United States Patent |
6,870,178
|
Asryan
,   et al.
|
March 22, 2005
|
Semiconductor laser with reduced temperature sensitivity
Abstract
A quantum dot (QD) laser having greatly reduced temperature sensitivity
employs resonant tunnel-injection of carriers into the QDs from a pair of
quantum wells (QWs). The carriers are injected through barrier layers.
Because the tunnel-injection process is essentially
temperature-independent, and because the tunnel-injection of carriers is
the dominant source of current through the device, temperature-dependent
currents are virtually eliminated, resulting in a device having a
temperature-independent threshold current. In an additional device,
carriers are injected into QDs from a pair of optical confinement layers
(OCLs), either by tunnelling or thermionic emission. Each barrier layer is
designed to have a low barrier height for carriers entering the QDs, and a
high barrier height for carriers exiting the QDs. As a result, parasitic
current from carriers leaving the QDs is greatly reduced, which enables
the device to have low temperature sensitivity even without using resonant
tunnel-injection and/or QWs.
Inventors:
|
Asryan; Levon V. (30 Emily Dr., Centereach, NY 11720);
Luryi; Serge (16 Holly La., Setauket, NY 11733)
|
Appl. No.:
|
468387 |
Filed:
|
February 17, 2004 |
PCT Filed:
|
February 28, 2002
|
PCT NO:
|
PCT/US02/06382
|
371 Date:
|
February 17, 2004
|
102(e) Date:
|
February 17, 2004
|
PCT PUB.NO.:
|
WO02/08260 |
PCT PUB. Date:
|
October 17, 2002 |
Current U.S. Class: |
257/14; 257/23; 257/431; 257/97; 372/43 |
Intern'l Class: |
H01L 029//06; H01S 005//00 |
Field of Search: |
257/9,14,21,23,25,431,97
372/43
|
References Cited [Referenced By]
U.S. Patent Documents
Other References
Yasuhiko Arakawa (1994) "Fabrication of quantum wires and dots by mocvd
selective growth", Solid-State Electronics, Elsevier Science Publishers:
pps. 523-528, Apr. 1994.
Asryan et al. (1997) "Charge Neutrality Violation in Quantum_Dot Lasers",
IEEE 3: pps. 148-157, Apr. 1997.
Huang et al. (2000) "Very Low Threshold Current Density Room Temperature
Continuous-Wave Lasing from a Single-Layer InAs Quantum-Dot Laser", IEEE
12: pps. 227-229, Mar. 2000.
Liu et al. (2000) "The Influence of Quantum-Well Composition on the
Performance of Quantum Dot Lasers Using InAs/InGaAs Dots-in-a-Well (DWELL)
Structures", IEEE 36: pps. 1272-1279, Nov. 2000.
|
Primary Examiner: Prenty; Mark V.
Attorney, Agent or Firm: Baker Botts LLP
Goverment Interests
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
This invention was partially made with U.S. Government support from the Air
Force Office of Scientific Research, MURI Grant No. F49620-00-1-0331.
Accordingly, the U.S. Government may have certain rights in this
invention.
Parent Case Text
CROSS-REFERENCE TO RELATED APPLICATION
This application claims priority to U.S. Provisional Patent Application
Ser. No. 60/272,202, entitled "Temperature-Insensitive Semiconductor
Laser," filed on Feb. 28, 2001, which is incorporated herein by reference
in its entirety. This application is a national phase application of
international application, PCT/US02/06382, filed Feb. 28, 2002.
Claims
What is claimed is:
1. A semiconductor structure, comprising:
a first barrier layer having a first barrier thickness;
a second barrier layer having a second barrier thickness;
a quantum dot layer including at least one quantum dot, the quantum dot
layer being disposed between the first and second barrier layers;
a first quantum well layer, the first barrier layer being disposed between
the first quantum well layer and the quantum dot layer, and the first
barrier thickness being sufficiently small to enable electrons to tunnel
from the first quantum well layer to the quantum dot layer; and
a second quantum well layer, the second barrier layer being disposed
between the second quantum well layer and the quantum dot layer, and the
second barrier thickness being sufficiently small to enable holes to
tunnel from the second quantum well layer to the quantum dot layer.
2. A semiconductor structure according to claim 1, further comprising:
a first optical confinement layer, the first quantum well layer being
disposed between the first optical confinement layer and the first barrier
layer; and
a second optical confinement layer, the second quantum well layer being
disposed between the second optical confinement layer and the second
barrier layer.
3. A semiconductor structure according to claim 2, further comprising:
an n-type cladding layer, the first optical confinement layer being
disposed between the n-type cladding layer and the first quantum well
layer; and
a p-type cladding layer, the second optical confinement layer being
disposed between the p-type cladding layer and the second quantum well
layer.
4. A semiconductor structure according to claim 3, wherein recombination of
electrons tunneled from the first quantum well layer with holes tunneled
from the second quantum well layer for generating photons takes place in
the at least one quantum dot.
5. A semiconductor structure according to claim 1, wherein recombination of
electrons tunneled from the first quantum well layer with holes tunneled
from the second quantum layer for generating photons takes place in the at
least one quantum dot.
6. A semiconductor structure according to claim 1, further comprising:
an n-type cladding layer, the first quantum well layer being disposed
between the n-type cladding layer and the first barrier layer; and
a p-type cladding layer, the second quantum well layer being disposed
between the p-type cladding layer and the second barrier layer.
7. A semiconductor structure according to claim 1, wherein the first
quantum well layer has an electron subband, the electron subband having a
lowest electron subband energy level, the at least one quantum dot having
a quantized electron energy level of at least one of a particular quantum
dot and an average-sized quantum dot, and the lowest electron subband
energy level being approximately equal to the quantized electron energy
level.
8. A semiconductor structure according to claim 7, wherein the second
quantum well layer has a hole subband, the hole subband having a lowest
hole subband energy level, the at least one quantum dot further having a
quantized hole energy level of at least one of a particular quantum dot
and an average-sized quantum dot, and the lowest hole subband energy level
being approximately equal to the quantized hole energy level.
9. A semiconductor structure according to claim 8, wherein the first
barrier layer has an electron energy barrier with respect to the lowest
electron subband energy level, the electron energy barrier having a height
with respect to the lowest electron subband energy level that is
sufficiently large to substantially prevent thermal emission of the
electrons from the first quantum well layer to the quantum dot layer, and
wherein the second barrier layer has a hole energy barrier with respect to
the lowest hole subband energy level, the hole energy barrier having a
height with respect to the lowest hole subband energy level that is
sufficiently large to substantially prevent thermal emission of the holes
from the second quantum well layer to the quantum dot layer.
10. A semiconductor structure according to claim 1, wherein the first
quantum well layer has an electron subband, the electron subband has a
lowest electron subband energy level, the first barrier layer having an
electron energy barrier with respect to the lowest electron subband energy
level, the electron energy barrier having a height with respect to the
lowest electron subband energy level that is sufficiently large to
substantially prevent thermal emission of the electrons from the first
quantum well layer to the quantum dot layer, and wherein the second
quantum well layer has a hole subband, the hole subband having a lowest
hole subband energy level, the second barrier layer having a hole energy
barrier with respect to the lowest hole subband energy level, and the hole
energy barrier having a height with respect to the lowest hole subband
energy level that is sufficiently large to substantially prevent thermal
emission of the holes from the second quantum well layer to the quantum
dot layer.
11. A semiconductor structure according to claim 1, wherein the second
quantum well layer has a hole subband, the hole subband having a lowest
hole subband energy level, the at least one quantum dot having a quantized
hole energy level of at least one of a particular quantum dot and an
average-sized quantum dot, and the lowest hole subband energy level being
approximately equal to the quantized hole energy level.
12. A semiconductor structure according to claim 1, wherein the at least
one quantum dot comprises at least a first quantum dot and a second
quantum dot, the first quantum dot having a first electron energy level
and a first hole energy level, the second quantum dot having a second
electron energy level and a second hole energy level, the first and second
quantum dots being separated by a dot separation distance that is
sufficiently large to prevent tunnel splitting of the first and second
electron energy levels and the first and second hole energy levels.
