Proceedings of National Laser Symposium (NLS-22), Manipal University, Manipal.

8-11 Jan 2014

Design and Implementation of Multilayered broad-Band Ultra-Low Reflective Coating

on Facets of Laser Diode to Suppress Optical Feedback
Sanjaykumar J. Patel1, Vipul Kheraj1*, C. J. Panchal2
1
Department of Applied Physics, Sardar Vallabhbhai National Institute of Technology, Surat-395 007, India
2
Applied Physics Department, Faculty of Technology & Engineering, M. S. University of Baroda,
Vadodara – 390 001, India
*Corresponding Author E mail: [email protected]
Abstract
The multilayered coating has been designed and implemented on the facets of edge-emitting laser diode
in order to achieve ultra-low reflectivity over a broad wavelength range for suppression of optical feed-back in
the laser cavity. The design of the multilayer coating has been obtained by a self-developed programme based
on genetic algoritm to achieve low-reflectivity of the order of 0.1% over a spectral width of 50 nm around the
central lasing wavelength of 818 nm and successfully implemented on the laser-diode facets. The effects of
facets coating on the optical power emission and the spectral response of the lasers have been investigated. It is
demonstrated that the simple design of multilayered coating obtained using a self-developed genetic algorithm
can successfully reduce the optical feedback from the cavity and prevent the lasing action from the structure,
which can be very usedful for the fabricaton of Superliuminescent diodes.
Introduction
Superluminescent Light Emitting Diode (SLED) is a semiconductor light source that shows relatively a
broad spectral linewidth of optical emission with low coherence like LEDs and high optical power like laser
diodes. SLED has a device structure apparently very similar to laser diodes, except that the optical feedback is
not efficient to achieve lasing action. Due to the elimination of optical feedback, the cavity modes are
suppressed which results in optical emission with a low ripple or smooth spectra. In other words, the SLED
combines the spatial coherence of a laser diode with the temporal incoherence of an LED. This combination is
extremely useful for a wide range of applications including fiber-optic gyroscopes [1], optical coherence
tomography (OCT) systems [2], optical time-domain reflectometers (OTDRs) [3], and wavelength-division
multiplexing (WDM) testing systems [4].
The low facet reflectivity is one of key concern to achieve high-power SLED operation in order to
prevent lasing even at the required high pumping levels. Several methods have been reported to suppress lasing
action such as an unpumped, absorbing region [5] or the proton implanted absorbing region [6] at the facet.
Tilting [7], or bending [8] the injection contact with respect to the output facet has also been used for reducing
the optical feedback. However, all these techniques require compound structure and involve complex
processing. Anti-Reflection (AR) facet coating [9] is one of the simplest ways of reducing facet reflectivity
besides protecting the emitting facet from the environmental influences.
Here we report the design and implementation of multilayered AR coating to achieve the ultra-low
reflectivity over a broad wavelength range. Although the standard schemes are available for AR coatings,
generally these schemes give low reflectivity over relatively a narrow wavelength band. However, the SLEDs
and other semiconductor light emitters require the reflectivity modulations over broader wavelength band for
which one requires a more efficient algorithm. We have used a self-developed programme based on genetic
algorithm to design the multilayered stack of the thin-film. The designed thin-film structure has been
implemented on both the facets of edge emitting laser diode in order to test its effectiveness in suppressing the
Proceedings of National Laser Symposium (NLS-22), Manipal University, Manipal. 8-11 Jan 2014

Fabry-Perot cavity modes. The L-I characteristics and spectral response of the laser diode were measured before
and after the AR coating on the facets.
Design of AR coating
A multilayer AR coating on a GaAs substrate was designed using genetic algorithm. GA is stochastic
global search optimization algorithm inspired by Darwin’s theory of natural selection [10]. It is essentially
mimicking the process of natural evolution underlying the idea of survival of fittest where the fitness of
individual is improved by successive iteration through the processes of selection, crossover and mutation. In the
beginning of GA, various parameters such as wavelength range, number of layers, name of coating materials,
range of minimum and maximum thickness of layers, probability of mutation and number of iteration are fed
into the program to search optimum design of multilayer AR coating. For the implementation of GA, a large
population of multilayer AR designs with fixed number of layers is created randomly. Each AR design is tested
by fitness function based on their figure of merit. We have used the averaged reflectivity of AR stack over the
wavelength range of interest as a fitness function, which is defined as,
 p 
 ∑ [ R cal ( λ k ) − R desired (λ k )
(1)
(λ ) =  
k =1
R averaged
p

Where, R cal ( λ k ) is the calculated value of reflectivity at wavelength ( λ k ) using transfer matrix method

described elsewhere [11] and R desired ( λ k ) is target reflectivity, which is taken to be zero for this case. p is the
total number of wavelength steps of spectrum. After evaluating the fitness of each AR design, they are followed
by selection, crossover and mutation operator to reproduce new AR design. This process is continued till
optimum design is obtained in terms of layer thickness. Figure 1 shows simplified flow-chart of the genetic
algorithm based programme. The selected design for the laser diode facets is shown in table 1.

