Addition of tin in 3% Si steel is a method to improve magnetic properties since tin in steel is known as an element that segregates at grain boundaries, inhibits grain growth and has a dragging effect on grain boundary movement. However, tin is generally known as an element that has a harmful effect on surface coating properties. The oxide subscale formed on the decarburization annealing of a 3% Si steel containing tin was investigated by SEM, EDS, GDS and FTIR. The forsterite film was evaluated by SEM, EDS and GDS. The higher tin content decreased thickness, oxygen and fayalite/silica ratio in the subscale and resulted into a discontinuous forsterite film with poor oxidation resistance. After secondary coating the samples presented a reddish color due to the formation of iron oxide and poor dielectric property.

The high permeability grain oriented silicon steel is widely used in transformers cores. Such application requires low core loss, high permeability and a proper insulation coating. Its excellent magnetic properties in the rolling direction results from a high degree of Goss texture (110)[001] obtained by secondary recrystallization, from a matrix of small grains and adequate texture stabilized by a dispersion of second phase. This material is produced by using AlN as grain growth inhibitor (acquired inhibitor) and small amounts of alloying elements like tin that acts as a supplemental inhibitor by means of grain boundary segregation and improves magnetic properties.1,2

Typically, high permeability grain oriented silicon steel has a primary insulation coating composed mainly of forsterite, formed by a solid-state reaction of the annealing separator MgO with silica and fayalite from the oxide subscale previously formed on the strip. The formation of silica and fayalite inclusions occurs during decarburization of the steel strip under a humid H2-N2 atmosphere. It has been proposed that the molecular structure of this subscale influences the forsterite film and the core losses in high permeability silicon steel. Excessively high Fe2SiO4/SiO2 ratio produces discontinuous forsterite film, with poor oxidation resistance and adherence, in addition, increases core loss.3–5 Some elements like tin can negatively influence the formation of the subscale. Tin concentrates at the surface during decarburization and retards subscale formation.2,3

The present work investigates the oxide subscale and forsterite formation in 3% Si steel with low and high tin content.

The samples (310 x 120 x 0,26 mm) used in this work were taken from two coils after decarburization, high temperature annealing and secondary coating process. The chemical composition is shown in Table I. Decarburization annealing process was performed in a mixed wet H2-N2 atmosphere with pH2O/pH2 = 0,52 at 850°C for 60 s in a continuous electrical furnace. After decarburization, the strip was nitrated in the gas mixture N2-H2-NH3 and then coated with a layer of 5 g/m2 of magnesia plus additives (mainly TiO2). The coated coils were annealed in the mixed H2-N2 atmosphere during the heating time and in 100% H2 during the soaking time at 1200°C for 15 h. In the next step, the strip was cleaned and coated with a layer of 8 g/m2 of secondary insulation coating (high tension coating), dried and finally submitted to a thermal flattening annealing at 800°C.

The morphology, microstructure and distribution of oxides in the subscale were investigated through the cross sectional observation by using image analyzer and optical and electronic microscopy (SEM - FEI - QUANTA FEG 250). In order to determine the oxide composition formed in the subscale, FTIR (Fourier transformed infrared Perkin Elmer - Spectrum 100) method was used. The oxide was carefully extracted from the internal oxidation layer by using the bromine-alcohol method and examined by FTIR.6 In-depth chemical composition was determined by glow discharge spectrometry (GDS - Horiba - Jobin Yvon - Gd-Profiler 2).

The microstructure and composition of the ceramic film of forsterite were analyzed by SEM, EDS (EDAX - TEAM - OCTANE SUPER 60mm) and GDS. The conventional ceramic film properties as surface color, adhesion, oxidation resistance and electrical resistance were investigated. The ceramic film adhesion to the base metal was characterized by bending a sample around cylinders of different diameters.3 The oxidation resistance of the ceramic film was analyzed by visual inspection (absence of oxides, medium oxide and heavy oxide) of the samples after heat treatment at 830°C for 1 h in air. The surface quality after secondary coating was characterized by visual inspection (uniform gray or no uniform color with the defects red and/or black oxides). The coating insulation was tested by Franklin test.7 

Figure 1 shows that the increase on Sn content decreased subsurface thickness, as else, fayalite/silica ratio. The achieved oxygen content was 769ppm and 681ppm, and carbon content was 7ppm and 19ppm for 0.06%Sn and 0.08%Sn, respectively, which can be related to the structural differences on the oxide subscale.8,9

The cross-sectional SEM images of the subscales are shown in figure 2. The basic structure of both samples was typical: oxides with a spherical shape in the upper region of the subscales and with a lamellar shape near the interface with the metal base.10,11

