Metall. Res. Technol. 111, 25–35 (2014)c© EDP Sciences, 2014DOI: 10.1051/metal/2014005www.metallurgical-research.org

Metallurgical Research&Technology

Prove of hydrogen formation through directpotential measurements in the rolling slitduring cold rolling

S.V. Merzlikin1,2, M. Wildau3,�, K. Steinhoff4 and A.W. Hassel2

1 Max Planck Institute for Iron Research, Max-Planck-Str. 1, 40237 Düsseldorf, Germany2 Institute for Chemical Technology of Inorganic Materials, Johannes Kepler University Linz,

Altenberger Str. 69, 4040 Linz, Austriae-mail: achimwalter.hassel@jku.at

3 Ingenieurbüro Dr.-Ing Monika Wildau, Am Kerper Weiher, 41352 Korschenbroich, Germany4 Steinhoff Kaltwalzen GmbH, Gerhard-Malina-Str. 65, Dinslaken, Germany

Key words:Cold rolling; forged rolls; hydro-gen embrittlement; hydrogen;open circuit potential; friction;tribo-corrosion; hydrogen wear

Received 28 August 2013Accepted 23 January 2014

Abstract – In this work, direct potential measurements during cold rolling of zinc andX20Cr13 stainless steel were carried out in the rolling slit to follow the tribologic and gal-vanic mechanisms of hydrogen formation and absorption on the surface of the workingrolls made of DHQ1 grade steel. An Ag/AgCl in 3.5 M KCl reference microelectrode wasused to record the open circuit potential of the electrochemical system roller-product im-mersed into commercially relevant electrolyte (rolling emulsion) with a pH value of 4.5and an electric conductivity 46 mS cm−1. The potential shift into either negative or posi-tive direction of the rolls-product system gives information on the processes taking place atthe surface in the course of the friction. A detailed discussion of the in-situ potentiometryexperiments reveals a stationary situation established between the destruction and repassi-vation of the surface structures during continuous cold rolling accompanied with intensivehydrogen evolution. Galvanic coupling of the working rolls with the product significantlyintensifies the hydrogen embrittlement related problems of the rolls. Atomic hydrogen isadsorbed on the surface and exhibits a pressure supported absorption into the rolls duringtheir whole lifetime.

C old rolling is an essential techno-logical step used in the steel indus-try to manufacture sheets, strips and

foils with extremely smooth surfaces and ac-curately controlled dimensions. Continuousrotation at high velocities and loads makesa mechanical failure of a forged roll verydangerous for employees and equipment ofthe plant. Among the possible damage rea-sons are irregularities of the roll, such asresidual stresses due to inclusions or seg-regations, which may lead to local heatingor local mechanical overloads during therolling process. Recently a number of me-chanical failures of the rolls during rollingattributed to hydrogen embrittlement (HE)were reported by the steel industry. It is wellknown from industrial applications such asfuel pipelines, parts of the machines in ag-gressive environment, hydrogen storages,

� Deceased on 23/12/2013.

power-plant boiler pipes etc. that steels suf-fer from HE. In fact, as it was pointed out byBernstein [1], hydrogen embrittlement is asevere environmental type of failure that af-fects not only steels but almost all metals andalloys. Experiments of Nagumo [2] showedthat high strength steels in particular arehighly susceptible to hydrogen embrittle-ment. Under certain conditions hydrogenexists in statu nascendi on the metal sur-face when it was just formed from wateror other hydrogen containing compounds.Before this atomic hydrogen combines tomolecular hydrogen, which can be releasedinto the adjacent atmosphere, it may be ab-sorbed by the metal. Hydrogen is inimitableamong all atoms, since its ionised form, theproton has no further electron and thus thediameter decreases dramatically to that ofa nucleus (in case of deuterium) or thatof an elementary particle. It can diffuse in

Article published by EDP Sciences

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

the metal and at particular places it can betrapped inside the material as it was shownby Hirth [3]. In presence of hydrogen steelsfail at far lower load levels compared withthose that steel free of hydrogen can sus-tain. Up to now there is no single concept ofHE accepted because all of the observationscannot be accounted for by a single mech-anism. There is a very informative reviewof Sofronis and Robertson [4] summarizingdifferent mechanisms of HE. The same be-haviour considering the degradation of themechanical properties was observed in ourprevious works [5, 6] for hardened forgedrolls for cold rolling, where molecular hy-drogen accumulated at some points in thenear-surface region of the rolls resulting inheavy spalling and blistering of the mate-rial. Numerous characteristic 0.5−1 mm longcracks formed on the working surface of theroll in the vicinity of blisters, as well as sub-surface (2 to 8 mm below the surface) fatiguecracks point at the hydrogen induction ofthis mechanical failures.

