1. Introduction

One of strengthening mechanisms of iron and its alloys iswell known to be solid-solution strengthening by interstitialatoms such as carbon and nitrogen. Since the solubility ofnitrogen in a-Fe is two times and five times (at least)greater than that of carbon at temperatures of 20 and590°C, respectively,1) thereby decreasing the tendency forprecipitation at a given level of strengthening. Furthermore,dissolved nitrogen is a more effective solid-solutionstrengthener than carbon at (or below) room temperature inaustenitic stainless steels.2,3)

To improve the surface hardness, wear and corrosion re-sistance, fatigue life, as well as global mechanical perfor-mance of pure iron and simple iron alloys, nitriding hasbeen widely used in industries. Many nitriding processeshave been applied to pure iron and low carbon steels, for instance, gas nitriding,4–10) salt bath nitriding,11) plasma nitriding,10,12–16) plasma immersion ion implantation(PIII),16–20) laser nitriding,21–25) and ion implantation nitrid-ing.26) Although the above methods are well established,some of them such as modern nitriding techniques still havesome disadvantages from an engineering viewpoint becausethey require the use of rather complicated and expensiveapparatus, and cannot, in general, produce thick nitridedlayers.

A large amount of nitrogen can be introduced into thesurface layers of iron and titanium by laser irradiation ofthe metals in liquid medium such as liquid nitrogen or am-

monia, but the laser irradiation did not succeed in introduc-ing a detectable amount of nitrogen into the metals whenconducted in 1 atm gas ambient.27) The white compoundlayer formed during gas nitriding is related to the quantityof nitrogen atoms in the interface between the gas mediumand workpiece surface, e.g., if the nitrogen supply exceedsthe absorbing capability of the workpiece, besides the diffu-sion zone, the nitrides layer is present.6)

Conventional industrial gas-nitriding is usually per-formed in NH3 gas at temperatures ranging from 450 to650°C for tens of hours. Nevertheless, the nitrided layerthickness is usually less than 0.5 mm. The gas nitriding canresult in a white brittle nitrites layer on the surface. In mostcases, this white layer must be removed by machining be-fore use. Another nitriding method, which is most common-ly used in industries, is salt-bath nitriding using liquid saltsconsisting of cyanide and cyanate. In fact, this process isactually a nitrocarburizing process, because the environ-ment of molten salt contains both carbon and nitrogen, andthe two elements diffuse into steel parts, simultaneously. Itsmajor problem in industrial application is the toxicity in-volved in the cyanide–cyanate salt. To the best of ourknowledge, no nitriding for steel without simultaneous car-burization has been realized with salt baths. Therefore, aneffort for finding a pure salt-bath nitriding of steel is neces-sary to be done.

Reported decomposition temperatures of potassium ni-trate (KNO3) are about 527–567°C28) and 628°C29) obtainedby differential scanning calorimetry and differential ther-

ISIJ International, Vol. 46 (2006), No. 1, pp. 111–120

111 © 2006 ISIJ

Nitriding of Interstitial Free Steel in Potassium–Nitrate Salt Bath

Yin Zhong SHEN,1) Kyu Hwan OH1) and Dong Nyung LEE1)

1) School of Materials Science and Engineering, College of Engineering, Seoul National University, Seoul 151-744, Korea.E-mail: shenliu8@snu.ac.kr, kyuhwan@snu.ac.kr, dnlee@snu.ac.kr

(Received on June 15, 2005; accepted on October 31, 2005 )

A study has been made of nitriding of interstitial free (IF) steel in the potassium–nitrate salt bath at tem-peratures ranging from 400 to 650°C. The salt is decomposed to generate nitrogen and oxygen. Nitrogendiffuses into steel, or steel is nitrided, while oxygen reacts on steel surface to form the oxide scale. Theoxide scale thickness is much smaller than the nitriding thickness. Most of nitrogen resides in steel as aform of interstitial solid-solution. For nitriding at higher temperatures, nitride precipitates (g�-Fe4N and z�-Fe2N) exist mostly in grain boundaries and partly in grains of the steel. The nitrate nitriding gives rise tomuch larger nitriding depth than other nitriding methods at similar nitriding temperature and time. The ni-trate nitriding of steel substantially increase its tensile strength as well as hardness, e.g., an IF steel speci-men nitrided at 650°C for 1.5 h shows a tensile strength of 916 MPa, which is 2.2 times higher than that ofnon-nitrided IF steel specimen, and an elongation of 20% at 70°C. Severe serrations are observed in flowcurves of nitrided steel specimens, mainly due to dynamic strain aging that occurs because of interactionbetween dissolved nitrogen and moving dislocations. The effective diffusion coefficient of nitrogen DN ob-tained from the nitriding data, DN�D0 exp(�Q/RT ) with D0�3.789�10�7 m2· s�1 and Q�76.62 kJ mol�1, isapproximately the same as that for diffusion of nitrogen in a -iron.

