U.S. patent application number 16/453349 was filed with the patent office on 2020-12-31 for enhancement mode saddle gate device.
The applicant listed for this patent is Northrop Grumman Systems Corporation. Invention is credited to Josephine Chang, Robert S. Howell, Ken Nagamatsu, Sarat Saluru.
Application Number | 20200411676 16/453349 |
Document ID | / |
Family ID | 1000005272725 |
Filed Date | 2020-12-31 |
United States Patent
Application |
20200411676 |
Kind Code |
A1 |
Chang; Josephine ; et
al. |
December 31, 2020 |
ENHANCEMENT MODE SADDLE GATE DEVICE
Abstract
An enhancement-mode (e-mode) field effect transistor (FET)
comprises a buffer layer, and a superlattice of conducting channels
on the buffer layer and including a trench that cuts down through
the superlattice into the buffer layer and separates the
superlattice into a source-access region and a drain-access region,
wherein the buffer layer forms a bottom of the trench. The e-mode
FET includes a source and a drain adjacent to the source-access
region and the drain-access region, respectively. The e-mode FET
further incudes a gate in the trench, such that a voltage above a
threshold voltage of the e-mode FET applied to the gate induces a
current channel in the buffer layer underneath the gate, which
electrically connects the source-access region to the drain-access
region to turn on the e-mode FET, and (ii) a voltage below the
threshold voltage applied to the gate eliminates the current
channel, which electrically disconnects the source-access region
from the drain-access region to turn off the e-mode FET.
Inventors: |
Chang; Josephine; (Ellicott
City, MD) ; Nagamatsu; Ken; (Ellicott City, MD)
; Howell; Robert S.; (Silver Spring, MD) ; Saluru;
Sarat; (Baltimore, MD) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Northrop Grumman Systems Corporation |
Falls Church |
VA |
US |
|
|
Family ID: |
1000005272725 |
Appl. No.: |
16/453349 |
Filed: |
June 26, 2019 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H01L 21/8252 20130101;
H01L 29/4236 20130101; H01L 27/0883 20130101; H01L 29/155 20130101;
H01L 21/28264 20130101; H01L 29/2003 20130101; H01L 29/66462
20130101; H01L 29/42376 20130101; H01L 29/7785 20130101; H01L
29/205 20130101 |
International
Class: |
H01L 29/778 20060101
H01L029/778; H01L 27/088 20060101 H01L027/088; H01L 29/15 20060101
H01L029/15; H01L 29/20 20060101 H01L029/20; H01L 29/205 20060101
H01L029/205; H01L 29/423 20060101 H01L029/423; H01L 21/28 20060101
H01L021/28; H01L 21/8252 20060101 H01L021/8252; H01L 29/66 20060101
H01L029/66 |
Claims
1. An enhancement-mode (e-mode) field effect transistor (FET),
comprising: a buffer layer; a superlattice of conducting channels
on the buffer layer and including a trench that cuts down through
the superlattice into the buffer layer and separates the
superlattice into a source-access region and a drain-access region
of the superlattice, wherein the buffer layer forms a bottom of the
trench, wherein the superlattice does not include any ridges formed
in the superlattice and does not include any trenches formed in the
superlattice other than the trench; a source and a drain adjacent
to the source-access region and the drain-access region,
respectively, so that the source and the drain are each
spaced-apart from the trench; and a gate disposed in the trench and
configured such that (i) a voltage above a threshold voltage of the
e-mode FET applied to the gate induces in the buffer layer
underneath the gate a current channel, which electrically connects
the source-access region to the drain-access region to turn on the
e-mode FET, and (ii) a voltage below the threshold voltage applied
to the gate eliminates the current channel, which electrically
disconnects the source-access region from the drain-access region
to turn off the e-mode FET.
2. The e-mode FET of claim 1, wherein the buffer layer is an
un-doped buffer layer.
3. (canceled)
4. The e-mode FET of claim 1, wherein the trench has a length that
separates the source-access region from the drain-access region
along a length of the trench that is orthogonal to a width of the
trench.
5. The e-mode FET of claim 4, wherein the source and the drain are
spaced-apart from each other and the trench in a direction of the
length.
6. The e-mode FET of claim 4, wherein: an edge of the source-access
region and an edge of the drain-access region that are spaced-apart
from each other across the length of the trench form a source-side
trench sidewall and a drain-side trench sidewall that rise from the
bottom of the trench to a topside of the superlattice,
respectively; and the gate fills the trench such that the gate has
a source-side gate sidewall and a drain-side gate sidewall adjacent
to the source-side trench sidewall and the drain-side trench
sidewall, respectively, and a gate bottom adjacent the bottom of
the trench, respectively.
7. The e-mode FET of claim 6, wherein: the gate is formed as a
T-gate that fills the trench completely and covers a periphery of
the edge of the source-access region and a periphery of the edge of
the drain-access region on the topside of the superlattice.