13. A semiconductor structure according to claim 1, wherein the at least
one quantum dot comprises at least a first quantum dot and a second
quantum dot, the quantum dot layer further including material separating
at least the first and second quantum dots, the first quantum dot having a
first quantum dot bandgap, the second quantum dot having a second quantum
dot bandgap, the material separating at least the first and second quantum
dots having a dot separation material bandgap, and the dot separation
material bandgap being sufficiently larger than the first and second
quantum dot bandgaps to substantially prevent tunneling of the electrons
and holes between the first and second quantum dots.
14. A semiconductor structure according to claim 1, wherein the at least
one quantum dot has a quantized electron energy level of at least one of a
particular quantum dot and an average-sized quantum dot, the at least one
quantum dot having a quantized hole energy level of at least one of a
particular quantum dot and an average-sized quantum dot, the first quantum
well layer having a hole subband, the hole subband having a lowest hole
subband energy level, the lowest hole subband energy level being greater
than the quantized hole energy level, the second quantum well layer having
an electron subband, the electron subband having a lowest electron subband
energy level, and the lowest electron subband energy level being greater
than the quantized electron energy level.
15. A semiconductor structure according to claim 1, wherein the at least
one quantum dot has a quantized electron energy level of at least one of a
particular quantum dot and an average-sized quantum dot, the second
quantum well layer having an electron subband, the electron subband having
a lowest electron subband energy level, and the lowest electron subband
energy level being greater than the quantized electron energy level.
16. A semiconductor structure according to claim 1, wherein the at least
one quantum dot has a quantized hole energy level of at least one of a
particular quantum dot and an average-sized quantum dot, the first quantum
well layer having a hole subband, the hole subband having a lowest hole
subband energy level, and the lowest hole subband energy level being
greater than the quantized hole energy level.
17. A semiconductor structure according to claim 1, wherein the at least
one quantum dot has a quantized electron energy level .epsilon..sub.n of
at least one of a particular quantum dot and an average-sized quantum dot,
and wherein the at least one quantum dot has a quantized hole energy level
.epsilon..sub.p of the at least one of a particular quantum dot and an
average-sized quantum dot, electrons in the first barrier layer having an
electron effective mass m.sub.c.sup.barrier, holes in the second barrier
layer having a hole effective mass m.sub.v.sup.barrier, a boundary between
the quantum dot layer and the first barrier layer having a conduction band
offset .DELTA.E.sub.c, a boundary between the quantum dot layer and the
second barrier layer having a valence band offset .DELTA.E.sub.v, and the
value of the expression m.sub.c.sup.barrier (.DELTA.E.sub.c
-.epsilon..sub.n) being approximately equal to the value of the expression
m.sub.v.sup.barrier (.DELTA.E.sub.v -.epsilon..sub.p).
18. A semiconductor structure, comprising:
a first barrier layer;
a second barrier layer;
a quantum dot layer including at least one quantum dot, the quantum dot
layer being disposed between the first and second barrier layers;
an n-side optical confinement layer, the first barrier layer being disposed
between the n-side optical confinement layer and the quantum dot layer,
the first barrier layer having an amount of transparency for electrons
being transported from the n-side optical confinement layer to the quantum
dot layer through the first barrier layer, and the first barrier layer
having an amount of transparency for holes being transported from the
quantum dot layer to the n-side optical confinement layer through the
first barrier layer; and
a p-side optical confinement layer, the second barrier layer being disposed
between the p-side optical confinement layer and the quantum dot layer,
the second barrier layer having an amount of transparency for holes being
transported from the p-side optical confinement layer to the quantum dot
layer through the second barrier layer, and the second barrier layer
having an amount of transparency for electrons being transported from the
quantum dot layer to the p-side optical confinement layer through the
second barrier layer, wherein the amount of transparency for electrons
being transported from the n-side optical confinement layer to the quantum
dot layer through the first barrier layer is greater than the amount of
transparency for electrons being transported from the quantum dot layer to
the p-side optical confinement layer through the second barrier layer, and
the amount of transparency for holes being transported from the p-side
optical confinement layer to the quantum dot layer through the second
barrier layer is greater than the amount of transparency for holes being
transported from the quantum dot layer to the n-side optical confinement
layer through the first barrier layer.
19. A semiconductor structure according to claim 18, further comprising:
an n-type cladding layer, the n-side optical confinement layer being
disposed between the n-type cladding layer and the first barrier layer;
and
a p-type cladding layer, the p-side optical confinement layer being
disposed between the p-type cladding layer and the second barrier layer.
20. A semiconductor structure according to claim 19, wherein recombination
of electrons being transported from the n-side optical confinement layer
to the quantum dot layer with holes being transported from the p-side
optical confinement layer to the quantum dot layer for generating photons
takes place in the at least one quantum dot.
21. A semiconductor structure according to claim 18, wherein recombination
of electrons being transported from the n-side optical confinement layer
to the quantum dot layer with holes being transported from the p-side
optical confinement layer to the quantum dot layer for generating photons
takes place in the at least one quantum dot.
22. A semiconductor structure according to claim 21, wherein the first
barrier layer has a first valence band hole energy level and a first
conduction band electron energy level, the second barrier layer has a
second valence band hole energy level and a second conduction band
electron energy level, the first valence band hole energy level being
greater than the second valence band hole energy level, and the second
conduction band electron energy level being greater than the first
conduction band electron energy level.
23. A semiconductor structure according to claim 18, wherein the first
barrier layer has a first valence band hole energy level and a first
conduction band electron energy level, the second barrier layer has a
second valence band hole energy level and a second conduction band
electron energy level, the first valence band hole energy level being
greater than the second valence band hole energy level, and the second
conduction band electron energy level being greater than the first
conduction band electron energy level.
24. A semiconductor structure according to claim 18, wherein the at least
one quantum dot comprises at least a first quantum dot and a second
quantum dot, the first quantum dot has a first electron energy level and a
first hole energy level, the second quantum dot has a second electron
energy level and a second hole energy level, the first and second quantum
dots being separated by a dot separation distance that is sufficiently
large to prevent tunnel splitting of the first and second electron energy
levels and the first and second hole energy levels.
25. A semiconductor structure according to claim 18, wherein the at least
one quantum dot comprises at least a first quantum dot and a second
quantum dot, the quantum dot layer further including material separating
at least the first and second quantum dots, the first quantum dot having a
first quantum dot bandgap, the second quantum dot having a second quantum
dot bandgap, the material separating at least the first and second quantum
dots having a dot separation material bandgap that is sufficiently larger
than the first and second quantum dot bandgaps to substantially suppress
tunneling of the electrons and holes between the first and second quantum
dots.
Description
BACKGROUND OF THE INVENTION
Semiconductor lasers are widely used in applications such as
telecommunications and optical data storage. However, the threshold
current of conventional laser depends upon the temperature at which the
laser is operated, and this temperature dependence causes variability in
the laser's optical output even if the driving current is constant.
Correction of the variability of the optical output can require
complicated and costly measures such as cooling systems and feedback
loops. It would therefore be preferable for the threshold current of a
laser to be temperature-independent. In some cases, quantum dot (QD)
lasers have demonstrated temperature insensitivity superior to that of
quantum well (QW) lasers. However, conventional QD lasers still exhibit
significant temperature sensitivity.
The threshold current in a semiconductor laser is the lowest injection
current at which lasing emission occurs. At the lasing threshold, the
optical gain of the active medium becomes equal to the total losses--the
total losses being equal to the sum of the mirror losses and the internal
losses. A major source of temperature dependence of threshold current in
QD lasers is parasitic recombination of carriers--i.e. electrons and
holes--outside the QDs. Such recombination occurs primarily in the optical
confinement layer (OCL) of the device. In most conventional QD lasers, the
OCL is a conductive material in which the QDs are embedded. For example,
FIG. illustrates a prior art structure 502 which includes n-type and
p-type cladding layers 504 and 512, and an OCL 514 comprising first and
second OCL portions 506 and 510. Self-organized QDs 508 are embedded
between the first and second OCL portions 506 and 510. The current flowing
through the device not only includes current I.sub.QD resulting from
carriers entering the QDs 508 and recombining to generate useful photons,
but also includes parasitic current resulting from recombination of
carriers in the OCL 514. The amount of this parasitic current depends on
the rate at which the carriers recombine in the OCL 514, which is
proportional to the populations of electrons and holes in the OCL 514. As
is well-known in the art, the ratio of the population of electrons in the
OCL 514 to the population of electrons in the QDs 508 increases with
temperature. Similarly, the ratio of the population of holes in the OCL
514 to the population of holes in the QDs 508 also increases with
temperature. As a result, the component of threshold current density
associated with recombination in the OCL 514 increases with temperature,
thereby causing the total threshold current also to increase with
temperature.