KEY WORDS: potasium nitrate; steel; nitriding; hardness; tensile strength; solid solution strengthening; dif-fusion coefficient; dynamic strain aging.

mal analyzer techniques, respectively. Potassium nitratemay decompose by some of the following reactions28,30–33):

KNO3→KNO2�1/2 O2 ........................(1)

2KNO2→K2O�NO2�NO......................(2)

2KNO2�2NO→2KNO3�N2....................(3)

KNO2�NO2→KNO3�NO .....................(4)

K2O�2NO2→KNO2�KNO3 ...................(5)

Reaction (2) has been observed at temperatures rangingfrom 517 to 537°C.28) In addition, potassium nitrate mayalso directly decompose into potassium oxide30,34,35) accord-ing to the following reaction (6) or (7):

2KNO3→K2O�2NO2�1/2 O2 ..................(6)

2KNO3→K2O�5/2 O2�N2.....................(7)

Thus, the thermal decomposition of KNO3 on heating canliberate nascent nitrogen and the nascent nitrogen can dif-fuse into the steel coupon. Therefore, there is a possibilityof nitriding in the liquid KNO3 salt bath under ambientpressure.

Interstitial-free (IF) steels are ultra low carbon and nitro-gen steels manufactured by adding carbide, nitride and sul-phide-forming elements such as titanium and aluminum toform precipitates such as TiC, TiN, AlN, and TiS, leadingto an elimination of the nitrogen, carbon and sulphur ele-ments from the solid solution like nearly pure iron. For thisreason, in order to investigate the effect of solid-solutionstrengthening by diffusion of nitrogen to the interstitial sitesof a-Fe lattice when KNO3 salt bath is used as nitriding ofiron and steels, we chose to use a Ti�Al stabilized IF steel,in which Ti and Al scavenge both carbon and nitrogen.

The purpose of this study is to investigate nitriding be-havior of IF steel in the KNO3 salt-bath, with emphasis onsolid-solution hardening and nitriding kinetics.

2. Experimental Procedures

A hot-rolled IF steel sheet of 4 mm in thickness was usedas the initial material. The chemical composition of thesteel is given in Table 1. The sheet was cold-rolled by 10,30, and 75% in thickness under lubrication.

Extra pure potassium nitrate (99.0 wt% min. purity) withimpurities of chloride (Cl, 40 ppm max.), sulfate (SO4,80 ppm max.), ammonium (NH4, 60 ppm max.), lead(10 ppm max.), iron (8 ppm max.), and calcium (60 ppmmax.) was used for nitriding of steel. The nitrate nitridingtemperature ranged from 400 to 650°C. Samples to be ni-trided were cleaned in hot 10% HCl solution to remove thesurface scales. Annealing of steel samples were carried outin the 67%CaCl2–33%NaCl salt bath (the chloride bath) toavoid any reaction between the samples and the bath. Afterthe heat treatments, all the samples were quenched in coldwater to remove surface scales. The scale-free specimenswere slightly ground on 2000-grit SiC emery paper to get abright surface before hardness test. Vickers-hardness testswere performed with Instron Wolpert Tester 930 underloads of 1 to 2 kg. The measured surface hardness valuesare mean values of ten measurements.

Optical and SEM micrographs were obtained from 3%nital-etched samples using Olympus PMG3 microscope andJSM-5600 scanning electron microscope (SEM) equippedwith OXFORD Link ISIS 300 Energy Dispersive X-raySpectrometer (EDS), respectively. Auger electron spec-troscopy (AES) of the samples was conducted by using aScanning Auger Microprobe System (model 660).

The hot-rolled IF steel sheet of 4 mm in thickness was ni-trided for 3 h in the KNO3 bath (the nitrate bath) at 650°C,and the nitrogen and oxygen concentrations in a 0.4 mmthick surface-layer (from 50 to 450 mm in depth) and a0.5 mm thick center-layer of the nitrided sample were mea-sured using a LECO TCH 600 Nitrogen/Oxygen/HydrogenDeterminator (an experimental error of less than 0.1 ppm).

Powders taken from the surface of samples heated in thenitrate bath were subjected to X-ray diffraction (XRD) withthe Cu-Ka1 radiation (l�0.154056 nm) to measure theirlattice parameters and phases.

Tensile specimens of 1 mm�6 mm�25 mm in gauge di-mensions, in accordance with ASTM E 8M-99, were spark-erosion-cut from steel sheets, with the tension axis parallelto the rolling direction. They were ground and polished upto a bright surface before tensile test. Tensions were con-ducted on an Instron-5582 testing machine at an initialstrain rate of 4.23�10�5 s�1. The stress drop magnitude inserration, which was defined as the engineering stress dif-ference between the maximum and minimum points of in-dividual serrations, was calculated using a special computerprogram edited for serration data analysis.

3. Results and Discussion

3.1. Formation of Oxide Scale

Figure 1 shows the XRD spectrum of the powder sampletaken from the surface of an IF steel sample heated for 1.5 hin the nitrate bath at 650°C. The XRD result indicates thatit consists of magnetite (Fe3O4), hematite (Fe2O3), andwustite (FeO). Therefore, oxygen must have been generatedin the nitrate bath.

