Diode-pumped CW molecular lasers

Abstract

First continuous laser oscillation on many lines in the range of 533–635 nm on different transitions of $\hbox {Na}_{2}$ and $\hbox {Te}_{2}$ molecules has been obtained, optically pumped with common cw blue emitting InGaN diode lasers operating around 445 and 460 nm. Spectral narrowing of the diode laser is achieved with a beamsplitter and grating setup, allowing use of more than 50 % of the diode power. Operation conditions and properties of the laser systems are presented, and perspectives for the realization of compact low cost molecular lasers are discussed.

1 Introduction

In the past, optical pumping of diatomic molecules (also referred to as dimers) has been investigated intensively for laser purposes, and for a variety of molecules as $\hbox {Li}_{2}$, $\hbox {Na}_{2}$, $\hbox {K}_{2}$, $\hbox {Se}_{2}$, $\hbox {Bi}_2 \hbox {Te}_{2}$ and $\hbox {I}_{2}$ continuous laser oscillation on a great number of lines in the visible and near infrared spectral range has been obtained (see refs. [1–4] and further refs. therein). So far these continuously operating molecular laser systems mostly have been pumped with Argon ion or Krypton ion lasers, although extremely low thresholds of less than 1 mW [1] also allow pumping with a low power He–Ne laser [5]. The development of new powerful laser diodes emitting in the violet (405 nm) or blue spectral range (445–460 nm), respectively, now opens new possibilities for optical pumping and spectroscopic investigations and the operation of low cost and compact lasers with some of these molecules. Cw operation on atomic alkali transitions pumped with cw near infrared diodes has been published [6, 7], but, to the best of our knowledge, cw operation of diatomic molecules pumped with blue emitting diodes has not been reported so far. In this contribution, we report on the first realization of cw laser oscillation on many lines at $B^1\varPi _u \rightarrow X^1\varSigma ^+_g$ transitions of $\hbox {Na}_{2}$, pumped at wavelength around 460 nm, and on $BO^+_u \rightarrow XO^+_g$ transitions of $\hbox {Te} _{2}$, pumped around 445 and 460 nm, respectively.

2 Operation of the diode pump lasers

For the experiments standard InGaN diode lasers at 445 and 460 nm (Nichia) with output powers up to about 1 W were used. The diodes are mounted on an actively cooled metal heat sink with a Peltier element and operated with a stabilized power supply. With an anamorphic prism pair a near $\hbox {TEM}_{00}$ mode is achieved. The spectral emission of such diodes is broadband with several nm width, and the central wavelength depends on temperature and current and shifts to longer wavelengths with increasing current. In detail, the emission spectrum consists of hundreds of modes, whose spacing is determined by the optical length of the diode cavity. Figure 1 shows such a broadband emission spectrum for the 460-nm diode, taken with a Fourier spectrometer [8]. The total emission width is almost 3 nm (at a diode current of 1000 mA, output power 600 mW; all laser powers measured with power meter Coherent LabMax), and the substructure shows cavity modes with about 43 GHz ($\sim$0.03 nm) mode spacing, corresponding to a diode cavity length of about 1.4 mm. Off course, such a broadband multimode laser is not at all well suited for optical pumping of narrow atomic or molecular transitions. For example, the Doppler width of rotational vibrational transitions between electronic states of $\hbox {Na} _{2}$ molecules is about 1.5 GHz and similar for atomic transitions. Consequently, optimum optical pumping requires tunable single frequency operation of the diode lasers. This can be achieved by applying external frequency selective feedback cavities, mostly using highly dispersive gratings in Littrow or Littman-Metcalf configuration, often in combination with anti-reflection coated diodes to reduce the influence of the diode cavity. In addition, careful stabilization and control of the feedback and diode cavity modes are necessary. For a detailed discussion and description see Refs. [9–12] and further refs. therein. Nowadays, narrowband tunable diode lasers with output powers in the mW to tens of mW range are commercially available at many wavelengths, but these systems are expensive and, to our knowledge, narrowband single frequency operation has not been demonstrated with the high-power diode types used here.

