Hydrogen and Temper Embrittlement of Medium Strength Steel
Vodikova in popustna krhkost jekla srednje trdnosti
Ule B.,1 V. Leskovšek, Institute of Metals and Technology, Ljubljana
The fracture ductility of high strength steel is strongly influenced by the presence of hydrogen, although hydrogen does not significantly affect the yieid strength. The deterioration of fracture ductility is particularly evident in low strain rate tension tests, but less pronounced at conventional crosshead speeds. Microfractographic investigations of fracture surfaces of hydrogen charged high strength 5Cr-1 Mo-0.3V steel from low strain rate tension test indicate that the grovvth and the coalescence of voids in the final stages of the fracture process are partly assisted by the decohesion of interfaces on vvhich hydrogen is adsorbed. Hovvever, such phenomena are not observed in the experimental medium strength steel. Although a strong interaction betvveen hydrogen and temper embrittlement vvas frequently observed in the alloyed steels'' and though the magnitude of such effect vvas directly related to the degree of intergranular phosphorus enrichmenf, such synergy vvas not found in the experimental steel vvith post-martensitic microstructure.
Key vvords: high strength steel, medium strength steel. hydrogen and temper embrittlement. fracture ductility, lovv strain rate tension test
Na lomno duktilnost jekla z visoko trdnostjo močno vpliva v jeklu prisoten vodik, čeprav slednji nima znatnejšega učinka na napetost tečenja jekla. Poslabšanje lomne duktilnosti je zlasti očitno pri majhni hitrosti natezanja, medtem, ko je pri običajni hitrosti natezanja manj izrazito. Mikrofraktografske preiskave prelomnih površin nastalih pri počasnem natezanju vodičenega jekla 5Cr-1 Mo-0.3V z visoko trdnostjo kažejo, da je rast in zlivanje por v končnih fazah procesa loma deloma podprta z dekohezijo medplastij, na katerih je adsorbiran vodik. Tega pojava nismo opazili pri preiskovanem jeklu srednje trdnosti, čeprav je bila pri legiranih jeklih često opažena močna interakcija med vodikovo in popustno krhkostjo12 in čeprav je bila intenzivnost tega učinkovanja neposredno povezana s stopnjo interkristalne obogatitve s fosforjem,2 pa takšne sinergije nismo našli pri preiskovanem jeklu s postmartenzitno mikrostrukturo.
Ključne besede: jeklo visoke trdnosti, jeklo srednje trdnosti, vodikova in popustna krhkost, lomna duktilnost, upočasnjeno natezanje
1. Introduction
The adverse effect of hydrogen on mechanical properties has long been recognised in various metallic materials, especially in high strength steels. Hydrogen embrittlement of such steels often in-volves both a change in fracture mode and a reduc-tion in ductility compared vvith the unhydrogenated condition3. It has also been established that the ki-netics of hydrogen embrittlement depends on the strain rate4"6. Hovvever, at constant static load, the
Dr. Boris ULE IMT, Lepi pot 11 61000 Ljubljana
failure of high strength hydrogen charged steels, knovvn as delayed failure, frequently occurs. This is caused by stress induced segregation of hydrogen and is characterised by the nucleation of a microc-rack vvhich then grovvs until it reaches a critical size, resulting in an abrupt failure. The incubation period, and the tirne to failure, are prolonged vvith a decrease in load until, at a sufficiently lovv load, delayed failure does not occur. Therefore, a threshold stress inten-sity factor Kth can be introduced vvhich is consider-ably lovver than the critical stress intensity factor or fracture toughness K,c of the uncharged steel.
Our preliminary investigations confirm that lovv concentrations of hydrogen (< 1 ppm by vveight) in
high strength steel have no substantial influence on its fracture toughness measured at conventional strain rate (1 mm min 1). Hovvever, the lovv strain rate (0.1 mm min 1) of hydrogen charged high strength steel decreases the fracture ductility, vvhich indicates the existence of the threshold stress intensity factor at semi-static testing conditions6.
