Skip to main content

Pavel Bolf


Pavel Bolf Katana kaji

Japanese sword making

Raw materials for making a Japanese sword blade.

The steel used for traditionally made Japanese swords is one of the factors that make the Japanese sword unique. It is typically made by reducing iron ore on charcoal in a tatara furnace. Another typical characteristic is the chemical purity of the steel thus reduced. In fact, it is an almost pure compound of iron and carbon. Other elements are present in very low levels. Due to the extremely low levels of chemical elements in the steel, these do not significantly affect the properties of the steel.

There are two types of steel used to make Japanese swords. Tamahagane, steel made by reducing iron ore to charcoal in a Tatara furnace. The second type of steel is Oroshigane. This steel is created by remelting unprocessed tamahagane steel of substandard quality or processed traditional steel (products made from it). For example, old nails, fittings, cast iron teapots and so on. Charcoal is used to remelt the steel and is done in a kiln.

The production of Tamahagane steel is quite demanding. It is made in a traditional Tatara furnace, built of clay. It is a shaft furnace with air intake at the bottom. The furnace is filled from the top with a charge of iron ore and charcoal. After the smelting is completed, the furnace must be demolished to remove the molten steel. The size of the furnace is proportional to the amount of steel required. In Japan, steel is currently produced at Shimane in Great Tatara. It produces about 2 tons of steel per plant. Such smelting is very demanding in terms of raw materials and time. The smelting process takes about three days, and a charge consists of about 20-30 tonnes of iron sands and charcoal. The resulting molten Kera steel ingot fills the bottom of the furnace and is about 2m long, 1m wide and 0.5m high. This huge piece of steel has an uneven carbon content. At the edges is high carbon steel with a carbon content of more than 1%. Towards the centre, the content decreases. The core of Kera is made up of low carbon steel with a carbon content of less than 0.5%. For this reason, it is necessary to break Kera into smaller pieces and sort the pieces according to quality. Steel with a carbon content of between 1-1.5% is suitable for sword making. Steel with a lower carbon content can be treated using the Oroshigane method or used for Shingane core steel. Steels with a carbon content higher than 1.5% are very difficult to forge, and as the carbon content increases, forging becomes impossible. Steel that is too carburized will tear during forging and be incoherent. It can be used in small quantities in the construction of the basic steel package for the sword hull. Small particles of such steel create a more distinctive structure in the folding structure of the Serpent. Overly carburized steel can again be modified by the Oroshigane method. Smaller furnaces can be used to increase efficiency in producing quality Tamahagane steel with ideal carbon content. A furnace with an internal shaft diameter of about 40cm and a charge of 30kg of quality iron ore and about 100kg of charcoal, will produce a Keru of 10-15kg after 4 hours of smelting. The quantity and quality of the steel is greatly influenced by the amount of air blown in and of course the quality and purity of the iron ore. Ideally, all Kera is made up of high carbon steel directly usable for processing into Hagane sword blade and sheath steel. Kera weighing more than 15 kg usually already contain steel with a lower carbon content in the core.

Nowadays it is also possible to use modern furnaces composed of several parts. These are made of a steel casing filled with refractory clay. These modern furnaces can be disassembled into individual stages after the melting process is completed. The molten steel is usually located at the bottom of the furnace. After breaking it out and removing it, the furnace can be reassembled and reused. Reuse damages the inner clay lining of the furnace. This needs to be repaired before each melting.

The Oroshigane technique can be used to adjust the carbon content of the steel. It allows the carbon content to be reduced and increased. The thickness of the charcoal layer between the air-entry chamber and the surface of the charcoal layer is essential for reducing the carbon in the steel.  If the charcoal layer is low, the melting steel is exposed to the oxygen contained in the air fed into the furnace. The oxygen binds with the carbon contained in the steel and thus the carbon content of the melting steel is reduced. If the charcoal layer is high, the melting steel in turn takes up the carbon contained in the carbon dioxide produced by the burning of the charcoal. This increases the carbon content of the steel. It is up to the experience of the swordsmith to determine the carbon content of the steel to be treated and to choose the appropriate charcoal layer height accordingly. Another important factor is the amount of air blown into the furnace. A greater amount of air accelerates the burning and thus the rate at which the steel inserted into the furnace melts and permeates from the surface to the bottom of the furnace under the air supply. The process of making Oroshigane is relatively simple. A shovelful of charcoal and a handful or so of pieces of remelted steel are gradually added alternately to the hot hearth. The size of the pieces of steel should be adjusted to a maximum of 3x3 cm so that the pieces are completely melted. It is better to melt smaller pieces of steel. It takes about 1 hour to melt 6 kg of steel. After the last pieces of steel have been inserted, only charcoal needs to be added for about 20 min.

