Dopant diffusion blocking for optoelectronic devices using...

Details Dopant diffusion blocking for optoelectronic devices using... Dopant diffusion blocking for optoelectronic devices using...

2000-03-31

2003-12-16

Lee, Eddie (Department: 2815)

Active solid-state devices (e.g., transistors, solid-state diode

Responsive to non-electrical signal

Electromagnetic or particle radiation

C257S436000, C257S461000

Reexamination Certificate

active

06664605

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates to a method for fabricating optoelectronic devices, such as lasers, modulators, optical amplifiers, and detectors, and in particular to a method and device for reducing the diffusion and/or interdiffusion of dopant atoms among differently doped regions of such optoelectronic devices.
BACKGROUND OF THE INVENTION
Blocking layers are increasingly important for optoelectronic devices. For example, in a buried heterostructure of a semiconductor laser diode, blocking layers confer superior characteristics, such as low oscillatory threshold value and stable oscillation transverse mode, as well as high quantum efficiency and high characteristic temperature. This is because, in the buried heterostructure laser diodes, a current blocking layer can be formed on both sides of an active laver formed between two clad layers having a large energy gap and a small refractive index. This way, current leakage during operation is substantially reduced, if not prevented.
A conventional method for the fabrication of semiconductor laser diodes having a semi-insulating buried ridge is exemplified in
FIGS. 1-7
and described below.
Referring to
FIG. 1
, the processing steps for fabricating a laser diode with a buried ridge begin with the formation of a multi layered structure
100
on an n-InP substrate
10
. The multi layered structure
100
is formed of a first n-InP cladding layer
12
, an active layer
14
, a second p-InP cladding layer
16
, and a layer
18
of a quaternary material (Q). Layers
12
,
14
,
16
and
18
are sequentially formed and successively epitaxially grown to complete a first crystal growth. The active layer
14
could be, for example, a multiple quantum well (MQW) structure formed of undoped InGaAs/InGaAsP pairs and formed by a Metal Organic Chemical Vapor Deposition (MOCVD) or Metal Organic Vapor Phase Epitaxy (MOVPE). Also, the second cladding layer
16
may be doped with a p-type dopant, the most common one being zinc (Zn).
Next, as shown in
FIG. 2
, a SiO
2
or Si
3
N
4
mask
20
is formed unto a stripe on the upper surface of layer
18
. Subsequently, the multi layered structure
100
is selectively etched down to the n-InP substrate
10
to produce a mesa stripe
50
, as illustrated in FIG.
3
. The mesa stripe
50
, which has the mask
20
on top, is then introduced into a growth system, such as a liquid phase epitaxial, a MOCVD, a molecular beam epitaxy (MBE), or vapor phase epitaxy (VPE) growth system, so that an InP current blocking layer
32
and an n-InP current blocking layer
34
are subsequently formed, as shown in FIG.
4
. The current blocking layers
32
and
34
surround the mesa stripe
50
and form a second crystal growth.
The first current blocking layer
32
may be doped with impurity ions, such as iron (Fe), ruthenium (Ru) or titanium (Ti), to form a semi-insulating (si) InP(Fe) blocking layer
32
. The addition of Fe-impurity ions increases the resistivity of the first current blocking layer
32
and reduces the leakage current that typically occurs at the interface between the substrate
10
and the first current blocking layer
32
. Similarly, the second current blocking layer
34
may be doped with impurity ions, such as silicon (Si), sulfur (S) or tin (Sn), to form an n-type InP-doped blocking layer
34
.
Referring now to
FIG. 5
, after removal of the mask
20
, a third crystal growth is performed on the upper surfaces of the second current blocking layer
34
and the Q layer
18
. Thus, a p-InP cladding layer
42
(also called a burying layer) and a p-InGaAsP or a p-InGaAs ohmic contact layer
44
are further grown to form a buried heterostructure. The cladding layer
42
may be also doped with p-type impurity ions, such as zinc (Zn), magnesium (Mg), or beryllium (Be), to form a p-type InP-doped cladding layer
42
. Since Zn is the most commonly used p-type dopant, the cladding layer
42
will be referred to as layer InP(Zn)-doped.