Initialize with random population

of multilayer AR designs Number of Layer
Materials
Layers thickness in Å
Evaluate the fitness for each multilayer AR Substrate GaAs ---
design over the defined wavelength range 1 MgF2 283
2 Si 203
3 MgF2 1530
Yes Incident medium Air ---
If optimum design is Stop
obtained? Total physical
2016
thickness (Å)
No Average
0.0138
Reflectivity (%)
Perform Selection
Reflectivity (%) at
Centre wavelength 0.0103
Perform Crossover (818 nm)

Perform Mutation

Fig.1: Flow chart of GA for AR stack design Table.1: AR stack design for facet coating

Experiemental details

The deposition of multilayer AR coating was accomplished under the high vacuum (10-6 mbar) using a 3
kW e-beam evaporation system equipped with 180 bend e-beam gun facility (Hind High Vacuum Co. (P) Ltd.).
The thickness and deposition rate of each layer in a stack is monitored by a quartz crystal oscillator. Since the
laser diode to be coated contains GaAs as the active layer material, the deposition of multilayered structures
Proceedings of National Laser Symposium (NLS-22), Manipal University, Manipal. 8-11 Jan 2014

were optimized on GaAs test substrate. After the optimization, the AR film was deposited on the facets of laser
diode with a GaAs substrate kept in close vicinity to the laser diode during the deposition for the estimation of
the reflectivity of coated thin film. The test-substrate and the laser diode are rotated with constant rpm inside the
vacuum chamber during deposition. The substrate temperature is attained using radiant heaters. The reflectivity
of the AR coated substrates were measured using ex-situ reflectivity measurement set up. The L-I characteristic
and spectral response of laser diodes were measured in pulse mode before and after the facets coating using
laser diode characterization facility developed at our laboratory [12].
Results and Discussion
The optimized multilayered AR coating was deposited on the laser diode facets as well as the GaAs test
substrate. Figure 2 shows the simulated and experimentally measured reflectivity spectra on coated test-
substrate. As shown in figure, the nature of simulated and experimental reflectivity spectra are in very good
agreement. We obtained as low as 0.1% reflectivity over the desired wavelength range as shown in the figure.
Although the obtained reflectivity is much higher than the simulated one, the overlaping nature of the simulated
experimental spectra indecate that the quality and the control of multilayer stack deposition is good. Also, the
higher value of reflectivity obtained experimentally above 850 nm is because of back reflection from the GaAs
test substrate as it becomes transperant whereas the simulatoin assumes the substrate to be infinte, thus by
neglecting the reflection from the back surface.

Fig.2: Simulated and Experimental Reflectivity Spectra for the optimized multilayer AR coating on GaAs test
substrate

In order to understand the effects of AR coating on the performance of the laser diode, we measured the L-I
characteristics of the laser diode before coating, after applying ultra-low reflective coatings on one facet and
after applying the coating on both the facets. The L-I characeterstics measurements were carried out in pulsed
mode with pulse with of 400 microsecond and the dutycycle of 0.25%. Figure 3 (a) shows the comparision of L-
I characteristics of the laser diode before coating and after one facet coating. As shown in the figure, the laser
diode shows a sharp threshold at about 330 mA with good slope efficiency. We obtained 450 mW optical power
per facet at 800 mA from uncoated facets as shown in figure. However, after applying the AR coating on one of
the facets, the L-I characteristics exhibit significant changes. In case of one facet coated diode, the reflectivity of
the coated facet is very low, of the order of 0.1% at 818 nm where as the uncoated facet has the normal
reflectivity of about 30% for GaAs-Air interface. Hence in this case, we measured the L-I characeteristics from
both, coated and uncoated facet, and the difference of power from the two facets is evident as shown in the
Proceedings of National Laser Symposium (NLS-22), Manipal University, Manipal. 8-11 Jan 2014

figure.3(a). The coated facets emit more light because of lower reflectance compared to the uncoated facet.
However, as seen from the figure, even after applying the AR coaitng on one of the facets, the sharp change in
the optical power is still observed from both the facets, albait at the higher current. The threshold current in
this case is found to have increased to 430 mA whereas the the overall power reduces from the original value
before coating. This can be explained by the fact that the mirror loss increases with decreasing the reflectivity of
one of the facets [13], which leads to the requirement of higher current for the transperancy condition.