An additional source of temperature-sensitivity in QD lasers is
non-uniformity of the sizes of the QDs 508. In a typical QD laser, the QDs
tend to exhibit significant size variation. The QD size variation causes
undesired pumping of non-lasing QDs, an effect which further contributes
to the temperature-dependence of the threshold current of the device.
Yet another cause of temperature-sensitivity in QD lasers is recombination
from non-lasing (typically higher-energy) carrier states in the quantum
dots. If a QD has electron and hole states other than the states being
used for lasing, the extra states can be populated by thermally-excited
carriers, an effect which is temperature-dependent. The carriers in the
extra states can recombine to generate parasitic current. This thermally
activated parasitic current adds to the temperature-dependence of the
threshold current of the device.
Still another source of temperature-sensitivity in QD lasers is the
violation of charge neutrality in individual quantum dots. The optical
gain of a QD laser is A=K.sub.1 (F.sub.n +F.sub.p -1), where K.sub.1 is a
constant, F.sub.n is the probability of occupancy of the lasing electron
state in a QD, and F.sub.p is the probability of occupancy of the lasing
hole state in a QD. The current associated with carrier recombination in
the QDs is I.sub.QD =K.sub.2 F.sub.n F.sub.p, where K.sub.2 is a constant.
If a QD is charge-neutral--i.e., if the number of electrons equals the
number of holes--then F.sub.n =F.sub.p, and therefore, A=K.sub.1 (2F.sub.n
-1) and I.sub.QD =K.sub.2 F.sub.n.sup.2. The amount of gain A required to
reach the lasing threshold is independent of temperature, and therefore,
the value of F.sub.n required to reach the lasing threshold is also
independent of temperature. Because I.sub.QD is a function of F.sub.n, the
threshold value of I.sub.QD is similarly temperature-independent. However,
if the above condition of charge neutrality is violated in a QD--i.e., if
there are one or more extra electrons, or one or more extra holes--then
F.sub.n and F.sub.p not only tend to be unequal, but as is well-known in
the art, F.sub.n and F.sub.p typically vary differently from each other as
functions of temperature. See L. V. Asryan and R. A. Suris, "Charge
Neutrality Violation in Quantum-Dot Lasers," IEEE Journal of Selected
Topics in Quantum Electronics, Vol. 3, No. 2, April 1997. As a result,
although the threshold value of A=K.sub.1 (F.sub.n +F.sub.p -1) does not
depend on temperature, the threshold value of I.sub.QD =K.sub.2 F.sub.n
F.sub.p can--and in fact, typically is--temperature-dependent. Therefore,
if charge neutrality is violated, the total threshold current is typically
temperature-dependent. Violation of charge neutrality is the dominant
cause of temperature sensitivity at low temperatures.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a QD laser in
which OCL recombination is reduced or eliminated, thereby providing
dramatic reduction of the temperature dependence of threshold current.
It is a further object of the present invention to provide a QD laser in
which the temperature sensitivity caused by QD size variation is reduced
or eliminated.
It is yet another object of the present invention to provide a QD laser in
which the temperature sensitivity caused by recombination of excited
carriers in non-lasing states in the QDs is reduced or eliminated.
It is still a further object of the present invention to provide a QD laser
in which the temperature sensitivity caused by the violation of charge
neutrality in individual QDs is reduced or eliminated.
These and other objects are accomplished by the following aspects of the
present invention.
In accordance with one aspect of the present invention, a semiconductor
laser structure comprises first and second barrier layers, a QD layer
including at least one QD, and first and second QW layers. The QD layer is
between the first and second barrier layers; the first barrier layer is
between the QD layer and the first QW layer; and the second barrier layer
is between the QD layer and the second QW layer. The first barrier layer
is sufficiently thin to enable electrons to tunnel from the first QW layer
to the QD layer. The second barrier layer is sufficiently thin to enable
holes to tunnel from the second QW layer to the QD layer.
In accordance with an additional aspect of the present invention, a
semiconductor laser structure comprises first and second barrier layers, a
QD layer including at least one QD, an n-side OCL, and a p-side OCL. The
QD layer is between the first and second barrier layers; the first barrier
layer is between the n-side OCL and the QD layer; and the second barrier
layer is between the p-side OCL and the QD layer. The first barrier layer
has an amount of transparency for electrons traveling from the n-side OCL
to the QD layer, and an amount of transparency for holes traveling from
the QD layer to the n-side OCL. The second barrier layer has an amount of
transparency for holes traveling from the p-side OCL to the QD layer, and
an amount of transparency for electrons traveling from the QD layer to the
p-side OCL. The amount of transparency for electrons traveling from the
n-side OCL to the QD layer is greater than the amount of transparency for
electrons traveling from the QD layer to the p-side OCL. The amount of
transparency for holes traveling from the p-side OCL to the QD layer is
greater than the amount of transparency for holes traveling from the QD
layer to the n-side OCL.
BRIEF DESCRIPTIONS OF THE DRAWINGS
Further objects, features, and advantages of the present invention will
become apparent from the following detailed description taken in
conjunction with the accompanying figures showing illustrative embodiments
of the invention, in which:
FIG. 1 is a diagram illustrating an exemplary QD laser structure, and
associated valence and conduction bands, in accordance with the present
invention;
FIG. 2 is a diagram illustrating portions of the valence and conduction
bands illustrated in FIG. 1, as well as carrier energy distributions
within the bands;
FIG. 3A is a diagram illustrating an exemplary QD layer for use in the QD
laser structure illustrated in FIG. 1, as well as valence and conduction
bands associated with the illustrated QD layer;
FIG. 3B is a diagram illustrating an additional exemplary QD layer for use
in the QD laser structure illustrated in FIG. 1, as well as valence and
conduction bands associated with the illustrated QD layer;
FIG. 4 is a band diagram associated with a current path passing through the
QD separation material in the QD layer of the QD laser structure
illustrated in FIG. 1;
FIG. 5 is a diagram illustrating an exemplary prior art QD laser structure;
FIG. 6 is a diagram illustrating an exemplary QD laser structure in
accordance with the present invention;
FIG. 7 is a diagram illustrating valence and conduction bands associated
with an exemplary QD laser structure in accordance with the present
invention;
FIG. 8 is a diagram illustrating valence and conduction bands associated
with an additional exemplary QD laser structure in accordance with the
present invention; and
FIG. 9 is a diagram illustrating yet another exemplary QD laser structure,
and associated valence and conduction bands, in accordance with the
present invention.
Throughout the figures, unless otherwise stated, the same reference
numerals and characters are used to denote like features, elements,
components, or portions of the illustrated embodiments.
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 illustrates an exemplary QD laser structure 100 in accordance with
the present invention. The structure 100 includes a QD layer 110 between a
first barrier layer 108 and a second barrier layer 116. The QD layer
includes one or more QDs 112 separated by a separation material 114 having
a bandgap 310 (see FIGS. 3A and 3B, discussed below) which is higher than
the bandgap 176 of the QDs. The structure 100 also includes n-side and
p-side QW layers 106 and 118 which are separated from the QD layer 110 by
the first and second barrier layers 108 and 116, respectively. The
thickness 160 of the first barrier layer 108 is sufficiently small to
enable electrons 128 to tunnel from the n-side QW layer 106 to the QD
layer 110. Similarly, the thickness 162 of the second barrier layer 116 is
sufficiently small to enable holes 130 to tunnel from the p-side QW layer
118 to the QD layer 110. The structure 100 also includes n-side and p-side
OCLs 104 and 120, and n-side and p-side cladding layers 102 and 122. The
n-side QW layer 106 is between the n-side OCL 104 and the first barrier
layer 108. The p-side QW layer 118 is between the p-side OCL 120 and the
second barrier layer 116. The n-side OCL 104 is between the n-side
cladding layer 102 and the n-side QW layer 106. The p-side OCL 120 is
between the p-side cladding layer 122 and the p-side QW layer 118.
