed to pure nickel. The niokel cathodes are eventual-iy melted into bars in a water-cooled silver boa\ as in the čase of iron. The limits of the impurity contents are given in the table of figure 4. On the whole, they are larger than in the čase of iron. But this is mainly due to the faot that they are based mainly on the only results of our laboratory7'8: the »safety limits« are not so well known as in the čase of iron vvhere they have been carefully discussed by severa! laboratories. The produetion of chromium has been started a few months ago. It has not yet reached the regu-lar pace arrived at for iron and niokel. Figure 5 shows the general prineiple of the refining process; figure 6 is a table of some, stili preliminary, ana-lytical results. je dosegel približno enako stopnjo razvoja kot za železo. Razlika je le v tem, da še nismo izvedli obsežnega programa primerjalnih kemičnih analiz, kot v primeru železa. Kot vidimo iz sheme na si. 3, tehnični karbo-nilni .nikelj elektrolitsko raztopimo v razredčeni HC1. Raztopino potem očistimo z ionskimi izmenjevalci, tj. z ustreznimi smolami, in sicer: — smola DOWEX 50 W X 8 Form H+ z granu-lacijo 100—200 mesh, — smola DOWEX 1 X 8 Form Cl- z granula-cijo 200—400 mesh. Iz tako dobljenega čistega nikljevega klorida pridobivamo čisti nikelj z elektrolizo. Če je to potrebno, Ni katode potem pretaldmo v vodno hlajeni srebrni ladjici, kot v primeru železa. Mejne vsebnosti posameznih elementov primesi so, kot kaže tabela na si. 4, nasploh večje kot v čistem železu. Vendar je to posledica dejstva, da imamo v tem primeru na razpolago le rezultate kemijske analize iz našega laboratorija7'8. »Varne meje« niso torej tako dobro znane, kot v primeru železa, kjer je pri analitskem delu sodelovalo več različnih laboratorijev. S proizvodnjo čistega kroma smo začeli šele pred nekaj meseci, zato postopek še ni standardiziran, kot pri železu in niklju. Splošno shemo rafinacijskega postopka za krom vidimo na si. 5, preliminarne analize pa kaže tabela na si. 6. C. VPLIV SELENA NA KINETIKO IN r -» a PREMENE ŽELEZA" 12 Priprava čistih materialov seveda ni bil naš končni cilj, temveč le prvi in nujni korak pri raziskavah vpliva nečistoč v jeklu, kar je neposredno povezano s praktičnimi problemi3. Po drugi strani pa so čisti materiali izredno pomembni za temeljne raziskave, kot lahko vidimo iz naslednjega primera. V okviru raziskav o tvorbi in možnih modifikacijah različnih vključkov v jeklu smo smatrali za potrebno, da znova preverimo nekatere dele ravnotežnega diagrama železo—selen9'10. V ta namen smo uporabili različne metode, med njimi tudi termično analizo z adiabatnim kalorimetrom11'12. Osnovni princip aparature je razviden iz sheme na si. 7. Preizkušanec E je v trdnem ali tekočem stanju v posebni celici, katere zunanje in notranje stene skrbno vzdržujemo pri enaki temperaturi. Tako smo popolnoma preprečili prenos toplote skozi stene od vzorca, oz. k njemu. Z grelnim elementom Fa dovajamo natančno določeno količino toplote preizkušancu E, katerega temperaturo merimo s termoelementom G. Ker je sistem toplotno izoliran, lahko natančno merimo količino dovedene toplote, ki je potrebna za določeni dvig temperature preizkušanca, oz. za kristalno premeno. Žal isti postopek ni možno uporabiti pri ohlajanju, ker ohlajanje toplotno izoliranega si- Značilne analize pretaljenega čistega kroma Typical analysis of the remelted pure chromium Vsebnost nečistoč v ut. ppm lmpurity content in wt. ppm 0 N 5 C Si iS IS C tO ■=20 c 10 Na 0,03 K -=Q2 Ca «=10 Mn 0,001 Fe C20 Co 1,7 Cu 'l' Zn c1,7 Ni ^2 Ca <0,003 As 0,01 Sb 0.U W 0,13 Pb -= 10 Al '10 Slika 6 Mejne vsebnosti nečistoč v čistem kromu Fig. 6 Limits of impurity contents in refined chromium. D. INFLUENCE OF SELENIUM ON THE KINETICS OF THE a y AND y a TRANSFORMATIONS OF IRON11-12 As pointed out previously, preparing pure metals is not our end goal: it is only a preliminary, necessary step, before we can investigate the influence of trace elements in steels, which is a ques-tion closely related to practical problems3. On the other