Coherent light generators – Particular resonant cavity – Distributed feedback
Reexamination Certificate
2000-11-03
2003-02-18
Davie, James (Department: 2828)
Coherent light generators
Particular resonant cavity
Distributed feedback
Reexamination Certificate
active
06522680
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention pertains to vertical cavity surface emitting lasers (VCSEL's) and particularly to VCSEL's made by a metal-organic chemical vapor deposition (MOCVD) process.
The perspective view shown in
FIG. 1
illustrates a typical structure for a vertical cavity surface emitting laser
10
. A gallium arsenide substrate
12
is disposed on an n type electrical contact
14
. A first mirror stack
16
and a bottom graded index region or lower spacer
18
are progressively disposed, in layers, on the substrate
12
.
Region
20
may have one or many quantum wells or may be a bulk active gain region. An active region
20
, having one or more quantum wells, is formed and a top graded index region or upper spacer
22
is disposed over active region
20
. The spacers are to provide the appropriate critical distance between the mirrors to provide the proper-sized resonant cavity for a given wavelength and the distance is related to that wavelength or a multiple thereof. Active region
20
has a gain that compensates for the leaking out of photons. Photons bounce back and forth and, due to ),imperfect mirrors
16
and
24
, eventually leak out of the device. Greater photon loss means more gain is needed.
A p type top mirror stack
24
is formed over active region
20
and a metal layer
26
forms an electrical contact. Current
21
can be caused to flow from the upper contact
26
to the lower contact
14
. This current
21
passes through the active region
20
. Upward arrows in
FIG. 1
illustrate the passage of light
23
through an aperture or hole
30
in the upper metal contact
26
. Downward arrows illustrate the passage of current
21
downward from the upper contact
26
through p type GaAs cap layer
8
, p type conduction layer
9
, p type upper mirror stack
24
and active region
20
. A hydrogen ion bombardment or implantation
40
forms an annular region of electrically resistant material. In order to confine the current flow
21
through active region
20
, device
10
uses a hydrogen ion implant technique to create electrically insulative regions around an electrically conductive opening extending therethrough. A central opening
42
of electrically conductive material remains undamaged during the ion implantation process. As a result, current
21
passing from upper contact
26
to lower contact
14
is caused or forced to flow through electrically conductive opening
42
and is thereby selectively directed or confined to pass through a preselected portion of active region
20
.
The present problem concerns active region
20
of the device. The issue relates to the reliability implications that result from the interaction between carbon and hydrogen in the VCSEL structure. There have been vertical cavity surface emitting lasers that have had short term degradation caused by hydrogen passivation or compensation-of carbon. Hydrogen compensates carbon acceptors in AlGaAs. This phenomenon is a byproduct of the MOCVD growth process and also results from proton implantation. Carbon ions are used in doping. Carbon doping brings in a significant amount of hydrogen. The results of hydrogen passivation are rapid degradation of the devices sometimes followed by rapid improvement, which is the result of the hydrogen moving through the structure under bias. Longer baking during the fabrication process drives out more hydrogen.
There are several kinds of doped structures. If low doping ≦5×10
17
/cm
3
(5el7) (curve
13
in
FIG. 2
a
) is used in p-spacer
22
of
FIGS. 1 and 3
, then the structure is sensitive to mobile hydrogen. The sensitivity to mobile hydrogen occurs because hydrogen acts as a donor and compensates the carbon. However, the hydrogen is very mobile and under field-aided diffusion, hydrogen H drifts towards active region
20
and compensates carbon C acceptors on the edge of active region
20
in
FIG. 2
a
. One way to overcome this compensation is to use higher p doping near active region
20
(see curve
15
in
FIG. 2
a
). The problem that arises with such doping is that there remains a large slope
19
in active region
20
even at lasing voltages in
FIG. 2
c
. The separation of the carriers resulting from energy band slope versus position
19
makes the recombination inefficient. To overcome this problem, a thick (
15
nanometers) effectively undoped region is placed on lower side
18
of active region
20
. The voltage is then allowed to drop across the undoped region.
The post growth anneal and the use of low arsine over pressures during growth have been other solutions attempted to prevent hydrogen passivation or compensation of carbon that causes at least short term degradation of VCSEL's. The present invention is a structural solution to the problem.
Schneider et al, U.S. Pat. No. 5,557,627, discloses a visible VCSEL that allows the use of Carbon as a p-type dopant, which improves the dopant precision and stability against thermal diff-usion during epitaxial growth and subsequent device processing. To allow the use of Carbon, a barrier or spacer layer is fabricated on the p-side of the active region, having little or no indium.
In Kobayashi et al, U.S. Pat. No. 5,513,202, a VCSEL having a substrate, a p-type mirror, a p-type spacer layer, an active region, an n-type spacer layer, and an n-type mirror is disclosed. The problem addressed in this patent concerns the high vertical resistance in a non-graded p-type mirror. One embodiment disclosed involves a p-type spacer layer with an impurity concentration higher than that of the p-type bottom mirror. More specifically, Kobayashi discloses that the p-type spacer layer has an impurity concentration of 3×10
18
cm
−3
. Additionally, the n-type spacer layer has a concentration of 3×10
18
cm
−3
, but may also be an undoped layer.
SUMMARY OF THE INVENTION
A structural way to make the VCSEL structure
10
less sensitive to the hydrogen passivation problem is to use heavily doped layers near active region
20
(of FIGS.
1
and
3
). These layers would be too heavily doped for the hydrogen to completely compensate. If this doping is not carefully performed, device
10
will not work because energy band structure
17
of active region
20
will have a residual tilt
19
even at lasing voltages (as shown in FIG.
2
c). This represents an electric field across active region
20
. The electric field causes the carriers of the opposite charge to preferentially seek one side or the other of active region
20
and radiative recombination becomes inefficient. Since radiative recombination is inefficient, parasitic recombination mechanisms dominate.
To eliminate this residual tilt of bands
17
, an effectively undoped section
18
in the n graded region (or close to active region
20
on the n-side which may include an n spacer) must be present and this undoped region
18
must be of sufficient extent. The term “effectively undoped” means that there are residual impurities in any material and that there is in reality no such thing as strictly undoped material. For purposes in this description and the claims, “unintentionally doped,” “effectively undoped” and “undoped” mean the same thing and may be used here interchangeably. Additionally, p doping can be added to active region
20
. Two dissimilar materials (i.e., n and p doped), when placed together, have different work functions and charges that flow from one to another. So before a bias voltage for lasing is applied, there is a built-in voltage between the p and n regions. This voltage causes an electric field/energy band slope which is reduced by reducing the charge at the junction. This is accomplished by introducing an undoped (uncharged) region. As voltage is applied the band flattens further and the electric field is reduced.
In summary, the invention has two features. It provides for relatively high doping in the p regions
22
down to and optionally through active region
20
, and it has a thick undoped region in the lower graded region
18
. The high p-doping in spacer
22
makes
Abeyta Andrew A.
Davie James
Honeywell Inc.
Shudy John G.
LandOfFree
VCSEL structure insensitive to mobile hydrogen does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with VCSEL structure insensitive to mobile hydrogen, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and VCSEL structure insensitive to mobile hydrogen will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3178432