1. Introduction

In recent decades, diode-pumped alkali lasers (DPALs) have been intensively studied to fully exploit alkali atoms’ inherently high quantum efficiency with technological advancement in the narrow-linewidth high power pump laser diodes [1–3]. DPALs have shown a significant potential for high-power scalability circumventing nonlinear optic limitations in other solid-state lasers whilst maintaining a good beam quality [4–6]. Various pumping schemes have been realized in DPALs to further multiplex pump lasers in a high-power single laser cavity [7,8]. Parallel to DPAL, diode-pumped alkali amplifiers (DPAAs) are also being investigated to increase the DPAL output power in a master oscillator power amplifier (MOPA) configuration [9–11]. Recently fiber optic delivery of both signal and pump has been attempted in a Cs DPAA and showed a high gain in the small-signal regime [12].

Intensive studies on amplified spontaneous emission (ASE) have followed in prior optical gain media after demonstrating lasers and amplifiers. For instance, ASE light sources have been investigated in compound semiconductors [13,14]. and rare-earth-doped optical fibers [15–17] after realization of their lasers and optical amplifiers. These ASE sources have provided a low coherence, a broad spectral bandwidth, and a high pump-to-ASE conversion efficiency [18] which found various applications such as low coherence interferometry [19] low-coherence tomography [20] and optical sensing [21] to name a few. However, in the case of alkali vapor gain media, ASE studies have been very scarce despite their successful demonstration of lasers and amplifiers. ASE from alkali atom vapors can bridge the spectral gap of prior ASE light sources and find practical applications in quantum computing based on alkali atomic qubits [22] ASE from alkali atoms has not been thoroughly investigated in experiments, and only numerical analyses have been available [23]. In this study, we experimentally reported ASE from a Cs atom vapor-cell in terms of its ∼4.7 GHz spectral width and a high output conversion efficiency over ∼30% in both forward and backward directions, for the first time to the authors’ best knowledge. Using the same vapor cell, we injected a 1mW signal to configure a single-pass optical amplifier, and we monitored the forward output both in temporal and spectral domains. Both the signal and laser were delivered through a single mode fiber to maximize the overlap in their intensity distributions along the Cs atom vapor cell. We found the signal laser efficiently suppressed the ASE, and we obtained a significant amplification factor over 700 at the pump power of 1.2 W, which is the largest value ever reported in DPAA experiments.

2. Measurements of ASE from a Cs atom vapor cell

2.1 Experimental set-up

The experimental setup to measure ASE is schematically shown in Fig. 1. We used a continuous-wave pump laser at the D2 transition of the Cs atom near λ=852 nm with a bandwidth of ∼10 kHz (Toptica, TA pro). The pump laser was coupled to a single mode fiber (SMF) which worked as a spatial filter making the laser output in the fundamental LP₀₁ mode. We used a collimator to make the SMF output a Gaussian-like beam with a diameter of ∼3.2 mm. The beam was further focused to the center of the Cs vapor cell using a single lens with a focal length of 100 mm. A cylindrical quartz cell filled with the Cs atom vapor and 500 torr ethane buffer gas was used as a gain medium. The cell had an axial length of 2 cm and a diameter of 2.5 cm. The windows on both facets of the cell were anti-reflection coated to suppress any optical feedback. The window showed a high transmission at both the pump at λ=852 nm and ASE near λ=894 nm. The cell temperature, T, was controlled by an electric heater from 110 to 150 °C. It is well-known that two distinctive features characterized ASE: 1) a nonlinear growth of its power without a pump power threshold, and 2) a spectral narrowing for increasing pump power [24–28].

 

Fig. 1. Schematic diagram of the experimental set-up for the output power and spectral characteristics of ASE in a Cs vapor cell and the energy diagram of Cs atom. Both facets of the cell were anti-reflection coated to prevent optical feedback. (SMF: single-mode fiber, DM: dichroic mirror, BPF: bandpass filter, BS: beam splitter, FPI: Fabry- Perot interferometer).