8. The e-mode FET of claim 6, wherein: the source-side trench
sidewall and the drain-side trench sidewall are parallel to the
source-side gate sidewall and the drain-side gate sidewall,
respectively, as the trench sidewalls and the gate sidewalls rise
from the bottom of the trench to the topside of the
superlattice.
9. The e-mode FET of claim 8, wherein: the gate sidewalls rise in a
direction that is vertical with respect to a direction of the width
and a direction of the length.
10. The e-mode FET of claim 6, wherein: the source-side trench
sidewall and the drain-side trench sidewall taper away from the
source-side gate sidewall and the drain-side gate sidewall as the
source-side trench sidewall and the drain-side trench sidewall rise
from the bottom of the trench to the topside of the superlattice,
respectively.
11. The e-mode FET of claim 6, further comprising a dielectric
layer between the gate sidewalls and respective ones of the trench
sidewalls, and between the gate bottom and the bottom of the
trench.
12. The e-mode FET of claim 1, wherein: the gate is formed as a
T-gate that fills the trench.
13. The e-mode FET of claim 1, wherein the conducting channels of
the superlattice include two-dimensional electron gas (2DEG)
channels.
14. The e-mode FET of claim 1, wherein the e-mode FET is
constructed on a first portion of the superlattice of conducting
channels that is adjacent to a second portion of the superlattice
of conducting channels on which a depletion-mode (d-mode) FET is
constructed.
15. The e-mode FET of claim 14, wherein the d-mode FET includes a
superlattice castellated FET (SLCFET).
16. A method of forming an enhancement-mode (e-mode) field effect
transistor (FET), comprising: forming a buffer layer; forming on
the buffer layer a superlattice of conducting channels including a
trench that cuts down through the superlattice into the buffer
layer and separates the superlattice into a source-access region
and a drain-access region of the superlattice, wherein the buffer
layer forms a bottom of the trench, and wherein the superlattice
does not include any ridges formed in the superlattice and does not
include any trenches formed in the superlattice other than the
trench; forming a source and a drain adjacent to the source-access
region and the drain-access region, respectively, so that the
source and the drain are each spaced-apart from the trench; and
forming a gate in the trench, the gate configured such that (i) a
voltage above a threshold voltage of the e-mode FET applied to the
gate induces in the buffer layer underneath the gate a current
channel, which electrically connects the source-access region to
the drain-access region to turn on the e-mode FET, and (ii) a
voltage below the threshold voltage applied to the gate eliminates
the current channel, which electrically disconnects the
source-access region from the drain-access region to turn off the
e-mode FET.
17. The method of forming the e-mode FET of claim 16, wherein: the
forming the superlattice includes forming the superlattice so that
an edge of the source-access region and an edge of the drain-access
region that are spaced-apart from each other across the trench form
a source-side trench sidewall and a drain-side trench sidewall that
rise from the bottom of the trench to a topside of the
superlattice, respectively; and the forming the gate includes
forming the gate fill the trench such that the gate has a
source-side gate sidewall and a drain-side gate sidewall adjacent
to the source-side trench sidewall and the drain-side trench
sidewall, respectively, and a gate bottom adjacent the bottom of
the trench, respectively.
18. The method of forming the e-mode FET of claim 17, wherein the
forming the gate further includes forming the gate as a T-gate that
fills the trench completely and covers a periphery of the edge of
the source-access region and a periphery of the edge of the
drain-access region on the topside of the superlattice.
19. An integrated circuit comprising: a buffer layer; a
superlattice of conducting channels on the buffer layer; an e-mode
field effect transistor (FET) constructed on a first portion of the
superlattice; and a superlattice castellated FET (SLCFET)
constructed on a second portion of the superlattice that is
spaced-apart from the first portion of the superlattice; wherein,
the e-mode FET includes: in the first portion of the superlattice,
a trench that cuts down through the first portion of the
superlattice into the buffer layer and that separates the first
portion of the superlattice into a source-access region and a
drain-access region of the first portion of the superlattice,
wherein the buffer layer forms a bottom of the trench; a source and
a drain adjacent to the source-access region and the drain-access
region; and a gate disposed in the trench and configured such that
(i) a voltage above a threshold voltage of the e-mode FET applied
to the gate induces in the buffer layer underneath the gate a
current channel, which electrically connects the source-access
region to the drain-access region to turn on the e-mode FET, and
(ii) a voltage below the threshold voltage applied to the gate
eliminates the current channel, which electrically disconnects the
source-access region from the drain-access region to turn off the
e-mode FET.
