Metal halide perovskites (i.e., ${\text{CsPbX}}_3$, X = Cl, Br, and I) materials have broad application prospects in the fields of high-definition displays, solar cells, abyss communications, and optoelectronic devices [1–7]. As a direct bandgap semiconductor, perovskite has the advantages of the adjustable bandgap, high optical gain, large absorption coefficient, high quantum yield, and low defect state density [8–13]. It has become an ideal material for designing low-threshold, color laser devices [14–19]. Despite such bright prospects, the rapid luminescence quenching and poor hydrothermal stability of perovskite nanocrystals (NCs) severely impede their practical applications in the laser field [20].

In order to optimize the stability of perovskite NCs, several different means have been employed [21–25]. Protesescu et al. reported ${\text{FAPbBr}}_3$ NCs with a strongly improved chemical stability during purification processes and good amplified spontaneous emission (ASE) properties [26]. Krieg et al. synthesized ${\text{CsPbBr}}_3$ NCs with a low ASE threshold capped with zwitterionic long-chain molecules, which further improved the long-term stability of NCs [27]. Yan et al. recently modified the ligands by replacing the oleic acid (OA) ligands with 2-hexyldecanoic acid (DA) to improve the stability of ${\text{CsPbBr}}_3$ NCs, and DA functionalized NCs showed a lower ASE threshold than OA affected NCs [28]. Although ligand modification, surface modification, and formation of core layer structure can effectively stabilize perovskite NCs [29–33], there is still much room for improvement in stabilizing ${\text{CsPbX}}_3$ NCs. Alternatively, since inorganic glass can protect ${\text{CsPbX}}_3$ NCs from the external environment, embedding ${\text{CsPbX}}_3$ NCs in glass has acted as one of the effective methods to improve the stability of perovskite NCs [34]. It is important that the perovskite NCs embedded in the glass not only exhibit ultrastable performance, but also have a low ASE threshold.

Although researchers have recently worked hard to adjust the ASE thresholds of perovskite NCs by adding organic molecular additives [35], zinc oxide nanoparticles [36], silver nanoparticles, [37], and changing the morphology and roughness of the sample [38], particle size also plays a vital role in regulating the ASE thresholds [39]. Fortunately, the size of perovskite NCs can be adjusted by controlling the doping of metal ions and the heat treatment temperatures. Furthermore, there are few reports linking the doping of metal ions and heat treatment conditions with the ASE thresholds of perovskite NCs at the same time. Based on the above, it is necessary to prepare a series of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses to study the effect of ${\mathrm{Gd}}^{3+}$ concentrations and heat treatment temperatures on the ASE thresholds of perovskite NCs. Excitingly, the prepared Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses can exhibit ASE characteristics, and it is found that doping metal ions and heat treatment temperatures have a regulatory function on the ASE thresholds of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses.

In this work, a series of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs emitting blue light were successfully synthesized in borosilicate glass by the traditional melting quenching method. The adjustment on the bandgap and wavelength of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses can be achieved by changing the doping concentration and heat treatment temperature. Moreover, the prepared Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass possesses good optical performance and excellent hydrothermal stability. Finally, under 800 nm fs pulsed laser at room temperature, not at low temperature, Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses can be used in the laser field, and the ASE threshold of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass is as low as $0.39{\mathrm{mJ}/\mathrm{cm}}^2$.

Cesium carbonate (${\mathrm{Cs}}_2{\mathrm{CO}}_3$, 99.9%), lead bromide (${\mathrm{PbBr}}_2$, 99%), lead chloride (${\mathrm{PbCl}}_2$, 98%), gadolinium fluoride (${\mathrm{GdF}}_3$, 99.9%), sodium chloride (NaCl, 99.5%), sodium bromide (NaBr, 99.9%), silicon dioxide (${\mathrm{SiO}}_2$, 99%), boron oxide (${\mathrm{B}}_2{\mathrm{O}}_3$, 99.9%), and zinc oxide (ZnO, 99%) were bought from Aladdin. All chemicals were used without further treatment.

