Investigation of upconversion phenomenon in Y _1-x Ln³⁺_x(BTC)(DMF)₂(H₂O) and Y_0.8-xYb_0.2Ln³⁺_x(BTC) (DMF)₂(H₂O) metal organic frameworks

INTRODUCTION

One of the pioneers of a successful study of the phenomenon of upconversion was Auzel, who in 1966 first raised the idea that energy transfer between rare earth RE ions can occur between two cations, both of which are in the excited state when energy transfer takes place in the initial stage [1]. Er³⁺–Yb³⁺ and, shortly aſterwards in another publication, Tm³⁺–Yb³⁺ trivalent ion pairs were the first among which Auzel experimentally observed the  conversion in 1966. These experiments were published separately by Ovsyankin with Feofilov in the same year, only they added a  pair of Ho³⁺–Yb³⁺ ions to these two pairs [2]. In addition to Er³⁺, Tm³⁺ and Ho³⁺ ions, which are still the most commonly used in conversion studies, Pr³⁺ and Nd³⁺ are used, much less some remaining Ln³⁺ (Tb³⁺, Eu³⁺, Sm³⁺, Dy³⁺), as well as several transition metals or actinides [1, 3].

Since 1997, a new class of the high-ordered and high-porosity materials, metal-organic frameworks (MOFs), began to be produced and investigated actively [4]. Such structures caused a great interest for researchers in regard to their unique structural, physical and chemical characteristics, providing an opportunity for their application in various areas of technology and science [5–8]. MOFs are very versatile – a modular structure, consisting of inorganic nodes connected between organic linkers, allows carrying out purposeful designing of material with the  specific functionality, predictable morphology and chemical properties. Synthesis capability of hybrid materials with predictable properties on the basis of MOF distinguishes them from classical solid-state materials, in particular from traditional porous materials, such as mesoporous silicon dioxide and activated carbon.

In the present day, the most attention of scientific investigations is put on the synthesis and structural characterisation of new MOFs, analysis of their physical and chemical properties and post-synthetic modification. MOFs can be used in gas storage and separation [9–12], serve as catalysts [13–17] and lu minescent materials [18–21]. However, today only about 14 MOFs are commercially available [22], mostly produced by Baden Aniline and Soda Factory (BASF), Germany. Because of their limited availability, MOFs could not force out usual adsorbents, catalysts and other materials yet.

In the present study, metal organic frameworks Y_1-xLn³⁺_x(BTC)(DMF)₂(H₂O) and Y0.8-xYb_0.2Ln³⁺_x(BTC) (DMF)₂(H₂O), which structure included lanthanide ions Ln³⁺ (Er³⁺, Ho³⁺, Tm³⁺, Yb³⁺, Nd³), possibly suitable for monitoring the upconversion phenomenon were synthesised and investigated.

EXPERIMENTAL

All reagents used in the synthesis were used without further purification. Benzene-1,3,5-tricarbo-xylic acid (BTC) was purchased from Glentham Life Sciences. Methanol (MeOH), N,N-dimethyl-formamide (DMF) and sodium acetate trihydrate (NaOAc·3H₂O) were purchased from Chempur. Yttrium (III) nitrate hexahydrate (Y(NO₃)₃·6H₂O) (99.9% pure) was purchased from Alfa Aesar. Other salts of lanthanides, ytterbium (III) nitrate pentahy-drate (Yb(NO₃)₃·5H₂O), thulium (III) nitrate pen-tahydrate (Tm(NO₃)₃·5H₂O), erbium (III) nitrate pentahydrate (Er(NO₃)₃·5H₂O), holmium (III) nitrate pentahydrate (Ho(NO₃)₃·5H₂O) and neodymium (III) nitrate hexahydrate (Nd(NO₃)₃·6H₂O), all 99.9% pure, were purchased from Sigma-Aldrich^®.

