HAL Id: jpa-00225416
https://hal.archives-ouvertes.fr/jpa-00225416
Submitted on 1 Jan 1985
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
HYDROGEN-INDUCED CRACKING IN PURE IRON
J. Armstrong, S. Carpenter
To cite this version:
JOURNAL DE PHYSIQUE
C o l l o q u e C I O , s u p p l é m e n t au n012, Tome 46, d é c e m b r e 1985 page CIO-139
HYDROGEN-INDUCED CRACKING IN PURE IRON
J. H. ARMSTRONG AND S.H. CARPENTER
U n i v e r s i t y o f Denver, Department o f Physics, Denver,
Colorado
80208, U . S . A .Abstract
-
The modulus and internal friction of Armco iron rere continuously m e a s u r e d d u r i n g c a t h o d i c charging rith hydtogen to investigate crack initiation and grorth. The observed modulus decrease ras attributed to crack initiation and grorth. The internal friction increase doring cathodic c h a r g i n g r a s atfributed to plastic deformation accompanying the crack formation. Both the modulos and internal friction behavior rere found to be a sum of tro parallel exponential processes. The tro exponential processes rare consistent rith difftrent sources of carbon for the crack-producing hydrogen bubble nucleation.1
-
INTRODUCTIONOne of the most prominent affects of hydrogen in pure iron is the cracking observed near the outer layer of cathodically charged simples. The p r i m r y purpore of this stady ras to determine if continuous, in situ. modulas measurements could be used to investigate the initiation and growth of cracks in pure iron. In addition, continuous internal friction measurements, rhich rere the by-product of the modulus measurements. rere used to investigate the mechanisms involved in the hydrogen- indaced csaaking.
Typically, the modulus of a material should decrease with the initiation and grorth o f cracks o r voids in the material. A theoretical treataent by B r i s t w (1) of the effective modulus of a material containing 'n' penny-shaped cracks per unit volume of average radins 'a' indicated suoh a decrease in modulus. Hydrogen-induoed cracking in iron is localized to the surface. Thus the effective modulus of the total sample could reflect a sample which consists of a oracked outer layer rith a solid inner oore. A cylindrical sample o f diameter 'd' rith a oracked layer of depth 'Ad' has an effective modulns 'Mn written as
3 na A d )
M
= Y o (1 - K I Twhere M is the modulus of the uncracked oore. The value of K1 depends upon ~oisron'% ratio for the unoracked naterial and the mode of vibration. As the change in modulus is proportional to the ohange in resonrnt frequoncy, equation (1) may be rerritten as
CIO-140 JOURNAL
DE
PHYSIQUErhere f is the initial resonant frequency. A f is the change in frequency and the constant K2 depends u p o n the material and the mode of deformation. Thus measured resonant frequency change should be directly related t o measurable metallographic quantities.
The diagram of the apparatus used t o measure the resonant frequency and internal friction has been described elserhere (2). The sample r a s driven in a s t a n d i n g r a v e at r e s o n a n t frequency (540 kHz) by a Marx type (3) composite oscillator of 112 rsvelength in a torsional mode of oscillation. T h e internal friction and vibratory strain amplitude rere measured using techniques found in the literature (4,s). The center of the quartz bars and the sample r e r e displacement nodes, thereby alloring electrical and support connections to be made rithout affeoting the standing ravetrain. The torsional mode of oscillation r a s chosen as it ras less susceptible to disturbance of the ravetrain rhile the sample ras in the charging solution. The samples rere ~ m u l t a n e o u s l y cathodically charged at current densities ranging from 10 to 3 0 malcm in a 1N H SO solution rith trace amounts of &,O3 and CS added to enhance hydrogen uptake.
f
3 h raveiength Al2o3 biffer rod r a s placed getreen the oscillator and the sample. The solution ras circnlated around the sample and the temperature of the solution r a s maintained by a heat exchanger placed in a reservoir. Samples rere maohined to one-half ravelength frcg Armco iron. In a11 cases. the samples rere annealed i n a vacuum better than 1 0torr for 6 hours at 5 0 0 ~ ~ after machining and before cathodic chaaging. The samples rere electropolished in H P O at a current density of 6 0 ma/cm prior to cathodic charging to provide uniform 3sur%ace conditions.