FIG. 1 also illustrates the conduction band 124 and valence band 126
associated with a current path passing through one of the QDs 112 in the
QD laser structure 100 described above. The band diagrams 124 and 126 are
illustrated in terms of electron potential energy; accordingly, the upward
direction in FIG. 1 corresponds to increasing electron energy, whereas the
downward direction in FIG. 1 corresponds to increasing hole energy. As is
illustrated in the drawing, the QDs 112 are formed from a material having
a bandgap 176 which is lower than the bandgaps 174 and 178 of the barrier
layers 108 and 116. The QW layers 106 and 118 are formed from materials
having bandgaps 172 and 198 which are lower than the bandgaps 174 and 178
of the barrier layers 108 and 116, and are also lower than the bandgaps
170 and 180 of the OCLs 104 and 120. The OCLs 104 and 120 have higher
indices of refraction than the respective cladding layers 102 and 122, in
order to provide confinement of light generated in the QD layer 110. The
bandgaps 168 and 182 of the cladding layers 102 and 122 are typically
greater than the bandgaps 170 and 180 of the OCLs 104 and 120, because in
most semiconductor systems used for diode lasers--e.g., InP-based systems
and GaAs-based systems--the higher-bandgap materials have lower refractive
indices than the lower-bandgap materials.
In bulk form, the material of the QDs 112 has a lowest electron energy
level 195 and a lowest hole energy level 192. However, as will be readily
understood by those skilled in the art, the quantum mechanical wave
function of a carrier within a QD restricts the possible energy states of
the carrier to a set of one or more discrete, quantized energy levels. For
example, an electron in the QD layer 110 can exist in one of a set of
quantized energy levels 132, 184, 186, etc. Similarly, a hole in the QD
layer 110 can exist in one of a set of quantized energy levels 134, 188,
190, etc. The number of electron and hole energy levels in the QD layer
110 depends upon the size(s) and geometry(ies) of the QDs 112, as well as
the band offsets at the heterojunction interfaces between the QDs 112 and
the respective barrier layers 108 and 116. The quantized electron and hole
energy levels 132, 184, 186, 134, 188, and 190 illustrated in FIG. 1
represent the energy levels occurring in a particular QD or an
average-sized QD in the QD layer 110. In thermal equilibrium, an electron
within the QD layer 110 has a highest probability of residing in the
lowest energy level 132 (a/k/a the electron "ground state") of the QD
layer 110, and a lower probability of residing in an excited (i.e.,
higher) state such as state 184, state 186, etc. Similarly, a hole within
the QD layer 110 is highly likely to reside in the hole ground state 134,
and less likely to reside in an excited state such as state 188, state
190, etc.
In bulk form, the materials of the QW layers 106 and 118 have lowest
electron and hole energy levels in their conduction and valence bands.
However, as will be readily understood by those skilled in the art,
because the QW layers 106 and 118 are very thin, the quantum mechanical
wave functions of carriers within the QW layers 106 and 118 have energies
which are restricted to values higher than the lowest energy states which
would exist in bulk samples of the materials. For example, the electron
subband 144 of the n-side QW layer 106 has a lowest electron energy 152
which is greater than the lowest electron energy 191 of the conduction
band of a bulk sample of the material from which the QW layer 106 is
formed. Similarly, the hole subband 150 of the n-side QW layer 106 has a
lowest hole energy 158 which is greater than the lowest hole energy 193 of
the valence band of a bulk sample of the material. Likewise, the electron
subband 148 of the p-side QW layer 118 has a lowest energy level 156 which
is greater than the lowest energy 194 of the conduction band of a bulk
sample of the material from which the QW layer 118 is formed, and the hole
subband 146 of the p-side QW layer 118 has a lowest hole energy level 154
which is greater than the lowest hole energy level 196 of the valence band
of such a bulk sample of the material.
An electron 140 in the electron subband 144 of the aside QW layer 106 can
tunnel through the first barrier layer 108 into a quantized electron
energy level 132 within the QD layer 110. Preferably, the lowest electron
energy level within the electron subband 144 is approximately equal to the
quantized electron energy level 132 into which the electron 140 is
tunneling, in order to enable resonant tunneling. Resonant tunneling not
only is more efficient than non-resonant tunneling, but has other
advantages which are discussed in detail below.
Similarly, a hole 142 can tunnel from the hole subband 146 of the p-side QW
layer 118 into a quantized hole energy level 134 within the QD layer 110.
Preferably, the lowest hole energy level 154 in the hole subband is
approximately equal to the quantized hole energy level 134 within the QD
layer 110, in order to enable resonant tunneling. Once electrons have
tunneled into the quantized electron energy level 132 of a QD 112 in the
QD layer 110, and holes have tunneled into the quantized hole energy level
134 of the QD 112, the electrons recombine with the holes--either
spontaneously or by stimulation from photons already in the laser--to
produce additional photons, as will be readily understood by those skilled
in the art. Each recombination 138 of an electron and a hole produces an
emitted photon 136.
Below and at the lasing threshold, the total injection current is the sum
of the radiative recombination current I.sub.QD in the QDs 112 and the
recombination current outside the QDs 112--i. e., the recombination
current in the QW layers 106 and 118 and the OCLs 104 and 120. As is
discussed above in the Background Section, the value of I.sub.QD required
to reach the threshold condition is independent of temperature, unless the
charge neutrality condition is violated in the QDs--a situation which is
discussed in further detail below. On the other hand, the recombination
current in the regions outside the QDs 112 is controlled by the minority
carrier density in these outside regions. If tunneling through the barrier
layers 108 and 116--discussed in further detail below--is the dominant
mechanism by which carriers travel through the structure, then any
minority carriers outside the QDs 112 are supplied primarily by the
tunneling process. Yet, the rate of tunneling of electrons and holes
through the barrier layers 108 and 116 does not depend on temperature. As
a result, the minority carrier density outside the QDs--and the
corresponding recombination current--are also independent of the
temperature. Therefore, the threshold current density of the laser of the
present invention is temperature-insensitive, provided that there are no
other potential sources of significant, temperature-dependent, parasitic
current.
In fact, in the laser structure 100 of the present invention, the only
sources of parasitic current are insignificant, temperature-insensitive,
or both. For example, consider the parasitic current caused by
recombination of electrons which have tunneled out of the QDs 112 and
through the second barrier layer 116 to become minority carriers on the
p-side of the structure. The recombination of these minority electrons
occurs primarily in the p-side QW layer 118 and OCL 120. Similarly,
consider holes tunneling out of the QDs 112 and through the first barrier
layer 108 to become minority carriers on the n side of the structure. The
recombination of these minority holes occurs primarily in the n-side QW
layer 106 and OCL 120. Such tunneling of carriers out of the QDs causes
parasitic current. However, the number of carriers tunneling out of the
QDs 112 is limited by the tunneling mechanism. Because the rate of
tunneling is temperature-independent, the parasitic current caused by
tunneling of carriers out of the QDs 112 is also temperature-independent.
Therefore, this parasitic current does not contribute to the
temperature-sensitivity of the threshold current of the laser.
On the other hand, although the tunneling of carriers out of the QDs 112
does not contribute to the temperature sensitivity of the device, it is
still desirable to reduce or eliminate this parasitic current in order to
reduce the threshold current and increase the efficiency of the device.
Accordingly, the lowest energy 156 of the electron subband 148 in the
p-side QW layer 118 is preferably designed to be greater than the lasing
electron energy 132 (i.e., the energy of the electrons which are
recombining with holes to generate the laser light). Similarly, in order
to reduce or eliminate current caused by holes tunneling out of the QDs
112 into the n-side QW layer 106, it is preferable for the lowest hole
energy 158 of the hole subband 150 in the n-side QW layer 106 to be
greater than the lasing hole energy 134 (i.e., the energy of the holes
which are recombining with electrons to generate the laser light).
