Zapozneli lom jekla z visoko trdnostjo Deiayed Fracture of High-strength Steel B. Ule*, F. Vodopivec*, J. Žvokelj*, M. Grašič* in L. Kosec** UDK: 669.14.018.2:539.56:620.192.3 ASM/SLA: Q26s, SG Ba, ST, 2-60, EGn, 3-66 V članku so opisane teoretične osnove, potrebne za razumevanje napetostno induciranega segregiranja vodika v jeklu z visoko trdnostjo. Opisano je merjenje kritičnega in mejnega napetostnega intenzitetnega faktorja ter na tej osnovi analiziran vpliv malih mikrostrukturnih variacij jekla na njegovo občutljivost k zapoznelemu lomu. 1. UVOD Ena od znanih oblik porušitve jekla z visoko mejo plastičnosti ter trdnostjo nad 1200 Nmm-2 je zapozneli lom, ki nastane zaradi napetostno induciranega segregiranja vodika v jeklu. Raziskave zapoznelega loma, ki jih je opravil G. L. Hanna s sodelavci1, kažejo, da obstaja inkubacijski čas do nastanka prve mikrorazpoke; ta počasi raste, vse dokler ne doseže kritične velikosti, kar privede do hipne porušitve. Tako inkubacijski čas, kot tudi čas do loma se podaljšujeta z zmanjševanjem obremenitve, vse dokler pri neki dovolj nizki obremenitvi zapozneli lom sploh izostane. A. R. Troiano2 je odkril, da nukleacija mikrorazpok izostane tudi v primeru pravočasne razbremenitve, vendar pa se mikrorazpoke pojavijo potem, ko je jeklo ponovno daljši čas mehansko obremenjeno. To vodi do sklepa, da je nukleacija mikrorazpok posledica elastične interakcije med mobilnimi atomi vodika v jeklu ter polji troosnih napetosti ob različnih diskonti-nuitetah kovine. Lokalno kopičenje vodika v zadostni količini pa poslabša kohezivnost mreže ter olajša nukleacijo mikrorazpok. 2. TEORETIČNI DEL Atomarni vodik v železu je bodisi na intersticijskih mrežnih mestih, bodisi ujet na različnih napakah kri- stalne mreže, ki jih imenujemo pasti. Nekaj vodika naj- demo vedno tudi v molekularni obliki v porah. Koncept pasti sta predlagala Darken in Smith3, da bi pojasnila vpliv temperature in koncentracije vodika v železu na njegovo difuzivnost. Oriani4 je kasneje na osnovi različnih eksperimentalnih podatkov izračunal ve-zavno energijo, s katero je vodik vezan v pasteh. Skoraj v vseh primerih je dobil vrednosti okrog 27 kJ na mol vodika. Podobne vrednosti so navedene tudi v referencah 5,6 in 7, čeprav sta Kumnick in Johnson8 odkrila tudi pasti z vezavno energijo približno 60 kJ na mol vodika. Gostoto teh pasti, vejetno so bili to dislokacijski pragovi, sta ocenila na 1023 m-3 v močno deformiranem železu. Danes je znano, da kot pasti delujejo skoraj vse nepravilnosti kristalne mreže kovin, tako dislokacije9-10, The paper presents theoretical fundamenta/s to un-derstand the stress-induced hydrogen segregation in high-strength steel. Measurements of the critical and of the threshold stress intensity factor are presented vvhich represent the basis for analysing the influence ofsmall microstruc-tural variatious of steel on its sensitivity to the delayed fracture. 1. INTRODUCTION Delayed fracture caused by stress-induced hydrogen segregation is one of the knovvn types of failure of steel vvith high yield strength and tensile strength above 1200 N mm-2. Investigations of the delayed fracture by G. L. Hanna and covvorkers1 shovved an incubation period before the nucieation of the first microcrack, a slovv growth of the microcrack and an instantaneous failure vvhen a critical size vvas reached. The incubation period as well as the tirne till failure occurs are prolonged vvith the decreased load until at a sufficiently lovv load the delayed fracture does not occur at ali. A. R. Troiano2 found that nucieation of microcracks does not appear if the unloading took plače in the due time, but microcraks occur after steel vvas again mechanical^ loaded for a longer time. This leads to the conclusion that the nucieation of microcraks is a conse-quence of elastic interaction betvveen the mobile hy-drogen atoms in steel and the triaxial stress fields at dif-ferent discontinuities in the metal. Local accumulations of hydrogen at a sufficient level diminish the cohesive forces in the lattice and facilitate the nucieation of cracks. 2. THEORY Atomic hydrogen in iron is found either in interstitial sites of the lattice or it is as trapped hydrogen bound on different imperfections of the crystal lattice, being called »traps«. Some hydrogen in molecular form is found al-ways in the microvoids too. Darken and Smith3 suggested the concept of traps to explain the influence of temperature and of concentra-tion on the diffusivity of hydrogen in iron. Later, on basis of experimental data, Oriani4 calculated the trap binding energy of hydrogen and obtained value of about 27 kJ/ mole hydrogen in almost ali the cases. Similar va-lues are quoted also in refs. 5, 6 and 7, vvhile Kumnick and Johnson8 discovered traps vvith the binding energy of about 60 kJ/mole hydrogen too. The density of traps, probably dislocation jogs, vvere estimated to 1023 m~3 in a heavily deformed iron. * Metalurški