Study of the Elemental Composition of Gadolinium–Aluminum Garnets—Obtaining Predictable Optical Properties

Abstract

For the first time, inductively coupled plasma mass spectrometry (ICP-MS) was developed for determining the target elemental composition of gadolinium–aluminum garnets with the varying composition Gd_3–xCe_xSc_yAl_5–yO₁₂, where x = 0.01–0.16 and y = 0.25–1.75. This fact has a fundamental importance for obtaining optical ceramics with predictable properties. Using the proposed acid mixture and temperature-time program, microwave digestion of these materials and complete transfer of the sample’s components into solution were possible. Moreover, we estimated the influence of the matrix composition, sample introduction system and collision cell on the limits of determination (LOD) of impurity elements by ICP-MS (Mg, Si, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Y, La, Pr, Nd, Sm, Eu, Tb, Er, Ho, Tm, Yb, and Lu). It has been shown that the conditions of mass spectral analysis proposed in this work provide LOD of target analytes in the range of 1∙10⁻⁶–4.15∙10⁻³ wt.%. The accuracy of the obtained results has been confirmed by the added-found method and by analyzing samples with known chemical composition. The standard deviation of repeatability (S_r) of the developed technique lies in the range from 1 to 6%. The developed analysis method is characterized by sensitivity, robustness and multi-elementality. It has application potential for other optical and ceramic materials of similar composition.

1. Introduction

Gadolinium–aluminum garnets (Gd₃Al₅O₁₂) are among the most promising functional materials due to their possible application as scintillators, solid-state lasers, LED lamps and others [1,2,3,4]. Most of the published research on Gd₃Al₅O₁₂ based materials focuses on the effect of alloying additives and production techniques on their optical properties [1]. Rare earth elements (REE), namely Sc, Ce, Sm, Gd, Tb, Dy and Er, frequently serve as alloying components [3,4,5,6,7,8,9,10].

Gadolinium–aluminum garnets doped with scandium and cerium (Gd_3–xCe_xSc_yAl_5–yO₁₂, where х = 0.01–0.16 and у = 0.25–1.75) can be used as X-ray luminescent composites and therefore, attract significant interest in the research community [9,10]. The simultaneous introduction of scandium and cerium into their structure increases the brightness of luminescence due to the preservation of a high quantum yield of internal luminescence [11]. To obtain the required optical properties during the development and production of doped gadolinium–aluminum garnets, it is of great importance to accurately determine the contents of alloying components and the impurity composition of the initial compounds, intermediates and final products [3,4,5,6,7,8,9,10]. In this regard, the development of multi-element, accurate and sensitive methods for analyzing these materials is a pressing task. The target impurity analytes are Mg, Si, Ti, V, Cr, Mn, Fe, Co, Ni, Cu and REE (Y, La, Pr, Nd, Sm, Eu, Tb, Er, Ho, Tm, Yb, and Lu). The list of target analytes is justified by the possibility of contamination of the final product with the above-mentioned elements at different stages of its production and the possible influence of these elements on the optical properties of the final material [1,11].

Among the analytical methods used to control the elemental composition of functional materials, including those based on REE, inductively coupled plasma mass spectrometry (ICP-MS) stands out as a highly promising technique. The advantages of this method include high sensitivity, a wide range of determined contents, multi-element properties and the possibility of using universal standard solutions to construct calibration curves and verify the accuracy and precision of the obtained results. However, to obtain reliable results when analyzing materials with complex compositions, such as gadolinium–aluminum garnets, it is necessary to study the influence of the analytical conditions on the signal intensity and selectivity, as well as the limits of determination (LOD) of the target elements [12,13,14,15,16,17,18,19,20,21,22,23]. This is due to the method limitations—the matrix effect and spectral interferences of various types, which affect the signal intensity and its selectivity of the target elements. The matrix effect is manifested in suppression or enhancement of the signal intensity of the analyte in the presence of a matrix element compared to the signal intensity of the same ions in the absence of a matrix element. Spectral interferences are divided into isobaric interferences (⁴⁰Ca⁺ and ⁴⁰Ar⁺, ¹⁴²Ce⁺ and ¹⁴²Nd⁺), doubly charged ion interferences (¹⁴²Ce⁺⁺ and ⁷¹Ga⁺) and polyatomic ions, formed from Ar, solvent and matrix elements [12].

