Propustna in lezna krhkost CrMoV jekla z okoli 1 % Cr zaradi oligoelementov
M. Krause,* Jin Yu**
Temper and crep embrittlement of a CrMoV steel vvith about 1 % chromium by residual elements
Eksperimentalno delo smo izvršili z jeklom 28 Cr-Mo-Ni-V 4 9 (StahI - Eisen - Werkstoffblatt 555), ki ima približno kemično sestavo: 0,28 % C, 1 % Cr, 0,9 % Mo, 0,7 % Ni in 0,3 % V. To je eno izmed dveh najpomembnejših jekel, odpornih proti lezenju, ki ju uporabljamo za rotorje srednje in visokotlačnih parnih turbin. Cilj raziskave je bil ugotoviti stopnjo popustne krhkosti, ki jo povzroče različne količine fosforja in kositra. Vzorci jekla so bili toplotno obdelani na marten-zitno in bainitno strukturo. Nadalje smo ugotovili vpliv bakra in aluminija na prelomno duktilnost po lezenju in po različnih časih na temperaturah, pri katerih to jeklo uporabljamo.
Na osnovi več raziskovalnih del je znano, da je reverzibilna popustna krhkost odvisna od stopnje izcejanja elementov P, Sn, As in Sb po kristalnih mejah1. Ti elementi znižujejo energijo kristalnih mej2-3 in s tem tudi energijo interkristalnega preloma4'5. Ugotavljajo, da povečujejo krhkost elementi v naslednjih redih P-Sn-Sb-As6 ali Sb-P-Sn-As7. Z Augerjevo spektroskopijo so bile ugotovljene izceje elementov P in Sn na kristalnih mejah krhkih jekel8. Nadalje je znano, da popustno krhkost povečajo Ni, Mn in Si, verjetno zato, ker ti elementi premaknejo ravnotežje med Mo v raztopini in Mo v karbidih na stran Mo karbidov8.
Pomemben je zlasti Mo, ki preprečuje vpliv P. Zmanjšanje koncentracije Mo v raztopini povečuje izceje P na kristalnih mejah in s tem tudi krhkost. Razen Mo vplivata enako tudi Cr in V, vendar manj močno kot Mo9.
Ker tvori ogljik karbide s Cr, Mo in V, je bilo ugotovljeno, da je vsaka važnejša komponenta kemične sestave v tem jeklu direktno ali indirekt-no povezana z reverzibilno popustno krhkostjo.
Količin Cr, Ni in Mn ne moremo preveč znižati, ker so potrebni za kaljivost, da se izognemo nastajanju ferita v jeklu velikih odkovkov. Mo in nekoliko manj tudi V povečata kaljivost in odpornost proti lezenju, Ugotavljajo, da je potrebna količina teh elementov v raztopini 0,2 do 0,25 % tudi še po dolgih časih obratovanja pri delovni obremenitvi in temperaturi, ne glede na njuno vsebnost v talini10. iNi še gotovo, da je ta količina obeh elementov zadostna za preprečitev popustne krhkosti. Z ozirom na ozke meje elementov C, Ni,
* Krupp StahI AG, Bochum
** Max-Planck-Institut fiir Eisenforschung, Dusseldorf, F. R. Germany
Experimental work has been carried out on the steel 28 Cr-MonNi-V 4 9 according to the German Stahl-Eisen-Werkstoffblatt 555 vvith a nominal chemical composition of about 0.28 % C, 1 % Cr, 0.9 °/o Mo, 0.7 % Ni and 0.3 % V, beeing one of the two most important creep resisting forging steels used for intermediate and high pressure steam turbine rotors. The aim of the investigations vvas to find out the degree of temper embrittlement of this steel caused by certain amounts of pho-sphorus and tin, in a martensitic and in a bainitic completely heat treated structure. Further, an influence vvas to be confirmed of copper and aluminium on the rupture ductility after creep of some duration under service temperature.
It is knovvn from numerous research vvorks that reversible temper embrittlement corresponds vvith the degree of segregation of the trace elements P, Sn, As and Sb along the grain boundaries1. These elements are knovvn to lovver grain boundary energy2-3 and subsequently the intergranular fracture energy4-5. The embrittling po-tency is reported to increase in the order P-Sn-Sb-As6 or Sb-P-Sn-As7. Segregation of the elements P and Sn has been detected at the grain boundaries ofembrittled steel by Auger-spectroscopy8. Further, it is knovvn that temper embrittlement is enchan-ced by Ni, Mn and Si probably because these elements shift the equilibrium betvveen molybdenum in solution and molybdenum in carbides to the side of molybdenum carbides8.
Molybdenum is of importance because it is a scavenger for phosphorus and therefore a reduced concentration of molybdenum in solution leads to a higher phosphorus segregation at the grain boundaries resulting in a higher trend to embrittlement. As compared to molybdenum the inter-action of chromium and vanadium vvith phosphorus is not so strong9.
