15. - 17. 5.

2013, Brno, Czech Republic, EU

THE EFFECT OF HEAT TREATMENT ON THE STRUCTURE OF SCALE LAYER AND ITS
REMOVING BY PICKLING IN REDUCING MELT

Petra VÁŇOVÁ1, Roman PĚNČÍK2, David ČEMPEL1, Kateřina KONEČNÁ1

1
VŠB – Technical University of Ostrava, Faculty of Metallurgy and Materials Engineering, 17. listopadu 15,
2
CZ 708 33 Ostrava, Czech Republic, Bochemie a.s., Lidická 326, Bohumín, [email protected]

Abstract

This paper deals with the structure and composition of scale layer of stainless steel AISI 316 and AISI 430
after heat treatment at temperatures of 800, 1000 and 1200 °C with holding time at temperature for 1 or 2
hours and after air and water cooling. Subsequently, pickling in reducing melt was carried out. The structure
of scale layers after heat treatment and after the following pickling in reducing melt, and its chemical
composition using X-ray microanalysis is described, the thickness of scale layer is measured and the weight
decrease after pickling in reducing melt is documented. After heat treatment, thicker and more compact
oxide layers were formed on the surface of the wire made of AISI 430 stainless steel than on the surface of
that made of AISI 316 steel. Oxide layers showed greater thickness after cooling in air than after cooling in
water. In the case of stainless steel AISI 316, the oxide layers contained relatively high amount of nickel,
which is rather difficult to eliminate by the action of reducing melt. As for AISI 430 steel, the layers are
formed on the iron oxide and chromium oxide basis. These layers are more easily distort when exposed to
reducing melt and they tend to crack and spall.
Keywords scale layer, oxide layer, heat treatment, steel pickling, reducing melt, X-ray microanalysis

1. INTRODUCTION

Scale layer is a thin oxide layer on the material surface, formed mainly during higher temperatures of heat
treatment. Formation of scales on the stainless steel surface depends on its composition (i.e. alloying
elements content) and also on the surface condition and the type of atmosphere, in which the steel is heat-
treated. The scales are formed by chemical composition of iron or other elements contained in steel with
oxygen from air or from the atmosphere of annealing oven. The estimation is that during steel annealing one
to five percent of material is changed into scale. The scale layer formed during steel heating at the
temperatures needed for its heat treatment is never chemically homogenous. It is made by two or three
stages with different composition and quality. The stages are separated from each other by quite acute
boundary. They are arranged as follows: the closer to the metal, the poorer in oxygen and the richer in iron.
This composition corresponds to the diffusion of iron and oxygen in the scale layer [1-3].
Iron and oxygen form together three stable oxides: iron (II) oxide FeO, magnetite Fe3O4 and iron (III) oxide
Fe2O3. Iron (II) oxide FeO is not invariable under temperatures lower than 575°C, because it slowly changes
into iron and magnetite Fe3O4. Therefore, the oxide layers formed under lower temperatures than mentioned
limit of 575°C do not contain FeO at all. It is important for pickling that iron (II) oxide FeO is easier to dissolve
in acids than other iron oxides. Thus the scale layers containing a great amount of iron (II) oxide are in
general easier to remove, while layers without this component are more difficult to remove, although they are
often much thinner. Magnetite Fe3O4 is the most invariable of the iron oxides and it does not change even in
a large range of temperatures. It is significantly much more difficult to dissolve in acids than iron (II) oxide.
Iron (III) oxide Fe2O3 is the iron oxide the richest in oxygen. It loses part of oxygen and turns into magnetite
during heating over 1200°C. Therefore, the scale layers formed under the temperatures higher than 1200°C
do not contain iron (III) oxide. Among the three oxides, this one is the most difficult to dissolve in acids [1-3].
 15. - 17. 5. 2013, Brno, Czech Republic, EU

