The Oxidation State of Europium in Halide Glasses

Abstract

The luminescent properties of divalent europium ions can be exploited to produce storage phosphors for x-ray imaging applications. The relatively high cost and limited availability of divalent europium halides makes it desirable to synthesize them from the readily available trivalent salts. In this work, samples of pure EuCl₃ and fluoride glass melts doped with EuCl₃ were processed at 700-800 °C in an inert atmosphere furnace. The Eu oxidation state in the resulting materials was determined using fluorescence and Mössbauer spectroscopy. Heat treatment of pure EuCl₃ for 10 minutes at 710 °C resulted in a material comprising approximately equal amounts of Eu²⁺ and Eu³⁺. Glasses made using mixtures of EuCl₂ and EuCl₃ in the starting material contained both oxidation states. This paper describes the sample preparation and analysis and discusses the results in the context of chemical equilibria in the melts.

1. Introduction

Glass-ceramic transparent storage phosphor materials can provide high resolution x-ray imaging and they are candidates for use in next-generation computed radiography systems [1-3]. The materials are based on europium-doped fluorochlorozirconate glasses. These glasses are heat treated to nucleate orthorhombic barium chloride nanoparticles that associate with divalent europium ions. The base glasses are the well-known ZBLAN family [4-6]. The glass is based on zirconium, barium, aluminum, lanthanum, and sodium fluorides. The modified ZBLAN compositions used for storage phosphor plates have melting and glass transition temperatures of approximately 700 and 230 °C, respectively [7,8]. The differences in thermal behavior between standard ZBLAN and the compositions investigated in this work are largely attributed to the substitution of BaCl₂ for BaF₂. The glass ceramic formed by heat treatment slightly above the glass transition temperature can convert ionizing radiation into stable electron-hole pairs that provide the ability to store x-ray images. The images can be developed by an exposure to light that activates recombination of the electron-hole pairs and leads to photostimulated luminescent (PSL) emission.

The current laboratory scale materials compare favorably with digital x-ray imaging technology [8]. Work is in progress to further optimize the imaging performance and scale-up production of the materials [7,8]. An important challenge is the development of large uniform plates that can be produced in quantity at low cost. A major cost is the divalent europium salts that are required to obtain the PSL effect. In-situ reduction of the relatively inexpensive EuCl₃ is a potential way to introduce divalent europium ions into the glass.

MacFarlane et al. [9] reduced trivalent europium by using metal hydrides and adding hydrogen to the process atmosphere. Hydrogen reacts with the EuCl₃ to produce gaseous hydrogen chloride and EuCl₂. Unfortunately, melting in a hydrogen-bearing atmosphere can partially reduce ZrF₄ and form dark grey precipitates in the glass. Phebus, et al. [10] showed that while trivalent europium was practically all reduced in a 4% hydrogen in nitrogen atmosphere, the resulting glass was of poor quality.

Coey, et al. [11] showed that fluoride glasses made using 10-20 molar % EuF₂ contained a significant proportion of EuF₃ after processing. Mössbauer spectra had a broad peak centered at -14 mm/s indicating spin relaxation of the Eu²⁺ ions. The Eu³⁺ produced a much narrower peak environment at 0 mm/s. Ball, et al. [12] also used Mössbauer spectroscopy to investigate thermal reduction of EuCl₃ processed in a vacuum furnace. They discovered that a mixture of EuCl₃, a chlorine deficient phase ~EuCl_2.8, and EuCl₂ was formed by treatment at 300 °C.

Glassner [13] compiled thermodynamic data for a variety of metal halides including the components of ZBLAN glass but not for europium dihalides. The data for zirconium shows that at all temperatures, ZrF₃ is the most stable fluoride. The thermodynamic stability of ZrF₃ in the glasses has not confirmed. However, MacFarlane et al. [9] processed melts in hydrogen-bearing atmospheres and found black precipitates in the glass that were attributed to reduced zirconium species. The precise mechanism of formation of these preciptitates is not yet understood. Broer [6] observed black precipitates in ZBLAN glasses and attributed them to partially reduced zirconium halides. Rard [14] reviewed the properties and thermodynamics of rare-earth fluorides and reported reduction of Eu³⁺ on heating. Massot, et al. [15] studied reduction of Eu³⁺ in molten salts by using cyclic voltammetry. They reported an equilibrium constant of 0.81 for [Eu²⁺]/[Eu³⁺] at 800 °C. This finding is consistent with Rard’s assessment and suggests that thermal reduction of EuCl₃ is a feasible route to synthesizing the divalent salt.