It has been reported that magnetite and hematite form on

ISIJ International, Vol. 46 (2006), No. 1

© 2006 ISIJ 112

Table 1. Chemical composition (wt%) of IF steel

Fig. 1. XRD spectrum of powder taken from surface of IF steelheated in KNO3 bath at 650°C for 1.5 h.

the surface of cold-rolled low carbon steel after carbonitrid-ing in the ABI salt bath consisting of MNO3, MNO2, andM2CO3, where M refers to Na and K, at 400°C.11)

Figure 2 shows a change in surface-oxide thickness of afull-annealed IF steel sample during heating for 16 h in thenitrate bath at 580, 600, and 640°C as well as for 4, 8, and16 h at 640°C. The full-annealed sample was made by 30%cold rolling and subsequent annealing for 2.5 h in the chlo-ride-bath at 650°C. The steel sample undergoes losses ofabout 154 mm (77 mm on one side) and 35 mm (17.5 mm onone side) in thickness after heating for 16 h at 640°C and580°C, respectively.

3.2. Nitriding

Figure 3 shows the Vickers hardness (load: 2 kg) of thesurface of 10 to 75% cold-rolled IF steel samples afterheating for 1 h in the nitrate bath at temperatures of 400 to650°C. The hardness increases with increasing reduction ata given heating temperature because of strain hardening. Asthe heating temperature increases, the hardness increases toa maximum value and then decreases, with the increasingrate being low up to about 550°C, above which the increas-ing rate becoming high. The 10, 30, and 75% cold-rolledsamples have their maximum hardness values at 600, 610,and 620°C, respectively. This hardness increase duringheating may be associated with nitriding.

In order to see if the hardness increase has to do with ni-triding, the same samples were annealed in the chloride

bath, which does not contain any nitrogen-generating com-ponents, at 650°C for 1 min to 4 h. The surface hardnessvalues of the chloride-bath annealed samples are shown inFig. 4. The hardness values decrease with increasing an-nealing time as expected. For the 10% cold-rolled sample,the hardness decreases a little even for an annealing time of3 h, whereas the 30% and 75% cold-rolled samples under-go drastic decreases in hardness. The differences in hard-ness values of the specimens are attributed to whether theyunderwent recrystallization or not. The 10% cold-rolledspecimen did not undergo recrystallization, because of lackof driving force for recrystallization, unlike the 30% and75% rolled specimens. Figure 5 shows some optical micro-graphs of center part in thickness direction of 30% rolledspecimens annealed in the chloride-bath. The hardness datain Fig. 4 and the microstructures in Fig. 5 indicate that re-crystallization was completed within about 1 h for the 75%rolled specimen and about 2 h for the 30% rolled specimen.Therefore, the hardness increases in Fig. 3 must be relatedto diffusion of nitrogen into the steel samples, or nitriding.

If nitriding takes place, a question arises why the hard-ness values decrease above about 620°C in Fig. 3. One pos-sibility may be recrystallization occurring above the tem-perature of 620°C. Figure 6 shows SEM micrographs ofthe cross-sections of 30 and 75% cold-rolled IF steel sam-ples after heating for 1 h in the nitrate bath at 600 and640°C, which are respectively 20°C lower and higher thanthe temperature at which the hardness values are maximal.We do not see any evidence of recrystallization in the speci-mens. This may be attributed to the nitrogen segregation ingrain boundaries and to interaction between dissolved nitro-gen and dislocations inside of grains, which pin grainboundaries and dislocations, resulting in prevention of re-crystallization. Therefore, the decreases in hardness above620°C in Fig. 3 cannot be due to recrystallization.

Figure 7 shows SEM micrographs of the full-annealedIF steel after nitriding at various conditions. White and darkparticles are observed along the grain boundaries. Theamount and size of the particles tend to increase with in-creasing nitriding time (Figs. 7(a)–7(c)), and are larger in aregion near the surface than in the center region (Figs. 7(d),7(e)) and at 640°C than at 620°C (Figs. 7(e), 7(f)). TheEDS and AES results indicate that the particles have highernitrogen concentrations than the matrices, as shown in Fig.

ISIJ International, Vol. 46 (2006), No. 1

113 © 2006 ISIJ

Fig. 2. Change in oxide thickness of full-annealed IF steel sam-ple during heating in KNO3 salt bath. � heating at 580 to640 °C for 16 h; � heating at 640 °C for 4 to 16 h.

Fig. 3. Vickers hardness (load: 2 kg) of surface of 10 to 75%cold-rolled IF steels after heating for 1 h in KNO3 saltbath at various temperatures.

Fig. 4. Vickers hardness (load: 1 kg) of surface of 10 to 75%cold-rolled IF steels after annealing in CaCl2–NaCl saltbath at 650°C.

8, possibly due to higher diffusion rates along the grainboundaries.

In order to identify the particles, XRD of the sample ni-trided at 640°C for 16 h was carried out after grinding away150 mm in depth from the surface to remove the surfacescale. The XRD result is represented in Fig. 9, which shows

g�-Fe4N and z-Fe2N phases. The XRD result along with theEDS and AES results in Fig. 8 indicate that the particlesdistributed along the grain boundaries seem to be z-Fe2Nand g�-Fe4N. In addition, it seems that there are fine nitridesin the matrix of the sample nitrided at higher temperatures,as shown in Figs. 7(d) and 7(e). Further studies such as

ISIJ International, Vol. 46 (2006), No. 1

© 2006 ISIJ 114

Fig. 5. Optical microstructures of midthickness of 30% cold-rolled IF steel after annealing in 67%CaCl2–33%NaCl bathat 650°C for (a) 0 min, (b) 5 min, (c) 15 min, (d) 1 h, (e) 2.5 h, and (f) 3 h.

