Multifunctional NaEu(WO₄)₂: defect-tuned red emission and acetone sensing at room temperature

Abstract

Rare-earth double tungstate NaEu(WO₄)₂ was synthesized via a trisodium citrate (Na₃cit)-assisted hydrothermal technique, followed by calcination, to promote crystallinity and detailed investigations on their crystal structures and luminescence properties. In this study, the structural evolution of our samples synthesized with different amounts of Na₃cit was studied by employing X-ray diffraction, Rietveld refinement, Fourier transform infrared and Raman spectroscopy techniques. It was found that NaEu(WO₄)₂ belongs to the scheelite family with Na and Eu atoms occupying the same sites and antisite defects deforming EuO₈ dodecahedra. The modulation of W–O, Eu–O and angle splitting in the presence of antisite defects was identified. From in-depth X-ray photoelectron spectroscopy, we validated the deformation of the EuO₈ dodecahedron due to the presence of oxygen vacancies (V_Os), which originated from antisite defects. Herein, we show that the band gap of NaEu(WO₄)₂ is highly sensitive to defects; however, the ⁵D₀–⁷F₂ transition of Eu³⁺ at 615 nm with color coordinates (0.67, 0.33) is very robust, making NaEu(WO₄)₂ a suitable red phosphor material for near UV-type light-emitting devices (LEDs). We also identified that V_Os present in the EuO₈ dodecahedron act as active sites for acetone sensing (∼68% response to 100 ppm) with a response and recovery time of ∼3.3/10 s at room temperature, suggesting the potency of NaEu(WO₄)₂ as a multifunctional material with applications in LEDs and acetone sensors. In order to validate our experimental observations theoretically, we calculated the band structure and density of states of bare and antisite defects containing NaEu(WO₄)₂ using ab initio density functional theory and identified the sensing mechanism. We believe that our studies will be helpful in introducing new multifunctional applications of NaEu(WO₄)₂, while theoretical calculations will provide new electronic insights that may be used to understand the features of other double rare-earth tungstate materials.

---

1. Introduction

Motivated by multifunctional materials in nature such as the human skin and bird feathers, the design and development of materials having multiple functionalities has become an upcoming field for various sustainable applications and has gained the attention of researchers in the recent past. Owing to their multiple features, scheelite-type alkali rare-earth double tungstates (A^ILn(WO₄)₂, A = alkali metal, Ln = lanthanide rare-earth metal) with C_4h symmetry and the I4₁/a space group have emerged as promising materials in many technologically diverse areas such as catalysis, energy storage, scintillators, and laser hosts. Because of the strong covalence of the W–O bond, scheelites act as good hosts that promote high solubility for lanthanide (Ln) doping, where they exhibit narrow emission characteristics of high spectral purity due to the strong f–f or f–d transition of Ln³⁺ ions.¹ It is worth mentioning that the absorption coefficient of Ln³⁺ ions is usually very low, resulting in a weak emission signal.² However, tungstates, because of their strong absorption coefficient, facilitate energy transfer to Ln³⁺, and subsequently, the emission intensity gets enhanced.³ Because of this, Ln³⁺-doped tungstate scheelites show maximum performance in the field of solid-state lighting devices, including field emission displays, plasma display panels, liquid crystal displays, and light-emitting diodes (LEDs). These lighting systems have several advantages of biocompatibility, low power consumption, long lifetime, remarkably high brightness, etc., and hence have high market value. As different Ln³⁺ ions have different energy levels and accordingly produce different colors, the choice of Ln³⁺ is most important to tune the color output of the lighting devices as per requirements and applications.⁴ In general, the energy transfer between tungstate and Ln³⁺ ions significantly depends on the type of Ln³⁺ and host matrix, wherein efforts are mostly given to explore different tungstate hosts for improved energy transfer. In contrast to Ln³⁺-doped tungstates, self-contained tungstate–lanthanide hosts have been found to be highly efficient for this process, wherein scheelite-structured alkali metals, i.e., Ln-binary tungstate with the generic formula A^ILn(WO₄)₂, a representative of new scheelite materials with high density and low phonon threshold,⁵ have drawn worldwide attention. Much more improved emission has been detected from A^ILn(WO₄)₂, suggesting their suitability for several lighting systems including laser crystals and phosphors.⁶ It has been found that ‘A’ is commonly responsible for crystal field distortion around Ln³⁺ ions, relaxing the selection rules for f–f or f–d transitions, which enhance emission intensity.⁷ Among several Ln³⁺ ions such as Gd³⁺, Lu³⁺, Tm³⁺, and Dy³⁺, several works have been carried out with Eu³⁺-doped binary tungstate as the designing red phosphor due to its high absorption coefficient in UV/near UV region, which circumvents low color rendering index in comparison with other Eu³⁺-doped metal oxides/sulphides (e.g., ZnO, SnO₂, and ZnS) and scheelites (e.g., SrWO₄).^8–11 Nevertheless, there is still a need to improve the Eu³⁺-red phosphor, wherein researchers are actively involved. In this regard, it may be stated that few fundamental studies are likely to explore the structure and optical emissions from NaEu(WO₄)₂, indicating its potency as a red phosphor material. As an example, J. Huang et al. have investigated the effect of alkali metal ions on the local structure of AEu(WO₄)₂ (A = Li, Na, K), where they identified broader red emission at 614 nm due to an effective charge transfer excitation in KEu(WO₄)₂ in comparison with that of LiEu(WO₄)₂ and NaEu(WO₄)₂.¹² Neeraj et al. observed a significant improvement in the light output in NaY_0.5Eu_0.5(WO₄)(MoO₄) with reference to the commercial red phosphor Y₂O₃S:Eu³⁺.¹³ Commonly, the A^ILn(WO₄)₂ class of materials is synthesised by several techniques such as solid state reaction, hydrothermal, solvothermal, and microwave synthesis.¹⁴ Several researchers, including our group, have found that defects depending on the synthesis protocols play an important role in tuning the distortion around Ln³⁺ ions and accordingly, the selection rules get relaxed. Therefore, it may be stated that emission is significantly dependent on the synthesis methods, where some studies are going on actively.¹⁵ As an example, the use of water and ethylene glycol during the solvothermal synthesis of NaEu(WO₄)₂ significantly alters the Eu–O bond lengths, which in consequence could cause a blue shift of the charge transfer band,¹⁶ while Munirathnappa et al. tuned the defects by electron beam radiation and observed a significant improvement of the luminescence property of NaEu(WO₄)₂.¹⁷ A careful literature review reveals that few efforts were undertaken to generate various micro/nanostructures of NaEu(WO₄)₂ using different surfactants such as ethylenediaminetetraacetic acid (EDTA), cetyltrimethyl ammonium bromide (CTAB), trisodium citrate (Na₃cit), polyvinylpyrrolidone (PVP), and ethylene glycol (EG) and the morphological influences on the emission characteristics were studied. However, there is no report to illustrate the effect of surfactant on the defects, particularly on the distortion of the crystal field around Eu³⁺ ions and related emissions.