hand, pure materials are also invaluable in fundamental research, as can be shown by the following example. As part of a research program concerning the formation and the possible modification of various inclusions in steels, we felt necessary to verify some parts of the iron-selenium equilibrium diagram9'10. Various methods have been used for this purpose and among them, thermal analysis with an adiabatic calorimeter11'12. The basic principle of the device can be expla-ined by the schematic representation of figure 7. The specimen E which can ibe either solid or liquid is placed in a special celi whose inside and outside walls can be rigorously maintained at the same temperature. This means that any heat flow through the walls of the celi, towards the specimen or from the specimen can be prevented. Thanks to a heating element Fa, well known quan-tities of heat can be injected into the specimen whose temperature is measured by a thermo-couple G. Consequently, as there is no heat loss through the walls of the celi, the amount of heat which is needed to increase the temperature of the specimen or to bring about a transformation can be measured accurately. Unfortunately, the same procedure does not work on cooling as heat loss of the specimen can be achieved only by heat loss through the waMs of the celi: the device is no more adiabatic. How-ever, carefully controlling the difference in the temperatures of the inside and outside walls of the celi, it is possible to determine, at least appro-ximately, the heat loss of the specimen. stema ni možno. Vendar je s skrbno kontrolo razlike temperatur zunanje in notranje stene možno vsaj približno določiti toplotne izgube preizku-šanca. Najprej smo opravili številne meritve, potrebne za preizkus in umerjanje aparature. Pri meritvah na srebru smo slučajno opazili, da je tališče neod- Slika 7 Merilna celica adiabatnega kalorimetra E — preizkušanec, F, — grelni element (volfram), Fb — — zaščitna korundna cev, Fc — vodniki (Mo žica), G — — termoelement Pt-Rh 10% - Pt, H, — adiabatna stena celice (Mo-pločevina), Hb r- diferenčni termoelement, Hc — mikro pečica, Hd — grelni element mikro pečice s korund-no izolacijo, I — krožni zasloni za zaščito pred toplotnim sevanjem, J — žica s korundno izolacijo. Fig. 7 Measuring celi of the adiabatic calorimeter. E — specimen, F. — heating vvire (tungsten), Fb — protect-ing sheath (alumina), Fc — electric wires (molybdenum), G — thermo-couple (Pt-Rh 10 % i- Pt), H. — adiabatic wall of the celi (moIybdenum sheets), Hb — thermo-couples (in opposition), Hc — jmicro-furnace, Ha — heating element of the micro-furnace, in alumina beads, I l— circular shields, J ^ sustaining vvire, in alumina beads. visno od količine dovedene toplote, tj. od hitrosti taljenja. Z drugimi besedami, temperatura na medfazni površini talina/trdna snov je vedno enaka in ustreza temperaturi ravnotežja talina/trdna snov. Pri taljenju temperatura medfazne površine ni odvisna od hitrosti, s katero se medfazna površina premika vzdolž preizkušanca. Potem smo začeli s preiskavami kristalnih premen a -> y in y ~> a v čistem železu. Premeni sta non-variantni, tako kot taljenje. Zato bi lahko A number of measurements have been perfor-med in order to test and to calibrate the device. When studying silver it was incidentally observed that the melting temperature is not affected if the heat input is changed, that is if the melting rate varies. In other vvords, if we consider the interface betvveen melted and solid silver, the temperature at the interface always equals the equili-brium temperature betvveen liquid and solid metal. On melting, the temperature at the interface does not at ali depend on the speed of the interface moving across the specimen. Then, the a y and y -» a transformations of pure iron vvere investigated. Like the melting process, such transformations are invariant ones. Therefore, we vvould expect the transformation T a A S, b c /T 1 p>p0 Po / P Pc) dobimo večjo temperaturo premene, kot vidimo na krivulji b. Če med poskusom spremenimo dotok