Laser driven conversion of MOFs to rare earth metal oxide nanoparticles

Rare earth metals and their compounds are critical for modern manufacturing, including the field of semiconductor production, electronics, and catalysis, commonly referred to as the “vitamin for industry.”^1–5 Rare earth metal oxides are usually prepared by chemical, physical, and hybrid methods. Those methods inevitably involve complicated solution-based precipitation, energy-demanding thermolysis, and/or vacuum environment.⁶ It is well-known that metal oxides show drastically improved performance in catalysis when they are produced in nanometer scale and uniform size.^7–10 However, it remains difficult to achieve such precise control, especially for rare earth metal oxides.^2,11

Terbium sesquioxide (Tb₂O₃) is an important rare earth material as a potential candidate for the next generation of semiconductors.¹² The attempt to synthesize Tb₂O₃ usually ended up in the formation of Tb₄O₇, which is more stable thermodynamically. There are only a few reports on the preparation of Tb₂O₃. Three polymorphic forms exist for Tb₂O₃, such as cubic (C-type, Ia-3), monoclinic (B-type, C2/m), and hexagonal (A-type, P-3m1), where their synthesis is sensitive to the reaction conditions.¹³ In comparison to A- and C-types of Tb₂O₃, the monoclinic phase of Tb₂O₃ has been shown to excel in luminescent property.¹⁰ Unlike the C-type converted from Tb oxide or salt calcination,^6,14 the B-type is hard to be obtained directly from the corresponding metal sources. Although the B-type Tb₂O₃ can be achieved by transformation of the C-type, extremely harsh conditions, such as high pressure (7 GPa) and temperature (1700 °C), are required [Fig. 1(a)]. The resulting particle size is usually random. It is highly desirable to develop a method to prepare uniformly distributed Tb₂O₃ in nanometer size under relatively mild conditions.

Here, we show that lasers, with fast production capability, can be utilized in the synthesis of rare earth metal sesquioxide nanoparticles using metal–organic frameworks (MOFs) as precursors [Figs. 1(b) and 1(c)]. MOFs, which can be rationally designed in topology with atomically dispersed metal sites, have been widely used in various fields.^15–19 It has been demonstrated as precursors for the production of metal nanoparticles, transition metal carbides, and bimetallic alloys.^20–26 The orderly arranged metal-containing secondary building units (SBUs) play a critical role in light adsorption and are responsible for the generation of particles with uniform nanometer size. This process also gives energy conversion efficiency unmatched by that of the traditional thermolysis. In this study, we demonstrate that the laser conversion method is also applicable for the synthesis of Tb₂O₃ with diverse phase types by using different Tb-MOFs as precursors, such as Tb₆O₈(C₄H₂O₄)₆ (Tb-MOF-Fum), Tb(C₂₇H₁₂O₆)(H₂O)₂(C₆H₁₂O) (Tb-MOF-BTB), and Tb₁₆(C₂₄H₁₂O₆N₃)₁₆ · (C₄H₉N₂O)₂₄] · (C₄H₉N₂O)₉₁(H₂O)₁₀₈ (Tb-MOF-TATB). The variation of topologies in these Tb-MOFs allows for precise modulation on the volumetric density of Tb metals. Tb-MOF-Fum with a Tb metal volumetric density of 61.3 mmol cm⁻³ gives rise to the C-type Tb₂O₃ after laser irradiation, while Tb-MOF-BTB and Tb-MOF-TATB with smaller Tb metal volumetric density, 0.095 and 17.3 mmol cm⁻³, lead to the direct formation of B-type Tb₂O₃ (Fig. 2 and Table S2). All the terbium sesquioxide nanoparticles obtained from the laser conversion of Tb-MOFs are uniform in particle size and exhibit excellent oxygen evolution reaction (OER) performance. Specifically, the B-type Tb₂O₃ obtained from Tb-MOF-TATB shows the particle size around 3 nm and offers a small overpotential of 331 mV in OER, best among all terbium sesquioxides.