inštitut, Lepi pot 11, Ljubljana ** Univerza Edvarda Kardelja, FNT — Montanistika, Aškerčeva 20, Ljubljana pore"'12, meje zrn13, mejne površine kovina-karbidni delci14, kot tudi površine nekovinskih vključkov15-l6. Matematični model za opis vodika v pasteh sta razvila Foster in McNabb17. Po njunem modelu je v pasteh ujeti vodik v lokalnem ravnotežju z intersticijskim vodikom. Interakcija med vodikom in pastmi je termično aktiviran proces z aktivacijsko energijo, ki jo sestavljata vezavna energija pasti ter aktivacijska energija za difuzijo vodika v idealni mreži železa, ki dosega vrednost 12 kJ na mol vodika. Gonila sila za intersticijsko difuzijo raztopljenega vodika v kovinah je gradient kemičnega potenciala vodika. K temu gradientu prispevajo koncentracijske razlike ter učinki polj elastičnih napetosti. Termodinamični učinek polj elastičnih napetosti je utemeljen z reverzi-bilno dilatacijo kristalne mreže kovin ter pozitivno spremembo volumna, ki spremlja vgnezdenje vodikovih intersticij v področja pozitivne deformacije, medtem, ko se področja s tlačno deformacijo z vodikom osiromašijo. Na ta način je z nehomogeno prerazporeditvijo vodika dosežen krajevno neodvisen kemični potencial vodika v nehomogenem polju elastičnih napetosti. Li, Oriani in Darken19 so s termodinamično analizo problema, pri čemer so vodik v železu obravnavali kot povsem mobilno komponento, izpeljali naslednjo enačbo: M-h = I^h + RT ln[H] — ah VH (1) v kateri prva dva člena določata kemični potencial vodika v odvisnosti od njegove koncentracije, VH je fenomenološki parcialni atomski volumen vodika v železu, ah pa je hidrostatična komponenta napetostnega tenzor-ja, s katerim popišemo troosno napetostno stanje. Ob predpostavki, da je v ravnotežju kemični potencial vodika krajevno neodvisen, z enačbo (1) izračunamo koncentracijo vodika v področju maksimalne hidrostatične komponente napetostnega tenzorja v razdalji r pred korenom zareze: [H]r = [H]exp(ahV„/RT), (2) kjer je [H] povprečna koncentracija vodika v preizku-šancu. Za ravninsko deformacijsko stanje, način obremenitve I ter za ravnino napredovanja razpok iz korena zareze uporabimo naslednjo enačbo20: ah = 2(l+v)K,/3y2jtr (3) Iz enačb (2) in (3) dobimo: [H]r = [H]exp2^1+V)/^_V" (4) 3 RT|/2ir Enačba (4) povezuje napetostni intenzitetni faktor K, ter koncentracijo vodika v razdalji r pred korenom zareze. Dejstvo, da pri apliciranem napetostnem intenzitet-nem faktorju K,, ki je nižji od mejnega napetostnega in-tenzitetnega faktorja KTH zapozneli lom sploh izostane, je Beachem21 zapisal v obliki kriterija, ki določa pogoje pojavljanja tega loma: KthKth=> [H]r = [H]" (6) At present, it is knovvn that almost ali the kinds of im-perfections in the crystal lattice of metals, both disloca-tions310, microvoids11,12, grain boundaries13, metal-car-bide interfaces14, and the surfaces of non-metalic inclu-sions15-16 can act as trapping sites. A mathematical model of hydrogen traps was deve-loped by Foster and McNabb17. According to the model, trapped hydrogen is locally in equilibrium vvith the inter-stitial hydrogen. The interaction betvveen hydrogen and the traps is a thermally activated process with an activa-tion energy constituted of the binding trap energy, and the activation energy of diffusion of hydrogen in an ideal iron lattice being 12 kJ/mole hydrogen18. The driving force for the interstitial diffusion of the dis-solved hydrogen in metals is the gradient of chemical potential. This gradient is influenced by the differences in hydrogen concentrations and by the effects of elastic-stress fields. The thermodynamic effect of the elastic-stress fields is caused by the reversible dilatation of the crystal lattice of metals, and the positive volume change accompanyed by the insertion of hydrogen interstitials into the areas of positive strain, vvhile compressively strained regions are impoverished vvith hydrogen. Thus a locally independent chemical potential of hydrogen in an inhomogeneous elastic-stress field is obtained through an inhomogeneous redistribution of hydrogen. The thermodynamic analysis of this process, under supposition that hydrogen is a completely mobile com-ponent, vvas established bi Li, Oriani and Darken19 who developed the follovving equation: Rh-^ + RT ln[H] — ohVH (1) The first two terms determine the chemical potential of hydrogen depending on its concentration, VH is the phenomenological partial atomic volume of hydrogen in iron, vvhile oh is the hydrostatic component of the stress tensor describing the triaxial state of stresses. Suppos-ing that the chemical potential of hydrogen in equilibrium is locally independent, the concentration of hydrogen in the region of the maximal hydrostatic component of stress tensor is given at a distance r from the notch root by (1): [H]r = [H]exp(ahVH/RT), (2) vvith [H] as an average concentration of hydrogen in the specimen. For a plane strain state, for a mode of loading I, and