The origin of hydrogen and the driv-ing force of hydrogen formation during coldrolling are of great practical and scientificinterest. The mixed open circuit potentialof the roll-product system during the coldrolling, along with the tribochemical aspects,was found to be one of the driving forces forhydrogen evolution by Shpenkov [7].

During cold rolling enormous frictionbetween metals in a metallic couple roll-product stripe immersed into rolling emul-sion allows building up an electrochemicalpotential difference. Two principle mecha-nisms are imaginable. Rolling a less noblemetal such as zinc or zinc coated steel witha steel roll generates conditions in which thetwo metals are on one hand in good electri-cal contact with each other and on the otherhand are both immersed in the same elec-trolyte which in this case is the rolling emul-sion. This mechanism is further referred toas galvanic mechanism. The second mech-anism is the tribochemical mechanism inwhich the high mechanical stress causes rup-ture of the passive film which may chemi-cally repassivate. Akiyama et al. [8], Abelevet al. [9], Hassel and Smith [10] found thatsuch a repassivation is going via splitting ofwater from the emulsion.

Table 1. Chemical composition of the DHQ 1steel.

Elemental concentration, wt.%Steel grade C Si Cr Mo

DHQ1 0.80 0.75 2.10 0.34

The stress or pressure applied betweena roll and a strip plays a great role in thetribomechanical mechanism. Weisz-Patraultet al. [11] showed the evaluated normal gappressure during the cold rolling to be in theorder of 1500 MPa with a maximum at thepoint of roll-product stripe contact, whichis the interface of the tribogalvanic elementwhere all electrochemical and tribochemicalprocesses take place. One metal becomes acathode and another acts as anode being de-stroyed by the mechanism of electrochem-ical corrosion. Galvanic coupling of differ-ent rolls and metals in the product stripare also playing an important role. The pro-cesses of corrosion and hydrogen formationattributed to it can be investigated by mea-suring the difference between the electrodepotentials of this metallic couple.

The novel approach presented in thiswork is to make use of these aspects andto measure an indication of hydrogen ab-sorption and hydrogen-related mechanicalfailure of the rolls through a direct measure-ment of the electrochemical potential in therolling slit during rolling.

1 Experimental

A pair of forged model rolls with diame-ter of 64 mm were specially designed andmade from DHQ 1 steel (Tab. 1, SteinhoffKalt-walzen) to fit the manual rolling mill (TypeK65, DIMA Maschinen GmbH), Figure 1.

One pair of the rolls made of 42CrMo4steel was hard chrome plated resulting in a70−100 μm thick hard chrome layer. Com-mercial rolling oil KT5.3 (Quackeroel) wasused and 2 wt.% of the oil was mixedwith deionized water to prepare the rollingemulsion. The electric conductivity and pHvalue of the emulsion were measured to be46 mS cm−1 and 4.5 (at 22 ◦C) correspond-ingly. Fresh emulsion has been continuouslystirred and added in the beginning of eachindividual rolling experiment, the stability

26

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

30 cm

Fig. 1. Manually powered rolling mill.

of the electrolyte was visually controlled.100 mm wide zinc foil and X20Cr13 stain-less steel strips were used as rolling stock.Rolling speed was 0.06 m/s (or 18 rpm), frontand back tensions were equal to zero, stripthickness at entry was 1.2 mm, at exit ap-prox. 1.1 mm. Since no direct measurementof contact pressure or load was available onthe mill used, the roll separation force hasbeen kept constant mechanically throughthe whole series of experiments by identicalsetting of the screwdown mechanism. Po-tentials were measured directly by using aAg/AgCl (3.5 M KCl, 205 mV vs. StandardHydrogen Electrode (SHE)) micro electrodewith solid electrolyte on agar basis, devel-oped by Hassel et al. [12], inside a 1 mm thickglass capillary as reference electrode. Signalconditioning, data acquisition and recordingwas performed through a potentiostat (Sim-Pot, HHU Düsseldorf, Germany) directly in thegap between the rolls during rolling of thematerial (Fig. 2).

