Development of a Variable Spectral-Width, Wavelength Tunable Laser Using A Super- Luminescent Diode

Development of a Variable Spectral-Width, Wavelength Tunable Laser Using A Super- Luminescent Diode

Jake Royal

University of Minnesota

December 19, 2016

Abstract: A super-luminescent diode (SLD) with an optical feedback apparatus was used to create a variable spectral-width, wavelength tunable laser. The threshold current required for lasing was measured to be 87.5 ± 0.5 mA. The power output was monitored using different angles of reflectance out of the cavity in order to determine the optimal angle of a glass slide for maximum power generated by the laser. This angle is between 40° and 50°. Implementation of a diffraction grating allowed for the measurement of 72 discrete spectral modes with a mode spacing of 0.35 ± 0.04 nm.

Introduction & Theory: In order to effectively understand lasing it is important to know the different types of emission and under what circumstances they occur. Spontaneous emission is when electrons in the excited state randomly decay to a lower state. This process emits the difference in energy of the two states in the form of a single photon. The produced photon’s energy (E) is described by Equation 1.

(1) 𝐸 = h𝜈

Where h is Plank’s constant and ν is the frequency.

Stimulated emission occurs when a feedback photon with energy greater than or equal to the band gap energy (Eg) is incident upon an electron in the higher state forcing it to decay to the lower state. By conservation of momentum this produces a photon that is in phase or coherent with the original incident photon. The process is shown in figure 1.

Figure 1: A diagram of stimulated emission. The feedback photon is incident upon an electron in the excited state forcing it to decay to the lower state. It combines with a hole in the lower band and releases another photon that is coherent with the feedback photon. The extra emitted photons will amplify the output therefore increasing the intensity of the signal(Taken from [2]) .

The structure of the semiconducting device used dictates the band gap energy (Eg) between the conduction and valence bands and is represented by equation 2.

(2)

An SLD is a semiconductor device made up of a pn-junction equipped with a waveguide as shown in figure 2.

Figure 2: A diagram of an edge emitting super-luminescent diode. A potential is applied and light is emitted spontaneously from the partially transmitting facet on the end of a waveguide. The waveguide simply restricts wave propagation to two dimensions. The partially transmitting facet is coated with an anti-reflective material (Figure from [1]).

When a voltage is applied, holes from the p-doped band are injected into the n-doped band, which ejects electrons into the p-type material. Energy from the electrons and holes combining is conserved and light is produced and restricted to populate in two dimensions by the waveguide. This light is created by both spontaneous and stimulated emission. In order for the device to lase, we need the stimulated emission to outweigh the produced spontaneous emission. This is achieved by adding an external feedback cavity allowing for more incident photons in the cavity, promoting more stimulated emission.

With the implementation of an external feedback cavity the devise will lase above a certain current threshold. The bandwidth will shrink allowing us to observe multiple fine spectral modes that exist above the gain threshold. The inclusion on an external feedback cavity presents boundary conditions for which modes are able to oscillate. Equation 3 shows the first of which (from [5]),

(3)

Here λ is wavelength, n is an integer, and L is cavity length. Equation 3 shows that only half wavelength integers are able to populate inside the cavity. Therefore the number of mode oscillations will be discrete. The spacing between frequency modes is then (from [4]),

(4)

Another restriction due to the cavity is that only modes that overtake the gain threshold of the SLD (the total loss) are able to survive. The gain loss inside the cavity will be due to absorption and scattering. In order to make the laser device tunable, we will use a diffraction grating in a Littrow configuration to measure and analyze the multiple modes produced.

Experimental Apparatus:

Figure 3: The first experimental apparatus diagram used to determine threshold current at required for lasing. The filter is replaced by a glass slide later in order to measure the best angle to output power from the laser. (Figure from [3])

Figure 3 shows the first apparatus used to induce feedback and to determine the best possible feedback power and threshold current required for lasing to occur. Current is supplied to the mounted SLD (ThorLabs SLD830S-A10). The SLD has a normal quoted central frequency of 830nm and a near-Gaussian spectrum. A ThorLabs temperature controller also monitors the SLD. Light is emitted through a collimating lens that directs the beam. It is then reflected back into the internal cavity by an adjustable gold mirror. The light is then split by a neutral density filter allowing us to dictate the feedback, monitor the spectrum, and the power using a the Ocean Optics spectrometer and a power meter respectively. The Neutral density filter was replaced by a glass slide and the incident beam's angle was changed while power was monitored to determine the best angle for outputing the most power attributed to the gain of the laser. The Fresnel equation for transverse electrical polarized light (TE) was used to calculate the respective Reflection coefficients in terms of the angle of incidence. The Fresnel equation for the TE reflection coefficient(from [5]) is,

(5)

Where n1 is the refraction index of air equal to 1, n2 is the refraction index of glass equal to 1.5, θ1 is the angle of incident light and θ2 is the angle of the refracted wave. Since the refracted wave lies in the plane of incidence Snell’s Law is satisfied. Snell’s Law is [5],

(6)

and θ2 can be found using the following equation[5].