for the plane crack propagation from the notch root, the follovving equation is given20: CTh = 2 (1 4-v) K,/3 l/2jtr (3) From equations (2) and (3) vve obtain: [H]r = [H]exp2'1+;)^H (4) 3 Rl]/2nr The equation (4) connects the stress intensity factor K, vvith the concentration of hydrogen at the distance r from the notch root. The fact, that the delayed fracture does not occur if the applied stress intensity factor K, is lovver than the threshold stress intensity factor KTH, has been applied by Beachem21 as the criterion to determine the conditions under vvhich the delayed fracture takes plače: Kth Kth=> [H]r=[Hr Combining Eqs. (4) and (6) we obtain: Kth = 3 RT ]/27tr 2(1+v)VH In [H] (6) (7) The equation (7) is correct only vvhen the plastic zone at the notch tip is limited to less than the grain diameter d. Considering that also hydrogen on the grain bounda-ries at the crack tip is involved, vve can vvrite: 3RTJ^ |nW 2 (1 + v) VH l [H] ; d>Rh (8) vvith R, as the size of the strain plastic zone under mode of loading I. On the basis of equation (8) it is not possible to ex-plain the connection betvveen the yield strength effect and Kth, though it has been established by experiments. If taking into account the influence of a slip-line field at the notch root, according to Gerberich22, vve obtain for Kth: Kth = ~ in ctVH m [H] 2a (9) It is necessary to mention that disagreements are of-ten observed betvveen Eq. (9) and experimental results at yield strength belovv 1200 N mm-2. These disagreements vvere explained partly by the dependence of the [H]cr/[H] ratio on the yield point. Namely Farrell and Quarrell25 ascertained that larger concentrations of hy-drogen are needed to produce embrittlement in steel vvith lovver yield strength, and postulated the relationship [H]cr oc 1/oys. Kim and Loginovv26 suggested that the content of sol-uble hydrogen in steel vvas proportional to the yield strength, therefore [H]«ays. Finally, vve obtain: (10) [H]*_ ft [H] oys' vvith P as constant for a single type of steel. 3. EXPERIMENTS AND RESULTS 3.1 Selection of Steel and the Geometry of Speci-mens The aim of the investigation vvas to establish the influence of microstructure on the hydrogen induced sus-ceptibility to cracking of a high-strength steel vvith the composition: 0.40 % C, 0.31 % Si, 0.71 % Mn, 0.019 % P, 0.006 % S, 1.03 % Cr, 0.21 % Mo, 0.26 % Cu, 0.009 % Al and 0.010 % Sn. The steel vvas manufactured by the VOD process, thus the content of sulphur vvas low and the concentration of residual hydrogen did not exceed 0.05 ppm. Tensile tests vvere made on notched tensile speci-mens vvith the geometry shovvn in Fig. 1. For these specimens, the relationship betvveen the stress intensity factor K|, the geometry, and the axial force P is given by the eguation27: K| = 5^i (-1.27+1,72 D/d) (11) macije dosežene v samem korenu zareze, kjer se sicer pojavljajo prve mikrorazpoke. Slika 1 Geometrija cilindričnih nateznih preiskušancev z zarezo po obodu. Fig. 1 Geometry of cylindrical round notched tensile specimens. under condition that: 0,5 < d/D <0,8 The ratio p/D, vvith p as the notch root radius, vvas close to the value 0.02. Moran and Noriš28 found by the computer simulation of the tension test vvith cylindrical, peripherally notched specimens, that the maximum stresses at fracture occur at about two notch-root radii belovv the surface vvhen the p/D ratio is 0.01 to 0.02. On the other hand the maximum strain occurs at the notch root vvhere also the first microcracks appear. 3.2 Thermal Treatment Thermal treatment of specimens consisted of a 30 mins. austenitisation at 850° C, quenching in vvater or in oil, and tempering. Fig. 2 shovvs the microstructure of the lath-formed martensite in the steel quenched in vvater. After tempering 2 hrs. at 480° C or 420°C yield strengths of 1185 and 1290 N mm-2 respectively vvere obtained. The hardness of oil quenched specimens vvas betvveen 52 and 53.7 HRC. This means, that the predominantly martensitic microstructure of oil quenched specimens (Fig. 3) stili contains up to 3 % of bainite. After tempering specimens quenched in oil 2 hrs. at 450° C, a yield strength of 1230 N mm-2 vvas obtained. 3.2 Toplotna obdelava Toplotna obdelava preizkušancev je obsegala 1/2-ur-no avstčnitizacijo pri 850°C s kaljenjem v vodi oziroma olju ter popuščanje. Slika 2 prikazuje mikrostrukturo letvastega martenzita, izoblikovanega pri kaljenju v vodi. S popuščanjem 2 uri pri 480 oziroma 420° C je bila dosežena meja plastičnosti 1185 oziroma 1290 Nmm-2. Slika 2 Avstenitizirano pri 850°C in kaljeno v vodi. Letvasti martenzit. Fig. 2 Austenitized at 850° C, and quenched in vvater. Lath-shaped martensite. Slika 3 Avstenitizirano pri 850°C in kaljeno v olju. Spodnji bainit v mart-enzitni osnovi. Fig. 3 Austenitized at 850°C, and quenched in oil. Lovver bainite in the martensitic matrix. 