In the present study, changes in the oxidation state that occur when europium halides are heated and melted either as pure salts or as part of a ZBLAN glass melt batch were investigated. Samples were made and examined using Mössbauer spectroscopy to determine the oxidation state of Eu in the resulting products.

2. Method

2.1 Sample preparation

Reagents of at least 99.9 % cation purity were purchased in sealed bottles (Aldrich Chemical Company, Milwaukee, WI) that were opened inside a glove box. The glove box was filled with argon that was purified using a recirculating getter system. The amount of oxygen, p(O₂), and moisture, p(H₂O), in the glove box atmosphere was monitored continuously and remained at values below the detection limit of 0.1 ppm throughout the experiments.

The base glass composition (in mole %) comprised: 51.0 ZrF₄, 20.0 BaCl₂, 3.5 LaF₃, 3.0 AlF₃, 20.0 NaF, 0.5 InF₃. Europium was added as pure EuCl₂, EuCl₃ or a mixture of these two compounds. Twenty gram batches of the base glass were made with additions of EuCl₂:EuCl₃ in the ratios of 0, 0.2, 0.4, 0.6, 0.8 and 1.0 in the starting mixture. Batches containing 2.5 % each of EuCl₂ and EuCl₃ and one with 5% EuCl₂ were also made. The experimental compositions and the heating cycles used are summarized in Table I. All the compositions were melted in a two-step process that involved mixing and melting only the metal fluorides together at a temperature of 800 °C. The resulting melt was cooled to form a crystalline product. The chlorides were added to this and the mixture was remelted at a temperature of 750 °C. The two-step melting procedure has been shown to significantly decrease loss of the more volatile chlorides from the glass batch and produce a uniform composition glass [16]. The pure EuCl₃ sample was heat treated at a temperature of 710 °C for a period of 10 minutes in an inert atmosphere. Samples were weighed into a platinum crucible inside a glove box. The reagents were all powders and they were thoroughly mixed by stirring the weighed components together in the crucible with a spatula. The crucible was covered with a platinum lid and mounted on a translating holder. The assembly was moved into the hot zone of a 100 mm diameter tube furnace that was sealed onto one end of the glove box. The temperature was measured close to the location of the crucible with a thermocouple. The pure europium chloride sample was cooled in the crucible by removing it from the furnace and standing it on a brass plate in the glove box. The mixed ZBLAN composition melts were cast into a heated brass mold to form plates a few centimeters sidelength and approximately 1 mm thick. Details of the glass casting are given in [7].

Table I.

Sample identification, molar percent divalent and trivalent europium used, processing temperature and processing time. Composition Z01 was pure EuCl₃, in Z02 through Z07, the europium salts were added to a base glass mixture comprising 51.0 ZrF₄, 20.0 BaCl₂, 3.5 LaF₃, 3.0 AlF₃, 20.0 NaF, 0.5 InF₃. Sample Z08 contained 53.0 ZrF₄.

Sample ID	Molar%EuCl2	Molar%EuCl3	Process Temp. (°C)	Process Time (min.)
Z01	-0-	100	710	10
Z02	-0-	2.0	Zr, La, Al, Na and In fluorides melted at 800 °C, then Ba and Eu chlorides added and mixture remelted at 750 °C	3.75 hours total heating time
Z03	0.4	1.6
Z04	0.8	1.2
Z05	1.2	0.8
Z06	1.6	0.4
Z07	2.0	-0-
Z08	-0-	-0-
Z09	2.5	2.5
Z10	5.0	-0-

The thermal behavior of the glasses was measured using a Netzsch F3 Maia Differential Scanning Calorimeter (DSC). The DSC measurements provided the glass transition temperature and the onset temperatures for crystallization of the glass.