Fig. 1

Emission spectrum of the diode laser (460 nm) taken with the Fourier spectrometer described in Ref. [8]. Diode power 600 mW at a current of 1000 mA, mode spacing about 0.03 nm (43 GHz)

It is the purpose of our present experiments to demonstrate the potential of pumping molecules as $\hbox {Na}_{2}$ and $\hbox {Te}_{2}$ with blue emitting diode lasers, but not to achieve single frequency operation. In fact, the dense and broad emission spectrum of the considered diodes on one side and the dense absorption spectrum of the considered molecules on the other side guarantees sufficient overlap for optical excitation of many transitions. Indeed, when pumping the molecules with the broadband emitting diodes, intense fluorescence spectra are easily obtained (similar to the spectra shown in Figs. 6, 9). Moreover, the spectral intensity of a specific mode of the broadband diode emission is even sufficient to pump a specific electronic molecular transition and generate laser emission. This has been observed for $\hbox {Na}_2$ pumped with the 460-nm broadband diode laser (see below). Nevertheless, a certain degree of spectral narrowing and tuning of the diode lasers is necessary to really explore the potential of diode pumping of molecules. Our experimental approach for this is given in Fig. 2. The output of the diode is split by a beamsplitter, with the feedback grating in Littrow configuration in one arm (similar as mentioned in [9]). In this way, the pump beam direction is not affected when adjusting and tuning of the grating and a higher usable power is possible as compared to output coupling via the first order of the grating. Typically, beam splitters with 40–60 % reflection are used. Also use of a prism (flint glass) with mirror feedback is possible, but the selectivity is slightly less. Therefore, in most experiments grating feedback was used. To control the feedback adjustment, the reflected beam from the grating is focused and sent divergently onto a solid Fabry–Perot (FP) plate of 0.3 mm thickness and about 80 % reflection coating on both sides. Feedback is indicated by the appearance of interference fringes (see the fringe pattern inserted in Fig. 2). Optimum feedback is especially adjusted by tilting the grating perpendicular to the x–y plane, while tilting the grating in the x–y plane mainly shifts the diode wavelength, which results in a moving of the interference fringes. For the given FP, a moving of one fringe corresponds to a diode wavelength shift of about 0.24 nm. The diode spectrum itself is taken with an optical fiber in combination with an optical spectrum analyzer (Ocean Optics USB-650), which has a resolution of about 2 nm. In addition, the spectral width of the diode emission can be measured with a variable distance plane-parallel scanning Fabry–Perot (reflection of the FP plates about 95 %; distance between 0.05 and 1 mm). To avoid feedback of both Fabry–Perot analyzers into the diode, apertures and slight tilts are used. The effect of the grating feedback for the 460-nm diode is demonstrated in Fig. 3, measured again with the Fourier spectrometer as mentioned above (Fig.1, Ref. [8]). A narrowband emission is seen, tunable in the vicinity of the still present residual weak broadband emission of the diode without feedback (compared with Fig. 1). The half-widths (FWHM) of the narrow emissions ranges from 100–160 GHz, which would allow 1–4 diode cavity modes within. Ripples seen within the narrow emissions may result from these cavity modes and in addition from modes of the coupled resonator (about 3 GHz for the optical length of the feedback cavity of about 5 cm), but due to fluctuations these modes can not be resolved by the Fourier spectrometer, which requires more or less stable frequency conditions for the measurement. The Fourier spectrometer allows absolute high resolution wavelength measurements (and has been used due to availability), but for routine and continuous observation of the diode emission at molecular fluorescence and laser experiments planned here, this spectrometer is not handy and too difficult to use. Therefore, for all further investigations the scanning FP (SFP) and the optical spectrum analyzer were used.