Based on our previous investigations¹⁸ and motivated by the fact that defects can be tuned by varying the surfactant during the hydrothermal process, an attempt has been made for the first time to examine the influence of Na₃cit on the defects and related deformation around the Eu³⁺ ions within NaEu(WO₄)₂. In this study, we have identified that Na₃cit plays a significant role to tune the antisite defects, which are subsequently counterbalanced by the oxygen vacancy (V_O). We have also emphasized the V_O-induced modification of the bond length and mixing of Eu 4f orbitals, which synergistically impact the emission properties and chromaticity coordinates. Most importantly, by varying the synthesis condition, we achieved the chromaticity coordinate (0.667, 0.333) from one of our synthesized samples, which perfectly matches with that of the National Television Standard Committee (NTSC)-prescribed chromaticity coordinate for red phosphor.

In the recent past, there has been an increasing interest in the detection of life-threatening toxic gases. Although techniques like as gas/liquid chromatography exist, because of their high cost, their wide application is limited. Herein, chemo-resistive-type gas sensors based on metal oxides have been viewed as one of the most economical, effective, and fast techniques for sensing applications. However, major issues still exist in their high sensitivity and high temperature (∼250–400 °C) operation.¹⁹ Hence, it is worthwhile to develop a room temperature sensor with low power consumption, safety, and long lifetime. Though sensors operating at higher temperature are largely reported, room temperature sensors are very rare. However, efforts are continuously underway to find out suitable materials for room-temperature sensors. Our previous investigations reveal that V_O often leads to the gas sensing property;^15,18 hence, we checked the gas sensing property of NaEu(WO₄)₂ and, most importantly, acetone sensing has been achieved at room temperature. Apart from the experimental investigations on the optical and gas sensing properties, ab initio density functional theory has been adopted to acquired knowledge of the band structure and partial density of states (PDOS) of NaEu(WO₄)₂ in order to understand antisite defect-induced changes in the optical and acetone sensing properties of NaEu(WO₄)₂. To the best of our knowledge, no such study exists for NaEu(WO₄)₂; hence, we believe that these experimental and theoretical findings will be helpful in realizing the multifunctional applications of NaEu(WO₄)₂.

2. Experimental section

2.1. Synthesis of NaEu(WO₄)₂

Sodium europium double tungstate (NaEu(WO₄)₂) was synthesized by a facile hydrothermal method. Initially, 1.00 mmol (0.338 g) of europium nitrate hydrate [Eu(NO₃)₃·H₂O, Alfa Aesar] and tri-sodium citrate dihydrate [Na₃cit, Merck] were dissolved in 50.00 mL DI solution and stirred for 20 minutes. Another 20.00 mL aqueous solution containing 2.00 mmol (0.660 g) sodium tungstate dihydrate [Na₂WO₄·2H₂O, Sigma Aldrich] was added dropwise to the previously prepared solution and stirred (30 minutes) for homogeneous mixing. Then, the mixture solution was transferred to a Teflon-lined stainless-steel autoclave of 100.00 mL capacity for hydrothermal reaction, which was carried out at 180 °C for 24 h. After the reaction, the temperature was cooled down and the powder sample was collected by means of centrifugation at 12 000 rpm, followed by washing with DI water and ethanol twice and drying at 70 °C for 24 h. The obtained powder was calcined at 800 °C under nitrogen (N₂) atmosphere for 5 h and the final product was obtained. In order to examine the effect of Na₃cit on the final product, we varied the Na₃cit (ca. 0.75, 0.50 and 0.25 mmol) amount during the hydrothermal reaction, keeping all other parameters same. Accordingly, they were named as NEWO_0.75, NEWO_0.50 and NEWO_0.25, respectively.

2.2. Characterization techniques

Structural analysis and phase confirmation were using the X-ray diffraction (XRD) patterns collected on a Rigaku (Japan) Ultima III powder diffractometer equipped with a Cu Kα radiation source (λ = 1.5406 Å). The Rietveld refinement was done with the help of FullProf Suite software to determine the structural and microstructural parameters. The background was fitted with linear interpolation, while the Pseudo-Voigt profile with asymmetry was used for the peak shape. The morphology was examined by field emission scanning electron microscopy (FESEM, S-4800, Hitachi Japan). Vibrational analysis was done using Fourier transform infrared spectroscopy (Shimadzu, IRPrestige-21) and Raman spectroscopy (alpha 300 Witec, 530 nm laser, 3 mW power and 2 m spot size). Elemental analysis was performed using X-ray photoelectron spectroscopy (PHI 5000 Versa Prob II, FEI Inc., Al Kα radiation, 1486 eV). The optical properties were investigated by a UV-vis (V-630, JASCO) and a photoluminescence spectrophotometer (RF-5301, Shimadzu). The measurement of luminescence lifetime was carried out with the help of the time-correlated single photon counting set up by Horiba Jobin-Yvon. The luminescence decay time was collected through a Hamamatsu MCP photomultiplier (R3809) and analysed using IBH DAS6 software.

2.3. Measurement of the gas-sensing properties of NaEu(WO₄)₂

Taguchi-type sensor modules were fabricated with a cylindrical alumina substrate wherein Pt wires and conducting gold paste were used for electroplating the alumina substrates. By mixing the as-synthesised materials with isopropyl alcohol, a slurry paste of desired consistency was prepared and the paste was drop-coated on the substrate and dried for 6 h to get rid of residue solvents. Ni–Cr wire, used as heating coils, was inserted inside the hollow of the alumina substrate. Then, the sensor modules, prepared by wielding the substrates on a six-pin socket, were used to evaluate the gas sensing properties in the presence of different volatile organic compounds (VOCs), namely, ammonia, acetone, ethanol, formalin, methanol, and isopropyl alcohol. The response was evaluated using the formulae where R_gas and R_air denote the resistance of the sensor in the absence and presence of the target gas, respectively. The response and recovery time of the sensors were determined from the time taken by the sensor to reach the 90% resistance change with respect to the equilibrium state resistance after exposing and removing the target gas.

2.4. Ab initio calculation using density functional theory

The first principles density functional theory (DFT) calculation was carried out in the plane wave pseudo-potential basis, as implemented in the Vienna ab initio simulation package the (VASP) using generalized gradient approximation (GGA) exchange–correlation potential. The cutoff energy of the plane wave basis was set to 500 eV, which was sufficient to achieve convergence, while 5 × 5 × 2 Monkhorst–Pack mesh was used for Brillouin zone sampling. The applied Hellmann–Feynman forces on each atom and EDIFF parameter were fixed to 0.015 eV Å⁻¹ and 10⁻⁶ eV for these calculations.

