Faktor mejne intenzitete napetosti pri počasnem natezanju navodičenega jekla z visoko trdnostjo
Threshold Stress /ntensity Factor at Siow-Strain-Rate Tension of High-Strength Hydrogen-Charged Steel
B. Ule1, F. Vodopivec1, L. Vehovar1 and L. Kosec2
Konstrukcijska jekla z visoko trdnostjo in visoko napetostjo tečenja se vedno bolj uporabljajo celo za izdelavo manj zahtevnih strojnih delov. Zaradi razmeroma nizke žilavosti tovrstnih jekel in slabo izraženega prehoda v krhko stanje postajajo toliko pomembnejše njihove lomne značilnosti.
Na izgubo lomne duktilnosti močno vpliva zlasti vodik v jeklu, čeprav pri tem ne učinkuje bistveno na napetost tečenja. Poslabšanje lomne duktilnosti pa je izrazito le pri počasnem natezanju navodičenega jekla, medtem ko ga pri konvencionainem nateznem preizkusu skoraj ne zaznamo. Mate koncentracije vodika v jeklu z visoko trdnostjo torej ne vplivajo na lomno žilavost takšnega jekla, peč pa imajo za posledico pojavljanje faktorja mejh-ne intenzitete napetosti.
1. UVOD
Ena od znanih oblik porušitve jekel z visoko trdnostjo je tako imenovani zapozneli lom statično obremenje-
Slika 1:
Tipičen zapis pojava zapoznelega loma na nateznem preizkuš-ancu z zarezo, obremenjenem s konstantno obremenitvijo. Zapis velja za vodičeno jeklo, vrisana pa je tudi trdnost ob zarezi
za jeklo brez vodika (preizkušano na zraku) /lit. (1)/. r .	F'9- I-'
iypical delayed-failure phenomenon for hydrogen-charged
notched tensile specimens at constant load. The notch tensile strength of uncharged steel (measured in air) is also shovvn (Ref. i).
Inštitut za kovinske materiale in tehnologije, Lepi pot 11, 61000 Ljubljana
~ Fakulteta za naravoslovje in tehnologijo, Oddelek za montani-stiko, 61000 Ljubljana
The use of structural steels vvith high-tensile and high-yieid strength is increasing even in the manufacture of less demanding machine parts. Because of their relativen lovv toughness and poorly expressed transition into brittle state, their fracture properties are of major im-portance.
The decrease in fracture ductility is in particular strongly influenced by hydrogen content in steel, al-though hydrogen does not essentially affect its yieid strength. Hovvever, the deterioration of fracture ductility is distinctive only at slovv-strainrate tension of hydrogen-charged steel, vvhereas it practically cannot be detected in a conventional tensile test. Consequently, the lovv concentration of hydrogen in high-strength steel does not influence its fracture toughness, but results in the appearance of a threshold stress intensity factor.
1. INTRODUCTION
Deiayed fracture caused by stress-induced hydrogen segregation is one of the knovvn types of fracture of high-strength steel. This problem is characterised by the nucieation of a microcrack, vvhich then grovvs until it achieves a critical size, resulting in an abrupt fracture (Fig. 1). The incubation period, as well as the delayed tirne to failure occurrence, are prolonged vvith the decrease in load until, at a sufficiently lovv load, the delayed failure does not occur. Therefore, we can speak of threshold of appiied stress or threshold stress intensity factor Kth, vvhich can be considerably lovver than the critical stress intensity factor or fracture toughness K/c of steel. In the čase of hydrogen embrittlement the threshold stress intensity factor is iikevvise denoted as KHE.
The threshold stress intensity factor is inversely pro-portionai to the hydrogen concentration in steel, vvhich leads to the idea that the mutual effect of hydrogen and applied stress provokes the nucieation of microcracks. It vvas also found that the incubation period strongly de-pends on the hydrogen concentration in steel, vvhiie the effect of the appiied stress magnitude is coniderably lovver.
Since the effect of hydrogen on the mechanicai properties of high-strength steel is manifested by the de-creased fracture ductility at slovv-strain-rate tension and since such decrease depends on crack nucieation as well as on crack propagation, it is therefore logical that a slovv-strain-rate tension test vvill by ali means prove to
' Institute of meta/s and Technology, Lepi pot 11, 61000 Ljubljana
" University of Ljubljana, Fac. of natural Science, Metal, dept., 61000 Ljubljana
nega jekla, ki je posledica napetostno induciranega se-gregiranja vodika v jeklu. Pri tem pojavu pride najprej do iniciiranja prve mikrorazpoke, ki nato počasi raste, vse dokler ne doseže kritične velikosti, kar povzroči hipno porušitev (si. 1)1. Inkubacijski čas kot tudi čas do loma se podaljšujeta z zniževanjem statično delujoče obremenitve, vse dokler pri neki dovolj nizki obremenitvi zapozneli lom izostane. Govorimo torej lahko o pragu delujoče napetosti oziroma o faktorju mejne intenzitete napetosti Kth (threshold stress intensity factor), ki je lahko tudi občutno nižji od faktorja kritične intenzitete napetosti, to je od lomne žilavosti jekla K|C. V razmerah vodikove krhkosti označimo faktor mejne intenzitete napetosti tudi kot Khe
Z znižanjem koncentracije vodika v jeklu se faktor mejne intenzitete napetosti zvišuje, kar napeljuje na misel, da je porajanje mikrorazpok posledica vzajemnega učinkovanja vodika in delujoče napetosti. Ugotovili so tudi2, da je inkubacijski čas zelo odvisen od koncentracije vodika v jeklu, le malo pa od velikosti delujoče napetosti.
Ker se učinek vodika na mehanske lastnosti visoko-trdnega jekla manifestira z izgubo lomne duktilnosti pri počasnem natezanju in ker je lomna duktilnost jekla odvisna tako od porajanja kot tudi napredovanja razpok, je logično, da bo počasni natezni preizkus vsekakor primernejši za ugotavljanje vpliva vodika na lastnosti jekla kot pa konvencionalni natezni preiskus. Če je hitrost deformacije pri natezanju tako velika, da pride do loma v času, ki je krajši od inkubacijskega časa, učinek vodika na lastnosti jekla ne bo zaznaven.
