4 Silicon Plasmonics
159
4.2.3.2 Silicon Plasmonic Detectors
At the end of each photonic/plasmonic communication link there is a need for a
detector which “converts” photons to electrons. While silicon detectors are widely
used in everyday life, they are not efficient for the detection of near-infrared signals
in the telecom regime. This is because the energy band gap of silicon is in the order
of 1.1 eV, whereas the energy of a photon in the telecom regime is only about 0.8 eV.
In other words, while the transparency of silicon at the telecom frequency band
allows its deployment as the material of choice for guiding, modulation and other
manipulations of light, its very same property prohibit silicon from being used as an
active absorbing material for detection of infrared optical signals. Over the years, the
silicon photonics industry developed approaches for circumventing this deficiency
by considering solutions such as the integration of germanium active layer with silicon platform [39–41], and fabrication of InGaAs/Si avalanche photodetectors using
wafer bonding technology [42, 43]. While these approaches provide a path towards
detection of light at the telecom wavelength regime, there is still a clear advantage in
developing monolithic CMOS compatible devices for the detection of light in silicon,
without the need to rely on other material systems. Recently, several approaches were
proposed and demonstrated along these lines including two-photon absorption [44,
45], insertion of midbandgap defect states into silicon lattice [46], using a polysilicon active layer [47] and cavity enhanced photocurrent generation [45, 47, 48].
Yet, another promising approach to detect infrared sub-bandgap optical radiation
in silicon is to employ the internal photoemission (IPE) process using a Schottky
barrier (SB) photodetector [49, 50]. In its simplest form, such a detector consists of
metal film on a lightly doped semiconductor (e.g. silicon) forming a Schottky contact
at metal-semiconductor interface with potential barrier ε B and rectifying electrical
characteristics. Typically, the obtained Schottky barrier (ε B ) is lower than the energy
bandgap of silicon [50], thus allowing detection of long-wavelength (infrared) photons via the internal photoemission process. More specifically, when optical radiation
below the bandgap is applied to the metal-silicon contact by top (trough the metal)
or back (through the semiconductor) illumination, the conduction electrons in the
metal absorb infrared photons with energy hλ exceeding the potential barrier at the
interface ε B (see Fig. 4.7). Gaining sufficient energy, these excited (hot) electrons
are able to cross over the SB [50], sweep out the depletion region of the semiconductor and be collected as a photocurrent under reverse bias (photoconductive mode)
operation.
A fundamental deficiency of the conventional SB based photodetector is related
to the fact that the volume in which the photons interact with electrons in the metal
is very small, namely only a small fraction of the incident photons actually causes
photoemission. Clearly, there is a need to confine the optical power at the boundary
between the materials forming the Schottky contact, thereby increasing the interaction of light with the metal in the vicinity of the interface where the photoemission
process takes place. This light localization could significantly improve the detection
capability of the system and potentially pave the way for device miniaturization and
realization of on-chip photodetectors on the nanoscale. As early as in 1970s the con-
159
4.2.3.2 Silicon Plasmonic Detectors
At the end of each photonic/plasmonic communication link there is a need for a
detector which “converts” photons to electrons. While silicon detectors are widely
used in everyday life, they are not efficient for the detection of near-infrared signals
in the telecom regime. This is because the energy band gap of silicon is in the order
of 1.1 eV, whereas the energy of a photon in the telecom regime is only about 0.8 eV.
In other words, while the transparency of silicon at the telecom frequency band
allows its deployment as the material of choice for guiding, modulation and other
manipulations of light, its very same property prohibit silicon from being used as an
active absorbing material for detection of infrared optical signals. Over the years, the
silicon photonics industry developed approaches for circumventing this deficiency
by considering solutions such as the integration of germanium active layer with silicon platform [39–41], and fabrication of InGaAs/Si avalanche photodetectors using
wafer bonding technology [42, 43]. While these approaches provide a path towards
detection of light at the telecom wavelength regime, there is still a clear advantage in
developing monolithic CMOS compatible devices for the detection of light in silicon,
without the need to rely on other material systems. Recently, several approaches were
proposed and demonstrated along these lines including two-photon absorption [44,
45], insertion of midbandgap defect states into silicon lattice [46], using a polysilicon active layer [47] and cavity enhanced photocurrent generation [45, 47, 48].
Yet, another promising approach to detect infrared sub-bandgap optical radiation
in silicon is to employ the internal photoemission (IPE) process using a Schottky
barrier (SB) photodetector [49, 50]. In its simplest form, such a detector consists of
metal film on a lightly doped semiconductor (e.g. silicon) forming a Schottky contact
at metal-semiconductor interface with potential barrier ε B and rectifying electrical
characteristics. Typically, the obtained Schottky barrier (ε B ) is lower than the energy
bandgap of silicon [50], thus allowing detection of long-wavelength (infrared) photons via the internal photoemission process. More specifically, when optical radiation
below the bandgap is applied to the metal-silicon contact by top (trough the metal)
or back (through the semiconductor) illumination, the conduction electrons in the
metal absorb infrared photons with energy hλ exceeding the potential barrier at the
interface ε B (see Fig. 4.7). Gaining sufficient energy, these excited (hot) electrons
are able to cross over the SB [50], sweep out the depletion region of the semiconductor and be collected as a photocurrent under reverse bias (photoconductive mode)
operation.
A fundamental deficiency of the conventional SB based photodetector is related
to the fact that the volume in which the photons interact with electrons in the metal
is very small, namely only a small fraction of the incident photons actually causes
photoemission. Clearly, there is a need to confine the optical power at the boundary
between the materials forming the Schottky contact, thereby increasing the interaction of light with the metal in the vicinity of the interface where the photoemission
process takes place. This light localization could significantly improve the detection
capability of the system and potentially pave the way for device miniaturization and
realization of on-chip photodetectors on the nanoscale. As early as in 1970s the con-