Furthermore, the adsorption energy of acetone on the slab surface was calculated to overcome the antisite defect-induced gas sensing operation. Herein, we considered a vacuum region of 15 Å to avoid interlayer interactions along the z-axis, while the 5 × 5 × 1 Monkhorst–Pack mesh and 500 eV cut off energy were employed to get converged free energy for the surface calculation. To understand the adsorption mechanism, isolated acetone molecules were modelled above the cleaved (0 0 4) surface (most stable surface with minimum free energy), keeping acetone at the centre of the unit cell using three configurations (i.e., vertical, parallel to the x-axis and parallel to the y-axis). The adsorption was determined from eqn (1)²⁰

E_ads = E_slab+gas − (E_slab + E_gas) (1)

where, E_slab+gas, E_slab and E_gas represent the total energy of the slab including gas, total energy of the slab and energy of the isolated gas, i.e., acetone. Within this definition, a negative binding energy is attributed to an exothermic process.

3. Results and discussion

3.1. Phase, crystal structure and micro-structural analyses of NaEu(WO₄)₂ by XRD and FESEM

The XRD patterns of all the samples (as shown in Fig. S1, ESI†) consist of strong, sharp diffraction peaks matched closely with tetragonal NaEu(WO₄)₂ (JCPDS no. 04-002-3849, space group I4₁/a).¹² The absence of any deleterious secondary peak in the diffraction patterns is imperative to point out the formation of the pure phase of the samples. Meanwhile, a crystallite size of ∼33–43 nm was obtained from the well-known Scherer's relation for all the samples, while the strain (ε), calculated using the relation where β and θ denote the full width at half maxima (FWHM) and Bragg's angle respectively, was calculated to be ∼3.40 × 10⁻³, 3.14 × 10⁻³, and 2.59 × 10⁻³ for NEWO_0.75, NEWO_0.50, and NEWO_0.25, respectively. It is noteworthy that Na₃cit does not have a significant effect on the particle size during the hydrothermal reaction; rather, it has significant impact on ‘ε’, which may be related to the structural distortion.

Rietveld refinement (shown in Fig. 1(a)–(c)), adopted to study long-range structural order–disorder, provides lattice parameters, unit cell volumes, and lengths of the links between (W–O)/(Na/Eu–O) of all the samples, as summarized in Table 1. Note that these parameters are very close to the data published in the literature^12,16 as well as goodness-of-fit (e.g., χ²), which indicates the reliability of the refinement process. The unit cell of NaEu(WO₄)₂, modelled using the refined parameters for the visualization in the electronic and structural analysis (VESTA) program, consists of the [WO₄] tetrahedron with T_d symmetry group and the (Na/Eu)O₈ dodecahedron, where Na⁺ and Eu³⁺ are randomly distributed over the same site positions with S₄ symmetry (without an inversion symmetry) (as shown in Fig. 1(d)) and corresponds to two different Eu–O bond lengths (shorter and longer), denoted by [Eu–O]_s and [Eu–O]_l, respectively, indicating the distortion of the (Na/Eu)O₈ dodecahedron.²¹ This distortion (e.g., Δ_Na/Eu–O), estimated from where 〈d_Na/Eu–O〉 denotes average Na/Eu–O bond length, (schematically shown in Fig. 1(e) for all the samples), decreases monotonically from NEWO_0.75 to NEWO_0.25.²¹ The peculiarity of the hydrothermal method, where varying the Na₃cit concentration commonly influences the organization of the [(Na/Eu)O₈] dodecahedron, which causes various structural defects in the form of V_O, distortion of the bonds, stresses, and strain on the crystallite lattices. Presently Δ_Na/Eu–O, ascribed to ‘ε’, is believed to be dependent on V_O, which can be explained as follows: we have noticed an enhancement in the (Na/Eu)–O bond length and reverse trend for W–O, which, according to Li et al., can be attributed to the decreasing dipole–dipole interactions due to charge entrapment at the V_O site, which gets formed due to the antisite defect at the 4a site, as given by .^22,23 Hence, the gradual decrease of Δ_Na/Eu–O from NEWO_0.75 to NEWO_0.25 may be ascribed to the reduced antisite defect. In our case, the hydrothermal reaction of Na⁺, Eu³⁺, and WO₄²⁻ ions resulted in the formation of NaEu(WO₄)₂ precipitates, which can be shown in eqn (2)–(6) below.^24–26

Na₂WO_4(aq) → 2Na⁺_(aq) + (WO₄)²⁻_(aq) (2)

Eu(NO₃)_3(aq) → Eu³⁺_(aq) + 3NO⁻_(aq) (3)

Na₃cit_(aq) → 3Na⁺_(aq) + cit³⁻_(aq) (4)

Eu³⁺_(aq) + cit³⁻_(aq) → [Eu³⁺–cit³⁻] complex_(aq) + NaNO_3(aq) (5)

2[Eu³⁺–cit³⁻] complex_(aq) + 2Na⁺_(aq) + 2(WO₄)²⁻_(aq) → 2 NaEu(WO₄)_2(S) + 2cit³⁻_(aq) (6)

The [Eu³⁺–cit³⁻] complex plays a pivotal role in the synthesis of NaEu(WO₄)₂, specifically on . Low Na₃cit concentration inhibits the formation of the [Eu³⁺–cit³⁻] complex, which as a result reduces the antisite defect. According to Shannon's data, the ionic radius of Na⁺ (1.18 Å) is larger than that of Eu³⁺ (0.947 Å); therefore, the equivalent radius of A-site ²⁷ where ‘x’ represents the fraction of the antisite defect, suggesting that a decrease of the antisite defect will increase the (Na/Eu)–O bond length, in good agreement with our experimental findings. The FESEM images of the samples shown in Fig. 2(a)–(c) illustrates the agglomerated nature of the particle, which may be attributed to the calcination effect at 800 °C.

Fig. 1 The observed and refined XRD pattern of NaEu(WO₄)₂ samples: (a) NEWO_0.75, (b) NEWO_0.50, (c) NEWO_0.25, (d) unit cell of the NaEu(WO₄)₂ tetragonal structure and (e) the schematic representation of the variation in the bond lengths of the Na/Eu–O dodecahedron in NaEu(WO₄)₂.

Table 1 Refined parameters from the Rietveld refinement method using FullProf suite software

Parameters		NEWO0.75	NEWO0.50	NEWO0.25
Lattice parameters	a = b (Å)	5.2586 (2)	5.24393 (8)	5.25153 (6)
	c (Å)	11.4182 (6)	11.3766 (3)	11.3979 (2)
Unit cell volume	V (Å^−3)	315.74 (2)	312.843 (14)	314.3378 (8)
Bond length (in Å)	Na/Eu–O	2.2565 (8)	2.34171 (5)	2.368 (9)
		2.33401 (9)	2.35401 (4)	2.380 (9)
	W–O	2.05794 (7)	1.94063 (3)	1.914 (8)
Bond distance distortion index (ΔNa/Eu–O)		16.88	2.62	2.53
R p (%)		5.10	4.63	4.74
R wp (%)		6.60	6.11	6.17
R exp (%)		5.75	4.90	5.59
χ ^2		1.32	1.56	1.22

---

Fig. 2 FESEM image of the NaEu(WO₄)₂ samples: (a) NEWO_0.75, (b) NEWO_0.50 and (c) NEWO_0.25.

