The synthesis of MIL-68(Al) and MIL-68(Fe) is the same as the literature [55]. The morphology of MIL-68(Al) crystals was observed by transmission electron microscopy (TEM), as shown in Figure 1A. According to the TEM image, the prepared MIL-68(Al) are clusters of needlelike crystals with different lengths [56]. Figure 1B shows the XRD pattern of MIL-68(Al) crystals. All the characteristic peaks of the MIL-68(Al) material are similar to those previous literatures [57, 58].

Similarly, the as-prepared MIL-68 (Fe) sample was also characterized. Figure 2A shows the TEM image and the size of the as-synthesized MIL-68(Fe) is about 1–3 μm. As shown in Figure 2B, the positions of the typical peaks of the XRD pattern matched well with the previous work, indicating that the crystal lattice parameters (cell length and angle) are the same but differ in relative intensity. However, the relative intensities are different influenced by the meritocratic orientation [59].

Figure 3 shows the measurement setup of the nonlinear absorptions. A self-made laser which generates pulse duration of 1.42 ps at 1.99 μm was used as one of the laser sources, as shown in Figure 3A. An optical output coupler with 50:50 fiber-pigtailed was used to simultaneously detect the reference signal and absorption. MIL-68(Al) or MIL-68(Fe) were coated on a CaF₂ window plate (F1) and the F2 was an uncoated CaF₂ window plate. Two detectors (D1 and D2) were used to measure the average powers. In addition, we used a homemade mode-locked fiber laser operating at 2.87 μm as the other laser source. The repetition rate is 18.39 MHz and a pulse duration is ∼20 ps. The balanced twin detector system was elaborated in Ref. [19], as shown in Figure 3B.

The parameters of the SAs were fitted with the following formula: R I = 1 − ∆ R ∙ exp − I I s a t − R n s , where R I indicates the reflectivity, ∆ R is the modulation depth, I is the incident peak intensity, I s a t is the saturation peak intensity and R n s represents the non-saturable loss [19]. Figure 4A shows the experimental results. The modulation depth, non-saturable loss, and saturation peak intensity of MIL-68(Al) were fitted to be 9.12%, 41.47%, and 0.3468 GW/cm². As shown in Figures 4B,C, the modulation depth, non-saturable loss, and saturation peak intensity of MIL-68(Fe) were fitted to be 18.89%/15.79%, 35.51%/52.47%, and 0.2198 GW/cm²/0.0044 GW/cm² under 1.99 μm and 2.87 μm laser irradiation, respectively.

The experimental setup of the passively Q-switched Tm³⁺-doped fiber laser is depicted in Figure 5. An all-fiber-integrated ring cavity configuration was adopted. The pump source was a 12 W commercial 793 nm diode laser (BWT). The gain fiber was a 5.8 m double-clad Tm³⁺-doped fiber (Coractive, 4 dB/m absorption at790 nm) which has a diameter of 128 μm and a numerical aperture (NA) of 0.22. The gain fiber was pumped via a 1 + 1 × 1 pump combiner. The fiber circulator guaranteed the one-way light propagation in the ring cavity with definite light propagation direction (1→2→3). In addition, the port-2 pigtail of the circulator was vertically cut to the fiber axis and positioned closely to the gold mirror coated with MOFs for optical modulation and feedback. A 25% port of a 25/75 fiber coupler was used to output the laser from the cavity. The average output power was measured with a power meter (Laserpoint). A 350-MHz-bandwidth digital oscilloscope was used to record the pulse trains and waveforms. The spectral profiles of output pulses were monitored by an optical spectrum analyzer (Yokogawa AQ6375, Japan).

The MIL-68(Al)-based passively Q-switched operation self-stared when the pump power was increased to 2.06 W, as shown in Figure 6A. A repetition rate of 38.81 kHz and a pulse duration of 5.73 μs were obtained. As the pump power gradually increased to 2.57 W, the pulse sequence maintained a stable Q-switched state with a repetition rate of 42.44 kHz and the minimum pulse duration of 2.13 μs. Once the pump power exceeded 2.57 W, the pulse trains started to grow erratic and then faded away. Nevertheless, stable Q-switching operation could be recaptured when the pump power was decreased, demonstrating that the MIL-68(Al) was not damaged by the photothermal effect [19, 32]. The high optical damage threshold of MIL-68(Al) was also confirmed [55]. The Q-switched output power increased from 1.37 mW to 3.08 mW and the maximum pulse energy was 0.073 μJ, as displayed in Figure 6B. Accordingly, the highest peak power was calculated to be 0.034 W. The inset shows that the center wavelength of the pulsed laser is 1988.5 nm. The repetition rate increased from 38.81 kHz to 42.44 kHz as the pump power increased, as displayed in Figure 6C. Specifically, the pulse duration decreased from 5.73 μs to 2.13 μs. As shown in the inset, the signal-to-noise ratio (SNR) of 41.1 dB was measured at the frequency of 42.44 kHz, indicating a stable Q-switched operation.