Three Tb-MOFs, denoted as Tb-MOF-Fum, Tb-MOF-BTB, and Tb-MOF-TATB, are used here to demonstrate the synthesis of terbium sesquioxide by laser conversion. These Tb-MOFs are composed of different organic linkers, such as fumaric acid (H₂Fum), 1,3,5-tris (4-carboxy phenyl) benzoic acid (H₃BTB), and triazine-1,3,5-tribenzoic acid (H₃TATB), respectively.^27–29 The Tb containing secondary building units (SBUs) are also different in these MOFs (Fig. 2). Tb-MOF-Fum is a microporous MOF, with the pore size around 6 Å, constructed by linking Tb₆O₈ hexanuclear SBU and fumaric acid [Fig. 2(a)]. The relatively small linker size results in a high Tb metal volumetric density of 61.31 mol cm⁻³. In contrast, the other two MOFs, Tb-MOF-BTB and Tb-MOF-TATB with a relatively larger pore size, constructed from tritopic linkers, H₃BTB and H₃TATB, give lower volumetric densities (17.36 and 0.095 mmol cm⁻³, respectively).

Prior to the laser conversion, the Tb-MOFs synthesized were activated by removing the solvent from the pores. Sharp peaks were observed in the powder x-ray diffraction (PXRD) patterns of the resulting solvent-free MOFs (Fig. 2). These PXRD patterns also matched well with those simulated from their corresponding crystal structures (Figs. 2 and S15), confirming the phase purity of these MOFs. N₂ adsorption isotherms were measured at 77 K, demonstrating their permanent porosity. The Brunauer–Emmett–Teller (BET) surface areas of Tb-MOF-Fum and Tb-MOF-TATB were 480 and 510 m² g⁻¹, respectively, consistent with the reported values (Figs. S10 and S11).²⁸ The alternating arrangement of terbium oxide clusters as oxidants and carbonaceous linkers as reductants at the molecular level in Tb-MOFs, makes these precursors ideally suited for the laser conversion.

The laser conversion process was carried out in the following sequence. First, the solvent-free Tb-MOF crystals were evenly packed and sandwiched between two glass slides with a sample thickness of 10 µm [Fig. 1(c)]. Then, laser scribing was applied line-by-line on the MOF layer to convert the MOF crystals into terbium oxides. The laser was programmed by the computer to give an 80 ns pulse duration and operated at a wavelength of 1064 nm at a frequency of 20 kHz. The parameters of the energy were precisely tuned for the conversion of different MOFs. The laser power was precisely tuned for these MOFs, 5.5 W for Tb-MOF-Fum and 9 W for Tb-MOF-TATB and Tb-MOF-BTB. Upon laser scribing, the white MOF powder instantly turned black, indicating the successful conversion. The entire laser conversion process was performed under ambient conditions without any facilitation from vacuum or protection by inert atmosphere. This straightforward process is suitable for industrial fabrication.

The black products detached from the glass slides were collected for further characterizations. The products converted from the corresponding MOF were termed Tb-MOF-laser. According to the PXRD patterns of all three products, they were assigned to Tb₂O₃, but with different crystal structures. Tb-MOF-Fum-laser was found to be dominated by the cubic phase (C-type) Tb₂O₃ with two different space groups, Ia-3 and Fm-3m [Fig. 2(a)]. In contrast, PXRD patterns of Tb-MOF-TATB-laser and Tb-MOF-BTB-laser samples showed characteristic peaks of Tb₂O₃ in B-type [Figs. 2(g) and 2(h)]. The full transformation from Tb-MOFs to Tb sesquioxides can also be reflected in their x-ray photoelectron spectroscopy (XPS) spectra. In the O 1s spectrum of the Tb-MOF-laser, a new peak besides 532 eV can be observed in three kinds of laser treated products, which indicated the phase transformation after laser (Figs. S17–S19). The resulted crystal phases of Tb₂O₃ might originate from the structural features of the original Tb-MOFs. One distinct difference was the volumetric density of the Tb metal in the MOF precursors, as a result of different atomic ratios between Tb and C. Tb-MOF-Fum with the largest Tb metal volumetric density of 61.31 mmol cm⁻³ favors the formation of Tb₂O₃ in the cubic phase, while Tb-MOF-BTB and Tb-MOF-TATB with a relatively smaller Tb metal volumetric density, 17.36 and 0.095 mmol cm⁻³, tend to generate B-type Tb₂O₃. Laser has been used to generate extremely high temperature and pressure, up to several thousands of kelvins and gigapascals,^30,31 which meets the requirements for the formation of B-type terbium oxides. Tb-MOF-BTB and Tb-MOF-TATB preserve lower volumetric densities of Tb, suggesting high atomic ratios between carbon and Tb atoms (C/Tb). The C/Tb ratios in Tb-MOF-BTB and Tb-MOF-TATB are 27 and 22, respectively, much higher than the value of 4 in Tb-MOF-Fum. The higher C/Tb ratios allow for more efficient absorbance of the laser in MOF, leading to much higher temperature, which favors the formation of B-type terbium sesquioxide. This is another example that the difference in the Tb metal volumetric density of the building blocks in the MOFs influences their macroscopic properties, in addition to the precise tuning of catalytic performance.³² Here, the rich chemistry in the MOF design provides a new way to synthesize novel nanomaterials by lasers, which is challenging to achieve via traditional synthetic methods.