Because of the role of carbon in forming carbides vvith chromium, molibdenum and vanadium, it can be stated, that any of the important components of the chemical composition of the steel under consideration is directly or indirectly connected vvith the reversible temper embrittlement.
From these components, Cr, Ni and Mn should not be reduced too much, because they are neces-sary to give a sufficient hardenability so that ferrite in the core of big forgings is avoided. Mo-lybdenum raises the hardenability and, next to vanadium, the creep resistance. It has been repor-
Mn, Cr, Mo in V v teh jeklih nastaja vprašanje, katere količine drugih elementov lahko dopuščamo brez nevarnosti za popustno krhkost. V našem delu smo se osredotočili na Si, P in Sn.
V 45-tkilogramski laboratorijski indukcijski peči smo izdelali taline, katerih osnovna sestava je navedena v tabeli 1. Številke kažejo najvišjo in najnižjo vsebnost legirnih elementov osmih talin, ki smo jih rabili pri prvih preizkusih našega programa.
Tabela 1: Kemična sestava preizkušanih jekel. Osnova: 28 Cr-Mo-Ni-V 4 9 (%) Table 1: Chemical composition of experimental steels. Base: 28 Cr-Mo-Ni-V 4 9 (%)
Osnovna sestava 8 preizkušanih jekel Base composition of 8 experimental steels
Mn
Cr Mo Ni
V
As Sb
N
0,29 0,56 0,007	0,94 0,90 0,64 0,31	0,004	0,001 0,008
0,33 0,70 0,009	1,03 0,95	0,68 0,35	0,005	0,002 0,010
Dodatno: Additionally:
Talina Melt	Si	P	Sn	Al	Cu
1	0,06	0,006	0,005	0,003	0,04
3	0,05	0,005	0,005	0,003	0,16
5	0,06	0,005	0,013	0,025	0,04
7	0,04	0,005	0,005	0,044	0,03
14	0,05	0,024	0,017	0,003	0,03
15	0,53	0,025	0,017	0,003	0,04
19	0,63	0,074	0,018	0,017	0,03
20	0,64	0,082	0,017	0,003	0,04
Ingoti so bili skovani v palice, ki so bile toplotno obdelane za prvo serijo preizkusov na marten-zitno strukturo, za drugo serijo pa na bainitno strukturo. Preizkusne palice z martenzitno strukturo so bile po popuščanju hlajene v olju za pridobitev začetnega žilavega stanja materiala kot osnove za krhkostne preizkuse. Preizkusi z bainitno strukturo pa naj bi simulirali stanje materiala pri velikih turbinskih rotorjih. Toplotna obdelava je bila torej izbrana v skladu s časovno-tempera-turno krivuljo za turbinski rotor, premera 1000 milimetrov, če je ta kaljen z 970 °C v olju in po-puščen. S stopenjskim hlajenjem pa smo preizkusili, koliko ta vpliva na pomik FATT (Fracture Appearance Transition Temperature — prehodna temperatura žilavosti, določena po videzih prelomnih površin).
Tabela 2 vsebuje rezultate vzorcev z martenzitno strukturo. Niti baker (talina 3) niti kositer (talina 5) v količinah, ki jih lahko najdemo v od-
ted to remain in solution at a level of 0.2 to 0.25 % irrespective of the initial bulk content of the melt, after long time exposure at service temperature and load10. It is not certain, whether this level is sufficient to avoid temper embrittlement in 1 % Cr-Mo-V steel. On the basis of the assump-tion that the chemical composition of melts in-tended for big creep resistant iforgings should remain within narrow limits regarding the elements C, Ni, Mn, Cr, Mo und V, the question arises, what level of other elements can be tole-rated without running a rise of temper embrittlement. In the first step, we concentrated on Si, P, and Sn.
45 kg laboratory induetion furnace melts have been made vvith a base composition shovvn in table 1. The ifigures in the upper line give the highest and lovvest content of the alloying elements of the eight melts used in the first step of our programme.
The ingots have been forged to bars and heat treated to obtain a martensitic microstructure for one series of the tests and a bainitic microstructure for another series. The martensitic test bars have been oil cooled after tempering to give a tough initial state of the material as a basis for embrittling experiments. The tests vvith bainite have been done to simulate the state of material near the rim of a big turbine rotor. Therefore, a heat treatment vvas choosen similar to the time-temperature curve near the rim of a turbine rotor of 1000 mm diameter vvhen austenitized at 970 °C, quenched in oil and tempered production-like. Step cooling has been used to developed a shift in FATT if possible (FATT — Fracture Appearance Transition Temperature).