3.2. Investigations of the defects in NaEu(WO₄)₂ by FTIR, Raman and X-ray photoelectron spectroscopy techniques

In order to judge the structural distortion of the dodecahedron due to the antisite defect, the FTIR and Raman spectra were taken as these spectroscopic techniques invaluably illustrate the short-range order–disorder phenomenon.²⁸ The FTIR spectra (Fig. S2, ESI†), measured in the range of 400–1000 cm⁻¹, consists of four absorption bands at 412–418, 438–440, 714–725 and 775–852 cm⁻¹ indexed to internal O–W–O symmetric stretching, O–W–O symmetric bending, O–W–O anti-symmetric stretching (1E_u), and O–W–O anti-symmetric stretching (1A_u) vibrations, respectively,^28,29 while the band at 515–555 cm⁻¹ is attributed to the Eu–O stretching vibration.^30,31 Deconvolution reveals two distinct peaks (shown in the inset of Fig. 3(a)–(c)) measured at 524.5 and 548.1 cm⁻¹ for NEWO_0.75, 524.3 and 547.2 cm⁻¹ for NEWO_0.50, and 521.9 and 542.3 cm⁻¹ for NEWO_0.25. Herein, we believe that these two peaks originate from the deformed EuO₈ dodecahedra containing the antisite defect, i.e., [EuO₈]_deformed and EuO₈ dodecahedra without the antisite defect, i.e., [EuO₈]_ordered.³² [EuO₈]_deformed was estimated from the FTIR spectra using the relation where A_deformed and A_ordered denote the area under the curves corresponding to [EuO₈]_deformed and [EuO₈]_ordered, showing a decreasing trend (∼0.80, 0.51 and 0.19 for NEWO_0.75, NEWO_0.50 and NEWO_0.25, respectively), which is in good agreement with the Rietveld results of the decreasing antisite defect.

Fig. 3 Deconvoluted peaks of the FTIR spectrum: (a) NEWO_0.75, (b) NEWO_0.50, and (c) NEWO_0.25. (d) Raman spectra of NEWO_0.75, and NEWO_0.50, NEWO_0.25. (e) Magnified Raman spectra (480–840 cm⁻¹).

According to group theory calculation, only thirteen Raman vibrational modes of tetragonal scheelite materials at the Γ-point are active and represented by eqn (7)²⁸

Γ = 3A_g + 5B_g + 5E_g (7)

where non-degenerate A_g, B_g modes and doubly degenerate E_g mode originate from various internal vibrations of [WO₄]²⁻ with the S₄ point symmetry or external vibrations of the (Na/Eu)O₈ dodecahedron with D_2d point symmetry. Herein, Raman spectrum of all sample (Fig. 3(d)), measured in between 100–1000 cm⁻¹, comprises of eight peaks, measured at 111, 147, 203, 329, 399, 763, 800 and 917 cm⁻¹, while NEWO_0.50 and NEWO_0.75 show three additional peaks. Very briefly, the peak at 111 cm⁻¹ is assigned to the E_g stretching vibration of the (Na/Eu)O₈ dodecahedron or the translational mode of Na⁺ and Eu³⁺,³³ while the peaks at 147 and 203 cm⁻¹ are indexed to the E_g and A_g free rotation of [WO₄]²⁻, respectively.³⁴ The peaks at 329 and 399 cm⁻¹ are ascribed to the A_g anti-symmetric and B_g symmetric bending of O–W–O,³⁵ while the peaks at 763 and 800 cm⁻¹ are ascertained to the B_g and E_g anti-symmetric stretching of the O–W–O vibrations, respectively.³⁶ The peak at 918 cm⁻¹ is attributed to the A_g symmetric O–W–O stretching.³⁷ In contrast to NEWO_0.25, three additional peaks of comparatively low intensity (zoomed-in image shown in Fig. 3(e)) have been identified at 484, 677 and 750 cm⁻¹ for NEWO_0.50 and NEWO_0.75, and they can be readily ascribed to [(Na/Eu)O₈]_deformed of NEWO_0.50 and NEWO_0.75, thus validating previous studies.³⁸

To accumulate more information about the chemical bonding characteristics of the constituent elements in the presence of [(Na/Eu)O₈]_deformed, we adopted XPS to examine the valence state of Na, Eu, W and O. Herein, all the measured binding energy data have been corrected with reference to C 1s with a binding energy at 284.6 eV, which appears as an environmental carbon. The XPS survey spectra (Fig. S3, ESI†), recorded in between 0 and 1100 eV, represents a core binding energy of Na, Eu, W and O of NEWO_0.75, NEWO_0.50, and NEWO_0.25, while the high-resolution spectra of Eu and O, W and Na were also recorded. As illustrated in Fig. 4(a), (c) and (e), the high-resolution Eu spectrum consists of two peaks at 135.9 and 141.8 eV corresponding to the spin–orbit splitting of 4d_5/2 and 4d_3/2 orbitals, respectively, in accordance with its trivalent oxidation state.³⁹ As these spectra are asymmetric, hence, a careful deconvolution demonstrates two peaks (denoted by E_a and E_b) for 4d_5/2 orbitals with a binding energy of 135.0, 136.8 eV for NEWO_0.75, 135.2, 137.0 eV for NEWO_0.5 and 136.0, 137.5 eV for NEWO_0.25. Similarly, the binding energies of 4d_3/2 orbitals, after deconvolution (e.g., peak E_c and E_d), were obtained at 140.4, 142.2 eV and 141.0, 142.5 and 141.5, 143.1 eV for the three respective samples. Herein, 4d_5/2 with lower binding energy (e.g., 135.0, 135.2 and 136.0 eV for NEWO_0.75, NEWO_0.5 and NEWO_0.25 respectively) is assigned to [EuO₈]_ordered, while a higher binding energy (e.g., 136.8, 137.0, and 137.5 eV for NEWO_0.75, NEWO_0.5 and NEWO_0.25, respectively) corresponds to [EuO₈]_deformed. Two different binding energies denote the presence of Eu³⁺ with two different polarizing fields, which can be normally expressed by where Z_eff and ‘r’ denote nuclear charge and ionic radius in the respective ligand field, respectively, while φ_eff is the effective ionic potential and denotes the polarizability of the Eu³⁺ ion.⁴⁰ In brief, due to the less difference between the electronegativity of O (3.44) and Eu (1.2), the Eu–O bond is covalent, where electrons are shared by Eu and O. In case of the [EuO₈]_ordered dodecahedron, the Eu³⁺ ions are surrounded by eight O atoms, whereas this number is less in the case of [EuO₈]_deformed due to the presence of , suggesting different ‘r’ for these two different dodecahedrons; rather, it will be less in [EuO₈]_deformed. Therefore, the higher binding energy of Eu 4d orbitals in [EuO₈]_deformed is ascribed to the higher φ_eff in the deformed dodecahedron. The [EuO₈]_deformed:[EuO₈]_ordered ratio, an indication of the polarizability around Eu³⁺ ions, was estimated from the area under the 4d_5/2 orbital, and this ratio was found to be 1.0, 0.7 and 0.6 for NEWO_0.75, NEWO_0.50 and NEWO_0.25, respectively, indicating a monotonic decrease of [EuO₈]_deformed. The covalence of the dodecahedron host, which primarily depends on ‘ϕ_eff’, is believed to be decrease from NEWO_0.75 to NEWO_0.25 as the decrease in the antisite defect makes the dodecahedron more ionic due to the larger electronegativity difference between Na (0.93) and O (3.44). In addition, the increasing binding energy of both 4d_5/2 and 4d_3/2 orbitals from NEWO_0.75 to NEWO_0.25 is correlated to the enhancement of the Eu–O bond lengths in good agreement with Rietveld analysis.

