EE234 Photonics Laboratory, Lab 4

Laser-to-Fiber Coupling
Carsten Langrock

John XJ Zhang
October 31, 2001
Abstract
In this lab, we were trying to couple light from a Fabry-Perot semiconductor
diode laser into a single-mode ber using an imaging telescope. Furthermore,
we tried to measure the mode size of the optical ber and the spectrum of the
emitted light. It turned out that even though launching light into a single-mode
ber is a challenging task, the theoretically predicted coupling eciency could
be achieved.

Author of this lab report

1. Procedure, Results, Analysis, and Discussion EE234 Photonics Lab, Lab 4
1 Procedure, Results, Analysis, and Discussion
1.1 Light source setup
Our rst objective in this laboratory experiment was to assure parallel alignment
of our light source, a Fabry-Perot semiconductor laser operating at a 1550 nm wave-
length (Mitsubishi ML976H6F). This laser diode was mounted into a special mount
(Newport 700C), which incorporated a collimating lens (Newport 700-42). The
position of the laser inside the mount can be changed by adjusting two screws on the
outside of the mount. Since the vertical and horizontal position of the lens is xed,
this adjustment allows us to change the angle of the emitted light. We will show in
Section 1.2 that it is important to know the direction of the emitted light to within a
few mrad. Therefore, we tried to align the beam parallel to the surface of the optical
table and also parallel to a row of holes.
To do this, we used the common iterative procedure. An iris is placed close to the
laser and adjusted, so that the beam passes through its center. Then, one displaces
the iris as far along the beam as possible. At this position, the beam will most likely
not hit the center of the iris anymore and the vertical axis of the diode laser has to
be adjusted to bring the beam back to the center. Following this, one returns the iris
to its initial position and adjusts its height again, etc. Going through these steps a
couple of times should make the beam travel parallel to the optical table. A similar
procedure can be used to make the beam travel parallel to a row of holes.
Since we wanted to operate at an output power of 2 mW, the diode driver
(Newport 501AN) had to be set to 24.5 mA (see lab reports 1, 2, and 3).
1.2 Theoretical coupling eciencies
Optimum coupling into a ber depends crucially on the relative position of the ber
with respect to the laser beam. The following expressions give estimates for the
coupling eciencies for some cases of misalignment.
Axial Oset & Modematching If we want to couple an elliptical Gaussian beam
into a single-mode ber, we can express the coupling eciency in the case of
axial oset D as follows.
T =
4

2
ber
xy

1 +

2
ber

2
x

2
+
k
2

2
ber
4R
2
x

1 +

2
ber

2
y

2
+
k
2

2
ber
4R
2
y
Author: Carsten Langrock 2
1.2. Theoretical coupling eciencies EE234 Photonics Lab, Lab 4
with

x,y
=
x
0
,y
0

1 +

2
x
0
,y
0

1/2
, R
x,y
= D

1 +

2
x
0
,y
0
D

, and k =
2

Modematching In the case that we dont have any mechanical misalignment, we

are still left with the problem of matching an elliptical Gaussian mode to the
circular Gaussian mode of the ber. We are therefore left with the following
expression for the coupling eciency.
T =
4
2
ber

2
x
+
2
ber

2
y
+
2
ber
(1)
Tilt In case the ber axis is tilted with respect to the optical axis of the incoming
beam, we get the following theoretical coupling eciency.
T =
4
2
ber

2
x
+
2
ber

2
y
+
2
ber
exp

2
2

2
ber

2
x

2
x

2
x
+
2
ber
+

2
y

2
y

2
y
+
2
ber

where
x,y
are the tilt angles in the x and y direction, respectively.
Lateral Oset In case of later oset between the axis of the optical ber and the
optical axis of the incoming beam, the coupling coecient computes to
T =
4
2
ber