However, it is well known that the effective mass of a hole in a given
semiconductor material tends to be greater than the effective mass of an
electron in that material. Accordingly, if the QDs 112 are formed from the
same material as the QW layers 106 and 118, then the n-side QW layer 106
typically should be thicker than the p-side QW layer 118 in order to align
the lowest edge 152 of the n-side electron subband 144 with the lasing
electron energy level 132 in the QD layer 112, while also aligning the
lowest edge 154 of the p-side hole subband 146 with the lasing hole energy
level 134 in the QD layer 110. FIG. 7 is a band diagram of an example of
such a structure. In the exemplary structure, the greater thickness of the
n-side QW layer 106 causes the lowest edge 158 of the n-side hole subband
150 to be at a lower energy, not a higher energy, than the lasing hole
energy level 134 in the QD layer 110. In such a structure, the
above-described technique of designing the lowest hole energy 158 in the
n-side QW layer 106 to be greater than the lasing hole energy 134 in the
QD layer 110 may not be practicable. Therefore, in a structure in which
the QDs 112 and the QW layers 106 and 118 are formed from the same
material, it is preferable to use an n-side barrier layer 108 which is
substantially thicker than the p-side barrier layer 116, in order to
reduce the rate of tunneling of holes out of the QDs 112 into the n-side
QW layer of 106. A thicker n-side barrier layer 108 can dramatically
reduce the rate of holes tunneling out of the QD layer 110, while having a
lesser effect on the rate of tunneling of electrons from the n-side QW
layer 106 into the QD layer 110, because the smaller mass of the electrons
makes them less susceptible to the effects of a thick barrier layer. No
such added barrier thickness is required on the p side of the structure to
prevent electrons from tunneling out of the QDs 112 into the p-side QW
layer 118, because it is typically possible to design the p-side QW layer
118 such that the lowest electron energy 156 in the p-side electron
subband 148 is greater than the lasing electron energy 132 in the QD layer
110, while still keeping the lowest hole energy 154 in the p-side hole
subband 146 aligned with the lasing hole energy 134 in the QD layer 110.
Those skilled in the art will recognize that in a typical tunneling
semiconductor structure, the tunneling component of current coexists with
non-tunneling, thermal (e.g., thermionic) components of the current. In
addition, thermally assisted tunneling contributes to the current flowing
through the structure. Thermally assisted tunneling is a combination of
thermal excitation and tunneling, in which a carrier is raised thermally
(i.e., by phonons) to a higher energy level which is still below the top
of the barrier. The thermally excited carrier tunnels through the barrier
more easily than a non-excited carrier would tunnel.
The relative magnitudes of the respective current components depend upon
the barrier height, the barrier thickness, and the temperature. For
example, consider the parasitic current caused by carriers which--because
of the thermal distribution of carrier energies--have sufficient energy to
pass through the respective barrier layers 108 and 116 without tunneling.
As is illustrated in FIG. 2, the first barrier layer 108 has a particular
energy height 164 with respect to the lowest electron energy 152 of the
electron subband 144 of the n-side QW layer 106. The distribution of the
energies of electrons in the QW layer 106 is temperature-dependent, as
will be readily understood by those skilled in the art. At extremely low
temperatures, the energy distribution of the electrons is approximately a
step function, such that the energy levels below a nearly horizontal line
202 (with energy being represented by the vertical axis of the graph) are
fully occupied, whereas the energy levels above the line 202 are empty. At
moderate temperatures, the distribution is a curve 204 which tails off at
higher energy levels. At higher temperatures, the distribution is a curve
206 which tails off at energy levels higher than those at which the
moderate-temperature curve 204 tails off. If the barrier height 164 is
sufficiently small, and the temperature is sufficiently high, a
significant number of the electrons will have energies, relative to the
lowest electron energy 152, which exceed the barrier height 164. Such
electrons can pass through the barrier layer 108 without tunneling. The
resulting thermally generated current is strongly temperature-dependent
due to the temperature dependence of the electron distribution, and it is
therefore desirable to suppress the thermally generated current so that it
is much smaller than the tunneling current. Accordingly, the material used
to form the barrier layer 108 is preferably selected to ensure that the
barrier height 164 is greater than the energies of all but an
insignificant number of electrons. Provided that the barrier height 164 is
sufficiently large, the thermal emission of electrons will be
substantially suppressed by the barrier. In other words, the parasitic
current caused by thermal emission of electrons--i.e., by energetic
electrons passing through the barrier layer 108 without tunneling--should
be negligible compared to the current caused by electrons tunneling
through the barrier layer 108.
Similarly, the energy distribution of the holes in the p-side QW layer 118
has a strong temperature dependence. At low temperatures, the energy
distribution of the holes is essentially a step function--i.e., a
horizontal line 208. At moderate temperatures, the distribution is a curve
210 which tails off at higher hole energies. At higher temperatures, the
distribution is a curve 212 which tails off at higher energy levels. If
the temperature is sufficiently high, and the barrier height 166 with
respect to the lowest hole energy level 154 of the hole subband 146 is
sufficiently small, then a significant number of the holes will have
energies exceeding the height 166 of the barrier. These holes can travel
through the barrier layer 116 without tunneling. In order to suppress such
thermally generated current--which, as discussed above, has a strong
temperature dependence--it is preferable for the barrier height 166 to be
greater than the energies of all but an insignificant number of holes.
Provided that the barrier height 166 is sufficiently large, the thermal
emission of holes will be substantially suppressed by the barrier. In
other words, the parasitic current caused by thermal emission of
holes--i.e., by energetic holes passing through the barrier layer 116
without tunneling--should be negligible compared to the current caused by
holes tunneling through the barrier layer 116.
Consider now the probability of a carrier such as an electron tunneling in
a single step through the QD layer 110 and both barrier layers 108 and
116. As is well-known in the art, the probability of a particle tunneling
through a barrier is a decreasing exponential function of the barrier
thickness, because of the exponential decay of the particle's wave
function across the barrier. In the case of an electron tunneling from the
n-side QW layer 106, through the first barrier layer 108, through one of
the QDs 112 in the QD layer 110, and through the second barrier layer 116
to the p-side QW layer 118, the wave function of the electron does not
significantly decay across the QD 112. Therefore, for this electron, the
effective combined thickness of the QD layer 110 and the barrier layers
108 and 116 is approximately equal to the combined thicknesses of the
barrier layers 108 and 116. Assuming that the barrier layers 108 and 116
are comparable and adequate in thickness, and have sufficient barrier
height for electrons, the rate of electron tunneling from the n-side QW
layer 106 directly to the p-side QW layer 118 is negligible compared to
the rate of tunneling through the first barrier layer 108 alone, because
the effective tunneling thickness for electrons tunneling directly from
the aside QW layer 106 to the p-side QW layer 118 is approximately double
the tunneling thickness for electrons tunneling from the n-side QW layer
106 to the QDs 112. Similarly, if the barrier layers 108 and 116 have
sufficient barrier height for holes, the rate of tunneling of holes
directly from the p-side QW layer 118 to the n-side QW layer 106 is
negligible compared to the rate of tunneling of holes through the second
barrier layer 116 alone, because the effective tunneling thickness for the
holes tunneling directly from the p-side QW layer 118 to the n-side QW
layer 106 is approximately double the tunneling thickness for holes
tunneling from the second QW layer 118 to the QDs 112.
In addition to the current paths passing through the QDs 112, there is a
possible current path passing through the material 114 separating the QDs
112. However, for a carrier to travel along this current path, it would be
necessary for the carrier to tunnel directly between the two QW layers 106
and 118. This tunneling process can be understood by reference to FIG. 4,
which illustrates the conduction band 124 and valence band 126 of the
parasitic path of current flowing through the material 114 separating the
QDs 112 in a QD layer 110. Consider, for example, the probability of an
electron 402 tunneling from the n-side QW layer 106 to the p-side QW layer
118 through the separation material 114. Similarly to the material used to
form the barrier layers 108 and 116, the separation material 114 in the QD
layer 110 has a large band gap. Consequently, the electron wave function
decays significantly all the way from the n-side QW layer 106, through
both barrier layers 108 and 116 and the QD layer 110, to the p-side QW
layer 118. For such an electron, the tunneling distance 406 is the
combined thicknesses of the QD layer 110 and both barrier layers 108 and
116. As discussed above with respect to single-step tunneling of carriers
through the QDs 112 and both barrier layers 108 and 116, the electron wave
function is an exponentially-decaying function of tunneling distance.
Therefore, because the total, combined thickness 406 of the QD layer 110
and the barrier layers 108 and 116 is large compared to the thickness 160
of the first barrier 108 alone, the rate at which electrons tunnel
directly through all three layers 108, 110, and 116 is very
low--insignificant, in fact--compared to the rate at which electrons
tunnel through the first barrier layer 108 into the QDs 112. Similarly,
consider the likelihood of a hole 404 tunneling from the p-side QW layer
118 directly to the n-side QW layer 106. The rate at which such holes
tunnel directly from the p-side QW layer 118 to the n-side QW layer 106 is
much lower than the rate at which holes tunnel from the p-side QW layer
118 to the QDs 112, because the combined thickness 406 of all three layers
108, 110, and 116 is much greater than the thickness of the second barrier
layer 116 alone. Therefore, in the device of the present invention, the
parasitic current caused by direct tunneling of carriers between the QW
layers 108 and 116 is negligible.
