We demonstrate up to 1.6 Watts of single frequency output from a tapered diode laser operating at 785 nm. The tapered diode laser is used in a rear end external cavity set-up where the external feedback element is a combination of a BaTiO3 phase-conjugating crystal and a high reflection mirror. The set-up presented inherently suppresses the self-wavelength scanning effect, which is well known when employing a phase-conjugating crystal as a dynamic wavelength selection component in a laser cavity. The experimental results have been discussed theoretically in some details by employing a two-beam coupling model.
© 2005 Optical Society of America
Phase-conjugation has long been recognized as an attractive method of correcting beam distortion and improving the spectral features of high power lasers. Several implementations have been used over the last two decades. One such implementation is the self-pumped phase-conjugator geometry1–6. This approach has particularly been explored in connection to broad area lasers (BALs)1–3. In the off-axis feedback set-up as much as 490 mW of 1.5 times diffraction limited power has been extracted from a 20-element laser diode array with 54% efficiency. This corresponded to more than 10 times increase in brightness. Furthermore the spectrum was narrowed to less than 0.5 nm. However different problems have been identified where self-wavelength scanning is one of the most remarkable phenomenons2–3. Although methods to circumvent the self-wavelength scanning have been demonstrated2,7, difficulties of simultaneously preventing the wavelength scanning and improving the spatial and spectral brightness properties complicates the set-up of the self-pumped phase-conjugator geometry.
We present here a new laser design based on a tapered diode laser in an external cavity configuration including a mirror/phase-conjugated crystal combination as the rear end reflector of the diode laser. The basic mechanism is that the standing wave in the external cavity induces an interference filter through two-beam coupling in the phase-conjugate medium. The generated filter now acts as a narrow band reflection filter which suppresses secondary lasing modes from reaching the threshold of lasing8–12. We demonstrate here at least 3 orders of magnitude improvement in the spectral brightness using the proposed method as compared to the arrangement using only a mirror as a feedback element. Furthermore the insertion of the wavelength selecting element, i.e. the phase-conjugate crystal does not decrease the output power noticeably, but merely increases the threshold current slightly when compared to a similar system only using a mirror.
In contrast to laser systems using passive frequency selective elements, e.g., a grating the current design lases in a single frequency mode at all levels of the diode drive current13. In addition, the proposed set-up does inherently not show the self-wavelength scanning effect, which normally is present when using phase-conjugating crystals in the laser cavity 2–3. Because the phase conjugate crystal is used as a rear end component and because of the big ratio between the optical powers exiting at the low power rear end and the high power front end of the tapered diode laser, the laser output power from the laser device is not limited by the maximum optical power allowed for the phase-conjugating crystal, but rather by the maximum allowed current for the tapered diode.
The presented laser system can be utilized in many applications, such as Raman spectroscopy, frequency doubling, interferometry and laser anemometry. We expect that, with a suitable choice of the phase-conjugating crystal and tapered diode laser, the presented set-up can easily be applied to make similar laser systems working at other wavelengths.
The tapered, 785 nm, diode in the proposed external cavity set-up is shown in Fig. 1. The tapered diode laser itself consists first of a narrow section with transverse dimensions supporting only a single spatial mode. The single mode section is followed by a tapered section, which acts as an amplifier for the optical output from the first single mode section. The tapered angle of the amplifier section is equal to the diffraction angle of the single mode output beam from the first section. High optical power is emitted from the output facet of the tapered amplifier whereas only the low optical power, allowed by the dimensions of single mode structure, is emitted from the rear end of the single mode section.