2
x
+
2
ber

2
y
+
2
ber
exp

2d
2
x

2
x
+
2
ber

2d
2
y

2
y
+
2
ber

where d
x,y
denotes the lateral oset in the x and y direction, respectively.
Since we tried our best to align the ber as well as possible, we are going to assume
that we dont have any tilt or lateral oset between the incoming laser beam and the
axis of the optical ber. Therefore, we will only pay attention to the expression that
deals with mode matching (Equation 1), since the mode of the laser beam is elliptical.
We measured this mode in Lab 3 and found the following values for the beam waists;

x
1.3 m and
y
1.32 m
The manufacturer of the laser diode doesnt quote the values for those two parameters,
but gives the parallel and perpendicular divergence angles
x
= 25

and
y
= 30

. It
is not immediately obvious, whether those values denote the full divergence angles or
Author: Carsten Langrock 3
1.2. Theoretical coupling eciencies EE234 Photonics Lab, Lab 4
0
0.2
0.4
0.6
0.8
1
0 0.01 0.02 0.03 0.04
Focal Length f (m)
C
o
u
p
l
i
n
g

C
o
e
f
f
i
c
i
e
n
t

T
Figure 1. Coupling eciency T as a function of focal length f for the case of modematching limited
coupling.
the half angles. Looking at the specsheet available on the internet, a graph of the far
eld pattern can be found on the last page. Looking at this graph it becomes clear
that
x,y
denote the half angles, the same way the divergence angle is usually dened.
This allows us to calculate the cross section of the laser diodes waveguide, assuming
that we have a diraction limited beam. With the relation

x,y
=

x,y
we nd

x
1.132 m and
y
0.942 m
For all future calculations, we will use these values for
x,y
. After propagation through
the rst lens with focal length 5 mm, the beam waists will be

x
2.179 mm and
y
2.619 mm
according to

x
2
,y
2
=
f

x
1
,y
1
(2)
To nd the optimum focal length for the second lens in the telescope conguration,
we simply insert Equation 2 into Equation 1, since we already know the mode size in
Author: Carsten Langrock 4
1.3. Experimental setup EE234 Photonics Lab, Lab 4
Laser mount
with TE-Cooler
Laser
Diode
Collimating
Lens, f
1
TE-Cooler
Controller
Current
Source
Collimating
Lens, f
2
x
y
z
Single-Mode Fiber
to power meter
x
y
z
Figure 2. Setup for laser-to-ber coupling experiment.
front of the second lens. We will assume that the beam waist lies one focal length in
front of the second lens and that the ber is placed one focal length after the second
lens. The resulting expression depends only on the focal length f (
ber
is given to
be 5.2 m). A plot of this function can be seen in Figure 1. The maximum is located
at f = 25.18 mm. At this point, the theoretical coupling eciency is T
25
= 99.16%.
Since we didnt have a lens with this focal length at the time of the experiment, we
used a lens (Newport F-L10B) with a 12 mm focal length. The theoretical coupling
eciency using this lens is T
12
= 60.03%.
1.3 Experimental setup
After having gone through the initial alignment procedure, as described in Section 1.1,
the ber coupler was placed approximately 10 cm in front of the laser mount. We
made sure that the axis of the coupler coincided with the axis of the laser beam. This
is done easiest, when the collimating lens and the ber are removed from the coupler.
The built-in iris helps to achieve a fairly precise placement of the xture.
Before installing the ber (Corning SMF-28) into the ber coupler, we had
to cleave the ber to assure an optically even surface (Fujikura CT-100B). Any
residual roughness would result in diuse surface reection, reducing the coupling
eciency vehemently. One end of the ber was placed inside the ber coupler ap-
proximately one focal length away from the second collimating lens, the other end was
brought in close contact with the power meter. To achieve close proximity with the
diode inside the power head, we removed its attenuator and adjusted the settings of
the power meter accordingly. A sketch of the setup is shown in Figure 2. To position
the ber with respect to the incoming laser beam, the ber coupler allowed to shift
Author: Carsten Langrock 5
1.3. Experimental setup EE234 Photonics Lab, Lab 4
the location of the negative lens, as well as the x-y-position of the ber itself. The
former allows for a ne adjustment of the incoming beam, while the latter can only
be used for coarse adjustment, due to the nature of the screws that have been im-
plemented. It turned out that backlash was a problem for either adjustable element.
This problem can certainly be overcome by more sophisticated translational stages.
As we all know, theoretical predictions are one thing, acquiring experimental
data is another thing. Launching light into a single-mode ber is certainly one of
the more complicated tasks and not far from being an art in itself. As so often, it
takes perseverance and a fair amount of luck to achieve optimum results that t the
theoretical predications closely. It must have been our turn this time, considering the
fact that the rest of the week was miserable enough. The maximum amount of light
that we were able to couple into the optical ber was
P = 1.16 mW
The transmittance of the second collimating lens is 94% at 1550 nm, according to the
Newport catalog. Assuming no further losses, the total incident power should have
been P
0
= 1.88 mW. This means that we achieved a coupling eciency of 61.7%,
which is slightly higher than the predicted theoretical value.
This result might be higher than initially expected, but there are some points that
have to be taken into account, besides the ingenious experimental capabilities of ours.
1. We should not forget that the beam waists that we measured in Lab 3 were
slightly larger than the ones that we used for the calculations of the theoretical
coupling eciency. Using those numbers, we calculate a maximum coupling
eciency of 77.7%, using a 12 mm focal length lens.
2. Furthermore, the mode size diameter of the ber is specied only within 0.8 m.
Takin this uncertainty into account, the theoretical maximum coupling e-
ciency lies between 54.5% and 66.1%.
3. We have to take into account that we dont really know exactly how much power
the laser diode was emitting at the moment of the measurement. Diode lasers
are know to exhibit uctuations from time to time. For future experiments it
would be advisable to have a means of constantly monitoring the diode laser.
This could be achieved by using a highly transmissive beam pick-o, which is
placed directly in front of the laser diode.
Author: Carsten Langrock 6
1.4. Mode size measurement EE234 Photonics Lab, Lab 4
4. Of minor concern is the fact that the theoretical expressions for the coupling
eciencies have been derived, assuming Gaussian mode propagation inside the
ber. This is certainly a fairly good approximation for the lowest order ber
mode and shouldnt cause major deviation from the exact values.
1.4 Mode size measurement
To measure the mode size diameter of the single-mode ber, we used a setup similar
to the one from Lab 3. We placed the tip of the ber one focal length in front of a
11 mm focal length lens and placed the mechanical chopper one focal length after the
lens. For convenience, we used the photodiode of the power head as the photoreceiver.
We measured the rise-time and fall-time for two dierent frequencies and calculated
the beam waist averaging those four values. Backing out the propagation of the light
through the lens, we arrived the following erroneous results;