Here, the rolls act as working electrode(WE) and the micro-reference electrode (RE)was immersed into the electrolyte (emul-sion). Since only potentiometry was per-formed a counter electrode was not nec-essary. Initial experiments were performedwith an electrical steel rolling mill (TypeK120/160, DIMA Maschinen GmbH) in whichrolls from the same material, also pre-pared specifically for these experiments,

were used. These experiments showed avery high noise level due to the unavoidablestrong vibrations generated during rollingwith the electric mill. Furthermore, the pre-cise positioning of the reference electrodewas not guaranteed because of these vi-brations. It therefore sometimes led to thedestruction of the RE. In case of the man-ual rolling mill a stable fixation of the REand a constant distance between the REand WE provided reliable in-situ potentialmeasurements.

Determination of the total hydrogen con-centration of the steel samples was obtainedthrough melt extraction method, using anELTRA OH 900 hydrogen analyzer (ELTRAGmbH, Neuss, Germany). Calibration of theanalyzer was carried out using helium gasand reference material (LECO Corporation,St. Joseph, MI, USA) with certified hydrogenconcentration. Prior to analysis the sampleswere cleaned with acetone, rinsed with ul-trapure deionised water (PURELAB PulseWater Purification System, ELGA Lab Wa-ter, Marlow, UK) and finally cleaned withmethanol and CCl4 (p. a., Sigma Aldrich).

2 Results and discussion

2.1 Potentiometric monitoringof the rolling process

First, the chrome-plated pair of rolls wasmounted onto the rolling mill. A zinc foiland a stripe made of stainless steel were pro-cessed using 2% KT5.3 emulsion. The changeof the open circuit potential during rollingis shown in Figure 3. After product mate-rial was fed into the rolls and wetted withthe electrolyte, the electrode potential of themetallic couple gradually shifted to its sta-ble value. After a steady potential was es-tablished, the rolls have been moved (thismoment marked with a red arrow in Fig. 3).Regardless of the product being rolled, thepotential changed rapidly in the direction ofnegative values. A sharp change of the po-tential at the beginning of rolling interactionis due to the friction-induced destruction ofthe topmost oxide layer and deformationalactivation of the metallic surfaces. The opencircuit potential of the system then restoredits initial value if given enough relaxation

27

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

WE

RE c)b) a)

c)

Fig. 2. Scheme (a,b) and (c) photo of the in-situ potentiometry experiment.

Fig. 3. Electrode potential (vs. SHE) during con-trolled cold rolling of (a) zinc foil, (b) stainlesssteel; working rolls – chrome-plated 42CrMo4.

time (some tens of minutes). This effect istypically observed for passivating materialsas it was demonstrated by Celis et al. [13]and Mischler et al. [14] in a series of tribo-corrosion experiments. The restoration of theopen circuit potential is also an indirect con-firmation of the emulsion stability duringthe rolling experiment.

If there is insufficient time for the systemto relax, the open circuit potential stabilizes

Fig. 4. Electrode potential (vs. SHE) during theshort cold rolling pulses of: (a) zinc foil, (b)stainless steel; working rolls – chrome-plated42CrMo4.

at lower values as for example shown inFigure 4 for short continuous pulses of coldrolling. This is a proof of a metastable stateof the surface under friction. The potentialwas then restored at a value somewhat lowerthan the initial one as a layer of native oxidewas formed on the surface of both rolls andthe product.

28

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

Fig. 5. Electrode potential (vs. SHE) duringthe short cold rolling pulses of: (a) zinc foil,(b) stainless steel; working rolls – DHQ 1 steel.

The whole contact area of the elec-trodes during the rolling is in the metastablestate in which an electrode potential de-pends on the velocity of rolling, as reportedby Shpenkov [7] and more recently, Rossiet al. [15] claimed also the dependence of thepotential on contact pressure. Celis et al. [13]found out that even the direction of the rota-tion influences electrode potential. This phe-nomenon was also seen during the investi-gation in Fig. 4b), where a lower rotationspeed of the rolls is marked with red arrows.The data can be found in the literature, thatlower contact pressure in such tribosystemresults in higher changes of the open cir-cuit potential, as for example in the work ofShyrokov and Vasyliv [16]. In present case,the contact pressure was the same for eachmeasurement as a very long stripe of theproduct was used without readjusting thesetup.