20. The integrated circuit of claim 19, wherein the SLCFET
includes: a source-access region and a drain-access region of the
second portion of the superlattice that are spaced-apart from each
other in a first direction; alternating multichannel ridges and
trenches formed in the second portion of the superlattice that
extend in parallel with each other in the first direction from the
source-access region to the drain-access region, and that undulate
in a second direction transverse to the first direction; a source
and a drain on the source-access region and the drain-access region
of the SLCFET, respectively; and a gate covering a portion of the
multichannel ridges and trenches.
Description
TECHNICAL FIELD
[0001] The present disclosure relates generally to transistor
devices and more specifically to field effect transistor (FET)
devices on superlattice structures.
BACKGROUND
[0002] While enhancement mode ("e-mode") Gallium Nitride (GaN)
field effect transistors (FETs) and high electron mobility
transistors (HEMTs) have been demonstrated, their conventional
device structures are not easily integrated into a superlattice
castellated FET (SLCFET) process flow. Moreover, methods to dope or
modify the surface and/or charge profile of the SLCFET to shift its
threshold voltage in a positive direction may not be easily
integrated into a standard depletion-mode (d-mode) SLCFET process
flow. Thus, constructing an integrated circuit (IC) that combines
such conventional e-mode and d-mode devices using a common SLCFET
process flow presents substantial technical and cost
challenges.
SUMMARY OF THE INVENTION
[0003] An enhancement-mode (e-mode) field effect transistor (FET)
comprises a buffer layer, and a superlattice of conducting channels
on the buffer layer and including a trench that cuts down through
the superlattice into the buffer layer and separates the
superlattice into a source-access region and a drain-access region
of the superlattice, wherein the buffer layer forms a bottom of the
trench. The e-mode FET includes a source and a drain adjacent to
the source-access region and the drain-access region, respectively,
so that the source and the drain are each spaced-apart from the
trench. The e-mode FET further incudes a gate disposed in the
trench and configured such that (i) a voltage applied to the gate
above a threshold voltage of the e-mode FET induces in the buffer
layer underneath the gate a current channel, which electrically
connects the source-access region to the drain-access region to
turn on the e-mode FET, and (ii) a voltage below the threshold
voltage applied to the gate eliminates the current channel, which
electrically disconnects the source-access region from the
drain-access region to turn off the e-mode FET.
[0004] An example embodiment of the invention is described below
with reference to the following drawing figures, in which like
reference numerals in the various figures are utilized to designate
like components.
BRIEF DESCRIPTION OF THE DRAWINGS
[0005] FIG. 1 is a top plan view of an example enhancement mode
("e-mode") saddle-gate field effect transistor (FET) (referred to
herein as an "e-mode FET") constructed on a superlattice
structure.
[0006] FIG. 2 is a cross-sectional view of the example e-mode FET
taken along the line 2-2 in FIG. 1.
[0007] FIG. 3 is a cross-sectional view of the example e-mode FET
taken along the line 3-3 of FIG. 1.
[0008] FIG. 4 is a cross-sectional view of the example e-mode FET
taken along the line 3-3 of FIG. 1, when the e-mode FET is
configured according to an alternative embodiment that employs
tapered trench sidewalls.
[0009] FIG. 5 is an illustration of the example integrated circuit
including the example e-mode FET and an example superlattice
castellated field effect transistor (SLCFET) constructed on
adjacent portions of a superlattice-based common structure.
[0010] FIG. 6 is a perspective of view of the example SLCFET in an
area 6-6 of FIG. 5.
[0011] FIG. 7 is a perspective view illustrating an example
epitaxial structure formed in initial stages of fabricating the
example e-mode FET.
[0012] FIG. 8 is a perspective view illustrating the example
epitaxial structure of FIG. 7 after a trench has been formed in the
epitaxial structure.
[0013] FIG. 9 is a perspective view illustrating the example
epitaxial structure of FIG. 8 after a conformal dielectric has been
formed on the epitaxial structure.
[0014] FIG. 10 is a perspective view illustrating the example
epitaxial structure of FIG. 7 after electrically conductive gate,
source, and drain contacts have been formed on the epitaxial
structure.
[0015] FIG. 11 is a flowchart of an example method of making the
example e-mode FET.
DESCRIPTION OF EXAMPLE EMBODIMENTS
[0016] Embodiments presented herein are directed to a structure and
a method of making an enhancement mode (e-mode) saddle gate field
effect transistor (FET) (more generally referred to as an "e-mode
FET") constructed on a base structure and a superlattice of
conducting layers. Other embodiments include a structure and a
method of making an integrated circuit (IC) including both the
e-mode FET and a depletion mode (d-mode) FET, such as a
superlattice castellated FET (SLCFET), constructed on a common base
structure and superlattice.
