ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Regular Article
Over-saturated Oxygen in Japanese Iron Nails of “Wakugi” for Wooden Structure
Yasuko Furunushi Kazuhiro Nagata
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML

2014 Volume 54 Issue 5 Pages 1074-1079

Details
Abstract

Ancient Japanese low-carbon steel, called “Hocho-tetsu” or “Wari-tetsu”, was made in “Okaji” process by decarburizing pig iron named “Zuku” produced in “Tatara” process. It is known that the low-carbon steel had higher corrosion-resistance and was much easier to forge-and-weld than modern steel. Japanese iron nails, called “Wakugi”, were made from “Hocho-tetsu” and had been used in shrines and temples until the Meiji period. The low carbon steel tends rapidly to make thin film of magnetite, called “Kurosabi”, on the surface to protect against corrosion under wet atmosphere or heating. The magnetite film is produced from the reaction of iron and oxygen. The oxygen and carbon concentrations in the iron matrix of ferrite in “Wakugi” were measured using EPMA to be about 0.15 to 0.38 mass% and 0.02 mass%, respectively. The oxygen concentration is over-saturated from the oxygen solubility of αFe and γFe. The over-saturated concentration of oxygen in “Wakugi” was caused from “Okaji” process without deoxidation of steel.

1. Introduction

Ancient Japanese low-carbon steel is called “Hocho -tetsu” or “Wari-tetsu” with about 0.1 mass% of carbon. Japanese iron nails, called “Wakugi”, had been used in shrines and temples until Meiji period and manufactured from “Hocho-tetsu”. “Hocho-tetsu” was made by decarburizing pig iron of “Zuku” in “Okaji” process. “Zuku” was produced in Tatara process.

A remarkable characteristic of “Hocho-tetsu” is high corrosion-resistance caused by the formation of magnetite thin film on the surface of steel. Although they are often heavily corroded on the surface, their cores contain sound metallic iron that is free of oxide scaling. Igaki1) carried out an enhanced corrosion test of the anodic polarization measurement in a natural aqueous solution using iron artifacts of various ages. Modern steel was poor in terms of corrosion resistance and showed heavy fluctuations on the passive current because of lack of stable passive film formation. Most old iron artifacts in the Kamakura period showed high corrosion-resistance and kept metallic luster under atmosphere for several months. Sugimoto et al.2) found that the surface covered with uniform thin layer of Fe3O4 acted as a protective film to prevent from further oxidizing. W.E.O’Grady3) investigated the relationship between the corrosion resistance and the crystal structure, and showed that the structure of ferrite phase exhibited high corrosion resistance. Furunushi4) focused on the oxide film formed on the surface of a nail that was manufactured by blacksmith. Using a transmission electron microscope, the fine polycrystals of FeO were found in the unit size about 10 nm at the interface between the steel matrix and the oxide film of scale.

Hocho-tetsu” was much easier to forge-and-weld than modern steel. Nagata et al.5) found that the interface of steel blocks are melted by the oxidation heat of steel at the interface in the time of forge-and-welding. Then, white sparks of “Wakibana” appear in the flame of charcoal burning in blacksmith’s furnace. The interface of steel blocks is filled with molten FeO and generates CO gas bubbles. The growth of hemisphere gas bubble hits molten steel at the interface to produce small iron particles. The iron particles are oxidized in air to make “Wakibana”. The molten steel at the interface of blocks is near in equilibrium with molten FeO in about 1470°C.

Nagata et al.6) realized that in the final stage of “Okaji” process, steel block was oxidized by air, resulting in heating and melting itself by oxidation heat, leading to decarburization less than 0.1 mass%C at about 1500°C. In this state, molten steel is covered with molten FeO and near in equilibrium. After then, the steel block was quickly forged to make steel plates of “Hocho-tetsu”.

Therefore, the oxygen concentration in “Hocho-tetsu” and forge-and-welded steel could be high. However, the conventional way to analyze oxygen content in “Hocho-tetsu” and forge-and-welded steel has been chemical analysis that includes oxygen in non-metallic inclusions as well as iron matrix. In the present work, The oxygen concentration in iron matrix without non-metallic inclusions in Japanese nails has been measured by an electron probe micro-analyzer (EPMA). The effect of high oxygen concentration in Japanese steel on the high corrosion-resistance of “Wakugi” will be discussed.