Other elements, contained in steel, mainly the so called alloying elements do take part in scales forming
processes. The easiest proportion we find in elements with similar chemical properties as iron. Manganese is
the typical example. The rapidity of its infiltration into scale layer is approximately the same as in the case of
iron. The rapidity of its oxidation is also the same as in the case of iron and the manganese oxides are
spread out in scale layer uniformly and in the amount corresponding to the percentage of manganese in
relevant material. As for other alloying elements, the diffusion is slower than in the case of iron, which leads
to settling of these elements in the lowest layer of the coating. If these metals are nobler (nickel, copper),
iron takes away the oxygen from its oxides and oxide layer therefore contains these elements in the form of
pure metals. Other elements (chromium, silicon, manganese) are in the oxide layer in the form of oxides. If
the steel is highly alloyed with e.g. chromium, the lowest layer of its oxide layer is so enriched with this metal,
so that the chromium oxide forms a significant component on the steel-scale boundary. Another iron atoms
are not able to penetrate through this boundary, which gives reasons for heat resistance of highly alloyed
steel [1-5].
Steel pickling is a part of final production process of some steel products. During this the scales, formed
during metal heat treatment under high temperatures, are removed from the surface of steel belts, wires and
other forms of steel by dissolving in acids or melts. Chemical removing of scales is based mainly on their
dissolving in appropriate solution, sometimes with the use of electric current. In the case of carbon steel the
hydrochloric acid (HCl) or the sulfuric acid (H2SO4) is commonly used, while the combination of hydrofluoric
acid and nitric acid (HF+HNO3) is often used for stainless steel. Acids dissolve scales chemically and at the
same time the iron oxides turn into corresponding iron salts. For pickling in the reducing melt the melt of
sodium hydroxide (NaOH), containing sodium hydride (NaH) as reducing agent, is used. Sodium hydride is a
highly reducing agent, changing iron oxides into lower oxides and iron. A great advantage of pickling in
sodium hydride is that it does not damage the basic metal. Oxidation melts are efficient in action of oxidation
agent from dissolved sodium hydroxide (NaOH) on the oxide layer. Above-valent oxides with larger specific
capacity are formed during oxidation, which provokes changes of structure of scales on the surface of the
material, tension and following cracking. Only a small part of scales is separated or dissolved in the melt,
where it settles like sediment. Greater part of scales is removed only during following operation – cooling and
rinsing in water – while the scales are partially torn away. The rest of scales is removed by final pickling in
acids [6-8].

2. EXPERIMENT DESCRIPTION

Two types of steel with different chemical composition and structures, i.e. AISI 316 and AISI 430, hot drawn
wires, have been used for the experimental finding of the effect of heat treatment on the formation of scale
layer on stainless steels. AISI 316 is austenitic chromium-nickel-molybdenum steel, AISI 430 is ferritic
stainless steel with 17% of chromium content. Chemical composition of both steels is given in Table 1.
2
Individual samples in form of 100 mm long and 7 mm diameter wires (sample surface was 0,0022 m ) were
pickled in reducing melt and afterwards in mixed acid in order to make their surface clean and ready for
following laboratory heat treatment.
Tab. 1 Chemical composition of AISI 316 and AISI 430 steels in weight %

AISI C Si Mn P S Cr Mo Ni

316 0.05 0.65 1.50 0.04 0.03 17.00 2.00 12.00

430 0.14 0.80 1.10 0.06 0.20 17.00 0.45 -

Samples warmed to 800°C and 1000°C temperatures were left in the oven during two hours, samples
warmed to 1200°C temperature during one hour. The holding time was followed by air or water cooling.
 15. - 17. 5. 2013, Brno, Czech Republic, EU