3.3 Hydrogen Charging After thermal treatment, the specimens vvere charged vvith hydrogen by etching 24 hrs. in a 0.1 N aqueous so-lution of hydrochloric acid. Chemical analysis of specimens immediately after the removal from the acid solution shovved a hydrogen con-centration of 2.9±0.1 ppm independently upon the yield Trdota v olju kaljenih preizkušancev je dosegala vrednosti med 52 in 53,7 HRc. Pomeni, da je v pretežno martenzitni mikrostrukturi tovrstnih preizkušancev po kaljenju v olju (si. 3) še tudi do 3 % bainita. S popuščanjem v olju kaljenih preizkušancev 2 uri pri 450° C je bila dosežena meja plastičnosti 1230 Nmm"l 3.3 Navodičenje Toplotni obdelavi je sledilo navodičenje preizkušancev z jedkanjem 24 ur v 0,1 N vodni raztopini solne kisline. Kemične analize vzorcev neposredno po odstranitvi iz kisline kažejo, da dobljena koncentracija vodika 2,9 ±0,1 ppm praktično ni odvisna od meje plastičnosti jekla. Z drugimi besedami: s 24-urnim jedkanjem še ni dosežena stacionarna koncentracija nasičenja jekla z vodikom. Zaključujemo, da del pasti še ni zaseden, saj bi v takšnem primeru bila koncentracija vodika različna v preizkušancih z različno mejo plastičnosti26'29. 24 ur po odstranitvi iz raztopine je koncentracija vodika v preizkušancih padla na 0,82 ±0,1 ppm ter 48 ur po odstranitvi na 0,58 ±0,08 ppm. Ob predpostavki, da je hitrost uhajanja vodika iz cilindričnih preizkušancev majhnega premera (D je približno 7 mm) premo sorazmerna razliki med trenutno ter residualno koncentracijo vodika v njih, dobimo za koncentracijo vodika v odvisnosti od časa po jedkanju (v urah) naslednji izraz: [H] = 0,55+ 2,35 exp( —0,09 t) (12) Polempiričen izraz (12) je uporaben za fenomenološki opis uhajanja vodika iz cilindričnih preizkušancev in ugotovljeno je bilo30, da v navodičenih preizkušancih ostaja še okrog 0,55 ppm vodika tudi dolgo časa po končanem jedkanju. 3.4 Določevanje kritičnega ter mejnega napetostnega intenzitetnega faktorja Razvit je bil eksperimentalni sklop za registriranje inkubacijskega časa, to je časa do porajanja prve mikro-razpoke ter za registriranje počasnega napredovanja mikrorazpok do hipnega loma statično obremenjenih preizkušancev z zarezo po obodu. Sestavljen je bil iz polovičnega Wheatstonovega mostička z variabilnim uporom, ki ga je predstavljal uporovni listič, nalepljen preko ustja zareze. Porajanje ter napredovanje mikrorazpok je bilo registrirano posredno z odpiranjem ustja zareze, kot sprememba upornosti aktivnega uporovnega strength of steel. In the other vvords, a 24 hrs. etching did not produce the saturation of steel with hydrogen. It was concluded that ali the traps were not filled since in such a čase the concentration of hydrogen in steel vvould be different in samples with different yield strengths26 Twenty-four hours after removal from the acid solution, the concentration of hydrogen in speci-mens dropped to 0.82 ±0.1 ppm and after 48 hours to 0.58±0.08 ppm. Supposing that the escape rate of hydrogen from the cylindric specimens vvith small diameter (D is approx. 7 mm) is proportional to the difference betvveen the actual and the residual hydrogen concentration, the follovv-ing equation can be derived for the variation of hydrogen concentration vvith tirne (hours) after the removal of samples from the acid solution: [H] = 0,55+ 2,35 exp (— 0,09 t) (12) This semiempirical equation is useful for the pheno-menological description of hydrogen losses from the cylindric specimens and as it has been established30 that the specimens charged vvith hydrogen stili contain residual hydrogen of about 0.55 ppm even a long tirne after the etching. 3.4 Determination of the Critical and the Threshold Stress lntensity Factor An experimental set-up was developed for the regis-tration of the incubation period i. e. of the tirne neces-sary for the nucleation of the first microcrack, as vvell as for the registration of the slovv propagation of micro-cracks to the instantaneous fracture of the round-notched specimens under static load. It consisted of a half-Wheatston bridge vvith a variable resistor represent-ed by a strain-gauge sticked across the notch opening. Nucleation of microcracks and their propagation were indirectly registered by the displacement of the notch opening as the change of the resistance of the active strain-gauge compared vvith the resistance of the reference strain-gauge. This experimental set-up (schemati-cally shovvn in Fig. (4)) permitted to detect the propagation steps of about 0.1 um. An almost similar set-up vvas used to measure the critical stress intensity factor i. e. fracture toughness of steel. Fig. 5 shovvs hovv the measurements vvere made on the "Instron" tensile machine vvith an accurate exten-someter mounted on the notch opening of the specimen for calibration of the strain-gauges. Slika 4 Eksperimentalni' sklop za zasledovanje porajanja ter napredovanja mikrorazpok (1 — natezni preiskušanec z uporovnim lističem, 2 — referečni uporovni listič, 3 — merilna enota z izvorom napetosti, galvanometrom