2. Experimental Methods

2.1. Specimens

The three “Wakugi” samples of No. 1, 2 and 3 were employed. No. 1 sample was used in the Saidai-ji temple in Nara Prefecture and manufactured in A.D.1300 of the Kamakura period. No. 2 sample was used in the Anumi Shrine in Ehime Prefecture and manufactured in A.D.1835 of the Edo period. The length of No. 1 was 150 mm and that of No. 2 was 170 mm. As shown in Fig. 1, they were formed in long pyramidal bars of steel with rolled head for No. 1 and bent head for No. 2, respectively. Their cross-sectional shape were squares of 6×5 mm for No. 1 and 8×7 mm for No. 2, respectively. Their surfaces were covered with a thin corroded layer of Fe3O4, FeOOH, FeO, and Fe2O3.

Fig. 1.

Overview of Japanese iron nails.

The nail of No. 3 was used in the Bicchu-kokubun-ji temple, where the temple was reconstructed between 1821 and 1835. The shape was already reported in the previous report.4) The size was 85 mm in length and 5×4.5 mm in cross-section.

2.2. Experimental Procedure

The iron nails were cut and mechanically polished. An electron micro-prove analyzer (EPMA) with a LaB6 electron gun was used to measure oxygen concentration in steel samples. The EPMA can analyze the concentrations of elements in matrix with the area of 1–5 μm in every direction excluding non-metallic inclusions. The microstructures of non-metallic inclusions in the nails were observed using a scanning electron microscope (SEM). The hardness of the nails was measured using a Vickers Hardness(Hv) tester (Hv 300 g).

3. Results

3.1. Hardness and Non-metallic Inclusions of Wakugi

The cross-sections of nails are shown Fig. 2 with Vickers hardness values. The gray parts in the figure are non-metallic inclusions that appear along the metal flow. The number of non-metallic inclusion in No. 1 sample is more than No. 2. The microstructures of both samples are single-phase ferrite crystals with carbon concentration of about 0.1 mass% or less.

Fig. 2.

Cross-sections and Vickers hardness of Japanese nails.

The Vickers hardness of No. 1 sample was in the range from 100 Hv to 135 Hv. The grain size at the 100 Hv area was about 100 μm, and 135 Hv, 20 μm. The Vickers hardness of No. 2 sample was in the range from 87 Hv to 186 Hv. The grain size at the 150 Hv area was under 10 μm. No. 3 sample as a reference was in the range from 115 Hv to 125 Hv. From these results, it is realized that carbon distribution in nails was not uniform and carbon content greatly fluctuates from place to place. Non-metallic inclusions, called “Noro”, were iron-rich slag with the composition of near fayalite. Figure 3 shows each grain and texture in the non-metallic inclusions, the chemical compositions of which are shown in Table 1. These grains and textures were composed of FeO (No. 1, No. 3),4) 2FeO·TiO2 (No. 2, No. 3), and textures, in fine dendrite structure. The grain of FeO consists of about 95 mass%(Fe+O), and the grain of 2FeO·TiO2, about 97 mass%(Fe+O+Ti). The concentrations of other elements are at the almost same level.

Fig. 3.

SEM images of microstructure of the inclusions.

Table 1. Chemical composition of the non-metallic inclusions in the Japanese nails measured by EPMA (mass%).
elementgraintexture
No.1No.2No.34)No.1No.2No.34)
Ti rich
mass%Si0.1870.0710.1701.68215.73215.43016.77
Mn0.1150.3100.1340.1840.2510.5340.282
P0.0040.0020.0000.0000.5572.4381.31
S0.0070.0080.0060.0140.1240.1390.193
Ti1.84515.4162.18012.5360.3170.4490.524
O27.42631.82826.34733.86138.50536.37140.224
Fe70.63550.37769.39249.03641.78037.60333.697
Al0.7512.8430.5916.6743.2592.4454.694
Mg0.1130.2390.3780.4490.7180.7821.011
Ca0.0000.0300.0000.7312.3324.7593.251
Total101.083101.12499.198105.167103.575100.950101.956
atomic%Ti1.39.61.57.30.20.20.3
O56.059.655.258.760.859.261.7
Fe41.327.041.724.418.917.514.8
Si0.20.10.21.714.214.314.6
O/Fe1.42.21.32.43.23.44.2

