In Situ AFM Imaging System for the Environmentally Induced

1. Introduction

Most damage issues in machines and structures are dueto material degradation induced by an operating environ-ment including corrosion fatigue and stress corrosioncracking (SCC) that involves both hydrogen embrittlement(HE) type SCC and active path corrosion (APC) type SCC.In order to clarify the fracture and the damage mechanisms,serial, high-magnification observations of damage initiationand growth processes are necessary. The conventionalmethod employed for such a purpose is a scanning electronmicroscope (SEM)1,2): the test is periodically interruptedand the sample surface is examined in a completely differ-ent environment of vacuum from the testing one, and therefore, in situ observation of serial changes in surfacedamage is impossible. The other method, except observingthe sample itself, is a replication technique, where the repli-ca of the sample surface is ex situ examined by an optical microscope or an SEM. This technique has been success-fully applied in particular to study the initiation processesof fatigue in air. In a corrosive environment, however, thereplication may be a serious problem, because the replica-tion procedures affect the sample surface condition, andtherefore the replication itself and the alternation of the environment during the replication procedure can affect thesuccessive degradation processes.

Recently, a special scanning electron microscope calledenvironmental SEM or low vacuum SEM3) has been devel-oped. This type of SEM can operate at a higher pressure of the specimen chamber up to about 10 to 20 torr (1 300–2 600 Pa). It is specifically designed to study wet bearing orinsulating materials, without prior specimen preparationsuch as conductive coating. They have been applied tostudy the corrosion behavior of metals or the crack initia-tion behavior of SCC.4,5) The pressure of the specimenchamber is much higher than that of the conventional SEM.However it is still lower than the atmospheric pressure, in-dicating that the testing environment is limited to low pres-sure gases such as water vapor. In addition, the vertical res-olution of both the conventional and the environmentalSEMs is insufficient for investigating the very early initia-tion stage of environmentally induced cracking and corro-sion damage.6)

In contrast with these techniques, a scanning tunnelingmicroscope (STM), first developed in 1982,7) gives a revo-lutionizing tool to the study of surface physics and electro-chemical researches.6) It is capable of imaging nanoscopictopography of surface not only in vacuum but also in air orin aqueous solutions. Although STM imaging of noncon-ducting surfaces is impossible, an atomic force microscope(AFM), that was developed in 1986,8) can image topogra-phy of nonconducting surface.9) Up to now, these micro-

ISIJ International, Vol. 43 (2003), No. 4, pp. 579–588

579 © 2003 ISIJ

In Situ AFM Imaging System for the Environmentally InducedDamage under Dynamic Loads in a Controlled Environment

Kohji MINOSHIMA, Yoshitaka OIE and Kenjiro KOMAI

Department of Mechanical Engineering, Graduate School of Engineering, Kyoto University, Yoshida-Honmachi, Sakyo-ku, Kyoto606-8501, Japan. E-mail: [email protected]

(Received on August 31, 2002; accepted in final form on November 12, 2002 )

The atomic force microscope (AFM) based system has been developed for in situ topographic imaging ofthe environmentally induced damage such as a stress corrosion crack under a dynamic loading condition.The system consists of an atomic force microscope (AFM), a mechanical testing machine, an X–Y–Z posi-tioning stage, an environment chamber, and their controllers. To reduce disturbing vibration, the system isequipped with isolators. A dynamic load can be applied to a sample by using an electromagnetic actuator: itprovides various loading waveforms including sinusoidal, triangular, and other arbitrary programmable ones,and has a maximum load capacity of �100 N. Nanoscopic in situ AFM observation can be conducted in acontrolled gaseous environment or in an aqueous solution. By using the developed system, in situ AFM ob-servation of the following growing crack was successfully performed in a high-strength stainless steel: a fa-tigue crack in dry air and a stress corrosion crack under high-frequency vibratory stresses superimposed ona sustained load (dynamic stress corrosion cracking), and under a low-frequency varying load (cyclic stresscorrosion cracking): the nanoscopic crack tips can be clearly visualized. The crack tip of the stress corrosioncrack under dynamic loads is sharp compared with a fatigue crack in dry air, with a larger scatter of the cracktip opening displacement than those of the fatigue crack in dry air.