x,y
0.85 m
This is wrong by a factor of 5, no matter how bad our measurements regarding the
radius of the chopping wheel, the distance between the ber and the lens, or the lens
and the chopping wheel might have been. At this point, we dont have a satisfying
explanation for this, but it is clear that a user error is most likely the culprit.
Repeating this measurement more carefully should deliver better results.
1.5 Laser diode spectrum
Let us turn our attention away from the mediocre result of the last section towards a
more pleasing one. We fed the output of the ber into an optical spectrum analyzer
(OSA) and measured the laser diodes spectrum between 1522 nm and 1557 nm. The
result can be seen in Figure 3. If there would have been any doubt, whether this diode
is of the Fabry-Perot or DFB kind, there should be no doubt any more after looking at
the plot. We are clearly seeing multimode operation, centered around 1540 nm. The
roundish shape of the peaks is a result of the lter shape. Reducing the resolution
bandwidth of the receiver (i.e. the OSA) should result in more pronounced peaks.
Comparing this result with the data provided by the manufacturer, we notice very
good agreement.
Author: Carsten Langrock 7
2. Conclusion EE234 Photonics Lab, Lab 4
Wa velength (nm)
15 25 15 30 15 35 15 40 15 45 15 50 15 55
S
i
g
n
a
l

A
m
p
l
i
t
u
d
e

(
d
B
m
)

-8 0
-7 0
-6 0
-5 0
-4 0
-3 0
Figure 3. Spectrum of a Fabry-Perot laser diode as measured by an optical spectrum analyzer.
2 Conclusion
Coupling light into a single-mode ber is a challenging experimental task. It depends
on many factors, such as knowledge of the laser diodes and bers mode sizes, as
well as endurance and persistence of the experimentalist. We were able to achieve a
fairly high coupling eciency, but didnt succeed in measuring the bers mode size
diameter accurately within a reasonable experimental error. The spectrum of the
laser diode showed us, that the diode is emitting several longitudinal modes, which
is common for Fabry-Perot laser diodes. The acquired skill how to launch light into
a ber will help us in future experimental setups. The problems we had in achieving
a high coupling eciency makes the need for standardized and precise ber coupling
connectors apparent. It wouldnt be economically, if every single ber had to be
treated the way we did.
Author: Carsten Langrock 8