Download Full Size | PDF

Using the experimental setup in Fig. 1, we measured the optical power of ASE using a powermeter and its spectrum by a scanning Fabry-Perot Interferometer (FPI) to confirm the above two characteristics. In the forward ASE measurement, a bandpass filter (BPF) with the peak transmission at λ=895.4 nm and a full-width at half-maximum of 20.3 nm was used to block the pump laser at λ=852 nm and transmit the ASE at λ= 894 nm. The backward ASE was monitored using a dichroic mirror (DM) that has a high reflection of 99% only for the ASE. The forward ASE was further characterized using a beam splitter (1:99) and an FPI (Thorlabs SA210-8B) in the spectral domain. The FPI had a free spectral range (FSR) of 10 GHz, a minimum finesse of 150, and a spectral resolution of 67 MHz.

2.2 Nonlinear increase of the ASE power

Using the setup shown in Fig. 1, we measured the output power and spectral characteristics of ASE from the heated Cs cell. We focused on the ASE parallel to the pump laser propagation in the forward and backward directions as in prior ASE reports [29–31]. The spontaneous emission in an optical gain medium is randomly directed. However, if the pumping of the gain medium is strong enough, the spontaneous emission is amplified via the “stimulated” emission process. The stimulated emission has a well-defined orientation, and in the case of a longitudinally pumped gain medium, only the spontaneous emission parallel to the pump direction is amplified to result in both forward and backward ASEs [28–30].

Figure 2(a) shows a plot of the ASE output power as a function of the Cs cell temperature when the pump power was set to 1.0 W. The ASE power in both directions rapidly increased in the T range of 120∼130 °C and saturated near T = 150 °C. At T = 150 °C, we found the maximum ASE power of 300 and 350 mW in the forward and backward directions, respectively. For T< 130 °C, the forward ASE output was larger than that of the backward ASE. In contrast, for T> 140 °C, the backward ASE output power was more significantly larger. Note that the focal point of the pump laser was fixed near the middle of the cell, and the above-mentioned behavior is not related to the pump beam focusing. It is well-known that the cell temperature directly affects the Cs atom density and, subsequently, the absorption of the pump laser [31]. In the lower T (T< 130 °C) with a less Cs atom density, the pump laser can reach further in the forward direction to contribute more in the forward ASE [32]. On the other hand, in the higher T (T> 140 °C), the Cs atom density increases exponentially, and the pump laser power decreases by the absorption of Cs atoms to contribute a higher level of population inversion in the backward direction. In Fig. 2(b), we plotted the ASE output power as a function of the pump power at T = 150 °C. In both directions, the ASE power increased in a nonlinear manner without a clear pump threshold. These behaviors are distinctive from DPAL cavities [33–37], where the clear signature of pump thresholds have been reported along with characteristic linear slopes.

 

Fig. 2. (a) ASE output power as a function of Cs vapor cell temperature T with a pump laser power of 1.0 W, (b) ASE output power as a function of the pump power at T = 150 °C.

Download Full Size | PDF

2.3 Spectral narrowing of the ASE

We then investigated the spectral narrowing characteristics of ASE, and the results are summarized in Fig. 3. At a low pump power of ∼ 50 mW, more than 20 modes were distributed over a bandwidth of ∼9 GHz, as shown in Fig. 3(a). The peak voltage of each mode randomly varied with time, and Fig. 3(a) shows one of the snapshots representing the spectral width of ASE. As the pump power increased to 1.0 Watt, the spectral distribution of ASE conspicuously changed, as shown in Fig. 3(b). The number of modes reduced to 5∼6, and they were within a spectral width of ∼4.7 GHz, which was about half of Fig. 3(a). In both cases, the spectral bandwidth of the individual modes comprising ASE was narrower than the spectral resolution of 67 MHz of our FPI. Experimental results in Fig. 2 and Fig. 3 confirmed ASE from a Cs atom vapor, which provided a linewidth of ∼4.7 GHz and an output power of ∼300mW in both forward and backward directions.

 

Fig. 3. ASE in the frequency domain measured by a scanning Fabry-Perot interferometer with a free spectral range (FSR) of 10 GHz. Red line is the piezoelectric drive voltage and detected ASE is shown in black. The pump power was (a) 50 mW, and (b) 1.0 W.