KEY WORDS: in situ observation; nanoscopic imaging; atomic force microscopy; stress corrosion cracking;hydrogen embrittlement; dynamic loading; dynamic SCC; cyclic SCC; fatigue; crack tip opening displace-ment; high-strength stainless steel.

scopes are called scanning probe microscopes (SPMs). Theadvantages of the SPMs are that they can operate in any en-vironment, and that the vertical resolution is extremelyhigher than the other microscopes.6)

The authors have already applied an SPM to imagingboth fracture surface (nano fractography)10) and sample sur-faces: by using the SPM, they investigated into the environ-mentally induced fracture mechanisms of various kinds ofmaterials including metals10–13) and micromaterials such asa microelement for micro electro mechanical systems(MEMS)14) and/or high-strength and high-elastic-modulusfiber15) from the nanoscopic view point. Micromaterials, inparticular, have characteristic dimensions on the order ofmm, and therefore, a nanoscopic observation tool is re-quired instead of a conventional mm-order damage evalua-tion tool such as an SEM. Allowing the advantage of theoperating environment of an AFM mentioned before, in situAFM observations were made, for example to investigatenot only the corrosion processes11) of such as intergranularcorrosion but also the mechanisms of SCC.12,13)

When SPMs are utilized for in situ imaging of the envi-ronmentally induced damage such as SCC, a loading deviceand a corrosion test cell must be installed in the SPM sys-tem. In general, the sample stage of the SPM is relativelysmall (up to about 20 mm), and the piezo-electric scanner isdirectly connected to the sample stage: an SPM image isobtained by raster scan of the sample stage. In this case, theloading device must be small and weigh light enough com-pared with the sample stage. To meet such conditions, theloading mechanism adopted is very simple, such as that bydriving screws.12,13) In this case, it is quite difficult to pre-cisely control the value of an applied stress and is impossi-ble to apply a dynamic load. For SCC, the influence of dy-namic load, such as low frequency varying load16) as well asvibratory stresses superimposed on a sustained stress16,17) isextremely important. The dynamic load sometimes causesthe enhancement of crack growth rate and the decrease inthe threshold stress of SCC. These phenomena are wellknown as cyclic SCC or dynamic SCC.16,17) These meanthat dynamic loading is one of the most important factors inclarifying the mechanism of the stress corrosion cracking.

Some SPMs, currently on the market, have a larger sam-ple stage without moving and a raster-scan SPM tip to ob-tain surface topography. Sugeta et al.18) have developed anin situ AFM instrument that consists of an AFM having alarger sample stage and piezo-electric driven actuator,which applies an out-of-plane bending load to the speci-men. They have applied the instrument for in situ observa-tion of fatigue crack growth behavior of metals in laborato-ry air. However, to investigate the mechanisms of the environmentally induced degradation of a high strength material, the testing environment should be controlled.Note that moisture in laboratory air can cause the environ-mentally induced degradation.

In this investigation, the in situ AFM imaging system hasbeen developed that consists of an environmental AFM, amechanical testing machine and an environment chamber:in situ AFM imaging of the environmentally induced dam-age, including crack initiation and its propagation, in a con-trolled environment is thereby possible under dynamicloading. The details of the developed system is first ex-

plained, and in situ images of a fatigue crack and a stresscorrosion (SC) crack under a dynamic loading condition,i.e., dynamic and cyclic SCC, were obtained in a high-strength stainless steel to examine the performance of thedeveloped system. The usefulness and the advantage of thedeveloped system are discussed.

2. In Situ AFM Imaging System

The developed system is composed of 1) an environmen-tal atomic force microscope (AFM), 2) a mechanical testingmachine, 3) an environment chamber, and 4) controllers for the AFM, the testing machine and so on. The AFM used was the Pico SPM® manufactured by the MolecularImaging Inc., USA. The AFM is so called a stand alonetype, and a piezo-electric scanner is directly connected withthe AFM tip holder: the AFM design put the piezo-electricscanner and the electronics above the sample, and therefore,it allows a large space below. It is also designed to preventany damage due to fluid leakage. The scanner unit and theAFM housing are environmentally sealed from the testingenvironment to prevent the leakage of vapor or gas from theenvironment chamber.

Figure 1 illustrates the view of the system developed.For imaging the sample surface under dynamic loading in acontrolled environment, a mechanical testing machine ishoused in the environment chamber, that mates with theAFM via an opening. The opening houses the AFM body,sealing onto a Viton® O-ring. That allows the imaging in acontrolled gaseous environment. The mechanical testingmachine is composed of a loading frame, load and displace-ment sensors, and an electromagnetic actuator that is excit-ed by a power amplifier. There are many kinds of actuatorsavailable. Of these, the electromagnetic actuator was select-ed in this system, that has a sufficient loading capacity andenables various loading waveforms with an arbitrary wave-form generator and a servo-controller. One of the most important and fascinating properties of the electromagneticactuator is to have quite low mechanical noise or vibration,and thereby the AFM imaging is possible, which will bediscussed later in detail. The mechanical testing machine isset on the X–Y–Z positioning stage, and a sample surfaceat an arbitrary position can be imaged with a help of a topviewer or an optical microscope installed in the AFM. Thehigh-precision Z positioning stage is used to let an AFM tipapproach the sample surface: the sample is first subjected todynamic loading in a controlled environment, and the sam-ple is held at the maximum load. And next the AFM tip isbrought to the sample surface by using the X–Y–Z posi-tioning stage. An AFM image is then obtained by using thepiezo-electric actuator installed in the AFM. The specifica-tions of the system are summarized as follows:

(1) Mechanical Testing Machinea) Loading actuator: Electromagnetic actuator,

Laterally driven typeb) Maximum load: �100 Nc) Maximum displacement: �2 mmd) Loading waveform: Sine, triangle, ramp, hold

wave, step, harversine, and other arbitrary wave-forms, generated by an HP 33120A function/arbi-

ISIJ International, Vol. 43 (2003), No. 4

© 2003 ISIJ 580


581 © 2003 ISIJ

(a)

(b)

(c) (d)

Fig. 1. External view of the developed AFM imaging system for the environmentally induced damage under dynamicloads in a controlled environment.(a) Overall view of the developed system (b) Mechanical testing machine and AFM (c) AFM placed on the top ofthe environment chamber (d) Specimen and AFM. Note that the polymeric isolator is not shown here for the sakeof simplicity.

trary waveform generator, Hewlett PackardCompany, USA.

e) Maximum loading frequency: about 10 Hz (sinu-soidal wave)

f ) Load and displacement control

(2) Sensorsa) Load cell: �100 N and �50 N (maximum capaci-

ty)b) Displacement sensor: Differential transformer,

�2 mm (maximum displacement)

(3) X–Y and Z Positioning Stagesa) X–Y positioning stage: �30 mm (maximum dis-

placement), 0.02 mm (resolution)b) Z positioning stage: �5 mm (maximum displace-

ment), 0.0005 mm (resolution)

(4) AFM (Pico SPM Manufactured by the MolecularImaging Inc.)a) Operating mode: Contact modeb) Maximum scanning range: �60 mm (horizontal

direction), �7 mm (vertical direction)As mentioned, the system uses the loading actuator with

low mechanical noise. However, for obtaining an AFMimage on the order of nanometer, isolating the mechanicalnoise which comes through the floor is very important. Forthis purpose, the system is mounted on air spring type iso-lators. In addition, to increase the isolation capacity, thespecially designed polymeric isolator is used. The effect ofthe polymeric isolator on the AFM images is shown inFigs. 2 and 3. In Figs. 2(a) and 3(a), the brightness of each

position expresses the height, and the brightest point corre-sponds to the highest point, and the darkest one the lowest.Figures 2(b) and 3(b) respectively show the cross-sectionsalong A–A� line shown in Figs. 2(a) and 3(a). The sampleused was a double cantilever beam specimen (Fig. 4(a)),and the polymeric isolator was set at the neck of the speci-men near the loading points. The isolator is so soft that itnever affect the value of the external load. It is clear that themechanical noise of about 40 to 100 nm in height could beseen, when only the air spring type isolators were used: seeFig. 2(b). One reason is that the double cantilever beamspecimen used here was easily vibrated by pivoting on the


© 2003 ISIJ 582

(a)

(b)

Fig. 3. Fatigue crack tip imaged with polymeric isolator.(a) AFM image (b) Cross section along A–A� line shownin Fig. 3(a)

(a)

(b)

Fig. 4. Shape and dimensions of the specimen. All dimensionsare in mm.(a) Double cantilever beam (DCB) specimen (b) Singleedge notched (SEN) specimen

(a)

(b)

Fig. 2. Fatigue crack tip imaged without polymeric isolator.(a) AFM image (b) Cross section along A–A� line shownin Fig. 2(a)

loading points. However, when the polymeric isolator wasused, the noise became less than about five nanometers inheight. This indicates that the polymeric isolator gives agood result for imaging. The following sections deal within situ AFM observation of the fatigue and the SC crackunder dynamic loading, to verify the usefulness and theperformance of the developed system.