toplote na prvotno vrednost (P = PG), znova dobimo prvotno temperaturo premene. Podobno kaže krivulja c, da se temperatura premene zniža, če je dotok toplote manjši, tj. če je P < PD. Lahko torej zaključimo, da je temperatura medfazne površine a/y odvisna od njene hitrosti gibanja vzdolž preizkušanca. Na si. 9, kjer je Te ravnotežna temperatura, tj. 1184 K, prikazujemo spremembo temperature premene (AT) v odvisnosti od dotoka toplote. V tem primeru dotok toplote 1W ustreza hitrosti 0,5 mm/min gibanja medfazne površine. Meritve pri zelo šibkih dotokih toplote so pokazale, da AT naglo narašča od 0 do približno 1,25 K, ko dotok toplote raste od 0 do 0,25 W. Pri večji intenziteti ogrevanja preizkušanca je rast AT vse do 10 W veliko počasnejša, kot vidimo iz slike 9. Pri ohlajanju imamo ustrezen padec temperature premene. Pri analizi teh rezultatov se je pojavilo vprašanje, ali nekatere nečistoče, npr. ogljik, vplivajo na temperaturo premene. Dejstvo je, da imamo med premeno a -» y in y -» a stalno prerazdelitev ogljika med a in y fazo, kar vpliva na temperaturo medfazne površine. Tej možnosti smo namenili posebno pozornost. Glede na to, da ima AT pri danem toplotnem dotoku P ves čas poskusa konstantno vrednost, domnevamo, da difuzija nima opaznega vpliva na rezultate meritev, tj. prerazdelitev oligoelementov med premeno nima pomembnejše vloge. Enake poskuse smo izvedli na zlitinah sistema železo-selen. Preizkušance lahko razdelimo po vsebnosti selena v skupini s 175 ppm in 550 ppm selena. V takih zlitinah imamo non-variantno premeno ki poteka pri 913 °C (slika 10). Kot v primeru železa, je temperatura potujoče medfazne površine odvisna od hitrosti premene, kar vidimo na si. 11. Primerjava s prej omenjenimi rezultati kaže, da majhne količine selena močno povečajo ta pojav pri ogrevanju. V primeru, ko je bila vsebnost selena v območju med 200 in 500 ppm, je bila razlika AT v temperaturi premene povečana za faktor 5. Presenetljivo je, da je ta vpliv selena veliko manjši pri ohlajanju. Pri teh zlitinah prerazdelitev selena med a in r fazo zanesljivo vpliva na temperaturo medfazne površine, vendar je koeficient difuzije selena tako majhen, da ta efekt verjetno lahko zanemarimo. Premena poteka tako, da pri masivni transformaciji hitrost premene, pa tudi prisotnost nečistoč vplivata na temperaturo potujoče medfazne površine. lovver than PD. Consequently, it turns out that the temperature at the moving interface depends on the speed of the interface. Figure 9 shovvs the numerical results. The value Te being the equilibrium temperature, that is 1184 K, the increase in the transformation temperature, AT, has been plotted as a function of the heat input. In the present čase, 1 W corres-ponds approximately to 0.5 mm/min. in the speed of the moving interface. Measurements vvith very lovv heat inputs have shovvn that AT rapidly incre-ases from 0 to about 1.25 K vvhen the heat input varies from 0 to 0.25 W. Then, as shovvn in figure 9 the increase in AT is much slovver, up to a heat input of 10 W. On cooling, there is a corresponding decrease in the transformation temperature. Concerning these results, the question has been raised as to vvhether the transformation temperature may be influenced by some impurity like carbon. As a matter of fact, during the 0,010,02 1 Fe 10 20 30 iO 50 Se(ut.%)(wt.°A) Slika 10 Ravnotežni diagram Fe-Se Fig. 10 Fe-Se equilibrium diagram. transformations, there is a permanent redistri-bution of carbon betvveen the tvvo phases vvhich effectively affects the temperature at the interface. This point has been carefully discussed. During an experiment vvith a given value of the heat input, AT has a constant value. Mainly because AT does not change during the experi-ment, it is felt that diffusion processes do not materially affect the results: the redistribution of residual elements should not play a significant role during the transformation. m j (K/min) Slika 11 Vpliv selena na temperaturo premene ai^f Fig. 11 Influence of Se on the a f=> y transformation temperature. D. VPLIV OLIGOELEMENTOV