Due to the single mode section in the diode, the laser system supports predominantly Gaussian mode operation. The output front facet of the tapered diode constitutes the output mirror of the laser cavity whereas the rear end of the laser diode is anti-reflection coated to allow for external feedback. The optical field emitted from the rear end is focused on a plane mirror using a high numerical lens. The plane mirror acts as a rear end mirror of the external cavity and is located approximately 100 mm behind the lens. The mirror is coated with aluminum having an intensity reflection coefficient of approximately 85%. The phase-conjugating crystal is positioned between the external rear end mirror and the focusing lens. A 6×6×6 mm BaTiO3 crystal from FEE in Germany is used in the experiments reported here. The crystal is aligned with an approximately 60 degree angle between the crystallographic C-axis and the axis of symmetry of the laser. This configuration yields the highest values of the electro-optic coefficient r424. The bulk crystal loss of this particular crystal was unusual high (2.3 cm-1), corresponding to a single pass transmission coefficient of only 18%. We believe that this crystal is doped with an impurity like Rh or Co, leading to a high absorption at 785 nm, and a grayish color of the crystal4. During operation, the tapered diode is actively cooled to a temperature of 20 degree Celsius by means of a peltier-element.
The alignment of the laser is achieved through a two step procedure. The phase-conjugating crystal is left out during the first alignment step where the external rear mirror is aligned for optimum laser power. Subsequently the photorefractive medium is inserted and the rear mirror is realigned in order to compensate for the extra optical path length introduced by the photorefractive material. The above procedure provides a simple alignment scheme of the external cavity design. The output power of the aligned laser assembly as a function of the diode current for both the case when having external mirror feed-back alone and for the complete set-up, including the BaTiO3 crystal is shown in Fig. 2. As can be seen, the slope-efficiency is essentially the same with or without the phase-conjugating crystal inserted. Only the threshold current is about 0.2 Amp higher when the phase-conjugating crystal is employed. The increased threshold current is due to a smaller effective reflection coefficient obtainable from the combined BaTiO3 – mirror reflector as seen from the diode laser rear end. The effective reflection coefficient was measured to be 50% for the BaTiO3 – mirror reflector compared to a reflection coefficient of 85% from the mirror alone. It is interesting to note that the single pass transmission of the BaTiO3 crystal is 18%, which we assume is primary due to electronic bulk absorption in the crystal. We will discuss to this subject further below.
The spectral properties of the laser system are investigated using a Burleigh interferometer (TLT 1500 series) with 3 nm free spectral range. Fig. 3 shows the spectral content of the laser output when running the diode by a drive current of 3 ampere. As can be seen, the laser operates in single frequency mode. The laser line width measurement is limited by the resolution of the instrument. A more detailed measurement shows a line width smaller than 2 pm, but still limited by the resolution of the instrument. The same spectrum as shown in Fig. 3 has been observed within the range from 1.2 ampere to 3 ampere. In fact, we have never seen multimode operation. This attractive property is in favor for the current set-up when compared to systems using passive feed-back elements, e.g., like gratings. In those systems, single mode operation is often restricted to certain current levels and thereby certain output powers13. When only the mirror is used in the external cavity, the laser has a line width in the order of a few nm depending on the precise alignment.
This system has been operated with up to 1.6 Watts of output power in single mode operation, only limited by the maximum allowed drive current for the tapered diode laser. We believe that the present set-up yields the highest single mode output power ever reported from a diode laser using phase-conjugation. Using the BaTiO3 crystal as a part of a rear end external cavity design of a tapered diode laser keeps the incident power to the crystal at a safe level, because the major amplification occurs in tapered output section of the diode. At this point no sign of crystal degradation has been observed in our experiments and tests. In the experiments, the optical power incident on the crystal has always been below 100 mW. Note that the external cavity power levels depend on the specific diode and the power levels may therefore vary from diode to diode.