[0017] Reference is now made to FIGS. 1-3. FIG. 1 is a top plan
view of an example e-mode FET 100 constructed on a superlattice
structure. FIGS. 2 and 3 are cross-sectional views of e-mode FET
100 taken along the lines 2-2 and 3-3 in FIG. 1. As best shown in
FIGS. 2 and 3, e-mode FET 100 includes a substrate layer 101 (e.g.,
Aluminum Gallium Nitride (AlGaN)), an insulating buffer layer 102
on the substrate layer, and a superlattice of parallel conducting
layers 104 (also referred to simply as superlattice 104) on the
buffer layer. Buffer layer 102 is an insulating layer configured to
enable high quality epitaxy. Buffer layer 102 may be a
semiconductor, but need not be a semiconductor. Un-doped GaN is a
good choice for buffer layer 102 because it is lattice matched to
GaN; however, a small amount of doping may be introduced into the
GaN (in which case buffer layer 102 is only lightly doped), such as
Aluminum (Al), Iron (Fe), or Carbon (C) to make the buffer layer
more insulating. The term "on" used herein with respect to a first
layer being "on" a second layer means that the first layer is
adjacent to, on top of, overlying, or covering the second layer,
and may be either in direct contact, or not in direct contact, with
the second layer (i.e. there may be one or more intervening layers
between the first and the second layers). In the embodiment of
FIGS. 1-3, substrate layer 101, buffer layer 102, and superlattice
104 are stacked one on top of the other in a vertical direction,
and each layer is in contact with the underlying layer.
[0018] FET 100 also includes a gate trench 106 that cuts down
vertically completely through superlattice 104 and into buffer
layer 102, such that the buffer layer forms a bottom B of the
trench. Trench 106 has a width W that defines an active width of
FET 100, a height H, and a length L that defines a gate length of
the FET, which is less than the width W in the example of FIG. 1.
Length L, Width W, and height H extend in x (length), y (width),
and z (height/vertical) directions, respectively, that are
transverse/orthogonal to each other. The cross section of trench
106 may be substantially rectangular, rounded, or shaped in
accordance with other cross section shapes. Trench 106 completely
separates or divides superlattice 104 across length L of the
trench, and along width W of the trench, into a source-access
region (SAR) 108 on a left-side of the trench and a drain-access
region (DAR) 110 on a right-side of the trench, as shown in FIGS. 1
and 2. In an example, trench 106 has a length of 10-500 nm and a
width of 1-500 um, buffer layer 102 has a depth of 50-100 nm, the
trench cuts down a depth 50-200 nm into the buffer layer, and 1-10
repetitions of this structure are electrically tied together to
form a multi-finger FET. Also depicted in FIG. 1 is a boundary-line
that defines an isolation region rectangle 107. The region outside
of/surrounding isolation region rectangle 107 (but not inside of
the rectangle) is an isolation region that is implanted to be
isolating. An active width of FET 100 is determined by the
isolation region, which is smaller than width W of trench 106.
[0019] FET 100 also includes gate contact G1 (also referred to as
gate G1) that fills width W of trench 106 along length L of the
trench, source contact S1 (also referred to as source S1) embedded
in/adjacent to source-access region 108 and spaced-apart from the
trench, and drain contact D1 (also referred to as drain D1)
embedded in/adjacent to drain-access region 110 and spaced-apart
from the trench. Gate G1 and trench 106 extends past an active
region of superlattice 104. Thus, source S1 and drain D1 are
positioned so that the source and the drain are spaced-apart from
each other in the length direction, i.e., on opposing sides of the
gate G1, and so that the source and the drain contact the parallel
conducting layers of the superlattice, described below. FET 100
also includes a dielectric layer 116 (e.g., Silicon Nitride
Si.sub.3N.sub.4) on superlattice 104 that covers source-access
region 108 between source S1 and trench 106, covers drain-access
region 110 between drain D1 and the trench, and covers/lines the
trench to form a dielectric layer between the trench and gate
G1.
[0020] As best shown in FIG. 3, superlattice 104 includes multiple
parallel heterostructures 304(1)-304(5) (collectively referred to
as "heterostructures 304") overlying one another over along a
height of the superlattice to form a vertical stack of horizontal
(parallel) heterostructures on buffer layer 102. Each
heterostructure forms a corresponding one of parallel conducting
layers or sheets of superlattice 104. Each heterostructure 304(i)
respectively includes a first layer 306(1) and a second layer
306(2) overlying the first layer so as to form a respective one of
the conducting layers of the superlattice at an interface between
the two layers. While a stack of 5 heterostructures are shown, it
is understood that a stack of more or fewer heterostructures (and
resulting superlattice layers) may be used.
[0021] In an example, first layer 306(1) of each heterostructure
304(i) comprises GaN, and second layer 306(2) of each
heterostructure 304(i) comprises AlGaN; however, a variety of
heterostructures may be employed as long as each heterostructure
comprises two layers of dissimilar materials configured to create a
sheet of electrons (i.e. a Two-Dimensional Electron Gas ("2DEG")
layer) or a sheet of holes (i.e., a Two-Dimensional Hole Gas
("2DHG") layer 2DHG) at the interface between the two dissimilar
materials. Various additional heterostructure materials include,
but are not limited to, Aluminum Gallium Arsenide (AlGaAs) and
Gallium Arsenide (GaAs), Indium Aluminum Nitride (InAlN) and GaN,
and alloys of Silicon (S1) and Germanium (Ge) overlying a base
structure.
