Chemistry Reference
In-Depth Information
chip) in 1960. Information-processing capability is directly linked to the number
of components (transistors, capacitors, and resistors) that can be manufactured on
the silicon substrate. Enabling technologies for chip manufacture are, for exam-
ple, thin film deposition, etching, and lithography. Lithography enables the chip
designer to transfer the electronic circuitry “blueprint” that is written on a photo-
mask onto the silicon wafer.
There are two types of photoresists: negative and positive photoresists. The neg-
ative resist reacts upon exposure to light to form an insoluble form of the photore-
sist. Addition of developer then dissolves the unexposed regions of the resist. A
positive resist reacts when irradiated to produce a soluble form of the resist that
is then washed away by the developer.
A typical lithographic positive photoresist process can be summarized as fol-
lows: Step 1: A photoresist is spin-coated onto a silicon wafer. Step 2: The photore-
sist is exposed by shining light through the photomask. Step 3: The photoresist is
developed by dissolving away the unreacted photoresist. Step 4: The exposed oxide
is etched away. Step 5: The remaining photoresist is stripped to give a patterned
silicon wafer. While there are several types of lithography (e.g., X-ray and electron-
beam), the electronics industry has invested much of its capital in UV-lithography.
The ability to pack more components on a chip is related to the feature size of
the components themselves. While the 1K DRAM chip relied on feature sizes of
around 10
m, the 256 M chip has been manufactured using feature sizes of
0.35
m, almost a two order of magnitude decrease. The industry has responded
to the call for smaller feature sizes by moving to shorter and shorter wavelength
UV light, from 436 nm (G-line) to 365 nm (I-line), and more recently to 248 nm
light. This is because resolution (
R
) is directly proportional to the wavelength (
)
of the incident irradiation [80]. For G- and I-line irradiation, positive-tone novolac-
diazanaphthoquinone photoresists are the industry standard.
Advances in photoresist technology have made the move to 248 nm wavelengths
possible. The use of novolac-diazanaphthoquinone resists at this wavelength is
limited by the system's strong UV absorbance. Poly(hydroxystyrene) (or PHS) was
found to be a suitable polymer backbone around which the 248 nm photoresists
could be built. PHS is marginally transparent at 248 nm. Without sufficient trans-
parency, light cannot penetrate to the bottom of the resist film, resulting in only a
partially developed resist. As wavelengths get shorter, the sources become dim-
mer. Efficient use of photons becomes paramount.
The use of poly(norbornene) polymers such as Duvcor in photolithographic ap-
plications is possible because of their transparency in the deep UV region
(193 nm), and the ability of the nickel and palladium catalysts to polymerize nor-
bornenes bearing a variety of functional groups. However, simply being transpar-
ent at the wavelength of interest is only the first hurdle in choosing a candidate
polymer backbone. A commercial photoresist must not only undergo the appropri-
ate solubility switch upon irradiation, but the insoluble polymer that remains
after aqueous base development must “resist” decomposition under plasma etch-
ing conditions, i.e., conditions in which the exposed portions of the silicon wafer
are etched away. Otherwise 3-D features, and therefore components, could not be
Search WWH ::
Custom Search