Poleg statičnih preizkusov s konstantno obremenitvijo (static delayed failure test) se za določevanje občutljivosti jekla za lom, induciran z vodikom, še največ uporablja natezni preizkus s cilindričnimi preizkušanci z zarezo po obodu. O tem priča opis Pollockove metode v reviji Metals Progress3. Pollock določa občutljivost jekla za lom, induciran z vodikom, z merjenjem sile loma cilindričnih preizkušancev z zarezo po obodu pri hitrosti na-teznaja 2x 10-4 mm s~1. Trdnost zarezanega preizkušan-ca je namreč v takšnih primerih manjša od trdnosti gladkega, kar neposredno odseva izgubo duktilnosti zaradi učinkovanja vodika.
V strokovni literaturi je opisanih še več različnih načinov kvalitativnega določevanja občutljivosti jekla za lom, induciran z vodikom, ki pa vsi temeljijo v glavnem na enostavnem merjenju stopnje poslabšanja kontrakcije pri natezanju jekla4 5.
Ker določevanje faktorja mejne intenzitete napetosti KTH le na osnovi rezultatov nateznega preizkusa v literaturi še ni ustrezno obdelano, smo raziskali rešitev tega problema. Izkoristili smo merljivo poslabšanje lomne duktilnosti jekla pri počasnem natezanju kot nadomestku za dolgotrajni statični natezni preizkus pri konstantni obremenitvi. Ob tem smo upoštevali hipotezo, po kateri poslabšanje lomne duktilnosti pri počasnem natezanju dejansko potrjuje obstoj faktorja mejne intenzitete napetosti, če se lomna žilavost takšnega jekla - merjena pri običajnih hitrostih natezanja - le malo ali pa sploh ne spremeni.
Razvoj in teoretična utemeljitev te metode omogočata določanje faktorja mejne intenzitete napetosti kar na osnovi rezultatov nateznega preizkusa, s tem pa je določevanje občutljivosti jekla za lom, induciran z vodikom, bistveno bolj objektivno, kot je pri sedaj uporabljanih metodah.
2. TEORETIČNI DEL
Vodik je v železu v atomarni obliki bodisi na intersti-
cijskih mrežnih mestih bodisi vezan v večji ali manjši me-
te more convenient in determining the influence of hy-drogen on steel properties than a conventionai tension test. If the deformation rate at tension is so iarge that failure occurs in a period shorter than the incubation one. the influence of hydrogen on the properties of steel will not be cognizable.
Besides static test at constant load (static delayed failure test), the tension test on cylindrical specimens vvith a circumferential notch is more frequently used for the determination of hydrogen induced fracture of steel. This is evident in the description of Pollock s method in Metals Progress3. Pollock determines the sensitivity of steel to hydrogen-induced fracture by measuring the fracture load at a crosshead speed of 2x 10~* mm s'1 using cylindrical notched tensiie specimens. In this čase the strength of notched specimens is tovver than that of smooth specimens. vvhich directly reflects the decrease of ductility due to the effect of hydrogen.
Many more methods for qualitative determination of hydrogen induced fracture of steel have been described in professional literature. Hovvever. aH of these are main-ly based on the measurement of the decrease of reduction of area at tension test of steel4 5. Since the determination of threshold stress intensity factor KTH only on the basis of the results of a tensiie test has not been ade-quateiy treated in literature, our attempts vvere aimed at finding a solution to this problem.
Instead of a long-term static test at constant load, we used the siovv-strain-rate tension test for determining the measurabie decrease in fracture ductility of steel. We follovved the hypothesis that the decrease in fracture ductility at siovv-strain-rate tension test actually confirms the existence of a threshold stress intensity factor if the fracture toughness of such steel - measured at conventionai strain rate - changes only slightly or doesn 't change at ali.
The development and theoretical justification of this method prove its usefuiness in determining the threshold stress intensity factor on the basis of results at-tained in a tensiie test. In this way, the determination of steel sensitivity to hydrogeninduced fracture is essen-tiaily more objective than in currently used methods.
2. THEORY
Hydrogen is present in iron either at interstitial sites at the lattice or bound as trapped hydrogen to different discontinuities of the cristall lattice - "traps" - and thus referred to as trapped hydrogen. Some hydrogen in moleč ular form is always found in the microvoids as vvell. The partial mola/ volume of hydrogen in iron as vveli as in most other metals is surprisingly high (appr. 2 cm3/mol of hydrogen or 0,33 nm3/atom)6-8. This results in a strong interaction betvveen hydrogen intersticials in the cristai lattice and elastic-stress fields of the loaded metal lattice.
A thermodynamic analysis of this process, based on the assumption that hydrogen is a completely mobile component, vvas performed by Li, Oriani and Darken9, who found the follovving relation:
u — /io= ct/j E;j dV	(1)
The distortion fieid around the hydrogen atom is described by the deformation tensor E:j; is the stress tensor vvhich determines the stress state originating from externai loads acting on cristai lattice. Ho is the chemical potential of hydrogen in the absence of exter-nal stress, vvhile u represents the chemical potential of hydrogen under externai stress. The difference betvveen
ri na različne diskontinuitete kristalne mreže, ki jih imenujemo s skupnim imenom pasti in od tod v pasteh ujeti vodik (trapped hydrogen). Nekaj vodika je v železu vedno tudi v porah v molekularni obliki. Parcialni molski volumen vodika v železu in večini drugih kovin je presenetljivo velik (približno 2 cm3/mol vodika, oziroma 0,33 nm^atom)6-8. Posledica tega je močna interakcija med vodikovimi intersticijami v kristalni mreži ter polji elastičnih napetosti v obremenjeni kristalni mreži kovin.
Li, Oriani in Darken9, so s termodinamično analizo tega problema, pri čemer so vodik v železu obravnavali kot povsem mobilno komponento, prišli do izraza:
u—H0=a,, Ejj d V
(1)
Deformacijski tenzor Ey opisuje deformacijsko polje okrog intersticijskega atoma vodika, nc^ je napetostni tenzor, ki opredeljuje napetostno stanje, izvirajoče od zunanje mehanske obremenitve kristalne mreže. Z |i0 je v zgornjem izrazu (1) označen kemijski potencial vodika v neobremenjeni kristalni mreži kovine, ji pa je kemijski potencial vodika v mehansko obremenjeni mreži kovine. Razlika potencialov je zato enaka delu, ki je potrebno za vgnezdenje intersticijskega vodika v polje delujočih napetosti.