(a) (b)

Fig.3(a):L-I characteristics of laser diode before AR coatings and after AR coating on one facet, (b)
comparision of L-I charcteristics before and after AR coating on both the facets.

The comparision of L-I characteristics of uncoated laser diode and the laser diode coated with AR coating on
both the facets is shown in figere 3 (b). As shown in the figure, the sharp threshold seen in the L-I characteristics
before coating totally dissappears after coating both the facets with the AR film and the L-I curve shows the
continuos increment in the slop. This indicates that the incresed mirror loss and reduced optical feedback from
the facets now prevents the lasing in the structure. To make this effect clearly visible, the plot is zoomed in near
the threshold current, which is shown in the inset of figure 3 (b). Also, the non-linear nature of the L-I curve
hints towards the presence of superluminescence emission [14].

(a) (b)

Fig.4: Spetral response of laser diode at different drive currents (a) before AR coatings and (b) after AR
coatings on both the facets.

To further investigate this behavior, we measured the spectral response of the laser diode before and after AR
coatings on both the facets, which is shown in figure 4. Figure 4 (a) shows the spectral response of laser diodes
at different drive currents near the threshold value before AR coating. As seen from the figure, the laser diode
exhibits distinct Fabri-parot cavity modes. As the current increases, one of the modes become dominent with a
Proceedings of National Laser Symposium (NLS-22), Manipal University, Manipal. 8-11 Jan 2014

few week modes still present. Thus, the peak is highly asymmetric because of presence of multiple oscillating
modes in the spectrum. However, as seen from figure 4 (b), the spetral response of laser diode after AR coatings
on both the facets is quite symmetric and smooth. This clearly indecate that the application of ultra-low
reflective coatings on both the facets effectively remove the optical feed back [15]. The lack of optical feedback
from the mirrors at the facets eliminates the multiple oscillating cavity modes and ripples to make the spectrum
smooth and symmetric.
Conclusion

We have designed and successfully implemented multilayer ultra-low AR films on the facets of laser
diode. The deposited AR film provides as low as 0.1 % reflectivity on the laser diode facets over the spectral
width of 50 nm around the central lasing wavelength at 818 nm. The obtained L-I characteristics after multilayer
AR coating of laser diodes clearly indicate the effect of reflectivity modulation on the facets. The suppression of
optical feedback is confirmed by observing the spectral response of laser diode before and after coating. It is
demonstrated that the simple dielectric AR coating on the facets could be a very useful technique for
suppression of optical feedback to eliminate the resonant cavity for fabrication of SLEDs.
Acknowledgement
Authors are thankful to Board of Research in Nuclear Science (BRNS), Department of Atomic Energy (DAE),
BARC, Mumbai for providing financial assistance.
References
1. W. K. Burns, C. I. Chen and R. P. Moeller, J Light wave Technol LT-1, 98 (1983).
2. D. Huang et.al, Science 254, 1178(1991).
3. K. Takada, I. Yokohama, K. China and J. Noda, Appl. Opt 26, 1603(1987).
4. K. Y. Liou and G. Raybon, IEEE Photon. Technol. Lett 7, 1025 (1995).
5. I. M. Joindot and C. Y. Boisrobert, IEEE J. Quantum Electron 25, 1659(1989).
6. C. S. Wang et.al, Appl. Phys. Lett 47, 587(1982).
7. T. Yamatoya, S. Mori, F. Koyama, and K. Iga, Jpn. J. Appl. Phys, 38, 5121(1999).
8. L. Fu, H. Schweizer et.al. IEEE J. Quantum Electron, 40, 1270 (2004).
9. Youngsheng Zhao et.al. Optical and Quantum Electronics, 28, 1685 (1996).
10. J. H. Holland, Adaptation in Natural and Artificial Systems (The university of Michigan Press, An
Arbor,1975).
11 V. A. Kheraj, C. J. Panchal, M. S. Desai, V. Potbhare, Pramana: Journal of Physics 72, 1011 (2009).
12. V. A. Kheraj, P. K. Patel, C. J. Panchal, and T. K. Sharma, Proc. of Sixth DAE-BRNS National Laser
Symposium, RRCAT, Indore, pp. 70 (2006).
13. V. A. Kheraj et.al. Journal of Optics and Laser Technology 39, 1395 (2007).
14. E. Feltin et.al. Appl. Phys. Lett. 95, 081107 (2009).
15. S. K. Kwong et.al. Appl.Phys. Lett. 51, 1879 (1987).