In Figure 5 the results for the pair ofrolls made of DHQ1 steel are shown. Redarrows in Figure 5a indicate rolling in the

Fig. 6. Electrode potential (vs. SHE) as a func-tion of length of the product’s initial posi-tion displacement during rolling: (a) zinc foil,(b) stainless steel; working rolls – chrome-plated 42CrMo4.

same direction at constant velocity, but withthe opposite change in the electrode poten-tial of the system. The direction of poten-tial change is opposite for the same work-ing rolls at the same rolling direction, force,load and speed. In order to explain such abehavior, the potential during rolling andthe one immediately after rolling (when asteady state was established) was plottedagainst the distance against initial positionof the product, that is its length (Figs. 6and 7). The total circumference of the roll’ssurface was 200 mm. An initial steady statepotential of −0.49 V to −0.35 V (zinc foil)and −0.07 V to −0.09 V (stainless steel) vs.SHE has settled at a level of mixed poten-tial of two metals (at the given conditions,according to the reference values of the stan-dard potentials in aqueous solutions takenfrom the Milazzo et al. [17], Bard et al. [18]and Bratsch [19]) at zero length. During coldrolling, the potential was decreasing to about

29

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

Fig. 7. Electrode potential (vs. SHE) as a functionof length of the product’s initial position dis-placement during rolling: (a) zinc foil, (b) stain-less steel; working rolls – DHQ 1 steel.

a half of the roll’s circumference, afterwardsincreasing. The effect was more pronouncedin the case of non-plated steel rolls (Fig. 7).Provided sufficient time for relaxation, thestable state potential tends to restore its orig-inal value. If one takes a look at the longermeasurement in Figure 7b, where potential’schanges along two full rounds of rollingare shown, a quasi-periodic oscillation ofthe open circuit potential with respect toan imaginary zero-length point on the rollcan be seen (dashed line). This is the re-flection of a dynamic equilibrium betweenthe destruction and recovery of the oxidelayer and a mass transport of the wear parti-cles originating from the metals in the elec-trolyte, this fact being in a good correspon-dence with the analysis of the electrochem-ically controlled tribocorrosion experimentby Landolt et al. [20], who took into accountthe formation, flux and ejection of these wearparticles from the sliding contact. The poten-tial shifts towards negative values because

of the destruction of the oxide surface layerof metal were observed also in the reviewon electrochemical methods in tribologicalsystems by Landolt et al. [21]. The same be-haviour of the potential was observed byOltra et al. [22], who followed an evolutionof the corrosion potential from the passiverange towards the active range during con-tinuous abrasion of the steel sample in di-luted sulphuric acid. This destruction phaseis followed by the phase of passivation of thesurface. The phase of repassivation of thebare steel surface submerged into the elec-trolyte manifests itself in a rise of the elec-trode potential. These phases replace eachother cyclically, following the rotation of therolls and thus causing the observed sharp os-cillations of the electrode potential. The sameobservation was found by Sholud’ko [23] fordifferent electrolytes. According to his re-sults the potential change depends to a largedegree on the material of the friction partsand on the presence of oxidizing/reducingagents in the electrolyte.

2.2 Hydrogen evolution during coldrolling

2.2.1 Thermodynamics

The changes in the potential are generallydue to the changes in the energy state of thesurface during and after friction. In orderto estimate a thermodynamic probability ofany phenomenon on the surface of metals inaqueous electrolyte solutions, Pourbaix di-agrams [24] can be used. The lines for theregions of stability, corrosion and passiva-tion of metals in Pourbaix diagrams (E vs.pH) obey under equilibrium conditions theNernst equation:

E = E0 − RTzF

ln(aRed

aOx

)

where E is the equilibrium potential at giventemperature, E0 is the standard potential, Fis the Faraday constant, R is the universalgas constant, T is the absolute temperature;z is the number of moles of electrons; aRed

and aOx are the chemical activities of the re-ductant and oxidant respectively. Since theactivity factor tends to unity at low concen-trations, activities in the Nernst equation can

30

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

Fig. 8. Measured intervals of the OCP change inthe system rolls-product.

be simply considered as concentrations ofthe components which are being reduced oroxidized during the reaction. During coldrolling, metals are being exposed to severefriction. As a result, Gibbs thermodynamicpotential of the metal changes on value ΔG,respectively changing the equilibrium po-tential E:

ΔE = −ΔGzF.