2.2 Mössbauer spectroscopy

Samples for Mössbauer spectroscopy were prepared as either thin glass plates formed by polishing 1 cm squares of glass to a thickness of ~150 μm, or from powders prepared inside a glove box. The thin plates were made by cutting a square of glass from a cast sheet, mounting it on an arbor and polishing it with 800 grit abrasive on a wheel. The powder samples were made by finely grinding material in an agate mortar inside a glove box. The powder was mixed with grease and uniformly coated onto plastic sample insert that was then mounted in the Mössbauer instrument. Thin glass plates were placed in polythene bags and mounted on a lead aperture in the Mössbauer instrument. The relatively large Zr content of the glasses resulted in a significant absorption of gamma rays. In order to increase sensitivity, measurements were made on glasses that contained a total of 5 mole % europium.

The Eu Mössbauer spectra were acquired for periods of 12-160 hours using a SEE Co (Minneapolis, MN) spectrometer. Measurements were made on some samples immediately after they were removed from the glove box and at later times after they had been exposed to the air. The data were acquired by computer and analyzed using software designed for fitting Mössbauer spectra using simple Lorentzian functions. The concentrations of the species were calculated by integrating the areas under the peaks in the Mössbauer spectra. The estimated error of the isomer shift is ± 2%, the estimated error for area is ± 3%. All of the Mössbauer spectra were acquired at ambient temperature. The spectrometer is being fitted with a cryostat and future measurements will be made at low temperatures. Low temperature measurements will help to understand the role of Eu²⁺ in the storage process since they can reveal the magnetic hyperfine structure of the ion [17].

Fluorescence spectra were measured for selected glass samples using a spectrofluorometer (Horiba Jobin Yvon FluoroLog-3). Glass samples were mounted in the spectrometer and fluorescence was excited by pumping them with a xenon lamp (450 W) at wavelengths of 280 and 394 nm. The resulting fluorescent spectra were detected with a cooled photomultiplier (Hamamatsu R928P). Emission at 400 and 613 nm results from excitation of Eu²⁺ and Eu³⁺, respectively. The ratio of these two intensities provides a relative measure of the activity of the two species in the host.

3. Results and Discussion

Heating pure EuCl₃ resulted in a color change from white to dark grey. The dark color of the heat treated sample suggests that it contains defects that result in strong absorption over the whole visible spectral range. Samples of the ZBLAN compositions that contained Eu additions were either glassy or white crystalline material. We note that for purposes of discussing the thermodynamics, we have considered only stoichiometric compounds. It is likely that the melts and resulting glasses are comprised of compounds with somewhat uncertain stoichiometries.

The mass of pure a EuCl₃ sample that was heated to 710 °C for 10 minutes decreased by 39.9 mass %. Loss of one chlorine atom per molecule of EuCl₃ would result in a mass loss of only 13.7 mass %. The additional loss is attributed to vaporization of the sample at the process temperature slightly above the melting point of EuCl₃, approximately 625 °C [9]. The prior work on the glasses has shown that evaporation of chlorides occurs during processing. Typical chloride loss from glass batches were 2-5 mass % during melting. The higher rate of evaporation from the pure chloride is consistent with its higher activity in the pure compound.

3.1 Mössbauer and Fluorescence Spectra

Mössbauer spectra for heat treated EuCl₃ samples are shown in Fig. 1 The spectra of pure EuCl₂ and EuCl₃ can easily be distinguished from their isomer shifts: around -13 mm/s for Eu²⁺ and close to 0 mm/s for Eu^3+.. The data for heat treated EuCl₃ immediately after heating to 710 °C for 10 minutes (Fig. 1(b)) showed approximately equal proportions of Eu²⁺ and Eu³⁺ when it was measured immediately after removal from the glove box. The primary phase equilibrium that can change in the oxidation state of Eu is:

$${\mathrm{EuCl}}_3={\mathrm{EuCl}}_2+1∕2{\mathrm{Cl}}_2$$ {1}

When the sample was exposed to air, the concentration of Eu²⁺ decreased and Eu³⁺ increased. Fig. 1(c) shows the spectrum after 10 weeks exposure to air, where the sample contains mostly Eu₂O₃, as shown by its isomer shift of about 1.0 mm/s. The decrease in Eu²⁺ is attributed to oxidation of the finely powdered material according to the reaction:

$$2{\mathrm{EuCl}}_2+2{\mathrm{H}}_2\mathrm{O}+1∕2{\mathrm{O}}_2={\mathrm{Eu}}_2{\mathrm{O}}_3+4\mathrm{HCl}$$ {2}

Fig. 1.