Afterwards transverse metallographic cuts were prepared in order to find the character and structure of scale
layer after different heat treatments.
As for the second set of samples, the scales were removed by pickling in hydride reducing melt, to which the
®
reducing component (NaH) was dosed using a special product Feropur . Samples with scale layers after the
heat treatment were put into reducing melt for 15 minutes and afterwards cooled by a thermal shock due to
water immersion. When removed from pickling bath, the samples were rid of alkalinity by washing in hot
water and drying. The samples were weighed before and after pickling and their weight loss was
documented. The samples without scale layers were also metallographically prepared in transverse cuts and
they were observed using scanning electron microscope.
The scale layers were observed using JEOL JSM-6490LV scanning electron microscope in backscatter
electrons (BE). Semi quantitative x-ray microanalysis was carried out as surface analysis, using Inca X-act
detector. Evaluation of scale layer in transverse cut using scanning electron microscopy (SEM) included
documentation of the scale layer in different parts of the wire and semi quantitative x-ray microanalysis. As
for the carrying out of microanalysis, different stages of oxide layer and basic material to the depth of 20 to
40 m under the surface were evaluated. Carbon was not included in the semi quantitative x-ray
microanalysis.

3. RESULTS

3.1 Evaluation of wires surface after the heat treatment

During heat treatment at 800°C temperature with holding time for 2 hours a layer of iron oxide (57 % Fe, 1 %
Cr, 1-3 % Ni, 1-2 % Mn, 38 % O) was formed on the AISI 316 steel wire surface. After air cooling the layer
was cracked (Fig. 1), while after water cooling it was more compact (Fig. 2). Chromium, iron (nickel,
molybdenum, silicon) oxides (27-35 % Cr, 20-24 % Fe, 5-7 % Ni, 3.5-5.5 % Mo, 33-37 % O) were
separated under the layer. The material right under the oxide layer was not chromium depleted. A thin layer
of chromium and iron oxides (44 % Cr, 17 % Fe, 7 % Mn, 32 % O) appeared on the surface of AISI 430 steel
wires after air cooling (Fig. 3), while after water cooling (Fig. 4) it was a thin layer of iron and chromium
oxides (52 % Fe, 20 % Cr, 7 % Mn, 20 % O). Right under the surface the matrix was chromium depleted
(14.5 % Cr).

Fig. 1 Oxide layer – AISI 316 steel Fig. 2 Oxide layer – AISI 316 steel
Heat treatment: 800°C/2 hours/air Heat treatment: 800°C/2 hours/water
 15. - 17. 5. 2013, Brno, Czech Republic, EU