Fig. 6. SEM micrographs of cross-sections of cold-rolled IF steel after heating for 1 h in KNO3 salt-bath. (a) 30% cold-rolled, heated at 600°C; (b) 30% cold-rolled, heated at 640°C; (c) 75% cold-rolled, heated at 600°C; (d) 75%cold-rolled, heated at 640°C.

ISIJ International, Vol. 46 (2006), No. 1

115 © 2006 ISIJ

TEM works are desirable for direct identification of theseprecipitates. The formation of g�-Fe4N is favorable for pro-longed nitriding at high temperatures.36) Usually, g�-Fe4Nforms a kink of coarse needle-like precipitates dispersed inthe ferritic matrix.13) For the conventional nitriding, g�-Fe4N and e-Fe2�3N are present in nitrided layers. The for-mation of z-Fe2N during nitriding is reported.16) The e-Fe2�3N phase is found to form at the grain boundaries (orjunctions) of the ultrafine-grained a-Fe phase, when nitrid-ed for 9 h at 300°C in ammonia gas.4) It follows from these

facts that the hardness drops above 620°C in Fig. 3 seem tobe due to g�-Fe4N.

Chemical analyses of nitrogen and oxygen in the full-an-nealed IF steel after heating for 3 h in the nitrate bath at650°C were performed. The analyses results are given inTable 2, which indicate that the nitrogen concentration ofthe surface layer is almost ten times higher than that of thecenter layer. The oxygen concentration varies little withdepth, indicating that most oxygen is consumed to form ox-ides at the surface. It is quite clear that nitriding took place

Fig. 7. SEM micrographs of cross-sections of full-annealed IF steel after heating in KNO3 salt-bath. (a) 640°C for 1 h,(b) 640°C for 4 h, (c) 640°C for 16 h, (d) 640°C for 16 h (near-surface), (e) 640°C for 16 h (center), (f) 620°C for16 h.

Fig. 8. (a) EDS and (b) AES results obtained from sample of Fig. 7(d).

in the sample. Figure 10 shows XRD results of 30% cold-rolled sam-

ples before and after heating for 3 h in the nitrate bath at650°C, which do not show any iron nitrides. This resultseems to contradict the result of Fig. 9. However, as the ni-trogen concentration is small, and thus, if nitrides had beenformed, the volume fraction of nitrides would therefore alsobe small, the XRD phase analysis may not be sensitiveenough to detect the presence of iron nitrides.

We have calculated the lattice parameters from the datain Fig. 10, and the calculated results are given in Table 3.The lattice parameter (0.28711 nm) of the heat-treated sam-ple is slightly larger than that (0.28679 nm) of the non-treated one, indicating that nitrogen is in interstitial sites.

The XRD work by Walkowicz et al.37) indicates that theaverage lattice parameter of a-Fe(N) phase formed in ISO35CrMoV5 steel nitrided for 8 h at 540°C is 0.28694 nm.The lattice constant a of a-Fe calculated from a�(4/ø3

–)rFe,

with the radius of Fe atom, rFe�0.124 nm, is 0.28637 nm,and the increment of lattice constant, D a, is about5.7�10�4 nm. In this study, the difference between the lat-tice parameters before and after the heat treatment,D a�3.2�10�4 nm, is comparable with their data. The dif-

ferences in peak intensities of the sample before and afterheating in Fig. 10 are due to a slight texture change afterheating.

We estimate the nitrogen concentration of the surface ofthe heat-treated sample using the lattice parameter data in Table 3. Let atomic radius and mass of element i be ri

and mi, respectively. Then, rFe�0.124 nm,38) mFe�55.85,39)

rN�0.071 nm,38) and mN�14.01.39) The lattice parameter ofFe is aFe�(4/ø3

–)rFe. Since the interstitial positions in bcc Fe

are 1/2 00, 0 1/2 0, 00 1/2, the interstitial gap in a-Fe is cal-culated to be [aFe�2rFe�0.03836 nm]. When one nitrogenatom intrudes in the gap, a lattice-parameter expansion of[2rN�0.03836 nm�0.10364 nm] takes place. The lattice-parameter expansion after nitriding, (0.28711–0.28679) nmfrom Table 3, is assumed to be caused by the nitrogen intru-sion. Then, 0.10364 XN�(0.28711–0.28679), with XN

being atomic fraction of nitrogen. We obtain XN�0.003088and XFe�0.9969, from which the nitrogen concentration iscalculated to be 0.0776 wt%. According to Wriedt andZwell,40) the unit cell parameter of a iron increases linearlyby 3.2�10�3 nm per wt% N dissolved. If the lattice-para-meter expansion after nitriding, 3.2�10�4 nm from Table 3,is assumed to be caused by the nitrogen intrusion, then thenitrogen concentration of the surface of the nitrate-bathheated (NBH) sample is estimated to be about 0.1 wt%. It isinteresting to note that these estimated nitrogen concentra-tions are similar to the solubility of nitrogen in a-Fe at650°C, about 0.09 wt%.41)

Tensile testing results of the chloride-bath annealed(CBA) sample and the NBH samples reveal a pronounceddifference, as shown in Fig. 11. The flow stresses of theNBH samples are more than two times higher than that ofthe CBA specimen. The tensile strengths of the NBH speci-

ISIJ International, Vol. 46 (2006), No. 1

© 2006 ISIJ 116

Fig. 9. XRD spectrum of ground surface of full-annealed IFsteel after nitriding for 16 h in KNO3 salt bath at 640°C.

