1. Fundamentals
The basic structure and operating principle of this device are covered in the Fundamentals chapter Transistors and Memory. The following sections describe fabrication in detail.
The basic structure and operating principle of this device are covered in the Fundamentals chapter Transistors and Memory. The following sections describe fabrication in detail.
The basis for an n-channel field-effect transistor is a p-doped silicon substrate, using boron as the dopant.
Silicon dioxide SiO2 (the gate oxide, GOX for short) is grown on the substrate by dry oxidation. It insulates the gate, deposited later, from the substrate.
Nitride is deposited in an LPCVD process; it later serves as a mask during field oxidation.
A photoresist is applied on top of the nitride, exposed, and developed, producing a structured resist layer that serves as an etch mask.
The resist masks the nitride; the exposed areas are removed by reactive ion etching.
After the pattern has been transferred into the nitride, the resist is removed wet-chemically using a developer solution.
The nitride acts as a mask; thermal wet oxidation occurs only where the nitride has been removed. The field oxide (FOX, e.g. 700 nm) provides lateral isolation from neighbouring devices.
After oxidation, the nitride is removed in a wet-chemical etching process.
Polycrystalline silicon is deposited in an LPCVD process.
A resist layer is patterned on top of the polysilicon to serve as an etch mask.
The photoresist again acts as an etch mask, and the silicon is patterned in a reactive ion etch step, forming the gate electrode that controls the transistor.
After etching, the resist is again removed wet-chemically.
A thin oxide, the post-oxide, is deposited next. It protects the gate electrode and also acts as a spacer for the source/drain implantation.
In an implantation step using phosphorus ions, the source and drain regions are n-doped. Since the gate electrode acts as an implantation mask, defining the width of the n-channel between source and drain, this is referred to as self-alignment. By adjusting the width of the spacers, the distance between the implanted dopant atoms and the gate can be precisely controlled according to the electrical requirements.
An oxide (the interlayer oxide, ILD for short, e.g. 700 nm) is deposited to insulate against the metallization layers above. This is done in an LPCVD process using TEOS, which provides good step coverage.
A further resist layer is patterned, and the edges of the contact holes are rounded off in an isotropic etch process.
The contact holes are then opened down to the n-doped regions and to the gate in an anisotropic etch process.
The contact holes are filled with aluminium in a sputtering process.
In a final lithography step, another resist layer is patterned.
The resist pattern is transferred into the underlying metallization layer in an anisotropic dry etch step.
Finally, the resist is removed, leaving behind aluminium interconnects that drive the transistor.
The actual structure of a transistor is considerably more complex. Additional planarization layers may be used to support the lithographic processes, and several drain/source implantations may be carried out to precisely set the threshold voltage. Additional spacers at the gate are likewise possible, in order to fine-tune the channel length or influence the doping profile.