Fig. 3 Oxide layer – AISI 430 steel Fig. 4 Oxide layer – AISI 430 steel
Heat treatment: 800°C/2 hours/air Heat treatment: 800°C/2 hours/water
During heat treatment at 1000°C temperature with holding time for 2 hours and air cooling a broader layer
was formed on the AISI 316 steel wire surface. On the very surface it was formed by rough crystals of iron
oxide (55 % Fe, 1.5 % Cr, 0,5 % Ni, 43 % O). Under this layer more porous sublayer of iron and chromium
oxides (30% Cr, 28 % Fe, 36 % O, 4 % Ni, Mn, Si) was found. This layer did not contain molybdenum. Right
under the layer the chromium depletion (13 % Cr) has taken place, but on the other hand the nickel and
molybdenum content (17 % Ni, 5 % Mo) increased. In the 20 m depth under the layer the matrix was still
chromium depleted (15.8 % Cr). After water cooling the layer seemed to be more compact. On the surface
there was a thin sublayer of tiny iron oxides crystals (52 % Fe, 4 % Mn, 5 % Ni, 1 % Cr). Rough crystals in
the layer were formed by iron oxides (63 % Fe, 1 % Cr, 36 %). The composition of the porous layer
corresponded to chromium and iron oxides (34 % Cr, 20 % Fe, 5.5 % Ni, 1,4 % Mn, 39 % O, Si).
Molybdenum was not separated in the layer. On the scale layer –basic material boundary chromium oxides
without molybdenum presence were found (52 % Cr, 37 % Fe, 4 % Mn, 51 % Ni, 4 % O). The chromium
depletion of basic material under the layer was minimal. In the case of AISI 430 steel wire the oxide layer
after air cooling was formed by a consistent iron and chromium oxides layer (57 % Fe, 4 % Cr, 39 % O).
After water cooling the oxide layer was more disrupted and the closer to the matrix, the higher the chromium
content in the scale layer. Right under the layer the matrix was chromium depleted (9-11 % Cr).
During heat treatment at 1200°C temperature with holding time for 1 hour and air cooling was the layer on
the AISI 316 steel wire surface horizontally cracked (Fig. 5). Chromium, iron and molybdenum oxides (35 %
Cr, 23 % Fe, 3 % Mo 1 % Ni, 1 % Mn, 38 % O) formed the scale layer. These oxides were interspersed with
discontinuous structures on the pure nickel and iron base (60 % Ni, 35 % Fe, 3 % Cr, 2 %). Under the
consistent layer tiny oxides interfering into the matrix were found. The matrix was chromium depleted
(11,5 % Cr). After the water cooling the scale layer was more consistent and less porous (Fig. 6). On the
surface it consisted of iron and nickel oxides (55 % Fe, 7 % Ni, 1.5 % Cr, 3 % Mn, 35 % O). There were more
compact chromium, iron and nickel oxides structures (25 % Cr, 25 % Fe, 7 % Ni, 35 % O, without Mo) in the
layer content. Chromium, iron and nickel oxides with lower oxygen content (27 % Cr, 28 % Fe, 17 % Ni,
28 % O) formed the porous parts of the layer. White structures represented reduced metal Ni, Fe, Mo (53 %
Ni, 17 % Fe. 15 % Mo – the O and Cr content was influenced by environs). Right under the layer the basic
material was chromium depleted (9.5 % Cr). In the case of AISI 430 steel wire after air cooling a consistent
thick scale layer was formed (Fig. 7). At the very surface it consisted of iron and chromium oxides (35 % Fe,
27 % Cr, 1 % Mn, 36 % O). The deeper, the more the Cr content in the layer increased, the more the Fe
content decreased and the more the Mo content increased. The oxygen content decreased at 27 % and
molybdenum (1.2 % Mo) was analyzed in the interlayer. Right under the layer the matrix was Cr depleted
(11 % Cr). In a more detailed scale Cr and Fe oxides (33 % Cr, 23 % Fe, 1 % Mo, 40 % O) and white
 15. - 17. 5. 2013, Brno, Czech Republic, EU

structures of almost pure iron (94 % Fe, 3 % Cr, 1% Mo, 1% Ni) formed the interlayer. Air cooling led to
spalling of the scale layer. The rests of the spalled layer were analyzed separately. In the scales the layer
was bright, crystalline based on iron (66 % Fe, 33 % O), smooth formed of Fe oxides (63 % Fe, 36 % O),
which on the reverse contained except Fe oxides also Cr (61 % Fe, 6.5 % Cr, 1 % Mn, 30 %). After water
cooling the layer of Cr, Fe and Mo was evident on the surface (Fig. 8). Interlayer composition and basic
material under the layer depletion was similar as in the case of air cooling.

Fig. 5 Oxide layer – AISI 316 steel Fig. 6 Oxide layer – AISI 316 steel
Heat treatment: 1200°C/1 hour/air Heat treatment: 1200°C/1 hour/water

Fig. 7 Oxide layer – AISI 430 steel Fig. 8 Oxide layer – AISI 430 steel
Heat treatment: 1200°C/1 hour/air Heat treatment: 1200°C/1 hour/water