X-ray diffraction (XRD) measurements of all synthesised compounds were recorded in the 2θ range between 5–70° with a Rigaku Miniflex II diffractom- eter (Cu Kɑ radiation with a graphite monochroma-tor). Thermogravimetric analysis (TGA) of the samples was performed with a Perkin Elmer STA6000 thermal analyzer in air up to 900°C at 5–10°C min^–1. The morphology of the samples was investigated with scanning electron microscopes  –  Hitachi TM3000 and Hitachi SU-70. For the upconversion data analysis, emission spectra were recorded using an Edin-burgh Instruments FLS 980 spectrofluorimeter. Measurements were made in a wavelength range of 350 to 800 nm by recording the measurement step every 0.5 nm. The compounds were excited by laser radiation at a wavelength of 980 nm, and the maximum laser current intensity used was up to I = 1.29 A.

H₃BTC (1 mmol), NaOAc·3H₂O (1 mmol) and two different lanthanide nitrate hydrate salts with a  total molar amount of 1  mmol, Y(NO₃)₃·6H₂O (0.94  mmol) and Ln(NO₃)₃·6H₂O (0.06  mmol), where Ln = Er³⁺, Ho³⁺, Tm³⁺, Yb³⁺ and Nd³⁺, were added to a 100 ml Erlenmeyer flask. 60 ml of a mixture of DMF and distilled water (2:1, v/v) was poured to the  flask which was covered with aluminium foil, mixed well and heated at 60°C for 24 h. Then, the reaction mixture was cooled to room temperature. The mixture of DMF and water was carefully decanted without pouring off the  precipitate formed on the bottom. Methanol (50 mL) was then added and the solution was kept for another 24 h, while replacing MeOH once with a  new amount and carefully decanting the previous portion again. The  reaction product (precipitate) is separated by vacuum filtration and MeOH is used to wash the precipitate. The precipitate is then dried under vacuum for a couple of hours at 70–80°C. The data of obtained Y_1-xLn³⁺_x(BTC)(DMF)₂(H₂O) synthesis products are summarised in Table 1. The  theoretical stoichiometric formula of the obtained MOFs is Y_0.94Ln³⁺_0.06(BTC)(DMF)₂(H₂O), where Ln³⁺ was an exchangeable cation during each synthesis.

For the  synthesis of MOFs with two active luminescent centres BTC (1  mmol), NaOAc·3H₂O (1 mmol) and three different lanthanide nitrate hydrate salts with a  total molar amount of 1 mmol, Yb(NO₃)₃·5H₂O (0.20  mmol), Y(NO₃)₃·6H₂O (0.80 – x_1–5 mmol) and Ln(NO₃)₃·5H₂O (x_1–5 mmol, where x_1–5 = 0.005, 0.01, 0, 02, 0.04 and 0.08 mmol), where Ln³⁺  =  Er³⁺, Ho³⁺ and Tm³⁺, were used. The rest of the synthesis procedure was performed as described above. The results of these syntheses are presented in Table 2. The theoretical stoichiometric formula of the  resulting MOFs corresponds to Y_0.8-xYb_0.2Ln³⁺_x(BTC)(DMF)₂(H₂O), where Ln³⁺ was an exchangeable cation with a variable molar ratio during each synthesis.

RESULTS AND DISCUSSION

The XRD patterns of YLnMOFs are presented in Fig. 1. Thestructure of synthesised YErMOF was confirmed by comparing the  obtained experimental XRD with the  results presented by Zhang et al. [23]. Following a successful repeated synthesis, monophasic completely new YLnMOFs, in which Er³⁺ was replaced by Ho³⁺, Tm³⁺, Yb³⁺ and Nd³⁺, were synthesised (see Fig. 1). The low background noise observed in the XRD patterns confirms that the  newly synthesised compounds are of high phase purity and have good crystallinity. It should be noted that regardless of the lanthanide cation being doped, the intensity and width of diffraction peaks do not change significantly. This can be explained by the fact that the structure of YMOF is doped with a small amount of Ln³⁺. In addition, the  difference of ionic radii of the  lanthanides is negligible [24].