The final product of the smelting process is Kera steel ingot in the case of both Tamahagane and Oroshigane. It is also referred to as 'sponge', due to its appearance resembling a loaf of bread or a sea sponge with a surface dotted with rough and sharp protrusions resembling coral formations. The underside of the Kera is often smoother and more compact.


I have done many experiments in steel melting. Their aim was to obtain the ratios of the amount of charge, the steel melted at the maximum possible quality achieved. The aim was to obtain steel in as homogeneous a piece as possible with as little variation in carbon content as possible in the different parts of the Kera. For Tamahagane I achieved the best results with a charge of 30kg of high grade iron ore (magnetite, iron sand 98.8% iron oxide), 110 KG charcoal. 10kg for furnace preheating, 100 kg for smelting. Duration of smelting approx. 3h 30 min. After two hours and after the iron ore is finished, I punch the slag through the hole in the bottom of the furnace. A modern three-segment furnace is usually used for smelting, the smear is made of fireclay. The inner diameter of the furnace is 40 cm. The bottom of the furnace is shaped in the form of a bowl of packed ash. This allows the steel to be cast into a compact piece in the shape of a loaf of bread. The resulting steel ingots range in weight from 12-15 kg. With faster burning (more air blown in), less steel is melted, with a lower carbon content. Too much air will accelerate the passage of the ore through the furnace. In the extreme case, the ore is not reduced to steel and only slag results from the smelting process. This material, which has not been reduced to steel, can be crushed after cooling and reused for the next charge. The slag discharged during successful smelting consists of molten minerals with no iron content. This is unusable for further smelting or is not a source of steel. However, a small amount of slag introduced into the furnace at the beginning of the smelting process may create a more favourable reducing environment at the bottom of the furnace to 'kick-start' a successful smelting process.

When making oroshigane, I use a 6 kg load of treated steel and about 10 kg of charcoal (in an already heated furnace).  The minimum amount of steel used for the charge is, in my opinion, 4kg. Using less than this makes the remelted Kera too jagged and difficult to process. More than 7kg of steel for remelting has a negative effect on the even distribution of carbon in the treated steel. The core of a large Kera has a lower carbon content and the lower part of the core is very high carbon cast iron.

After the steel has finished setting, I add charcoal for about 20 min. About 10 min before the end of the melting process I cover the furnace with a steel plate. This will raise the temperature in the lower part of the furnace and the higher temperature will cause the crust to fuse into a more compact unit.

When the melting is finished, I break the Kera out of the bottom of the furnace and throw it into the water while it is still hot. This cleans the Kera of most of the slag and other adhering impurities.

When making Oroshigane, I usually pile the rest of the hot coals back into the firebox after removing the Kera and do another smelting. Sometimes a third melt can be done. Afterwards, the hearth needs to be allowed to cool and cleaned of the slag that has accumulated on the bottom of the hearth during successive smelts.

The quality of traditionally manufactured steel can vary. The steel used in Japan to make Japanese Tamahagane swords is chemically very pure. This means that its composition is almost entirely iron and carbon. This is due to the extreme chemical purity of the iron sands that are used as a feedstock in the reduction of the steel. So we try to use ore or steel of similar quality in the production of Japanese swords. The iron ores found in Europe often contain a higher proportion of chemical elements. Some, such as manganese, increase the malleability and hardness of the steel. This need not be a negative characteristic in small quantities. On the contrary, manganese increases the hardenability and thus the hardness and cutting power of steel.  Other elements such as phosphorus or sulphur affect the properties of steel used for sword blades negatively.

