How a Katana Is Made: From Ore to Polished Blade

The smith, and why one name is misleading

The person who forges and hardens a Japanese sword blade is the swordsmith, 刀鍛冶 (katanakaji), also honored as 刀匠 (tosho) or 刀工 (toko). On a finished sword it is the smith's name that is signed into the tang, and it is the smith's reputation that a collector buys. That signature is the source of a persistent misunderstanding: that a katana is the work of one man in one forge.

It is not. A blade passes through several distinct trades, most of them lifelong specialties in their own right. The smith forges, folds, shapes, and hardens the steel. But the mirror surface, the visible grain, and the temper line are brought out by a separate specialist, the polisher. The metal collar that locks the blade into its mounts is made by another. The scabbard is carved by another. The hilt is wrapped by yet another. A smith of forty years may never have polished a blade to finish, and a master polisher may never have forged one.

This page follows the steel from raw material to finished, mounted blade, and names the specialist responsible at each stage. For the material itself, the iron-sand steel called tamahagane and the clay furnace it comes from, see our companion reference on tamahagane and the tatara furnace.

One more reason the single-name picture endures is legal and cultural. In Japan today a new sword blade can only be made lawfully by a government-licensed swordsmith working by traditional methods, and licensed smiths are limited in how many long blades they may produce. That licensing framework attaches the craft, and the fame, to the individual smith. It says nothing, though, about the polisher and the fitting makers, who remain essential and are simply less visible in the way blades are named and traded. Keeping the whole chain in view is the difference between admiring a sword and actually understanding how it came to exist.

Forge-welding and folding: orikaeshi tanren

The smith begins with a heterogeneous raw steel that varies in carbon content from piece to piece. The first task is to turn that inconsistent material into something uniform and clean enough to trust in a blade. The method is repeated forge-welding and folding, called 折り返し鍛錬 (orikaeshi tanren), literally forging by folding back. The broader forging step is also simply called 鍛錬 (tanren) or 鍛え (kitae).

The smith heats the stacked steel to welding temperature, hammers it into a single mass, then cuts a notch and folds the billet back on itself, welding it again under the hammer. Each cycle is one fold. The steel is folded lengthwise and crosswise in turn, which is why the finished grain, the hada, can run in different patterns depending on how the smith worked.

In genuine historical and modern licensed practice, a blade is folded on the order of ten to sixteen times. That is the number the physics rewards, and it is nowhere near the figures that circulate online.

There is a practical reason the folding is done in two directions. Folding only lengthwise would stack the layers like the pages of a book, and any slag or weakness would run in continuous lines down the blade. By alternating lengthwise and crosswise folds, the smith breaks those lines up and distributes any remaining inclusions in every direction, which is why a well-forged billet fails gracefully rather than splitting along a seam. The visible result is the surface grain, the hada, which a trained eye reads on a finished, polished blade as a signature of how the steel was worked. Different forging traditions produced recognizably different grain patterns, and the pattern is one of the features by which old blades are attributed.

Historically the raw steel a smith started with was not uniform to begin with. The smith would often separate the harder, higher-carbon pieces destined for the edge from the softer, lower-carbon material destined for the core, and forge and fold each separately before combining them. Folding, in other words, is not one operation applied once to one lump of metal. It is a sustained campaign of heating, welding, and refining, repeated on more than one batch of steel, before the blade is even shaped.

What folding actually does, and what it does not

Folding does three real, verifiable things. It homogenizes the steel, spreading carbon more evenly through a material that started out uneven. It disperses and reduces slag and non-metallic inclusions, driving impurities out or breaking them up so they no longer form weak points. And it builds up many thin layers, which contributes to the surface grain that connoisseurs read on a finished blade.

Here is the arithmetic that gets distorted. Each fold roughly doubles the number of layers, so the count grows as two to the power of the number of folds. Fifteen folds give two to the fifteenth, which is 32,768 layers. That is where the impressive layer counts come from, and they are honest as a description of layering.

Differential construction: hard skin, soft heart

A sword has to be two contradictory things at once: hard enough at the edge to take and hold a keen bevel, and tough enough in the body to survive impact without snapping. No single steel does both well. The traditional answer is to combine steels of different carbon content in one blade, a soft low-carbon core and a hard high-carbon skin. Several named constructions exist. Three are well attested in the historical record.