No significant background noise is also seen in the XRD patterns of YYbLnMOFs-2 (Fig. 2). This again shows that the  new MOFs possess a  high phase purity and good crystallinity. Also, no big differences between the  XRD patterns of MOFs doped with different cations are seen. However, the amount of Yb³⁺ in the YYbLnMOFs-2 has effect on the crystallinity of synthesised compounds. Comparing the influence of Yb³⁺, which has a high concentration in YYbLnMOF compounds up to 20%, an obvious difference is observed between the 2Θ values of 15 and 25 degrees, where the intensity of the peaks increases significantly with increasing an amount of the Ln³⁺ (Fig. 3).

The TG/DTG/DSC curves recorded for the synthesised MOFs are presented in Fig. 4. The  mass losses are clearly indicated in the TG/DTG curves. As seen, with increasing temperature up to 350°C three 3 identifiable mass losses could be deter-mined. The first mass loss, visible from 64 to 95°C, is associated with evaporation of methanol (boiling point  =  64.7°C  [25]) and a methanol-water azeotropic mixture. Residual MeOH in the  MOF sample is possible due to changing the  solvents during the  purification step and the  mixture of DMF and water solvents. The  second mass loss can be attributed to the evaporation of DMF (boiling temp. = 153.0°C [26]), which takes place over a  wider temperature range starting at ~150°C. The last mass loss in the specified range up to 350°C represents degradation of the  whole MOF structure (BTC melting point  =  374–376°C declared by the manufacturer Alfa Aesar [27]), it occurs in a temperature range of 300–360°C. All of these processes account for about 24–25% of the weight loss. Solvent evaporation accounts for MeOH up to 6.5% and DMF up to 7.5% mass loss. In terms of heat flow, the processes are slightly endothermic (up to a maximum of 15 mW of heat absorbed), which is characteristic of evaporation and melting.

The  most intensive mass loss in the  TG/DTG curves is observed at 510–514°C. About 42–43% of the material is lost during this stage. The metal organic framework of benzene-1,3,5-tricarboxylic acid completely decomposes. This is demonstrated by the strongly exothermic process (heat flow value about –265  mW)  –  thermal decarboxylation. Further decomposition products in this temperature range are likely to correspond to Ln₂O₃ and Ln₂(CO₃)₃ [28, 29]. The remaining ~31–33% of the material consists of inorganic lanthanide oxides and carbonates. MOF maintains thermal stability up to ~150°C when DMF evaporation begins. According to the literature [30], a consistent drop in the mass loss curve represents MOF decomposition.

The SEM data revealed (Fig. 5) that the  MOFs with two luminescent centres form rod-shaped crystalline structures with Er³⁺ and Ho³⁺, ranging in size about 1.0–2.0 μm. However, the MOF with Tm³⁺ (see Fig. 5c) is composed of rectangular crystals about 1.0 μm in size.

In the  MOF compounds with one luminescent centre (Y_1-xLn³⁺_x(BTC)(DMF)₂(H₂O)) the  shape of the crystals varies from cubic (with Er³⁺) to irregular rectangles (with Ho³⁺), but their size remains unchanged in a  range of 1.0–2.0  μm. The  MOF crystals with Tm³⁺ have an irregular microstructure, having mostly a rectangular shape and a size about 1.0  µm. The  surface microstructure of synthesised MOFs particles with Nd³⁺ and Yb³⁺ is very similar to the microstructure of MOF crystals with Tm³⁺.

Thus, the  microstructure of synthesised Y_1-xLn³⁺_x(BTC)(DMF)₂(H₂O) and Y0.8-xYb_0.2Ln³⁺_x (BTC)(DMF)₂(H₂O) MOFs is slightly dependent on the nature of a lanthanide element.