This article investigated the ICP-MS method, aiming to expand the range of determined analytes and overcome the limitations associated with the complex composition of the gadolinium–aluminum garnet matrix.

2. Materials and Methods

The objects of study were samples of ceramics based on gadolinium–aluminum garnet doped with scandium and cerium of the composition Gd_3-xCe_xSc_yAl_5-yO₁₂, where x = 0.01–0.16 and y = 0.25–1.75. Samples were provided by the A.M. Prokhorov General Physics Institute of the Russian Academy of Sciences (GPI RAS).

The following acids in the given concentrations were used for microwave digestion of the samples: high purity nitric acid HNO₃ (nitric acid—70%), high purity hydrochloric acid HCl (hydrogen chloride—35–38%), high purity sulfuric acid H₂SO₄ (sulfuric acid—93.5–95.6%) and high purity hydrofluoric acid HF (hydrofluoric acid—46–49%). All dissolutions and dilutions were conducted using deionized water with a resistivity of 18.2 MΩ∙cm at 25 °C.

Aqueous calibration solutions were prepared from multi-element standards (High-Purity Standards, Charleston, SC, USA) by serial dilution to different volumes with a 2% HNO₃ mixture. Calibration solutions for ICP-MS were prepared in a concentration range of 1–100 μg/L. The influence of the acid system used for microwave digestion of samples was taken into account in a blank experiment.

The microwave digestion of samples was carried out using a MARS 6 laboratory system (CEM Corp., Matthews, NC, USA). Easy-Prep iWave 30 mL vessels were used for the experiment. These vessels allow the pressure to be maintained up to 100 atm and carried out with a wider range of acid ratios (aqueous solutions, concentrated acids) and temperatures (up to 300 °C). Inductively coupled plasma optical emission spectrometry (ICP-OES) was used to control the recovery of the elements. The determination of the matrix and alloying components in the solutions obtained during the development of the decomposition method was carried out using a Thermo Scientific iCAP PRO XP spectrometer (Thermo Scientific, Waltham, MA, USA) with a vertical torch, a purged Eschelle polychromator and a Charge Injector Device array detector. The following analytical lines were used during the analysis: Al—396.152 nm, Sc—335.373 и 188.060 nm, Ce—446.021 nm, Gd—369.619 and 335.047 nm.

The study of the conditions of analysis of Gd_3-xCe_xSc_yAl_5-yO₁₂ for the determination of target analytes by the ICP-MS method was performed using a NexION^® 1000 mass spectrometer (PerkinElmer, Waltham, MA, USA) with a conical spray chamber cooled to 5 °C and a Meinhard glass nebulizer. Instrumental operating conditions are summarized in

Table 1. Instrumental operating conditions of the mass spectrometer.

Operating Conditions	Value
Forward power, W	1300
Coolant gas flow, L∙min−1	15
Auxiliary gas flow, L∙min−1	1.2
Nebulizer gas flow, L∙min−1	0.90
Sample flow rate, rpm	30
Sampling depth, relative units	0
Helium flow rate, mL∙min−1	6.5
Level of oxide ions, %	<2
Level of doubly charged ions, %	<1.5
Measurement mode	Peak hopping
Isotopes of elements to be determined, m/z	24/25Mg, 29Si, 45Sc, 47Ti, 51V, 52/53Cr, 55Mn, 59Co, 58/60Ni, 63/65Cu, 89Y, 139La, 140Ce, 141Pr, 143/145Nd, 147/149Sm, 151Eu, 159Tb, 162/163Dy, 165Ho, 166/167Er, 169Tm, 171Yb, 175Lu
Internal standard	103Rh, 115In

.

3. Results and Discussion

3.1. Optimization of Microwave Digestion Conditions

The impact of the acid mixture composition and temperature-time programs on element recovery was investigated using samples of gadolinium–aluminum garnet with a certified chemical composition. Ceramic samples (Gd_2.73Ce_0.02Sc_1.0Al_4.25O₁₂) contained: Al—14.63 wt.%, Ce—0.36 wt.%, Sc—5.75 wt.%, Gd—54.72 wt.%.

