Self-injection locking of an extraordinarily widebroad-area diode laser

with a 1000-mm-wide emitter

Mingjun Chi, Birgitte Thestrup, and Paul Michael PetersenOptics and Plasma Research Department, Risø National Laboratory, DK-4000 Roskilde, Denmark

Received November 11, 2004

The experimental results of self-injection locking of an antireflection-coated broad-area diode laser with a1000-mm-wide emitting area are presented. To our knowledge, it is the broadest single-element diode laserthat has been used in an external-feedback cavity until now. Usually, wide diode lasers suffer from filamen-tation, which leads to poor spatial beam quality. We show, however, that the beam quality of the diode laseris improved significantly when we use asymmetric self-injection locking. An output power of 2.05 W is ob-tained with a beam quality factor M2 of 2.7. The self-injection locking technique improves the beam qualityby a factor of 107. By comparing the results with those obtained with an ordinarily coated diode laser witha 1000-mm-wide emitter we show that antireflection coating on the front facet is decisive for this improve-ment in the beam quality. © 2005 Optical Society of America

OCIS codes: 140.0140, 140.2020, 140.5960, 070.6110.

Broad-area diode lasers (BALs) can produce largeamounts of optical power, and they are attractive be-cause of their compactness, long lifetimes, and rela-tively low cost. However, these devices suffer frompoor spatial coherence owing to the broad emitter ap-erture in the slow axis, typically several tens to a fewhundred micrometers wide. Various methods for im-proving the beam quality of the multimode output ofBALs have been suggested. These methods fall intotwo categories: injection locking to an externalsingle-mode master laser1–3 and external-cavity feed-back with a spatial filter, also called self-injectionlocking (SIL).4–9 The SIL technique is the more at-tractive method because no additional single-modelaser is required and therefore a compact system canbe set up.

The commercially available BALs can produce sev-eral watts of optical power, whereas diode laser barscan produce power from several tens to hundreds ofwatts. Unfortunately, the techniques mentionedabove are not suitable for diode laser bars, as it is dif-ficult to lock the different noncoherent emitters in adiode laser bar simultaneously with a single-modemaster laser or with the SIL technique. The outputpower of single-element BALs scales with the widthof the emitting area. However, very broad BALs (thewidth of the emitter is larger than 500 mm) sufferfrom lateral lasing and filamentation. A way to elimi-nate these phenomena could be injection locking witha master laser or SIL technique. To obtain severalwatts of optical power from a BAL and at the sametime significantly reduce beam quality factor M2, wesuggest applying the injection locking or the SILtechnique to a BAL with an extraordinarily broadstripe width.

Here we present the experimental results of SIL ina custom-made 1000-mm-wide antireflection-coatedBAL. To our knowledge, it is the broadest single-element diode laser used until now. An output powerof 2.05 W is obtained with a value of beam qualityfactor M2 of 2.7, whereas the M2 value is 292 for thefreely running laser at the same operating current.

The diode laser used in the experiment was an830-nm gain-guided BAL. The emitting area of thelaser chip was 1 mm31000 mm, and the length of thelaser cavity was 2.5 mm. The reflectivity of the frontfacet of the diode was originally ,6%, the thresholdcurrent was 4.1 A, and the diode laser could emit10 W of power at an injected current of 13.0 A. Weapplied an additional antireflection coating sR,0.01% d to the front facet of the diode, which in-creased the threshold current to 7.2 A.

The experimental setup is similar to that describedin Ref. 7 and shown in Fig. 1. The first lens is used tocollimate the laser beam in the fast axis. It is an as-pherical microlens of 0.91-mm focal length with aN.A. of 0.85. A cylindrical lens of 60-mm focal lengthis used to collimate the beam in the slow axis and totransform the near field to the far field simulta-neously with respect to the slow axis. These twolenses are antireflection coated for the near-infraredwavelengths. At the pseudo-far-field plane, a mirrorwith a sharp edge combined with a razor blade isused as a mirror stripe for selection of a spatial modeand its reflection back into the active area of the di-ode laser; the width of the stripe is ,500 mm. A beamsplitter is inserted into the external cavity, and thereflected beam is used as a diagnostic beam. Anotheradjustable aperture just behind the mirror is used toblock the sidelobes, and the main lobe that passesthrough the aperture is the output beam of this SIL

Fig. 1. Experimental setup of the BAL system with SIL:NF, near field; FF, far field; HMR, half-mirror and razorblade (the units are millimeters).

May 15, 2005 / Vol. 30, No. 10 / OPTICS LETTERS 1147

0146-9592/05/101147-3/$15.00 © 2005 Optical Society of America

laser system. The width of the aperture is 800 mm.The output power and the beam quality factor aremeasured behind this aperture.