Gradient napetosti torej povzroči gradient kemijskega potenciala vodika, le-ta pa predstavlja gonilno silo za difuzijo intersticijsko raztopljenega vodika. Rezultat tega je segregiranje vodika v neenakomernem polju napetosti: vodik se zaradi reverzibilne dilatacije kristalne mreže s pripadajočo pozitivno spremembo volumna, ki spremlja vgnezdenje vodikovih intersticij, koncentrira v področjih prevladujočih nateznih napetosti, medtem ko se področja s prevladujočimi tlačnimi napetostmi z vodikom osiromašijo. Prerazporejanje vodika v obremenjeni kristalni mreži poteka toliko časa, dokler ni dosežena v vseh točkah mreže ravtežna koncentracija vodika, določena z izrazom:
a E dV [H] = [H]0 exp ^uJaHjl
n I
(2)
pri čemer je [H]a koncentracija enakomerno porazdeljenega vodika v neobremenjeni kristalni mreži.
Če upoštevamo le volumsko spremembo v okolici vrinjenih vodikovih atomov, lahko izraz (2) zapišemo v obliki:
[H] = [H]0 exp ■
RT
(3)
kjer je z am označena hidrostatična komponenta napetostnega tenzorja [am=1/3 (ax + ay + arz), VH pa je parcialni molski volumen vodika v železu.
Z enačbo (3) je mogoče izračunati koncentracijo vodika v lokaliziranem področju, na primer v zoženem vratu nateznega preiskušanca, kjer deluje hidrostatična napetost am. Ko koncentracija vodika [H] na tem mestu doseže kritično vrednost [H]cr, ko je torej K| = KTH, moramo računati z iniciiranjem mikrorazpok in zapoznelim lomom jekla. Problem je analitično rešil Gerberich10, ki je za faktor mejne intenzitete napetosti izpeljal izraz:
Kth =
M m M
[H]0
(4)
aVH	2 a
pri tem ima a eksperimentalno ugotovljeno vrednost 2/5 mm"1'2.
Odvisnost (4) je eksperimentalno dobro potrjena, vendar pa pri napetostih tečenja, ki so nižje od 1200 MPa, pogosto prihaja do neujemanja med enačbo (4) in rezultati eksperimentov. To neujemanje lahko deloma razložimo z odvisnostjo razmerja [H]cr/[H]0 od napetosti tečenja jekla. Farrell and Ouarrell4 sta namreč ugo-
the potentials is the vvork needed to ptace the hydrogen into the active stress field.
The gradient of chemical potential of hydrogen is therefore caused by the stress gradient and represents the driving force for the diffusion of interstitially dissolved hydrogen, resulting in hydrogen segregation in non-uniform stress field. Hydrogen concentrates in the areas of predominantly tensiie stresses due to the rever-sibie diiatations of the cristal tattice vvith the correspond-ing volume changes accompanied by the insertion of hy-drogen interstitiais, vvhile the compressiveiy strained re-gions become impoverished vvith hydrogen. The redis-tribution of hydrogen in the strained cristal iattice takes plače until an equilibrium concentration of hydrogen is achieved in aH points of the cristal Iattice. This is expressed by:
[H]=[H]0expa«EJdV HT
(2)
vvhere [H]0 is the concentration of hydrogen, uniformly distributed vvithin the unstrained cristal Iattice.
If only the volume change around the inserted hy-drogen atoms is considered, the equation (2) may be ex-pressed as:
[H]=[H]0exp
dVH RT
(3)
vvhere om is the hydrostatic component of stress tensor om= 1/3 (ax+ ay+ oz) and VH is the partiai molal volume of hydrogen in iron.
Equation (3) may be used to calculate the concentration of hydrogen in a localised area, as for example in the narrovved neck of tensiie specimens vvith hydrostatic stress <7m.
Microcracks nucieation and delayed fracture of steel can be expected vvhen the hydrogen concentration [H] in this region achieves the critical value [H]cr, i.e. vvhen K,= Kth. This problem vvas so/ved analyticaily by Gerberich10, who expressed the threshold stress intensity factor in the form of:
KTH=*JLln[±!lcr. a VH [H]0
2 a
(4)
vvhere factor a reaches the experimentally determined value of 2/5 mm~1/2.
The reiation (4) is experimentally vvell confirmed, ai-though some discrepancies can often be observed betvveen Eq. (4) and the experimental resuits at yield strength belovv 1200 MPa. These discrepancies can be partly explained by the dependance of the [H]cr/[H]0 ratio on the yield point. Namely Farrell and Ouarrell4 ascer-tained that larger concentrations of hydrogen are needed to produce embrittlement in steel vvith Iovver yield strength, vvhich they expressed vvith the reiation [H]cr °o 1/ays.
Kim and Loginow11 proved that the content of solu-ble hydrogen in steel vvas proportionai to the yield strength, thus [H]0°o cry5. If both statements are taken into account this can be vvritten as:
(5)
[HL_ P
[H]o Pys
vvhere p is a constant for specific types of steel vvith determined hydrogen concentration.
By substituting (5) for (4), we arrive at the vvell-knovvn Gerberich eguation for threshold stress intensity factor in its final form:
Kt,
Min-L■
aVH avs
2a
(6)
tovila, da so ze doseganje krhkosti v jeklih z nižjo napetostjo tečenja potrebne višje koncentracije vodika, kar sta zapisala kot [H]croo1/ays. Kim and Loginovv" pa sta dokazala, da se v jeklih z višjo napetostjo tečenja topi več vodika, torej [H]0ooays. Z upoštevanjem obeh navedenih ugotovitev lahko zapišemo:
[H]Cr_ P [H]0 Pys
pri čemer je p konstanta za posamično vrsto jekla in za določeno vsebnost vodika v njem.
Ko substituiramo (5) v (4), dobimo znano Gerberic-hovo enačbo za faktor mejne intenzitete napetosti v njeni končni obliki:
(5)

aV„
2a
(6)
/0,05 E, n2 E crvs K|C= |/-o-(MPa m12)
(7)
P = CTysexp
aVH
f
0,05 E, n2 E Gy,
2a
(10
Since hydrogen in steel mostly affects the fracture ductility at slovv-strain-rate tension, it would therefore seem adequate for further analysis to determine the re-lation betvveen fracture toughness Klc and the parameters of tensiie test. Such a relation, knovvn as the Hahn-Rosenfield correlation12 13, is given by:
K,
0,05 E, rf Eov.