3.2. Concentrations of Oxygen and Other Elements Dissolved in Iron Matrix of Wakugi

Table 2 shows the concentrations of oxygen and other elements in iron matrix of inner sample and near surface. The oxygen content of iron matrix was 0.153 to 0.187 mass% for No. 1 and 2 and 0.383 mass% for No. 3 by EPMA. The metal structure of iron matrix in samples measured by EPMA is single phase of ferrite that has the maximum solubility of carbon of 0.02 mass%. The carbon concentration of No. 3 by chemical analysis in Table 3 was 0.04 mass%. The difference means that pearlite structure was partially in sample because of fluctuation of carbon concentration.

Table 2. Chemical composition in iron matrix of the Japanese nails measured by EPMA (mass%).
elementSurface PartInner Part
No. 1No. 2No. 34)No. 1No. 2No. 34)
Si0.0200.002000.002
Mn000000
P0.010.0190.0490.0160.0810.047
S000.009000.008
Ti00.0070000
O0.180.1870.1710.1750.1530.383
Fe98.04398.62298.45798.25898.3397.635
Al000000
Mg0.00100000
Ca000000
Table 3. Chemical composition of the Japanese nails (mass%).
Wooden structureperiodADCSiMnPSTiORef.
Horyu-jiKon-doAsuka/Nara6070.100.004tr.0.0330.004<0.0100.0148)
Byodo-inHeian10530.350.0390.010.0300.003tr.0.0438)
0.190.0980.010.01tr.tr.0.1478)
0.200.082tr.0.0140.0030.1450.2208)
0.210.052tr.0.0070.0030.0470.2408)
Horyu-jiKon-doKamakura12830.090.013tr.0.0270.0030.0100.0768)
Horyu-jiKon-doEdo16030.250.0080.2300.0180.063<0.0100.0098)
Byodo-inEdo16700.300.030tr.0.0300.0020.0440.1908)
Bicchu-Kokubun-ji(1821)0.040.0210.0070.0680.0040.0830.4904)
Konko-in(ori)Edo17000.090.0030.0030.0410.0050.0020.064
0.020.0330.0030.0040.0040.0020.0044)
0.040.0640.0030.0240.0040.0180.3504)
Senjyu-ji17290.240.0290.0050.0380.0040.0010.160
Daigo-ji17700.160.006tr.0.0380.0010.0250.0128)
Otsuka-shuzou19000.070.0050.8100.0550.0280.0010.032
SLCM(Yakushi-ji)Gendai0.090.010.0100.0010.0020.0039)
b.f.steel(SPHC)20000.04<0.0080.2100.0020.0130.0010.0024)

Note: Oxygen concentrations were analyzed by chemical analysis.

4. Discussions

4.1. The Metallurgical Property of Japanese Nails

Metallurgical investigations of nails were first performed on Horyu-ji temple nails by Nishimura et al.7) and Horikawa et al.8) Horikawa examined twenty-eight nails with the production time extends to 1800 from 607 years. The hardness distribution in these nails is in the range of 80 to 250 Hv. The hardness of No. 1 and No. 2 samples are lower than these nails. The reason could be that the metal structure of No. 1 and 2 samples is ferrite single phase.

4.2. Chemical Composition of Wakugi

Table 3 shows the chemical composition of Japanese nails in published literatures.4,8,9) The concentration of Ti was over 0.01 mass% in the Japanese nails made from hocho-tetsu that was produced from pig iron by Tatara furnace using iron sands. The concentration of Mn in the nails was lower by two orders of magnitude than that of modern blast furnace iron. Also, the concentration of sulfur in the nails was observed to be very low. The low concentrations of these elements come from high oxygen potentials of about 1×10–12 atm at 1350°C in Tatara furnace and of about 1×10–9 atm at 1500°C in Okaji furnace that are near the equilibrium oxygen potential with iron and FeO, while oxygen potential in a modern blast furnace is low of about 1×10–16 atm at 1500°C in equilibrium with carbon and CO gas. Thus, in the blast furnace, gangue in iron ore is easily reduced and dissolved into pig iron produced. In addition, it should be noted that the iron produced using charcoal has less sulfur that makes iron brittle.

