Scripta Materialia 53 (2005) 1345–1349
www.actamat-journals.com
Nitriding of steel in potassium nitrate salt bath
Y.Z. Shen, K.H. Oh, D.N. Lee
*
School of Materials Science and Engineering, Seoul National University, Shinrim-dong, Gwanak-gu, Seoul 151-744, Republic of Korea
Received 24 January 2005; accepted 26 August 2005
Abstract
A potassium nitrate salt bath has been used for nitriding of interstitial-free steel. The nitriding behavior can be reasonably well
described by nitrogen diffusion in iron. Most nitrogen is dissolved interstitially. During nitriding, a slight oxidation of the steel surface
also takes place. The nitrided specimen achieves a pronounced solid–solution strengthening.
2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Potassium nitrate; Nitriding; IF steel; Interstitial diffusion; Dynamic strain aging
1. Introduction
Nitriding is a surface treatment technique used to introduce nitrogen into metallic materials to improve their
surface hardness, mechanical properties, wear and corrosion resistance, as well as fatigue life. Established nitriding
methods include gas nitriding, plasma nitriding, laser
nitriding, reactive magnetron sputtering and nitrogen
implantation, and plasma immersion ion implantation [1].
Although the above methods are well established, some
of them have disadvantages from an engineering viewpoint,
for example, they may require the use of rather complicated and/or expensive apparatus. One of the traditional
and most commonly applied nitriding methods for steel
parts is salt bath nitriding by means of liquid salts containing cyanide and cyanate. In fact, this process is actually a
nitrocarburizing process [2], since the environment of molten salt contains both carbon and nitrogen and the two
elements generally diffuse into the surface of steel parts,
simultaneously. To our knowledge, up to the present no
nitriding process for steel without simultaneous carburization has been realized with salt baths.
The reported decomposition temperatures of potassium
nitrate (KNO3) are about 527–567 C [3] and 628 C [4] ob*
Corresponding author. Tel.: +8228807093; fax: +8228876388.
E-mail address: [email protected] (D.N. Lee).
tained by differential scanning calorimetry and differential
thermal analysis techniques, respectively, and potassium
nitrate may directly decompose to oxide (K2O) [5–7]
according to the following reactions (1) or (2),
2 KNO3 ! K2 O þ 2 NO2 þ 1=2 O2
ð1Þ
2 KNO3 ! K2 O þ 5=2 O2 þ N2
ð2Þ
Therefore, the thermal decomposition of KNO3 on heating
may liberate nascent nitrogen before the formation of
molecular nitrogen (N2). The nascent nitrogen can diffuse
into the steel coupon.
The objective of this study was to investigate the possibility of nitriding steel in a KNO3 salt bath.
2. Experimental procedure
A 4 mm thick hot-rolled interstitial-free (IF) steel provided by POSCO, Korea was used in this study. Its chemical composition is given in Table 1. Vickers hardness test
specimens of 2.8 mm thickness produced from IF steel
sheets cold rolled to a thickness reduction of 30% were annealed for 1–300 min at 650 C in two kinds of salt baths,
the KNO3 bath (the nitrate bath) and the 67% CaCl2–
33% NaCl bath (the chloride bath), to provide two different
atmospheres, namely a possibly nitriding atmosphere and
an inert atmosphere. After annealing, the specimens were
1359-6462/$ - see front matter 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.scriptamat.2005.08.032
0.037
0.027
0.050
0.002
0.008
water-quenched to remove surface scales. The scale-free
specimens were slightly ground on 2000-grit SiC emery
paper to get a bright surface before hardness test. Vickers
hardness tests were carried out with an Instron Wolpert
Tester 930 at a load of 5 kg and a holding time of 15 s.
The reported surface hardness of each specimen is the average of 10 measurements.
The surface scales were collected and pulverized and
then subjected to X-ray diffraction (XRD) to identify their
structures. Sheet specimens were polished before carrying
out the XRD scan. XRD measurements were performed
using a MAC Science Co. M18XHF-SRA diffractometer
with Cu Ka1 radiation (k = 0.154056 nm). The instrument
was set up for Bragg–Brentano geometry with a line focus
and a graphite monochromator in diffracted beam arm.
The slit configuration was as follows: divergence slit = 1;
receiving slit = 0.3 mm; scatter slit = 1. XRD patterns
were scanned in steps of 0.02 (2h), in the 2h range from
20 to 119. Silicon (a-Si) was used as an external standard.
Optical micrographs were taken from 3% nital–etched samples using an Olympus PMG3 microscope.
The nitrogen and oxygen concentrations in a 0.4 mm
thick surface-layer and a 0.5 mm thick center-layer of the
sample taken from the hot-rolled sheet after annealing
for 3 h in the KNO3 bath at 650 C were measured using
a LECO TCH 600 Nitrogen/Oxygen/Hydrogen Determinator (with an experimental error of less than 0.1 ppm).