(MPa m1/2)
(7)
vvhere E, is the fracture ductiiity, calculated from the ac-tual reduction of area Z using the eguation:
Ef= in [1/(1 -Z)]
(8)
vvhereas the strain hardening exponent n can be calculated from the uniform elongation eu using the eguation:
Ker vodik v jeklu še najbolj vpliva na lomno duktilnost jekla pri počasnem natezanju, je za nadaljnjo teoretično analizo smiselno poiskati soodvisnost med lomno žila-vostjo Klc in parametri nateznega preizkusa. Takšno soodvisnost poznamo pod imenom Hahn-Rosenfieldova korelacija1213, ki ima naslednjo obliko:
n= In (1+ ej
(9)
Pri tem je E, lomna duktilnost, ki jo izračunamo iz znane kontrakcije jekla Z po formuli:
Ef = In [1/(1—Z)]	(8)
medtem, ko eksponent deformacijskega utrjevanja n izračunamo iz enakomernega raztezka eu po formuli:
n = ln(1+ej	(9)
Ker pri običajnih hitrostih obremenjevanja, kakršne uporabljamo pri merjenju faktorja kritične intenzitete napetosti KIC, ne zaznamo opaznejšega poslabšanja duktilnost) jekla, ki bi ga sicer lahko pripisali vplivu majhnih koncentracij vodika v jeklu (okoli 1 ppm)4 14, se zdi utemeljena hipoteza, da poslabšanje lomne duktilnosti jekla pri počasnem natezanju dejansko odraža eksistenco faktorja mejne intenzitete napetosti KTH.
V skladu s to hipotezo bi lahko Hahn-Rosenfieldovo korelacijo (7) uporabili kar za izračunavanje faktorja KTH, potem ko bi v enačbo (7) vstavili vrednosti, izmerjene pri počasnem natezanju. Ker pa je poznana tudi teoretično izpeljana Gerberichova enačba za KTH (6), v kateri je neznana le vrednost p, je mogoče po izenačenju enačb (6) in (7) vrednost p izraziti eksplicitno:
Since the evident vvorsening of fracture ductiiity, vvhich may be attributed to the small amounts of hydrog-en in steel (appr. 1 ppm), cannot be detected4 14 by con-ventional strain-rate used in measurements of critical stress intensity factor KIC, it therefore seems that the hy-pothesis according to vvhich the decreased fracture ductility at siovv-strain-rate tension reflects the exis-tence of threshold stress intensity factor is justified.
In accordance vvith this hypothesis. the Hahn-Rosen-field correlation (7) could be used for the calculation of Kth after the vaiues measured at slovv-strain-rate have been inserted into equation (7). Since Gerberich s theor-etically developed equation for KTH (6) is also knovvn (P being the only unknovvn value). it is therefore possible to express the value of p explicitly after the balance of Eq. (6) and (7):
P= cTyS exp
oVh R T
f-t
05 E, n2 Ea,
is
Oj
2 a
(10)
RT1'	3
V izrazu (10) so veljavne vrednosti za ays, E, ter n, kot že rečeno, izmerjene pri počasnem nateznem preizkusu.
Verificiranje postavljene hipoteze se bo torej reduciralo na ugotavljanje konstantnosti veličine p, ki mora biti neodvisna od napetosti tečenja jekla cys. Konstantna vrednost p pomeni, da je postavljena hipoteza pravilna in da s podatki počasnega nateznega preizkusa lahko izračunamo Kth kar z enačbo (7).
3. EKSPERIMENTALNI DEL
Za eksperimentalno delo smo izbrali jeklo Č.4751 z naslednjo kemijsko sestavo: 0,38% C, 0,99% Si, 0,38% Mn, 0,012% P, 0,010% S, 5,19% Cr, 1,17% Mo ter 0,23% V. Po homogenizacijskem žarjenju in normalizaciji smo iz kovanih palic s premerom 16 mm za natezni preizkus
/4s already mentioned, the relevant vaiues of oys, E, and n in Eq. (10) are measured in a slovv-strain-rate tension test.
The verification of the postulated hypothesis vvill thus be reduced to the measurement of constancy of the p value. vvhich has to be independent of the yield stress oys of steel. The constant p value means that the hypo-thesis is correct and that KTH can be calculated using the Eq. (7) on the basis of siovv-strain-rate tensiie test data.
3. EXPERIMENTAL
Steel Č.4751 containing (wt-%) 0.38% C, 0,99% Si, 0.38% Mn. 0,012% P, 0.010% S. 5.19% Cr, 1,17% Mo and 0,23% V has been chosen for experimental vvork. Cylin-drical tensiie specimens vvith a diameter of 10 mm, gauge length of 100 mm and total length of 250 mm vvere machined from the forged rod after it had been homoneously anneaied and normalized. Specimens vvere thermally treated in a vacuum annealing furnace. They were austenitised at 980° C for a short period. quenched in a flovv of gaseous nitrogen and then tempered at temperatures of 620° C, 640° C and 670° C re-spectively. Thus. three separate and distinct classes of yield strength - 1220 MPa. 1020 MPa and 900 MPa re-spectively vvere achieved.
The cathodic charging of thermally treated tensiie specimens vvas carried out for 1 hour in 1 N sulfuric acid at a current density of 0,3 mA/cm2. The experimentai set-up for cathodic polarisation of tensiie specimens, as shovvn in Fig. 2 is composed of a potentiostat and corrosion celi vvith electrodes.
izdelali 250 mm dolge cilindrične preiskušance s premerom 10 mm in dolžino 100 mm. Preiskušance smo toplotno obdelali v vakuumski žarilni peči, tako da smo jih po kratkotrajni austenitizaciji pri 980° C kalili v toku plinastega dušika, nato pa popuščali pri temperaturah 620°C, 640° C oziroma 670° C. Na ta način smo dobili tri ločene in dobro definirane trdnostne razrede z napetostjo tečenja 1220 MPa, 1020 MPa oziroma 900 MPa.
Toplotno obdelane preiskušance za natezni preizkus smo navodičili z enournim katodnim polariziranjem v 1 N raztopini žveplene kisline pri gostoti toka 0,3 mA/cm2. Eksperimentalni sklop s katodnim polariziranjem preizkušancev, sestavljen iz potenciostata in korozijske celice z elektrodami, je prikazan na sliki 2.