It is well known that the output of a gain-guidedBAL consists of a set of transverse spatial modes.10,11

Each mode has a double-lobed intensity profile in thefar field, and these modes are distinguished at differ-ent positions in the far field. The purpose of the SILtechnique is to select one lobe of a spatial mode at thefar field and reflect it back into the laser cavity.Thereby the other lobe of this mode is amplified andcoupled out of the external cavity as the output beamof the laser system, and all the other modes are sup-pressed effectively.4–9

The far-field profiles along the slow axis are mea-sured with and without the SIL at different operatingcurrents. Figure 2 shows the measured results atcurrents of 7.0 A [Fig. 2(a)] and 11.0 A [Fig. 2(b)].The profiles are measured in the diagnostic beam byuse of a photon beam analyzer (Photon, Inc., model0180). A cylindrical lens of 40-mm focal length im-ages the pseudo far field at the position of the feed-back mirror to the beam analyzer.

At an operating current of 7.0 A, just below thethreshold current for the freely running laser, asshown in Fig. 2(a), chiefly spontaneous emission con-tributes to the freely running laser output. However,strong lasing takes place when the SIL is applied.This indicates that the application of SIL reduces thethreshold of the diode laser significantly. At this cur-rent level the effect of SIL is a 40-times reduction inthe output beam’s width and a 15-times increase inthe peak intensity compared with the freely running

condition. At an injected current of 11.0 A, as shownin Fig. 2(b), the effect of SIL is a 36-times reductionin the output beam’s width and an 11-times increasein the peak intensity compared with the freely run-ning condition. In the freely running condition theobserved relatively uniform far-field pattern consistsof an incoherent superposition of multiple spatialmodes.10,11 The far-field intensity profile changes to astrongly asymmetric profile when SIL is applied.This situation holds for all the setups with asymmet-ric feedback.4–9 The output powers from the SIL di-ode laser system at operating currents of 7.0 A and11.0 are 0.59 and 2.05 W, respectively.

One estimates the beam quality of the output beamalong the slow axis by measuring beam quality factorM2 for the SIL diode laser system. A cylindrical lenswith a 60-mm focal length is used to focus a part ofthe output beam reflected by a beam splitter. Thenthe beam width, W (FWHM), is measured at variousrecorded positions along the optical axis on both sidesof the beam waist. The value of M2 is obtained by fit-ting the measured data with a hyperbola.8 Figure 3shows the measured beam widths and the fittedcurves at injected currents of 7.0 and 11.0 A. The es-timated M2 values are 2.20±0.26 and 2.72±0.28, re-spectively. For clarity, we have shifted the spatial po-sitions of the curves in the figure; in the experimentsthe beam waist is located at almost the same positionat both injected currents. The M2 values along theslow axis of the freely running laser are estimated aswell, with the same method. However, in this casethe beam width is measured in the diagnostic beamand the beam is focused by a cylindrical lens of 40-mm focal length. The M2 values are 236.6±19.9 and292.9±13.6 at operating currents of 7.0 and 11.0 A,respectively. By use of SIL the M2 value of the diodelaser system is improved by a factor of 107 at bothoperating currents.

Spatial mode selection and spatial mode suppres-sion by the SIL technique become increasingly diffi-cult when the emitter width is large. For a gain-guided BAL, emission angle a for a spatial mode oforder n scales approximately with order as an>nl / s2Dd.9,11 Here D is the full width of the emittingarea, i.e., 1000 mm for our diode laser. Thus the emis-sion angle of the same order mode in a 1000-mm-wide

Fig. 2. Far-field profiles measured along the slow axis forthe freely running laser (dotted curves) and the SIL lasersystem (solid curves) at injected current of (a) 7.0 A and (b)11.0 A. For clarity, the data for the freely running laserhave been multiplied by 10.

Fig. 3. Beam width (FWHM) measurement of the outputbeam from the SIL laser system for the slow axis with in-jected currents of 7.0 A (squares) and 11.0 A (circles). Thecurves represent hyperbolic fits to the data. The relativepositions of these curves are offset for clarity.

1148 OPTICS LETTERS / Vol. 30, No. 10 / May 15, 2005

BAL is, e.g., five times smaller than that in a 200-mm-wide BAL (3–4-W output power). Correspond-ingly, the angular distance between the adjacentmodes is 0.024° for the 1000-mm-wide BAL, whereasit is 0.12° for a 200-mm-wide BAL. For our broad di-ode laser the far-field angle in the slow axis is ap-proximately 7°–8°, which is almost the same as thatof a diode with a few-hundred-micrometer emitter.Therefore ,150 modes oscillate in our diode (freelyrunning condition), whereas, for comparison, only,30 modes oscillate in a 200-mm-wide BAL.