Fig. 4 High-resolution XPS spectra of the Eu 4d state and O 1s state of (a) and (b) NEWO_0.75. (c) and (d) NEWO_0.50 and (e) and (f) NEWO_0.25.

For oxygen, the high-resolution O 1s spectra of the three samples are shown in Fig. 4(b), (d) and (f), which is measured at 530.0 eV of the typical asymmetric characteristics, indicating the presence of in the samples.⁴¹ After deconvolution, two different bonding states of oxygen with binding energies of 530.3, 531.0 eV for NEWO_0.75, 530.4, 532.0 eV for NEWO_0.50 and 530.7, 532.6 eV for NEWO_0.25 were observed. Low binding energy corresponds to the lattice oxygen (O_lat) species of [EuO₈]_ordered, while the higher binding energies are attributed to the adsorbed oxygen (O₂(ads)) species at the site within [EuO₈]_deformed. An increase in the binding energy of O_lat is ascribed to the reduced deformation of the [EuO₈]_deformed dodecahedron. However, the O₂(ads)/O_lat molar ratio was calculated to be ∼0.44, 0.29 and 0.20 for NEWO_0.75, NEWO_0.50 and NEWO_0.25, respectively; hence, the result indicates a monotonic decrease in from NEWO_0.75 to NEWO_0.25. As XPS is highly sensitive to the surface, therefore, we believe that these mostly resides on the sample surface. In contrast, the spectra of Na and W have been found in the Gaussian shape and were deconvoluted (as shown in Fig. S4, ESI†). From the above FTIR and XPS experiments, we have confirmed that the antisite defect mainly arose due to the Eu elements, which play a key role in the defects.

3.3. Investigation of the optical properties of NaEu(WO₄)₂

In order to gain insight into the mechanism on the effect of the antisite defect on the optical properties, the UV-vis absorption spectra (shown in Fig. S5, ESI†) were recorded and analysed to calculate the band gap (E_g) using Tauc plot.⁴² Herein, the direct E_g of NEWO (discussed later) was estimated to be ∼2.24, 2.27 and 2.58 eV for NEWO_0.75, NEWO_0.50, and NEWO_0.25, respectively (shown in Fig. 5(a)–(c)). Such a trend of increasing E_g may be assigned to a decrease in the covalence of the host lattice due to [EuO₈]_deformed and can be explained as follows: mostly, E_g depends on the polarizability of cations and deformation of anions. Greater the cationic polarization and anionic deformation, stronger the covalent links between the anion and cation, thus narrowing the E_g. Our XPS studies reveal that the covalence of the host matrix decreases from NEWO_0.75 to NEWO_0.25; thus, the enhancement of E_g can be certainly ascribed to the reduced covalence of the host matrix.

Fig. 5 Tauc plot: (a) NEWO_0.75, (b) NEWO_0.50, and (c) NEWO_0.25. (d) PL spectra of NEWO_0.75, NEWO_0.50, and NEWO_0.25. (e) Bar diagram of the normalized PL intensity of samples and (f) CIE diagram of NEWO_0.75, NEWO_0.50, and NEWO_0.25 samples.

Prior to examining the emission, we recorded the photoluminescence excitation (PLE) spectra of the as-synthesized samples (normalized spectra are shown in Fig. S6(a), ESI†). It is clearly visible that the PLE spectra of all the as-prepared materials consist of three peaks, measured at 380, 392 and 463 nm, which are assigned to ⁷F₀–⁵L₇, ⁷F₀–⁵L₆ and ⁷F₀–⁵D₂ intra-configurational 4f⁶ transitions of Eu³⁺. One peak, recorded at 360 nm for NEWO_0.50 and NEWO_0.25, is ascribed to the ⁷F₀–⁵D₄ transition of Eu³⁺, while another peak of relatively low intensity at 413 nm, observed only for NEWO_0.25, demonstrates the ⁷F₀–⁵D₃ transition of Eu³⁺. Careful literature survey reveals that our measured PLE spectra are not very sharp in comparison with other observations and may be ascribed to the distorted dodecahedron.⁴³ In addition, a sharp peak was also identified at 306 nm, which may be assigned to the charge transfer band (CTB) from the surrounding oxygen anions to the WO₄ tetrahedron. In this context, it may be stated that this is an interesting result for this particular phosphor material because of the fact that they have intense excitation peaks at 380, 392 and 463 nm in the ultra-violet (UV), near-UV, and UV-blue region, respectively, can act as efficient pumping sources for the red emission from Eu³⁺ ions.