As discussed above, it is possible for a QD to have more than one electron
energy state and/or more than one hole energy state. For example, the QDs
112 illustrated in FIG. 1 have energy levels 132, 184, 186, 134, 188, and
190. In conventional QD lasers, at typical operating temperatures the
ratio of the carrier population in a higher energy level to the carrier
population in a lower energy level tends to increase as an exponential
function of temperature. If one of the energy levels is intended to be
used for lasing, and the other level is not, then the presence of
electrons in the non-lasing level results in the parasitic effect of
carrier recombination producing spontaneous light emission which does not
contribute to the laser light. Not only does such parasitic recombination
cause excess current through the device, but the value of this current is
temperature-dependent. Accordingly, in a conventional structure, because
the excess current varies with temperature, the parasitic recombination
increases the temperature dependence of the threshold current.
In contrast, in the structure 100 illustrated in FIG. 1, lasing occurs by
resonant tunneling from the QW layers 106 and 118 into the QDs 112. By
choosing an appropriate thickness for the n-side QW layer 106, the lowest
edge 152 of the electron subband 144 in the n-side QW layer 106 can be
tuned so that it is equal to, or slightly less than, the QD electron
energy level selected for lasing (e.g., the lowest energy level 132 in the
QD layer 110). When the laser is operated, electrons resonantly tunnel
from the QW electron subband 144 to the selected QD energy level 132. Once
such resonant tunneling occurs, the effective resistance between the QW
layer 106 and the QD layer 110 becomes extremely small, thereby
suppressing further changes in voltage between the QW layer 106 and the QD
layer 110. Because further voltage changes between the QW layer 106 and
the QD layer 110 are suppressed, the lowest edge 152 of the electron
subband 144 locks onto the selected electron energy level 132 in the QD
layer 110. Even if the lowest edge 152 of the electron subband 144 is
slightly lower than the selected energy level 132 in the QD layer 110 when
no bias voltage is applied to the device, the application of a forward
bias voltage raises the lowest edge 152 of the electron subband 142 until
this edge 152 matches and locks onto the selected lasing energy level 132
in the QD layer 110, thereby giving rise to resonant tunneling. Similarly,
the lowest edge 154 of the hole subband 146 in the p-side QW layer 118 is
preferably tuned so that it is equal to, or slightly less than, the
selected hole lasing energy level (e.g., energy level 134). When a forward
bias voltage is applied to the device, the lowest edge 154 of the hole
subband 146 locks onto the selected hole energy level 134 in the QD layer
110, resulting in resonant tunneling. Once such resonant tunneling occurs
between the p-side QW layer 118 and the QDs 112, the effective resistance
between the p-side QW layer 118 and the QD layer 110 becomes very small,
thereby suppressing further changes in the voltage between the QW layer
118 and the QD layer 110. As a result, in a device which operates by
resonant tunneling, essentially all carriers are injected into the
selected electron and hole energy levels (e.g., electron energy level 132
and hole energy level 134). Furthermore, the exchange of carriers among QD
electron levels (e.g., levels 132, 184, 186, etc.) and among hole levels
(e.g., levels 134, 188, 190, etc.) is relatively slow, provided that the
energy separations between energy levels do not equal the energy of an
optical phonon, as is discussed below. Therefore, because of the resonant
tunneling injection of electrons into level 132, the other levels (i.e.,
levels 184, 186, etc.) will remain substantially empty, rather than in
thermal equilibrium with the lasing energy 132. Similarly, the non-lasing
hole levels 188, 190, etc. will be substantially empty, rather than in
thermal equilibrium with the hole level 134 being pumped by the resonant
tunneling current. The effect of temperature-dependent carrier populations
in the respective QD energy levels is thus mitigated by the structure of
the present invention.
In order to take fill advantage of the above-described benefit of the
resonant tunneling structure of the present invention, the energy
separations among the QD electron levels (e.g., the separation between
levels 132 and 184 or the separation between levels 132 and 186) and the
energy separations among the QD hole energy levels (e.g., the separation
between levels 134 and 188 or the separation between levels 134 and 190)
should not be equal to the energy of an optical phonon or any other phonon
mode having a high density of states and strong interaction with the
carriers.
An additional advantage of the laser structure of the present invention is
that the resonant tunneling mechanism suppresses violation of charge
neutrality in the QDs, which, as discussed above, is one of the mechanisms
contributing to temperature dependence of threshold current in
conventional QD lasers. In the resonant tunneling structure of the present
invention, any charge imbalance in a QD shifts all energy levels of the QD
relative to the energy levels of the injecting QW layers 106 and 118. For
example, the QD electron level 132 is shifted relative to the electron
subband 144 in QW layer 106, and the QD hole level 134 is shifted relative
to the hole subband 146 in QW layer 118. Due to the small electrical
capacitance of a single QD, the energy shifts caused by an imbalance even
as small as the charge of a single electron or a single hole are large
enough to substantially suppress resonant tunneling in that QD. Because
any QD having non-neutral charge is thus effectively deactivated, current
tends to flow only through QDs into which an equal number of electrons and
holes have tunneled--i.e., QDs in which charge neutrality has not been
violated. Thus, violation of change neutrality, and its effect on the
threshold current of the device, is greatly reduced in the resonant
tunneling structure 100 of the present invention.
The resonant tunneling mechanism of the laser structure of the present
invention has the further advantage of suppressing the effects of
inhomogeneity in the characteristics of the QDs 112. In particular, due to
resonant tunneling, the lowest edge 152 of the electron subband 144 in the
n-side QW layer 106 tends to lock on to the lasing electron energy level
132 of an average-sized QD. Similarly, the lowest edge 154 of the hole
subband 146 in the p-side QW layer 118 tends to lock onto the lasing hole
energy level 134 of the average-sized QD. As a result, QDs having sizes
which deviate significantly from the average size become inactive.
Therefore, the structure 100 of the present invention reduces or
eliminates the additional temperature-dependence of threshold current
associated with QD size variation.
If the QDs 112 in the QD layer 110 are sufficiently close together,
tunneling can occur among the QDs 112. For example, FIG. 3A illustrates a
pair of QDs 352 and 354 in a QD layer 110. The QDs 352 and 354 are
separated by a dot separation distance 302. A separation material 114
fills the volume between the QDs 352 and 354. The separation material 114
has a bandgap 310 which is larger than the bandgap 176 of the QDs 352 and
354. Depending upon the dot separation distance 302 and the difference
between the dot separation material bandgap 310 and the QD bandgap 176, it
may be possible for electrons 348 and/or holes 350 to tunnel between the
QDs. For example, if a first QD 352 has an electron energy level 312, and
a second QD 354 has its own electron energy level 314--which is typically
approximately equal to the electron energy level 312 of the first QD
352--it is possible for an electron 348 to tunnel from the first QD 352 to
the second QD 354, or vice versa. Similarly, if the first QD 352 has a
hole energy level 318, and the second QD 354 has an approximately equal
hole energy level 320, then it is possible for a hole 350 to tunnel from
the first QD 352 to the second QD 354, or vice versa. However, if the
separation distance 302 between the respective QDs is sufficiently large,
and/or the dot separation material bandgap 310 is sufficiently larger than
the dot bandgap 176, then such inter-dot tunneling is substantially
suppressed. In other words, the tunneling of carriers among the dots will
be suppressed to a level at which the tunneling will have a negligible
effect upon the energy states 312, 314, 318, and 320 of the QDs 352 and
354.
On the other hand, as illustrated in FIG. 3B, if the separation distance
302 is sufficiently small, inter-dot tunneling causes the carrier states
in the QDs 352 and 354 to interact with each other. The resulting,
coupled, two-dot system has energy levels which are different from those
of the uncoupled QDs 352 and 354. For example, in the illustrated band
diagram, electron states 312 and 314 become electron states 324 and 326,
and hole states 318 and 320 become hole states 336 and 338. This
phenomenon is often referred to as "tunnel splitting" of energy levels.