Fig. 2

Scheme of the diode feedback arrangement and measuring system. OSA optical spectrum analyzer, SFP scanning Fabry–Perot, OSCI oscilloscope

Fig. 3

Emission spectrum of the diode laser (460 nm) taken with the Fourier spectrometer for 3 different settings of the feedback grating. Diode power about 350 mW behind beamsplitter (R=60 %)

Figure 4 shows as an example the signals of the SFP (for a mirror separation of 0.06 mm; free spectral range 2500 GHz, corresponding to about 1.76 nm) for operation of the 460-nm diode without feedback and with feedback. The SFP signals clearly show the narrowing of the spectral emission. The trace (A) in Fig. 4 indicates an emission width (FHWM) of about 160 GHz (about 0.1 nm), which is comparable to the data obtained with the Fourier spectrometer (Fig. 3). The scope trace without the feedback yields an emission width of about half a free spectral range (0.9 nm). The spectrometer can not resolve the narrow emission (resolution 2 nm) and only shows an increase in the signal with feedback and, depending on the grating adjustment, a wavelength shift. Within the narrow emission, diode cavity modes with a spacing of 43 GHz can clearly be seen. An individual mode will pump a narrow molecular transition and generate laser emission. However, the diode modes strongly shift with the diode temperature, typically at a rate of 70 GHz (0.05 nm)/$^\circ$C. Consequently, stabilization of the diode temperature to better than $0.01^\circ$ C is necessary to keep an individual mode within the molecular absorption line. In summary, the described setup (Fig. 2) offers a simple and low cost way to generate (and measure) high-power tunable narrowband (not single frequency at present) diode laser emission and proved well suited for the fluorescence and laser investigations at $\hbox {Na} _{2}$ and $\hbox {Te} _{2}$ molecules described in the following.

Fig. 4

Oscilloscope traces of the SFP signals (at 0.06 mm mirror spacing) for a with grating feedback and b without feedback for the 460-nm diode at a power of 360 mW

3 Setup for fluorescence and laser experiments

For experiments with $\hbox {Na} _{2}$, Na vapor has to be generated and handled at higher temperatures, which can well be done in a heatpipe, as described in Refs. [13] or [1]. Our experimental setup for spectroscopic and laser experiments is shown in Fig. 5. The heatpipe consists of a stainless steel tube of about 350 mm length and 20 mm diameter with Brewster angle windows. The length of the vapor zone is about 100 mm, defined mainly by the heating element. As buffer gas Argon is used. At typical operating temperatures of 530 $^{\circ } \hbox {C}$, the total vapor pressure (Na + Na₂) is about 10 mbar, with a dimer partial pressure of about 0.4 mbar [14]. The diode pump radiation is focused into the center of the heatpipe by a lens (mirror M1), and fluorescence from the vapor is coupled into a fiber and sent into the optical spectrum analyzer. If necessary, a cutoff filter is used to remove too strong pump radiation. To fully avoid direct pump radiation getting into the fiber, the fiber can also be placed into an off-axis position A or B. For straightforward laser experiments, a concentric resonator is applied, consisting of two mirrors with radii of 250 mm at a distance slightly below 500 mm. The mirror M1 also serves as focusing lens in such a way that the radius of curvature equals the focusing length of 250 mm. Both mirrors have high reflectivity for the expected laser range (500–650 nm for $\hbox {Na} _2$ and $\hbox {Te} _2$, see below), while mirror M1 in addition has high transmission and mirror M2 high reflection for the pump radiation. The concentric resonator is well suited for first laser experiments, but has the disadvantage that pump radiation is coupled back into the diode and thus influences the feedback. To avoid this, the ring resonator scheme is used, with input coupling of the pump radiation at the plane mirror M3 (high transmission for the pump radiation; high reflection 500–650 nm) and focusing into the heatpipe by mirror M4. Mirrors 4–6 have high reflection for the range 440–650 nm. The mirrors M4 and M5 have radii of 500 mm and are placed at a distance slightly above 500 mm. In case of Te, the setup is almost identical, with the heatpipe replaced by a sealed-off cell with Brewster angle windows, located in an oven, and a change of the radii of the curved mirrors M4 and M5 (see further details below).