Download Full Size | PDF

2.4 Diode-pumped alkali amplifier (DPAA)

As a next step, we built a Cs DPAA where a signal at λ_s= 894 nm was combined with the pump laser at λ_p= 852 nm using a polarization beam splitter, as shown in Fig. 4(a). Here we used a laser diode as the signal with an output power of 1 mW and a spectral width of 0.5 MHz. We coupled the pump and signal lasers simultaneously into the SMF where both were guided in the fundamental LP₀₁ mode, to realize an optimal spatial overlap between them similar to a prior report [12]. The laser beams were then focused at the center of the Cs cell using a collimator and a lens similar to the setup in Fig. 1. Using a BPF the pump beam was blocked and only the amplified signal, and ASE were monitored. The signal was modulated using a mechanical chopper at a frequency of 50 Hz, and the photodetector (PD) detected both the amplified signal and ASE [38]. Figure 4(b) shows the temporal response of the DPAA. Here we fixed the pump power at 1.2 W. Both signal pulses and ASE background were measured for various Cs cell temperatures. It is noted that the ASE background monotonically increased as T increased from 110 to 150 °C, which is consistent with Fig. 2(a). However, the amplified signal reached the largest peak at T= 140 °C (the pink curve) and decreased at an elevated temperature T = 150 °C (the green curve). We measured the FPI spectra of the amplified signal at T = 140 °C for the pump power of 1.2 W. The results are shown in Fig. 4(c). In contrast to the sporadic multi-line ASE spectra in Fig. 3, the amplified signal showed a single stable peak within the FSR of 10 GHz with a signal-to-noise ratio better than 30:1. We varied the signal input power in a range from 0.2 to 1.0 mW, and we did not observe traces of ASE. In a recent simulation of a DPAA transversely pumped by multiple pump lasers, it has been reported that a high-intensity signal suppresses ASE sufficiently low due to a gain saturation of the narrow spectral band of Cs atoms [23]. In our longitudinally pumped DPAA with a very high modal overlap between the pump and the signal, it is speculated that the amplifier was also in a saturated regime, and the ASE was sufficiently suppressed [39–42], consistent with the prior theoretical reports [23]. Further quantification of the saturated gain of our DPAA is being pursued by the authors.

 

Fig. 4. (a) Schematic diagram of the proposed diode-pumped alkali amplifier (DPAA) set-up (PBS: polarization beam splitter, PD: photo-detector, BPF: bandpass filter). (b) Temporal response of the signal modulated at 50 Hz for various cell temperatures, at pump power = 1.2 W. (c) Spectral response of the amplified signal measured by FPI. (d) Amplified output power as a function of the pump power for various cell temperatures.

Download Full Size | PDF

The amplified signal power is plotted as a function of the pump power for various cell temperatures in Fig. 4(d), for the signal power of 1mW. Consistent with Fig. 4(b), the DPAA showed the largest amplified signal power of 700 mW at T= 140 °C with the pump power of 1.2 W, corresponding to an optical gain of 700. See the pink curve in Fig. 4(d). The curve was well-fitted to a line and its slope was ∼58%. In contrast, lower T =100, 120, and 130 °C cases, the amplified signal did not show a linear response. At T= 150 °C, the signal grew with the pump power in a linear manner but with a less slope than at T= 140 °C. Compared with the previous small-signal gain study [12,23], the optical gain of ∼700 achieved in this study is an order of magnitude higher to show a significant potential in high power scaling in DPAA [43].

3. Conclusions

In summary, we experimentally measured amplified spontaneous emission in a hot Cs atom vapor cell, which showed both a nonlinear optical power growth and a spectral narrowing with increasing pump power. The spectra of ASE reduced from ∼9 GHz to ∼4.7 GHz as the pump power increased from 50 mW to 1 W. At a cell temperature of 150 °C, we obtained an ASE light source with a bandwidth of 4.7 GHz and an output power of 300∼350 mW at the pump power of 1.0 Watt. We also experimentally demonstrated a power scalability of Cs atom vapor optical amplifier using a signal laser of 0.5 MHz linewidth and 1mW optical power. At the cell temperature of 140 °C, the amplifier produced an output of 700 mW for the pump power of 1.2 W with a linear slope of ∼58%. In the spectral domain we confirmed that the ASE was efficiently suppressed by the amplified signal in a saturated regime, and the signal to noise ratio over 30:1 was measured. We found the very high spatial overlap between the pump and signal by propagating them in a single mode fiber played an important role in efficient ASE generation and high output amplification.