Natezne preizkuse smo opravili na nateznem trgal-nem stroju INSTRON, potem ko smo natezne preizku-šance po končanem navodičenju 24 ur zadrževali na zraku, da so se koncentracije vodika v jeklu približale resi-dualnim vrednostim (približno 0,7 ppm), ki se nato časovno skoraj niso več spreminjale.
Za hitrost natezanja smo izbrali tako hitrost 1 mm/ min, značilno za običajni natezni preizkus, kot tudi hitrost 0,1 mm/min, značilno za počasno natezanje. Merili smo napetost tečenja ays (MPa), natezno trdnost ctts (MPa), maksimalni enakomerni raztezek eu ( x 100%) ter kontrakcijo jekla Z ( x 100%).
Lomno duktilnost E, in eksponent deformacijskega utrjevanja n smo izračunali z enačbama (8) in (9).
Mikrofraktografske preiskave prelomnih površin vo-dičenih in pri hitrosti natezanja 0,1 mm/min obremenje-vanih nateznih preizkušancev smo opravili s scanning elektronskim mikroskopom JEOL JSM-35 (SEM).
4. REZULTATI
Izmerjene mehanske lastnosti navodičenega jekla kot tudi jekla brez vodika (pod 0,05 ppm) so zbrane v tabeli 1. V tej tabeli so zbrane še lomne žilavosti K,c jekla, izračunane s Hahn-Rosenfieldovo korelacijo (7) na osnovi rezultatov običajnih nateznih preizkusov pri hitrosti natezanja 1 mm/min ter nadalje še faktorji mejne intenzitete napetosti KTH vodičenega jekla, izračunani z isto enačbo (7), vendar na osnovi rezultatov počasnega natezanja pri hitrosti 0,1 mm/min.
Slika 2:
Eksperimentalni sklop za vodičenje nateznih preiskušancev s
katodnim polariziranjem. (P-potenciostat, D-natezni preizkušanec, G-grafitni protielek-trodi, K-kalomelova elektroda in R-rotometer).
Fig. 2:
Experimentai set-up for hydrogen charging of tensiie specimens with cathodic poiarisation. (P-potenciostat. D-tensiie specimen. G-graphite eiectrodes. K-kaiomei eiectrode. Ft-rotameter).
The tension tests vvere made on an INSTRON testing machine. after hydrogen charging of specimens vvas compieted and the specimens exposed to air for 24 hours. This enabied the concentrations of hydrogen in steel to approach the residuai vaiues (appr. 0,7 ppm), vvhich remained near/y time-independent.
The tension tests vvere performed at conventional strain rate i.e. at a crosshead speed of 1 mm/min as well as at iovver-strain-rate i.e. at a crosshead speed of 0,1 mm/min. The yie/d strength cjys (MPa), tensiie strength aTS (MPa), max. uniform elongation eu (x100%) and the reduction ofarea Z (x100%) ivere measured. The fracture ductiiity E, and the strain hardening exponent n vvere calculated using equations (8) and (9) respectiveiy.
The fracture surfaces of tensiie specimens tested at a crosshead speed of 0,1 mm/min vvere examined in the scanning electron microscope JEOL JSM-35 (SEM).
R
Tabela 1: Mehanske lastnosti jekla brez vodika in istega jekla po navodičenju Table 1: Mechanicai properties of uncharged and hydrogen-charged steel
Hitrost natezanja 1 min/min Crosshead speed 1 mm/min
Napetost tečenja Yield strength ays(MPa)
Enakomerni raztezek Uniform elongation e„ x 100%
Kontrakcija
Reduction
ofarea Zx 100%
Lomna žilavost Facture toughness
Hitrost natezanja 0,1 mm/min Crosshead speed 0.1 mm/min
MPa. m1'
Napetost tečenja Yield strength oys(MPa)
Enakomerni raztezek Uniform elongation eux 100%
Kontrakcija
Reduction
of area Zx 100%
Faktor mejne inten. napet. Threshold stress inten. factor
MPa • m1
Konstanta Constant
/En (10)/ /Eq. (10)/
MPa
Jeklo brez vodika, uncharged steel
924 8,7 52	126,9	910	8.5	51		
1010 7,4 51,3	112,5	1027	6,5	50,3		
1270 6,4 50	107,5	1214	6,2	50,3		
Navodičeno jeklo, hydrogen charged steel						
885 8.4 50.3	117,2	899	8,1	47,7	109,8	4005
1082 7,2 49,3	110,1	1078	6,5	42,7	90,1	4223
1209 6,1 47,3	96,3	1226	6,0	27,3	67,4	4037
V tabeli 1 so prikazane tudi vrednosti za konstanto p, izračunane s pomočjo enačbe (10) na osnovi rezultatov počasnega natezanja.
Slika 3:
Diagram sila-deformacija pri počasnem natezanju jekla trdnostnega razreda 1300 MPa; a) brez vodika (pod 0,05 ppm) in b) 24 ur po vodičenju (ca. 0,7 ppm vodika).
Fig. 3:
Load-deformation diagram obtained at slow-strain-rate tension
test of steel vvith yieid strength of 1300 MPa. (a) vvithout hydrogen (iess than 0.05 ppm) and (b) 24 hours after hydrogen charging (appr. 0.7 ppm hydrogen).
V diagramu sila-deformacija na sliki 3 je prikazana odvisnost med silo in raztezkom pri počasnem natezanju jekla s trdnostjo ca. 1300 MPa. Odvisnost, označena z a), velja za jeklo brez vodika (pod 0,05 ppm), z napetostjo tečenja 1070 MPa. trdnostjo 1286 MPa, enakomernim raztezkom eu = 6% in kontrakcijo Z = 49%, medtem ko velja odvisnost, označena z b), za jeklo istega trdnostnega razreda, ki pa je bilo natezano 24 ur po vodičenju (ca. 0,7 ppm vodika). V tem primeru smo namerili napetost tečenja 1090 MPa, trdnost 1284 MPa, enakomerni raztezek eu = 5,6% ter kontrakcijo Z = 39%.
Mikrofraktografske preiskave prelomnih površin na-vodičenih nateznih preizkušancev kažejo, da z vodikom inducirani lom pri počasnem natezanju takšnih preizkušancev ne ostane povsem duktilnega tipa, celo pri preiz-kušancih z relativno nizko napetostjo tečenja ne (slika 4). Pri višji napetosti tečenja so prelomne površine navodičenega ter počasi natezanega jekla mešane narave; poleg kvazicepilnih ploskev najdemo na prelomnih površinah tudi jamičasta duktilna področja ter številne grebene, nastale s trganjem (slika 5).