Because the interaction betvveen hydrogen and temper embrittlement vvas frequently observed in lovv and medium alloyed steels12, the synergism betvveen both embrittlements vvas investigated in a 5Cr-1 Mo-0.3V steel vvith post-martensitic microstructure. The aim of the present vvork is therefore not only to demonstrate the applicability of lovv strain rate tension test in order to study the hydrogen embrittlement phenomena in high strength steel, but also to study the possible interaction betvveen hydro-gen and temper embrittlement in medium strength steel.
2. Experimental
A secondary hardening steel containing 0.38 C, 0.99 Si, 0.38 Mn, 0.012 P, 0.01 S, 5.19 Cr, 1.17 Mo, and 0.23 V (ali wt-%) vvas used. Cylindrical tensile specimens vvith 10 mm dia. vvere machined from a forged rod after being homogeneously annealed and normalised. Some specimens vvere austenised at 980 C for 30 minutes in a vacuum furnace, quenched in a flovv of gaseous nitrogen at a pressure of 0.5 MPa, and then tempered at temperatures of 620, 640 or 670r C. Three separate classes of yield stress, 1220, 1020 and 900 MPa respectively, vvere achieved.
The remaining specimens vvere quenched under the same conditions, then tempered tvvice for 2 hours at 710 C vvith intermediate undercooling (yield stress: 668 - 679 MPa, Charpy V-notch impact ener-gy: 72 J), vvhereas some of these specimens vvere additionally tempered for 24 hours at 570°C, i.e. in the temperature range of reversible temper embrittlement78 (yield stress: 648 - 665 MPa, Charpy V-notch impact energy: 37 J). In such cases, the experimental steel had a post-martensitic microstructure.
The cathodic charging of tensile specimens vvas performed for 1 h in 1 N sulphuric acid at a current density of 0.3 mA cm 2. The concentration of hydro-gen in some specimens (bulk concentrations) vvas determined using a high temperature (1050 C) vacuum extraction technique vvith gas chromatograph-ic analysis. Tension tests vvere performed after the hydrogen charging of specimens had been com-pleted and the specimens had been exposed to air for 24 hours. This enabled the concentrations of hy-drogen to approach the residual value of approxi-
mately 0.8 ppm by vveight vvhich remained almost tirne independent.
The tension tests vvere performed at a conventional strain rate, i.e. at a crosshead speed of 1 mn min and at a lovver strain rate, i.e. at a crosshead speed of 0.1 mm min1.
The fracture surfaces of the tensile specimens vvere examined in a scanning electron microscope (SEM).
3. Results
3.1 Tensile Properties
The results obtained from different strain rate tension tests for both uncharged and hydrogen-charged steel are pointed out in Table I and II. The results in Table I refer to the specimens vvhich vvere quenched and tempered in a temperature range from 670 to 620° C vvith a class of yield stress of approx. 900 MPa, 1020 MPa and 1220 MPa, respectively.
Table I: Mechanical properties of uncharged and hydrogen charged steel, quenched and tempered up to high yield stresses.
Tabela I: Mehanske lastnosti nevodičenega in vodičenega jekla, ki je bilo kaljeno in popuščano na visoko napetost tečenja
Crosshead speed 1 mm min"1 Crosshead speed 0.1 mm min
Yield Uniform Reduction Yield Uniform Reduction stress elongation of area stress elongation of area MPa %	% MPa %	%
Uncharged steel
924	8.7	52	910	8.5	51
1010	7.4	51.3	1027	6.5	50.3
1270	6.4	50	1214	6.2	50.3
Hydrogen-charged steel
885	8.4	50.3	899	8.1	47.7
1082	7.2	49.3	1078	6.5	42.7
1209	6.1	47.3	1226	6.0	27.3
The decrease of the crosshead speed at tension had no influence on the mechanical properties of the uncharged steel, vvhereas it essentially influenced the reduction of area in the hydrogen charged steel (approx. 0.8 ppm hydrogen). The loss of ductility vvas more pronounced in steel vvith a higher yield stress.