Table 2 represents the results obtained on martensitic specimens. Neither copper (melt 3) nor tin (melt 5) in amounts possibly encountered in forgings, change the FATT, and no shift of FATT vvas produced by step cooling. If a relatively high phosphorus — and tin — content are combined (melt 14), there may be some indi-cation of an increase of FATT by step cooling. If the silieon content is increased (melt 15), an embrittlement becomes more distinet. At high silieon — and phosphorus — contents i. e. beyond techni-cal limits (melt 19 and 20), an embrittlement by step cooling ist quite obvious.
Table 3 represents the results obtained on bainitic specimens shovving that silieon and phosphorus contents beyond the technical level and a relatively high tin content result in temper embrittlement in bainitic structure of a turbine rotor. It should be pointed out, that the structure tested here consisted of a mixture of lovver and upper bainite vvith lovver bainite being predomi-nant. Moreover, it might be ooncluded that no temper embrittlement — as revealed by step cooling — vvould develop in a fully upper bainitic microstructure like that near the core area of a
kovkih, ne spremenita F ATT. Tudi ni opaziti nobene spremembe FATT pri stopenjskem ohlajanju. V primeru sorazmerno visokih vsebnosti P in Sn (talina 14) je nekaj indikacij, da se pri stopenjskem ohlajanju poviša FATT. če se poveča količina Si (talina 15), postane krhkost bolj očitna. Pri količinah S in P nad tehničnimi mejami (talini 19 in 20) postane krhkost po stopenjskem ohlajanju zelo izrazita.
Tabela 2: Vpliv Si, P in Sn na zarezno žilavost. Jeklo 28 Cr-Mo-Ni-V 4 9
Table 2: Influence of Si, P and Sn on notch toughness. Steel 28 Cr-Mo-Ni-V 4 9
e2s
Vsebnost oligoelementov Content of residual
elements '%
Rm 20 °C N/rnm2
AFATT
FATT oq S t open j -
olie s ko Ou hlajeno Step cooling
1		760	—93	0
3	0,16 Cu	785	—98	0
5	0,013 Sn	790	—95	~0
7		790	—80	0
14	0,024 P 0,017 Sn	770	—90	30
15	0,53 Si 0,025 P 0,017 Sn	790	—72	50
19	0,63 Si 0,074 P 0,018 Sn	810	—40	135
20	0,64 Si 0,082 P 0,017 Sn	810	—55	135
Toplotna obdelava, heat treatment: 970° 3 ure, hrs/olje, oil, + 720° 15 ur, hrs/olje, oil, presek, section 25 X 25 mm
Mikrostruktura, microstructure: Martenzit, mar-tensite
Iz tabele 3, ki vsebuje rezultate vzorcev z bai-nitno strukturo, je razvidno, da je potrebno več Si in P, kot ju vsebujejo tehnična jekla, in sorazmerno visoka vsebnost Sn, da se pojavi popustna krhkost pri bainitni strukturi turbinskega rotorja. Treba je poudariti, da je bila struktura, ki jo obravnavamo, mešanica spodnjega in zgornjega bai-nita, z večino spodnjega. Celo lahko zaključimo, da se popustna krhkost ne bo razvila (kot se kaže pri stopenjskem ohlajanju) v popolnoma zgornji bainitni strukturi, kot je tista v jedru odkovka, če je kemična sestava ista, kot smo jo že omenili.
Prvi primer krhkega loma zaradi lezenja je znan s konca tridesetih let, ko so se na prvem navoju pri glavi lomili prirobniški vijaki pri visokotlačnih parnih ceveh pri delovni temperaturi 500 °C. Vijaki so bili izdelani iz jekla s približno 0,12 °/o C, 1,5 % Ni, 0,7 % Cr in 0,5 % Mo in so bili poboljšani na natezno trdnost okoli 900 N/mm2. Sprva so verjeli, da sta za te lome odločilni vsebnost niklja in neke vrste popustna krhkost. Danes
big forging, provided that the chemical composition lie within the limits mentioned above.
Tabela 3: Vpliv Si, P in Sn na zarezno žilavost. Jeklo 28 Cr-Mo-Ni-V 4 9 Table 3: Influence of Si, P and Sn on notch toughness. Steel 28 Cr-Mo-Ni-V 4 9
■3
hS
Vsebnost oligoelementov Content of residual elements
Rm
20 °C N/mm2
AFATT
°C
FATT Stopenjsko hlajeno
Step cooiling
°C
1
3 5 7
14
15
19
20
0,16 Cu 0,013 Sn
0,024 P 0,017 Sn 0,53 Si 0,025 P 0,017 Sn 0,63 Si 0,074 P 0,018 Sn 0,64 Si 0,082 P 0,017 Sn
770 795 815 805 770 825 820 830
+ 75 + 62 + 65 + 76 + 80 + 92 + 104 + 110
0 ~ 0 ~ 0 + 10 ~ o
+ 25 + 60 + 60
Toplotna obdelava, heat treatment: 970°, 18 ur, hrs/rob, border 1000 mm, hlajeno v olju, oil cooling, 710° 10 ur, hrs/35° na uro, pro hr Mikrostruktura, microstructure: bainit, bainite
The first cases of brittle creep failure became known at the end of the thirties when high pressure steam tube flange bolts at a service temperature of 500 °C fractured in the first thread of the screw head. The bolts were made of a steel vvith about 0.12 % C, 1.5 % Ni, 0.7 % Cr and 0.5 % Mo,
1000 800 600
400
2!
