Distributed Bragg reflector (DBR) laser

A distributed Bragg reflector (DBR) laser is a semiconductor laser whose cavity is closed at one or both ends by a passive Bragg grating instead of a cleaved facet. The grating, a periodic corrugation of the waveguide beside the gain section, reflects strongly only near its Bragg wavelength (Bragg condition),

$${\lambda }_B=2n_{\text{eff}}\mathrm{\Lambda },$$

where $n_{\text{eff}}$ is the effective index of the waveguide mode and $\mathrm{\Lambda }$ the grating period, 242 nm for 1550 nm at $n_{\text{eff}}$ = 3.2. Of the cavity's longitudinal modes, the one closest to that wavelength lases. In a DFB laser the grating runs along the gain region itself; in a DBR laser the gratings are etched in passive waveguide sections, transparent at the lasing wavelength and with electrodes of their own, so the wavelength can be tuned without changing the gain. A tunable DBR laser on indium phosphide typically has a gain section, a phase section and one or two grating sections, each with its own current, and often an integrated semiconductor optical amplifier at the output to set the power.

How the grating selects the mode

A uniform grating of length $L_g$ and coupling coefficient $\kappa$ reflects a fraction ${\tanh }^2(\kappa L_g)$ of the power at the Bragg wavelength, over a band whose width between the first zeros is

$$\mathrm{\Delta }\lambda =\frac{{\lambda }^2}{\pi n_g{L}_g}\sqrt{(\kappa L_g{)}^2+{\pi }^2},$$

with $n_g$ the group index. The light is reflected progressively along the grating rather than at one plane, and for the spacing of the cavity modes the grating acts as a mirror placed an effective length $L_{\text{eff}}=\tanh (\kappa L_g)\mathrm{/}2\kappa$ inside it. As an example at 1550 nm with a group index of 3.6, a rear grating 300 µm long with $\kappa$ = 50 cm⁻¹ ($\kappa L_g$ = 1.5) reflects 82% at its peak, over a band 2.5 nm wide between the zeros (1.6 nm at half maximum), and has an effective length of 91 µm. With a 400 µm gain section and a 150 µm phase section, the effective cavity length is 641 µm and the modes are 0.52 nm apart, ${\lambda }^2\mathrm{/}(2n_{g}L)$, so about five modes fall inside the reflection band. When one mode sits at the peak, its neighbors see a reflectivity of 70% and the next ones 10%, and the higher mirror loss holds them below threshold (side-mode suppression ratio).

Tuning

Current injected into a grating section raises its carrier density, which lowers the refractive index through the free-carrier effect and moves the Bragg wavelength to shorter wavelengths. Heating raises the index and moves it the other way, by about 0.09 nm/K at 1550 nm, as in a DFB laser (wavelength temperature coefficient). In either case

$$\frac{\mathrm{\Delta }{\lambda }_B}{{\lambda }_B}=\frac{\mathrm{\Delta }{n}_{\text{eff}}}{n_g},$$

so an index change of −0.01 moves the Bragg peak by 4.3 nm at 1550 nm. The range is limited by how far injection can lower the index before the heat dissipated by the current, which raises it, takes over, and a single grating section tunes by several nanometers.

Tuning the grating alone does not tune the laser smoothly. A cavity mode moves only in proportion to the part of the cavity whose index changes: in the example the grating's effective length is 14% of the cavity, so a mode moves by 14% of the Bragg shift. The peak therefore slides past the modes, and the laser hops to the next mode, 0.52 nm away, each time the peak has moved by 0.61 nm, tuning in a staircase with 0.09 nm of continuous tuning between hops. A phase section, a passive section whose index is tuned in the same way, shifts the modes independently. For the laser to tune continuously, the optical length of the cavity has to change by the same fraction as the Bragg wavelength, and since the gain section's index is held fixed by the clamped carrier density above threshold, a phase section of length $L_p$ beside a gain section of length $L_a$ needs

$$\mathrm{\Delta }{n}_p{L}_p=\mathrm{\Delta }{n}_B(L_a+L_p),$$

where $\mathrm{\Delta }{n}_p$ and $\mathrm{\Delta }{n}_B$ are the index changes in the phase and grating sections. In the example the phase section needs 3.7 times the index change of the grating, so if both are limited to 0.01, continuous tuning covers 1.2 nm of the 4.3 nm the grating can reach. A mode can still be placed at any wavelength within the full range, since shifting the modes by one spacing takes an index change of only 0.005 in the phase section. Three-section lasers are operated this way, with each wavelength set by a pair of currents and hops between neighboring settings.