The results in Table II refer to the steel vvith a post-martensitic microstructure, i.e. specimens vvhich vvere quenched and tempered tvvice at 710 C vvith intermediate undercooling, and specimens vvhich
Qa = 160.3 t 4.5 kJ m	oT<
	
	
	
11
1J2
1 .„3
y - W3,KJ
vvhere R.A^ is the reduction of area at conventional strain rate tensile test, i.e. at a crosshead speed of 1 mm min 1 and R.A.0, is the reduction of area at low strain rate tensile test, i.e. at a crosshead speed of 0.1 mm min"1.
Table II: Mechanical properties of uncharged and hydrogen charged steel quenched and tempered tvvice at 710 C and of the same steel, additionally tempered for 24 hours at 570°C
Tabela II: Mehanske lastnosti nevodičenega in vodičenega jekla, ki je bilo kaljeno in dvakrat po-puščano pri 710°C ter istega jekla, ki je bilo dodatno popuščano 24 ur pri 570 C
Crosshead speed 1 mm min1 Crosshead speed 0.1 mm min1
Figure 1: Evaluation of the activation energy of segregation of phosphorus according to Arrhenius equation (from Ref. 7) Slika 1: Izvrednotenje aktivacijske energije za segregiranje fosforja z uporabo Arrheniusove enačbe (Ref. 7)
vvere further, additionally, tempered for 24 hours at 570 C, both vvith a class of yield stress of approx. 660 MPa. A weak reversible temper embrittlement of experimental steel, having a high tempered post-martensitic microstructure, vvas produced vvith addi-tional tempering as confirmed by our preliminary in-vestigations of the time-temperature relationship of the Charpy impact energy reduction resulting from such tempering7. An activation energy of about 160 kJ mol"1, vvhich is very close to that for bulk diffusion of phosphorus in ferrite, vvas derived from the slope of a log-log plot of tirne versus reciprocal tempering temperature (Fig. 1).
Indeed, it has already been confirmed, using Auger spectroscopy, that in particular phosphorus segre-gates in this type of steel. Romhanyi and covvorkers9 found up to 6 % P and 1 % S at grain boundaries, and a strong carbon peak vvith carbide structure is also remarkable in Auger spectra of such steel, austeni-tized at 1100 C, quenched and tempered at 600 C (Fig. 2). Hovvever, the segregations in our experi-mental steel, quenched from much lovver temperature, vvas not so intense, and the embrittlement phe-nomena could be detected only by impact testing and not at ali at semi-static tensile tests.
The results from both Tables are also shovvn in Diagram (Fig. 3), vvhere the tendency for hydrogen embrittlement is formulated in accordance vvith Morimoto and Ashida10:
R.A.-R.A.01
Degree of embrittlement =-x 100% (1)
R. A. -j
Yield Uniform Reduction stress elongation of area MPa %	%
Yield Uniform Reduction stress elongation of area MPa %	%
Ouenched and tempered tvvice at 710 C(Charpy V-notch energy:72 J) Uncharged steel
679 12.4 55 668	11.8 55.7
Hydrogen-charged steel
668 11.2 45
661
10.3
43.1
Ouenched and tempered tvvice at 710 C further, additional-ly, tempered for 24 hours at 570 C (Charpy V-notch energy: 37 J)
Uncharged steel
665 11.1 51.7 664 11.7 51.7
Hydrogen-charged steel
648 11.1 52.5
640
10.9
53
dN(E) dE "
A 3
703 Fe
E (eV)
Figure 2: Auger spectrum of the intergranular surface of steel vvith 5 % Cr, austenitized at 1100 C, quenched and tempered at 600 C (from Ref. 9) Slika 2: Augerjev spekter z intergranularne površine jekla s 5% kroma, austenitizirano pri 1000 C, kaljeno in popuščano pri 800 C (Ref. 9)
rate test and the fracture surface of hydrogen charged low- and conventional-strain rate test specimens of medium strength, i.e. the fracture surface of hydrogen charged steel, either quenched and tem-pered twice at 710 C or of the same steel addition-ally tempered for 24 hours at 570; C, are totally duetile (Fig. 6). Of course, the fracture surface of