3 §
& Ž
200
100
				D uk ti t no		
				Ductile		
3	------					
		2		24°		
					30*	
						251
5 M0
o c
e
.o o
<u
200
100
					
			Zarezr Notch	to občut sensitivt	Ijivo
■JO					3
					
					
					
JO
10
10J
10*
105
Čas obremenjevanja (h)-Loading tirne (hrs)
Slika 1
Lezna lomna trdnost pri duktilnem in zarezno občutljivem stanju jekla (shematično)
Fig. 1
Creep rupture strength at ductile and notch sensitive steel state (shematically)
je znano, da taki lomi lahko nastanejo pri vrsti kovinskih materialov, od nizko legiranih konstrukcijskih jekel do jekel, odpornih proti lezenju, zlitin na osnovi niklja in do neželeznih zlitin. Slika 1 daje osnovno razlago takega loma. V zgornjem delu slike so prikazani rezultati preizkusov lezne lomne trdnosti za žilavo stanje materiala: pri vseh časih in enakih obremenitvah je pri zarezanih vzorcih čas loma daljši kot pri gladkih vzorcih. Številke pri gladkih vzorcih so lomni raztezki. Vrednosti so zadosti visoke in se ne spreminjajo sistematično. Na spodnjem delu slike pa so prikazani rezultati preizkusov lezne lomne trdnosti za krhko stanje materiala. Opazimo lahko dva značilna pojava: v določenem intervalu, katerega trajanje je odvisno od stanja materiala in razmer prezkušanja lezenja, je pri zarezanih vzorcih čas do loma pri enakih obremenitvah krajši kot pri gladkih vzorcih in lomni raztezki gladkih vzorcev
S
I
e £fc
Slika 2
Vpliv natezne trdnosti, bakra in aluminija na lezno zarez-no občutljivost pri Cr-Mo-V jeklih z okoli 1 % Cr
Fig. 2
Influence of tensile strength, cooper and aluminium on creep limit notch sensitivity of Cr-Mo-V steels vvith appr. 1 % Cr
heat treated to a tensile strength of about 900 N/mm2. At first, the nickel content and some kind of temper embrittlement vvere believed to be responsible for the failures. Today, hovvever, it is knovvn, that failures of this kind may be obser-ved at a broad variety of metallic materials, rang-ing from low alloy construction steels to creep resistant steels, niokel base alloys and to non ferrous alloys. Figure 1 shovvs the principal featu-res of this failure. In the upper part of the picture results of creep rupture tests at a tough state of material are shovvn: at any time, the rupture lifes of notched specimens are longer than those of plain specimens at the same load. The figures at the plain specimens represent the rupture elongations. The values are sufficiently high and do not change in a systematic manner. In the lovver part of the picture, results of creep ruptur tests at a brittle state of the same material are represented. Tvvo typical features may be obser-ved: At a certain interval, the duration of vvhich depends on the state of the material and the test conditions, the rupture times of notched specimens are shorter than those of plain specimens vvith the same load, and the rupture elongations of the plain specimens, in the same interval, decrease to very lovv values. The underlying mechanism includes the relative creep strength of grain volume and grain boundary and further aspect vvhich vvill be mentioned later on. In addition to that, sometimes an influence of impurities has been presumed or found in experiments. Since 1967, scarce references have been made regarding an influence of aluminium or copper on the rupture ductility: Ratcliff and Brovvn11 reported about a deleterious effect of 0.020 % soluble aluminium on the rupture ductility of a 1 % Cr-Mo-V steel. Their specimens had tensile strengths of about 1100 to 1300'N/mm2 vvhich is outside tech-nical limits. Beneš and Skvor12 noted a decrease in rupture duetility of a Cr-Mo-V Steel in creep tests at 600 °C by 0.25 % Cu. Hopkins and al13 in-vestigated a steel similar to 28 Cr-Mo-Ni-V 4 9 vvith bainitic microstructure and found longer rupture life and higher rupture ductility for a high purity laboratory bar material as compared to steel of conventional purity. The specimens had tensile strengths of about 1000 N/mm2 vvhich is too high for this type of steel. As a consequence, the rupture elongations of the bainitic specimens rema-ined lovv for the pure steel (6 % at 13 000 hrs rupture time) as for the impure steel as vvell (3.9 % at 4500 hrs rupture time). Visvvanathan, in his re-vievv article14, again mentioned the influence of Al and Cu, among others on rupture ductility.