3.2 Evaluation of wires surfaces after pickling in reducing melt

Evaluation of scale layer after pickling in reducing melt was carried out by the same method as after heat
treatment, i.e. by using scanning electron microscopy (SEM) on the transverse cuts. Only the air cooled
samples were evaluated.
The oxides were present in surface defects in the case of AISI 316 steel samples, heat treated in
800°C/2 hours/air mode and subsequently pickled in reducing melt (Fig. 9). Calcium and sodium as remains
of the pickling process and molybdenum were analyzed in the oxides. The surface without oxide layer was
not Cr depleted. A thin layer of Cr and Fe oxides (27-36 % Cr, 19-23 % Fe, 2-3 % Ni, 35 % O, without Mo)
appeared in the case of samples heat treated in 1000°C/2 hours/air mode and subsequently pickled in
reducing melt. Right under the layer the Cr depletion (13.5 % Cr) was still apparent. In the case of samples
 15. - 17. 5. 2013, Brno, Czech Republic, EU

heat treated in 1200°C/1 hour/air mode and subsequently pickled in reducing melt (Fig. 10), the wire surface
showed more significant oxide remains, somewhere based on Fe and Ni (52 % Fe, 11 % Ni, 4 % Cr, 23 %
O), somewhere based on Fe and Cr (42 % Fe, 20 % Cr, 38 % O), or the original interlayer with presence of
pure nickel and iron.
The iron, chromium and manganese oxides (29 % Fe, 25 % Cr, 18 % Mn, 26 % O, without Mo) were present
in surface inequalities in the case of AISI 430 steel samples, heat treated in 800°C/2 hours/air mode and
subsequently pickled in reducing melt (Fig. 11). Calcium and sodium as remains of the pickling process were
analyzed in the oxides. The surface without oxide layer was slightly Cr depleted (14 % Cr). A thin layer of Cr
and Fe oxides (46 % Cr, 19 % Fe, 30 % O) appeared in the case of samples heat treated in 1000°C/2
hours/air mode and subsequently pickled in reducing melt. In some parts of the surface thin layers of pure
metal (66 % Fe, 18 % Cr, 11 % Ni) were apparent. Right under the layer the Cr depletion (13.8 % Cr) was
still apparent. In the case of samples heat treated in 1200°C/1 hour/air mode and subsequently pickled in
reducing melt (Fig. 12), the wire surface showed a broader layer which morphology corresponded to the
interlayer of annealed samples. There were Cr and Fe oxides (37 % Cr, 26 % Fe, 38 % O) on the surface.

Fig. 9 AISI 316 steel sample after pickling Fig. 10 AISI 430 steel sample after pickling
Original heat treatment: 800°C/2 hours/air Original heat treatment: 800°C/2 hours/air

Fig. 11 AISI 316 steel sample after pickling Fig. 12 AISI 430 steel sample after pickling
Original heat treatment: 1200°C/1 hour/air Original heat treatment: 1200°C/1 hour/air
 15. - 17. 5. 2013, Brno, Czech Republic, EU

3.3 Scale layer thickness and weight loss

As apparent from given figures (Fig. 1 – Fig. 8), the scale layer was thicker with higher temperatures of the
heat treatment. After samples air cooling the oxide layer grew thicker because of additional oxidation of the
surface, which often led to spalling of scales. From Fig. 13 it is evident that scale layer thickness on AISI 430
steel is bigger than in the case of AISI 316 steel despite the loss of a part of layer during air cooling.
The weight loss was determined by samples weighing before and after the pickling. The higher the
temperature of heat treatment was, the bigger was the weight loss of given material after the pickling in
reducing melt. After previous heat treatment at the same temperature the weight loss was lower in the case
of air cooling than in the case of water cooling. There were certain divergences in the case of samples
treated at the 1200°C temperature when during air cooling from the 1200°C temperature massive cracking
and spalling of scales from the material surface appeared, mainly in the case of AISI 430 steel. That led to
bigger weight loss. It is evident from Tables (Fig. 14) that there was bigger weight loss in the case of AISI
430 steel.