The first step was to investigate the influence of the concentration and ratio of inorganic acids on the degree of digestion of the main and alloying components. Experiments on the selection of the acid ratio were carried out at a hold temperature of 170 °C and a hold time of 60 min. In addition to the acids under investigation, 5 mL of deionized water was added to each autoclave. The following acids were selected for the series of experiments: HNO₃, HCl and H₂SO₄. The results are accurately presented in Figure 1.

The results demonstrated that an increase in the volume of HNO₃ and HCl led to the digestion of Gd, Al and Sc into solution, yet the recovery of Ce did not exceed 20 relative % (Figure 1). At the same time, HNO₃ promoted better dissolution of Al, Sc and Gd. An increase in the volume of H₂SO₄ resulted in an enhanced degree of digestion of Al and Sc: however, the recovery of Gd did not exceed 70 relative %, and Ce of 10 relative % (Figure 1). It is known that the use of micro-quantities of HF can improve the recovery of Ce [16]. Based on this, a series of experiments were conducted to select the optimal volume of HF. The studies were carried out using a solution of HNO₃ (5 mL H₂O, 4 mL HNO₃). It has been established that the optimal volume of HF is 0.025 mL. A smaller amount of HF resulted in incomplete digestion of Ce, while the increase in volume reduced the solubility of Gd and Sc. During this study, it was found that the complete dissolution of the samples occurred in the system: 5 mL H₂O, 4 mL HNO₃ and 0.025 mL HF.

The second stage of the study examined the impact of temperature-time programs on the completeness of dissolution of gadolinium–aluminum garnet (Figure 2). The obtained results showed that the complete digestion of all components of the sample occurred in no less than 60 min at a temperature of 200 °C. Additionally, the effect of the holding temperature on reducing the dissolution time was studied. The obtained results demonstrated that at a temperature of 250 °C, the complete digestion of all elements occurred within 30 min (Figure 2).

3.2. ICP-MS Determination of Elements

The ICP-MS analysis of functional materials based on rare earth metals, including gadolinium–aluminum garnet, is affected by a matrix effect and a wide number of spectral interferences, which lead to an increase in the LOD of impurity elements [22].

The matrix effect is expressed as the suppression of the signal intensity of the determined elements with increasing concentration of the matrix element [22,23]. We simulated the composition of the material under study (Gd_2.73Ce_0.02Sc_1.0Al_4.25O₁₂) and showed (Figure 3) the dependence of the signal intensity of the target elements characterizing the mass scale (⁹Be, ²⁵Mg, ⁶⁵Cu, ¹³⁷Ba, ²⁰³Tl, ²⁰⁷Pb, and ²³²Th) on the concentration of the main elements. The magnitude of the matrix effect was calculated as I_i/I₀, where I_i is the signal intensity of the analyte isotope in a solution with matrix elements (0–500 μg/mL matrix), and I₀ is the signal intensity of the analyte isotope in a pure nitric acid solution.

When two elements are introduced as internal standards, a noticeable decrease in signal intensity (above 10 relative %) is observed for solutions containing matrix elements at a concentration of 400 mg/L and above (Figure 3). This allows for the analysis of more concentrated solutions (up to 500 mg/L). A more pronounced matrix effect is observed for light elements (⁹Be, ²⁵Mg, and ⁶⁵Cu). This can be explained by a number of processes, such as the collision of analyte ions with matrix elements in the region of supersonic expansion and the space charge effect in the optical system, which leads to the defocusing of the ion beam and a decrease in the overall sensitivity of the spectrometer [24].

Increasing the robustness of the mass spectrometer to the content of the main components in the sample can be achieved by varying the parameters of the sample inlet system [25] (Figure 4).

As can be seen in Figure 4, increasing the nebulizer gas flow from 0.6 to 0.90 L∙min⁻¹ lead to an increase in the I_i/I₀ ratio. With a further increase in the gas flow rate through the atomizer, the value of I_i/I₀ ratio decreased, and the level of doubly charged ions ¹³⁸Ba⁺⁺/¹³⁸Ba⁺ increased. The minimal matrix effect and the maximum analytical signals for most elements were obtained with a nebulizer flow of 0.90 L∙min⁻¹ (Figure 4). The level of doubly charged ions did not exceed 1.5% (Figure 4).

The conducted study showed that changing the plasma sampling depth led to a decrease in the intensity of the useful signal of the determined elements, while changing the sample flow rate did not affect the magnitude of the matrix effect. Based on this, it was decided to use standard values of these parameters—0 relative units (axis z) for the plasma sampling depth and 30 rpm for the sample flow rate.