When MIL-68(Al) was replaced with MIL-68(Fe) in the same experimental setup, the MIL-68(Fe)-based passively Q-switched operation self-started as the pump power raised to 1.59 W, as shown in Figure 7A. The repetition rate was 18.89 kHz. Stable Q-switching operation without any adjustment was sustained until the launched pump power of 2.79 W. The repetition rate and pulse duration were 50.32 kHz and 3.37 µs, respectively. With further increasing the launched pump power, the Q-switching began unstable and then disappeared. When we reduced the pump power to less than 2.79 W again, stable Q-switched operation could be observed again. As the launched pump power rising from 1.59 W to 2.79 W, the Q-switched output power and pulse energy both increased, as shown in Figure 7B. The maximum output power of 13.3 mW and the maximum pulse energy of 0.26 μJ were received. Accordingly, the highest peak power of 0.078 W was obtained. The center wavelength was 1973.3 nm, as shown in the inset. Figure 7C shows the variation of Q-switched pulses over the same pump range. The repetition rate raised from 18.89 kHz to 50.32 kHz while the pulse duration decreased from 5.72 μs to 3.37 μs. The measured signal-to-noise ratio (SNR) was 52.5 dB at the frequency of 50.32 kHz, indicating a stable Q-switched operation.

Mode-locked operation of the two fiber lasers were not observed in our experiments. This may be related to the parameters of SAs and/or the current laser resonator design. By comparing selected all-fiber Q-switched lasers operating in the 2 μm region in Table 1, the minimum pulse width of MIL-68(Al) in this work is better and the maximum pulse energy of MIL-68(Fe) measured in this system is higher than most of materials reported previously [20–22, 30, 31]. Compared with traditional SESAM, MOFs has the characteristics of simple manufacturing process and compared with low-dimensional materials like BP and TMDs, MOFs exhibits better physiochemical stability with high laser damage threshold, as well as the high temperature stability (∼500°C). In addition, the pulse width obtained by using MIL-68(Al) as SA is narrower and the pulse energy is higher compared to the one when using MIL-68(Fe) as SA, while the average output power and peak power obtained by using MIL-68(Fe) as SA are both higher compared to the ones when using MIL-68(Al) as SA.

We have also studied the optical performance in the 3-μm waveband of MIL-68(Fe). We established the MIL-68(Fe) SAs enabled passively Q-switched mid-infrared fiber laser. Figure 8 shows the experimental setup of the passively Q-switched fiber laser based on MIL-68(Fe). The pump light was a commercial laser diodes (LD) (Eagleyard Photonics) operating at 1,150 nm. A 5.5 m long Ho³⁺/Pr³⁺ co-doped ZBLAN fiber (FiberLabs) was used as the gain fiber. The core diameter is 10 µm and numerical aperture (NA) is 0.2. The front end of fiber was vertically cut to the fiber axis to provide a 4% feedback. A dichroic mirror (DM) which has a high reflectivity at 2.8 μm and high transparency at 1,150 nm was 45°placed to be used as the output coupler. An anti-reflection CaF₂ plano-convex lens (L1: f = 20 mm) was employed to collimate the laser beam. The coupling efficiency was estimated to be 82% [30]. The output light beam from the angle-cleaved fiber end was collimated and focused onto the gold mirror coated through MIL-68(Fe) with a pair of ZnSe objective lens (L2: f = 12 mm, L3: f = 6 mm).

The fiber laser started Q-switching operation when the incident pump power was increased to 0.46 W, as shown in Figure 9A. The repetition rate and pulse duration were 182.0 kHz and 1.26 μs. The Q-switching operation can be maintained until the pump power of 1.26 W. The shortest pulse width of 497 ns was obtained with a repetition rate of 189.9 kHz. Figure 9B shows the Q-switched average output power and pulse energy over the same incident pump power. The output power increased linearly with the pump power and the maximum output power was 107.2 mW. The maximum pulse energy of 0.56 μJ and peak power of 1.14 W were obtained. The inset shows the Q-switched pulse spectrum and the center wavelength locates at 2,864.3 nm. The full width at half maxima (FWHM) was 1.6 nm. The repetition rate raised from 182.0 kHz to 189.9 kHz while the pulse width decreased from 1.26 μs to 497 ns, as displayed in Figure 9C. The radio-frequency (RF) spectrum with a signal-to-noise ratio (SNR) of 33.1 dB was measured at the repetition rate of 189.9 kHz, as shown in the inset of Figure 9C.

In contrast to the output characteristics of 2.9 µm Q-switched Ho³⁺ and Ho³⁺/Po³⁺ co-doped fiber lasers modulated by the typical 2D materials, a stable passive Q-switched laser with a short pulse width of 497 ns, which is the shortest one, as far as we know, was investigated based on a novel MIL-68(Fe)-SA. As can be seen in Table 2, The generated pulse peak power manifested the advantages compared with materials like graphene [17], PtSe₂ [19], WS₂ [23], antimonene [32] etc. In addition, the MIL-68(Al) and MIL-68(Fe) SAs used in our experiments were stored in the thermostat for two to 3 months and then the Q-switched laser experiments were repeated. Stable Q-switching pulses can still be achieved, although the pulse performances are slightly different at a given pump power. The results show that MIL-68(Al) and MIL-68(Fe) are SA materials with long-term stability in the mid-infrared spectral range.