NA NASTANEK KOROZIJSKIH JAMIC NA POVRŠINI AVSTENITNIH NERJAVNIH JEKEL 13,14 Kljub našemu zanimanju za temeljne raziskave je proizvodnja čistih kovin v našem laboratoriju prvenstveno povezana s prakso, tj. z uporabnimi lastnostmi jekel. V zvezi z mehanskimi lastnostmi smo že omenili raziskave krhkosti^2.3 železa in jekla. Sedaj bomo obravnavali nekatere pojave korozije pri nerjavnih jeklih. Izmed večjih problemov v zvezi z uporabo av-stenitnih nerjavnih jekel je nastanek korozijskih jamic na površini, t. im. pitting korozija, eden najpomembnejših. Pri tej posebni vrsti korozije je nastanek korozijskih jamic na površini bistvena faza korozijskega procesa. Ko že nastane majhna površinska jamica, ta nadalje neizogibno raste in se razvija. Sedaj je že na splošno sprejeto, da nekovinski vključki inicirajo nastanek korozijskih površinskih jamic. Verjetnost nastanka jamice je odvisna od vrste vključka. S praktičnega stališča je torej pomembno, da poznamo obnašanje različnih vrst vključkov, ki lahko nastopajo v nerjavnih jeklih. To lahko spoznamo s korozijskimi poizkusi, pri katerih preštejemo jamice, nastale na površini preiakušanca, ki je bil izpostavljen določenemu korozijskemu vplivu v skrbno kontroliranih pogojih. Lahko pa dobimo enakovredne, če ne boljše podatke in v veliko krajšem času s pomočjo t. im. The same experiments have been performed vvith iron-selenium alloys. Tvvo kinds of specimen have been used vvith lovv selenium contents of 175 and 500 ppm respectively. In such alloys, there is an invariant a;±r transformation at 913 °C (Fig. 10). As in the čase of iron, the temperature at the moving interface depends on the transformation rate. This is shovvn in figure 11. Now, the compa-rison vvith the foregoing results shovvs that small contents of selenium considerably enhance the phenomenon on heating. In the present čase, vvith selenium contents in the range of some 200 to 500 ppm, the difference in the transformation temperature is approximately multiplied by a factor of 5. On cooling, the influence of selenium is surpri-singly much more limited. In these alloys, the temperature at the interface is certainly affected by the redistribution of selenium betvveen the a and y phase. Hovvever, the diffusion coefficient of selenium is so lovv that this effect is likely to be negligible. The transformation proceeds in such a way that a and y phase have practically the same composition on either side of the interface: this situa-tion is characteristic of a so called »massive transformation«. From a fundamental point of vievv, these results shovv hovv the temperature at the moving interface during a massive transformation is influenced both by the transformation rate and by the presence of trace elements in the metal. D. INFLUENCE OF RESIDUALS (AS SOLUTES OR INCLUSIONS) ON THE PITTING INITIATION IN AUSTENITIC STAINLESS STEELS 13 u In spite of the interest of fundamental investi-gations, it has already been pointed out, that the preparation of pure metals in our laboratory is primarily concerned vvith practical properties of steels. Regarding mechanical properties, research vvorks about the brittleness of iron and steel have been mentioned1'2-3. Some corrosion phenomena in stainless steels vvill be discussed novv. Among difficult problems concerning austeni-tic stainless steels, pitting is one of the most important. In that particular kind of corrosion the initiation of pits is the essential step because once a small corrosion hole has been formed on the surface, it vvill inevitably develop. Then, it is novv generally agreed that corrosion pits initiate on non-metallic inalusions. Hovvever, depending on the nature of the inclusion this phenomenon is more or less likely to happen. From a practical point of vievv, it is therefore important to knovv the behaviour of the various Ikinds of inclusions vvhich can be present in a stainless steel. This can be done by corrosion tests, counting the number »polarizacij skih krivulj«. Vzemimo primer preiz-kušanca, ki je potopljen v korozijsko raztopino. Če je električno izoliran, prevzame določen potencial glede na raztopimo, to je t. im. ravnotežni potencial Ec. Če sedaj z ustrezno napravo dovedemo preizkušancu drugačen potencial, se pojavi tok od preizkušanca k raztopini ali obratno. S spreminjanjem tega potenciala lahko dobimo krivuljo, ki ima tri glavna območja, kot vidimo na si. 12: ' Aktivnost Pasivnost Transpasivnost Activity Passivity Transpassivity I I I I I I I I /£"e Ea Ep E Slika 12 Tipična polarizacijska krivulja jekla Fig. 12 Typical ipolarization icurve of a steel. 