丸 (maru): one steel throughout

丸 (maru), also called 無垢 (muku), is mono-steel construction: a single grade of steel forms the whole blade, with no soft core. It is the simplest to make and relies entirely on differential hardening, covered below, to get a hard edge and a more forgiving spine out of one material.

甲伏せ (kobuse): hard jacket over soft core

甲伏せ (kobuse), roughly armor-covered, wraps a hard high-carbon skin, the 刃鉄 (hagane), over a softer low-carbon core, the 心鉄 (shingane). The hard skin takes the edge; the soft heart absorbs shock. This is among the most common traditional constructions and a sound engineering compromise.

本三枚 (honsanmai): a three-part sandwich

本三枚 (honsanmai), true three-plate, builds the blade from a hard edge steel, a soft core, and separate side plates, the 皮鉄 (kawagane). It is more labor and more forge-welds, and it lets the smith place each steel exactly where its properties are wanted.

You will see other named methods online, sometimes presented as ancient recipes. State the record plainly: terms such as a defined multi-layer Soshu construction are best understood as a stylistic tradition or school rather than a fixed lamination recipe, and some four-sided or split-edge constructions appear far more often in modern maker catalogs than in scholarly sword literature. Where a listing quotes anachronistic modern steel grades for a supposedly classical method, that is a sign the term is contemporary marketing, not documented history.

Drawing out the blank: sunobe

Once the steels are welded into their intended arrangement, the smith draws the billet out into a long flat bar in the rough proportions of a blade. This shaping stage is 素延べ (sunobe). From the sunobe the smith establishes the fundamental form: the taper from base to tip, the ridge line, the width, and the rough thickness. The point is shaped, the back and edge are defined, and files and a draw-knife-like tool refine the surfaces while the steel is still relatively soft, before hardening locks everything in place.

At the end of shaping the blade is essentially the right form but still soft, straight or only slightly curved, and unhardened. Everything that gives it its final character as a cutting instrument happens in the next two steps.

It is worth pausing on what has and has not happened at this point, because it corrects another common misconception. The blade already has its construction, its steels arranged as core and skin, and it already has its profile and its ridge geometry. What it does not yet have is its hardness or its final curve. A viewer who saw the blade now would recognize the shape of a sword but could bend it, and it would hold almost no edge. The transformation from a shaped bar of soft steel into a weapon is still ahead, and it happens in a few violent seconds in the quench. This is why the smith treats shaping as preparation and the quench as the true test: months of forging can be undone in an instant if the hardening goes wrong.

The coating that controls the fire: tsuchioki

Before hardening, the smith paints the blade with a clay slurry, a step called 土置き (tsuchioki), placing the clay. The mixture, made of clay with additions such as charcoal powder and stone powder and known as 焼刃土 (yakiba-tsuchi), is applied unevenly on purpose. It is laid thin along the cutting edge and thick over the spine and body.

The clay is an insulator. In the quench that follows, the thinly coated edge will lose heat almost instantly, while the thickly coated spine will cool much more slowly. That single difference in cooling rate is what produces both the hard edge and the temper line. The smith also uses the clay to draw the outline of the future temper line, so the pattern seen on a finished blade is in part a design decision made here with a brush.

Yaki-ire: the hamon and the curve are born here

The decisive moment is the differential quench, 焼き入れ (yaki-ire). Working in low light so the color of the glowing steel can be read precisely, the smith heats the clay-coated blade to a critical temperature and then plunges it into a trough of water. Because the edge is barely insulated and the spine is heavily insulated, the two regions cool at completely different rates, and two things happen at once.

First, the 刃文 (hamon) appears. This is the temper line, the visible boundary running the length of the blade between the hardened edge and the softer body. Its shape follows the clay line the smith painted, and it is one of the primary features by which a blade is judged and attributed.

Second, the blade gains its curvature, the 反り (sori). This is not bent in by hand. During the quench the edge transforms first and the blade momentarily flexes; then, as the differently structured spine finishes cooling, the edge region expands relative to the spine and the blade bends back, lifting the tip and deepening the curve. The graceful arc of a katana is in large part a byproduct of the physics of this quench. The result is checked and, if necessary, gently corrected afterward, but the curve itself is created in the water.

Water-quenching this way is unforgiving. The thermal shock can and does crack blades outright, and a cracked blade is scrapped. That risk is one honest reason a fully traditional blade is expensive: some percentage of the smith's work is lost at the quench.