Based on the  literature review, synthesised MOFs would be consistent with the  GSA/ETU conversion mechanism. Emission spectra of 6% (optically inactive Y³⁺ ions filling the  rest of the  crystalline matrix) of Er³⁺, Ho³⁺, Tm³⁺, Nd³⁺ and Yb³⁺ MOFs compounds were recorded aſter excitation with 980 nm laser radiation. Of the five different lanthanide cations, only Er³⁺ containing Y_0.94Er_0.06(BTC)(DMF)₂(H₂O) MOF exhibited up-conversion properties (Fig. 6).

In the  emission spectrum of Y_0.94Er_0.06(BTC) (DMF)₂(H₂O) three distinct peaks of different types (triplet, doublet and triplet forms, respectively) can be identified. The  first two peaks correspond to a green emission (at 523 and 547 nm, the latter being the most intense), and the third peak represents a red emission (669 nm). Based on the literature data [3, 31], electron transitions ²H_11/2 → ⁴I_15/2,⁴S_3/2 → ⁴I15/2 and ⁴F_9/2  →  ⁴I_15/2, respectively, were observed. All emission peaks are quite narrow and correspond to the Er³⁺ specific emission. A very low intensity peak at 412 nm can also be observed in the emission spectrum. According to the literature [32], this may correspond to a ²H_9/2 transition.

Another peak of even lower intensity in this spectrum is at 701 nm (red emission). However, the intensity of this peak is very low and is not attributed to the characteristic emission of Er³⁺ ions.

Emission spectra of Yb³⁺, Ho³⁺, Tm³⁺ and Nd³⁺ metal organic frameworks yielded no informative data. It can be stated that the latter ions do not have favourable energy levels for successful upconversion under the single fluorescent centre GSA/ETU conversion mechanism. It can also be stated by analogy that the BTC anion is not suitable for upconversion. This carboxylic acid does not have a sufficient conjugated π system in the benzene ring along with a carboxylic acid moiety, which would allow BTC to act as a sen-sitizer in upconversion initiation by the  TTA-UC mechanism.

Following these results, it was decided to try to synthesise MOF compounds with two fluorescent centre that could correspond to a GSA/ETU type conversion mechanism. In this case, the most effective lanthanide pairs discussed in the literature from the activator-sensitizer were selected: Er³⁺–Yb³⁺, Ho³⁺–Yb³⁺ and Tm³⁺–Yb³⁺ (Y³⁺ ions were filling the rest crystalline matrix). The amount of Yb³⁺ ions was maintained in two fluorescent centre MOFs at 0.2 mmol, and acted as a sensitizer in these compounds. To estimate which concentration of activators gives the highest intensity in the  emission spectra, their concentrations were doubled from 0.005 to 0.08  mmol each time. A  total of five different concentrations of MOF compounds for each ion pair of two luminescent centres were investigated. Aſter the  excitation with 980  nm radiation, emission spectra were measured. However, an unusual phenomenon was observed. During the measurement by focusing the laser beam on the powder in the cuvette, MOFs of all three different ion pairs, regardless of the minimum or maximum (0.1–1 W) laser power intensity, burned to charring. In this way, it was not possible to determine the same laser power intensity and focus for different samples. Therefore, a comparison of the intensities of different activator concentrations could not be objective because of the different measurement conditions for each compound in the series.

No justification could be found in the literature for this charring phenomenon. The essential condition that varied between measurements at one (up to 6% concentration) and two luminescence centres (20% Yb³⁺ and 0.5–8% replaceable activator Ln³⁺) was an increase in the Yb³⁺ ion concentration of more than three times (compared to not charred Y_0.94Yb_0.06(BTC)(DMF)₂(H₂O)). Yb³⁺ ions effectively absorb 980 nm laser radiation. It can be hypothesised that, due to unknown processes, the concentration of this ion was too high in these MOF compounds. Tm³⁺–Yb³⁺ MOFs did not exhibit upconversion properties and did not emit a visible spectrum. However, the upconver-sion phenomenon was recorded using Er³⁺ and Ho³⁺ as an activator. The  emission spectrum of Y_0.76Yb_0.20Er_0.04(BTC)(DMF)₂(H₂O) is shown in Fig. 7.