The 2.8 mm thick sheet was further rolled to about
1 mm thickness to make tensile specimens with gauge
dimensions of 1 · 6 · 25 mm, in accordance with ASTM
E 8M-99, with the tension axis parallel to the rolling direction. One specimen was heated in the nitrate bath and another one was heated in the chloride bath at 650 C for 3 h.
They were subjected to tensile tests on an Instron-5582 at
an initial strain rate of 4.23 · 105 s1 at 70 C.
3. Results and discussion
Fig. 1 shows the XRD spectrum of a powder sample of
the surface scale removed from the surface of the 2.8 mm
thick IF steel annealed for 3 h at 650 C in the nitrate bath.
The XRD result indicates that the powder sample consisted
of magnetite (Fe3O4), hematite (Fe2O3), and wüstite (FeO).
Therefore, oxygen must have been generated in the nitrate
bath during annealing by decomposition of KNO3 according to the reactions (1) and/or (2). The sample underwent a
loss of about 0.05 mm in thickness on each side due to the
surface oxidation upon annealing in the nitrate bath for 3 h
at 650 C.
Fig. 2 shows the Vickers hardness of the surface of the
2.8 mm thick IF steel as a function of annealing time at
650 C in the nitrate and chloride baths. The surface scales
800
600
400
200
80
90
FeO(420)
1.23
Fe2O3(413)
<0.003
Fe2O3(2110)
S
Fe3O4(533)
N
Fe2O3(226)
FeO(400)
P
Fe2O3(1012)
FeO(311)
Ti
Fe3O4(220)
Al
0
20
30
40
50
60
70
100 110 120
2θ (deg)
Fig. 1. XRD spectrum of powder from surface of IF steel sample
annealed in KNO3 bath at 650 C for 3 h.
400
(a)
350
Vickers Hardness
Mn
Intensity (cps)
C
Fe2O3(116)
Fe3O4(422), Fe2O3(018)
1000
Fe3O4(400), Fe2O3(202)
Table 1
Chemical composition (wt.%) of IF steel
Fe3O4(440), Fe2O3(214), FeO(220)
Y.Z. Shen et al. / Scripta Materialia 53 (2005) 1345–1349
Fe3O4(311), Fe2O3(110), FeO(111)
1346
(b)
300
250
200
150
100
50
0
50
100
150
200
250
300
Annealing Time (min)
Fig. 2. Vickers hardness of surface of 30% cold-rolled IF steel as a
function of annealing time at 650 C in (a) KNO3 and (b) CaCl2–NaCl
baths.
of the specimens annealed in the nitrate bath were removed
before hardness testing. For the sample annealed in the
nitrate bath, the hardness drops a little due to recovery
and then increases with increasing time. On the other hand,
for the chloride-bath-annealed (CBA) sample, the hardness
first decreases to a minimum value due to recovery and
slightly increases again to a maximum peak, followed by
a decrease with increasing time as expected when recrystallization and subsequent grain growth take place (Fig. 3).
The weak hardness peak in the CBA sample may be an
experimental error or due to strain aging (not dynamic
strain aging). This needs further study. However, the two
different hardness evolutions at the later stage of annealing
suggest that different reactions take place in the nitrate and
chloride-bath-annealed samples.
The unusual increase in the hardness of the nitrate-bathannealed (NBA) sample was supposed to be related to
diffusion of nitrogen into the steel sample. If the nitrogen
diffusion occurred in the sample, a hardness gradient is to
be expected. The hardness distribution of the NBA sample
is shown in Fig. 4. Indeed, the hardness decreases with
Y.Z. Shen et al. / Scripta Materialia 53 (2005) 1345–1349
1347
Table 2
Concentration (wt.%) of nitrogen and oxygen in IF steel after annealing in
KNO3 bath at 650 C for 3 h
400
α-Fe (200)
Fig. 3. Optical micrographs of 30% cold-rolled IF steel after annealing in
67% CaCl2–33% NaCl bath at 650 C for 0 and 150 min (from left).
Intensity (a.u.)
1
(b)
3
2
α-Fe (310)
0.0071
0.0041
α-Fe (220)
O
0.0556
0.0064
α-Fe (211)
N
0.4 mm surface layer
0.5 mm center layer
α-Fe (110)
Location
4
5
Vickers Hardness
350
(a)
300
250
20
200
40
60
80
2θ (deg)
100
120
Fig. 5. XRD spectra of IF steel (a) before and (b) after annealing in
KNO3 bath at 650 C for 3 h.