Increasing the number of coupled QDs increases the number of different
energies at which carriers exist.
For most applications, it is preferred for each QD to have only a single
predominant electron state and a single predominant hole state, in order
to ensure that only one lasing energy gap is active in the device. In
fact, if many of the QDs are very strongly coupled to each other, the
number of discrete energy levels in the QD layer can be so numerous that
the levels form a nearly continuous energy band. If so, the QD layer will
behave more like a QW layer than a true QD layer. Therefore, in most
cases, it is preferable that the separation distance 302 among the QDs 112
is sufficiently large to prevent significant tunneling/coupling among the
QDs 112. A suitable separation distance 302 is typically between 300 .ANG.
and 700 .ANG..
It is to be noted that even if the average separation between QDs is
sufficiently large that tunnel splitting is negligible, it may still be
possible for a small number of QD pairs to be close enough to cause
significant tunnel splitting in these pairs. The effect of such tunnel
splitting in a small number of QDs is similar to the effect of QD size
variation discussed above. However, in the resonant-tunneling structure of
the present invention, the tunnel-split levels of a pair of closely spaced
QDs will not be in resonance with the carriers being injected into the
QDs, and the QD pair will therefore not be effectively pumped by the bias
current. Accordingly, temperature-sensitivity associated with coupled QDs
is reduced or eliminated in the resonant tunneling structure 100 of the
present invention.
The QDs 112 illustrated in FIG. 1 are prismatic in shape. However, QDs in
accordance with the present invention can also have a pyramidal geometry,
as illustrated in FIG. 6. In particular, as is well-known in the art,
typical methods of forming self-organized QDs result in pyramidal QDs with
square bases. If such pyramidal QDs 112 are used in the structure 100 of
the present invention, then the distance 602 between the tips 604 of the
QDs 112 and the barrier layer adjacent to the tips 604 (e.g., the p-side
barrier layer 116) should be small enough to allow carriers (e.g., holes)
to easily tunnel into the QDs 112 from the QW layer (e.g., layer 118) on
the other side of the barrier layer (e.g., layer 116).
In accordance with a preferred embodiment of the laser structure of the
present invention, the QD material, the respective QW materials, and the
respective barrier materials can be selected such that an electron or hole
is subjected to a low barrier for entering the QDs 112 and a high barrier
for exiting the QDs 112. FIG. 8 is a band diagram of such a structure. In
this structure, the n-side barrier layer 108 and the p-side barrier layer
116 are formed from different materials. Exemplary materials which can be
used to form such a structure are listed in Table I, which is discussed in
further detail below. In the illustrated example, the n-side barrier layer
108 has relatively low conduction band offsets, but relatively high
valence band offsets, with respect to the QDs 112 and the n-side QW layer
106. As a result, electrons entering the QDs 112 are subjected to a low
conduction band barrier energy 802, whereas holes leaving the QDs 112 are
subjected to a high valence band barrier energy 806. In contrast, the
p-side barrier layer 116 has relatively high conduction band offsets, but
relatively low valence band offsets, with respect to the QDs 112 and the
p-side QW layer 118. As a result, electrons leaving the QDs 112 are
subjected to a high conduction band barrier energy 804, whereas holes
entering the QDs are subjected to a low valence band barrier energy 808.
It can therefore be seen that it is easy for electrons and holes to enter
the QDs 112 via tunneling (inducing thermally assisted tunneling) and/or
thermionic emission, because the conduction band energy 802 of the n-side
barrier layer 108 is quite low and the valence band energy 808 of the
p-side barrier layer 116 is also quite low; yet, it is difficult for
electrons and holes to exit the QDs 112 because the conduction band energy
804 of the p-side barrier layer 116 is quite high and the valence band
energy 806 of the n-side barrier layer 108 is also quite high.
Accordingly, because it is easy for carriers to enter the QDs 112, but
difficult for the carriers to exit the QDs 112, the parasitic current
caused by carriers exiting the QDs 112 is greatly reduced.
The materials and geometries of the QDs 112 and the barrier layers 108 and
116 should be selected so that both electrons and holes can readily tunnel
into the QDs 112. This is best accomplished by selecting materials and
geometries based on the following equation:
m.sub.c.sup.barrier (.DELTA.E.sub.c
-.epsilon..sub.n).apprxeq.m.sub.v.sup.barrier (.DELTA.E.sub.v
-.epsilon..sub.p), (1)
where: m.sub.c.sup.barrier is the electron effective mass in the n-side
barrier layer 108; m.sub.v.sup.barrier is the hole effective mass in the
p-side barrier layer 116; .DELTA.E.sub.c is the conduction band offset at
the heteroboundary between the QD layer 110 (in particular, the QDs 112)
and the n-side barrier layer 108; .DELTA.E.sub.v is the valence band
offset at the heteroboundary between the QD layer 110 (in particular, the
QDs 112) and the p-side barrier layer 116; .epsilon..sub.n is the
quantized electron energy level in a particular QD or an average-sized QD;
and .epsilon..sub.p is the quantized hole energy level in a particular QD
or an average-sized QD.
Table I lists a set of exemplary materials which can be used to fabricate a
QD laser structure 100 having the bands 124 and 126 illustrated in FIG. 8,
and also lists the corresponding bandgap E.sub.g of each of these
materials. Table II lists the resulting conduction band offset
.DELTA.E.sub.c and valence band offset .DELTA.E.sub.v for each of the
heteroboundaries resulting from the use of the materials listed in Table
I. The listed materials are preferably formed using molecular beam epitaxy
(MBE) and/or metal-organic chemical vapor deposition (MOCVD). MBE and
MOCVD are well-known processes for thin semiconductor film growth.
TABLE I
Layer Material E.sub.g (eV)
Cladding layers (102 and 122) InP 1.35
OCLs (104 and 120) Ga.sub.0.2 In.sub.0.8 As.sub.0.44 P.sub.0.56
1.01
and QD separation mat'l (114)
QW layers (106 and 118) Ga.sub.0.47 In.sub.0.53 As 0.717
Barrier.sub.2 (116) A1.sub.0.48 In.sub.0.52 As 1.4
Barrier.sub.1 (108) Ga.sub.0.14 In.sub.0.86 As.sub.0.3 P.sub.0.7
1.108
QDs (112) InAs 0.36
TABLE II
Boundary .DELTA.E.sub.c (eV) .DELTA.E.sub.v (eV)
n-Cladding/OCL.sub.1 (102/104) 0.133 0.207
OCL.sub.1 /QW.sub.1 (104/106) 0.114 0.179
QW.sub.1 /barrier.sub.1 (106/108) 0.152 0.239
Barrier.sub.2 /QW.sub.2 (116/118) 0.52 0.163
QW.sub.2 /OCL.sub.2 (118/120) 0.114 0.179
OCL.sub.2 /p-cladding (120/122) 0.133 0.207
If the materials listed in Table I are used, the thickness of the QD layer
112 is preferably in the range of 50-150 .ANG.. If the QDs are
self-organized, each QD is typically pyramidal with a square base. The
mean size of the QDs typically ranges from 50-150 .ANG., with
approximately a 10% statistical variation in the QD sizes. The surface
density of QDs in the QD layer 110 is typically 2.times.10.sup.10
-1.times.10.sup.11 cm.sup.-2, which corresponds to a separation distance
of 700-300 .ANG. between the centers of adjacent QDs. A QD separation
distance within this range is typically sufficient to prevent tunnel
splitting of the carrier states in the QDs.
The thicknesses of the QW layers 106 and 118 and the size(s) of the QDs 112
determine, respectively, the positions of the QW subband edges 152, 154,
156, and 158 and the quantized QD energy levels 132, 134, 184, 186, 188,
and 190. The appropriate QW thicknesses can be determined using
conventional numerical techniques which calculate the QD energy levels
based on QD shape and size. Estimates of the appropriate QW thicknesses
can be obtained by neglecting the penetration of the QW electron and hole
wave functions into the barrier regions. Based on such an approximation, a
suitable value of the n-side QW thickness is 165 .ANG., and a suitable
value of the p-side QW thickness is 95 .ANG.. Typically, the preferred
thicknesses of the QW layers 106 and 118 are in the range of 20-200 .ANG..