Two steels in one blade, explained simply

Underneath the ceremony is straightforward metallurgy. Heat the steel high enough and its internal structure changes into a form called austenite that can hold carbon in solution. What that austenite becomes on cooling depends entirely on how fast it cools.

• The edge, cooled fast, becomes martensite. With almost no clay to slow it, the edge cools so quickly that the carbon is trapped in place, forming martensite, a very hard but brittle structure. This is what lets the edge take and hold a keen bevel.
• The spine, cooled slowly, stays softer. Insulated by thick clay, the spine cools slowly enough to form softer, pearlite-based structures rather than martensite. Precisely which structures form depends on the cooling rate, and metallurgists note the slow-cooled zone is typically pearlite along with intermediate phases, so the accurate description is softer, tougher, and more flexible than the edge.

The hamon is simply the visible frontier between these two worlds: hard martensite below the line, softer structure above it. That is why the temper line is not decoration applied to the surface but a record of where the steel changed phase. And because martensite occupies slightly more volume than the structures forming in the spine, the edge effectively grows a little relative to the back, which is the mechanism behind the curvature described above. One blade, two steels, produced not by welding two bars together but by cooling one bar at two different speeds.

The polisher and the stones: togishi

A blade leaves the smith hardened and shaped but dull and grey. Turning it into the mirror-and-mist object people picture is a separate profession, the polisher or 研ぎ師 (togishi). Polishing a sword is not sharpening in the kitchen sense. It is the establishment of the final geometry and the revealing of everything the steel contains, and on an important blade it can take weeks.

The work moves through a long progression of stones, from coarse foundation stones that set and correct the geometry, through a series of increasingly fine stones, to the finishing stages, the 仕上げ (shiage), where tiny stone fingertip fragments and burnishing tools are used. Only in these final stages do the grain of the folded steel, the hada, and the crystalline detail of the hamon become fully visible. A great polish can make a good blade look extraordinary; a poor polish can hide a fine blade's qualities or, worse, remove metal that can never be put back. This is why serious blades are entrusted only to trained polishers.

The other trades, and common questions

A polished blade is still not a finished sword. Mounting it draws in three more specialist crafts, each a lifelong trade.

Add the smith and the polisher to these three and you have at minimum five distinct specialists, often more, contributing to a single finished sword. The signature in the tang honors the smith, but the object in your hands is a collaboration.

Frequently asked questions

Who actually makes a katana?

A chain of specialists, not one person. The swordsmith (katanakaji) forges, folds, shapes, and hardens the blade. A separate polisher (togishi) sets the final geometry and surface. A collar maker (shirogane-shi), a scabbard maker (saya-shi), and a hilt wrapper (tsukamaki-shi) complete the mounted sword. Five or more trades contribute to one blade.

How many times is a katana folded?

Roughly ten to sixteen times in genuine practice. Each fold about doubles the layer count, so fifteen folds produce about 32,768 layers. The popular claim of folding a thousand or a million times is a myth: by around twenty folds the layering stops mattering, and further folding burns off carbon and harms the steel.

Does folding make a katana sharper?

No. Folding homogenizes the steel, disperses slag and impurities, and builds the grain, which improves consistency and toughness. Sharpness comes from blade geometry and the final edge work done by the polisher, not from the number of folds or layers.

Why is a katana curved?

The curve, called sori, is created mainly during the differential quench (yaki-ire). The clay-coated edge and spine cool at different rates, the edge hardens into martensite and expands slightly relative to the spine, and the blade bends, lifting the tip. It is a byproduct of the metallurgy, not a shape simply bent in by hand.

What creates the hamon, the wavy line on the blade?

The hamon is the visible boundary between the hard martensite of the quench-hardened edge and the softer, pearlite-based structure of the body. The smith controls its shape by painting insulating clay (tsuchioki) thin on the edge and thick on the spine before quenching. It is a record of where the steel changed phase, not a surface decoration.

Sources

Published reference material on Japanese swordsmithing and on hamon formation and differential hardening; nihonto reference and craft-association materials on construction methods, polishing, and sword fittings; and general metallurgical writing on the hardening and tempering of carbon steel. Where popular sources and scholarship disagree, in particular on folding counts and on certain named construction methods, we have followed the scholarly record and flagged the difference in the text.