This fact in its turn leads to the shift ofthe lines and regions of the Pourbaix dia-grams. One can therefore introduce the spe-cific Pourbaix diagram of the system underfriction (e.g. cold rolling) as proposed byCelis et al. [13]. Dedy et al. [25] in his workmade some attempts to create the frictionmaps in conjunction with the equilibriumPourbaix diagram for the copper-water sys-tem. But unfortunately, there is no detailedinformation or atlases of Pourbaix diagramsfor the vast majority of such tribo-systems.The thermodynamic state of the metal un-der non stationary conditions, including thechanges in electrode potential can be stillevaluated by the classical equilibrium Pour-baix diagrams. Pourbaix diagrams for Fe-H2O, Zn-H2O, and Cr-H2O, systems wereconsidered for the surfaces contacted underfriction during the cold-rolling process.

From the diagrams for Fe and Zn in aque-ous solutions with pH = 4.5, at the valuesof electrode potential measured during thecold rolling (see Fig. 8) the dissolution of baremetal can occur through the following an-odic reaction, Me→Me2++2e−, where Me is

Fe or Zn. It is accompanied with the cathodicreaction H2O+e− → Had+OH−, where Had isa hydrogen atomically adsorbed to the sur-face of the metal. Anderson et al. [26] studiedthe behaviour of surface potentials of differ-ent metallic electrodes, scraped inside thesolutions with different pH and describedthe phenomenon of an accelerated hydrogenevolution on freshly cleaved metallic sur-faces in aqueous medium; it has been widelyreported since then in the literature. For ex-ample, Burstein and Kearns [27] did an ex-cellent study on describing an acceleratedevolution of hydrogen from freshly gener-ated surfaces of Cr, Fe and their alloys.

From the thermodynamic analysis ofchromium the mechanisms of anodic disso-lution and passivation were elaborated byTsuru [28], Bjornkvist and Olefjord [29] didit for acidic solutions. The experiments onthe system Cr-H2O to establish the productsof chromium dissolution were made also byDrazic and Popic [30]. From their results itis clear that a layer of Cr(OH)2 is formed onchromium surface at the potentials of Cr-Fecouple from Figure 8. The mechanism is thefollowing:

Cr +H2O↔ CrOHad +H+ + e−

CrOHad +H2O→ Cr(OH)2,ad +H+ + e−

Cr(OH)2,ad + 2H+ ↔ Cr2+(aq) + 2H2O

in case of dissolution

Cr(OH)2,ad ↔ CrOOHad +H+ + e−

in case of passivation.

In the region of the Pourbaix diagram, whichis corresponding to the Cr-Zn couple duringthe cold rolling experiment, the final step isthe passivation.

As it has been shown above, with theinitiation of the rolling process the sur-face oxide layer is being destroyed and afresh metallic surface is exposed to the elec-trolyte at the side of RE (Fig. 9). In thecourse of further rolling this film is con-tinuously destroyed again and again. Af-ter a while, due to the self-organization ofthe destroyed friction surface, the passivefilm is restored in the interaction betweenthe activated friction surface and the elec-trolyte. These processes are inevitable dur-ing the cold rolling. As shown by Andersonet al. [26] and Burstein and Kearns [27] they

31

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

Fig. 9. Scheme of friction surface of metals dur-ing the cold rolling.

are always accompanied by the formationof hydrogen on the freshly exposed metal-lic surface in the presence of aqueous mediacontaining hydrogen ions.

Moreover, as it can be seen in Figures 3and 4 for chrome plated working rolls, aswell as in Figures 5 and 7 for DHQ-1 rolls,the sign of the potential during zinc andstainless steel rolling is opposite: negativefor zinc and positive for stainless steel. Thisdifference in potentials can also impact theprocess of hydrogen ingress into the work-ing roll. Hydrogen diffuses into the bulk ofthe material from the surface because of theconcentration gradient, since there is morehydrogen on the surface, where hydrogen isbeing formed, adsorbed and absorbed. Thesign of the potential can only play a rolein accelerating the proton adsorption on thesurface of the electrode. If the open potentialof the metal is more negative than the elec-trode potential at point of zero charge (pzc),than the surface is negatively charged and at-tracts cations. Thus, hydrogen atoms will bereadily adsorbed on it. For iron Epzc is equalto−0.35 V (SHE) [31], being clearly less nega-tive than values of the potential in Figure 7afor Fe-Zn couple in steady state and dur-ing friction (–0.36 to −0.75 V (SHE) accord-ingly), favoring H+ad adsorption on the roll.In case of stainless steel strip (Fe-Fe situationin Fig. 7b), observed potential (which was0.09 V in steady state and became negativedown to −0.2 V during rolling) were morepositive than Epzc for iron, showing that(thermodynamically) the surface of the elec-trode has to be charged positively, i.e. protonrepulsive. But during the friction the openpotential of the system tends to go deeper