Figure 5: Detail from Fig. 4 (SEM) Slika 5: Detail iz slike 4 (SEM)
Figure 4: Fracture surface of hydrogen charged low-strain rate tensile specimen vvith yield stress of 1226 MPa (SEM)
Slika 4: Prelomna površina vodičenega in počasi natezanega preizkušanca z napetostjo tečenja 1226 MPa (SEM)
Figure 6: Fracture surface of hydrogen charged low-strain rate tensile specimen vvith yield stress of 640 MPa (additionally tempered for 24 hours at 570 C) (SEM)
Slika 6: Prelomna površina vodičenega in počasi natezanega preizkušanca z napetostjo tečenja 640 MPa (dodatno popuščano 24 ur pri temperaturi 570 C (SEM)
temper embrittled* hydrogen charged
uneharged
Figure 3: Degree of embrittlement of hydrogen charged
steel as funetion of yield stress Slika 3: Stopnja krhkosti vodičenega jekla v odvisnosti od napetosti tečenja
3.2 Fractography
The micromorphology of the typical fracture surface of hydrogen charged lovv-strain rate tensile specimen vvith yield stress of 1226 MPa, is shovvn in Fig. 4 and 5. It can be concluded that the hydrogen-induced fracture is locally duetile, tearing type of fracture vvith some quasicleavage details on the periph-ery of larger and deeper tunnel-type dimples.
The fracture surface of hydrogen charged high strength specimens obtained at conventional strain
600 700 800 900 1000 1100 Yietd stress. rfys (MPa)
1200 1300
0,8(um
additionally tempered impact specimens as for instance Charpy V-notch specimens - even uncharged - are of mixed mode. After an additional tempering at 570 C for 24 hours, the crack propagation path changes and sporadic intergranular fracture along preaustenite grain boundaries, and quasicleavage fracture details, and single ductile tearings can be regularly observed7.
4. Discussion
Lovv concentration of hydrogen in high strength steel have no significant influence on the mobility of the dislocations in earlier stages of the tensile defor-mation process. Hydrogen has almost no effect on
Figure 7: Schematic representation of microvoid formation, grovvth, and coalescence along grain boundaries in vvhich hydrogen is adsorbed (from Ref. 13)
Slika 7: Shematski prikaz tvorbe mikropor, njihove rasti in zlivanja vzdolž meja zrn, na katerih je adsorbiran vodik (Ref. 13)
the yield stress or on the uniform elongation of steel, and it only affects the reduction of area. Hovvever, the reduction of area decreases only if the strain rate is lovv enough to enable the Cottrell atmosphere of the hydrogen atoms pinned on the dislocations to penetrate deep into the plastic zone of the tensile specimens. Since the size / of the plastic zone of a hydrogen charged specimen is approximately half the diameter of the neck (1=3 mm) at fracture, and the crosshead speed v is 1.6 x 103 mm s 1 (0.1 mm min"1), a value of strain rate e = v/l = 5.3 x 104 s1 is obtained. In earlier literature11 higher e values are quoted for stainless steel. Hovvever, the investigations performed by Nakano et aV2 on hydrogen charged steel vvith yield stress of 500 MPa using lovv strain rate measurements shovv that at sufficient concentration of hydrogen in steel the reduction of area asymptotically approaches the lovver value even at a critical strain rate of e = 104 s1, vvhich is of the same order of magnitude as in the present investigations.
Microfractographic examinations shovv that hydrogen charged lovv strain rate tensile specimens
exhibited some interfacial separation on the fracture surface. The grovvth and the coalescence of mi-crovoids along the grain boundary, schematically shovvn in Fig. 7, are accelerated by separating inter-nal interfaces vvhere hydrogen is adsorbed1314.