We had the opportunity to use a igreat amount of data kindly submitted by the German Coopera-tive Creep Testing Committee to find out vvhether a relation exists betvveen the copper and aluminium content of ruptured creep specimens and their rupture behaviour. A qualitative relationship could be found. Figure 2 shovvs a diagram based
160 preizkusnih serij (BSCC) jekla 10CrMoV25 brez zmanjšanja lezne kontrakcije	0
160 experimental series of BSCC onJOCrMoV
2 5 without diminiched creep reduction of area
0 0.05 0.10 0,15 0,20 0.25 0,30 Cu+5x Altop - Cu* 5 x A/so/ (%)
a a a.,
a Malo duktilen prelom - Low ductility fraeture
o Duktilen prelom - Ductile fraeture a a Pri preizkusu lezenja - In creep test
aa
A ZSAA
se v istem intervalu znižajo do zelo majhnih vrednosti. Osnovni mehanizem vsebuje relativno trdnost lezenja notranjosti zrna in kristalne meje, nadaljnji vidiki pa bodo omenjeni kasneje. Dodatno k temu pa včasih domnevajo tudi na vpliv nečistoč, ali pa so ta vpliv ugotovili v eksperimentih. Od leta 1967 je le malo podatkov o vplivu aluminija ali bakra na lomno duktilnost. Ratcliff in Brovvn11 poročata o škodljivem vplivu 0,020 % topnega Al na lomno duktilnost 1 % Cr-Mo-V jekla. Njuni vzorci so imeli natezno trdnost 1100 do 1300 N/mm2, kar je zunaj tehničnih mej. Beneš in Skvor12 poročata, da 0,25 % Cu v Cr-Mo-V jeklu zniža lomno duktilnost pri preizkusih lezenja na temperaturi 600 °C. Hopkins in sodelavci13 so ugotovili pri jeklu, podobnem 28 Cr-Mo-Ni-V 4 9, ki je imelo bainitno strukturo, daljši čas do loma in večjo lomno duktilnost, če je bilo to jeklo izdelano laboratorijsko in zelo čisto, v primerjavi z jeklom običajne čistoče. Vzorci so imeli natezno trdnost okoli 1000 N/mm2, kar je preveč za to vrsto jekla. Posledica je bil nizek lomni raztezek pri čistem jeklu (6 % po 13000 urah do loma) in prav tako tudi pri nečistem jeklu (3,9 % po 4500 urah do loma). Tudi Visvvanathan v svojem delu14 omenja vpliv Al in Cu ter drugih elementov na lomno duktilnost.
Priložnost imamo uporabiti veliko podatkov, ki jih je prijazno dal na razpolago nemški Komite za preizkušanje lezenja, iz katerih je razvidna zveza med vsebnostjo Cu in Al v vzorcih in njihovim vedenjem pri lomih. Ta najdena zveza je kvalitativna. Slika 2 kaže diagram, osnovan na rezultatih okoli 120 serij preizkusov lezenja različnih Cr-Mo-V jekel. Ordinata predstavlja natezne trdnosti vzorcev, x-os pa poljubno kombinacijo vsebnosti Cu in Al. Temperatura preizkušanja lezenja je bila večinoma 550 °C. Če so bili raztezki ali kon-trakcije po lomu -majhni, ali če so bili časi do loma zarezanih vzorcev krajši od tistih pri gladkih vzorcih (pri isti obremenitvi), je rezultat preizkusa označen s trikotno točko. Žilavi lomi so označeni s krogi. Vidi se, da mora biti pri večjih vsebnostih Cu in Al natezna trdnost nižja, da dobimo žilav prelom, in obratno. V to shemo je zajetih 80 % preizkusnih serij.
Predpostavljamo, da Al in Cu ne znižujeta površinske energije faz na kristalnih mejah, kot so karbidi ali druge. Nastane vprašanje, na katerem mehanizmu je osnovana ta zveza med obema elementoma, natezno trdnostjo in prelomom vzorcev pri preizkusih lezenja. Da bi se temu posvetili bolj podrobno, smo napravili preizkuse lezenja s talinami različnih čistosti.