ZnO (molar percentage: 14.27%), ${\mathrm{B}}_2{\mathrm{O}}_3$ (32.80%), and ${\mathrm{SiO}}_2$ (28.27%) were first weighed as the glass matrix in the preparation of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses. The three ingredients were mixed and milled uniformly in the mortar, and then NaBr (5.58%), NaCl (5.58%), ${\mathrm{PbBr}}_2$ (2.48%), ${\mathrm{PbCl}}_2$ (2.48%), and ${\mathrm{Cs}}_2{\mathrm{CO}}_3$ (8.54%) were put into the alumina crucible. The crucible was put in the heating furnace and was heated to 1200°C for 10 min to ensure that everything melts. Then the crucible was taken out and poured into a grinding tool preheated to 400°C. After the glass was slightly shaped, it was quickly put in an oven at 400°C and heated for 3 h and cooled to room temperature naturally. This heat preservation process is a necessary process, called annealing. It is designed to eliminate the internal stress of the ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs precursor glass and prevent the newly formed glass melt from breaking due to a sudden drop in temperature. A series of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses were obtained by regulating the concentration of ${\mathrm{GdF}}_3$ (GdF₃ concentration of 0%, 0.5%, 1%, 3%, respectively marked as S1, S2, S3, and S4).

Using Bruker D8 Advance measurement, the phase structure of the samples was characterized by X-ray diffraction (XRD). It is pointed out that the ${\mathrm{Cu}\mathrm{K}}_{\alpha }$ radiation of Bruker D8 Advance measurements was running in the boundary of 10°–60° ($2\theta$) at the velocity of $0.02$° per second (the current as 40 mA and the voltage as 40 kV). The ultraviolet–visible (UV-vis) spectra were surveyed by a PerkinElmer 750 UV-vis spectrometer, and the recording scope was 400–600 nm. The JEOL JEm-2100F transmission electron microscope (TEM) was used to observe the microstructure of perovskite NCs. The distribution of elements was tested with TEM-assisted energy dispersive spectrometer (EDS). The temperature-dependent photoluminescence (PL) spectra were measured with THMS 600. Under the excitation of the commercially available Ti:sapphire regenerative amplifier (800 nm, 35 fs, 1 kHz) at room temperature, the ASE of perovskite NCs glasses has been actively studied. A cylindrical lens with a focal length of 10 cm was used to focus the laser beam onto the surface of the sample. Adjustable slits could be used to precisely control the fringe length of the light spot. The ASE was detected by a fiber optic spectrometer (Marine Optics) with a spectral resolution of 1 nm. The fluorescence and absorption measurements are all under the same excitation wavelength of 365 nm.

The network structure of borosilicate glass is silicon oxide tetrahedron, which is mainly composed of bridged and non-bridged oxygen bonds. In the precursor borosilicate glass, ${\mathrm{Cs}}^{+}$, ${\mathrm{Pb}}^{2+}$, ${\mathrm{Gd}}^{3+}$, ${\mathrm{F}}^{-}$, ${\mathrm{Cl}}^{-}$, and ${\mathrm{Br}}^{-}$ ions are distributed in the borosilicate glass. Once external conditions give these ions enough energy to reach the formation energy of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs, ${\mathrm{Cs}}^{+}$, ${\mathrm{Pb}}^{2+}$, ${\mathrm{Cl}}^{-}$, and ${\mathrm{Br}}^{-}$ ions will migrate to the network voids to nucleate and grow. However, due to the inconsistent size of the network voids, the size of the formed perovskite NCs is also not uniform [40]. To explore the optimal ${\mathrm{Gd}}^{3+}$ concentration in ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses, a series of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses were synthesized, and they all emitted bright blue light under UV light [Fig. 1(c)]. Then, the PL of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses was tested. As shown in Fig. 1(a), a certain amount of ${\mathrm{Gd}}^{3+}$ doping promotes the increase of the PL intensity of ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses, but excessive ${\mathrm{Gd}}^{3+}$ doping quenches the PL intensity of ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses. It can be concluded that the best ${\mathrm{Gd}}^{3+}$ doping in ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses is found to be 1% (S3). The quantum efficiencies of S1–S4 are 5.1%, 10.3%, 12%, and 2%, respectively. ${\mathrm{Gd}}^{3+}$ doping improves the optical performance of ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses because ${\mathrm{Gd}}^{3+}$ ions reduce defects, which enhances radiative recombination and reduces non-radiative recombination paths [41]. As for ${\mathrm{F}}^{-}$ ion, ${\mathrm{F}}^{-}$ ion promotes a part of ${\mathrm{Gd}}^{3+}$ doping into ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses by fracturing the silicon–oxygen bonds and generating non-bridging oxygen bonds [42]. In addition, the PL emission peak position has a slight blueshift as the ${\mathrm{Gd}}^{3+}$ concentration increases, which may be because part of ${\mathrm{Gd}}^{3+}$ replaces ${\mathrm{Pb}}^{2+}$ caused by the reduction of the particle size. Figure 1(b) shows that as the ${\mathrm{Gd}}^{3+}$ concentration increases, the bandgap energy increases, which is caused by more ${\mathrm{Gd}}^{3+}$ entering the lattice leading to the shrinkage of the perovskite NCs lattice. Since perovskite NCs are direct interband transitions, the formula can be used to calculate the optical bandgap ${\mathit{E}}_g$ of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses as follows [43]: $${(\alpha h\nu )}^2=A(h\nu -E_g),$$where $A$ is the parameter concerning the effective mass of the valence and conduction bands, $h$ is Planck’s constant, $\alpha$ is the absorption coefficient, and $\nu$ represents the optical frequency.

