How were the terracotta soldiers of Qin Shi Huang made?

The soldiers were built from separately formed parts, and each face was finished by hand, which is why the faces differ. One detail needs a correction: the main body parts were joined before firing; only heads were sometimes fired on their own.

Terracotta soldiers: separate sections, individual faces

The terracotta soldiers discovered near Xi'an in 1974 combine repeated forms with individual finishing. They belonged to the burial complex of Qin Shi Huang, who died in 210 BCE. Thousands of warriors, horses and accompanying objects gave the emperor a military presence in the afterlife. UNESCO emphasizes both their realistic differences and the evidence they preserve for pottery and bronze working. Understanding their manufacture means looking beneath the finished faces to the clay structure, firing, surface coatings and equipment.

The main body parts of the terracotta soldiers were joined before firing. Separate formation did not mean that every arm, leg and torso was baked independently and then attached. Bo Rong and colleagues' 2026 study of officer figurines compares their production with the warriors: craftspeople used molds to form sections, assembled them, let them dry and finally fired them in kilns. Modeling and joining therefore preceded the hardening of the main body. Painting belonged later in the process, after the ceramic form had been made durable by heat.

Heads were sometimes fired separately and attached afterward. Archaeologist Marcos Martinón-Torres's National Geographic manufacturing reconstruction illustrates that exception through its firing and assembly stages: a separate head is placed beside the body for firing, then fitted onto the warrior. This is a different distinction from making a head in its own mold. A separately shaped head could still be attached before firing; a separately fired head had already hardened. Keeping those operations distinct reconciles the body sequence with the qualification about heads.

The finished figures combined recognizable military types with varied features. UNESCO draws attention to their uniforms, weapons and even the horses' halters as historical evidence. These details let a viewer compare the equipment of a formation rather than seeing an undifferentiated crowd. The terracotta soldiers could share a general appearance while differing in their modeled faces and clothing. Craftspeople finished facial details by hand. Repeated components and individual finishing belonged to the same undertaking: standard forms supplied a framework within which craftspeople could change the final surface.

Material preparation also mattered. Patrick Sean Quinn and colleagues' 2017 microscopic study identified sand added to the clay of sampled warriors and acrobats. The researchers suggest that this temper made fine clay easier to shape and opened pores through which water could escape. That helps explain why a manufacturing account needs more than a description of face molds. A convincing face depended on a clay body that could be built, dried and heated successfully; the ingredients were part of that process.

Drying was another controlled stage. Quinn's team found evidence of careful drying in the thick ceramic walls before they were heated. They also favored a kiln that heated gradually over an uncontrolled open fire. Those are interpretations of material evidence, rather than a surviving instruction sheet from a Qin workshop. Their practical implication is clear: producing a large figure required attention to its interior as well as its visible outline. Care during preparation and heating helped turn joined clay sections into a standing sculpture.

Color completed the terracotta soldiers. Rong and colleagues describe a lacquer ground beneath mineral pigments, a layered system shared by the officer figurines and the warriors. Red cinnabar, green malachite, blue azurite and black carbon belonged to this palette. The lacquer and pigment layers were distinct from the structural clay beneath them. Their sensitivity to changes in humidity helps explain the fragile painted surfaces encountered during excavation. An apparently plain clay warrior can therefore preserve only part of the original combination of modeled form and applied color.

The equipment was a further manufacturing task. The 2019 study by Martinón-Torres and colleagues describes functional bronze weapons, including spears, lances, swords and crossbow triggers. Its account of the partial excavation of Pit 1 lists up to 40,000 arrowheads, typically in bundles of 100 associated with individual quivers. Little remains of the wooden shafts, scabbards and other organic components. The surviving bronze objects are therefore parts of originally composite weapons, just as the exposed clay is only one layer of the finished sculpture.

That same study reassessed chromium detected on some weapons. Its analyses linked the chromium on the metal to lacquer contamination during burial, rather than an intentional protective coating applied by weapon makers. The authors instead pointed to the bronze alloy and the fine-textured, alkaline ground as likely contributors to preservation. A chemical trace on an excavated object need not describe an ancient production step. Here the distinction separates the techniques used to make the weapons from changes that occurred while they lay beside the painted figures underground.

The scale of the work did not require every terracotta soldier to pass through a single workshop. Martinón-Torres describes small workshop groups producing figures at the same time, with painting completed before transport to the pits. Weapon production could proceed in other armories before the equipment was assigned to the figures. This reconstruction connects repeated forms with skilled handwork and coordinated delivery. The manufacturing achievement lay in bringing suitable clay, formed sections, detailed faces, controlled firing, painted surfaces and real equipment together in an ordered army.

Terracotta soldiers: separate sections, individual faces

Terracotta soldiers: separate sections, individual faces

The Question

The Answer

Read next