Normalized photoluminescence signals under 392 nm excitation wavelength are presented in (Fig. 5(d)). Herein, the spectrum of each sample shows a typical emission band of Eu³⁺ in the range of 555–635 nm, comprising one prominent peak at 615 nm and two relatively less intense peaks at 565 and 592 nm. The peak at 615 nm corresponds to the ⁵D₀ → ⁷F₂ electric dipole transition of Eu³⁺, which is hypersensitive to the local structures of Eu³⁺, particularly in the non-centrosymmetric crystal site, according to Judd–Ofelt's theory. In contrast, the emission peak at 592 nm, assigned to the ⁵D₀ → ⁷F₁ magnetic dipole transition of Eu³⁺, does not depend on the surrounding environment.⁴⁴ It is noteworthy that the ground state electronic configuration of Eu³⁺ has ⁷F₀ non-degenerate and non-overlapping ^2S+1L_J multiplets; thus, Eu³⁺ is mostly used as an optical probe to investigate the crystal field surrounding them, where the intensity ratio between ⁵D₀ → ⁷F₂ and ⁵D₀ → ⁷F₁ is used to estimate the asymmetry of the local environment.⁴⁵ A very careful estimation reveals that the area under the 615 nm emission curve increases from NEWO_0.75 to NEWO_0.25, indicating the reduction of [EuO₈]_deformed, in good agreement with the previous XRD and FTIR results. Hence, it may be concluded that the lowering of the Na₃cit concentration reduces the antisite defect as well as , which, as a result, increase the symmetry of the crystal structure. The emission peak at 565 nm, attributed to the ⁵D₀ → ⁷F₀ transition of Eu³⁺, has been observed in few other NaEu(WO₄)₂-like host matrices including the LaAlO₃:Eu³⁺ nanophosphor, LaPO₄:Eu³⁺ nanoparticles, SrB₂O₄:Eu³⁺ phosphor, and potassium–aluminoborotellurite.^46–48 In principle, he ⁵D₀ → ⁷F₀ transition of Eu³⁺ is prohibited according to the selection rule of magnetic dipole transition (J = 0 → J′ = 0). However, the presence of this transition indicates the violation of the Judd–Ofelt's selection rule, which may be ascribed to the asymmetry of the [EuO₈]_deformed dodecahedron.⁴⁹ Herein, we believe that the asymmetry of the crystal lattice leads to wavefunction mixing between the ⁷F₀ and ⁷F₂ states (discussed later). In this context, the second-order parameter (B₂₀) of the crystal field expansion plays a pivotal role in the mixing of the wavefunctions and is related to the intensity of ⁵D₀ → ⁷F₀ and ⁵D₀ → ⁷F₂ transitions by the following eqn (8).⁵⁰

where Δ₂₀ (=1439 cm⁻¹) denotes the energy gap between the ⁷F₀ and ⁷F₂ levels. In this calculation, B₂₀ was found to be 623, 1393 and 2492 cm⁻¹ for NEWO_0.25, NEWO_0.50 and NEWO_0.75, respectively. These values of the B₂₀ parameter are within the range of few other materials such as YPO4:Eu³⁺ (1365 cm⁻¹), Ca(PO₃)₂:Eu³⁺ (1600 cm⁻¹), and K–Al–B–Te:Eu³⁺ (829 cm⁻¹), which suggests the reliability of our estimated B₂₀ parameters.⁵¹ In this context, it may be stated that a smaller magnitude of the B₂₀ parameter indicates a low degree of crystal field perturbation (denoted by the asymmetric ⁵D₀ → ⁷F₀ transition at 565 nm, shown in the bar diagram in Fig. 5(e)); thus, the decreasing trend of B₂₀ from NEWO_0.75 to NEWO_0.25 can be certainly attributed to the lowering of [EuO₈]_deformed in our synthesized samples and corroborates highly with the Rietveld, FTIR and XPS studies. The quantum yield (QY), of NEWO_0.25, calculated from the photoluminescence decay profile (shown in Fig. S6(b), ESI†) using the relation (where, τ₁ and τ₂ denote non-radiative and radiative recombination times), is ∼50%, which is slightly less than the QY of other scheelite materials.^52,53

Furthermore, the colour feature, indicated by the color coordinate (x, y) in the Commission International de I’Eclariage (CIE) chromaticity diagram (shown in Fig. 5(f)), has been examined for our synthesized samples. As presented in Table 2, the color coordinates of all the samples are observed within the red region, which may be attributed to the strong emission at 615 nm. Most importantly, NEWO_0.25 exhibits improved colour coordinate from the commercially available red Y₂O₂S:Eu³⁺ phosphor (0.63, 0.35),⁵⁴ which perfectly matches with the standard color coordinate (0.67, 0.33) according to NTSC. Colour purity (CP) of ∼86.7, 93.6 and 97.1% for NEWO_0.75, NEWO_0.50 and NEWO_0.25 was calculated using the following eqn (9)³⁸

where (x, y) is the CIE coordinate of the NaEu(WO₄)₂ samples, (x_i, y_i) is the illuminate point of the 1931 CIE Standard Source C with the colour coordinates of (0.3101, 0.3162), and (x_d, y_d) is the colour coordinate of the dominant wavelength. Herein, the correlated colour temperatures (CCTs) of ∼1747, 2365 and 2957 K for NEWO_0.75, NEWO_0.50 and NEWO_0.25, respectively, were calculated using the McCamy relation eqn (10)⁵⁵

CCT = 449n³ + 3525n² + 6823.3n + 5520.3 (10)

where ; (x, y) represent the chromaticity coordinate. These results suggest that NaEu(WO₄)₂ may be a potential candidate for red emission with excellent color purity and color coordinates close to the NTSC norms.

Table 2 List of the CIE co-ordinates of this work, commercial red phosphors and few synthesised red phosphors

Materials name		CIE coordinate		Ref.
		x	y
NEWO0.25		0.667	0.333	This work
NEWO0.50		0.653	0.347	This work
NEWO0.75		0.625	0.374	This work
NTSC value		0.67	0.33	56
Commercial red phosphors	Y2O2S:Eu^3+	0.63	0.35	54
	Y2O3:Eu^3+	0.49	0.32	57
Ba2Gd0.1Eu0.9(BO3)2Cl		0.581	0.355	57
Ba2Bi0.4Eu0.6V3O11		0.65	0.35	58
KGd0.7Eu0.3TiO4		0.65	0.34	59
Sr3Y0.94 (PO4)3:0.06 Eu^3+		0.64	0.33	56
Sr1.7Zn0.3CeO4:Eu^3+		0.60	0.34	60
Ca2Ga2GeO7:0.07Eu^3+,0.07Na^+		0.64	0.359	61
CdWO4:0.5%Eu^3+		0.6465	0.347	62
Ca3(PO4)2:0.04Eu^3+		0.5776	0.2840	63
YBO3:Eu^3+		0.655	0.345	64
NaCaBO3:0.11Eu^3+		0.608	0.387	65
Sr2SiO4:Eu^3+		0.6106	0.3517	66

---

4. Investigation of the gas sensing properties of NaEu(WO₄)₂

The gas sensing property of any metal oxide mostly depends on the gas–solid interaction with target gas molecules, wherein the adsorption/desorption of the gas molecule leads to a significant change in the resistance of the material. Previous studies by several researchers including our group reveal that V_O traps electrons (e′), acts as a donor level within E_g and plays a pivotal role in both optical and gas sensing.^{15,18,67–72} For metal oxide sensors, V_O facilitates receptor and transducer functions, which in turn tune the interaction of the target gas molecule on the sensor surface and transportation of the generated electronic signal within the sensor. It is an accepted mechanism that V_O enhances the chemisorption of atmospheric O₂, as described by O₂(gas) + e⁻ ↔ O₂⁻(ads)⁷³ and facilitates the receptor activity by oxidizing more of the target gas molecules. Herein, a careful literature survey shows that O₂⁻(ads) can facilitate the following redox reactions between NEWO samples and acetone (CH₃COCH₃), ethanol (C₂H₅OH), ammonia (NH₃), methanol (CH₃OH), and formaldehyde (HCHO), as presented by eqn (11)–(15).