3. In Situ AFM Observation of Stress Corrosion CrackGrowth under Dynamic Loading

3.1. Experimental Procedures

The material used to make in situ AFM observation was a high-strength stainless steel, HSL 18019). The chemi-cal composition and the mechanical properties are listed inTables 1 and 2, respectively. The materials of 70 mm in di-ameter were machined into a double cantilever beam(DCB) specimen (Fig. 4(a)) in the C–R crack plane orienta-tion, whose two letter code is defined in the ASTM stan-dard E-399-90 (reapproved in 1997): the first letter desig-nates the direction normal to the crack plane, and the sec-ond letter the expected direction of crack propagation,where “C” and “R” denote the circumference or tangentialdirection and the radial direction of the material, respective-ly. The side groove was machined on the one-side of speci-men surface to let the crack grow straight. AFM observa-tion was made on the back side of the side-grooved surface,and the time required for the AFM imaging was 216 s. Thestress intensity factor was calculated by replacing the thick-ness with the effective one obtained by the geometric meanof the specimen thickness or the distance between sides ofthe specimen (2 mm) and the net thickness at the root of theside groove (1 mm).

In order to examine the quality of the AFM images ob-tained by the developed system, an atomic force micro-scope having a larger sample stage, Dimension 3000 manu-factured by the Veeco Instruments Inc., former the DigitalInstruments Inc., was used together with a static in-planebending device shown in Fig. 5. The sample used in thissystem was a single edge notched (SEN) specimen ma-chined in the L–R crack plane orientation (Fig. 4(b)), where“L” denotes the direction of the maximum grain flow or thelongitudinal direction. By using the in-plane bending de-vice, in situ AFM observation under static loading is possi-ble in an aqueous solution or in laboratory air but not in acontrolled gaseous environment. The SEN specimen is setin the loading device, and an in-plane bending moment isapplied to the specimen by driving the screw in the bendingdevice. The stress intensity factor is controlled by monitor-ing a load by using a load cell (loading capacity: 2 kN).Note that the sample set in the loading device cannot besubjected to dynamic loading but to only static one. Forboth specimens, the specimen surface was ground withemery papers (#2000) followed with final finish with a dia-mond paste, giving mirror surface.

In situ AFM observation was made first for the fatiguecrack in dry air. The environment chamber was evacuatedto about 100 Pa, and then dry air (dew point: �74°C) wasintroduced to the chamber. The fatigue test was conductedafter the sample was conditioned in dry air for 10 h. Thedew point of dry air during the AFM observation was about

�52°C. The test was conducted at a stress ratio of 0.1 and astress cycle frequency of 3 Hz with a sinusoidal waveform.

In situ AFM observation was also made for an SC crackunder a dynamic loading condition. The environment of a3.5% NaCl�3 g/l NH4SCN solution was prepared with re-gent grade NaCl, NH4SCN and deionized water (specificresistance�1 MW · cm). The solution at 21�1°C was circu-lated between a corrosion test cell and a solution reservoirby a vane pump made of synthetic resin. The AFM obser-vation was made for dynamic SCC under high-frequencyvibratory loads superimposed on a sustained load and forcyclic SCC under a low-frequency varying load. The dy-namic SCC tests were conducted using a sinusoidal wave-form at a stress ratio of 0.9 and a stress cycle frequency of30 Hz, whereas the cyclic SCC tests under a negative pulsewaveform at a stress ratio of 0.1 and a stress cycle frequen-cy of 0.1 Hz, where the ratio of the holding time of themaximum load to that of the minimum load was nine, andthe time required for the minimum load to the maximumload and for the maximum load to the minimum load was0.2 s. The tests in the solution were conducted at the freecorrosion potential.


583 © 2003 ISIJ

Fig. 5. Dimension 3000 AFM system together with an SEN sam-ple set in the static loading device, that is mounted on theAFM sample stage.

Table 1. Chemical composition of HSL 180 steel, mass%.

Table 2. Mechanical properties of HSL 180 steel.

3.2. Crack Growth Behavior

The fatigue crack growth rate of the DCB specimens indry air was the same as that of the CT specimens (width:40 mm, thickness: 12.5 mm)20) that are frequently used forinvestigating the crack growth characteristics of metallicmaterials. The other interesting thing is that the crackgrowth rate in the corrosive environment under the pulsewaveform is higher than that of the fatigue in dry air. Thisis because the growth rate was enhanced by the hydrogenembrittlement, or so-called cyclic SCC.16,17,20) Under high-frequency vibratory stress superimposed on a sustainedstress, the growth rate is also higher than that of the staticSCC under a sustained load, in particular in a lower stressintensity factor region, i.e., Kmax�20 MPa ·m1/2, showingthe dynamic SCC behavior. As far as the present experi-ments were concerned, the crack growth rate of the staticSCC was equal to or less than from 310�10 to 210�9 m/s. This may indicate that the crack in the solutionmay extend about 60 to 400 nm, or less during the AFMimaging.