In order to further confirm the crystal structures of the obtained terbium oxide nanoparticles, microscopic tests were performed. Scanning electron microscopy (SEM) images revealed that Tb-MOF-laser samples were quite uniform in the size of nanoparticles across a large area, and the sizes were tunable from 3 to 12 nm (Fig. S16). The nanoparticle sizes of Tb-MOF-BTB-laser and Tb-MOF-TATB-laser samples were much smaller than those of the Tb-MOF-Fum-laser sample. Transmission electron microscopy (TEM) was also used to confirm the phases of the resulting terbium oxide nanoparticles [Figs. 3(a), 3(f), and 3(k)]. The presence of B- and C-type terbium oxides can be unambiguously distinguished in high-resolution TEM images by their distinctive lattice fringes [Figs. 3(c), 3(h), and 3(m)]. In the case of Tb-MOF-Fum-laser, the lattice distance of the nanoparticles was 0.32 nm, consistent with the 222 facet of Tb₂O₃ in C-type. The lattice distances of nanoparticles in Tb-MOF-BTB-laser and Tb-MOF-TATB-laser were 0.31 and 0.27 nm, representing the 111 and 11-2 facets of Tb₂O₃ in B-type, respectively. The selected area electron diffraction (SAED) patterns of three Tb-MOF-laser samples also matched well with the corresponding terbium oxides structures revealed by high-resolution TEM images [Figs. 3(b), 3(g), and 3(l)]. Statistical analysis of the TEM images showed that the nanoparticle size of the laser treated samples from Tb-MOF-Fum, Tb-MOF-BTB, and Tb-MOF-TATB was 12, 6, and 3 nm, respectively, with a narrow size distribution [Figs. 3(e), 3(j), and 3(o)].

In addition, few-layered graphene was also found surrounding the terbium oxide nanoparticles as evidenced by the characteristic lattice distance of 0.34 nm. The presence of few-layered graphene could be identified by its fingerprint 2D band, which peaked at 2700 cm⁻¹ in the Raman spectrum [Figs. 3(d), 3(i), and 3(n)]. Graphene has been widely used as support for metal/metal oxide nanoparticles because of its unique interaction with the nanoparticles.²² The presence of graphene provides ideal anchor sites for these ultrasmall nanoparticles and precludes their aggregation under harsh application environments. The BET surface areas of Tb-MOF-Fum-laser, Tb-MOF-BTB-laser, and Tb-MOF-TATB-laser were 30, 153, and 50 m²/g, respectively (Figs. S23–S25).