V tabeli 4 je preliminaren pregled dela teh preizkusov. Vključena je talina z veliko čistostjo in druga, ki vsebuje Cu in Al. V tabeli so navedena avstenitizacijska temperatura TA, natezna trdnost Rm in rezultati preizkusov lezenja pri štirih nape-
on the results of about 120 -series of creep tests of different Cr-Mo-V steels. The ordinate represents the tensile strength of the specimens, the x-axis represents some arbitrary combination of their copper and aluminium contents. The creep testing temperature was mainly 550 °C. If rupture elonga-tions or reductions of area were of a few percents only or if the rupture life of a notched specimen was shorter than that of a smooth specimen at the same load, the result is marked by a triangle. Tough ruptures are marked by a circle. It can be seen that the higher the oopper and aluminium content, the lovver must the tensile strength be to give tough rupture and vice versa. 80 percent of the test series fit into the scheme, so there >should be something significant behind it.
It is supposed, that aluminium and copper are not effective in reducing the surface energy of phases like carbides or others in the grain boun-dary. So the question arises whether there is some mechanism by which the relation between these elements, the tensile strength and the rupture behaviour of creep specimens is established. To follow this into more detail, we have carried out creep tests with melts of -different purity. In table 4, a preliminary survey of a part of this investi-gation is given. A melt of high purity and another one vvith copper and aluminium impurities are in-cluded. The austenitizing temperature, TA, the tensile strength, Rm, and the results of creep tests at four different levels of creep load are noted. The creep temperature vvas 550 °C in ali cases. Combined specimens vvith a plain and a notched part vvere used. The rupture life of both parts is given. Besides, the rupture elongation (A) and re-duction of area (Z) are also given.
970 °C -is generally accepted as being the upper limit of austenitizing temperatures for this steel in practice. It is seen, that, using it, a tensile strength of 1140 N/mm2 is too high, if creep embrittlement should be avoided even in a high purity metal 1050 °C is knovvn to be an austenitizing temperature enchancing creep embrittlement, if the material is tempered too lovv and hence has a high tensile strength. From line 2 and 3 it may be deduced, that under these circumstances, a tensile strength of 970 N/mm2 is too high, and leads to creep embrittlement vvhile this seems not to be the čase vvith a tensile strength of 770 N/mm2. With the irnpure material, an austenitizing temperature of 1050 °C, vvhich is knovvn to be too high, leads to embrittlement rather early even at a tensile strength of 830 N/mm2, vvhich is vvithin the practical limits. A lovver austenitizing temperature of 890 °C has to be used and rather impractical lovv tensile strengths, namely 550 or 650 N/mm2, to avoid creep embrittlement.
Temper embrittlement and creep embrittlement follovv different mechanisms. As to the first, the cohesive forces betvveen adjacent grains are lovvered by impurities such as P, Sn, Sb, and others, vvhich segregated to the grain boundaries.
Tabela 4: Preizkusi lezenja za pojav lezne krhkosti
Table 4: Creep experiments for the process of creep embrittlement
Rm
N/mm2
400
Napetost, Stress, N/mm2 310	260
200
150
0,01 Cu
970°
1140
O 366 A 565 A 14,7 Z 29,5
o 3555 a 3380 A 6,9 Z 8,5
O 6020 a 5571 A 5,5 Z 3,7
< 0,003 Al
0,001 Sb 0,005 Sn
970
1050°
770
O	838
A	208 A 1,5 Z 3,5
O 1788 A 486 A 1,2 Z 0,6
O 176 A 412 A 28,5 Z 76,6
O	744
A	989
A	15,9
Z	44,2
O
100
o
100
A
550
0,15 Cu 0,031 Al 0,003 Sb 0,014 Sn
890°
650
O 32
A	292 A 38,1 Z 80,4
O	149
A	864
A	43,4
Z	79,8
O 839 A 2969 A 21,1 Z 63,1
O	7386
A	7698
A	13,6
Z	31,6
1050°
830
O 208 A 299 A 15,5 Z 35,6
O	931
A	514 A 6,4 Z 3,9
O 6217 A 5426 A 6,6 Z 15,6
100: O 29694 A 19012 A 9,1 Z 11,4
O Gladki vzorci, plain specimens A Zarezani vzorci, notched specimens
tostih. Temperatura preizikušanja je bila v vseh primerih 550 °C. Uporabljeni so bili gladki in zarezani vzorci. Pri obeh vrstah vzorcev so navedeni časi do preloma. Razen tega so v tabeli zabeleženi tudi raztezki (A) in kontrakcije (Z).