Reduced steel, keru, needs to be visually inspected. Its quality can be judged quite accurately by its surface structure. This varies depending on the carbon content from foamed (approx. 0.4-0.6%C), to sponge-like pumice (0.7-1.5%), to a slice with a smooth surface (1.6-2.5%C)

Low carbon parts can be used for core steel and sword guards, or steel components. However, it cannot be used as cutting steel due to its low ductility. After several folding operations, the carbon content is reduced below 0.5% and the steel can no longer be hardened to a higher hardness.

Steel with a carbon content of around 1% is usually used for the sheath and cutting edge of the blade. During the process of homogenization and cleaning by forging, the carbon content is reduced to a final 0.6-0.7%. The carbon content can be controlled by various methods during the reloading process. These are wrapping the packet in burnt straw ash during forging and protecting the packet from oxidation during heating to welding temperature with a layer of water-clay solution. However, there are more factors affecting the change in carbon content during the steelmaking process. For example, the welding temperatures reached, where the higher the temperature, the greater the carbon loss. The location of the packet relative to the air supply to the furnace also has an influence. If it is placed directly in front of the air inlet, decarburization occurs faster. However, the most important factor is the carbon content of the packet at the beginning of the process. I prefer steel with a higher carbon content (approx. 1.3%). It is a little more difficult to work with such steel at the beginning. It tears more and is harder to join. However, after the fourth folding, the properties change for the better due to the lower carbon content. The steel is more malleable and starts to weld well. Each folding reduces the carbon content of the packet by about 0.05%. This changes the properties of the steel for the better. At the same time, however, care must be taken not to over-carbonize the steel and thus cause it to deteriorate. After ten cycles of folding, the carbon content is reduced by approx. This is usually sufficient and the steel is ready for final package construction, which combines the cladding, cutting edge and core steel of the desired style. Changes in the quality of the steel during reloading can be well observed at the chopping and bending points of the packet. Well-treated steel does not tear in the bend, is ductile and the welded layers do not separate at the intersection. This condition can usually be achieved after eight to ten folding operations. Steel processed in this way is easy to shape. It retains these properties for a further two to three folding operations. Then its malleability and formability begin to deteriorate rapidly. It begins to tear at the edges during forging. Further folding leads to complete degradation of the steel. I usually fold the steel 8-10 times and then put the final package together. The number of layers in the steel processed in this way varies between 3500-30000 layers depending on the number of layers in the base packet at the beginning of the process.

Reduced sponge, or segments of broken bark of larger size , are usually irregularly shaped formations. Tamahagane produced in Japan comes in pieces approximately the size of a fist. This applies to high quality. Lower quality steel usually comes in smaller pieces. These pieces of steel need to be heated to a temperature of around 900°C and cut into slabs about 0,5 cm thick. These are then hammered and broken into smaller segments, which make up the packet. Larger pieces of raw steel can be cut into 1-3 cm thick blocks. These can then be used as a base for a packet made of small steel segments after welding to the handling bar, or the packet can be built directly from them.

The tamahagane and oroshigane produced in my forge in a forge or small tare weighs between 3 and 12Kg. Oroshigane I try to melt in about 4kg bushes . Based on many experiments, this size is ideal for its quality, i.e. carburization and compactness. Also the shape and weight is ideal for handling and further processing. It can be easily recast into a relatively homogeneous block and further divided into smaller pieces for packet construction. Kernels weighing more than 5 kg have to be heated and cut into halves for further processing in a similar way. However, the heating of large chips is quite demanding due to their size. When we have pieces of steel of the appropriate size, the processing process begins. Its purpose is to clean the steel of slag and other impurities. Homogenising it and adjusting its carbon content to values suitable for blade hardening.

Before we begin to discuss the technique of working steel using the forging technique, I will briefly mention the other types of steel used in sword making and explain the basic differences between them. For now, we've covered the traditional steel used for making Japanese Tamahagane and Oroshigane swords. These steels are reduced or remelted on charcoal. The processes result in a steel sponge, kera.  Namban tetsu is mentioned in the history of Japanese sword making. Translated, it's the steel of the southern barbarians. It was steel imported into Japan. It was brought in by ships from the south, and to the Japanese, anything outside their territory was considered barbaric. Whether the steel came from Korea, China or India or European countries. Steel was usually imported in the form of forged steel ingots. What is crucial, however, is the fact that it was steel produced in a similar way to tamahagane. That is, by reducing the ore on charcoal in a shaft furnace. The smelted pieces were then recast into ingots for ease of handling and to save space. Unlike the steel produced in Japan, imported steels were less chemically pure. For example, most ores from deposits in Europe contained manganese. There are swords whose spines say they were made from namban tetsu.