EDS (Energy-dispersive X-ray spectroscopy) analysis performed on the oxides at the subsurface confirmed the chemical composition already discussed by other authors: at the near surface a thin film around nanometers, mostly composed by Fe2SiO4 and FeO. Beneath this film, dark globular particles (dark oxides) with Si content around 6% embedded in a lighter region with a Si content lower than the composition of the steel. Some globular particles with a light gray ring with around 2%Si, probably correlated with fayalite.11,10 Besides Fe, Si and O, Mn was detected in some regions, which indicated the formation of (MnFe)2SiO4.13–15 

Figure 3 presents GDS profiles for elements Si, Al, Mn and Sn on the subscales. The sample with 0.06%Sn presents an enriched zone of Si of about 3μm. Despite the enrichment at the surface, a nearly constant level of Si up to 2μm deep can be observed, followed by an almost linear decrease to the interface with the base metal. The Al profile follows the Si profile. The Mn concentration presents a peak near surface and deeper depletion with a minimum at a distance of 2.5μm from the surface.14 In this sample, enrichment in Sn occurs only in very near surface while in the sample with 0.08% it occurs also close the interface with the metal base with a maximum concentration around 2μm. The high content of Sn does not change the qualitative shape of the Si, Al and Mn profiles, but reduces the depth of the enrichment zone.

The cross-sectional SEM images of the forsterite film are shown in figure 4. The 0.06% Sn sample presented a typical morphology: thin continuous film with particles in the subsurface.16 The increase in the tin content to 0.08% changed the glass film morphology. In this sample a discontinuous and heterogeneous film was obtained with regions in order of 400nm and others with thickness greater than 2.5μm. The amount of subsurface particles is lower in case of high Sn respect low Sn case.

The main components of the ceramic film in the 3%Si steel with AlN as inhibitor is forsterite (Mg2SiO4) and the spinel MgAl2O4.17 The stoichiometric ratios of the forsterite components are Mg/Si = 1,7 (in weight) and Mg/O=0,8. The stoichiometric ratios of the spinel MgAl2O4 components are Mg/Al = 0,44 and Mg/O=0,5.

The EDS analysis of the ceramic film on both samples are very similar regarding the elements O, Fe, Si Al and Mg. The thin film of the sample with 0,06%Sn presented with a composition of 31%O, 27%Mg, 1%Al and 15%Si and the particles at the subsurface with 15%O, 11%Mg, 12%Al and 3%Si. The thin (spot 1, figure 4a) and thick film (spot 2, figure 4a) of this sample presented a chemical composition similar with a little bit higher aluminum in the second one. These facts suggest that the continuous film of the sample with low tin is composed mainly of forsterite while the particles (spot 3, figure 4a) are mainly spinel. The increase of tin changed the composition of the surface film. It was possible to find composition like 42%O, 32%Mg, 3%Al and 19%Si in the thicker region (spot 4, figure 4b) and also some segregation of Sn in the thinner region of the surface film (spot 5, figure 4b) with a chemical composition (27,2%O, 18,3%Mg, 2,3%Si, 0,7%Sn, 0,1%Ca, 0,7%Mn and 7,3%Cu) that does not fit forsterite neither spinel, as shown in figure 5.

Figure 6a presents the GDS chemical composition profiles of the forsterite considering the elements Si, Mg, Al and Sn for the sample with 0.06%Sn. The Mg peak appears near the outer surface while Si and Al peaks appear 1μm depth from the surface.17 Sn profile follows silicon profile. Figure 6b shows that the increase of Sn decreased the enrichment of Mg, Si and Al at the near surface, as else it decreased the lag between the Mg and Al peaks and made the Al decrease softer towards the oxidation front.

Concerning coating properties, Table II shows the main properties of the ceramic film before and after the secondary coating. The sample with low tin presented better results of adhesion, oxidation resistance (figure 7) and insulation after secondary coating. Figure 7 shows the appearance of both samples (bottom surface) after the test of oxidation resistance. The sample with higher tin presented strong red oxides while the sample with low tin did not present any oxides.

It is well known that the oxides in the subscale obtained after decarburization of 3%Si-Fe at 850°C in high dew point (pH2O/pH2=0.52) are mainly composed of silica and small amount of fayalite.3 The crystalline fayalite can be found in the surface region within several ten nanometers (as a film) and with another shape in the middle region of the subscale within micrometers from surface, surrounding spherical silica. Amorphous silica can be found both in the upper region of the subscale with spherical shape and in the interface with metal base as lamellar shape. The chemical composition profiles of the alloying elements are already known: a peak near surface and a small enrichment up to 2μm-depth for Si and Al.10,12 Enrichment in Mn appears near surface and deeper depletion at the end of the subscale.14 The main novelty of this work is to discuss the Sn profile at the subsurface oxidation.

The 0.06% Sn sample showed only a very superficial tin enrichment and a Sn concentration along the subscale less than the base metal. The increase of Sn turned into a subscale with a Sn concentration higher than the substrate and with a peak at 2μm from the surface. A comparison of figures 2b and 3b shows that the segregation of tin occurred in the oxidation front. Studies have shown that the formation of precipitates at this interface may retard and even prevent the progress of oxidation.9 The hypothesis to be considered is that the segregation of tin at this interface would be responsible for the lower oxidation of the material and for the higher final carbon content.