V splošnem velja, da je za to jeklo v praksi zgornja temperatura avstenitizacije 970 °C. Če uporabimo to temperaturo, kaže, da je natezna trdnost 1140 N/mm2 previsoka, če se hočemo celo pri zelo čisti talini izogniti krhkosti pri lezenju. Znano je, da temperatura avstenitizacije 1050° C poveča krhkost pri lezenju, če je bil material prenizko popuščen in ima zato visoko natezno trdnost Iz druge in tretje vodoravne vrste v tabeli 4 lahko sklepamo, da je v teh okoliščinah natezna trdnost 970 N/mm2 previsoka in vodi do lezne krhkosti, medtem ko do tega ne pride pri natezni trdnosti 770N/mm2. Pri nečistem materialu povzroči avstenitizacijska temperatura 1050 °C krhkost celo pri natezni trdnosti 830 N/mm2, ki je znotraj prak-
As to the seoond, cavity nucleation at grain boun-dary carbides or other phases during grain boun-dary creep plays the main role. On the other hand, the cohesive force betvveen grain boundary phases and the matrix vvill also be lovvered by impurities. It is knovvn that brittle intergranular fracture has its origin in nucleation and growth of cavities mainly at the interface between carbides and matrix. Novv the critical free energy AG of cavity nucleation is
cr2 '
vvere y is the surface energy per unit area, F a geometrical factor, and a the stress acting at the plače of the nucleating cavity. Due to the exponent 3 of AG is highly sensitive to changes of the surface energy, for instance by impurities. More-over, the matrix vvith a higher strength has the higher average stress along the grain boundary and subsequently higher void grovvth rates16. This
tičnih mej. Treba je bilo uporabiti nižjo avsteniti-zacijsko temperaturo 890 °C in jeklo popustiti na prenizke natezne trdnosti 550 ali 650 N/mm2, da se ni pojavila lezna krhkost.
Popustna in lezna krhkost sledita različnim mehanizmom. Prvič, elementi P, Sn, Sb in drugi, ki se izcejajo na kristalnih mejah, znižajo kohezivne sile med sosednjimi zrni. Drugič, med lezenjem po kristalnih mejah igra najvažnejšo vlogo nukleacija praznin pri karbidih in drugih fazah na kristalnih mejah. Na drugi strani pa nečistoče znižajo kohezivne sile med fazami na kristalnih mejah in matrico. Znano je, da ima krhki inter-kristalni prelom izvor pretežno v nukleaciji in rasti praznin na mejni plosikvi med .karbidi in matrico. Kritična prosta energija AG za nukleacijo praznin je
kjer je r površina energije na enoto ploskve, F geometrijski faktor in cr napetost, ki deluje na mestu nukleacije praznine. Zaradi eksponenta 3 pri y je AG zelo občutljiv za spremembe površinske energije, na primer zaradi nečistoč. Poleg tega ima trdnejša matrica večje povprečne napetosti vzdolž kristalnih mej in s tem večje hitrosti rasti praznin16. To je v skladu z opažanji, da je krhki lom pri lezenju bolj verjeten pri visoki trdnosti jekla kot pri nizki. Vlogo določenih nečistoč pri nastajanju lezne krhkosti moramo še nadalje zasledovati, zlasti s preizkusi lezenja pri materialih z martenzitno mikrostrukturo.
is in agreement with the observation that brittle creep failure is more likely to occur at high ten-sile strength than at a low one. Therefore, the effect of certain impurities on creep embrittlement shall be investigated mainly vvith creep tests on materials vvith martensitic microstructure vvhich had shovvn temper embrittlement as men-tioned above.
Literatura - References
1.	Guttmann, M.: Phil. Trans. R. Soc. Lond. vol. A 295, 1980, p. 169.
2.	Hondros, E. D.: Proc. R. Soc. vol. 286, 1965, p. 479.
3.	Seah, M. P., E. D. Hondros: Proc. R. Soc. vol. 335, 1973, p. 191.
4.	McMahon, C. J., V. Vitek: Acta Met., vol. 27, 1979, p. 509.
5.	Kameda, J., C. J. McMahon: Met. Trans., vod. /11 A, 1980, p. 91.
6.	Ohtani, H., H. C. Feng, C. J. Mc Mahon: Met. Trans, vol. 7 A, 1976, p. 1123.
7.	Fischmeister, H. F., I. Olefjord: Berg- und Hiitten-mannische Monatsh., vol. 123, 1978, p. 75.
8.	Seah, M. P., E. D. Hondros: Script. Met. vol. 7, 1973, p. 735.
9.	Guttmann, M.: Phil. Trans. R. Soc. Lond. A. 295, 1980 p. 169.
10.	Kriisch, A., F. K. Naumann, H. Keller, H. Kudielka: Archiv Einsenhiittenvv. vol. 42, 1971, p. 353.
11.	Ratoliff, J. H., R. M. Brovvn: Trans'. ASM vol. 60, 1967, p. 176.
12.	Beneš, F., P. Skvor: iHutn. Listy, vol. 3, 1972, p. 197.
13.	Hopkins, B. E., H. R. Tipler, G. D. Branch: Journ. ISI, vol. 209, 1971, p. 745.
14.	Visvvanathan, R.: Metals Engng. Quart. Nov. 1975, p. 50.
15.	Raj., R., M. F. Ashby: Acta Met. vol. 23, 1975, p. 653.
16.	Pope, D. P., et. lal.: EPRI Interim Report 1979, Penn-sylvania Univ.
RAZPRAVA
C. Goux, ENSM, Saint Etienne
Kakšen je razlog za nizko vsebnost silicija, ki je omenjena v eni od vaših slik?