A very attractive feature of the set-up presented is the possibility of suppressing the self-scanning of the laser wavelength by placing the BaTiO3 close to the rear end mirror. Figure 4 demonstrates this feature. The laser was not temperature stabilized before turning it on and the drift observed in Fig. 4 is believed to be due to the fact that all the optical and mechanical components needs to achieve their steady state temperature before the laser will reach its long term steady state frequency. The highly absorbing BaTiO3 crystal, which was not thermostatised, could in fact be the major source of the observed change. Other tests show that the system approaches a steady state after a few hours of operation. The fluctuating points we attribute primarily to mechanical vibrations originating from a cooling fan nearby and the resolution of the instrument. The external cavity is mounted on separate mounting posts and is not yet integrated with the laser diode assembly. In comparison, typical self-scanning laser systems shows wavelength scans of approximately 3 nm.7 Moving the BaTiO3 crystal to a position halfway between the external rear mirror and the laser diode, causes the laser wavelength to execute the well known wavelength scans. The reason for the suppression of the self-wavelength scanning when situating the crystal close the rear end mirror is not known at present. However, additional experiments suggest that a small change in the alignment of the rear end mirror can lead to self-scanning toward both blue and red wavelengths. So it is not inherently given in which direction the laser will scan in this set-up.
As mentioned above, the effective reflection from the crystal-mirror combination was 50%. With a single pass transmission of the crystal of only 18% it is clear that the mirror can only provide a negligible contribution to the combined reflection coefficient as seen from the diode laser medium because the light should pass the crystal twice. On the other hand, the laser operation stops within 1 to 2 seconds if the cavity is blocked by an obstruction in between the rear mirror and the crystal. During the blocked period the output power from the laser diode decreases from the initial 700 mW (2 ampere drive current) to approximately 80 mW, which corresponds to the output power emitted from the laser diode in the absence of any rear-end feed-back. The estimated light transmitted through the crystal is 0.6 mW in steady state when approximately 60 mW is incident on the crystal from the diode’s side. From an experimental point of view it appears surprising that a beam of only 0.6 mW, that further needs to make a second pass through a crystal with 18% transmission, is absolutely vital for the laser device to function. In order to explain and validate these observations we will in the next section propose a theoretical model based on two-wave (counter propagating) coupling in a phase conjugate medium including loss.
4. Theoretical discussion
In order to discuss the influence of the strongly attenuated optical signal reflected from the rear end mirror, a two-beam coupling model including loss is suggested14,
where the length of the crystal is normalized to a unity length of 1 (See Fig. 5). I 1=I 1 (z) and I 2=I 2 (z) are the forward and backward propagating beam intensities respectively at position z in the crystal. αn is the normalized bulk absorption coefficient of the BaTiO3 crystal. Γ n is the normalized gain coefficient of the BaTiO3 crystal, which is assumed real in this model. The boundary condition is RM I1 (l)=I 2 (l), where RM is the power reflection coefficient of the rear end mirror and l is the unity length of the crystal. Further I 1 is normalized so that I 1 (0)=1. Thereby I 2 (0) equals the effective reflection coefficient from the combined feed-back element and the transmission coefficient through the crystal equals I 1 (1). The normalized absorption coefficient, αn , is measured to be equal to 1.7. This value was measured in a separate experiment. The coefficient, Γ n , is found from fitting the theory to the experiments where the effective reflection coefficient is measured to be 50% and RM=83% respectively. In Fig. 6 the calculated normalized intensities are shown. From a computer simulation, using Eq. (1), we obtain a normalized Γ n of 5.1. Since the crystal has a thickness of 6 mm and is tilted 60 degree, we can calculate the photorefractive gain to a value of 8 cm-1. This is a quite respectable value supporting our hypothesis of two-beam coupling as the underlying process14. From Fig. 6 it is also observed that the major part of the two-beam coupling takes place in only a fraction of the crystal i.e. approx. 30% of the crystal closest to the laser diode. Actually by observing the crystal visually, this appears to be the case, since scattered light can be visually seen from this part of the crystal.
From the model, the transmission coefficient is calculated to be 0.44%. With 60 mW incident on the crystal a transmitted power of 0.26 mW is expected. This number should be compared with a measured value of 0.6 mW. The experimental value and the calculated values are not in full agreement, but we attribute the difference to be due to experimental difficulties. As a last experiment we rotated the crystal 180 degrees to reverse the sign of the electro-optic coefficient. Thereby the flow of energy in the two-beam coupling should change. Indeed, by doing so, the reflection coefficient dropped to a level where only 250 mW of output power could be obtained from the laser system when driven by 2 ampere. The effective reflection coefficient is in this case approximately 20%. This experiment thus supports the two-beam coupling model.