III
-
BXPERIMENTAL RESDLTSTypical internal friction and sample resonant frequenoy behav o r d u r i n g
2
cathodic charging o f pure Araico iron at a carrent density of 30 malcm at 6'~ are s h w n in figure 1. In al1 cases, the resonant frequency decreased as a function of time. The decay r a s rapid during the first 2 0 minutes of charging, follored by a s l w decrease for op to 900 minutes of charging. The internal friction increased significantly during the first 20 minutes. bat tends to decrease for longer chargins time. A micrograph of a sample cathodically charged a current density of 30 malcm for 6 0 minutes at a temperature of 6'~ is shorn in figure 2. Note that the cracking ras essentially intergranirlar and ras localized to a layer near the surface.The careful removal of the craoked outer layer of tro samplei cathodically charged for 5 and 480 minutes respectively resulted in the recovery of the resonant frequency loss observed during cathodic charging as shorn in figure 3. The slight decrease i n resonant frequency belor its original value is mort likely due to uncertainty caused by the necesssry removal, machining and regluing of the sample.
Samples rere charged for a variety of current densities, vibratory strain amplitudes, oharging solution temperatures and oharging times. A11 samples were subjected to metallographic analysis to determine the average c r y k radius, depth of cracking and nuaber density of cracks. A plot the qaantity na A d l d (equation 2) versos the corresponding change in resonant frequency is gigen in figure 4. The data yield a fair linear fit. r i t h a slope o f -1.33
X
1 0 and a oorrelation coefficient of 0.822. The line ras constrained t o go through the origin as no frequency loss ras observed rithout cracking. An increase in the charging oirrrent density, rhich inoreases the effective hydrogen pressure, affects the magnitude and kinetio behavior of the frequency decay (2).tails' used in determining decay constants for mixed radioactive sources, the frequency r a s found t o be parallel sum of a rapid and a slow exponential decay process. Regardless of charging time, current density, vibratory strain amplitude or charging solution temperature, a sum of tro parallel processes ras obtained. A typical frequency process is shorn in figure 5. Likerise, the internal friction ras found to be a of tro parallel exponential processes as shorn in figure 6.
Both processes have b e e n shorn (2) to be the result o f the initiation and growth of cracks. Thus. a plausible explanation for the existence o f a t r o part parallel process may be that the hydrogen bubbles responsible for the initiation and grorth of cracks nucleate at different rates. Carpenter and Parks (6) shored that annealed samples exhibited a more rapid. greater frequency decay than samples that aere unannealed. As both samples rere carefully machined from an ingot of iron not subjected to severe deformation, one can assume that the only difference r a s the source of grain boundary carbon o n rhich the hydrogen bnbbles rould nucleate. The unannealrd sample wonld show a majority of the carbon in the form of carbides rhile the annealed sample rould exhibit more free carbon. A similar behavibr ras noted at elevated temperatures in steels by Weshphal and Worzala (7).
V
-
CONCLUSIONSThere are a number of conclnsions rhich may be made from this investigation. First, the method of oontinuous monitoring of the resonant frequency and internal friction o f a material c a n be useful in detecting the presence o f cracks in a material. Second, there is a definite relationship betaeen the metallographic data and the modulas loss. I n al1 cases, the resonant frequency and internal friction behavior rere found to be parallel exponential in nature. The behavior r a s consistent w f t h the nacleation o f hydrogen bnbbles at different rates on either grain boundary carbon or carbides.
This research ras funded by the U.S. Department of Energy, Basic Sciences Division.
1. Bristor, J.R., Brit. J. Aool. Phvs.,
U r
81 (1960).2. Armstrong, J.H., Ph.D. thesis. University of Denver, June 1985. 3. Marx, 3 . .
Bsv.
Soi. Instrum,,a,
503 (1951).4. Robinson. W.H. and ûdgar.
A.,
PEL Technibal Note No. 204,NZ,
DSIR (1970). 5. Robinson, W.H., Carpenter S.E. and Tallon, J.L..J.
Anol. P ~ V S L . G r 1975(1974).
6. Carpenter, S.E. and Parks, J.E., Scriota Metall,
U,
699 (1981).(210-142 JOURNAL
DE
PHYSIQUEF i g u r e . 1
-
Typioal resonant frequenoy Figure 2-
Micrograph of o a t h o d i o a l l y and i n t e r n a 1 f r i c t i o n . charged Armoo i r o n .3
Figure 3
-
Frequenoy reoovery by removal Figure 4-
P l o t of ~ a A d / d v e r s a s of t h e o a t e r oraoked l a y e r . resonant freqnenoy l o s s .F
..-
...
..e*L(*.
W...- -
- . d l . . d A i s t i i l d i t .Figure 5