CH₃COCH₃(gas) + 8O₂⁻(ads) ↔ 3H₂O(gas) + 3CO₂(gas) + 8e⁻ (11)

CH₃CH₂OH(gas) + O₂⁻(ads) ↔ 2H₂O(gas) + C₂H₂O + 2e⁻ (12)

4NH₃(gas) + 3O₂⁻(ads) ↔ 6H₂O(gas) + 2N₂(gas) + 3e⁻ (13)

CH₃OH(gas) + O₂⁻(ads) ↔ 2H₂O(gas) + CO(gas) + 2e⁻ (14)

HCHO(gas) + O₂⁻(ads) ↔ H₂O(gas) + CO₂(gas) + 2e⁻ (15)

These electrons released at the defect state at the conduction band minima act as donors, thus improving the response significantly. As XPS analyses reveals the presence of O₂(ads) formed within the [EuO₈]_deformed dodecahedron, hence, the gas sensing property can be expected from NEWO samples. In order to validate the gas sensing ability of our synthesized NEWO samples as a proof-of-concept, we carried out a series of resistive-type sensing measurements using NEWO_0.75, NEWO_0.50 and NEWO_0.25 as sensor probes against NH₃, CH₃COCH₃, C₂H₅OH, CH₃OH, HCHO and IPA (C₃H₈O) at room temperature. As shown in Fig. 6(a), NEWO_0.75 exhibits highly selective sensitivity against CH₃COCH₃ in comparison with other target gases such as etc. as its resistance changes, while no significant variation was observed for other gases ( which was 5%, 3%, 8%, 4% and 6% for C₂H₅OH, NH₃, HCHO, CH₃OH and C₃H₈O, respectively). Room temperature operational condition is ascribed to the low bond-dissociation energy of CH₃COCH₃ (393 kJ mol⁻¹), in comparison with C₂H₅OH (458 kJ mol⁻¹), NH₃ (435 kJ mol⁻¹), and CH₃OH (439 kJ mol⁻¹), while the absence of sensing activity of NEWO_0.50 and NEWO_0.25 can be assigned to the low [EuO₈]_deformed content and can be explained as follows. The electron-depleted layer (L_D), formed due to O₂⁻(ads), also plays a pivotal role in transducing the electronic signal in sensing response. Normally, L_D is prominent for the V_O-enriched sample, leading to superior sensing response.⁷¹ Our XPS studies reveal the highest content of NEWO_0.75, indicating the most efficient signal transduction in comparison with NEWO_0.25 and NEWO_0.50. Therefore, the high content of NEWO_0.75 is believed to play a significant role in sensing CH₃COCH₃, one of the highly flammable VOCs that cause difficulties in breathing, nausea, vomiting and symptoms of neurological poisoning and also as a biomarker of diabetes found in exhaled human breath.^74,75

Fig. 6 (a) Selectivity of theNEWO_0.75 sensor toward acetone at RT. (b) Dynamic response curve of NEWO_0.75 when exposed to 1–100 ppm acetone at RT. (c) Acetone response value of NEWO_0.75 as a function of acetone concentration at RT (inset: isotherm fitting of the acetone response value of NEWO_0.75 with linear regression). (d) Response and recovery time of the NEWO_0.75 sensor for 100 ppm acetone at RT. (e) Response data at different %RH values to 100 ppm acetone at RT. (f) Long-term stability data of the NEWO_0.75 sensor to 100 ppm acetone for 120 days.

Dynamic response, tested by exposing the sensor to different concentrations of CH₃COCH₃ ranging from 1 to 100 ppm (Fig. 6(b)), illustrates that the response increases gradually with increasing CH₃COCH₃ concentration, while Fig. 6(c) depicts the response of NEWO_0.75 as a function of CH₃COCH₃ concentration. Herein, it may be concluded from Fig. 6(c) that the NEWO_0.75 sensor has the capability to detect CH₃COCH₃ down to 1 ppm with changes in the resistance of ∼12% at room temperature, while the response increases to ∼68% at 100 ppm CH₃COCH₃ saturating above 100 ppm. To understand the adsorption/desorption mechanism, we fitted the response data with the Freundlich adsorption isotherm, as given by eqn (16) and (17).

S_g = 1 + aC^b_g (16)

i.e. log(S_g − 1) = a + b log(C_g) (17)

where, S_g and C_g are the sensitivity and concentration of the target gas, while “a” and “b” denote the constants that depend on the electrical charge of the surface species and stoichiometry of the surface reaction. As is known from previous reports, b < 0.5 indicates physisorption process, while b > 1 denotes chemisorption mechanism.⁷⁶ Currently, the linear characteristic between log(S_g − 1) and log(C_g), as shown in the inset of Fig. 6(c), provides b = 0.38, indicating the physisorption of CH₃COCH₃ on the surface of NEWO_0.75. Response and recovery time of ∼3.3 s/10 s in the presence of 100 ppm CH₃COCH₃ indicate that the sensor is suitable for ultrafast application, as shown in Fig. 6(d). The repeatability, studied by exposing the sensor to seven consecutive cycles of exposure to 10 ppm acetone at room temperature (shown in Fig. S7, ESI†), gives almost similar characteristics (≤4% variation) for each measurement, yielding the excellent repeatability of NEWO_0.75 as a sensor material. The performance of the sensor was also checked under different humidity conditions, wherein our sensor shows highly robust performance under humidity (Fig. 6(e)). The long-term stability of the senor (presented in Fig. 6(f)), tested over 120 days, shows the negligible deterioration of the response. In this context, we compared the CH₃COCH₃ response characteristics in the previous literature and have presented a study in Table 3. It can be stated here that NaEu(WO₄)₂ exhibits the best response characteristics (∼68% to 100 ppm acetone) with very short response–recovery time, and most importantly, it works at room temperature, indicating the reliability of the NaEu(WO₄)₂-based sensor for different non-invasive applications.