M. Krause
Nizko vsebnost silicija v nekaterih od preizkušenih materialov smo izbrali, da bi določili, če silicij pospešuje popustno krhkost v jeklih te vrste. Na drugi strani je mogoče, da bo to jeklo dezoksidirano po postopku vakuumske dezoiksi-dacije z ogljikom na talinah, ki so praktično brez dodatka silicija.
S. Engineer, Thyssen Edelstahlvverke, Krefeld
Kako pomembna je udarna žilavost pri sobni temperaturi za lezne lastnosti in prelom teh jekel?
M. Krause
Če mislite kako značilne so nekatere žilavosti pri sobni temperaturi s stališča leznih lastnosti pri delovni temperaturi, naj vas spomnim, da ima ve-
DISCUSSION
C. Goux, ENSM, Saint Etienne
What is the reason for a very lovv silicon content mentioned in one of your slides?
M. Krause
The lovv isilicon content in some of the test materials vvas chosen as a basis for determining vvhether silicon might promote temper embrittlement in this type of steel. On the other hand, it could be possible that this steel vvill be deoxidised by the vacuum carbon deoxidation procedure vvhich vvould be done on heats vvith nearly no addition of silicon.
S. Engineer, Thyssen Edelstahlvverke, Krefeld
Hovv important are the impact toughness properties at room temperature for the creep rupture properties of these steels?
M. Krause
If you mean hovv indicative certain values of the room temperature impact toughness are re-garding the creep rupture properties at service temperature, you may recollect, that high impact
liko udarno žilavost mikrostruktura iz martenzita ali spodnjega bainita, ki pa imata oba majhno lezno trdnost, medtem ko ima mikrostruktura iz zgornjega bainita nizko žilavost in največjo odpornost proti lezenju v jeklih te vrste.
Zato so sprejemljive nizke žilavosti do 10 J za izkovke za turbinske rotorje za delo pri visoki temperaturi.
F. Vodopivec
Imam vprašanje, ki morda ni neposredno povezano s temo vašega predavanja. V ohišjih ventilov za visokotlačne parne kotle smo opazili, da se relativno velike razpoke (dolžine nad 100 mm, globine do 15 mm) lahko razvijejo med letnimi ali dvoletnimi revizijami. Ohišja so iz jeklene litine podobne sestave, kot jekla katera obravnavate v vašem delu, le vsebnost ogljika je nižja, okoli 0,15%. Delovna temperatura je približno 515 °C, pritisk pa 125 at. Elektrarna je namenjena za pokrivanje energetskih konic. Ali mislite, da so razpoke izključno posledica malociklične utrujenosti, ali je vmes tudi vpliv krhkosti?
M. Krause
Ohišja ventilov so obremenjena z malociklično utrujenostjo, posebno v koničnih centralah. To bi bilo bolj nevarno za krhko kot za žilavo jeklo. Kaj je v Vašem primeru razlog za pokanje se lahko določi samo na osnovi mikrostrukturnih raziskav, popolnega poznavanja mehanskih lastnosti, kemične sestave, termične obdelave in delovnih karakteristik ventila.
toughness is correlated with martensite or martensite plus lower bainite, both with poor creep resi-stance, and upper bainite, vvith lovv impact values, represents the microstructure vvith the highest creep resistance in this type of steel. Therefore, if 1 % Cr-Mo-V turbine rotor forgings for high temperature service are considered, impact values as lovv as about 10 J vvill be sufficient.
F. Vodopivec
I have a question vvhich is not maybe direetly related to the subject of your paper. We observed that in valve casings for high pressure steam po-vver station relatively great craoks (length over 100 mm and depth to 15 mm) could develop betvveen annual or biannual revisions. The casings are manufactured from čast steel of similar composition as the steel vvhich you discussed in your communication, only the content of carbon is lovver, appr. 0.15 %. The vvorking temperature is appr. 515 °C and the pressure 125 atm. The povver station is intended to cover peak energy requirements. Do you mean that the craoks are due only to lovv cycle fatigue, or brittleness pheno-mena are involved also?
M. Krause
Valve bodies are exposed to low cycle fatigue especially in povver stations covering consumption peaks. This vvould be more harmfull in a brittle material than in a tough one. But vvhat, in your čase, has been the reason for craoking can only be assessed by microstructural examinations and complete knovvledge of the mechanical properties, the ehemical composition, the heat treatment of the material and the service caracteristics of the valve.