To obtain correct results in the ICP-MS analysis of gadolinium–aluminum garnet ceramics, it is necessary to take into account the influence of polyatomic ions formed from macro components, solvent elements and argon on the signal intensity of target analytes. A study to investigate and quantitatively characterize the influence of polyatomic ions was carried out using model solutions containing 500 mg/L of matrix elements. The first step was to study the spectral interference in the standard spectrometer mode. The obtained results showed that during the determination of Ni, Eu, Tb, Ho, Dy, and Er, it was possible to select isotopes without spectral interference from macro components or with a slight increase in LOD (^58/60Ni, ¹⁵¹Eu, ¹⁵⁹Tb, ^162/163Dy, ¹⁶⁵Ho, and ^166/167Er). This allowed the determination of these elements with LODs from n∙10⁻⁶–n∙10⁻⁴ wt.%. The greatest influence on LOD was exerted by oxide and hydroxide ions (ⁿGd¹⁶О⁺, ⁿGd¹⁶О¹Н⁺) formed from Gd (

Table 2. Polyatomic ions that interfere with the determination of target analytes in a solution containing 500 mg/L of main elements simulating the composition Gd_2.73Ce_0.02Sc_1.0Al_4.25O₁₂.

Element	Isotope	Polyatomic Ion	Apparent Concentration of the Element, μg/L
Ni	61Ni	45Sc16O+	108
	62Ni	45Sc16O1H+	7.80
Eu	153Eu	136Ce16O1H+	0.95
Tb	159Tb	142Ce16O1H+, 158Gd1H+	0.45
Dy	161Dy	160Gd1H+	0.70
	164Dy	152Gd12C+	0.07
Ho	165Ho	152Gd13C+	0.003
Er	166Er	152Gd14N+, 154Gd12C+	0.09
	167Er	152Gd15N+, 154Gd13C+	0.09
	168Er	152Gd16O+	2.50
Tm	169Tm	152Gd16O1H+	2.75
Yb	171Yb	154Gd16O1H+, 155Gd16O+	178
	172Yb	155Gd16O1H+, 156Gd16O+	174
	173Yb	156Gd16O1H+, 157Gd16O+	183
	174Yb	157Gd16O1H+, 158Gd16O+	169
Lu	175Lu	158Gd16O1H+	15.1

). This led to a significant increase in the apparent concentration of Tm, Yb and Lu elements (Table 2).

One of the most common techniques to reduce the influence of polyatomic ions on the LOD of analytes is the use of a collision cell with an inert gas—He (KED mode). Based on this, the second stage of the study examined the impact of the helium consumption rate on the apparent concentration of ⁿGd¹⁶O⁺, ⁿGd¹⁶O¹H⁺ ions during the determination of Tm, Yb and Lu in the KED mode. The studied solutions contained 500 mg/L of the main elements simulating the composition of Gd_2.73Ce_0.02Sc_1.0Al_4.25O₁₂.

As expected, the apparent concentration of ⁿGd¹⁶О⁺ and ⁿGd¹⁶О¹Н⁺ ions decreased as the helium flow rate increased (Figure 5).

This is due to the collision of polyatomic ions in the cell with He, their fragmentation (dissociation) and partial loss of kinetic energy, which leads to the rejection of ions by a potential barrier at the exit of the ion beam from the cell according to the principle of discrimination by kinetic energy [25]. Nevertheless, the signal intensity of the elements being determined decreased: at a helium flow rate—5 mL∙min⁻¹—by 7–8 times, at 6.5 mL/min—by 30–35 times, at 7.25 mL∙min⁻¹—by 75–80 times, depending on the analytes (Figure 5). The optimal helium flow rate was found to be 6.5 mL∙min⁻¹. The apparent concentration of ⁿGd¹⁶О⁺ and ⁿGd¹⁶О¹H⁺ ions decreased by an order of magnitude (Figure 5). The optimal conditions of the ICP-MS analysis selected during the study provide the LOD of target impurity elements at the level of n∙10⁻⁶–n∙10⁻³ wt.% (

Table 3. Limits of determination and results of analysis of gadolinium–aluminum garnet using the ICP-MS method; Х ± Δ, wt.% (P = 0.95, n = 3).