3.3. In Situ AFM Images of Fatigue Crack in Dry Air

An example of in situ AFM images of a fatigue crack bythe developed system is shown in Fig. 6: the load beingheld at the maximum load of the fatigue test, the AFM ob-servation was made without fully unloading, as remainingthe environment of dry air. Figure 7 shows an AFM imageof the fatigue crack tip of the SEN specimen, taken by theDimension 3000 AFM system. In this case, the fatigue testwas conducted in dry air under a three-point in-plane bend-ing load by using an electro-hydraulic fatigue testing ma-chine (maximum capacity: 50 kN). The test was then inter-rupted and the applied load was fully unloaded. The AFMimage of the crack was obtained after the specimen wassubjected to the same maximum stress intensity factor asthe fatigue test by using the in-plane loading device (Fig.5).

When the Dimension 3000 system is used, the disturbingmechanical noise is small compared with that by using theAFM system developed here. This is because the load wasapplied to the SEN specimen by a mechanically drivenscrew. In addition, the in situ AFM imaging system has theenvironmental chamber and the X–Y–Z positioning stage.These can enlarge the mechanical noise. However, theDimension 3000 system does not use such additional in-struments. This indicates that the AFM images obtained bythe Dimension 3000 system (Fig. 5) may have the similarquality as that obtained by usual AFM observation withoutany loading. From these figures, the quality of the AFM im-ages was almost the same between the two systems: the fa-tigue crack tip is clearly visualized by the developed sys-tem. This indicates that the developed system works in agood and precise manner.

Secondly, it is clear that the localized plastic zone isformed ahead of the crack, where the specimen surfaceformed a hollow. The fatigue crack microscopically grew ina zig-zag manner, showing a microstructure dependentcrack growth. The degree of the zig-zag path was largerwith an increase in the applied stress intensity factor range.The magnification of the AFM images is high enough, and

the crack opening displacement can be measured. In thecase of the fatigue crack in dry air, the crack opening dis-placement measured 10 mm behind the crack tip was re-spectively 700 nm for the crack shown in Fig. 6(a) and1 520 nm for the crack shown in Fig. 6(b). This indicates


© 2003 ISIJ 584

(a)

(b)

Fig. 6. In situ AFM images of a fatigued crack tip in dry air.Arrow shows the crack tip. (DCB specimen)(a) Kmax13 MPa ·m1/2 (b) Kmax21 MPa ·m1/2

Fig. 7. An AFM image of a fatigue crack tip in dry air. Arrowshows the crack tip. (SEN specimen, Kmax21 MPa ·m1/2)Note that the fatigue test was interrupted, and the cracktip was imaged by the Dimension 3000 AFM system byreloading using the in-plane bending jig, shown in Fig. 5.

that the crack opening displacement became larger with anincrease in the applied stress intensity factor range, whichwill be discussed later in detail.

3.4. In Situ AFM Images of Dynamic Stress CorrosionCrack

An example of in situ imaging of a dynamic SC crack isshown in Fig. 8. In the case of the environmentally inducedcrack, as is mentioned before, the changes in testing condi-tions during imaging should be minimized. From this viewpoint, the developed system gives us a better experimentalconditions compared with those in conventional ex situimaging.

In the case of the dynamic SCC, two types of crack mor-phologies were observed. One is that the crack grewstraight, which was not the case of the fatigue crack, andthe plastic deformation, or the hollow ahead of the crackwas smaller, compared with that of the fatigue crack (seeFig. 8(a)). This type of crack morphology was also obtainedin a static SC crack under a sustained load at the free corro-sion potential,21) where the fracture surface was solely cov-ered with intergranular cracking.20,21) The other is that theplastic deformation or the hollow of the surface ahead ofthe crack tip was larger, and this large plastic deformationwas formed on both sides of the crack (Fig. 8(b)). An inter-esting thing is that the crack did not grow straight, but in a

zig-zag manner, when a large hollow was formed ahead ofthe crack tip. The depth of the hollow ahead of the crackwas larger in this case than those of both the fatigue crackand the dynamic SCC shown in Fig. 8(a). The depth of thehollow of the fatigue crack tip increased with an increase inapplied stress intensity factor, and the depth measured 2 mmahead of the crack tip was about 180 to 190 nm at Kmax13MPa ·m1/2. However in the case of the dynamic SCC, it wasabout 289 nm for the crack shown in Fig. 8(b), whereas itwas 71 to 96 nm for the case shown in Fig. 8(a).