Mössbauer spectra of: (a) pureEuCl₃, (b) pure EuCl₂, (c) finely powdered EuCl₃ immediately after heating to 710 °C for 10 minutes: the relative areas of the Eu²⁺ and Eu³⁺ signals are 44:56 (d) after 10 weeks exposure to air the sample is mostly Eu₂O₃, (e) pure Eu₂O₃.

Fluorochloride glases are more sentitive to environmental damage than their pure fluoride counterparts. The hydrolysis reaction can also proceed in the glass plates but it will be diffusion limited so the oxidation of Eu²⁺ will only occur near the surface of the plate. Using a diffusion coefficient of 10⁻¹⁶ cm²/s at 25 °C for moisture/oxygen in the glass, internal oxidation would proceed at a rate of approximately10 µm/year. In an x-ray image plate application, the glass would be coated with dielectric layers and encapsulated in a polymer that would further reduce the rate of oxidation. Commercial x-ray image plates that use Eu²⁺-doped barium halide particles in a polymer matrix do not exhibit significant time dependent loss of performance [18]. Samples exhibit consistent X-ray storage performance over a period of more than a year of storage in air.

Mössbauer spectra of Eu in ZBLAN glass are shown in Fig. 2. Fig. 2(a)is for a glass made with5 mole % EuCl_2. The ratio of the areas under the peaks for Eu²⁺ and Eu³⁺ is 78:22. Fig 2(b) is for a glass containing 2.5 mole % EuCl₂ and 2.5 mole % EuCl₃. The ratio of the areas under the peaks for Eu²⁺ and Eu³⁺ is 37:63 indicating that 13% of the Eu²⁺ was oxidized during the melt processing.

Fig. 2.

Mössbauer spectra of ZBLAN samples made with (a) 5.0% EuCl₂ and (b) 2.5 % EuCl₂ and 2.5% EuCl₃.

Eu fluorescence spectra measured for a series of glasses made up with different initial concentrations of Eu²⁺ and Eu³⁺ are presented in Fig. 3 The ratio of Eu²⁺/Eu³⁺ emission increases monotonically with the increased ratio of these components in the glass batch. This result suggests that the ratio of divalent to trivalent Eu in the glass is principally determined by the composition of the starting material, not by redox reactions during melt processing.

Fig. 3.

Ratio of Eu²⁺ to Eu³⁺ fluorescence as a function of the EuCl₂ and EuCl₃ content in the starting materials used to make the glass. The total Eu content in the glass was maintained at 2 mol%. The as-made glass was heat treated at 290 °C prior to making the measurement.

3.2 Thermodynamics

The equilibrium constant, K, for the reaction {1}can be defined in terms of the free energy of the reaction, ΔG^Θ, at the process temperature T. This can be used to calculate the activities, a, of the components according to:

$$K=e^{-\frac{\Delta G_T^{ϴ}}{\mathit{RT}}}=\frac{a\left[{\mathrm{EuCl}}_2\right]{{}_p}^{1∕2}\left[{\mathrm{Cl}}_2\right]}{a\left[{\mathrm{Eucl}}_3\right]}$$ {3}

In the case of heating pure EuCl₃ in an inert atmosphere chlorine is removed from the system promoting reduction of the trichloride. The reported value of 0.81 for the [Eu²⁺]/[Eu³⁺] ratio at 800 °C [15] means that heating the pure trichloride would result in a equilibrium mixture containing approximately 45% Eu²⁺ and 55% Eu³⁺. The presence of the sub-stoichiometric EuCl_2.8 phase as observed by Ball et al. [12] may result in additional equilibria that could appear to decrease the Eu³⁺ concentration. The Mössbauer results are consistent with the known thermodynamics of EuCl₃ decomposition.