in the negative direction towards the scribepotential of the Fe (–0.44 V SHE), where thesurface again becomes negatively chargedand proton attractive. Moreover, in real sit-uation, where fresh and passivated metal(iron) surface immersed simultaneously intoone electrolyte, a passivated iron surface isan anode, whereas fresh surface is a cathodewhere hydrogen evolution takes place.

2.2.2 Galvanic coupling

While in the case when stainless steel is cold-rolled with working rolls made of DHQ 1,hydrogen evolution appears mainly becauseof the fresh formed metallic surfaces ex-posed to the emulsion, other roll-productcombinations have their own peculiarities.When the rolls-product system is built upfrom different metals, galvanic coupling con-tributes additionally to the changes in opencircuit potential of the system and can leadto an enhanced hydrogen production. In thegalvanic couple Fe-Cr, both metals are closeto each other in standard potential series,thus developing only a very small poten-tial difference. Hard chromium plating ofthe working rolls is known to have goodcorrosion resistance, thus successfully pre-venting hydrogen diffusion into the rolls.The surface of an average chromium platedroll was investigated by Simao and Aspin-wall [32]. They found chromium deposit tohave various microscopic cracks resultingin a network of open cracks on the sur-face of the coated layer. Fedrizzi et al. [33]claimed pitting-like corrosion morphologycan be possibly initiated at these pre-existingcracks. However, the corrosion phenom-ena were found to be of secondary impor-tance, because the material removal duringthe wear-corrosion occurred mainly throughan adhesive mechanism. An important as-pect of hard chromium coating of the rollsis however, the environmental risk associ-ated with the hexavalent chromium used inthe plating process, which is a carcinogenand has several serious concerns like longterm legal liability for worker health andsafety, waste disposal and soil contamina-tion. Recent regulation efforts, including theOSHA Cr6+ PEL in the USA and the Eu-ropean RoHS, ELV and WEEE regulations

32

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

directives, resulted in new tougher rulesbeen introduced worldwide. These latestdevelopments present a formidable arrayof restrictions for manufacturers and sur-face finishers, and the ban of the hexava-lent chromium is to be expected in the nearfuture. There are already working know-hows for hard chromium functional coat-ings using trivalent chromium electroplat-ing [34]. Unlike for the hexavalent chromiumplating, where cathodic DC current is be-ing used, a modified waveform includingcathodic and anodic DC current, followedby relaxation phase was used in this process.An equivalent to hexavalent chromium mi-crostructure and hardness, decreased poros-ity and good adhesion at a 250 μm thicknesswas obtained [35]. Since during electroplat-ing Cr3+ is reduced at the cathode (surfaceof the roll) to metallic chromium by gain-ing three electrons, no change in the corro-sion potential is expected. Therefore, mainlymechanical properties of the chromium hardcoating, such as surface stability and hard-ness play an important role in the preventionof hydrogen diffusion into the roll, good per-formance of chromium-coated rolls with re-spect to HE will not suffer from replacementfrom Cr 6+ to Cr 3+ plating bathes. The mostsevere case of hydrogen evolution occurs ingalvanic couples with zinc. The contact be-tween these metals is so intimate that eventhe formability depends on the crystallo-graphic texture as it was shown in the workby Signorelli et al. [36], and thus is strongenough to allow for strong galvanic cou-pling. The Fe-Zn couple at pH = 4.5 (the casewhen zinc foil is cold rolled with normal,non-chrome plated rolls) the product (Zn)acts as a sacrificial anode Zn→ Zn2+ + 2e−,since iron is more noble than zinc accord-ing to the electrochemical series (Milazzoet al. [17], Bard et al. [18] and Bratsch [19]).The evolution of hydrogen takes place at thecathode, which is the surface of the roll, ac-cording to the cathodic half reaction:

2H+ + 2e− → H2 ↑H+ + e− → Had

O2 + 4e− + 4H+ → 2H2O.