Next, the phase structure and microstructure of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses were analyzed. It can be seen from Fig. 2(d) that the diffraction peaks are located between the standard diffraction cards (PDF#54–0752 is the standard diffraction card of cubic ${\text{CsPbBr}}_3$ and PDF#18–0366 is the standard diffraction card of cubic ${\text{CsPbCl}}_3$), demonstrating that the ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs were successfully precipitated in the borosilicate glass matrix. In addition, it can also be seen that the diffraction peak intensities of doped and undoped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses are quite strong, indicating that the crystallinity of the perovskite NCs in S1–S4 is good. It needs to be emphasized that, with the increase of ${\mathrm{Gd}}^{3+}$ content, the XRD peaks of ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses move slightly to a larger angle. This shift is attributed to the contraction of the crystal lattice, which is due to the partial replacement of ${\mathrm{Pb}}^{2+}$ with a larger radius by ${\mathrm{Gd}}^{3+}$ with a smaller radius. From the perspective of microstructure, it can be clearly seen from the TEM image [Fig. 2(a)] that Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs with a circular shape are embedded in the glass matrix. By comparing the high-resolution transmission electron microscope (HRTEM) of doped (S3) and undoped ${\mathrm{Gd}}^{3+}$ (S1), it is concluded that the (1 1 1) lattice fringe spacing of ${\mathrm{Gd}}^{3+}$ doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass is obviously reduced [Figs. 2(b) and 2(c)]. This is the result from partial ${\mathrm{Gd}}^{3+}$ doped into ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass that causes the crystal lattice to shrink [44]. Figures 2(e)–2(j) show the EDS element distribution of a part of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass, which confirms the presence of F, Cl, Br, Cs, Pb, and Gd in Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass.

The particle size of ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs can not only be controlled by metal ions doping, but temperature is also one of the factors that affect the size of the perovskite NCs [45]. Therefore, the different heat treatment temperatures (470°C–10 h, 500°C–10 h, and 530°C–10 h) of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses were explored. As shown in Fig. 3(b), the intensity of PL first increases and then decreases. This is because a proper increase of the heat-treatment temperature is beneficial to the formation and growth of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs, which will cause the intensity of PL to increase. However, the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs become larger when the heat-treatment temperature exceeds the optimal temperature, causing the aggregation of perovskite NCs, which will lead to concentration quenching and weaken the intensity of PL [46]. As the heat-treatment temperature increases from 470°C to 530°C, the intensity of XRD diffraction peaks becomes stronger and tends to shift slightly to a low angle [Fig. 3(a)] for the increase in size of perovskite NCs. The increase in the size of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs will result in a redshift of the PL peak position [Fig. 3(b)] and a decrease in the bandgap [Fig. 3(d)]. In Fig. 3(c), due to the Stokes shift [47], the edge of the absorption spectrum will be located near the PL peak of the corresponding sample. The precise control of the absorption [Fig. 3(c)] and bandgap [Fig. 3(d)] of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses can be successfully achieved by changing the heat treatment temperature.