ter ojačevalcem, 4 — registrator). Fig. 4 Experimental set-up for the detection of crack nucleation and propagation (1 — tensile specimen vvith strain-gauge, 2 — reference strain-gauge, 3 — measuring unit vvith povver source, galvanometer and amplifier, 4 — recorder). Slika 5 Merjenje lomne žilavosti. Ekstenzometer na preiskušancu služi kalibriranju uporovnih lističev. Fig. 5 Measurement of fracture toughness. Extensometer on the ten-sile specimen used for the calibration of the strain-gauges. lističa glede na upornost referenčnega uporovnega lističa. Eksperimentalni sklop (shematsko prikazan na sliki 4) je dovoljeval zaznavanje koraka propagacije okrog 0,1 ji,m. Skoraj podoben sklop opreme je bil uporabljen za merjenje kritičnega napetostnega intenzitetnega faktorja, t. j. lomne žilavosti jekla. Na sliki 5 je prikazana izvedba merjenja na trgalnem stroju »Instron« s preciznim ekstenzometrom, montiranim preko ustja zareze preizkušanca ter uporabljenim za kalibriranje uporovnih lističev. Z merjenjem časa do loma preizkušancev v odvisnosti od uporabljene obremenitve je bil eksperimentalno določen mejni napetostni intenzitetni faktor KXH, to je mejna statična obremenitev, pri kateri še ne pride do nukleacije mikrorazpok. Kritični napetostni-intenzitetni faktor Klc, t. j. lomna žilavost jekla je bila izmerjena na cilindričnih preizku-šancih z zarezo ter utrujenostno razpoko v korenu zareze. Tako kot za izračun mejnega, je bila tudi za izračun kritičnega napetostnega intenzitetnega faktorja uporabljena formula (11), globina utrujenostne propagacije mikrorazpoke pa je bila izmerjena z optičnim mikroskopom po vsakokratnem preizkusu. Za preverjanje rezultatov je bila lomna žilavost izračunana še s korelacijo Rolfe-Novak31 za takoimenovano upper shelf področje. Odvisnost med izmerjenimi faktorji (Kth, K,c) ter mejo plastičnosti jekla je prikazana S -C* "E J3C O >4— E -z. C i- 4-> c C c <11 S C J) £ 1/1 i> o 1— c to o u č S KJ 4-. L. b •D C C O X I t— ^ "O E a? O) Z -C t- Slika 6 Odvisnost med napetostnim intenzitetnim faktorjem (KTH, l|C) in mejo plastičnosti preiskovanega jekla. Fig. 6 Relationship betvveen the stress intensity factor (KTH, K,c) and the yield strength of the investigated steel. The threshold stress intensity factor KTH vvhich repre-sents the limiting value of static load belovv vvhich micro-cracks do not appear, was experimentally determined by measuring the tirne till fracture occurs related to the ap-plied load. The critical stress intensity factor K,0 i. e. the fracture toughness of steel was measured on round notched tensile specimens vvith fatigue crack at the notch tip. The Eq. (11) was applied to calculate both the threshold and the critical stress intensity factor. The vvidth of the fatigue crack vvas measured vvith an optical microscope after each experiment. To check the obtained results, the fracture toughness vvas also calculated by the Rolfe-Novak correlation31 for the upper shelf region. The relation betvveen the measured factors (Kth, K,c) and the yield strength of the investigated steel is shown in Fig. 6. The plot presents also the relation by eq. (9) calculated on the basis of measured values for steel vvith fully martensitic microstructure after quenching (points M). Considering the equa-tion (10), a vaule of 5770 N mm-2 vvas calculated for the constant p. The straight line for [H]cr/[H] = const. proves that linear interpolation is acceptable at yield strengths above 1200 N mm"2. The threshold stress intensity factor KTH for tempered martensitic-bainitic microstructure vvith only fevv per-cents of bainite after quenching (point M + B) has the same value as that for a fully martensitic tempered microstructure vvith the same yield strength (KTH = = 2100 N mm-3'2). Hydrogen has no noticeable influence, if any at ali, on the fracture toughness of the investigated steel. How-ever, at the same yield strength, the fracture toughness, M'. TH] - konstx N o 0,05 ppm hydrogen j}0,55 ppm hydrogen K™-avHlnmr 2ot [H]rr„ 5770 - IhT 800 900 1000 1100 1200 1300 1400 Meja plastičnosti, Yield strength cSys(Nmm2) na sliki 6. V diagramu je vrisana tudi odvisnost (9), izračunana na osnovi izmerjenih vrednosti za jeklo s povsem martenzitno mikrostrukturo po kaljenju (točki M). Upoštevaje izraz (10) ima konstanta p vrednost 5770 Nmm-2. Premica za [H]cr/[H] = konst. dokazuje, da je pri meji plastičnosti nad 1200 Nmm-2 dopustna linearna interpolacija. Mejni napetostni intenzitetni faktor KXH za popušče-no martenzitno-bainitno mikrostrukturo z le nekaj odstotki bainita po kaljenju (točka M + B) ima enako vrednost, kot je bila določena za jeklo z mikrostrukturo popuščenega martenzita ter enako mejo plastičnosti (Kth = 2100 Nmm_3/2). Vodik nima večjega, če ima sploh kakšen vpliv na lomno žilavost preiskovanega jekla. Pri enaki meji plastičnosti pa je lomna žilavost, v nasprotju z mejnim napetostnim intenzitetnim faktorjem, nekoliko odvisna od majhnih mikrostrukturnih variacij jekla. Tako ima jeklo s popuščeno martenzitno-bainitno mikrostrukturo ter mejo plastičnosti 1230 Nmm-2 lomno žilavost med 2310 in 2360 Nmm~3/2, kar je nekoliko več od z linearno interpolacijo določene lomne žilavosti za popuščeno martenzitno mikrostrukturo enake meje plastičnosti (K]c med 2250 in 2320 Nmm"3/2). 