The fracture surface of the dynamic SCC exhibited amixed mode of intergranular and transgranular cracking,whereas the fracture surface of the static SCC exhibited in-tergranular cracking.20,21) Considering the fracture morphol-ogy and the similarity of the crack shape morphology of thedynamic SCC shown in Fig. 8(a) and the static SCC,21) the crack shown in Fig. 8(a) is considered to growalong a grain boundary, whereas the crack morphologyshown in Fig. 8(b) reflects the transgranular crack growth.These indicate that when the crack grows along a grainboundary, the plastic deformation, or the formation of the hollow is inhibited. The reason why the transgranulardynamic SC crack formed a larger hollow ahead of thecrack is required to be further investigated into.

3.5. In Situ AFM Images of Cyclic Stress CorrosionCrack

Figure 9 shows in situ AFM images of a cyclic SC cracktip. The images were taken in keeping the maximum load ina single pulse wave load. Note that they were not takenunder fully unloaded condition. Figure 9(a) shows thecyclic SC crack tip, which was associated with a elongatedplastic deformation ahead of the crack tip. A noticeablepoint is the hollow was not formed on both sides of thecyclic SC crack: it had grown along the edge of the hollow.This type of crack morphology was sometimes observed. Inthe case of Fig. 9(b), the surface hollow was formed onboth sides of the crack, similarly to the case of the fatiguecrack in dry air. The depth of the hollow measured 2 mmahead of the crack tip was 139 nm (Fig. 9(a)) and 141 nm(Fig. 9(b)), and they were almost equal to or somewhatsmaller than those of the fatigue crack. This leaning forma-tion of the hollow shown in Fig. 9(a) may be due to thecrack plane not normal to the observed surface but at a cer-tain angle.

3.6. Crack Tip Opening Displacement

As discussed before, the developed system is very pow-erful for in situ imaging of the crack tip. The advantage ofthe system is that the system uses the stand-alone typeAFM, and then leaves large space enough to combine themechanical testing frame below the AFM. Therefore, afracture mechanics specimen whose stress intensity factoris easily controlled can be used under a dynamic loadingcondition. At the same time, the nanoscopic imaging can bedone with minimizing the influence of changing the testingconditions on both the crack growth behavior and the AFMimages. This means that the AFM images taken by the de-veloped system reflect the true three-dimensional shape in amore precise manner compared with the ex situ observationby using, for example, a scanning electron microscope or a


585 © 2003 ISIJ

(a)

(b)

Fig. 8. In situ AFM images of a dynamic SC crack tip in a 3.5%NaCl�3 g/l NH4SCN solution. Arrow shows the cracktip.(a) Kmax11.1 MPa ·m1/2 (b) Kmax13 MPa ·m1/2

scanning probe microscope. As shown before, the crack tip opening displacement

(CTOD) increased with an increase in an applied stress in-tensity factor (see Fig. 6). To examine the influence of envi-ronment on the CTOD, the crack opening was measured byusing the AFM images taken by the developed system: theCTOD used here was defined as the crack opening dis-placement measured 2 mm or 10 mm behind the crack tip,and the relation between the CTOD and the maximumstress intensity factor is shown in Fig. 10. In the measure-ment, the important but difficult thing is the identificationof the crack tip. This is sometimes difficult from only thetopographic data, leading to measurement errors. For thispurpose, the differential image, or error signal image wasused to identify the crack tip. The image is based upon thedifference of the actuation voltage of the piezoelectric actu-ator during imaging, and can be regarded as a kind of dif-ferential image. By using this image, the crack tip is clearlyidentified, which is reported elsewhere.21)

The CTOD values of the static SCC under a sustainedload21) are also plotted as a reference, that were obtained byusing in situ AFM images of the SEN specimens immersedin the 3.5% NaCl�3 g/l NH4SCN solution at the free corro-sion potential. From the figure, it is clear that the CTOD

value of the fatigue crack in dry air increased with an in-crease in the applied stress intensity factor, and the valueobtained by using the DCB specimens agreed with thoseobtained by using the SEN specimens. This indicates thatthe CTOD value of the fatigue crack is determined by theapplied stress intensity factor. This is a logical result fromthe view point of the fracture mechanics. Note that the me-chanical conditions adopted in this investigation satisfiedboth small scale yielding and plane strain conditions.