Fig. 10. XRD spectra of 30% cold-rolled IF steel (a) before and(b) after nitriding in KNO3 bath at 650°C for 3 h.

Fig. 11. Stress–strain curves of 75% cold rolled IF steel speci-mens after heating in salt baths at 650°C, with initialtensile strain rate of 4.23�10�5 s�1 at 70°C.

Table 2. Concentration (wt%) of nitrogen and oxygen in full-annealed IF steel after heating for 3 h in KNO3 bathat 650°C.

Table 3. Identification of XRD peaks in Fig. 10 and calculatedlattice parameters a and mean values a.

mens nitrided for 0.5, 1.5, and 3 h at 650°C are 1.7, 2.2, and2.15 times higher than that of the CBA specimen, respec-tively. It is noted that the specimens (2) and (4) in Fig. 11are heat-treated for the same period of time in differentbaths at the same temperature. However the NBH sample(2) shows a much higher tensile strength than the CBAsample (4), indicating a pronounced solid-solution strength-ening in NBH sample.

Furthermore, severe serrations occur in the flow curvesof NBH samples, whereas little serrations are observed inthe CBA specimen (Table 4). The serrations in the NBHsamples are characterized by fine serrations superimposedcoarse stress drops. The fine serrations are due to dynamicstrain aging (DSA), which is related to nitrogen dissolvedin the NBH specimen. The coarse stress drops may be asso-ciated with shearing of some nitride particles or by inhomo-geneous nitrogen concentration.

The studies on DSA in iron and steel by Keh et al.42) in-dicate that the serration phenomena in a low carbonrimmed steel (with alloying elements of 0.035% C,0.36Mn, 0.006P, 0.016S, 0.005Si, 0.005Cu, 0.002Ni,0.008Cr, 0.003N) (LCRS) appear at 68°C, and most severe-ly occur at 162°C at a tensile strain rate of 4.23�10�5 s�1.The carbon concentration in the present IF steel is muchlower than that of LCRS, but the amount of nitrogen is sim-ilar to LCRS. The weak serration behavior in the CBAspecimen may be caused by a little dissolved (unscavengedby Ti and Al) carbon and nitrogen solutes.

According to the temperature dependence of diffusivitiesof carbon and nitrogen, and the values of activation energyfor diffusion of carbon and nitrogen in a-Fe,1) the diffusioncoefficients of carbon and nitrogen at 70°C are calculated tobe 1.316�10�18 and 1.760�10�18 m2 s�1, respectively, indi-cating that the diffusion velocity of nitrogen in a-Fe isslightly higher than that of carbon. Since a substantial DSAcan occur at a carbon level of around 0.002 wt%, and as lit-tle as 0.001–0.002 wt% N can result in severe strain aging,1)

the serrated flow of the present IF steel at 70°C is caused bynitrogen solute.

In contrast to the classical model of solute–dislocationinteractions, the serrated flow is also believed to be directlycaused by the shearing of precipitates,43) i.e., by the interac-tions of precipitates with mobile dislocations. In light of thefact that the same amount of fine particles, such as TiN,TiC, and AlN, are present in both the CBA and NBH spec-imens, the differences in serration amplitude between theCBA and NBH specimens must be caused by dissolved ni-trogen.

Since the recrystallization may be prevented in the NBHspecimens as discussed before, their strengths may not re-flect a pure nitriding effect. Therefore, the hardness valuesof 75% cold-rolled, 1 mm thick IF steels before and afterheating for 3 h in the nitrate and chloride baths at 650°Cwere measured and compared in Fig. 12. The nitrided spec-imen has higher hardness than the as-rolled specimen.

3.3. Nitriding Kinetics

Figure 13 shows the surface hardness of the full-an-nealed IF steel samples after nitriding for 1 to 16 h at tem-peratures of 560 to 640°C. The surface hardness increaseswith nitriding time and temperature. However, for the sam-ples nitrided at higher temperatures of 620 and 640°C, thehardness increases to a peak value and decreases, with timeto reach the peak decreasing with increasing temperature.These phenomena are attributable to increases in nitrogenconcentration and grain size with increasing nitriding tem-perature and time. The increase in interstitial-nitrogen con-centration raises the hardness, while the formation of Fe4Ndecreases the hardness.44) The decreases in the surfacehardness of the specimens nitrided at 620 and 640°C seemto be caused by the formation of Fe4N as discussed in theprevious section.

Figure 14 shows the hardness profiles through the thick-ness of the full-annealed IF steel samples after nitriding for1 to 16 h at 560 to 640°C. The hardness decreases with in-creasing distance from the surface and decreasing tempera-ture. For the samples nitrided for 1 to 16 h at 640°C, theoverall hardness level increases with annealing time, espe-cially for the samples nitrided for 8 and 16 h, and is muchhigher than that of the samples nitrided at other tempera-tures, implying the higher nitrogen concentration through-out the samples. Therefore, we may achieve a relativelyuniform nitrogen-concentration profile through the thick-ness, if the nitriding time is sufficiently long.