3.5 Mikromorfologija prelomov Mikrofraktografske preiskave prelomnih površin na-vodičenih cilindričnih nateznih preizkušancev z zarezo po obodu, so bile opravljene z vrstičnim elektronskim mikroskopom. Slika 7 kaže del prelomne površine nateznega preizkušanca z zarezo, z mikrostrukturo popuščenega martenzita ter mejo plastičnosti 1185 Nmm-2. Statična Slika 7 Z zapoznelim lomom nastala prelomna površina preiskušanca z mikrostrukturo popuščenega martenzita ter mejo plastičnosti 1185 N mm-2 posneta z vrstičnim elektronskim mikroskopom. Fig. 7 Scanning electron micrographs of the delayed fracture surfaces of specimen vvith the tempered martensitic microstructure and yield strength 1185 N mm-2. contrary to the threshold stress intensity factor, de-pends slightly on small microstructure variations of steel. For instance, steel vvith the martensitic-bainitic microstructure vvith yield strength 1230 N mrrr2 has fracture toughness betvveen 2310 and 2360 N mm~3/2, vvhich is slightly above the value of K,c being betvveen 2250 and 2320 N mm"3'2 found by linear interpolation for the tempered martensitic microstructure vvith the same yield strength. 3.5 Micromorphology of Fractures Microfractographic investigations of fracture surfaces of round notched tensile specimens charged vvith hy-drogen were performed in a scanning electron micro-scope. Fig. 7 shovvs a part of fracture surface of a round notched tensile specimen vvith fully martensitic tempered microstructure and vvith yield strength 1185 N mm-2. Static load i. e. applied stress intensity factor 2190 N mm-3'2 vvas close to the limiting value (KTH = 2180 N mm-3'2) vvhich caused the delayed fracture of the specimen after 181 hours. Right along the notch (above), an area of slovv crack propagation can be seen separated by an unsharp boundary from the fracture surface formed by an instantaneous failure (belovv). The area of slovv crack propagation, vvhich is com-pletely undefined at low magnification, is shovvn in Fig. 8 at a higher magnification. This area is predominantly ductile and irregularly shaped dimples are found next to the well defined dimples. Irregular dimples indicate that decohesion occurred at a very lovv plastic deformation. It is also possible that some details are a consequence of cleavage too. A similar micromorphology of the area of slovv crack propagation is observed on samples vvith the tempered Slika 8 Področje počasne propagacije s slike 7. Pretežno duktilna oblika preloma. Fig. 8 The area of slovv crack propagation from Fig. 7. Predominantly ductile fracture. Slika 9 Področje počasne propagacije na preiskušancu s popuščeno martenzitno mikrostrukturo ter mejo plastičnosti 1290 N mm"2. Duktilna oblika preloma s posameznimi cepilnimi ploskvami. Fig. 9 The area of slovv crack propagation in specimen vvith the tem-pered martensitic microstructure and yield strength 1290 N mm"2. Ductile fracture vvith single cleavage facets. obremenitev, namreč aplicirani napetostni intenzitetni faktor 2190 Nmm~3/2, je bila blizu mejni vrednosti (KTH = 2180 Nmm-3/2), kar je povzročilo zapozneli lom preizkušanca po 181 urah. Neposredno ob zarezi (zgoraj) je moč videti cono počasnega napredovanja mikrorazpok, ne ostro ločeno od prelomne površine, nastale s hipno porušitvijo (spodaj). Cona počasnega napredovanja mikrorazpok, ki je pri nizki povečavi povsem neopredeljiva, je pri večji povečavi prikazana na sliki 8. To področje je pretežno duk-tilno, poleg dobro definiranih jamic pa najdemo tudi jamice nepravilnih oblik. Nepravilne jamice kažejo,, da se je dekohezija izvršila z zelo malo plastične deformacije in prav mogoče je, da so nekateri detajli tudi proizvod cepljenja. Podobno mikromorfologijo preloma v coni počasnega napredovanja mikrorazpok zasledimo tudi na preizkušancih s popuščeno martenzitno-bainitno mikrostrukturo, medtem ko je na preizkušancih s popuščeno martenzitno mikrostrukturo ter mejo plastičnosti 1290 Nmm~2 že tudi občutnejši delež cepilnih ali kvazicepilnih ploskvic (slika 9). Področje naglo zlomljenega osrednjega dela preizkušanca s slike 7 je pri večji povečavi prikazano na sliki 10. Prevladujejo področja duktilnega tipa preloma, čeprav je opaziti tudi cepilne oziroma kvazicepilne ploskvice, a v manjšem obsegu. Podobna je mikromorfologija preloma osrednjega, naglo zlomljenega dela preizkušancev s popuščeno martenzitno-bainitno mikrostrukturo, kot tudi preizkušancev s popuščeno martenzitno mikrostrukturo ter mejo plastičnosti 1290 Nmm"2, čeprav je v slednjem primeru število kvazicepilnih ploskvic povečano. Pojavljanje cepilnega oziroma kvazicepilnega tipa preloma v