[0022] Still referring primarily to FIG. 3, substantially vertical
edges of source-access region 108 and drain-access region 110 form
substantially vertical source-side and drain-side edges or
sidewalls 308 and 310 of trench 106 (i.e., source-side and
drain-side trench sidewalls 308 and 310), respectively,
spaced-apart by length L. Trench sidewalls 308 and 310 rise
vertically from trench bottom B and past a topside 312 of
superlattice 104. Gate G1 includes a gate post 314 and a gate top
316 connected to the post and that together form a T-gate. Gate
post 314 is shaped and sized so that the post has a gate length
that completely fills trench 106 (as covered by dielectric layer
116), and so that the post covers trench bottom B and trench
sidewalls 308 and 310. Gate top 316 spills over opposing
source-side and drain-side peripheries 318, 320 of superlattice
topside 312, adjacent trench sidewalls 308, 310, respectively. Gate
post 314 has substantially vertical opposing source-side and
drain-side gate sidewalls 322, 324 adjacent substantially vertical
source-side and drain-side trench sidewalls 308, 310, respectively.
That is, trench sidewalls 308, 310 and gate sidewalls 322, 324 are
parallel to each other as they rise through superlattice 104. Thus,
a respective separation between each trench sidewall 308, 310 and
each gate sidewall 322, 324, respectively, which is equal to a
thickness/length of intervening dielectric layer 116, is
substantially constant as the trench sidewalls rise from trench
bottom B to and past superlattice topside 312.
[0023] Operation of FET 100 is now described. Trench 106 physically
severs the conducting layers of superlattice 104 in source-access
region 108 from those in drain-access region 110. When a voltage
(i.e., a "gate voltage") less than a threshold voltage of FET 100
is applied to gate G1, no current can flow from source S1 to drain
D1 through superlattice 104 because source S1 is cut-off from drain
D1, and there is no path for current to flow from the source to the
drain. Thus, a sub-threshold voltage applied to gate G1 turns off
FET 100.
[0024] In contrast, when a voltage greater than (i.e. above) a
threshold voltage of FET 100 is applied to gate G1, where the
threshold voltage may be zero or positive, for example, band
bending in buffer layer 102 below superlattice 104 results in an
accumulation of a channel of electrons AL (also referred to as a
"charge accumulation layer" and a "current channel") at an
interface of the buffer layer just below and near gate G1. The
shape of current channel AL is similar to that of a saddle beneath
gate G1, hence the name "saddle" to identify FET 100 as an "e-mode
saddle gate FET 100." Thus, the voltage above the threshold voltage
applied to the gate induces current channel AL in buffer layer 102
just underneath trench bottom B and near gate 214. Current channel
AL electrically connects source-access region 108 to drain-access
region 110, and provides a current path through which current flows
from the source-access region (and source S1) to the drain-access
region (and drain D1), to turn on FET 100. In other words, current
flows from source S1 to drain D1 through (i) source-access region
108 (i.e., the conducting layers of superlattice 104 in the
source-access region), (ii) current channel AL, and (iii)
drain-access region 110 (i.e., conducting layers of the
superlattice in the drain-access region). Current channel AL
represents an active channel that switches off when the gate
voltage is biased below (i.e., to be less than) the threshold
voltage to electrically disconnect source S1 from drain D1, and
switches on when the gate voltage is above the threshold voltage to
electrically connect the source to the drain. When the voltage
applied to gate G1 is equal to the threshold voltage, FET 100 is in
transition between conducting and not conducting as described
above. The conducting layers of superlattice 104 are part of the
source and drain access regions, not the active channel.
[0025] The gate-trench sidewall arrangement of the embodiment shown
in FIGS. 2 and 3 can result in a relatively high gate-to-source and
gate-to-drain capacitances. Described below is an e-mode FET
according to an alternative embodiment configured to reduce the
gate-to-source and the gate-to-drain capacitances relative to the
embodiment shown in FIGS. 2 and 3, described above.