ISIJ International, Vol. 46 (2006), No. 1

117 © 2006 ISIJ

Fig. 12. Vickers hardness values (load: 1 kg) of 75% cold rolled,1 mm thick IF steel sheets before and after heating in650°C KNO3 and CaCl2–NaCl baths for 3 h.

Fig. 13. Vickers hardness (load: 2 kg) of surface of full-annealedIF steel as a function of nitriding time in KNO3 salt bathat various temperatures.

Table 4. Serration characteristics of CBA and NBH speci-mens.

The data in Figs. 13 and 14 and results in Sec. 3.2 indi-cate that the hardness data are related to the nitrogen diffu-sion. Tong et al.4) found the coincidence in the profileshapes between nitrogen concentration and hardness alongthe nitrided depth. Simmons45) observed a linear relationbetween the tensile strength and the nitrogen concentrationdissolved in a austenitic stainless steel. Therefore, we de-fine a nitriding depth as the distance between the surfaceand the position where the hardness value is equaled by10% of the difference between the surface hardness and theinitial hardness of the center of sample plus the initial hard-ness of the center of sample.

From the hardness data in Fig. 14, we obtain the nitridingdepth data: 1.067, 1.261 and 1.385 mm for nitriding for 8 hat 580, 600, and 620°C, respectively, and 1.374 mm for ni-triding for 16 h at 560°C. The data at higher temperaturesare excluded in this kinetics of nitriding to avoid complexi-ty due to the formation of nitrides and due to the limit ofspecimen thickness.

Since the thickness of IF steel samples concerned is larg-er than the nitriding thickness, and the quantity of the ni-trate bath is high enough for the generated nitrogen concen-tration to be constant, the nitrogen diffusion during nitrid-ing can be described by the following equation46):

..................(8)

where C(x, t) represents the concentration at depth x aftertime t, Cs the concentration at the surface (x�0), and D thediffusion coefficient. The expression erf(x/2øDt

—) is the

Gaussian error function. Based on the present definition ofnitriding depth, setting C(x, t)�Cs/10, we obtain

..............................(9)

For nitriding, x, t, and DN are the nitriding depth andtime, and the effective diffusion coefficient of nitrogen, re-spectively. The DN value may obey an Arrhenius-type equa-tion:

.......................(10)

where D0 is the temperature-independent pre-exponentialfactor, Q is the activation energy for diffusion, R�8.3143 (Jmol�1 K�1) is the ideal gas constant, and T is the absolutetemperature (K).

The DN values calculated using the nitriding depth, time,and temperature data are plotted in Fig. 15. The DN–T datafit the Arrhenius equation. It follows from Fig. 15 that

....(11)

The activation energy, Q�76.62 kJ mol�1, is in good agree-ment with the activation energy for diffusion of nitrogen ina-Fe, 77.8 kJ mol�1, which Fast and Verrijp47) obtained attemperatures of 500–600°C, and 76.2 kJ mol�1, whichWert48) obtained over a much smaller temperature range.The pre-exponential factor, D0�3.789�10�7 m2 s�1, is sim-ilar to 3�10�7 m2 s�1 48) and a little lower than 6.6�10�7 m2

s�1.47) Therefore, we conclude that the diffusion behavior ofnitrogen for the KNO3-nitriding of the IF steel sample is similar to that for the nitrogen diffusion in a-Fe at temperatures below 620°C. It is also noted that Przyleskiand Maladazinski49) obtained the self-diffusion coeffi-cient of nitrogen in a-Fe, DN

(a)*�6.6�10�7· exp (77.9/RT)m2 s�1.

The nitrogen concentration in a layer from 50 to 450 mmin depth of the full-annealed steel specimen after nitridingat 650°C for 3 h was measured to be 0.0556% N. We wantto calculate the nitrogen concentration using Eq. (8). Thediffusion coefficient of nitrogen in the sample at 650°C iscalculated from Eq. (11) to be 1.748�10�11 m2 s�1. Thesolubility of nitrogen in a-Fe at 650°C is about 0.09%.41)

This concentration is assumed to remain constant andequivalent to Cs in Eq. (8). The average nitrogen concentra-tion in the layer from 50 to 450 mm in depth is calculatedusing the following trapezoidal rule.50)

DRTN

2(m s� ��

� �3 789 1076 627 1. exp

.)

D DQ

RTN � �0 exp

x D t�2 32. N

C x t Cx

Dt( , )� �s erf1

2

ISIJ International, Vol. 46 (2006), No. 1

© 2006 ISIJ 118

Fig. 14. Vickers hardness (load: 1 kg) as a function of distancefrom surface for full-annealed IF steel after nitriding inKNO3 salt bath. (1) 640°C for 16 h, (2) 640°C for 8 h,(3) 640°C for 4 h, (4) 640°C for 2 h, (5) 640°C for 1 h,(6) 620°C for 8 h, (7) 600°C for 8 h, (8) 580°C for 8 h,(9) 560°C for 16 h, (10) 560°C for 8 h.

Fig. 15. Arrhenius-type plot for diffusion coefficient of nitrogenin full-annealed IF steel, DN, during nitriding in KNO3

salt-bath. Activation energy is determined from slope ofstraight line.