5. RAZPRAVA
Analiza rezultatov mehanskih preizkusov (Tabela 1) kaže, da je lomna žilavost K,c, s katodnim polariziranjem navodičenega jekla z visoko trdnostjo, le malo manjša od lomne žilavosti enakega jekla brez vodika, kot o tem tudi sicer lahko sklepamo iz diagrama na sliki 1.
Počasnejše natezanje pri jeklu brez vodika ne povzroči kakšnih opaznejših sprememb, medtem ko se pri
Slika 4:
Jamičasta duktilna prelomna površina s posameznimi kvazice-pilnimi detajli (B) pri vodičenem in počasi natezanem jeklu z napetostjo tečenja ca. 900 MPa. Fig. 4:
Dimpied ductile fracture area vvith some quasicleavage details (B) in hydrogen-charged and slow-strain-rate tested steel vvith yield strength of 900 MPa.
Slika 5:
Mešana oblika preloma vodičenega jekla z napetostjo tečenja ca. 1070 MPa.
Poleg cepilnih oziroma kvazicepilnih ploskev (B) je na prelomni površini moč zaslediti tudi jamičasta duktilna področja ter številne grebene, nastale s trganjem.
Fig. 5:
Mixed fracture mode on hydrogen-charged steel vvith yield
strength of 1070 MPa. Besides cleavage and quasicleavage facets (B). dimpied ductile areas and many tear ridges can also be observed.
navodičenem jeklu močno poslabša lomna duktilnost, t.j. kontrakcija jekla, ne pa tudi enakomerni raztezek, trdnost in napetost tečenja takšnega jekla. Poslabšanje lomne duktilnosti navodičenega jekla pri počasnem na-tezanju dejansko kaže na obstoj faktorja mejne intenzitete napetosti KTH (KHE). S Hahn-Rosenfieldovo korelaci-jo (7) izračunane vrednosti KTH dajejo namreč, po substituciji v Gerberichovo enačbo (6), za konstanto [} vrednost približno 4100 MPa. Ta vrednost je neodvisna od napetosti tečenja jekla in zato v okviru eksperimentalne natančnosti merjenja res konstantna količina, skladno z Gerberichovim modelom.
Eksperimenti so nadalje pokazali, da je bila uporabljena hitrost natezanja 0,1 mm/min že dovolj majhna, da smo lahko iz poslabšanja lomne duktilnosti navodičenega jekla izračunali take vrednosti faktorja mejne intenzi-tete napetosti KTH. za katere je p konstanta.
Če upoštevamo, da velikost plastične cone v trenutku loma navodičenega preizkušanca dosega velikost približno polovice vratu preiskušanca (I = 3 mm), dobimo za Ec, upoštevaje hitrost natezanja v=1,6x10~3mm s-1 (0,1 mm/min), vrednost Ec = v/I = 5,3 x 10"4 s-1.
V strokovni literaturi15 navajajo za nerjavna jekla nekoliko višje vrednosti Ec, približno 10_1 s-1. Raziskave Nakana in sodelavcev16, opravljene s počasnim natezanjem vodičenega jekla z napetostjo tečenja 500 MPa, pa kažejo, da se pri zadostni koncentraciji vodika v jeklu kontrakcija jekla asimptotično približa neki znižani vrednosti že pri kritični hitrosti deformacije Ec= 10-4 s-1, to pa je že velikostni red naših izmerjenih vrednosti. Ta hitrost deformacije je namreč že dovolj majhna, da Cottrellovi oblaki vodikovih atomov lahko potujejo skupaj z disloka-cijami globoko v plastično cono nateznih preizkušancev.
Diagram na sliki 3 potrjuje, da vodik ne vpliva bistveno na mobilnost dislokacij v zgodnjih fazah deformacij-skega procesa pri natezanju, saj skoraj ne učinkuje na napetost tečenja, trdnost in enakomerni raztezek jekla, pač pa le na kontrakcijo jekla. Vodik torej spreminja obliko diagrama sila-deformacija šele od pojavljanja plastične nestabilnosti dalje, kar se dobro ujema z navedbami iz različnih literaturnih virov17-25. Po teh navedbah vodik ne vpliva niti na zgodnje nukleiranje mikropor, niti na gostoto mikropor, ko se stopnja deformacije približuje lomni deformaciji. Očitno je zato vpliv vodika zaznaven šele v fazi rasti mikropor in/ali fazi njihovega združevanja. Do pospešene rasti in koalescence mikropor v tej fazi pa lahko pride tudi z mehanizmom ločevanja prostih površin, na katerih je adsorbiran vodik26. Na sliki 6 je shemat-sko prikazana rast in koalescenca mikropor vzdolž meje dveh kristalnih zrn. Mehanizem koalescence mikropor z ločevanjem prostih površin, na katerih je adsorbiran vodik, prične delovati, ko se oblikuje troosno napetostno stanje v zoženem delu nateznega preizkušanca. Posledica tega je že opisano "zgoščevanje" zadnje faze plastične deformacije pri počasnem natezanju navodičenega jekla.
Mikrofraktografske preiskave samo še ilustrirajo pravkar opisani mehanizem loma. Pojasnjujejo namreč lome jamičaste duktilne vrste, pri katerih pa kažejo stene in dna jamic posamične mikromorfološke značilnosti cepilnega oziroma kvazicepilnega loma. Res smo tudi pri naših raziskavah vodičenega in počasi natezanega jekla višjega trdnostnega razreda opazili poleg duktilnih jami-častih področij še trganja (tearing), ki so sicer značilna za jekla z zadostno duktilnostjo in dovolj majhno napetostjo tečenja, da do porušitve lahko pride s plastično deformacijo. Poleg detajlov takšne vrste pa smo opazili na prelomnih površinah tudi področja kvazicepilne narave, često na samem obrobju večjih in globjih, lijakasto
4.	RESULTS
The mechanical properties of hydrogen-charged as vvell as hydrogen uncharged steel (/ess than 0,05ppm) are presented in Table 1 The fracture toughness of steel K,c, ca/culated according to the Hahn-Rosenfieid correlation (7) on the basis of conventional tension tests made at a crosshead speed of 1 mm/min, as vveil as the thresho/d stress intensity factor KTH of cathodic charged steel. also calculated using eguation (7), but on the basis of results obtained at a crosshead speed of 0,1 mm/ min, are also shovvn in Table 1.