Fig. 13 Scale layer thickness Fig. 14 Weight loss after pickling

4. CONCLUSION

During the heat treatment of given stainless steel wires a oxide layer was formed on their surface. The
higher temperature was used, the thicker the scale layer was. The very surface of the scale layer contained
iron oxides, but approaching the matrix the chromium (Cr) fixed in oxides content was increasing. In high
annealing temperatures (1200°C) the molybdenum (Mo) was fixed in oxides too. In this temperature an
interlayer with pure metal particles based on nickel (Ni) and iron (Fe) in the case of AISI 316 steel or pure
iron only in the case of AISI 430 steel was formed between the matrix and the oxide layer. The layer seemed
to be more cracked after air cooling, it often spalled and was therefore thinner in the final state. On the
contrary, after water cooling the oxide layers stayed thicker and more compact. Chromium (Cr) depletion
right under the scale layer was more significant by AISI 430 steel.
The oxide layer decreased after pickling in reducing melt depending on its thickness in the after heat
treatment state. Thin layers formed under low temperature annealing were removed by pickling and
therefore the depletion on the material surface was the least possible. The remains of oxides stuck only in
deeper surface defects. After the pickling of layers formed under higher temperature annealing, in some
parts of the surface areas of reduced oxides or small layers of pure metals were observed. In the interlayer
area a higher amount of pure metal particles was evident.
In the case of AISI 316 steel, the oxide layers contained relatively high amount of nickel (Ni), which is quite
difficult to eliminate by the sodium hydride action. As for the AISI 430 steel, the layer is more complex, but
formed mainly by iron and chromium oxides. Such a layer is more easily distort by hydride action and it
cracks and spalls. The rest of scales is removed during final pickling in acids.
 15. - 17. 5. 2013, Brno, Czech Republic, EU

ACKNOWLEDGEMENT

The authors are grateful to the Ministry of Education of the Czech Republic for the financial support
of the project No. CZ.1.05/2.1.00/01.0040 “Regional Materials Science and Technology Centre” –
research activity “New sources of strength and toughness of materials for high technological
applications” within the frame of the operation program “Research and Development for
Innovations” financed by the Structural Funds and from the state budget of the Czech Republic.

LITERATURE
[1] RONEŠ, J.; JAROŠ, M.. Moření oceli a litiny. 1. vyd. Praha: SNTL- Nakladatelství technické literatury, 1955.
252 s.
[2] ŠTURC, J. a kol. Moření ocelí. 1. vyd. Praha: SNTL – Nakladatelství technické literatury, 1985. 260 s.
[3] RITUPER, R. Beizen von Metallen, 1. vyd. Weingarten; Eugen G. Leuze Verlag, 1993. 524 s. ISBN-10:
3874800784.
[4] YANG, C. W.; KIM, J. H.; TRIAMBULO, R., E.; et al.. The mechanical property of the oxide scale on Fe-Cr alloy
steels. Journal of alloys and compounds, 2013, vol. 549, p. 6-10. ISSN: 0925-8388.
[5] Matsuhashi, T; Okada, H; Kiya, S: Effects of Si, Mn contents on the descalability behavior of the scale of
annealed austenitic stainless steels. Tetsu to hagane-journal of the iron and steel institute of japan, 2004, vol. 90,
Is. 7, p. 487-493. ISSN: 0021-1575
[6] ULLRYCH, J.; TRŽIL, J.. Mechanismus odokujování ocelí v taveninové redukční lázni. In Moření ocelí v
redukčních taveninách. 26.04.1995. Frýdek-Místek: Ekomor; Bochemie; ZP HS Válcoven plechu, 1995. 31 s.
[7] LINDELL, D.; PETTERSSON, R.. Pickling of Process-Oxidised Austenitic Stainless Steels in HNO3-HF Mixed
Acid. Steel research international, 2010, vol.: 81, Is. 7. p. 542-551. ISSN: 1611-3683.
[8] LI, L. F.; CELIS, J. P.. Pickling of austenitic stainless steels (a review). Canadian metallurgical quarterly, 2003,
vol. 42, Is. 3, p. 365-376. ISSN: 0008-4433.