I have come across the opinion that wootz (true damask) was also imported to Japan from India. This steel is produced by a completely different method. It is the refining of steel by remelting it in a crucible. This was made of refractory clay (e.g. fireclay). The crucible was filled with pieces of steel, supplemented with ground charcoal to carburise the steel and glass was added to the surface. The glass prevents oxidation after the contents of the crucible are melted and impurities are picked up. The crucible was sealed after filling and placed in the furnace. It was then heated for several hours until the contents were completely liquefied. It was then allowed to cool slowly. The slow cooling crystallises the typical structure of this steel. If the melting is done correctly, the resulting ingot is already clean, without slag. It is also compact. Such steel can be forged directly, without the need for reloading. However, it is very hard, difficult to forge and requires specific processing techniques. Failure to do so results in ingot breakage. At high forging temperatures, the typical wootz structure of the steel is also removed and it becomes 'ordinary' high carbon steel. If idiopathic wootz was indeed used in Japan, it is possible that the technological processes for processing this type of steel were unknown in Japan and therefore it was only used as high-carbon steel mixed with other types of steel (e.g. oroshigane).

Another type of steel used in swordsmithing is mechanical Damascus. This type of steel originated as an attempt to imitate real damascus (wootz). It is created by combining steels of different properties and layering them by forging. This process is similar to the reloading of steel for Japanese swords. The fundamental difference is the choice of steel for the base package. It combines steels with a significant difference in composition. For example, low-carbon and high-carbon steel. After repeated folding, the packet develops a characteristic distinctive structure of mixed layers. This can be influenced by the forging method, by extruding the layers or by torsion (twisting) of the bars made of this steel. After polishing the blade, this structure is clearly visible. Outside Japan, the structure of damascus steel was further enhanced by etching in acids. The technique of creating complex damascus patterns, combinations of bars with different structures, and complex blade constructions from differently processed parts were widespread in Europe and mastered at a high technological level. In Japan, this type of steel processing is not very common. It was, however, used, for example, by the Gassan school in the creation of its typical wavy steel structure (Ayasugi hada).

Nowadays, the damascene steel production technique is widely used in the cutlery industry. Due to the technological possibilities and advances in metallurgy and metallurgy, we have a selection of steels with much greater colour contrast and performance characteristics superior to historic steels. This allows the creation of works of art of a high aesthetic standard and the continued development in this field.

Modern metallurgical methods are also capable of producing large quantities of steel with a chemical composition similar to traditional tamahagane steel. It is essentially an unalloyed high carbon steel. Such steel is generally recognised as tool steel. It is used for the production of cutting tools. This steel can be used to make a blade for a Japanese sword. In the period after the 2nd World War, it was briefly used by Japanese sword makers due to the lack of traditional steel produced in Japan. However, due to its properties in its raw state and the mechanical and visual differences in blades made from it, such blades are not considered Japanese swords. Currently, the use of modern steel in the production of swords is not allowed in Japan.

Steel folding.

Traditionally made raw steel is not a very high quality material. It is a porous piece of steel with an unevenly distributed carbon content and many inclusions, impurities, slag and charcoal pieces. Most of the impurities are on the surface of the bark. By recasting it into an ingot, most of these surface impurities are removed. Heating the batch to welding temperature and recoating it produces relatively clean steel. Kers with a carbon content above 1.2% are more difficult to forge and will tear during forging. This negative characteristic is removed as the carbon content decreases during the recoating process. However, after the first or second folding we can already obtain steel suitable for blade (knife) production. The impurities in the steel increase the corrosiveness of the steel. To thoroughly clean the steel and homogenise it, the steel needs to be recast at least 6 times. To further refine it and thus increase the quality of the final sword blade, the final number of folding is about 10 times. Some sources indicate up to 15 folding of the steel. The evolution of steel quality during the reloading process can be monitored by observing several factors. The first is the malleability of the steel. During the first folding, the steel is hard, unyielding, and tears when forged at higher temperatures. Usually the ductility changes significantly after the 4th-5th folding. Due to homogenization and reduction of carbon content by about 0.2% it becomes more pliable, easy to shape. It retains these properties until approx. 12-13th folding. It has been my repeated experience that after that, the steel starts to degrade rapidly during the next folding. Its malleability deteriorates significantly. Although it is easy to shape and soft, it starts to tear when the packet is pulled. At first, short cracks form at the edges of the packet. On further folding, the cracks are larger and deeper. The steel becomes essentially unworkable.