Figure 4 shows the intensity profiles of the nearfield for the freely running laser and the SIL lasersystem with an operating current of 11.0 A. The nearfield is measured in the diagnostic beam with thebeam analyzer. A cylindrical lens of 60-mm focallength is used to transform the pseudo-far-field pat-tern into a near-field image. For the freely runninglaser the superposition of the multiple spatial modesproduces a relatively flat near-field distribution.When SIL is applied, the near field becomes slightlyasymmetric, a situation similar to asymmetric SIL ofa BAL with a narrow emitter.8,9 In addition, the near-field modulation increases when SIL is applied. Webelieve that this modulational increase is due to theimproved spatial coherence of the laser; however, themaximum intensity peak at the facet for the feedbacklaser is less than 50% larger than the intensity in thefreely running laser. This is important because theintensity distribution at the facet may lead to cata-strophic damage of the laser.

To investigate how the additional antireflectioncoating deposited on the BAL affects the beam qual-ity, we applied the SIL technique to a similar 1000-mm-wide BAL with a standard antireflection coatingas well.12 Table 1 lists the results of the beam qualitymeasurements in the SIL setup based on the two dif-ferent BALs. For comparison, the M2 values are mea-sured at similar output levels. At an output power of2 W, the additional antireflection coating improvesthe beam quality value by a factor of 4 compared withthat of the standard coating. A special antireflectioncoating can increase the thresholds of laser modesand improve the coupling efficiency of the selectedmode to the active area in the laser chip. Thereby,contributions of unselected modes can easily be sup-pressed. The result indicates that a low-reflection,nonstandard antireflection coating on the front facet

of the diode laser is decisive for obtaining high outputpower and good beam quality with this laser system.This situation is similar to the SIL of a BAL with anarrow emitter.13

In summary, the beam quality of an antireflection-coated 1000-mm-wide BAL is improved significantlyby use of the SIL technique. We believe that this isthe widest single-element BAL used in an external-feedback cavity until now. The far field, the M2 value,and the near field were measured with and withoutSIL at injected currents of 7.0 and 11.0 A. SIL im-proves the beam quality factor M2 by a factor of 107at these currents. An output power of 2.05 W with anM2 value of 2.7 is obtained from the SIL diode lasersystem with the operating current of 11.0 A. An anti-reflection coating on the front facet is important forreducing the M2 value of this very broad diode laser.The results presented in this Letter indicate that it ispossible to obtain high output power and high beamquality by using extraordinarily wide broad-area di-ode lasers.

This research was financed partly by Esko-Graphics, DK-8520 Lystrup, Denmark. M. Chi’se-mail address is mingjun.chi@risoe.dk.References

1. L. Goldberg and M. K. Chun, Appl. Phys. Lett. 53, 1900(1988).

2. S. MacCormack, J. Feinberg, and M. H. Garrett, Opt.Lett. 19, 120 (1994).

3. H. Horiuchi, T. Shimura, T. Omatsu, O. Matoba, and K.Kuroda, Appl. Phys. B 68, 1021 (1999).

4. C. J. Chang-Hasnain, J. Berger, D. R. Scifres, W.Streifer, J. R. Whinnery, and A. Dienes, Appl. Phys.Lett. 50, 1465 (1987).

5. M. Løbel, P. M. Petersen, and P. M. Johansen, Opt.Lett. 23, 825 (1998).

6. S. Juul Jensen, M. Løbel, and P. M. Petersen, Appl.Phys. Lett. 76, 535 (2000).

7. V. Raab and R. Menzel, Opt. Lett. 27, 167 (2002).8. M. Chi, N.-S. Bøgh, B. Thestrup, and P. M. Petersen,

Appl. Phys. Lett. 85, 1107 (2004).9. S. Wolff, A. Rodionov, V. E. Sherstobitov, and H.

Fouckhardt, IEEE J. Quantum Electron. 39, 448(2003).

10. J.-M. Verdiell and R. Frey, IEEE J. Quantum Electron.26, 270 (1990).

11. R. M. R. Pillai and E. M. Garmire, IEEE J. QuantumElectron. 32, 996 (1996).

12. M. Chi, B. Thestrup, and P. M. Petersen, Proc. SPIE5336, 33 (2004).

13. S. Wolff, D. Messerschmidt, and H. Fouckhardt, Opt.Express 5, 32 (1999), http://www.opticsexpress.org.

Fig. 4. Near-field profile measurements in the slow axisfor the freely running laser system (left curve) and the SILlaser system (right curve) at an operating current of 11.0 A.

Table 1. M2 Values and Output Powers of the SILSystem Based on BALs with Different Coatings

Parameter

Type of Coating

StandardsR,6% d

SpecialsR,0.01% d

Current (A) 9.0 11.0M2 value 11.1±3.5 2.72±0.28Power (W) 2.03 2.05

May 15, 2005 / Vol. 30, No. 10 / OPTICS LETTERS 1149