coni počasnega napredovanja mikrorazpok je le sporadično, v nasprotju s prevladujočo duktilno obliko preloma, zato sklepamo, da je nukleacija mikro- Sllka 10 Področje naglega loma iz slike 7. Duktilno, cepilno in kvazi-ce-pilno. Fig. 10 Region of fast fracture from Fig. 7. Ductile, cleavage and quasi-cleavage. martensitic-bainitic microstructure, vvhile a noticeable amount of cleavage or quasi-cleavage facets appear in the samples vvith the martensitic microstructure and the yield strength 1290 N mm"2 (Fig. 9). The area of fast fracture, already shovvn in Fig. 7, is shovvn again in Fig. 10 at a higher magnification. Here, the ductile type of fracture prevails though cleavage or quasi-cleavage facets can also be observed but in small-er extent. A similar micromorphology of the areas of fast fracture is also observed on the samples vvith the tempered martensitic-bainitic microstructure as vvell as on the samples vvith the tempered martenistic microstructure and vvith yield strength 1290 N mm-2, though in the lat-ter čase the number of quasi-cleavage facets is larger. Cleavage or quasi-cleavage type of fracture in the area of slovv crack propagation is merely sporadic in comparison to the prevailing ductile type of the fracture, thus the conclusion can be made that the crack nucieation as vvell as the slovv crack propagation are mainly strain induced processes related to the decrease of fracture ductility i. e. decrease of the microplasticity in the area of stress induced segregation of hydrogen at the crack tip. 4 CONCLUSIONS An appropriate method vvas developed for the detec-tion of the nucieation and the propagation of micro-cracks at the notch tip of hydrogen charged static loaded cylindrical round notched tensile specimens. The limit of the detectability of microcrak propagation vvas about 0.1 (im. Measurements of the threshold stress intensity factor Kth of the hydrogen charged chromium-molybdenum Č.4732 steel vvith the tempered martensitic microstruc- razpok ter njih počasno napredovanje deformacijsko induciran proces povezan s poslabšanjem lomne duktil-nosti, t. j. poslabšanjem mikroplastičnosti v področju napetostno induciranega segregiranja vodika ob konici razpoke. 4. ZAKLJUČKI V okviru opravljenega dela je bila razvita primerna metoda za preučevanje nastajanja ter napredovanja mi-krorazpok iz korena zareze na obodu navodičenih ter statično obremenjenih cilindričnih nateznih preizkušan-cev z zarezo. Najmanjši korak propagacije mikroraz-pok, ki ga je bilo moč zaslediti, je znašal okoli 0,1 ^m. Merjenja mejnega napetostnega intenzitetnega faktorja Kth navodičenega krom-molibdenskega jekla, vrste Č.4732, z mikrostrukturo popuščenega martenzita oziroma popuščeno martenzitno-bainitno mikrostrukturo enake meje plastičnosti, kažejo, da majhne mikro-strukturne variacije preiskovanega jekla ne vplivajo na k-th- Ti rezultati se ujemajo z ugotovitvami Nakasata in Terasakija32, ki potrjujeta, da mejni napetostni intenzi-tetni faktor KIscc pri enaki trdnosti jekla ni odvisen od mikrostrukturnih variacij visokotrdnega jekla. Če je ta ugotovitev splošna, potem je utemeljena hipoteza, po kateri je nukleacija mikrorazpok, ki povzroče zapozneli lom jekla, vedno omejena na martenzitne dele mikrostrukture (najprej dosežena [H]cr). Le na ta način namreč lahko razložimo, da majhni deleži bainita v pretežno martenzitni mikrostrukturi popuščenega visokotrdnega jekla nimajo vpliva na mejni napetostni intenzite-tni faktor. Zdi se, da se različna jekla pri enaki vsebnosti vodika ter enaki meji plastičnosti v pogledu nuklea-cije mikrorazpok obnašajo kot elastični kontitfuum. Merjenja kritičnega napetostnega intenzitetnega faktorja kažejo, da majhne koncentracije vodika v preiskovanem jeklu nimajo opaznejšega vpliva na lomno žila-vost jekla. Pač pa je pri enaki meji plastičnosti lomna žilavost (v nasprotju z mejnim napetostnim intenzite-tnim faktorjem) odvisna tudi od majhnih mikrostrukturnih variacij jekla, saj ima jeklo s popuščeno martenzitno-bainitno mikrostrukturo nekoliko višjo lomno žilavost, kot isto jeklo s popuščeno martenzitno mikrostrukturo enake meje plastičnosti. Rezultati teh preiskav se ujemajo s podatki, ki so jih objavili Ohtani, Terasaki in Kunitake3 , ki so povečano žilavost duplex mikrostrukture razlagali s koristno vlogo majhnih deležev bainita pri zmanjšanju delov posameznih avstenitnih zrn, ki se transformirajo v martenzit. Takšna mikrostruktura ima povečano odpornost proti napredovanju razpok, t. j. manjšo občutljivost k zapoznelemu lomu. Porajanje mikrorazpok v področju maksimalnih deformacij, kot tudi pretežno duktilna oblika preloma v coni počasnega napredovanja mikrorazpok navajata k sklepu, da je nukleacija mikrorazpok deformacijsko induciran proces, povezan s poslabšanjem lomne duktil-nosti med trajanjem obremenjevanja. ture or the tempered martensitic-bainitic microstructure and with the same yield strength shovved that small microstructure variations