150
100
50
0
500
1000 1500 2000 2500 3000 3500
Distance From Surface (µm)
Fig. 4. Vickers hardness distribution of IF steel specimen annealed in
KNO3 bath at 650 C for 3 h.
increasing distance from the surface. In order to prove the
presence of nitrogen in the sample, a chemical analysis was
performed. The results in Table 2 indicate that the nitrogen
concentration of the surface layer is almost 10 times higher
than that of the center layer. The oxygen concentration
varies little with depth, indicating that most oxygen is consumed to form oxides at the surface. It is quite clear that
nitriding took place in the sample.
In order to investigate the state of nitrogen in the sample, XRD analyses of samples were performed before and
after annealing in the nitrate bath at 650 C for 3 h. The results are shown in Fig. 5. The lattice parameters calculated
from the XRD data are given in Table 3. The diffraction
pattern (Fig. 5) does not contain reflections of iron nitrides.
However, as the nitrogen concentration is rather small (cf.
Table 2), and thus, if nitrides had been formed, the volume
fraction of nitrides would therefore also be small, X-ray
diffraction phase analysis may not be sensitive enough to
detect the presence of iron nitrides. On the other hand,
the values in Table 3 indicate that the lattice parameter
of the annealed sample is slightly higher than that of the
unannealed sample. This finding suggests that most nitrogen occupies interstitial sites in the annealed sample. The
differences in peak intensities of the sample before and after
annealing are due to a texture change after annealing.
We estimate the nitrogen concentration at the surface of
the NBA sample using the lattice parameter data in Table
3. According to Wriedt and Zwell [8], the unit cell parameter of a iron increases linearly by 3.2 · 103 nm per
wt.% N dissolved. If the lattice-parameter expansion after
nitriding (2.3 · 104 nm from Table 3) is assumed to be
caused by the nitrogen intrusion, then the nitrogen concentration of the surface of the NBA sample is estimated to be
about 0.072 wt.%.
The concentration determination via the lattice expansion measured by X-ray diffraction implies that residual
stresses do not significantly affect measured diffraction line
positions. Though nitrogen up-take generally leads to the
generation of stresses, it is likely that no significant residual
stresses are present in the investigated specimens, as the
concentration of 0.072 wt.% determined from the measured
lattice expansion is compatible with the surface concentration estimated from the average concentration in a surface
layer (as determined by chemical analysis) on the basis of
nitrogen diffusion (see below).
Next, we want to check if the measured nitrogen concentration in Table 2 is reasonable in terms of diffusion.
Assuming that the diffusion behavior of nitrogen in the
nitrate-nitriding of the IF steel sample is the same as that
for the nitrogen diffusion in a-Fe, the diffusion coefficient
at 650 C can be calculated to be 2.605 · 1011 m2 s1 from
the following equation [9]:
Q
DaN ¼ 6:6 107 exp
ð3Þ
ðm2 s1 Þ
RT
ðaÞ
where DN is the diffusion coefficient of nitrogen in a-Fe,
Q = 77.8 kJ mole1 the activation energy for nitrogen
1348
Y.Z. Shen et al. / Scripta Materialia 53 (2005) 1345–1349
Table 3
Identification of XRD spectra and calculated lattice parameters a
Sample
Peak
2h
(hkl)
Int
I/I0
a (nm)
a (nm)
IF steel before annealing
1
2
3
4
5
44.68
64.94
82.26
98.92
116.08
110
200
211
220
310
6333
1160
2569
185
253
100
18.3
40.5
2.9
3.9
0.28659
0.28696
0.28684
0.28669
0.28710
0.28684
IF steel after annealing in KNO3 bath at 650 C for 3 h
1
2
3
4
5
44.62
64.86
82.16
98.80
116.12
110
200
211
220
310
6233
165
574
359
342
100
2.6
9.2
5.7
5.4
0.28694
0.28728
0.28713
0.28694
0.28704
0.28707
where C(x, t) and Cs represent the concentrations of diffusion species at depth x after time t and at x = 0, respectively, and
pffiffiffiffiffi D is the diffusion coefficient. The expression
erfðx=2 DtÞ is the Gaussian error function.
We calculate the surface concentration Cs from the mea
sured average nitrogen concentration Cð¼
0:0556 wt.%Þ of
the 0.4 mm surface layer and C(x, t) calculated using Eq.
can be obtained by the trapezoi(4). The expression for C
dal rule [11].
R xn
cðx; tÞ dx
x
ð5Þ
C¼ 0
nDx
with the depth interval Dx = 0.01 mm, the number of intervals n = 40, x0 = 0, xn = 0.4 mm, D = 2.605 · 10 11 m2
s1, and t = 3 h. Then, we obtain Cs = 0.070 wt.% for
¼ 0:0556%. The Cs value is very close to the 0.072
C
wt.% obtained from XRD measurements.