Another aspect is the formation of scales on the surface of the packet. Scales need to be removed from the surface of the welded surfaces. Usually, a stream of water is sprayed under the packet between the hot steel surface and the anvil to remove them. The water rapidly cools the scale layer, partially separating it from the steel surface. Subsequently, the steam pressure generated between the hot packet surface pressed against the anvil surface will pull the scale away from the surface.  If the scale remains enclosed within the folded packet, it can prevent the steel surfaces from joining and a weld defect (ware, kitae ware) is formed at this point. Even a 1 mm thick layer of scale is formed. These are sometimes difficult to remove from the steel surface and mechanical aids are needed. For example, the edge of a hammer or various scrapers.During the reloading process, the amount and thickness of the scale decreases. In the end, the scales take the form of small flakes that fall off spontaneously during the forging of the package.



Utsuri is an interesting metallurgical effect. It appears as a shadow over the hamon line. Its specific form may vary. There are many types of utsuri. According to the shape, we can distinguish bo-utsuri, as a straight line. Midare utsuri has the shape of waves, choji utsuri like choji hamon resembles clove flowers. Sometimes the utsuri is scattered over the blade area in separate patches. In some cases it follows the line of the hamon and is identical in type to it. Other times the hamon and utsuri differ in type. Most often, the utsuri appears as a dark shadow. Sometimes this shadow is bounded by a band of reflective particles or is directly formed by reflective particles. Then we just refer to it as non-utsuri. This brightens significantly when the blade is illuminated at a certain angle. Utsuri ideally follows the hamon line the entire length of the blade. This is not always the case.

There are also types of utsuri that occur only at the beginning of the hamon line. The first is the mizukage. This phenomenon usually manifests as a shadow rising above the beginning of the hamon line at a steep angle toward the back of the blade. Translated, mizukage means water shadow. Mizukage is thought to be an effect produced by the plunging of the sword into the water during tempering where the water level was. It would represent the transition between the cooled and uncooled part of the blade during hardening. In reality, however, this transition leaves only faint metallurgical changes on the blade.  The way the blade is heated has a much greater influence on the formation of this effect. Significant misukage occurs at the point where the temperature of the blade drops at the nakago during the hardening heating. If the temperature difference in the area of the blade setting before hardening is significant, a significant misukage is produced. It is therefore more of an utsuri type. Mizukage is listed as a defect in some sources. There is no justification for this. Only when it occurs in shortened blades is it a sign of blade over-hardening. Then it may be considered a manifestation of a defect, a non-native hardening.

Another interesting type, or rather part of utsuri, is the cluster of reflective particles at the beginning of utsuri, called koshiba. It forms in the same area as the mizukage. It resembles a small cloud or band of mist above the beginning of the hamon line. The reflective particles are usually very fine. Utsuri then usually continues on above the hamon line. This effect is not very common. It occurs almost exclusively on fine 12th-13th century blades. This effect is directly related to the heating of the blade during hardening. Like mizukage, it occurs at the point where the blade undergoes significant temperature changes when heated, i.e. in the area behind the hamachi blade setting. On some blades, this may be displaced to the unpolished part of the nakago arbor. In many antique blades, these metallurgical features are then lost due to the shortening of the blade.