of the investigated steel had no influence on Kth. These results confirm the Nakasato's and Terasaki's statements32 according to vvhich the threshold stress in-tensity factor K|SCC at the same tensile strength does not depend on the microstructure variations of high strength steel. If this statement is general, the hypothesis sug-gesting that the microcrack nucleation leading to the de-layed fracture is always confined to martensitic areas of the microstructure ([H]cr is at first reached) has argument. It seems that this is the only way to explain the lack of influence of small portion of bainite in a predomi-nantly martensitic microstructure of the tempered high strength steel on the threshold stress intensity factor. As far as the crack nucleation is concerned, it seems that various steels vvith the same yield strength and the same hydrogen concentration behave as an elastic con-tinuum. The measurements of the critical stress intensity factor show that small concentrations of hydrogen in the investigated steel has no noticeable influence on the fracture toughness of steel. Hovvever, at the same yield strength the fracture toughness (contrary to the threshold stress intensity factor) depends also on small microstructure variations of steel, since steel vvith the tempered martensitic-bainitic microstructure has slightly higher fracture toughness than the same steel vvith the tempered martensitic microstructure and vvith the same yield strength. The results of this investigation agree vvith the data published by Ohtani, Terasaki and Kunitake33 who explained the higher toughness of the duplex microstructure by the beneficial effect of small quantity of bainite vvhich reduces the size of single austenite-grain parts in vvhich the martensitic transformation takes plače. Such a microstructure has a better resistance to crack propagation i. e. it is less sensitive to the delayed fracture. The nucleation of the crack in the region of maximal strain as well as the predominantly ductile type of fracture in that area suggest the conclusion that the nucleation of microcracks is a strain induced process related to the decrease of fracture ductility during the loading. LITERATURA/REFERENCES 1. G. L. Hanna, A. R. Troiano in E. A. Steigerwald: Transacti-ons of the ASM, 57, 1964, 658-671 2. A. R. Troiano: Transactions of the ASM, 52, 1960, 54 3. L. S. Darken in R. P. Smith: Corrosion 5,1949,1 4. R. A. Oriani: Acta Metali. 18, 1970,147 5. V. Sakamoto in J. Eguchi: Proc. Japan Congress on Materials Research 19,1976, 91 6. A. P. Miodovvnik: Stress corrosion cracking and hydrogen embrittlement of-lron base Aloys, NACE, Huston, 1977 7. A. Zielinski, B. Lunarska in M. Smialovvski: Acta Metali. 25, 1977, 551 8. A. J. Kumnick in H. H. Johnson: Acta Metali. 28, 1980, 33 9. A. M. Adair in R. E. Hook: Acta Metali. 10,1962, 741 10. A. J. Kumnick in H. H. Johnson: Metallurgical Transactions 5A, 1974, 1199 11. G. M. Evans in E. C. Rollason: Japan Iron Steel Inst., 1969, 1484 12. D. M. Allen-Booth in J. Hevvitt: Acta Metali. 22, 1974,'171 13. H. Hargi, V. Hayashi in L. L. Shreir: Corrosion Science 11, 1971,25 14. G. W. Hong in J. Y. Lee: J. Mat. Sci. 18, 1983, 271 15. T. Asaoka, G. Lapasset in M. Aucouturier: Corrosion 34, 1978, 39 16. G. M. Pressouyre in I. M. Bernstein: Metallurgical Transactions 9A, 1978, 1571 17. A. McNabb in P. K. Foster: Trans. Am. Inst. Min. Engrs. 227, 1963, 618 18. W. Tyson: Canadian Metallurgical Quarterly, Vol. 18, 1979, 1-11 19. J. C. M. Li, L. S. Darken in R. A. Oriani: Zeitschrift fur Physt-kalische Chemie Neue Folge 9,1966, 271 20. H. L. Ewalds in R. J. H. Wanhill: Fracture mechanics, Edvvard Arnold Ltd., 1985 21. povzeto po A. G. Guy: Essentials of Materials Science, McGraw-Hill, 1976 22. W. W. Gerberich: Effect of hydrogen on high-strength and martensitic steels, Hydrogen in Metals — Proceedings of an international conference, 23,—27. September 1973, Seven Springs Conference Center, Champion, Pa. USA, Library of Congress Catalog Card Number: 73— 86455 23. R. A. Oriani: Ber. der Buns. Gesell. 76, 1972, 848 24. C. D. Beachem: Metallurgical Transactions 3, 1972, 437 25. K. Farrell in A. G. Ouarrell: J. Iron Steel Inst., 202, 1964, 1002 26. C. D. Kim in A. W. Loginovv: Corrosion 24,1968, 313 27. K. Heckel: Einfiihrung in die technische Anvvendung der Bruchmechanik, Carl Hanser Verlag, Munchen 1970 28. B. Moran in D. M. Noriš: Metallurgical Transactions A, 1978, 1685 29. L. S. Darken in R. P. Smith: Corrosion 5, 1949, 60 30. B. Ule: Zapozneli lom zaradi vodika v jeklu, Magistrsko delo, Univerza E. Kardelja v Ljubljani, 1987 31. S. T. Rolfe in S. R. Novak: Slow-bend K,c testing of medium-strength high-toughness steels, STP 463, 1970, 124—159 kot tudi R. B. Scarlin in M. Shakeshaft: Metals Technology, Jan. 1981, 1—9 32. F. Nakasato in E. Terasaki: Transactions ISIJ, 15, 1975, 290—291 33. H. Ohtani, F. Terasaki in T. Kunitake: Transactions ISIJ, 12, 1972,185