In summary, we have obtained up to 1.6 Watts of output power from a set-up with a tapered diode laser in a rear end external cavity set-up where the feed-back element is a combination of a BaTiO3 phase-conjugating crystal and a high reflection mirror. Only a slight increase in the threshold current was observed when comparing the proposed design with a system only employing mirror feed-back. It was verified experimentally that the laser is lasing in a single frequency mode with a line width of less than 2 pm, where the measurement is limited by bandwidth of the Fabery-Perot interferometer. Single frequency operation was observed at all power levels of the laser. By locating the crystal near to the rear end mirror, the otherwise well known self-wavelength scanning effect could be suppressed. The experimental results have been explained in reasonable details by a two-beam coupling model including loses in the crystal.
We would like to acknowledge Torsana Laser Systems A/S for using their BaTiO3 crystals.
References and links
2. S. J. Jensen, M. Løbel, and Paul M. Petersen, “Stability of the sibgle-mode output of a laser diode array with phase conjugate feedback,” Appl. Phys. Lett. , 76, 535–537 (2000). [CrossRef]
3. S. Weiss, M. Segev, and B. Fisher, “Line narrowing and self frequency scanning of laser diode arrays coupled to a photorefractive oscillator,” IEEE J. Quantum Electron. 24, 706–708 (1988). [CrossRef]
4. G.W. Ross, P. Hribrik, R.W. Eason, M.H. Garrett, and D. Rytz, “Impurity enhanged self-pumped phase conjugation in the near infrared in “blue” BaTiO3,” Opt. Comm. 101, 60–64 (1993). [CrossRef]
5. Y. Zhu, D. Zhang, X. Niu, and X. Wu, “High efficiency self-pumped phase conjugation from 45°-cut Ce:BaTiO3,” Opt. Commun. 134, 205–210 (1997) [CrossRef]
6. N. Hout, J.M.C. Jonathan, G. Roosen, and D. Rytz, “Characterization and optimization of a ring self-pumped phase-conjugate mirror at 1.06µm with BaTiO3:Rh,” J. Opt. Am. B, 15 (1998).
7. M. Løbel, P.M. Petersen, and P.M. Johansen, “Suppressing self-induced frequency scanning of a phase conjugate diode laser array with using counterbalance dispersion,” Appl. Phys. Lett. 72, 1263–1265 (1998). [CrossRef]
9. L. Meilhac, N. Dubreuil, G. Pauliat, and G. Roosen, “Modeling of laser mode self-adapted filtering by photorefractive Fabry-Perot interferometers,” Opt. Mat. 18, 37–40 (2001). [CrossRef]
10. A. Shiratori and M. Obara, “Wavelength-stable, narrow-spectral-width oscillation of an AlGaInP diode laser coupled to a BaTiO3:Co stimulated photorefractive backscattering phase conjugator,” Appl. Phys. B , 15, 329–333 (1997). [CrossRef]
11. N. Hout, J.M. Jonathan, G. Pauliat, P. Georges, and G. Roosen, “Laser mode minipulation by intracavity dynamic holography: Application to mode selection,” Appl. Phys. B 69, 155–157(1999). [CrossRef]
12. A. Godard, G. Pauliat, G. Roosen, P. Graindorge, and P. Martin, “Relaxation of the aligment tolerance of a 1.55µm extended-cavity semiconductor laser by use of an intracavity photorefractive filter,” Opt. Lett. 26, 1955–1957 (2001). [CrossRef]
13. A. Godard, G. Pauliat, G. Roosen, P. Graindorge, and P. Martin, “Single-mode gain in grating-tuned extended-cavity semiconductor lasers: Investigation of stable single-mode opration conditions,” IEEE J. Quantum Electron.38, NO. 4, (2002).
14. Pochi Yeh, Introduction to Photorefractive Nonlinear Optics, (John Wiley & Sons Inc., 1993)