Abe [12] has reported that the substitutional–interstitial
dipoles are formed in the alpha solid solution of iron if the
first-order interaction coefficient between a substitutional
solute and an interstitial solute is a negative value, and suggested that Mn–N dipoles form in the alpha solid solution
of iron, due to a negative interaction coefficient of N with
Mn. Therefore, the formation of Mn–N dipoles may affect
the result for the determination of the nitrogen content.
However, the good agreement between the Cs value and
the surface concentration estimated by the XRD measurements suggests that, if the formation of Mn–N dipoles
1000
IF Steel
o
Cold-rolled + 650 C x 3 h
(1) In KNO3 Bath
(2) In CaCl2-NaCl Bath
800
Stress (MPa)
diffusion, R = 8.3143 (J mol1 K1) the ideal gas constant,
and T the absolute temperature (K).
Since the nitriding thickness of IF steel sample nitrided
at 650 C for 3 h can be estimated
as x = 0.53 mm accordpffiffiffiffiffiffiffiffi
ing to the equation x ¼ DN t [10] with effective diffusion
coefficient of nitrogen DN = 2.605 · 1011 m2 s1 and
nitriding time t = 3 h, which is obviously smaller than the
half of thickness of IF steel sample, and the quantity of
KNO3 in the nitrate bath is high enough for the generated
nitrogen concentration to be constant on the surface of the
sample, the nitrogen diffusion during nitriding can be
described by the following equation [10]:
x
Cðx; tÞ ¼ C s 1 erf pffiffiffiffiffi
ð4Þ
2 Dt
-5
-1
o
Tension: 4.23 x 10 s , 70 C
600
(1)
400
200
(2)
0
0
5
10
15
20
25
30
35
40
Strain (%)
Fig. 6. Stress–strain curves of IF steel annealed for 3 h at 650 C in (1)
nitrate and (2) chloride baths at tensile strain rate of 4.23 · 105 s1 at
70 C.
occurs, it does not significantly affect either the amount
of interstitially dissolved nitrogen or its diffusion behavior.
The good agreement thus also confirms that the concentration of interstitially dissolved nitrogen determined via the
lattice expansion measured by XRD is not significantly
affected by the presence of a state of residual stress.
Tensile testing results of the CBA and NBA specimens
reveal a pronounced difference in the deformation behavior
as shown in Fig. 6. The flow stress of NBA specimen is
more than two times higher than that of the CBA specimen. Furthermore, severe serration (with an average
stress-drop of 18.5 MPa and a maximum stress-drop of
50.45 MPa) occurs in the NBA specimen, whereas very fine
serrations (with an average stress-drop of 1.05 MPa and a
maximum stress-drop of 2.3 MPa) are observed in the CBA
specimen. The severe serration implies dynamic strain
aging due to nitrogen in solid solution in the NBA specimen. These results suggest that a KNO3 bath can be used
for nitriding of steel to achieve solid–solution strengthening of steel.
4. Conclusions
The study of nitriding of IF steel specimens in a KNO3
salt bath for 3 h at 650 C led to the following conclusions.
Y.Z. Shen et al. / Scripta Materialia 53 (2005) 1345–1349
The nitriding behavior of IF steel in the KNO3bath can
be reasonably well described by nitrogen diffusion in iron,
with the surface concentration of nitrogen being about
0.07 wt.% at 650 C. Most nitrogen atoms occupy interstitial sites. The 1 mm thick nitrided specimen shows a
tensile strength of 890 MPa and an elongation to fracture
of 15% at 70 C. The KNO3 bath can thus be used for
nitriding of steel to achieve solid–solution strengthening
of steel.
Acknowledgement
This study was supported by the BK21 Materials Education and Research Division, and the Texture Control
Laboratory (NRL), Seoul National University.
1349
References
[1] Schaaf P. Prog Mater Sci 2002;47:1.
[2] ASM. Metals hand book. Metals Park, (OH): American Society for
Metals; 1978.
[3] Kramer CM, Munir ZA, Volponi JV. Thermochim Acta 1982;55:11.
[4] Gordon S, Campbell C. Anal Chem 1955;27:1102.
[5] Stern KH. J Phys Chem Ref Data 1972;1:747.
[6] Freeman ES. J Phys Chem 1956;60:1487.
[7] Freeman ES. J Am Chem Soc 1957;79:838.
[8] Wriedt HA, Zwell K. Trans Metall Soc AIME 1962;224:1242.
[9] Fast JD, Verrijp MB. J Iron Steel Inst 1954;176:24.
[10] Shewmon P. Diffusion in solids. 2nd ed. Warrendale, PA: TMS;
1989. p. 22.
[11] Thomas GB, Finney RL. Calculus and analytic geometry. 8th
ed. Don Mills, Ontario: Addison-Wesley; 1992. p. 298.
[12] Abe H. ICOTOM-8 Proc. Warrendale, PA: TMS; 1988. p. 661–6.