.......(12)

where the depth interval Dx�10 mm, the number of inter-vals n�40, x0�0 mm, xn�400 mm were used, and c(xi, t)was calculated using Eq. (8) with Cs�0.09%, D�1.748�10�11 m2 s�1, t�3 h. The average nitrogen concen-trationis calculated to be 0.0674%, which is comparablewith the measured one, 0.0556%. This analysis implies thatEq. (11) is applicable to nitriding of IF steel or a-iron.

It was pointed out in Sec. 3.1 that steel samples formedoxide scales of about 154 mm (77 mm each side) and 35 mm(17.5 mm each side) in thickness after nitriding for 16 h at640°C and 580°C, respectively. Under the same nitridingconditions, the nitriding thickness is calculated from Eqs.(11) and (9) to be about 2.21 mm at 640°C and 1.55 mm at580°C. The nitriding thicknesses are much larger than theoxide scale thicknesses.

Various nitriding results are compared in Table 5. It canbe seen that the nitriding layer in this study saliently thick-en compared to other methods. It should be noted that thenitriding thickness from the references is the distance fromthe surface to the location where the hardness value (or ni-trogen concentration) reduces to the core hardness (or nitro-gen concentration). In addition, bubbling of gaseous prod-ucts, such as NO and NO2, was not observed during nitrid-ing. Therefore, this nitriding method is environment-friend-ly.

From a technical viewpoint, therefore, the KNO3 salt-bath nitriding can be as a pure nitriding of iron and steelsthat can be used in industrial applications, because it is easyto control the nitrogen balance in the steel coupon and the

processing optimization, along with simple apparatus,cheap salt, and non-toxicity.

4. Conclusions

Potassium nitrate was successfully used for pure nitridingof steel in salt bath, for the first time. Nitriding of IF steelspecimens in the KNO3 salt-bath at 400 to 650°C lead tothe following conclusions.

Most of nitrogen resides in steel as a form of interstitialsolid-solution. The hardness of the steel specimens increas-es with increasing nitriding time and temperature due to in-creasing nitrogen concentration. However, the nitridingprocess with too high temperatures and longer time can de-crease the surface hardness due to the formation of soft ni-tride. The hardness decreases with increasing distance fromthe surface due to gradient profiles of the nitrogen concen-tration. The nitriding kinetics is controlled by diffusion ofnitrogen.

The KNO3 bath nitriding gives rise to a much thicker ni-triding depth than existing nitriding methods at similar ni-triding temperature and time. A pronounced solid-solution-strengthening effect is therefore obtained, i.e., the 1 mm-thick nitrided IF steel specimen shows a tensile strength of916 MPa and an elongation of 20% at 70°C after nitridingfor 1.5 h at 650°C. Thus, the KNO3 bath can be used for ni-triding of steels to achieve their solid-solution strengthen-ing.

During nitriding of steel, the oxide scale forms, but theoxide thickness is only a few percents of the nitriding thick-ness. For nitriding at high temperatures of 640°C, nitrideprecipitates (g�-Fe4N and z-Fe2N) seem to form mostly ingrain boundaries.

Acknowledgement

One of the authors (Shen) acknowledges financial sup-port from BK21 Materials Education and ResearchDivision, Seoul National University.

REFERENCES

1) R. W. K. Honeycombe: Steels—Microstructure and Properties,Edward Arnold, London, (1981), 1.

2) M. L. G. Byrnes, M. Grujicic and W. S. Owen: Acta Metall., 35(1987), 1853.

3) E. Werner: Mater. Sci. Eng. A, A101 (1988), 93.4) W. P. Tong, N. R. Tao, Z. B. Wang, J. Lu and K. Lu: Science, 229

(2003), 686.5) M. A. J. Somers and E. J. Mittemeijer: Metall. Mater. Trans. A, 26

(1995), 57.6) J. Baranowska and M. Wysiecki: Surf. Coat. Technol., 125 (2000),

30. 7) L. Torchane, P. Bilger, J. Dulcy and M. Gantois: Metall. Mater.

Trans. A, 27 (1995), 1823.8) S. S. Hosmani, R. E. Schacherl and E. J. Mittemeijer: Acta Mater.,

53 (2005), 2069.9) J. Walkowicz, J. Smolik and K. Miernik: Surf. Coat. Technol.,

116–119 (1999), 361.10) F. Ashrafizadeh: Surf. Coat. Technol., 173–174 (2003), 1196.11) K. Kurosawa, H. L. Li, Y. Ujihira, K. Nomura, E. Mochizuki and H.

Hayashi: Mater. Charact., 34 (1995), 241.12) E. J. Miola, S. D. De Souza and M. Olzon-Dionysio: Surf. Coat.

Technol., 167 (2003), 33.13) L. C. Gontijo, R. Machado, E. J. Miola, L. C. Casteletti and P. A. P.

Nascente: Surf. Coat. Technol., 183 (2004), 10.14) S. C. Mishra, B. B. Nayak and B. C. Mohanty: Surf. Coat. Technol.,

� � � ��2 22 1c x t c x t c x tn n( , ) ( , ) ( , )]L

C

c x t dx

n x nc x t c x t

x

xn

� � �01

220 1

∫ ( , )

[ ( , ) ( , )D

ISIJ International, Vol. 46 (2006), No. 1

119 © 2006 ISIJ

Table 5. Comparison between nitriding-thickness data mea-sured in full-annealed IF steel and data from variousreferences.