Another critical point is the high overpoten-tial of hydrogen evolution on zinc which suc-cessfully hinders hydrogen evolution and

Fig. 10. Total hydrogen amount (±0,01 μg.g−1)of the samples made of DHQ 1 in 2% KT5.3emulsion, with and without contact to zinc foil.

therefore redirects the hydrogen formationto the susceptible rolls.

The following simple experiment illus-trates this behaviour. DHQ 1 steel samplesin electric contact with a zinc foil were im-mersed into the 2% KT5.3 emulsion used forcold rolling. The change of the total hydro-gen concentration of the steel samples wasmeasured as a function of time spent in thesolution. As it can be seen from Figure 10, thetotal hydrogen content of the steel samplesconnected to zinc rises significantly strongerthan that of uncoupled samples.

Formed in this way nascent hydrogen isinitially adsorbed on the surface of the rollsin its atomic rather than molecular form.The concentration gradient of hydrogen be-tween bulk and the surface of the rolls de-livers the driving force for the hydrogen dif-fusion, further supported by the enormouspressure during the cold rolling. In caseswhere the rolling emulsion contains surfac-tants the surface concentration of atomic hy-drogen can be further increased by hinder-ing the recombination of adsorbed atomichydrogen to the molecule. The concentra-tion of Had on the surface of the rolls in-creases thus promoting its absorption. De-formation and high temperatures on slidingmetallic surfaces increase atomic hydrogen’ssolubility and diffusion. Since the hydrogenformation is continuously happening duringthe cold-rolling, large amounts of hydrogendiffuse into the rolls increasing the risk ofhydrogen-induced embrittlement.

33

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

Main stages of hydrogen-induced me-chanic failure of the working rolls can beformulated:

Hydrogen evolution from hydrogenouslubricants due to the tribochemical reactionsand high surface temperature in the fric-tion zone + hydrogen evolution from thefreshly formed surfaces of the friction zonesubmerged in the aqueous rolling emul-sion → accompanied with the adsorptionof atomic hydrogen on the surface of therolls → followed by diffusion of the atomichydrogen into the rolls supported by theconcentration, temperature and stress gra-dients from the surface to the bulk → ab-sorption and subsequent recombination ofthe atomic hydrogen inside the defects of thesurface stressed layer→ resulting in a brit-tle mechanic failure as a result of formationand coalescence of the individual subsurfi-cial cracks.

3 Conclusions

Through in situ measurement of the opencircuit potential directly in the rolling slit ofthe rolling mill during the cold-rolling in-formation on the surface in the course ofthe friction can be obtained. The potentialcan shift into either negative (due to the de-struction of the surface oxide layer) or posi-tive (due to the subsequent repassivation ofthe surface) direction depending on the pro-cesses taking place at the surface.

The change in open circuit potential dur-ing cold-rolling is a complex synergy of tri-bochemical, electrochemical and mechani-cal properties of the rolling mill system. Anequilibrium potential of a metal under me-chanical stress is different from that of theunstressed metal because of the changes inthe free energy. Additionally, different me-chanical loading of the electrodes (rolleror product) leads to a mixed potential be-tween the loaded and unloaded areas. Thespeed of the cold rolling and the mechani-cal stress during rolling affect the electrodepotential of the system. Still, a mass trans-port of the charge carriers inside the elec-trolyte to and from the electrode plays a greatrole in open circuit potential since it affectsthe composition of the double layer nearthe electrode surface. A stationary situation

is established between the destruction andrepassivation of the surface structures arisesduring the continuous cold rolling process.The electrode potential in this progressingstate depends on the mode of friction. Un-der severe conditions leading to the cyclicdestruction of the surface and catastrophicwear, followed by its restoration, sharposcillations of the electrode potential wereobserved.

The destruction of the oxide layers andexposure of the fresh metallic surface tothe aqueous emulsion is accompanied withintensive hydrogen evolution. Atomic hy-drogen is adsorbed on the surface and ex-hibits a pressure supported absorption intothe rolls during their whole lifetime in therolling mill. Galvanic coupling of the work-ing rolls with the product intensifies the hy-drogen embrittlement related problems ofthe rolls. Brittle mechanic failure and blister-ing of the forged rolls are the consequencesof the hydrogen-related wear.

Acknowledgements

The financial support of this work provided bythe German Federal Ministry of Economics andTechnology under grant number KF 2001501SU8is gratefully acknowledged.