The method of fabricating the above structure poses three major drawbacks, all of them relating to the diffusion and interdiffusion of dopant atoms, particularly those of zinc, since zinc is the most common and widely used p-type dopant in the optoelectronic industry.
First, zinc diffusion occurs into the active region of the semi-insulating buried ridge.
FIG. 5
shows the diffusion of zinc in the direction of arrow A, from the doped p-InP(Zn) second cladding layer
16
into the active layer
14
, because of the direct contact between the two layers. The high diffusivity of zinc leads to an undesirable shift in the emitting wavelength, up to tenths of microns. The reshaping of the overall zinc distribution profile further impacts the electrical characteristics of the optoelectronic device. The excess of zinc in the active region
14
of the device structure also results in the degradation of various device characteristics, such as the extinction ratio and the junction capacitance of the electro-absorption modulator strictures.
Second, iron-zinc (Fe—Zn) interdiffusion occurs at the interface between the doped p-InP(Zn) second cladding layer
16
and the semi-insulating InP(Fe) first current blocking layer
32
.
FIG. 5
shows the diffusion of zinc in the direction of arrow B
1
, from the p-InP(Zn) second cladding layer
16
into the InP(Fe) first current blocking layer
32
. Similarly, arrow B
2
of
FIG. 5
illustrates the diffusion of iron from the InP(Fe) first current blocking layer
32
into the p-InP(Zn) second cladding layer
16
.
Third, iron-zinc (Fe—Zn) interdiffusion occurs in the blocking structures of the laser devices, more precisely at the interface between the semi-insulating InP (Fe) first current blocking layer
32
and the p-InP(Zn) cladding layer
42
. The problem arises because the Fe-doped InP current blocking layer
32
, which was initially covered by the mask
20
, comes into contact with the Zn-doped InP cladding layer
42
after the removal of the mask
20
. The contact regions are exemplified in
FIGS. 5
as regions D, situated on lateral sides of the mesa stripe
50
. The interdiffusion of Fe and Zn atoms at the regions D can significantly increase the leakage current and degrade the device, leading to a poor manufacturing yield. In addition, if the active layer
14
has a multiple quantum well (MQW) structure, the Zn impurities in the Zn-doped InP cladding layer
42
can enter the active layer
14
to form mixed crystals therein and practically reduce the quantum effect to zero.
In an effort to suppress the diffusion and interdiffusion of Zn dopant atoms, different techniques have been introduced in the IC fabrication. For example, one technique of the prior art, shown in
FIG. 6
, considered the incorporation of a zinc doping set-back into the device structure, such as an undoped InP layer
52
. The undoped InP layer
52
is grown after the growth of tie active layer
14
, but before the growth of the p-InP second cladding layer
16
, to prevent therefore the direct contact between zinc and the active region. In lieu of the undoped InP layer
52
, a silicon doped n-InP(Si) layer may be used also as a dopant set-back.
Although the above technique has good results in preventing tie Zn diffusion, its processing steps require extremely sensitive parameters, such as doping level and thickness, of the zinc-doped cladding and contact layers. Also, growth conditions, such as growth rate and temperature, must be very narrowly tailored so that the set-back is optimized for each device structure and for each reactor. Further, this method does not allow control over the shape of the final zinc distribution. Finally, when a silicon doped n-InP(Si) layer is alternatively used as a dopant set-back, the incorporated silicon, which is an n-type dopant, forms an additional and undesirable p-n junction on the p-side of the device.
Another technique of the prior art that tried to minimize the zinc-iron interdiffusion is exemplified in FIG.
7
. This technique contemplates the insertion of an intrinsic or undoped InP layer
70
between the Fe-doped In

Affiliated with

Also associated with

Rating

Dopant diffusion blocking for optoelectronic devices using... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Dopant diffusion blocking for optoelectronic devices using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dopant diffusion blocking for optoelectronic devices using... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3154007