4.3. Oxygen Concentration in Iron Matrix of Wakugi

EPMA with LaB6 electron gun used in the present work had high accuracy in quantification. As the resolution area of secondary electrons is about 5 nm, inclusions with the size of 25 nm can be recognized but inclusion with the size less than 25 nm could not be recognized. Figure 4 shows the iron matrix and FeO inclusions with the size of about 1–3 μm including slightly Si, P and Ti elements.

Fig. 4.

SEM images of the fine inclusions of FeO including slightly Si, P and Ti. The mark + is the measurement points. Si concentration in the black area (B) is relatively larger than (A). P in (A) is relatively larger than (B).

The iron matrix composition of nails of No. 1 and 2 measured by EPMA is shown in Table 2. The analysis was carried out at the center and near the surface of samples, respectively. The oxygen concentration of the iron matrix ranged from 0.153 to 0.189 mass% and approximately lager by one order than modern steel. The data in literature4) shows the oxygen concentration of 0.490 mass% for Bicchu-Kokubun-ji and 0.350 mass% for Konko-in that oxygen concentrations are analyzed by chemical analysis and affected by non-metallic inclusions in nails.

Figure 5 shows the relation between the carbon and the oxygen concentrations in nails measured by chemical analysis and EPMA reflecting on the phase diagram of Fe–O–C system at 1528°C. The white circles are the concentrations determined by chemical analyses. The samples with the oxygen concentration less than 0.04 mass% are in liquid iron phase and δFe. The other samples are, however, out of liquid phase and have the oxygen concentration including non-metallic inclusions. The data by EPMA analysis are also in liquid phase. In “Okaji” process, a steel block was heated over 1500°C and partially melt by the oxidation heat of iron. After carbon concentration in steel decreased less than 0.1 mass%, the block was taken out to forge and quickly formed “Hocho-tetsu” in plate. Thus, the state of decarburized steel remained oxygen concentration in liquid iron into matrix without non-metallic inclusions.

Fig. 5.

Carbon and oxygen concentrations in Japanese nails with phase diagram of Fe–C–O system at 1528°C.

4.4. Chronological Change of Oxygen Concentration in Japanese Nails

Figure 6 shows the chronological change of oxygen concentration in nails.4,8,9) The nail of Horyu-ji temple in Nara city dated in A.D. 607 has low oxygen concentration of 0.014 mass% with 0.10 mass% of carbon. The state of liquid steel at high temperature is in equilibrium with δ-Fe, as shown in Fig. 5. After then, the oxygen concentration in nails chronologically increased to about 0.2 mass%. The state of most of nails is out of liquid iron and has non-metallic inclusions.

Fig. 6.

Chronological change of oxygen concentration in Japanese nails and modern steel analyzed by chemical analysis and EPMA.

The chronological change in the state of steel from the Nara age to the early age of Heian could be due to the development of technique of “Okaji” process, that is increase in operation temperature. In order to increase temperature and to proceed decarburization of steel in “Honba” of the second stage of “Okaji”, iron was oxidized and produced more FeO and non-metallic inclusions.

4.5. Over-saturated Oxygen in Steel during “Okaji” and Forge-and-welding Processes

In Tatara process, steel “Kera” and pig iron “Zuku” were produced. Pig iron and low quality of steel “Bugera” in these products were decarburized with air by hand blowing to produce steel plates with low carbon concentration, called “Hocho-tetsu” or “Wari-tetsu”. The decarburization in premodern refining process, called “Okaji” was composed of two stages of “Sageba” and “Honba”.10) In the “Sageba”, pig iron was decarburized to steel of about 0.7 mass%C, called “Sagegane” and the yield was almost 100%. In the “Honba”, the “Sagegane” was decarburized to about 0.1 mass%C, called “Orosigane” and the yield was 60 to 70%. The “Orosigane” was spontaneously forged to make plates without deoxidation.