The vaiues of constant ji are also given in Table 1. These vvere calculated using equation (10), on the basis of slow-strain-rate tensile test data.
The /oad-deformation diagram (Fig. 3) shovvs the re-lation betvveen load and elongation at siovv-strain-rate tension of steel vvith a tensile strength of appr. 1300 MPa. Curve a) denotes the uncharged steel (less than 0,05 ppm hydrogsn), having a yield strength of 1070 MPa, tensile strength of 1286 MPa, uniform elongation eu=6% and reduction of area Z=49%, vvhereas Curve b) denotes hydrogen-charged steel (appr. 0,7ppm hydrogen) of the same strength class. being tested 24 hours after it had been charged. In this čase the yield strength of 1090 MPa, tensile strength of 1284 MPa, uniform elongation eu = 5,6% and reduction of area Z= 39% vvere measured.
The microfractographic axaminations of fracture sur-faces of hydrogen-charged specimens confirm that hy-drogen-induced fracture at slovv-strain-rate tension does not remain predominantly ductile. even in specimens vvith a relatively lovv yield strength (Fig. 4). The fracture surfaces of hydrogen-charged steel vvith higher yield strength, stretched at slovv-strain-rate tension, are of mixed mode. In addition to quasicleavage details, ductile dimpled areas and numerous tear ridges have also been observed (Fig. 5).
5.	DISCUSSION
The anaiysis of mechanical testing data (Table 1) shovvs that the fracture toughness K/c of hydrogen-charged high-strength steel is only slightly lovver than that of the uncharged steel. vhich can also be con-cluded from the diagram in Fig. 1.
The slovv-strain-rate tension of uncharged steel does not provoke any noticeable changes, vvhereas that of hy-drogen-charged steel strongly decreases the fracture ductility. i.e. the reduction of area. but does not affect the uniform elongation, tensile strength and yield strength of such steel. The deterioration of fracture duc-tility of hydrogen-charged steel at slovv-strain-rate tension indicates the existence of the threshold stress in-tensity factor KTH (KHE). Namely, the KTH - vaiues, calculated vvith the Hahn-Rosenfieid correlation (7) after being substituted into Gerberichš equation (6), give an approximate value of 4100 MPa for the /)-constant. With-in the experimental error of the measure- ments, the obtained value is constant and, in accordance vvith Gerber-ich 's model, independent of the yield strength of steel.
The experiments further shovved that the applied crosshead speed of 0,1 mm/min vvas sufficiently low to enable the calculation of KTH - vaiues from the decrease in fracture ductility of hydrogen-charged steel, i.e. the calculation of KTH - vaiues for vvhich f) is a constant.
Considering that the size of the plastic zone of a hy-drogen- charged specimen is approximately half of the neck diameter (1=3 mm) at fracture, the crosshead speed is v= 1,6x 10~3 mm s~' (0.1 mm/min), then a value of Ec= v/l= 5,3x 10-4 s'1 is obtained.
Troosno napetostno stanje zaradi pojavljanja vratu
Slika 6:
Shematski prikaz nastajanja por, njihove rasti in koalescence vzdolž meja zrn, na katerih je adsorbiran vodik /lit. (26)/.
Fig. 6:
Schematic representation of microvoid formation. growth and coaiescence aiong grain boundaries where hydrogen is ad-sorbed (Ref. 26).
oblikovanih jamic (slika 4, detajl B), pa tudi kot povsem samostojna plitvejša področja.
6. SKLEPI
Na osnovi opravljenih raziskav smo ugotovili, da izgubo lomne duktilnosti pri počasnem natezanju navodi-čenega jekla z visoko trdnostjo lahko uspešno izkoristimo za kvantitativno določevanje faktorja mejne intenzitete napetosti KTH (KHE)- Ugotovili smo namreč, da majhne koncentracije vodika (pod 1 ppm) v jeklu, ki je bilo navo-dičeno s katodnim polariziranjem, ne vplivajo bistveno na lomno žilavost, merjeno pri običajnih hitrostih nateza-nja (1 mm/min). Počasno natezanje (0,1 mm/min) navo-dičenega jekla z visoko trdnostjo pa poslabša lomno duktilnost, kar nakazuje obstoj faktorja mejne intenzitete napetosti KTH. Z izračunanjem faktorjev KTH s pomočjo Hahn-Rosenfieldove korelacije (7) in vstavljanjem teh vrednosti v Gerberichovo enačbo (6) smo ovrednotili parameter P /enačba (10), tabela 1/. Dobili smo konstantno vrednost okoli 4100 MPa, neodvisno od napetosti tečenja jekla, kot to tudi zahteva Gerberichov model za KTH.
Mikrofraktografske preiskave prelomnih površin počasi natezanega vodičenega jekla z visoko trdnostjo ka-čejo, da je prelom lokalno še vedno tudi duktilne vrste. Kljub detajlom kvazicepilne narave smo v vseh primerih našli še duktilne grebene, nastale s trganjem, in jamiča-sta področja duktilnega tipa. Kvazicepilna oblika loma na obrobju večjih in globjih lijakasto oblikovanih jamic dokazuje, da so le-te rasle in se medsebojno zlivale tudi z mehanizmom ločevanja prostih površin, na katerih je bil adsorbiran vodik.
Professionai literature15 quotes somewhat higher Ec -values for stainless steels, approximately 10~1 s~'. How-ever, the investigations performed by Nakano and co-workersw on hydrogen-charged steel vvith yield strength of 500 MPa using siow-strain-rate measurements shovv that at sufficient concentration of hydrogen in steel the reduction of area asymptotically approaches the re-duced value already at a criticai deformation rate of Ec= 10-4 s"', vvhose magnitude is of the same order as found in our investigations. This deformation rate is, in fact, iovv enough to enable the Cottrell atmosphere of hydrogen atoms pinned on dislocations to penetrate deep into the piastic zone of tensiie specimens.