In the corrosive environment, the CTOD values increasedwith an increase in stress intensity factor. An interestingthing is that the values of both cyclic and dynamic SCCwere smaller than those of fatigue. Secondly, the scatterband of the CTOD in the corrosive environment becamelarger compared with those of fatigue. In the case of thestatic SCC, similar results can be obtained: the CTOD val-ues are smaller than those of the fatigue, and a large scatterexhibited. As mentioned before, the dynamic SCC had the


© 2003 ISIJ 586

(a)

(b)

Fig. 9. In situ AFM images of a cyclic SC crack tip in a 3.5%NaCl�3 g/l NH4SCN solution. Arrow shows the cracktip.(a) Kmax11.1 MPa ·m1/2 (b) Kmax13 MPa ·m1/2

(a)

(b)

Fig. 10. Relationship between the crack tip opening displace-ment (CTOD) and the maximum applied stress intensityfactor. Note that the superscript “∗” denotes that theCTOD value was obtained by using SEN specimens.The superscript “s” denotes the data for the dynamic SCcrack growing straight with smaller hollow ahead of thecrack tip (see Fig. 8(a)). (a) CTOD measured 2 mm behind the crack tip. (b)CTOD measured 10 mm behind the crack tip.

two types of crack morphology: one is that the crack grewstraight with a smaller hollow ahead of the crack tip, andthe other is that the crack grew in a zig-zag manner with alarger hollow ahead of the crack tip. However, the differ-ence of the CTOD between the two was small, and waswithin the scatter band.

The authors13) have already reported for a high-strength7075 Al alloy that after the sample having a growing stresscorrosion crack in laboratory air was conditioned in vacu-um, the crack tip became blunt and the crack was retardedfor a while. Although the AFM imaging was made in labo-ratory air, this result indicates that the crack tip displace-ment of 7075-T651 Al alloy is influenced by an environ-ment: the laboratory air containing water vapor decreasedthe CTOD value.13) In other words, the growing static SCcrack tip is sharper than those conditioned in vacuum envi-ronment. This result agrees with the results obtained in thepresent experiments, although the material tested was dif-ferent. In the case of the material tested here, hydrogen em-brittlement causes the SC crack growth and an enhance-ment of crack growth under dynamic loading. In dry air,hydrogen content in the material was small, compared withthose tested in the corrosive environment. This indicatesthat the hydrogen in the material may inhibit localized plas-ticity of the growing crack, leading to smaller crack tipopening displacement, or sharper crack.

Kinaev et al.,5) however, reported a contradictive result.They measured the crack tip opening displacement of a4340 high-strength steel by using an environmental SEM.They introduced a fatigue precrack in laboratory air, andthen the sample, which was a bolt-loaded DCB specimen,was set in the environmental SEM. The SEM observationwas made in water vapor and hydrogen at a pressure of upto 15 torr (2 kPa), with a rising load condition and a sus-tained load condition. They reported that the crack openingdisplacement in water vapor or hydrogen was larger thanthose observed in high vacuum, or a conventional SEM op-erating condition. In their case, the testing duration waslimited to only by 400 min, and considering the relativelymild environment, the hydrogen content at the crack tipmay be extremely small. The most important thing is thatthe observed crack conditions are different. In our case, thecrack tip visualized by the AFM was a growing crack.However, they did not observe the crack extension duringthe observation: in a sense, they observed the influence ofenvironment on the crack opening behavior of the fatiguecrack, not of the SC crack. Of cause further experiments arerequired to investigate into the influence of environment, orthe hydrogen content at the crack tip on the CTOD value. Inaddition, the scatter of the CTOD values were larger in thecorrosive environment than those of fatigue in dry air. Thisis due to the crack growth is influenced by microstructuresand the amount of the hydrogen content at the crack tip,that is discussed in detail in Ref. 21.

4. Conclusions

In this investigation, the in situ AFM imaging system hasbeen developed for visualizing the nanoscopic topographiesof environmentally induced damage under dynamic loads ina controlled environment. The system is composed of 1) an

atomic force microscope (AFM), 2) a mechanical testingmachine with a low-noise actuator, 3) an environmentchamber, and 4) controllers for the testing machine, theAFM, and so on. The AFM used is a stand alone type, andthe piezoelectric scanner is connected with the AFM tipholder, and therefore, the AFM allows a large space below.By using this characteristic AFM and low-noise actuatorsthe system has been developed. The developed system wasapplied to the in situ observation of a stress corrosion cracktip of a high-strength stainless steel. The investigationyielded the following conclusions.