Element	LOD, wt.%	Certified Value, wt.%	Content, wt.%
Mg	2.0∙10−5	(1.45 ± 0.05)·10−3	(1.47 ± 0.06)·10−3
Sc	1.0∙10−5	5.75 ± 0.85	5.58 ± 0.20
Si	2.0·10−3	(2.50 ± 0.15)·10−3	(2.60 ± 0.15)·10−3
Ti	1.0∙10−4	<1.0∙10−4	(4.08 ± 0.20)·10−5
V	1.0∙10−4	<1.0∙10−4	<1.0∙10−4
Cr	4.0∙10−6	(2.30 ± 0.10)·10−4	(2.15 ± 0.15)·10−4
Mn	8.0∙10−6	(6.80 ± 0.30)·10−4	(6.74 ± 0.27)·10−4
Fe *	7.5∙10−5	(3.50 ± 0.20)·10−4	(3.58 ± 0.26)·10−4
Co	5.0∙10−6	(1.65 ± 0.10)·10−4	(1.62 ± 0.12)·10−4
Ni	1.0∙10−5	(2.80 ± 0.15)·10−4	(2.76 ± 0.12)·10−4
Cu	3.0∙10−6	(1.40 ± 0.05)·10−4	(1.37 ± 0.06)·10−4
Y	2.0∙10−6	(3.10 ± 0.10)·10−4	(3.06 ± 0.08)·10−4
La	3.0∙10−6	<1.0∙10−4	(1.02 ± 0.04)·10−5
Ce	5.0∙10−6	(3.60 ± 0.20)·10−1	(3.50 ± 0.19)·10−1
Pr	2.0∙10−6	<3.0∙10−4	(6.24 ± 0.16)·10−5
Nd	4.0∙10−6	<2.0∙10−4	(2.17 ± 0.08)·10−5
Sm	2.0∙10−6	<1.0∙10−4	(3.62 ± 0.09)·10−5
Eu	2.0∙10−6	(1.00 ± 0.04)·10−4	(1.13 ± 0.03)·10−5
Tb	1.5∙10−4	(3.40 ± 0.20)·10−4	(3.35 ± 0.16)·10−4
Dy	1.0∙10−6	<1.0∙10−4	<1.0∙10−6
Ho	1.0∙10−6	<1.0∙10−4	<1.0∙10−6
Er	2.5∙10−5	(4.70 ± 0.30)·10−4	(4.65 ± 0.26)·10−4
Tm *	7.5∙10−5	<1.0∙10−4	<7.5∙10−5
Yb *	4.15∙10−3	(8.00 ± 0.30)·10−3	(8.12 ± 0.40)·10−3
Lu *	3.6∙10−4	(5.08 ± 0.10)·10−3	(5.02 ± 0.12)·10−3

* The results obtained using a collision cell for the determination of isotopes subject to spectral interference from polyatomic oxygen ions are presented.

).

3.3. The Accuracy of the Obtained Results

The accuracy of the ICP-MS analysis of gadolinium–aluminum garnets was confirmed using the added-found method and analysis of a sample with known chemical composition (Table 3). For the added-found method, solutions containing 1.0, 10.0 and 50.0 μg/L of target impurity elements were used in the analysis (Mg, Si, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Y, La, Pr, Nd, Sm, Eu, Tb, Er, Ho, Tm, Yb, and Lu). The repeatability standard deviation (S_r) for the target analytes was 1–6%.

4. Conclusions

A novel method for microwave digestion of gadolinium–aluminum garnets doped with scandium and cerium has been developed. The proposed acid system contains 5 mL H₂O, 4 mL HNO₃, 0.025 mL HF; a temperature-time program suggests hold temperature of 250 °C and hold time of 30 min. This method allows for the complete transfer of all components of the sample into solution.

The conditions of ICP-MS analysis were studied and selected, specifically: nebulizer gas flow—0.90 L∙min⁻¹, sampling depth—0 conventional units, sample flow rate—30 rpm, helium flow rate—6.5 mL/min. The use of a collision cell resulted in a reduction of LOD of Tm, Yb and Lu by an order of magnitude. The proposed “robust conditions” and two internal standards ensured high sensitivity and accuracy of the obtained results when determining target analytes in solutions with matrix element concentrations of up to 500 mg/L.