The electrocatalytic oxygen evolution reaction (OER) is applied to demonstrate the difference between the terbium oxide nanoparticles obtained from different MOFs. The OER activities of all products were measured in a 1.0M KOH solution using a standard three-electrode setup.³³ The linear sweep voltammetry (LSV) of three samples showed distinct rates in oxygen productivity, as evidenced by the current densities [Fig. 4(a)]. The Tb-MOF-TATB-laser sample exhibited the largest current density at a fixed potential. An overpotential of 331 mV was observed in the Tb-MOF-TATB-laser sample at a current density of 10 mA cm⁻² [Fig. 4(b)], superior to that of the Tb-MOF-Fum-laser (459 mV) and Tb-MOF-BTB-laser (370 mV). The overpotential of the Tb-MOF-TATB-laser was in the same ball park of commercially available IrO₂ catalysts (314 mV).³⁴ The Tafel slopes of these samples were also examined to identify their OER kinetics. As shown in Fig. 4(c), the Tafel slope of the Tb-MOF-TATB-laser (80 mV dec⁻¹) was less than that of Tb-MOF-Fum-laser (95 mV dec⁻¹), indicating an improved OER kinetics. Electrochemical impedance spectroscopy (EIS) tests were also carried out to investigate the charge-transfer kinetics. As shown in Fig. 4(d) and Table S5, the EIS curves revealed that the Tb-MOF-TATB-laser catalyst preserved an impedance of 153.4 Ω, much smaller than those of Tb-MOF-Fum-laser and Tb-MOF-BTB-laser (304.2 and 166.0 Ω, respectively). The smallest impedance in the Tb-MOF-TATB-laser catalyst manifested its efficient charge transfer kinetics, ascribing to the small particle sizes of Tb-MOF-TATB-laser, which provided rich contact between Tb₂O₃ nanoparticles and the conductive graphene agent. In addition, the smaller particle size also provided more active sites for the catalytic reaction, which, in turn, enhanced the OER activity. This showed that the rational design of the MOF precursors is feasible to prepare advanced nanomaterials to meet with applications on demand. The advantages of laser conversion to generate rare earth metal oxide nanoparticles in air were reflected in the short reaction time, low energy consumption, small particle sizes, and challenging phase hard to be obtained from traditional methods (Table S6). The combination of the rare earth MOF and laser provides the great potential for the discovery of novel rare earth metal oxides in the nanometer scale.

Two types of terbium sesquioxides are successfully prepared by laser conversion using different Tb-MOFs as precursors. The fast heating and subsequent cooling feature of laser allows for the production of terbium sesquioxide with the particle size ranging from 3 to 12 nm and with a narrow size distribution. MOFs, designable in different topologies, enable the synthesis of Tb₂O₃ in different phases. Specifically, the B-type Tb₂O₃ in the nanometer scale can be directly transformed from the MOF by laser irradiation under ambient conditions, unprecedented in conventional heating exchange and high-pressure method. The B-type Tb₂O₃ particles exhibited the best OER performance with the particle size down to 3 nm. This high speed and energy-saving way to produce terbium sesquioxide set an example for the potential synthesis of other rare earth metal oxides.

The synthesis of Tb-MOF-Fum was achieved with slight modification from previous studies.^27,35 Specifically, fumaric acid (FA, 0.522 mmol), Tb(NO₃)₃ · 5H₂O (0.261 mmol), and 2-fluorobenzoic acid (FBA, 4.176 mmol) were dissolved in the solution of N,N-dimethylformamide (DMF, 8.1 ml) and deionized H₂O (2.1 ml) in a 20 ml vial. The vial was sealed and heated to 115 °C for 72 h to yield colorless polyhedral crystals. Sequential solvent exchange was done before the activation of the resulting Tb-MOF-Fum crystals. Typically, the as synthesized MOF crystals were washed by DMF (10 ml) for three times, followed by washing with 10 ml of ethanol for another week to remove the unreacted metal salts and organic linkers in the pores. During this period, ethanol was refreshed three times per day. The Tb-MOF-Fum crystals were dried at 100 °C under vacuum for 10 h to fully activate the MOF sample.

The synthesis of Tb-MOF-TATB was carried out with slight modifications from the previous work.²⁸ Tb(NO₃)₃ · 5H₂O (0.345 mmol) and 2,4,6-tris(4-carboxyphenyl)-s-triazine (H₃TATB, 0.113 mmol) were mixed and dissolved in the solution of N,N-dimethylacetamide (DMA, 10 ml) and methanol (2 ml) in a 20 ml vial. Then, it was sealed and incubated under 105 °C for 48 h to generate colorless crystals in polyhedral shape. Before the activation of Tb-MOF-TATB crystals, solvent exchange was applied. Generally, the obtained MOF crystals were washed with 10 ml of anhydrous DMF for three times, followed by immersion in 10 ml of ethanol for 7 days, during which ethanol was replaced three times per day. The MOF crystals were heated to 150 °C under vacuum for 12 h to yield the activated Tb-MOF-TATB sample.