Fig. 5

Setup for fluorescence and laser experiments with a heatpipe for the generation of Na/Na₂ vapor or a cell for Te/Te₂ vapor and linear or ring resonator configurations. Diode pump laser setup as given in Fig. 2 (BFT: birefringent tuner)

4 Results for $\hbox {Na} _{2}$

Upon focusing of the pump radiation into the heatpipe, intense green fluorescence can easily be seen, starting already at temperatures around 300 $^{\circ } \hbox {C}$ and having its maximum in the range of 530–560 $^{\circ } \hbox {C}$. At higher temperatures the pump radiation is strongly absorbed. As an example Fig. 6 shows the fluorescence spectrum of $\hbox {Na}_2$ pumped with the diode laser at 460 nm (with feedback, reflection of beamsplitter 60 %), with the optical fiber in the position B (see Fig. 5) to avoid direct pump light. A clear vibrational fluorescence progression can be seen, peaking at 550 nm. As the progression shows no emission on the blue side of the 460 nm line (no anti-stokes line), we assume a pump transition starting from vibrational level $v''=0$ of the $X^1\varSigma ^+_g$ ground state. Indeed, the molecular data of $\hbox {Na}_2$ given by Kusch and Hessel [15] suggest a transition from the electronic ground state $X^1\varSigma ^+_g (v''=0)$ to the $v'=13$ state of the excited electronic state $B^1\varPi ^+_u$, with subsequent fluorescence lines $B^1\varPi ^+_u (v'=13)\longrightarrow X^1\varSigma ^+_g(v''= 0,1,2....n)$. Figure 7 shows the corresponding $\hbox {Na} _2$ molecular level scheme with indicated pump and fluorescence transitions up to v”=30. However, as the accuracy of the spectrometer is only about 1 nm, an exact identification of the pump and fluorescence transitions is not possible at present, and within the emission line of the diode further nearby pump transitions may contribute as well. Fig. 7 demonstrates the good match of blue diode laser emission wavelengths for optical pumping of the $\hbox {Na} _2$ X-B band. Before, $\hbox {Na}_2$ X-B transitions have been pumped by Argon laser lines at 488.0, 476.5 and 472.7 nm [1]. Laser oscillation was first tested and achieved with the concentric resonator (Fig. 5) at optimum heatpipe temperatures around 530$^{\circ } \hbox {C}$ and typical buffer gas pressures of 8 mbar to 10 mbar. For broadband excitation (no feedback), laser oscillation was achieved at a single line only (554 nm, see Table 1) at a pump power around 500 mW (measured in front of the input mirror M1) while with feedback much lower pump powers were sufficient to achieve simultaneous oscillation on many lines. Figure 8 shows examples of multiline laser oscillation for different settings (pump wavelengths) of the grating feedback. The spectra of the laser emissions strongly fluctuate, which is due to fluctuations of the pump spectrum (diode and feedback cavity modes) and also of the laser cavity modes. Moreover, fluctuations are due to back-coupling from the resonator mirrors. This back-coupling can be avoided by use of the ring resonator setup of Fig 5. Oscillation with the ring resonator was as easy to achieve as for the linear resonator and shows the characteristic feature of these molecular lasers, namely unidirectional oscillation (in direction of the pump), which is due to the inherent Raman gain contribution [1]. By optimization of the ring resonator (overlap between pump and laser beam, separation between mirror M4–M5 53 cm, total length of resonator 166 cm) and especially by setting of the pump frequency, thresholds for ring laser oscillation of less than 3 mW (behind the input mirror M3) were achieved, which agrees with threshold data obtained for pumping of $\hbox {Na}_2$ with Argon laser lines [1]. At these power levels the diode itself operates almost at threshold (current about 70 mA) and diode oscillation only occurs with feedback. The ring resonator setup is also well suited to insert optical elements into the almost parallel beam between the plane mirrors M3 and M6 (Fig. 5). So with a single stage birefringent tuner tunable oscillation on single lines was possible. Investigations on the output power for individual lines have not been performed so far. However, with an output coupling mirror of 50 % (mirror M6), indicating the high gain of the laser, an output power of 4 mW (multiline at about 300 mW pump power behind mirror M3) has been measured. Table 1 summarizes so far obtained pump and laser lines. The identification of the pump transitions is preliminary due to the resolution of the spectrometer and does not allow an identification of rotational levels. With the molecular data provided by [15], the vibrational transitions can be identified in good agreement.