[0026] With reference to FIG. 4, there is a cross section view
taken along the cut line 3-3 of the e-mode FET of the alternative
embodiment. The alternative embodiment of the e-mode FET is similar
to the embodiment shown in FIGS. 2 and 3, except for the
differences described below. In the embodiment of FIG. 4, trench
106' includes source-side and drain-side trench sidewalls 308' and
310' (i.e., the edges of source-access and drain-access regions 108
and 110, respectively) that are not vertical. Instead, trench
sidewalls 308' and 310' are tapered away from vertical sidewalls
322 and 324 of gate G1, respectively, as the trench sidewalls rise
from trench bottom B to superlattice topside 312. Thus, a
separation between each gate sidewall 322, 324 and the respective
adjacent trench sidewall 308', 310' increases as the respective
adjacent trench sidewall rises. The tapered trench sidewalls
decreases gate-to-source and gate-to-drain capacitances relative to
an arrangement that does not include the tapered trench sidewalls,
e.g., the vertical trench sidewalls shown in the embodiment of
FIGS. 2 and 3. The alternative embodiment also includes a
source-side dielectric wedge 404 wedged between trench sidewall
308' and gate sidewall 322, and a drain-side dielectric wedge 406
wedged between trench sidewall 310' and gate sidewall 324. Wedges
404, and 406 follow the tapers of trench walls 308' and 310'.
[0027] Having described single e-mode FET 100 above, an integrated
circuit (IC) that includes the e-mode FET alongside a different
type of FET is now described. An IC typically includes multiple
FETs of different types that interact to perform specialized
circuit functions. The FETs may include one or more e-mode FETs and
one or more d-mode FETs to implement "E-D logic." Accordingly, an
embodiment presented herein is directed to an IC that includes an
e-mode FET and a d-mode FET constructed on a common structure that
includes a superlattice, as described below in connection with
FIGS. 5 and 6. The d-mode FET may be configured as a SLCFET. The
d-mode and e-mode FETs may be fabricated simultaneously using
substantially identical process flows. The d-mode and e-mode FETs
may be interconnected with metal wiring or they may share one or
more source and/or drain contact regions. By way of overview, the
SLCFET includes a superlattice (2DEG) on a buffer layer (e.g., an
un-doped or lightly doped GaN buffer layer). The superlattice has
many parallel castellation trenches formed in the superlattice,
which completely cut through the superlattice and into (e.g.,
50-100 nm into) the buffer layer, leaving ridges on either sides of
each trench. A gate covers the trenches and the ridges. The
castellation trenches and ridges are parallel to a direction of
current flow from a source to a drain of the SLCFET. When a
negative bias is applied to the gate, a side-pinching action
depletes carriers from the superlattice remaining in the ridges to
turn off current flow between the source and the drain. A
particular SLCFET structure is described below.
[0028] With reference to FIG. 5, there is an illustration of top
plan view of a portion of an example IC 500 that includes e-mode
FET 100 and a SLCFET 502 constructed on adjacent respective
portions P1 and P2 of a superlattice-based common structure 504 of
the IC, including substrate layer 101, buffer layer 102, and
superlattice 104. An advantage associated with IC 50 is that both
e-mode FET 100 and SLCFET 502 may be constructed on common
structure 504 in a single SLCFET process flow. FIG. 6 is a
perspective view of SLCFET 502 in an area 6-6 of the SLCFET from
FIG. 5. With reference to both FIGS. 5 and 6, SLCFET 502 includes
portions of substrate layer 101, buffer layer 102, and superlattice
104 adjacent to corresponding portions of those same layers of
e-mode FET 100. SLCFET 502 also includes a source-access region 506
of superlattice 104 (in portion P1 of the superlattice), a
drain-access region 508 of the superlattice (in portion P2)
spaced-apart from the source-access region by a distance d1 in the
x-direction (i.e., length direction), a source S2 embedded in and
in contact with the source-access region, and a drain D2 embedded
in and in contact with the drain-access region.
[0029] SLCFET 502 also includes alternating ridges 518 and trenches
516 formed in superlattice 104 and referred to as ridge and source
"castellations," and that extend parallel to each other from
source-access region 506 to drain-access region 508 over distance
d1 along the x-direction (i.e., along the length direction).
Alternating ridges 518 and trenches 516 undulate in the y-direction
across a width of SLCFET 502 that is transverse to the length, such
that the alternating ridges and trenches define a castellated
region of length d1 from source-access region 506 to drain-access
region 508 throughout the length of the SLCFET. Each of ridges 514
provides a respective layered stack of 2DEG channels that connects
source-access region 508 to drain-access region 510 (and thus
source S2 to drain D2). Thus, each ridge may be referred to as a
"multichannel" ridge. Also depicted in FIG. 5 is a boundary-line
that defines an isolation region rectangle 519 for SLCFET 502. The
region outside of isolation region rectangle 519 (but not inside of
the rectangle) is an isolation region that is implanted to be
isolating. An active width of SLCFET 502 is determined by the
isolation region.
[0030] SLCFET 502 includes a gate G2 that covers, and undulates
with, ridges 518 and trenches 516 along the width of the SLCFET.
Gate G2 has a length d2 in the x-direction that is less than the
length d1. Thus, gate G2 represents an undulating strip of
metallization of length d2 in the x-direction, positioned between
and spaced-apart from the source-access region 508 and the
drain-access region 510 in the length direction, which spans and
undulates along the width of the SLCFET in the y-direction. A
negative voltage applied to gate G2 selectively turns off (i.e.,
depletes) the 2DEG channels to correspondingly turn off the
SLCFET.