1. aktivacijsko območje med ravnotežnim potencialom Ee in potencialom Ea, ki ga imenujemo »aktivacijski«, »pasivacijski« ali »Flade potencial«; 2. pasivacijsko območje, ki je na zgornji strani omejeno s »porušnim potencialom« :Ep; 3. »transpasivacijsko« območje. V pasivacijskem območju je kovina zaščitena s »pasivacijskim slojem«. Pasivacijski tok ip je v splošnem tako majhen, da korozija skoraj ne poteka. Kolikor daljše je to območje, toliko večja je odpornost proti koroziji. Porušni potencial je opredeljen: — s porušitvijo pasivacijskega sloja ali — z začetnim pojavom korozije na nekem vključku. Ta druga opredelitev pojasnjuje, zakaj porušni potencial pogosto imenujemo »pitting potencial«, tj. potencial, pri katerem nastajajo korozijske jamice na površini jekla. Lahko se zgodi, da je pitting potencial vključkov večji od porušnega potenciala, pri katerem se pretrga pasivacijski sloj, V tem primeru se pri vrednosti pitting potenciala pojavi porast korozijskega toka v transpasivacij-skem območju. Vključki so na splošno dobro opredeljeni s svojim pitting potencialom v določenem koroziv-nam okolju. Vključki so toliko bolj škodljivi, oz. nevarni, kolikor manjši je njihov pitting potencial. of pits which are formed on a specimen under well defined conditions. Hovvever, it is possible to get a similar if not better information using a much less time consuming procedure by means of the so called »polarization curves«. Let us consider a specimen which is dipped in a given corrosive medium. If it is electrically iso-lated it takes on a certain potential with respect to ithe solution: this is the equilibrium potential Ee. Now, if a different potential is applied to the specimen by means of a suitable device, a current will flow from the specimen into the solution or in the opposite direotion. Changing the potential, one gets a complete curve vvhich exhibits three main regions (Fig. 12): 1. the activity region, betwen Ee and Ea. The latter potential is called the »activation potential« or »passivation potential« or »Flade potential«. 2. the passivity region limited on the high side by the »breakdovvn potential« Ep. 3. the »transpassMty« region. In the passivity range, the metal is protected by a »passive layer«. The passivation current ip is generally so low that there is no significant cor-rosion of the metal. Generalily speaking, the lon-ger is this region, the better is the steel. Further-more, the rupture potential is determined: — either by the breaikdown of the passive layer. — or by the onset of corrosion on some inclusion. The second mechanism explains why the breakdovvn potential is frequently called, in the čase of stainless steels, the »pitting potential«. However, it can happen that the pitting potential of the inclusions is higher than the true rupture potential of the passive layer. In such a čase, there will be an increase in the corrosion current at the value of the pitting potential, in the region of transpassivity. On the whoIe, in a given medium, the inclusions are charaoterized by their pitting potential. The lower it is, the worse are the inclusions. Therefore, it is obviously of interest to classify the inclusions with respect to their pitting poten-tials. It is hardly possible to achieve an accurate classification using industrial steels because there are in general several kinds of inclusions with a whole range of compositions. On the other hand, this can be done vvtith high purity metals contain-ing one vvell defined category of inclusions. Such a research