145 (2001), 24.15) Y. Sun and T. Bell: Mater. Sci. Eng. A, A224 (1997), 33.16) G. W. Malaczynski, C. H. Leung, A. A. Elmoursi, A. H. Hamdi, A.

B. Campbell, M. P. Balogh, M. C. Militello, S. J. Simko and R. A.Waldo: Mater. Sci. Eng. A, A262 (1999), 289.

17) M. K. Lei and Z. L. Zhang: Surf. Coat. Technol., 91 (1997), 25.18) C. Blawert, B. L. Mordike, U. Rensch, R. Wunsch, R. Wiedemann

and H. Oettel: Surf. Coat. Technol., 131 (2000), 334.19) G. Schreiber, U. Rensch, H. Ottel, C. Blawert and B. L. Mordike:

Surf. Coat. Technol., 167–170 (2003), 447.20) A. Mitsuo, S. Uchida and T. Aizawa: Surf. Coat. Technol., 186

(2004), 196.21) P. Schaaf: Prog. Mater. Sci., 47 (2002), 1.22) P. Schaaf, A. Emmel, C. Illgner, K. P. Lieb, E. Schubert and H. W.

Bergmann: Mater. Sci. Eng. A, A197 (1995), L1.23) F. Landry, K-P. Lieb and P. Schaaf: J. Appl. Phys., 86 (1999), 168.24) E. Carpene and P. Schaaf: Appl. Surf. Sci., 186 (2002), 100.25) C. J. Copola, I. Avram, M. C. Terzzoli, S. Duhalde, C. Morales, T.

Perez, F. Audebert, Ph. Delaporte and M. Sentis: Appl. Surf. Sci.,197–198 (2002), 896.

26) P. J. Wilbur and B. W. Buchholtz: Surf. Coat. Technol., 79 (1996), 1.27) S. B. Ogale, A. Polman, F. O. P. Quentin, S. Roorda and F. W. Saris:

Appl. Phys. Lett., 50 (1987), 138.28) C. M. Kramer, Z. A. Munir and J. V. Volponi: Thermochim. Acta, 55

(1982), 11. 29) S. Gordon and C. Campbell: Anal. Chem., 27 (1955), 1102.30) E. S. Freeman: J. Am. Chem. Soc., 79 (1957), 838.31) T. M. Oza: J. Indian Chem. Soc., 22 (1945), 173.32) V. J. Szper and K. Z. Fiszmon: Anorg. Allg. Chem., 206 (1932), 257.33) T. M. Oza and S. A. Patel: J. Indian Chem. Soc., 31 (1954), 520.34) K. H. Stern: J. Phys. Chem. Ref. Data, 1 (1972), 747.35) E. S. Freeman: J. Phys. Chem., 60 (1956), 1487.

36) Y. Inokuti, N. Nishida and N. Ohashi: Metall. Trans. A, 6A (1975),773.

37) J. Walkowicz, J. Smolik and K. Miernik: Surf. Coat. Technol.,116–119 (1999), 361.

38) J. A. Shackelford: Introduction to Materials Science for Engineers,4th ed., Prentice Hall International, Inc., New Jersey, (1996), 624.

39) R. C. Weast: Handbook of Chemistry and Physics, 58th ed., CRCPress, Boca Raton, Fla., (1977).

40) H. A. Wriedt and K. Zwell: Trans. Metall. Soc. AIME, 224 (1962),1242.

41) H. A. Wriedt, N. A. Gokcen and R. H. Nafziger: Binary Alloy PhaseDiagrams, 2nd ed., ed. by T. B. Massalski, H. Okamoto, P. R.Subramanian and L. Kacprzak, ASM, Metals Park, Ohio, (1990),1729.

42) A. S. Keh, Y. Nakada and W. C. Leslie: Dislocation Dynamics, ed.by A. R. Rosenfield, G. T. Hahn, A. L. Bement, Jr. and R. I. Jaffee,McGraw-Hill, New York, (1968), 381.

43) F. Chmelík, E. Pink, J. Król, J. Balík, J. Pesikca and P. Lukác: ActaMater., 46 (1998), 4435.

44) C. Blawert, B. L. Mordike, U. Rensch and H. Oettel: Surf. Coat.Technol., 142–144 (2001), 376.

45) J. W. Simmons: Unpublished result, U. S. Bureau of Mines, Albany,OR 97321.

46) P. Shewmon: Diffusion in Solids, 2nd ed., TMS, Warrendale,Pennsylvania, (1989), 22.

47) J. D. Fast and M. B. Verrijp: J. Iron Steel Inst., 176 (1954), 24.48) C. A. Wert: J. Appl. Phys., 21 (1950), 1196.49) Z. Przyleski and L. Maladazinski: Proc. of 4th Int. Conf. on

Carbides, Nitrides and Borides, Pozman Kolobrzeg, Poland, (1989),153.

50) G. B. Thomas, Jr and R. L. Finney: Calculus and Analytic Geometry,8th ed., Addison-Wesley, Don Mills, Ontario, (1992), 298.

ISIJ International, Vol. 46 (2006), No. 1

© 2006 ISIJ 120