If the above-listed dimensions are used, then the gap between the lowest
electron energy level 152 and the lowest hole energy level 158 in the
n-side QW layer 106 is 0.754 eV, and the gap between the lowest electron
energy level 156 and the lowest hole energy level 154 in the p-side QW
layer 118 is approximately 0.829 eV. The height 164 of the barrier in the
first barrier layer 108 with respect to the lowest electron energy level
152 in the n-side QW layer 106 is 0.118 eV, and the height 166 of the
barrier in the second barrier layer 116 with respect to the lowest hole
energy level 154 in the p-side QW layer 118 is approximately 0.153 eV. The
gap between the electron energy level 132 and the hole energy level 134 in
the QD layer 110 is approximately 0.80 eV, and as a result, photons
generated by lasing in the device have an energy of 0.80 eV. This photon
energy corresponds to a photon wavelength .lambda. of 1.55 .mu.m, which is
currently the most useful wavelength for telecommunications applications.
As is well known to those skilled in the art, the optimal thickness of the
OCL is chosen so as to maximize the optical confinement factor. For a
laser emitting at .lambda.=1.55 .mu.m, the thickness of each of the OCLs
104 and 120 is typically 1000-1500 .ANG..
As is discussed above, the n-side barrier layer 108 is preferably thin
enough to allow significant tunneling of electrons from the n-side QW
layer 106 into the QDs 112, yet thick enough to significantly suppress the
tunneling of holes from the QDs 112 into the n-side QW layer 106.
Similarly, the p-side barrier layer 116 is preferably thin enough to allow
significant tunneling of holes from the p-side QW layer 118 into the QDs
112, yet thick enough to significantly suppress the tunneling of electrons
from the QDs 112 into the p-side QW layer 118. The tunneling rates R.sub.e
and R.sub.h of electrons and holes through a barrier of thickness d can be
estimated using the following equations:
##EQU1##
For the materials listed in Table I, an application of Eqs. (2) and (3)
indicates that the preferred thicknesses of the barrier layers 108 and 116
are approximately 10-20 .ANG..
In accordance with an additional aspect of the present invention, carriers
can originate on one side of a laser structure, and can be injected
non-resonantly into a QD layer, yet can be blocked from reaching the other
side of the structure and becoming minority carriers. A band diagram of
such a structure is provided in FIG. 9. Similarly to the structure 100
illustrated in FIG. 1, the structure 900 illustrated in FIG. 9 includes a
QD layer 110 having QDs 112 separated by a separation material 114;
barrier layers 108 and 116; and n-type and p-type cladding layers 102 and
122. However, rather than having both OCL layers and QW layers, the
structure 900 illustrated in FIG. 9 includes only OCL layers 902 and 904
which are preferably formed from low band gap materials, but need not be
thin enough to exhibit significant QW behavior. Similarly to the structure
discussed above with reference to FIG. 8, the structure 900 illustrated in
FIG. 9 includes a n-side barrier layer 108 having a low conduction band
electron energy 802 and a high valence band hole energy 806. The structure
900 also includes a p-side barrier layer 116 having a high conduction band
electron energy 804 and a low valence band hole energy 808. An electron
128 on the n-side of the structure can enter the n-side OCL layer 902, and
because the conduction band electron energy 802 of the aside barrier layer
108 is very low, it is easy for the electron 128 to pass through the
barrier layer 108 into the QD layer 110 by tunneling (including
thermally-assisted tunneling) or thermionic injection. The n-side barrier
layer 108 has a high "transparency" for electrons traveling from OCL layer
902 to the QD layer 110--i.e., the layer 108: (a) is thin enough and/or
has a sufficiently low conduction band electron energy 802 to enable a
significant number of the electrons to tunnel through it by thermally
assisted tunneling and/or non-thermally-assisted tunneling; and/or (b) has
a sufficiently low conduction band electron energy 802 to enable a
significant number of the electrons to be thermally emitted from OCL layer
902 to the QD layer 110. Similarly, a hole originating on the p-side of
the structure can enter the p-side OCL layer 904. The p-side barrier layer
116 has a high transparency for holes traveling from OCL layer 904 to the
QD layer 110--i.e., the layer 116: (a) is thin enough and/or has a
sufficiently low valence band hole energy 808 to enable a significant
number of holes to tunnel through it by thermally assisted tunneling
and/or non-thermally-assisted tunneling; and/or (b) has a sufficiently low
valence band hole energy 808 to enable significant thermal emission of
holes from OCL layer 904 into the QD layer 110.
Yet, the n-side barrier layer 108 should have a low transparency for holes
traveling from the QD layer 110 to OCL layer 902, and the p-side barrier
layer 116 should have a low transparency for electrons traveling from the
QD layer 110 to OCL layer 904. In other words, the n-side barrier layer
108 should be thick enough to substantially suppress tunneling of holes
out of the QD layer 110 into OCL layer 902, and/or should have a valence
band hole energy 806 which is sufficiently high to substantially suppress
thermal emission of holes from the QD layer 110 to OCL layer 902;
similarly, the p-side barrier layer 116 should be thick enough to
substantially suppress tunneling of electrons out of the QD layer 110 into
OCL layer 904 and/or should have a sufficiently high conduction band
electron energy 804 to substantially suppress thermal emission of
electrons from the QD layer 110 to OCL layer 904. There is, of course, a
trade-off between making the barrier layers 108 and 116 thick enough to
suppress the tunneling of carriers out of the QD layer 110, while being
thin enough to enable tunneling of carriers into the QD layer 110.
However, as discussed above, the tunneling mechanism is not temperature
dependent, and therefore, suppression of tunneling of carriers out of the
QD layer 110 is less important than the suppression of thermal emission of
carriers out of the QD layer 110. Accordingly, in order to maintain the
temperature-insensitivity of the laser structure, it is sufficient to
ensure that the n-side barrier layer 108 has a high valence band hole
energy 806, and that the p-side barrier layer 116 has a high conduction
band electron energy 804. Because the high conduction band electron energy
804 of the p-side barrier layer 116 prevents electrons from reaching the
p-side of the structure, and because the high valence band hole energy 806
of the n-side barrier layer 108 prevents holes from reaching the n-side of
the structure, recombination of carriers outside the QD layer 110 is
suppressed. As a result, the above-described structure has greatly reduced
temperature sensitivity. Preferably, the conduction band electron energy
804 of the p-side barrier layer 116 is greater than the conduction band
electron energy 802 of the n-side barrier layer 108, and the valence band
hole energy 806 of the n-side barrier layer 108 is greater than the
valence band hole energy 808 of the p-side barrier layer 116, in order to
ensure that carriers can easily enter the QD layer 110, but cannot easily
leave the QD layer 110.
It is further to be noted that the structure illustrated in FIG. 9 has the
additional advantage of not requiring any special alignment of QW and QD
energy levels.
Table III lists exemplary materials suitable for forming the structure
illustrated in FIG. 9, and also lists the corresponding bandgap E.sub.g of
each of these materials. Table IV lists the resulting conduction band
offset .DELTA.E.sub.c and valence band offset .DELTA.E.sub.v for each of
the heteroboundaries resulting from the use of the materials listed in
Table III.
TABLE III
Layer Material E.sub.g (eV)
Cladding layers (102 and 122) InP 1.35
QD separation mat'l (114) Ga.sub.0.2 In.sub.0.8 As.sub.0.44 P.sub.0.56
1.01
OCL layers (902 and 904) Ga.sub.0.47 In.sub.0.53 As 0.717
Barrier.sub.2 (116) Al.sub.0.48 In.sub.0.52 As 1.4
Barrier.sub.1 (108) Ga.sub.0.14 In.sub.0.86 As.sub.0.3 P.sub.0.7
1.108
QDs (112) InAs 0.36
TABLE IV
Boundary .DELTA.E.sub.c (eV) .DELTA.E.sub.v (eV)
n-Cladding/n-side 0.247 0.386
OCL layer (102/902)
n-side 0.152 0.239
OCL layer/barrier.sub.1 (902/108)
Barrier.sub.2 /p-side 0.52 0.163
OCL layer (116/904)
p-side OCL layer/ 0.247 0.386
p-cladding (904/122)
Although the present invention has been described in connection with
specific exemplary embodiments, it should be understood that various
changes, substitutions, and alterations apparent to those skilled in the
art can be made to the disclosed embodiments without departing from the
spirit and scope of the invention as set forth in the appended claims.
* * * * *