work has been performed in our laboratory. The base steel was of grade 18 to 10. Thanks to suitable doping, various kinds of inclusions have been produced: oxides, sulphides, selenides, tellurides. As a general rule, the speci-mens have been water-quenched at 1150°C. How-ever, some other treatments, including deformation at high or low temperature have also been used. Zato je zanimiva razvrstitev vključkov po njihovem pitting potencialu. Seveda je nemogoča natančna klasifikacija vključkov v nerjavnem industrijskem jeklu, ker tehnična nerjavna jekla vsebujejo celo vrsto različnih Vključkov s spremenljivo kemično sestavo, ki se lahko spreminja v določenem območju. Za tako klasifikacijo rabimo torej čiste materiale, ki vsebujejo le eno, dobro opredeljeno vrsto viključkov. V našem laboratoriju smo zato izvedli ustrezne raziskave na čistem jeklu 18—10, v katerem smo s skrbnim dodajanjem zelo majhnih količin določenih nečistoč proizvedli točno določene vrste vključkov: okside, sulfide, selenide in teluride. Vsi preizkušanci so bili gašeni v vodi s 1150 °C. Uporabljali smo tudi druge vrste obdelave, oz. predelave, vključno z deformacijo pri višji ali nižji temperaturi. Pitting potenciale smo določali na preizkušancih, ki so bili potopljeni v vodni raztopini natrijevega klorida (30 g NaCl na liter raztopine), ki je bila podobna morski vodi. Izmerjene potenciale za oksidne vključke kaže si. 13. Ker so najbolj nevarni vključki, ki imajo najmanjši potencial, lahko rezultate teh poizkusov povzamemo v naslednji obliki: 1200 1100 1000 900 . 800 £ 700 Uj g 600 £ 500 ^ 400 300 200 100 £ C ^ 11 { £ 0 1 £ £ t a + o o 5 o <0 o Ž1 o s? o B Vključki g J Inclusions I Slika 13 Pitting potenciali različnih oksidnih vključkov: mV/ECS — elektrodni potencial kovine merjen z nasičeno kalomel referenčno elektrodo, HT — toplotno obdelano (kaljeno v vodi pri 1150 °C), H — homogenizirano (žarjeno pri 1250 "C), D — žarjeno pri 800 "C za odpravo napetosti. Fig. 13 Pitting potentials of oxide Inclusions: mV/ECS — electrode potential of the metal, measured with a saturated calo-mel reference electrode, HT >— heat treated (water qu-enched at 1150 °C), H — homogenized (annealed at 1250 °C), D — stress relieved (at 800 °C). 1. Najmanj nevarni so Ti02 vključki, ki imajo potencial približno 800 mV. 2. Najbolj nevarni so MgO vključki, ki imajo potencial približno 400 mV. 3. Različni načini obdelave in predelave skoraj ne vplivajo na potencial in s tem na škodljivost MgO vključkov. The pitting potentials have been determined in an aqueous solution of sodium chloride simiilar to sea water, that is with 30 g of sodium chloride per litre of solution. Figure 13 shows the results for oxide inclusions. Remembening that the most dangerous inclusions correspond to the lovvest pitting potentials, this figure can be commented as follows: 1. Ti02, vvith a pitting potential of about 800 mV is the least dangerous inclusion. 2. MgO, with a pitting potential of about 400 mV is the most dangerous one. 3. Considering the MgO inclusions, it is appa-rent that different treatments do not materially change their pitting potential. 4. Introducing copper or molybdenum into the metal changes te pitting potential of chromite inclusions, although the composition of the inclusions themselves remains unaltered. This shows that the pitting potential of the inclusions depends on the behaviour of the passive layer. For instance, molybdenum significantly improves the resistance to pitting, a fact that has long been iknovvn in practice. Figure 14 shows the results for sulphides. On the whole, the pitting potentials are much lower than in the čase of oxides. Manganese sulphides are especially dangerous. As in the former čase, molybdenum has a favourable influence. Considering the inclusions of titanium sulphide, the pre-sence of manganese in the matrix seems to have an adverse effect, contrary to molybdenum. 1000 900 800 __ 700 ^ 600 £ 500 e 400 ^ 300 200 100 0 £ I £ d + 10 5: + to i i I ♦ I * to i 5 tj ♦ I + tO S S S S <3 I to I £ to 1 £ t? "> <0 C ■cr S I I S S <0 t