The Y_0.76Yb_0.20Er_0.04(BTC)(DMF)₂(H₂O) emission spectrum displayed doublet (at 627 and 639  nm) and triplet (at 654, 668 and 682  nm) peaks. The  most intensive triplet peak coincided with the  most intensive peak of the Y_0.94Er_0.06(BTC)(DMF)₂(H₂O) triplet. Thus, in the  Y_0.76Yb_0.20Er_0.04(BTC)(DMF)₂(H₂O) case, this emission spectrum peak corresponds to the characteristic ⁴F_9/2  →  ⁴I15/2 electron transition. An intensive doublet in the  emission spectrum is not entirely characteristic of the Er³⁺ ion. According to the Dieke diagram [33], an erbium trivalent ion does not have an energetic level similar to that of ⁴F9/2, which could be attributed to the  intensive peak in the spectrum doublet (see Fig. 7). Therefore, the  relaxation jump of this peak may not occur to the unexcited state of ⁴I15/2 at the lowest energy level. This can be explained by following the Dieke diagram. The closest and higher energy than the  ⁴F_9/2 excited level is only the  ⁴S_3/2 level. However, the relaxation transition from this level to the unexcited state is a typical green emission. The  emission spectrum of Y_0.76Yb_0.20Er_0.04(BTC)(DMF)₂(H₂O) also did not display a  green emission in comparison to the  emission spectrum of Y_0.94Er_0.06(BTC)(DMF)₂(H₂O).

According to the  Dieke diagram, this transition would be difficult to attribute to a  specific transition towards an unexcited level, because in the 26–30·10³ cm^–1 energy range a trivalent Ho³⁺ ion has many similar energy levels. The remaining peaks of Y_0.79Yb_0.20 Ho_0.01(BTC)(DMF)₂(H₂O) emission are located at 594 and 615 nm (orange), as well as at the 640 and 652 nm peak (red). The latter corresponds to a characteristic (mentioned in the  literature review at 647  nm) electron transition ⁵I₄ → ⁵I₈. According to the Dieke diagram [33], the  640  nm peak corresponds to the  transition from the excited state of ⁵F₅ to ⁵I₈.

In conclusion of this study, XRD analysis results displayed the  monophasic Y_1-xLn³⁺_x(BTC) (DMF)₂(H₂O) MOFs of one luminescence centre (Ln³⁺   =  Yb³⁺, Ho³⁺, Tm³⁺ and Nd³⁺) and Y0.8-x Yb_0.2Ln³⁺_x(BTC)(DMF)₂(H₂O) MOFs of two luminescence centres (Ln³⁺ = Er³⁺, Ho³⁺ and Tm³⁺) were successfully synthesised using a simple crystallisation method. The obtained MOFs were found to be thermally stable up to 150°C. The upconver-sion properties of the synthesised MOFs were investigated. It has been found that in the case of one luminescent centre, Er³⁺ can be used as an activator for successful upconversion, and in the case of two luminescent centers, Yb³⁺–Er³⁺ and Yb³⁺–Ho³⁺ can be used as sensitizer-activator pairs. It was concluded that Yb³⁺, Ho³⁺, Tm³⁺ and Nd³⁺ cations are unsuitable as activators for the  upconversion for one luminescent centre, and the  Yb³⁺–Tm³⁺ pair is unsuitable for two luminescent centres of MOF type. It has been also shown that the benzene-1,3,5-tricarboxylic acid (BTC) anion cannot act as a up-conversion sensitizer for the TTA-UC mechanism.

ACKNOWLEDGEMENTS

This work was supported by a  Research Grant NEGEMAT (No. S-MIP-19-59) from the Research Council of Lithuania.

Received 15 March 2021

Accepted 22 April 2021