Fig. 6

$\hbox {Na} _{2}$ Fluorescence spectrum optically pumped with the spectrally narrowed 460 nm diode laser at a pump power of about 300 mW. The vibrational structure is not fully resolved. Indicated is a calculated number v”=30 for a transition at 562 nm (see also Fig. 7)

Fig. 7

$\hbox {Na} _{2}$ energy level scheme (after Ref. [15]) with 460 nm diode pump laser transition and possible fluorescence and laser lines

Fig. 8

Simultaneously oscillating $\hbox {Na}_{2}$ laser lines, pump wavelengths 460, 462 and 464 nm, linear resonator, pump power about 300 mW; spectra measured behind mirror M2

Table 1 Suggested $\hbox {Na}_2$ $X^1\varSigma ^+_g(v'')\rightarrow B^1\varPi _u(v')$ pump transition and observed laser lines

5 Results for $\hbox {Te} _{2}$

For experiments with $\hbox {Te}_2$, the experimental setup of Fig. 5 is used, with the heatpipe replaced by a sealed-off cell of rectangular size (13 mm $\times$ 13 mm) with Brewster angle windows and a length of about 85 mm. The cell has a side arm which contains a small amount of metal Te (no buffer gas) and is placed in an oven with two chambers, one for the side arm (reservoir) and one for the Brewster cell part. The temperature T1 in the side arm defines the Te vapor pressure, while the temperature T2 of the cell itself is kept about 50–70 $^{\circ } \hbox {C}$ higher, in order to prevent condensation of the vapor at the cell windows. At a typical operation temperature of the reservoir of about 530$^{\circ } \hbox {C}$, the Te vapor pressure is about 7 mbar [14] and fully consists of $\hbox {Te}_2$ molecules. As the cell is shorter than the heatpipe, the curved mirrors M1 and M2 for the linear resonator are replaced by mirrors with radii of 148 mm, allowing a bit stronger focusing (focal length of M1 is 148 mm) and more compact resonators. In case of the ring resonator mirrors M4 and M5 have radii of 353 mm and are placed at a distance of about 370 mm (total resonator length about 1260 mm). Observation of fluorescence without a filter or influence of the pump radiation is possible through a small side window of the oven chamber for the Brewster cell. Figure 9 shows fluorescence spectra of $\hbox {Te}_2$ optically pumped with a laser diode at 443, 445 nm and with the 461 nm diode laser used for the $\hbox {Na} _2$ experiments. For the 443/445 nm laser diode a beamsplitter with about 40 % reflection was used. The fluorescence progressions extend from about 420 nm to almost 650 nm. In case of the 461-nm pump (Fig. 9) 6 vibrational lines on the anti-stokes side of the pump are clearly recognizable, indicating pumping from a vibrational level $v''=7$ of the ground state, while in case of the 443- and 445-nm pump the starting levels seem to be $v''=6$ and $v''=9$ (Fig. 9). Figure 10 shows the corresponding molecular level scheme with indicated pump and fluorescence transitions between the $XO^+_g$ ground and the $BO^+_u$ excited electronic state. According to the present spectral resolution, the 445-nm pump line fits well a $XO^+_g (v''=6)$ to $BO^+_u (v'=11)$ transition, the 443 nm a transition $XO^+_g (v''=8)$ to $BO^+_u (v'=14)$ and the 460-nm line a $XO^+_g (v''=7)$ to $BO^+_u (v'=7)$ transition. In the past, $\hbox {Te} _2$ has been pumped with the 476.5-nm and 457.9-nm lines of the Argon ion laser and the 406.7-nm line of the Krypton ion laser [3, 4].