All of these effects, utsuri, nie utsuri, mizukage, koshiba, are found predominantly in swords of the Koto period. Particularly Kamakura period swords. They are typical of some sword schools and regions. For example, Bizen, Ichimonji. They are a sign of the sensitivity of the steel used and its ability to create metallurgical structures of various types. They also indicate the skill and ability of the sword makers who created such metallurgically rich blades. To a certain extent, these blades are indicative of the techniques of steel processing and hardening in the Koto period. By the Shinto period (after 1600) the ability to produce utsuri had essentially disappeared. Today, some sword makers have successfully reproduced this effect on their blades.


Utsuri is an interesting and beautiful manifestation of the metallurgical changes of steel during hardening. Two factors are essential for its formation. The first is the ability of the steel to produce these secondary manifestations of hardening. I have had the opportunity to work repeatedly with contemporary Japanese tamahagane and the blades produced very little or no utsuri. In contrast, on the rare occasion I have produced a blade from Japanese tamahagane produced before WWII, it has produced extremely strong khoji utsuri. I use tamahagane and oroshigane of my own production when making my blades, and consider utsuri to be a common manifestation of hardening. In my opinion, this is paradoxically due to the inferior quality or purity of the steel I produce. I produce steel in sponges of about 5kg for oroshigane and about 12kg for tamahagane. The most striking utsuri produces blades from smaller forged pieces. This may be due to surface contamination. While this is thoroughly cleaned during the reloading process, residual contamination by trace amounts of dirt can affect the sensitivity of the steel. Similarly, encapsulation of the welded steel package in straw ash affects the ability to produce richer metallurgy.

Tamahagane from current Japanese production was an excellent material to work with. The high quality steel pieces were clean, with no signs of oxidation, charcoal chunks or any other impurities. Perhaps it is this purity and perfection of the starting material that leads to a certain sterility in the creation of metallurgical structures. In trying to reproduce a Koto-style sword, it might have been interesting to use lower quality Japanese tamahagane.

The second factor is how the blade is heated. Utsuri is just the temperature range in which the steel recrystallized. Just like the hamon line where recrystallization into a martensite structure occurred. I'm mostly involved in reproducing Ichimonji School swords. In creating the Ichimonji style hamon line, clay paste is not used. This prevents water from reaching the steel during hardening and thus defines the boundary and pattern of the hamon line drawn on the blade by the clay swordsmith. Without the use of paste, it is necessary to heat the blade prior to hardening so that distinct temperature bands develop from the blade towards the spine. A temperature band at the blade at around 800°C recrystallises into martensite to form hamon. Another temperature band at a lower temperature above the hamon then forms utsuri. This principle works the same when quenching with or without paste. The more pronounced the temperature bands the swordsmith can create during blade heating, the more pronounced the utsuri will be. This method of heating the blade is very fast. However, the prominence and form of the utsuri depends on the sensitivity of the steel.

Personally, I consider utsuri to be a secondary effect. Although it is desirable in Koto style sword reproductions, it is not a good idea to focus too much on its production for a particular blade. I always try to produce a good hamon line first and foremost. When using the technique described above, the utsuri will almost always produce itself, without much effort to produce it. In experimentation, I have created a blade with a focus on producing a strong utsuri. While this has been successful, I do not consider this to be the correct direction.

Many swordsmiths nowadays place great emphasis on the application of clay paste. When hardening, they then heat the blade for a long time until they reach a uniform temperature throughout the blade. No temperature bands are created. This technique makes it possible to produce beautiful and well-controlled hamon lines. At the same time, however, uniform heating of the blade throughout the entire profile greatly reduces the possibility of utsuri. To a limited extent, utsuri can be achieved with paste by thinning the clay layer above the hamon line.

The presence of utsuri can be considered as a manifestation of blade qualities. In my opinion, however, it is more a manifestation of aesthetic qualities, or the quality of approaching visually the metallurgy of Koto swords in an attempt to reproduce them. The mere presence of utsuri does not seem to affect the utility of the blade in any way. Likewise, a blade without utsuri may be superior in quality to a blade with utsuri. It always depends on the particular blade and other aspects of steel working and metallurgy must be considered.

Polishing also has a significant effect on the distinctiveness of the Utsuri. Many blades have some degree of utsuri present. However, it is not noticeable due to the polishing method. It is up to the polisher to determine what qualities of the blade he or she wants to bring out.