References

[1] I.M. Bernstein, Hydrogen Effects inMaterials, in: A.W. Thompson, N.R. Moody(Eds.), TMS, Warrendale, PA, 1996, p. 3

[2] M. Nagumo, ISIJ Int. 41 (2001) 590-598[3] J.P. Hirth, Metall. Trans. A 11 (1980) 861-890[4] P. Sofronis, I.M. Robertson, AIP Conference

Proceedings 837 (2006) 64-70[5] S.V. Merzlikin, A.W. Hassel, K. Steinhoff,

M. Wildau, Practical Metallography 07 (2011)365-375

[6] A.W. Hassel, M.M. Lohrengel, Electrochim.Acta 40 (1995) 433-437

[7] G.P. Shpenkov, in Tribology Series. 29 D.Dowson (Ed.), Friction Surface Phenomena,Elsevier, 1995

[8] E. Akiyama, M. Stratmann, A.W. Hassel, J.Phys. D 39 (2006) 3157-3164

[9] E. Abelev, A.J. Smith, A.W. Hassel, Y. Ein Eli,J. Electrochem. Soc. 153 (2006) B337-B343

[10] A.W. Hassel, A.J. Smith, Corros. Sci. 49 (2007)231-239

[11] D. Weisz-Patrault, A. Ehrlacher, N. Legrand,J. Mater. Process. Technol. 211 (2011) 1500-1509

34

S.V. Merzlikin et al.: Metall. Res. Technol. 111, 25–35 (2014)

[12] A.W. Hassel, K. Fushimi, M. Seo,Electrochem. Commun. 1 (1999) 180-183

[13] J.-P. Celis, P. Ponthiaux, F. Wenger, Wear 261(2006) 939-946

[14] S. Mischler, S. Debaud, D. Landolt, J.Electrochem. Soc. 145 (1998) 750-758

[15] S. Rossi, F. Deflorian, M. Zen, L. Fedrizzi,Mater. Corros. 51 (2000) 552-556

[16] V.V. Shyrokov, Kh.B. Vasyliv, Mater. Sci. 44(2008) 646-652

[17] G. Milazzo, S. Caroli, V.K. Sharma, Tablesof Standard Electrode Potentials, Wiley,Chichester, 1978

[18] A.J. Bard, R. Parsons, J. Jordan, StandardPotentials in Aqueous Solutions, MarcelDekker, New York, 1985

[19] S.G. Bratsch, J. Phys. Chem. Ref. Data 18 (1989)1–21

[20] D. Landolt, S. Mischler, M. Stemp, S. Barril,Wear 256 (2004) 517–524

[21] D. Landolt, S. Mischler, M. Stemp,Electrochim. Acta 46 (2001) 3913–3929

[22] R. Oltra, B. Chapey, L. Renaud, Wear 186-187(1995) 533-541

[23] M. Pourbaix, Atlas of ElectrochemicalEquilibria in Aqueous Solutions, PergamonPress, Oxford, 1966

[24] V.P. Sholud’ko, Fiz.-Khim. Mekh. Mater. 18(1982) 89–92

[25] Dedy Ng, Tapajyoti Sen, Feng Gao, HongLiang, J. Electrochem. Soc. 155 (2008) H520-H524

[26] T.N. Anderson, J.L. Anderson, H. Eyring, J.Phys. Chem. 73 (1969) 3562-3570

[27] G.T. Burstein, M.A. Kearns, J. Electrochem.Soc. 131 (1984) 991

[28] T. Tsuru, Mater. Sci. Eng. A 146 (1991) 1-14[29] L. Bjornkvist, I. Olefjord, Corros. Sci. 32

(1991) 231-242[30] D.M. Drazic, J.P. Popic, Corrosion 60 (2003)

297-303[31] S. Trasatti, J. Electroanal. Chem. 33 (1971) 351-

378[32] J. Simao, D.K. Aspinwall, J. Mater. Process.

Technol. 92-93 (1999) 281-287[33] L. Fedrizzi, S. Rossi, F. Bellei, F. Deflorian,

Wear 253 (2002) 1173-1181[34] FARADAYIC r© Process for Chromium

Plating from a Trivalent Bath, http://www.faradaytechnology.com/PDFfiles/IndustrialCoatings

[35] EPA, Final Report: A Cost-CompetitiveFunctional Trivalent Chromium PlatingProcess To Replace Hexavalent ChromiumPlating http://cfpub.epa.gov

[36] J.W. Signorelli, M.J. Serenelli, M.A.Bertinetti, J. Mater. Process. Technol. 211(2011) 1500-1509

35