Japanese steels of “Watetsu” produced by “Tatara” process, such as “Tama-hagane” of high quality steel and “Hocho-tetsu”, have been characterized by fluctuation of carbon concentration. Therefore, “Watetsu” is forge-and-welded by hammering to make a plate. The fluctuation of carbon concentration in steel dispersed into fine areas. In the forge-and- welding process, steel plates are heated up to about 1300°C by charcoal burning. When “Wakibana5) of white sparks are appeared in flame, blacksmith takes out the plates to forge and welded each other. White sparks of “Wakibana” are small iron particles oxidized by air. The iron particles come from the interface of steel plates where the surface of steel is heated near 1500°C and melted by oxidation of iron. FeO melt produced by oxidation of iron reacts with carbon in steel to generate CO bas bubbles. The growth of CO gas bubble is accompanied with production of fine iron particles.

White sparks of “Wakibana” always appear in flame when Japanese steel is melted and accompanied with oxidation of iron to produce FeO melt. FeO melt contacts with molten steel near 1500°C. Therefore, the oxygen concentration in iron matrix becomes about 0.2 mass% and FeO inclusions are included.

As shown in Table 3, the oxygen concentration in modern steels is very low because of deoxidation of molten steel by aluminum before solidification. However, there is not deoxidation process to make low carbon steel in “Okaji” process.

4.6. High Corrosion-resistance of “Wakugi

The maximum solubility of oxygen in δFe is 0.0084 mass% at 1528°C, 0.003 mass% in γFe at 1371°C and lower in αFe than γFe. Oxygen concentration in Japanese steels of “Watetsu” is always in over-saturation. Under 560°C, αFe with over-saturated oxygen decomposes into ferrite and magnetite by some triggers, such as humidity and heating. Oxygen in iron is a surface active element and tends to concentrate at the surface of αFe plate. Therefore, the surface of “Hocho-tetsu” with over-saturated oxygen is quickly covered by magnetite film at a trigger. The thin and uniformly dense oxide film of magnetite acts as a protective layer that behaves high corrosion-resistance.

Over-saturated oxygen also precipitates fine oxide particles, mainly wüstite of FeO in steel during forge-and welding process. These fine particles could be a material to form some beautiful patterns on Japanese swords as well as metallic crystals of metallography.

5. Conclusions

Japanese nails were mainly composed of ferrite and have over-saturated oxygen of about 0.15 to 0.38 mass% in the iron matrix. The nails were made from “Hocho-tetsu” that was produced in “Okaji” works. In “Okaji” process, pig iron and low grade steel block produced in Tatara works were heated and melted by charcoal burning and decarburized by air near 1500°C. The decarburized steel block was immediately forged to form the plates of “Hocho-tetsu” with fluctuated carbon concentration. Over-saturated oxygen in iron matrix of nails quickly forms a thin film of magnetite on the surface of plate that is caused of high corrosion-resistance.

Acknowledgement

Authors thank Dr. Takako Yamashita, JFE Steel Corp, for her help with our experiments.

References
  • 1)   K.  Igaki: Kobunkazai-no-Kagaku, 28 (1984), 18.
  • 2)   K.  Sugimoto,  S.  Matsda,  M.  Isshiki,  T.  Ejima and  K.  Igaki: J. Jpn. Inst. Met, 46 (1982), 155.
  • 3)   W. E.  O’Grady: J. Electrochem. Soc., 127 (1980), 555.
  • 4)   Y.  Furunushi: Tetsu-to-Hagané, 91 (2005), No. 1, 91.
  • 5)   K.  Nagata,  T.  Watanabe and  N.  Kugiya: Tetsu-to-Hagané, 97 (2011), No. 12, 637.
  • 6)   K.  Nagata and  T.  Watanabe: Abstract of BUMA8, D-03, ISIJ, Tokyo, (2013).
  • 7)   H.  Nishimura and  N.  Aoki: Tetsu-to-Hagané, 41 (1955), No. 3, 289.
  • 8)   K.  Horikawa and  Y.  Umezawa: Tetsu-to-Hagané, 48 (1962), No. 1, 4.
  • 9)   O.  Umezawa: Bull. Iron Steel Inst. Jpn., 6 (2001), No. 10, 49.
  • 10)   K.  Tawara: Old Smelting Method of Iron Sand in Meiji Period, Maruzen, Tokyo, reprint with come by M. Tate, Keiyousha, Tokyo, (2007).
 
© 2014 by The Iron and Steel Institute of Japan
feedback
Top