[0031] Methods of FET fabrication are now described. An example
fabrication of e-mode FET 100 is described below in connection with
FIGS. 7-10.
[0032] FIG. 7 is an illustration of an epitaxial structure formed
in initial stages of the fabrication. The epitaxial structure
includes superlattice 104 and a base structure 701. Base structure
701 may comprise a base layer 702, a nucleation layer 704 overlying
the base layer (where the base layer and the nucleation layer may
collectively form substrate layer 101), and a buffer layer 706
(e.g., buffer layer 102), overlying the nucleation layer, of, for
example, Silicon Carbide (SiC) or Si.sub.3N.sub.4, AlGaN, and
un-doped (or lightly doped) GaN, respectively. Other base
structures will be apparent to those skilled in the art.
[0033] The epitaxial structure also includes heterostructure layers
of superlattice 104 formed (e.g., deposited) on base structure 701
using successive applications of an epitaxial growth process. The
growth process may include Metal Organic Chemical Vapor Deposition
(MOCVD) or Molecular Beam Epitaxy (MBE), for example.
[0034] FIG. 8 is an illustration of the epitaxial structure of FIG.
7 after trench 106 has been formed in the epitaxial structure. The
afore-mentioned structures may be etched into the superlattice
using Reactive Ion Etching (ME), for example.
[0035] FIG. 9 is an illustration of the epitaxial structure of FIG.
8 after dielectric layer 116 has been formed on a top surface of
the epitaxial structure. Dielectric layer 116 may be deposited
using low-pressure chemical vapor deposition (LPCVD), or
plasma-enhanced chemical vapor deposition (PECVD), for example.
[0036] FIG. 10 is an illustration of the epitaxial structure of
FIG. 9 after gate G1 has been formed on the epitaxial structure.
Gate G1 may be formed, e.g., using electron beam lithography, metal
deposition by reactive direct current (DC) or radio frequency (RF)
sputtering, thermal evaporation, and liftoff. In the example shown
in FIG. 8, gate G1 forms gate post 314 that fills trench 106, and
includes gate top 316 connected to the gate post. Similar
techniques may be used to form source S1 and drain D1 on the
epitaxial structure.
[0037] FIG. 11 is a flowchart of an example method of
making/forming an e-mode FET (e.g., e-mode FET 100).
[0038] Operation 1102 includes forming a base structure including a
buffer layer that is un-doped or only lightly doped as described
above.
[0039] Operation 1104 includes forming on the buffer layer a
superlattice of conducting channels including a trench that cuts
down through the superlattice into the buffer layer and separates a
portion of the superlattice into a source-access region and a
drain-access region of the superlattice, wherein the buffer layer
forms a bottom of the trench. The superlattice of conducting
channels may be formed on the buffer layer using successive
applications of an epitaxial growth process. The growth process may
include MOCVD or MBE, for example. The trench may be etched into
the superlattice using ME, for example. Operation 1104 further
includes forming a dielectric layer on the superlattice of
conducting channels and the trench. The conformal dielectric layer
may be deposited using LPCVD or PECVD, for example.
[0040] Operation 1106 includes forming a source and a drain
embedded in the source-access region and the drain-access region,
respectively, so that the source and the drain are each
spaced-apart from the trench. The source and the drain may be
formed on the underlying dielectric and superlattice layers using
electron beam lithography, metal deposition by reactive direct
current (DC) or RF sputtering, thermal evaporation, and liftoff,
for example.
[0041] Operation 1108 includes forming a gate in the trench,
wherein (i) when a voltage above a threshold voltage of the FET is
applied to the gate, a current channel is induced in the buffer
layer, underneath the gate, which electrically connects the
source-access region to the drain-access region to turn on the
e-mode FET, and (ii) when a voltage below the threshold voltage is
applied to the gate, the current channel is eliminated, which
electrically disconnects the source-access region from the
drain-access region to turn off the e-mode FET. The gate may be
formed using the same techniques used to form the source and the
drain, for example.
[0042] A method of making a SLCFET (e.g., SLCFET 502) is similar to
the method of making the e-mode FET. For example, the method
includes forming a substrate including a buffer layer (which may be
the same buffer layer on which the e-mode FET is constructed), and
then forming a superlattice of conducting channels on the buffer
layer, such that the superlattice includes spaced-apart source and
drain access regions, and alternating (castellated) ridges and
trenches in parallel with each other extending between the source
and drain access regions. The method includes forming source and
drain metal on the source and drain access regions, and forming a
castellated strip of gate metal across the castellated ridges and
trenches. Moreover, the method of forming the e-mode FET and the
method of forming the SLCFET may be combined into a SLCFET process
flow for forming an integrated circuit that includes the e-mode FET
and the SLCFET on separated portions of the same buffer layer and
superlattice.