(1) Clear AFM images of the crack tip can be success-fully obtained by using the developed system under loadingand in controlled environments. For AFM imaging, one ofthe most important issues is to minimize the mechanicalnoise or vibration. For this purpose, the specially designedpolymeric isolator is used, giving a good vibration isolatingperformance together with the air-spring type isolators: insitu observation was successfully conducted both in dry airand in a corrosive environment under both sustained loads,fatigue or low frequency varying loads, and small vibratoryloads superimposed on a sustained load.

(2) In the case of the dynamic SCC, two types of crackmorphologies were observed. One is that the crack grewstraight, and that the plastic deformation, or the hollowahead of the crack tip was small. The other is that plasticdeformation or the hollow ahead of the crack tip was larger,with zig-zag crack path. In the latter case, the depth of thehollow was larger than those of fatigue, cyclic SCC and dy-namic SCC with straight crack growth.

(3) The crack tip opening displacement (CTOD) of thefatigue crack in dry air is solely determined by appliedstress intensity factor. However, in the corrosive environ-ment, the CTOD value scatters larger, but is lower thanthose of the fatigue crack. This indicates that the stress cor-rosion crack is sharper than fatigue crack in dry air.

Acknowledgments

This investigation is supported by the Special Coordina-tion Fund for the Promoting Science and Technology,“Foundation of Hydrogen in Environmental Degradation ofStructural Materials”, Ministry of Education, Culture,Sports, Science and Technology of Japan, and the ScientificFund of the Ministry of Education, Culture, Sports, Scienceand Technology of Japan for the fiscal years, 2000 and 2001(Contract No.: 12650084) and 2001 and 2002 (ContractNo.: 13450046). Thanks are extend to Dr. N. Nakama,Sumitomo Precision Products Co., Ltd., for the donation ofthe test materials. Their support is greatly appreciated.

REFERENCES

1) K. Komai, K. Minoshima, S. Kinoshita and G. Kim: JSME Int. J., 31(1988), 606.

2) K. Minoshima, I. Nagashima and K. Komai: Fatigue Fract. Eng.Mater. Struct., 21 (1998), 1435.

3) G. D. Danilatos: Microscopy Res. Tech., 25 (1993), 354.4) A. Atrens and Z. F. Wang: J. Mater. Sci., 33 (1998), 405.5) N. K. Kinaev, D. R. Cousens and A. Atrens: J. Mater. Sci., 34 (1999),

4931.6) K. Minoshima and K. Komai: Mater. Sci. Res. Int., 2 (1996), 209.7) G. Binnig, H. Rohrer, C. Gerber and E. Weibel: Appl. Phys. Lett., 49

(1982), 57.


587 © 2003 ISIJ

8) G. Binnig, C. F. Quate and C. Gerber: Phys. Rev. Lett., 56 (1986),930

9) G. Meyer and N. M. Amer: Appl. Phys. Lett., 56 (1990), 2100.10) K. Minoshima, M. Endo, T. Miyawaki and K. Komai: Trans. Jpn.

Soc. Mech. Eng. Series A, 61 (1995), 1587.11) K. Komai, K. Minoshima and M. Itoh: J. Mater. Sci. Jpn., 43 (1994),

329.12) K. Komai, K. Minoshima and M. Itoh: J. Mater. Sci. Jpn., 43 (1994),

336.13) K. Komai, K. Minoshima and T. Miyawaki: JSME Int. J. A, 41

(1998), 49.14) K. Minoshima, T. Terada and K. Komai: Fatigue Fract. Eng. Mater.

Struct., 23 (2000), 1035.15) K. Minoshima, Y. Maekawa and K. Komai: Int. J. Fatigue, 22

(2000), 757.16) K. Komai and K. Minoshima: JSME Int. J. Series I, 32 (1989), 1.17) K. Minoshima and K. Komai: Localized Corrosion (Current

Japanese Materials Research Vol. 4), ed. by F. Hine, K. Komai andK. Yamakawa, Elsevier Applied Science Publishers, Barking,England, (1988), 123.

18) A. Sugeta, K. Hirose, Y. Uematsu and M. Jono: Proc. 51st AnnualMeeting of Soc. Mater. Sci., JSMS, Kyoto, (2002), 310.

19) T. Uehara, R. Watanabe and N. Nakama: US Patent, US5288347(1994), US5358577 (1994).

20) K. Minoshima, K Takuno and K. Komai: Proc. 77th Annual Meetingof JSME Kansai, Osaka, No.024-1, (2002), 8-7.

21) K. Minoshima, Y. Oie and K. Komai: Proc.2002 Annual Meeting ofJSME/MMD, JSME, Tokyo, No.02-05, (2002), 633.


© 2003 ISIJ 588

Documents

In Situ AFM Imaging System for the Environmentally Induced