Microvoid coalescence and the separation of inter-nal interfaces due to adsorbed hydrogen become op-erative vvhen the triaxial stress state in the narrovv neck of the tensile specimen is formed (Fig. 7, se-quences 3 to 5), resulting in the "condensation" of the last stage of plastic deformation in the lovv strain rate tension testing of high strength hydrogen charged steel. Hovvever, such phenomena are not observed in medium strength steel. Although a strong interac-tion betvveen hydrogen and temper embrittlement vvas frequently observed in such alloyed steels12 and though the magnitude of such effect vvas directly re-lated to the degree of intergranular phosphorus en-richment2, such synergy vvas not found in the exper-imental steel vvith post-martensitic microstructure. In studying the influence of bulk and grain boundary phosphorus content on hydrogen induced cracking in lovv strength steel, Dayal and Grabke15 also found that the effect of phosphorus is related to the bulk content and not to the grain boundary concentration. Obviously, in the čase of the experimental steel vvith post-martensitic microstructure, the influence of hy-drogen decreases and becomes more complicated due to some particular effect of the microstructure. In agreement vvith Charbonnierand Pressouyre16these results shovv that the nearer is the actual microstruc-tural state to the state of the thermodynamical equi-librium, the less susceptible is the steel to hydrogen embrittlement.
5. Conclusions
The relevance of the lovv-strain rate tension test to establish the hydrogen embrittlement susceptibility of both high and medium strength steel is demon-strated. The applicable formula (1) for the estimation of such susceptibility10, based on the reduction of area measurements at the lovv and the conventional strain rate tensile test, is also successfully adopted. The synergism betvveen hydrogen and temper embrittlement vvas not found in the experimental, addi-tionally tempered steel vvith post-martensitic microstructure. On the contrary, such steel is less susceptible to the influence of hydrogen, due to some particular effects of the microstructure vvhich is close to the thermodynamical equilibrium.
References
1	C. A. Hippsley and N. P. Haworth: Mater. Sci. Tech., 4, 1988, 791-802
2	C. A. Hippsley: Mater. Sci. Tech., 3, 1987, 912-922
31. M. Bernstein and A. W. Thompson: Int. Metali. Rev., 21, 1976, 269
"J. K.Tien, A. W. Thompson, I. M. Bernstein and R. J. Richards: Metali. Trans., 7A, 1976, 821
5	M. Hashimoto and R. Latanision: Metali. Trans., 19A, 1988, 2799
6	B. Ule, F. Vodopivec, L. Vehovar, J. Žvokelj and L. Kosec: Mater. Sel. Tech.. 9, 1993, 1009-1013
7	B. Ule, F. Vodopivec, M. Pristavec and F. Grešovnik: Mater. Sci. Tech.. 6, 1990, 1181-1185
8	J. Janovec, P. Ševc and M. Koutnik: Kov. Zlit. Teh.. 29, 1995, 40
9	K. Romhanyi, Zs. Szasz Csih, G. Gergely and M. Menyhard: Kristali Tech., 15, 1980, 471-477
10	H. Morimoto and Y. Ashida: Transactions ISIJ, 23,1983, B-325
11	M. B. VVhiteman and A. R. Troiano: Corrosion. 21, 1965. 53-56
12	K. Nakano, M. Kanao and T. Aoki: Trans. Nat. Res. Inst. Met. (Jpn), 29, 1987, 34-43
13	H. Cialone and R. J. Asaro: Metali. Trans.. 10A. 1979. 367-375
14	H. Cialone and R. J. Asaro: Metali. Trans.. 12A. 1981. 1373
15	R. K. Dayal and H. J. Grabke: Steel Research, 58.1987, 179-185
16	J. C.Charbonnier and G. M. Pressouyre: Residual hy-drogen in steels, 4th International Conference "Residuals and Trace Elements in Iron and Steel", Portorož, Yugoslavia, October 1985, Proceedings, pp. 81-103