Sampled gratings and the Vernier effect

A sampled grating is a Bragg grating interrupted periodically, with bursts of grating repeated every sampling period $Z$. It reflects a comb of peaks spaced by

$$\mathrm{\Delta }{\lambda }_s=\frac{{\lambda }^2}{2n_{g}Z}\mathrm{.}$$

An SG-DBR laser has a sampled grating at each end, with slightly different sampling periods, so the two combs have slightly different spacings and their peaks coincide at only one wavelength within a wide range, where the laser operates. Shifting one comb against the other by the difference in spacing moves the coincidence to the next pair of peaks, a Vernier effect that turns a small index change into a large change of wavelength. As an example at 1550 nm with a group index of 3.6, sampling periods of 60 and 55 µm give combs spaced 5.56 and 6.07 nm apart. The coincidence repeats only every 67 nm, $\mathrm{\Delta }{\lambda }_1\mathrm{\Delta }{\lambda }_2\mathrm{/}(\mathrm{\Delta }{\lambda }_2-\mathrm{\Delta }{\lambda }_1)$, and shifting one comb by 0.51 nm moves the lasing wavelength by 5.6 nm, eleven times as far. Tuning both mirrors together fills in between the peaks, and the phase section aligns a cavity mode with the coincidence, so three tuning currents, besides the gain current, set the wavelength; the currents for each channel are measured for each laser and stored in a table. Jayaraman and colleagues set out the theory and design of sampled-grating lasers, and superstructure gratings, whose period is chirped within each sampling period, have tuned a DBR laser over more than 100 nm (Tohmori and colleagues). Other widely tunable lasers, such as the digital supermode DBR and the modulated-grating Y-branch laser, use different mirror arrangements for the same purpose.

Linewidth and control

Current injected into a tuning section adds noise of its own: random fluctuations of the carrier density make the index, and with it the lasing frequency, fluctuate, so the linewidth of a DBR laser broadens as tuning current is applied (Amann and Schimpe). The lasing wavelength also depends on the temperature and, through heating, on every current, so a DBR laser in a transmitter is held on its channel by a wavelength locker, an etalon whose transmission changes with frequency, together with a thermoelectric cooler.

DBR and DFB lasers

A DFB laser has one drive current and tunes only through its temperature and current, by a few nanometers, but it has no mode hops within that range and needs no calibration. A DBR laser separates wavelength selection from gain, so it can be tuned electronically at constant output power, at the cost of more currents, a calibration table and mode hops that have to be controlled. The comparison table in the DFB laser entry sets the two against the Fabry-Perot laser, and external cavity lasers reach narrower linewidths and wider continuous tuning with a separate filter. The name is also used for fiber lasers whose mirrors are two fiber Bragg gratings written around a short doped section (fiber laser), and the mirrors of a VCSEL are distributed Bragg reflectors of the thin-film kind (distributed Bragg reflector).

Common questions

What is a DBR laser?

A semiconductor laser whose mirrors are passive Bragg gratings beside the gain section. The grating reflects only near its Bragg wavelength, so one longitudinal mode lases, and the wavelength can be tuned by current through the grating sections.

What is the difference between a DBR laser and a DFB laser?

In a DFB laser the grating runs along the gain region and there is one drive current; in a DBR laser the gratings are separate passive sections with their own currents, which can be tuned without changing the gain. DFB lasers are simpler and tune by a few nanometers through temperature; DBR lasers tune further, electronically, but need calibration to avoid mode hops.

How is a DBR laser tuned?

By current injected into the grating section, which lowers its index and shifts the Bragg wavelength to shorter wavelengths, and by temperature, about 0.09 nm/K at 1550 nm. A phase section moves the cavity modes with the grating, so that the laser tunes without hopping over part of the range and can reach any wavelength within it.

What is an SG-DBR laser?

A DBR laser with two sampled gratings, each reflecting a comb of peaks with a slightly different spacing. The laser operates where two peaks coincide, and shifting one comb moves the coincidence by many times the shift, which gives tuning over tens of nanometers with three tuning currents.

References: L. A. Coldren et al., "Tunable semiconductor lasers: a tutorial," Journal of Lightwave Technology 22, 193 (2004). L. A. Coldren, S. W. Corzine and M. L. Mašanović, Diode Lasers and Photonic Integrated Circuits, 2nd ed. (Wiley, 2012). J. Buus, M.-C. Amann and D. J. Blumenthal, Tunable Laser Diodes and Related Optical Sources, 2nd ed. (Wiley-IEEE Press, 2005). V. Jayaraman, Z.-M. Chuang and L. A. Coldren, "Theory, design, and performance of extended tuning range semiconductor lasers with sampled gratings," IEEE Journal of Quantum Electronics 29, 1824 (1993). Y. Tohmori et al., "Over 100 nm wavelength tuning in superstructure grating (SSG) DBR lasers," Electronics Letters 29, 352 (1993). M.-C. Amann and R. Schimpe, "Excess linewidth broadening in wavelength-tunable laser diodes," Electronics Letters 26, 279 (1990).