The quality of the prepared perovskite NCs should be considered, including good optical properties and good stability. Therefore, it is necessary to test the hydrothermal stability of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass (S3). The heat resistance of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass is pictured in Fig. 4(a). Even after heating, cooling, and reheating, the optical properties of the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass can be restored to the original state, showing excellent thermal cycling performance. As shown in Fig. 4(c), the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass (S3) was immersed in water, and it can be observed that the S3 remains bright blue light after being stored in water for 30 days without any significant changes, because the borosilicate glass substrate as an ideal protective layer can effectively isolate the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs from the harsh external environment. PL measurements were performed at different times [Fig. 4(b)], and the results show that even after 30 days, the intensity of PL remains around 96%. The above indicates that the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass has excellent water resistance. All in all, the prepared Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass has excellent water/heat resistance.

The laser mechanism of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses is that randomly arranged Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs work as both a gain medium and multiple scattering sources. The emitted light undergoes multiple scattering at the grain boundaries and finally returns to the path where they were scattered before, thereby forming a closed-loop cavity. When the optical gain exceeds the loss, laser oscillation is generated at the resonant frequency of the corresponding feedback loop [48]. A series of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses were subjected to 800 nm fs laser experiments at room temperature to explore the relationship between metal ion doping and ASE thresholds. As shown in Figs. 5(a), 5(c), 5(e), and 5(g), when the pump density is lower than the power density that can reach the threshold, the emission spectrum shows a broad emission, namely spontaneous emission (SE), which dominates the PL spectrum of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass. But when the pump density exceeds the power density that can achieve the threshold, the sharp increase in the output intensity of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass is accompanied by a sharp decrease in full width at half-maximum (FWHM), which is the transition from SE to ASE [49]. In the ASE measurement, the blueshift of SE can indeed be attributed to the phase separation of the perovskite NCs. This is because the local electric field destroys the ionic bonds in the mixed halogen NCs [50]. In addition, ASE is located at $\sim 510\mathrm{nm}$, which indeed comes from the ${\text{CsPbBr}}_3$-rich phase. In addition, it is found that the ASE peak has a redshift relative to the SE peak, which is caused by the recombination of double excitons [51,52]. Figure 5(b) shows the intensity (blue line) and FWHM (red line) versus pump density. It can be observed that when the pump density exceeds $0.54{\mathrm{mJ}/\mathrm{cm}}^2$, the spectra become narrower and the output intensity increases, so ${\mathit{P}}_{\mathrm{th}}=0.54{\mathrm{mJ}/\mathrm{cm}}^2$ is regarded as the ASE threshold of S1. By analogy, the thresholds of S2–S4 are $0.81{\mathrm{mJ}/\mathrm{cm}}^2$, $0.85{\mathrm{mJ}/\mathrm{cm}}^2$, and $1.22{\mathrm{mJ}/\mathrm{cm}}^2$, respectively [Figs. 5(d), 5(f), and 5(h)]. The gradual increase in the thresholds of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses (S1–S4) is caused by the decrease in the particle size of the perovskite NCs, which also confirms the fact that partial ${\mathrm{Gd}}^{3+}$ ions have been doped into ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs.

The effect of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses with different heat treatment temperatures on the ASE thresholds was also researched. As shown in Figs. 6(a)–6(c), the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses with different heat-treatment temperatures have ASE phenomena when the pump density is increased. In addition, the threshold of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses decreases with the increase of heat treatment temperature (470°C–530°C), and the ASE threshold can be as low as $0.39{\mathrm{mJ}/\mathrm{cm}}^2$ [Figs. 6(d)–6(f)]. This is the result of the increase in the heat treatment temperature that causes the Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs particles to become larger. In short, increasing the heat treatment temperature can obtain Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass with better crystallinity and larger grain size, making it more suitable for laser applications.

In summary, a series of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses in borosilicate glass (${\mathrm{SiO}}_2\text{−}{\mathrm{B}}_2{\mathrm{O}}_3\text{−}\mathrm{ZnO}$) were successfully prepared. The PL intensity of ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses can be improved after doping with ${\mathrm{Gd}}^{3+}$ ions. Specifically, the structure and optical aspect of Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses were studied deeply. Furthermore, Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glass shows excellent hydrothermal resistance. It is uplifting that Gd-doped ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses can be applied in the ASE field. By studying the effect of different ${\mathrm{Gd}}^{3+}$ doping concentrations and heat treatment temperatures on the ${\text{CsPbCl}}_{1.5}{\mathrm{Br}}_{1.5}$ NCs glasses, it is clarified that the ASE thresholds of perovskite NCs can be adjusted, which lays the foundation for creating more possibilities for the application of perovskite NCs in the laser field in the future.