The structure and composition of the subscale formed in the decarburization are the key in order to obtain an adherent ceramic film with good resistance to oxidation.3 Table II shows that the sample with 0.06%Sn presented a more adherent film with higher oxidation resistance, better appearance and higher insulation after the secondary coating than the steel with 0.08%Sn.

The poor oxidation resistance of the sample with high tin can be related to the discontinuities of the film. The poor adhesion can be related to the thick region (spot 4) of the film as can be observed in figure 4b. Natori named that region of the film as “aggregated of fosterite”.18 The dimensions of this defect were described as: a thickness of 2 times or more than the average thickness of the film and a size L in a direction parallel to the surface steel of 3 μm or more. This “aggregated of forsterite” can be observed in figure 4b. The results of the present work show that this aggregate is mainly spinel. So, the poor adhesion of the ceramic film of the sample with high tin can be related with this defect now named “aggregate of spinel plus forsterite” probably with a less dense structure than the typical continuous thin forsterite film.

The poor coating properties obtained in the sample with high tin can be related not only to the structure and composition of the oxide subscale, but also to the detrimental effect of Sn in the formation of the forsterite film.

The analysis of the chemical composition of the Mg, Al, Si and Sn profile in the ceramic film shows that Sn altered the profile mainly of Mg and Al in the ceramic film. The lower Sn sample showed a substantially surface-free aluminum, which may be related to pure forsterite, suggesting that the forsterite film is formed first. The sample with higher Sn showed peaks of Mg and Al closer to each other, suggesting poor or later formation of the film. Previous work has shown in order to obtain a good film, the forsterite must form first and then the spinel MgAl2O4. Masui et al reported that for a good film formation, traces of forsterite must appear around 950°C and the traces of spinel around 1000°C during the high temperature annealing of samples coated with magnesia with Sb as additive.17 

Increasing the addition of tin from 0.06% to 0.08% decreases the thickness, the oxygen content and the fayalite/silica ratio in the oxide subscale formed at 850°C with high dew point.

The poor coating properties obtained in the sample with high tin can be related not only to the structure and composition of the oxide subscale, but also to the detrimental effect of Sn in the formation of the forsterite film.

Sn segregates at the surface and generates a discontinuous, heterogeneous and non-adherent forsterite film with a richer composition of spinel than the typical surface film.

The forsterite film obtained with low tin presented a pure forsterite film on the outer surface, which suggests that forsterite film formation occurs at lower temperature than the spinel during the high temperature annealing. This result is compatible with previous studies.

1.
N.
Takahashi
,
Y.
Suga
, and
H.
Kobayashi
,
Journal of Magnetism and Magnetic Materials
160
,
98
(
1996
).
2.
K.
Iwayama
,
Y.
Kuroki
,
Y.
Yoshitomi
,
K.
Homma
, and
T. J.
Wada
,
Appl. Phys.
55
(
6
),
2136
(
1984
).
3.
M.
Komatsubara
, Int. Cl6. H01F1/04. U.S.P. n°5,571,342. 5 Nov 1996.
4.
M. G. M. M.
Cesar
and
M.
Mantel
,
Journal of Magnetism and Magnetic Materials
254
,
337
(
2003
).
5.
C. C.
Silveira
,
M. A.
Cunha
, and
V. T. L.
Buono
,
Journal of Magnetism and Magnetic Materials
358
,
65
(
2014
).
6.
N.
Morito
and
T.
Ichida
,
Corrosion Science
17
,
961
(
1977
).
7.
American Society for Testing Materials, ASTM A717/A717M. (2012).
8.
T.
Yamasaki
,
Transactions ISIJ
9
,
66
(
1969
).
9.
W. F.
Block
 et al,
Materials Science and Technology
2
,
22
(
1986
).
10.
H.
Toda
 et al,
Journal of Materials Engineering and Performance
6
,
722
(
1997
).
11.
S.
Jung
 et al,
ISIJ International
51
(
8
),
1163
(
2011
).
12.
D. C. L.
Vasconcelos
 et al,
Materials Research
2
(
3
),
159
(
1999
).
13.
Y.
Suzuki
 et al,
ISIJ International
49
(
4
),
564
(
2009
).
14.
D.
Huin
 et al,
Oxidation of Metals
64
(
112
),
131
(
2005
).
15.
S.
Jung
 et al,
Surf. Interface Anal.
45
,
1119
(
2013
).
16.
Cunha
M. A.
 et al,
IEEE Trans. Mag.
30
,
6
(
1994
).
17.
Masui
H.
 et al, Int. CL6. B32B15/04.U.S.P. n°5,565,272. 15 Out. 1996.
18.
Natori
Y.
 et al, U.S.P. n°8,920,581, 30 Dec. 2014.