Fig. 9

$\hbox {Te} _{2}$ Fluorescence spectra at various pump wavelengths. Reservoir 530 $^{\circ } \hbox {C}$, cell 620 $^{\circ } \hbox {C}$, pump power about 300 mW

Fig. 10

Energy level diagram of $\hbox {Te} _{2}$ (after [16]), with some pump and laser transitions

Here, laser oscillation was achieved with the linear and the ring resonator (unidirectional oscillation) as well, on many lines in the range of 538–635 nm (Table 2). Figure 11 shows a multiline laser oscillation spectrum for the 445 nm pump diode. Similar as for $\hbox {Na}_2$, pump thresholds below 10 mW (behind the plane input mirror M3) for the ring resonator have been obtained, again close to the threshold of diode laser oscillation (about 70 mA). Table 2 summarizes so far identified pump transitions and observed laser lines. Again the identification is preliminary as discussed above for $\hbox {Na}_2$. Similar as for $\hbox {Na}_2$ single line operation in the ring resonator by use of a birefringent tuner is possible but investigations on output power and output coupling have not been performed so far. Compared to the $\hbox {Na} _2$ system, the $\hbox {Te} _2$ system is much easier to operate, as the sealed-off cell allows reproducible adjustment of operation conditions, whereas the heatpipe conditions will always slightly change. Moreover, the $\hbox {Te} _{2}$ spectrum covers a broader range, with potential for laser lines from blue to red. The Te used here for the cell is natural Te, which consists of a number of isotopes [17]. Consequently, different isotopes will be excited and contribute to the spectra, which however is not resolved here. Favorable clean spectroscopic conditions are possible by use of the available $^{130}$Te isotope. This has been demonstrated in optical pump experiments with the 476.5- and 457.9-nm lines of the Argon ion laser [3, 4]. Moreover, it may be of interest to pump $\hbox {Te} _{2}$ with violet emitting laser diodes around 407 nm. In this case pumping from the $v''=0$ vibrational ground state will be possible, as it has been demonstrated by pumping with the 406.7 nm line of the Krypton ion laser [4]. In some first pump experiments with a broadband 405-nm laser diode a clear $\hbox {Te} _{2}$ fluorescence progression was obtained, starting from vibrational level v”= 0 of the electronic ground state.

Fig. 11

$\hbox {Te} _{2}$ Fluorescence spectrum and simultaneously oscillating laser lines pumped with 443 nm diode laser. Measured behind mirror M6 of the ring resonator. Pump power about 300 mW

Table 2 Suggested $\hbox {Te} _{2}$ $XO^+_g (v'') \rightarrow BO^+_u (v')$ pump transitions and observed laser lines

6 Conclusions

The performed experiments clearly show that blue emitting diode lasers can be well applied for optical pumping of molecules as $\hbox {Na} _{2}$ and $\hbox {Te} _{2}$ and operation of lasers at many lines. The diode emission has to be spectrally narrowed for reliable optical pumping, which can easily be done by feedback techniques. However, optimum excitation of individual transitions and stable laser operation requires stable tunable single frequency diode lasers. Techniques for this are known, and corresponding commercial systems are available. With such pump lasers output powers on single molecular lines in the mW range are expected. Due to the Raman features of these molecular lasers, the individual laser frequencies will follow the pump laser frequency. Consequently, when pumping $\hbox {Te} _{2}$ with a stabilized blue emitting single frequency diode laser in combination with a stabilized ring resonator a set of stabilized lines covering the blue to red spectral range will be possible, as has been demonstrated for the yellow to infrared range in case of $I_{2}$, pumped with a stable frequency doubled Nd:YAG laser system [18]. Such systems will be of interest for molecular spectroscopy and precision experiments. Less ambitious lower cost systems will be possible for demonstration and educational purposes in molecular physics. Here a system with $\hbox {I}_{2}$ molecules, which can be handled in a sealed-off cell at room temperature, is especially simple, and a ring laser system has recently been realized by us, pumped with a modified 40-mW green laser pointer at 532 nm [19].