(7)

Finally the reflection coefficient is [5],

(8)

This coefficient is used to calculate initial power of the SLD in terms of angle. We then measure the power reflected on both sides of the device in order to measure power attributed to the gain of the laser for certain incident angles. The tunable laser setup is shown in figure 4.

Figure 4: The experimental setup used for wavelength tuning. (Figure from [3])

Figure 4 shows the apparatus for the final wavelength tunable laser devise. A diffraction grating is implanted this time to reflect light creating the external cavity. The diffraction grating allows us to tune to different mode frequencies by adjusting the angle.

Results & Analysis:

With maximum feedback applied we increased the injection current and observed the optical power until the current threshold for lasing to occur was found by plotting optical intensity vs injection current shown in figure 5.

Figure 6: The power vs injection current observed.

When maximum feedback is induced, there is a spike in optical power that represents the onset of lasing and the dominant emission form is stimulated. The observed current threshold required for lasing was 87.5 ± 0.5 mA.

The neutral density filter in the first setup was replaced by a glass slide in order to measure the best angle of incidence to ensure the power out of the cavity was maximized. The next figure shows the graph of initial power vs incident angle of light compared with the observed power attributed to gain from the laser.

Figure 7: On the left is power attributed to gain from the feedback cavity vs the angle of incidence. On the right is the usable initial power from the SLD vs the angle of incidence.

The maximum initial power is achieved between 40 and 50 degrees producing about 12mW of power. The power attributed to gain is also maximized between 40 and 50 degrees providing about 9mW of usable power. After this area, as the angle increases, more power is lost due to scattering and eventually at around 80 degrees the angle is to large to effectively provide feedback and lasing is ceased. We have shown that it is best to deflect light out of the cavity at around 45° of incidence in order to maximize usable power.

We were able to observe 72 separate discrete modes from our tunable laser. The supplied current was 126mA in order to ensure lasing. The modes observed are shown in the following figure.

Figure 8: 72 observed frequency modes at 126mA of injection current.

Unfortunately, we were unable to avoid clipping of the spectrometer at higher intensity levels. However, it is clear that the observed modes represent the makeup of a Gaussian gain envelope between 822nm and 848nm. The spacing of these modes were calculated by plotting using LSQfit the first half of the gain envelope’s peak mode values vs the number of modes. The slope of this LSQfit line is equal to the mode spacing and is reported to be 0.35 ± 0.04 nm.

Conclusion: A tunable laser was created using a super-luminescent diode and an external feedback apparatus. The measured threshold current for the device to lase was found to be 87.5 ± 0.5 mA. A glass slide was used to deflect light out of the cavity and the Fresnel equation for transverse electric polarized light was used to determine the reflectance coefficient dependent upon the angle of the slide. Output power of the laser was measured and used along with the reflective coefficient to plot the usable power attributed to gain of the laser. It was observed that the best angle of incident light used to deflect power out of the cavity is between 40° and 50°. A diffraction grating was then implemented to reflect zero order light back into the SLD acting as our feedback apparatus. The angle of the grating was changed in order to tune the laser to different wavelengths. We were able to observe 72 discrete modes from our laser at 126mA of supplied current. The mode spacing was determined using LSQfit and is reported to be 0.35 ± 0.04 nm, with a reduced χ2 equal to 0.65 indicating an overestimation of error but an acceptable value of the spacing for the observed gain envelope. This experiment could be improved by determining a lower injection current that does not produce clipping of the spectrometer when finding the number of observable modes.

Acknowledgments: A special thank you to Kurt Wick for his guidance during the course of this experiment.

References:

[1] Nave, Carl Rod. "The P-N Junction." Hyper Physics. 1999. Web. 03 Mar. 2016

14

[2]

"Stimulated Emission." Wikipedia. Wikimedia Foundation, n.d. Web. 26 Oct. 2016.

[3] DeMars, Luke A., and Erik L. Tilseth. "S15SLD - MXP." S15SLD - MXP. N.p., n.d. Web. 13

Oct. 2016.

[4] Sheu, F., & Luo, P. (2007). Development of a variable spectral-width, wavelength-tunable

light source using a superluminescent diode with optical feedback. American Journal of Physics,

769-769.

[5] Saleh, B., & Teich, M. (1991). Fundamentals of Photonics. New York: Wiley.