Method for aligning optical devices with substrate

Details Method for aligning optical devices with substrate Method for aligning optical devices with substrate

1999-11-05

2002-02-05

Gulakowski, Randy (Department: 1746)

Etching a substrate: processes

Etching of semiconductor material to produce an article...

C216S024000, C216S039000, C216S057000

Reexamination Certificate

active

06344148

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates to a semiconductor technology; and, more particularly, to a method for aligning an optical device with an optical fiber and a planar lightwave circuit (PLC) by using a passive alignment method.
DESCRIPTION OF THE PRIOR ART
There is a need for an exact alignment between optical axes to get the maximum optical coupling between optical devices such as an optical fiber and a laser diode (LD) Conventionally, when the laser diode is aligned with an optical device, it is required that the tolerance of the optical alignment accuracy is within about 1 &mgr;m. In general, the alignment method is classified into an active alignment and a passive alignment. The active alignment is a method of achieving an optimum alignment by controlling the position of the optical waveguide and the LD based on the intensity observed with turning on the LD. This method can provide an optimum optical alignment easily, however, it is expensive to implement due to its complex process. Whereas the passive alignment is a method of carrying out the alignment of the optical axes with turning off the LD. In this method, it is difficult to obtain an optimum alignment of the optical axes, however, it is cost-effective due to its simple structure.
The passive alignment method is further classified into three methods as follows:
The first is a mechanical alignment method that optical devices are mounted on a silicon substrate after cutting grooves exactly to have the optical devices mounted on the silicon substrate and to fabricate the optical devices in an exact dimension. In this method, it is difficult that the optical devices and the silicon substrate are fabricated to have the tolerance of the optical alignment accuracy within about 1 &mgr;m.
The second is a flip-chip bonding method through the use of the surface tension of solder bumps that the solder bumps on the silicon substrate and solder pads on the optical device are exactly formed on their corresponding positions, respectively, by a conventional photolithography method. After the solder pads are aligned with the solder bumps, and then they are heated, the solder bumps are reflowed to be changed into the most stable shape and have the optical devices exactly positioned. However, in this method the reflow condition of the solder bumps is strict and it is required to prevent the oxidation of the solder.
Finally, the third is a mask alignment method that can be referred as the prototype of this invention, which alignment marks are formed on the silicon substrate and the optical devices and the optical devices are aligned with the substrate by using the alignment marks, this method will be described in detail hereinbelow.
FIG. 1
a
is an exemplary cross sectional view illustrating a method of connecting a laser diode to an optical fiber by using a conventional mark alignment method,
FIG. 1
b
represents a plan view of a laser diode and
FIG. 1
c
is a plan view of a substrate.
Referring to the drawings, a silicon wafer
10
is generally used as a substrate and a V-shaped groove
11
is used for clamping an optical fiber. The V-shaped groove
11
is obtained by wet-etching a silicon single crystal in KOH solution. The silicon single crystal has a different etching rate according to its crystalline plane, that is, since an etching ratio between a crystalline plane 111 and the other crystalline plane is in the order of 1/100, it is possible to obtain the V-shaped groove with a predetermined angle, e.g., 54.7°, after the etching is carried out for enough period of time, wherein the predetermined angle is an angle between the surface of the substrate and the surface of the V-shaped groove. Once the V-shaped groove
11
is formed, in case that it is still immersed in the KOH solution, the angle of the V-shaped groove
11
is hardly changed due to its negligible etching amount.
As shown in
FIG. 2
a,
a silicon wafer
20
of crystalline plane 100 is generally used, after being etched by KOH solution, to obtain an angle of 54.7° between the surface of the silicon wafer
20
and an exposed crystalline plane
100
. Typically, a process for forming the V-shaped grooves A, B is as follows: first, depositing on top of the silicon wafer
20
a material such as a silicon nitride (Si
3
N
4
) layer and a silica layer (SiO
2
) and the like which does not react with KOH solution; selectively removing portions of the material where the V-shaped grooves are to be formed, to thereby form a V-shaped etching window; and immersing the silicon wafer
20
into KOH solution, thereby forming the V-shaped grooves A, B under the removed portions of the material. At this time, the width and the depth of the V-shaped grooves A, B are determined by the width of the V-shaped etching window. After two windows are formed on the silicon wafer
20
, each window having a different width, and by etching the silicon wafer
20
in KOH solution, the V-shaped grooves A, B in
FIG. 2
a
can be obtained, simultaneously. In this case, the V-shaped groove B of the small window width is first formed and even though it is still immersed into KOH solution until the V-shaped groove A of the large window width is formed, the V-shaped groove B is not being etched further. Therefore, a width and a depth of a V-shaped groove can be obtained by controlling a window width. Thereafter, as shown in
FIG. 2
b,
an optical fiber
22
can be exactly aligned by positioning it on a V-shaped groove
23
.
Referring back to
FIGS. 1
a
to
1
c,
there are alignment marks
13
a,
13
b
for defining a position of an optical device such as a LD
12
. After cross-shaped marks
13
a,
13
b
are formed on the silicon substrate
10
and the LD
12
, respectively, these marks
13
a,
13
b
are aligned by using an optical apparatus such as a microscope to fix the LD
12
.
FIG. 3
is a view exemplifying a method for aligning an optical device
31
with a silicon substrate
30
by using an infrared ray, which has low attenuation in a substrate
30
and optical device
31
. The reference numerals
32
and
33
represent the alignment marks and the objective lens of the microscope, respectively.
Further, a plurality of solder bumps
14
and metal pads
7
can be provided on an upper surface of the silicon wafer
10
to attach the LD
12
to the silicon wafer
10
and to apply electrical signals to the LD
12
. As shown in
FIG. 1
a,
there is a under bump metallurgy (UBM)
15
for bonding the solder bump
14
to the silicon wafer
10
and a solder dam
16
made of a material such as silicon nitride which is non-adhesive to the solder bump
14
. The LD
12
is electrically connected to the metal line
17
through the use of the solder bump
14
. Also, a Au support
8
is formed to match a height of a core
18
of the LD
12
to that of a core
9
of the optical fiber
19
.
As described hereinabove, the passive alignment method for aligning an optical device and optical fiber and the passive alignment for aligning an optical device and a PLC is well known, however, the passive alignment of three devices, e.g., the laser diode, the PLC and the optical fiber, is very difficult.
On the other hand, the prior art method employs a metal layer to obtain an optical contrast for the alignment mark. As shown in
FIG. 4
, since a structure of silicon PLC
41
based on a silicon substrate
40
has a non-planarized surface due to a step caused by its structure, it is impossible to obtain a planarized surface by spin-coating a thin photoresist on top of the non-planarized surface. Therefore, it is required that a very thick photoresist
42
should be coated on top of the non-planarized surface to obtain a planarized surface of the photoresist
42
. Alignment marks
43
a
are formed by a lift-off method as follows: selectively removing portions of photoresist
42
where the alignment marks
43
a
are formed; depositing a metal layer
43
on top of the photoresist
42
; and removing the photoresist
42
.
Since, however, the very thick photoresist
42
is used as a mask in this method, there is a shortc

Affiliated with

Also associated with

Rating

Method for aligning optical devices with substrate does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Method for aligning optical devices with substrate, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for aligning optical devices with substrate will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2956787