In the glass melt, more complex equilibria can occur between the components. Zirconium and indium in particular provide a source of fluorine due to reactions such as:

$${\mathrm{ZrF}}_4={\mathrm{ZrF}}_3+1∕2{\mathrm{F}}_2$$ {4}

$${\mathrm{ZrF}}_4={\mathrm{ZrF}}_2+{\mathrm{F}}_2$$ {5}

$${\mathrm{InF}}_3=\mathrm{InF}+{\mathrm{F}}_2$$ {6}

In the case of zirconium, ZrF₃ is the most stable fluoride at 700 °C [13]. The free energies of formation of the difluoride, trifuoride and tetrafluoride are -410, -418 and - 395 kJmol⁻¹, respectively, at 725 °C [12]. These free energy values yield an equilibrium constant of 15.9 at 725°C, meaning that at equilibrium the concentration of ZrF₃ would be ~16 times higher than ZrF₄. Since clear glasses are readily formed from the base ZBLAN composition [7-10] it appears that this reaction is suppressed in the glass melt by the addition of indium trifluoride.

The available fluorine can potentially oxidize EuCl₂ according to:

$${\mathrm{EuCl}}_2+1∕2{\mathrm{F}}_2={\mathrm{EuCl}}_2\mathrm{F}$$ {7}

The results of both the Mössbauer measurements shown in Figs. 1d and e and the fluorescence data shown in Fig. 3 suggest that the Eu²⁺/Eu³⁺ ratio is determined by the starting composition of the glass rather than by melt reactions.

There are reports that the Eu oxidation states in the melt can be controlled using reducing agents. MacFarlane et al. [9] used a hydrogen-bearing atmosphere or metal hydrides to reduce Eu³⁺ in a melt. Phebus et al. [10] confirmed the reduction by hydrides but found that the resulting glasses were of poor optical quality. While chemical reduction is a feasible route, it is not suitable for production of optical quality glass plates since black particles attributed to reduced zirconium speicies are formed.

3.3 Glass Formation

The base fluoride glass that does not contain any europium is readily formed into plates that show no evidence of crystalline regions. The addition of europium chloride makes production of quality glass plates significantly more difficult. Coey et al. [11] reported that the Eu²⁺ acts as a network modifier. The Eu³⁺ ions were found to behave as a network former.

Results of thermal analysis of undoped glasses and those that contain Eu²⁺ and Eu³⁺ are shown in Fig. 4 Both the base glass and the composition containing Eu³⁺ show similar behavior. The onset of the glass transition occurs at ~210 °C and peaks at 230-240 °C. In the composition that contained Eu²⁺, the onset of the glass transition and the peak occur at lower temperatures. The peak height is approximately 25% smaller because the glass transition is aborted by the onset of crystallization in the Eu²⁺ doped glass. In the other glasses, the onset of crystallization is clearly separated from the glass transition regions (see inset Fig. 4)

Fig. 4.

DSC measurements made at a heating rate of 10 °C/min. on three glasses corresponding to Z02, Z07 and Z08. The inset shows an expanded view over the temperature range that includes crystallization of a barium chloride phase. The onset of crystallization results in a reduction of the peak height in the Eu²⁺ sample.

While the analysis is insufficient to determine a fragility index for the melts, it indicates substantial differences in the glass forming behavior of the composition that contain Eu²⁺ species. The effects of a low concentration of Eu²⁺ ions is more significant in the rather fragile ZBLAN melts than addition of modifier to a strong liquid. For example, in the case strong network former, silica, several % sodium ions are required to have an appreciable effect on the melt viscosity and glass forming behavior [19,20].

4. Conclusions

1. Heating pure EuCl₃ results in a product that is approximately 50% EuCl₂. This finding shows that in-situ formation of EuCl₂ is a feasible method of producing divalent europium from the trichoride. The finely divided material oxidizes to form Eu₂O₃ on exposure to moist air.

2. When ZBLAN glasses are made by a two-step melting process in which the europium is added to a premelted mixture of fluorides, the ratio of Eu²⁺ to Eu³⁺ in the glass remains essentially the same as in the starting materials. Exchange of fluorine between zirconium and europium appears to be a relatively slow process in the melt.

3. Based on the available thermodynamic data, at equilibrium, ZrF₄ present in the starting glass batch would be reduced to ZrF₃ during melting The formation of clear glasses suggests that the zirconium can be stabilized in the 4+ state in the glasses.

4. The addition of Eu²⁺ to the melt decreases the glass forming tendency. This effect is attributed to the divalent ions acting as network modifiers.