The diagram on Fig. 3 confirms that hydrogen has no essential influence on the mobility of dislocations in ear-iier phases of the deformation process at tension. It has aimost no effect on the yield strength, tensiie strength and uniform eiongation of steel, but only on the reduction of area. Thus hydrogen changes the shape of a ioad-deformation diagram only after the appearance of piastic instability, as aiready found by a number of au-thors,7~ss. According to these sources. hydrogen has no effect on the early nucieation of microvoids, nor on microvoid density, when the deformation approaches the fracture deformation. Therefore, the effect of hydrogen becomes obvious only at the stage of microvoid grovvth and/or during their coaiescence. The grovvth of microvoid and their coaiescence can also be acceierated by a mechanism of separation of internat interfaces vvhere hy-drogen is adsorbed26. The grovvth and coaiescence of microvoids along the grain boundary are schematically shovvn in Figure 6. The mechanism of microvoid coaiescence and the separation of internaI interfaces due to adsorbed hydrogen becomes operative vvhen the triaxial stress state in the narrovv neck of the tensiie specimen is formed, resulting in the previousiy described "conden-sation" of the tast stage of piastic deformation at siovv-strain-rate tension of hydrogen-charged steel.
The microfractographic investigations are an addi-tionai illustration of above-mentioned fracture mechanism. They explain the ductile-dimpled types of fracture in vvhich the vvalls and bottoms of dimpies exhibit individual micromorphological characteristics of the cleavage or quasi-cleavage type of fracture. It is true that our investigations of hydrogen-charged high-strength steel at slow-strain-rate also shovved, in addition to ductile-dim-pled areas. tearing regions typicai for steels vvith sufficientiy high ductiiity and yield strength iovv enough that the fracture may be the result of piastic deformation. Be-sides the detaiis of such ductiie types. we also observed quasicieavage areas on fracture surfaces. most often on the very periphery of iarger. deeper funnel-type dimpies (Fig. 4, detaii B) and also as entirely independent shal-low regions.
6. CONCLUSIONS
On the basis of the performed investigations. we have ascertained that the ioss of fracture ductility at slow-strain-rate tension of hydrogen-charged high-strength steel can be successfully used for the quantita-tive determination of the threshold stress intensity factor Kth (KHe>- It was estabiished that small concentrations of hydrogen (less than 1 ppm) in cathodically-charged steel have no substantial influence on the fracture toughness, as measured at conventional strain-rate (1 mm/min). Hovvever, the slow-strain-rate (0.1 mm/min) of hydrogen-charged high-strengthsteei vveakens the fracture ductility, which refiects the existence of the threshold stress intensity factor KTH. The parameter fi
/Eq. (10), Table 1/ vvas determined by Inserting the KTH -values, calculated using the Hahn-Rosenfield correlation (7), into Gerberich s equation (6). We thus obtained a constant vaiue of about 4100 MPa, independent of the yield strength of steel, as requested by Gerberich s model for KTH.
Microfractographic investigations of fracture sur-faces of highstrength hydrogen-charged steel tested at siovv-strain-rate indi- cate that. locally. the fracture is stili of ductile type. Despite the quasicleavage details, ductile ridges as a result of tearing as well as ductile dimpied areas vvere found in ali cases. The quasicleavage type of fracture on the periphery of larger. deeper and funnel-type dimpies proves that the grovvth and coales-cence of voids are also the consequence of the mechan-ism causing the separation of internal interfaces vvhere hydrogen is adsorbed.
LITERATURA / REFERENCES
1.	C.S. Kortovich in E.A. Steigervvald, Eng. Fract. Mech., 4, 637 (1972).
2.	G.L. Hanna. A.R. Troiano in E.A. Steigervvald, Trans, of the ASM, 57.658-671 (1964).
3.	Nevv hydrogen-embrittlement test. Advanced Materials and Processes inc., Metal Progress, 7, 10-11 (1988).
4.	K. Farrell in A.G. Ouarrell, J. Iron Steel Inst., 202, 1002 (1964).
5.	H. Morimoto in Y. Ashida. Trans. ISIJ, 23. B-352 (1983),
6.	H. VVagenblast in H.A. VVriedt, Metali. Trans., 2, 1393-1397 (1971).
7.	J.OM. Bockris, W. Beck, M.A. Genshavv. P.K. Subramanyan in F.S. VVilliams. Acta metali,, 19, 1209-1218 (1971).
8.	H. Peisl v Hydrogen in Metals, Topics in Applied Physics, Ed.: G. Alefeld and J. Volkl. vol. 28, 53-74 (1978).
9.	L.C.M. Li, R.A. Oriani in L.S. Darken. Z. Phzsik. Chem., 49, 271 (1966).
10.	W.W. Gerberich v Effect of Hydrogen in high-strength and martensitic steels, Hydrogen in Metals. ASM-Ohio, 115 (1974), kot tudi:
W.W. Gerberich in S.T. Chen, Metali. Trans., 6A. 271 (1975).
11.	C.D. Kim in A.W. Loginow, Corrosion, 24, 313 (1968).
12.	G.T. Hahn in A.R. Rosenfield, Sources of Fracture Toughness: The Relation betvveen KIC and the Ordinary Tensile
Properties of Metals, Applications Related Phenomena in Titanium Alloys, ASTM STP 432, 5-32, Philadelphia, (1968).
13.	G.T. Hahn in A.R. Rosenfield, Metali. Trans., 6A 653-668 (1975).
14.	G.T. Hahn in A.R. Rosenfield, Trans.. ASM, 59, 909 (1966).
15.	M.B. VVhiteman in A.R. Troiano, Corrosion, 21, 53-56 (1965).
16.	K. Nakano, M. Kanao in T. Aoki, Trans, of National Research, Institute for Metals, 29, No. 2, 34-43 (1987).
17.	R. Garber. M. Bernstein in A.W. Thompson, Ser Metali 10 341-345 (1976).
18.	A.S. Argon in J. Im, Metali. Trans., 6A, 839-851 (1975).
19.	J.R. Rice in D.M. Tracey, J. Mech. Phys. Solids 17 201-217 (1969).
20.	D.M. Tracey, Eng. Fract. Mech., 3, 301-315 (1971).
21.	A.S. Argon, J. Im in R. Safoglu, Metali. Trans., 6A, 825-838 (1975).
22.	C.D. Beachem, Metali. Trans., 3, 437-451 (1972).
23.	T.D. Lee, T. Goldenberg in J.P. Hirth, Metali. Trans 10A 199-208 (1979).
24.	In-Gyn Park in A.W. Thompson, Metali. Trans., 21A 465-477 (1990).
25.	D. Kwon in R.J. Asaro, Metali. Trans., 21 A, 117-134 (1990).
26.	H. Cialone in R.J. Asaro, Metali. Trans., 10A 367-375 (1979).