[0043] E-mode and d-mode combined logic (i.e., E-D logic) requires
a technology with both enhancement and depletion mode device types
available. The addition of even a simple superlattice-based e-mode
device (e.g., FET) along-side a superlattice-based d-mode device
(e.g., FET) is transformative because it enables simple E-D digital
logic for applications such as, but not limited to, control signal
routing. The saddle-gate e-mode FET with superlattice access
regions presented herein is an e-mode FET that is easily integrated
into a standard SLCFET process flow, enabling superlattice-based
E-D logic. For example, when integrating monolithic microwave ICs
(MMICs) into a package, the number of inputs/outputs needed can be
an important factor in determining the complexity and cost of the
final package. The number of direct current (DC) control signals
required for a SLCFET in the package can be reduced by 50% or more
if a simple on-chip digital inverter were available to generate
complementary gate bias signals. An IC that combines the e-mode FET
presented herein (e.g., to implement the on-chip digital inverter)
in combination with SLCFET(s) on a common substrate/superlattice
provides a cost effective and simplified solution for IC
integration.
[0044] In summary, in one aspect, an e-mode FET is provided
comprising: a buffer layer; a superlattice of conducting channels
on the buffer layer and including a trench that cuts down through
the superlattice into the buffer layer and separates the
superlattice into a source-access region and a drain-access region
of the superlattice, wherein the buffer layer forms a bottom of the
trench; a source and a drain adjacent to the source-access region
and the drain-access region, respectively, so that the source and
the drain are each spaced-apart from the trench; and a gate
disposed in the trench and configured such that (i) a voltage above
a threshold voltage of the e-mode FET applied to the gate induces
in the buffer layer underneath the gate a current channel, which
electrically connects the source-access region to the drain-access
region to turn on the e-mode FET, and (ii) a voltage below the
threshold voltage applied to the gate eliminates the current
channel, which electrically disconnects the source-access region
from the drain-access region to turn off the e-mode FET.
[0045] In an embodiment, the e-mode FET is constructed on a first
portion of the superlattice of conducting channels that is adjacent
to a second portion of the superlattice of conducting channels on
which a depletion-mode (d-mode) FET is constructed.
[0046] In another embodiment, the superlattice comprises
heterostructures that form the conducting channels.
[0047] In yet another embodiment, each heterostructure includes an
Aluminum (Al) Gallium (Ga) Nitride (N) layer and a GaN layer.
[0048] In another aspect, a method of making an e-mode FET is
provided comprising: forming a buffer layer; forming on the buffer
layer a superlattice of conducting channels including a trench that
cuts down through the superlattice into the buffer layer and
separates the superlattice into a source-access region and a
drain-access region of the superlattice, wherein the buffer layer
forms a bottom of the trench; forming a source and a drain adjacent
to the source-access region and the drain-access region,
respectively, so that the source and the drain are each
spaced-apart from the trench; and forming a gate in the trench, the
gate configured such that (i) a voltage above a threshold voltage
of the e-mode FET applied to the gate induces in the buffer layer
underneath the gate a current channel, which electrically connects
the source-access region to the drain-access region to turn on the
e-mode FET, and (ii) a voltage below the threshold voltage applied
to the gate eliminates the current channel, which electrically
disconnects the source-access region from the drain-access region
to turn off the e-mode FET.
[0049] In yet another aspect, an integrated circuit is provided
comprising: a buffer layer; a superlattice of conducting channels
on the buffer layer; an e-mode field effect transistor (FET)
constructed on a first portion of the superlattice; and a
superlattice castellated FET (SLCFET) constructed on a second
portion of the superlattice spaced-apart from the first portion of
the superlattice; wherein, the e-mode FET includes: in the first
portion of the superlattice, a trench that cuts down through the
first portion of the superlattice into the buffer layer and that
separates the first portion of the superlattice into a
source-access region and a drain-access region of the first portion
of the superlattice, wherein the buffer layer forms a bottom of the
trench; a source and a drain adjacent to the source-access region
and the drain-access region; and a gate disposed in the trench and
configured such that (i) a voltage above a threshold voltage of the
e-mode FET applied to the gate induces in the buffer layer
underneath the gate a current channel, which electrically connects
the source-access region to the drain-access region to turn on the
e-mode FET, and (ii) a voltage below the threshold voltage applied
to the gate eliminates the current channel, which electrically
disconnects the source-access region from the drain-access region
to turn off the e-mode FET.
[0050] The above description is intended by way of example only.
The description is not intended to be exhaustive nor is the
invention intended to be limited to the disclosed example
embodiment(s). Many modifications and variations will be apparent
to those of ordinary skill in the art without departing from the
scope and spirit of the invention.
* * * * *