The synthesis and activation of Tb-MOF-BTB were performed based on a previous report with modifications.³⁶ 1,3,5-benzenetrisbenzoic acid (H₃BTB, 0.68 mmol), Tb(NO₃)₃ · 5H₂O (0.682 mmol), and NaOH (2M, 0.6 ml) were dissolved in H₂O (7.5 ml) in an autoclave. The mixture was heated at 200 °C for 4 days to produce colorless polyhedral crystals. The as synthesized MOF crystals were washed with 10 ml of anhydrous DMF for three times and then were immersed in 10 ml of ethanol for three times. The activation of the Tb-MOF-BTB crystals was achieved by thermal treatment at 120 °C for 12 h under vacuum.

There are two critical steps in the laser conversion process—sample preparation and tuning of laser parameters. In the sample preparation step, 8.0 mg of the activated Tb-MOF crystals was placed in the center of a copper foil with a hole of 12 mm in diameter. Then, the samples were pressed between two glass slides to achieve an even packing, followed by wrapping taps on the sample-free sides to fix the samples. The thickness of the MOF sample was roughly 10 µm. In the laser adjustment step, a nanosecond pulsed 1064 nm laser was used in this study as the energy source (YLP-0.5-80-10, IPG photonics). It was operated with a frequency of 20 kHz, and the pulse duration was set up to 80 ns. The laser beam was focused by F-Theta lens (LINOS), and the beam scribing pathway was programmable by computer software. The output power and the scribing speed of the laser were precisely tuned for each Tb-MOF sample. In order to guarantee the accuracy, the output power of the laser was calibrated by an optical power meter prior to the laser processing. In a typical trial, the object distance and the distance between the sample and the focus lens were carefully adjusted to 330 mm. The laser power was tuned systematically for each MOF to achieve complete conversion.

PXRD was performed in a SmartLab diffractometer (Rigaku). Cu Kα radiation (λ = 1.5405 Å) was used for measurement using a beam filter was to minimize influences from another wavelength. The PXRD patterns of the corresponding MOF and rare earth metal oxides were simulated using Materials Studio based on crystal structures from database. XPS measurement was carried out on ESCALAB250Xi instrument (Thermo Fisher). The peaks of all samples were calibrated based on C 1s spectra at a binding energy of 284.8 eV. Solid-state UV–vis near-infrared (NIR) absorption measurement was performed on a UV-3600 (Shimadzu) UV/vis/NIR spectrometer. Raman spectra were taken on an inVia confocal Raman spectrometer (RENISHAW, Britain). The laser wavelength was selected at 532 nm, and the irradiation process was set as a microscope with a 50× LP objective lens (Olympus MPlan, Japan). N₂ adsorption isotherms at 77 K were measured using an Autosorb iQ₂ analyzer (Quantachrome). A field emission SEM (Verious 460, FEI) and a LaB₆ TEM, JEM-2100Plus (JEOL, Japan) were used to collect the SEM and TEM images of the samples, respectively. Thermogravimetric analysis was conducted under ambient conditions in a thermal gravimetric analyzer (TGA, STA449F5, NETZSCH, Germany).

The electrochemical performance tests were performed on a CHI 760E electrochemical workstation (Chenhua, Shanghai), utilizing a three-electrode configuration under ambient conditions. First, Nafion solution (0.6 ml of 10 wt. %) was mixed with the metal oxide nanoparticles (3 mg) to prepare the catalyst ink, followed by ultrasonication for 1 h. In a typical test, the working electrode was prepared by casting 80 µl of the catalyst ink onto nickel foam with an area of 0.2 cm². The overall mass loading of the catalyst was 2 mg m⁻². The Hg/HgO electrode and carbon rod were adopted here as reference and counter electrodes, respectively. The LSV test was performed in 1.0M KOH solution with a potential window between 1.22 and 1.77 V. The scanning speed was set to 1 mV s⁻¹ in saturated O₂ environment. The potentials were calibrated on the basis of reversible hydrogen electrode (RHE) by the following formula: E (vs RHE) = E (vs Hg/HgO) + 0.059 × pH + 0.098. The potential of the Hg/HgO electrode (vs RHE) was calibrated using the Pt/C electrode (20 wt. % loading of Pt).

See the supplementary material for additional experimental detail and analysis.

This project has received financial support from the National Natural Science Foundation of China (NSFC) (Grant Nos. 91545205, 91622103, 21971199, and 22025106), the National Key Research and Development Project (Grant No. 2018YFA0704000), the National Key Basic Research Program of China (Grant No. 2014CB239203), and the Innovation Team of Wuhan University (Grant No. 2042017kf0232).

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding author upon reasonable request.