Table 3 The sensing performance of various metal oxides toward acetone gas

Material name (type)	Operating temperature (°C)	Response/recovery time (s)	Acetone response value (in ppm)	Ref.
NEWO0.75	Room temperature	3.3/10	68 (100 ppm)	This work
0.5% Fe–ZnO	365	3/12	105.7 (100 ppm)	77
Pt–Fe2O3	139	3/22	25.7 (100 ppm)	78
ZnFe2O4	220	23/9	64.9 (50 ppm)	79
SnO2	370	9.7/5.8	50.2 (200 ppm)	80
WO3	320	5/5	32 (100 ppm)	81
TiO2	400	0.75/0.5	21.6 (200 ppm)	82
Co3O4	240	2/5	4.88 (500 ppm)	83
PrFeO3	180	5/5	32.5 (50 ppm)	84
In2O3	300	0.7/14	29.8 (50 ppm)	85
La0.7Sr0.3FeO3	275	20/270	0.7 (500 ppm)	86

---

Presently, the high selectivity of NaEu(WO₄)₂ towards CH₃COCH₃ sensing can be understood as the lowering of bond dissociation energy (BDE) in comparison with other VOCs. According to the previous reports,^87–90 the BDE of acetone is 393 kJ mol⁻¹, whereas that of other gases is as follows: ethanol, 458 kJ mol⁻¹; ammonia, 435 kJ mol⁻¹; methanol, 439 kJ mol⁻¹; formaldehyde, 364 kJ mol⁻¹; and IPA, 440 kJ mol⁻¹. During the adsorption of the gas molecules, acetone molecules will break more easily, allowing them to participate in the reaction process. However, the gas molecules having higher BDE would be less likely to react with the sensor material at low operating temperature (RT). It is interesting to note that even though the BDE of formaldehyde is in the range comparable to that of acetone, the sensor showed lower response to formaldehyde. This can be attributed to the electron donating effect during the reduction process of the target gases. As given in eqn (11)–(15), it is evident that during the reduction process of acetone, a greater number of electrons are released that contribute to the highest sensitivity and makes the sensor selective to acetone than any other target gas. Additionally, acetone molecules are known to have a large dipole moment (μ = 2.91 D).^91,92 Therefore, there will be a strong interaction between acetone and the sensing layer of NEWO_0.75, resulting in high response.

5. Computational analysis of ordered and disordered NaEu(WO₄)₂

To perceive a deep insight on the effect of the antisite defect on optical and gas sensing properties, we carried out a series of ab initio calculations of electronic structure using density functional theory. Prior to examining the antisite effects, we calculated the electronic structure along high symmetry points within the Brillouin zone, total density of states (TDOS), and angular momentum projected partial density of states (PDOS) of perfectly ordered NEWO (NEWO_order). In this calculation, we optimized the unit cell structure and obtained lattice parameters a = b ∼ 5.215 Å and c ∼ 11.511 Å, Eu–O bond lengths ∼2.461 and 2.460 Å, W–O ∼1.783 Å and Δ_Eu–O = 0.04, which match closely with the experimentally measured parameters, indicating the good choice of the exchange potential for our calculations. In these calculations, the top of the valence band maxima (VBM) was set at zero. The electronic band structure (shown in Fig. 7(a)) of NEWO_order exhibits E_g ∼ 3.29 eV with both valence band maxima (VBM) and conduction band minima (CBM) at the Γ-point. TDOS and PDOS calculations (shown in Fig. 7(b)–(f)) demonstrate the predominant contribution of O 2p_x, 2p_y and Eu 4f orbitals in following order 4f_z³ > 4f_xz² > 4f_zx³ > 4f_y³x² > 4f_yz², 4f_x³, 4f_xyz to the upper part of the valence band, while the lower part mostly consists of W 4d_xz, 4d_z², 4d_xy and O 2p_z. The lower part of the conduction band mostly comprises the hybridization of W 4d_z², 4d_x² O 2p_z orbitals, whereas the upper part mostly consists of Eu 4f–O 2p_x, 2p_y orbitals. As the electronegativity of Eu (1.2) is lower than that of W (2.36) on the empirical Pauling electronegativity scale, hence, Eu 4f orbitals should lie above the W 4d orbitals, which corroborates well with our calculations. Among different orbitals, the VBM and CBM of NEWO mostly originate from the hybridization of Eu 4f_xyz–O 2p_x, 2p_y and W 4d_z²–O 2p_z, respectively. Therefore, it may be stated that the VB to CB electronic transition is associated with the charge transfer from the EuO₈ dodecahedron to the WO₄ tetrahedron. The band structure, TDOS and PDOS calculations of the disordered samples (shown in Fig. S8–S10, ESI†) give E_g ∼ 1.77, 1.47 and 0.77 eV for NEWO_0.25, NEWO_0.50 and NEWO_0.75, respectively; thus, the decreasing trend of E_g agrees well with the experimental results. Though the VBM remains unchanged at Γ-point for all the samples, however, the CBM is shifted to the M-point for NEWO_0.50 and NEWO_0.75, indicating a change in the E_g from direct to indirect. PDOS and TDOS illustrate that all the orbitals get delocalized in the disordered samples with respect to NEWO_order, indicating enhanced bonding hybridization among the Eu 4f, O 2p and W 4d orbitals. Therefore, the observed phenomenon of enhanced second-order parameter (B₂₀) is certainly ascribed to the mixing of the orbitals 4f_z³, 4f_xz², 4f_y³x², 4f_yz², 4f_x³, 4f_xyz while, most importantly, the 4f_zx² orbital does not contribute to this variation. However, significant modifications have been noted for the conduction band; more specifically, the delocalization and red shifting observed for W 4d_z², 4d_x² and O 2p_z orbitals are attributed to the increase in the W–O bond length; thus, the reduction of E_g is assigned to the red shifting of the W 4d_z², 4d_x² and O 2p_z states. The high density of states of Eu 4f in the conduction band, found in the deformed samples and assigned to the higher antisite defect, agree well with the enhanced emission corresponding to the f–f transitions of Eu³⁺ ions.

Fig. 7 DFT of NEWO_order: (a) band structure, (b) TDOS, (c) DOS of Na, (d) DOS of Eu, (e) DOS of W, and (f) DOS of O.

Furthermore, in order to understand the effect of antisite effects on the sensing performances, we examined CH₃COCH₃ adsorption on NEWO_order and NEWO_0.75, while the calculation was performed using three different orientations of the CH₃COCH₃ molecule (shown in Fig. 8(a)–(c)) on the (0 0 4) plane as the model plane for adsorption. As depicted in Table 4, CH₃COCH₃ adsorption has been found to be stable in vertical configuration and most importantly, we notice the most negative CH₃COCH₃ adsorption energy for NEWO_0.75 in comparison with NEWO_order, indicating the greater stability of CH₃COCH₃ on the (0 0 4) plane of NEWO_0.75. We observed an optimized distance between the adsorbed CH₃COCH₃ and sample surface of ∼2.9–3.2 Å, which is larger than the sum of the corresponding atomic covalent radii; hence, it may be stated that no direct chemical is formed between CH₃COCH₃ and the sample. It is noted (shown in Fig. 8(d)–(f)) that CH₃COCH₃ adsorption gives rise to additional O 2p_x, 2p_y and 2p_z orbitals as impurity levels in between the valence and conduction bands. In addition, we also observed that the PDOS calculations of Eu 4f and W 4d orbitals get red-shifted, indicating a significant change in the conductivity of NEWO in the presence of CH₃COCH₃ (shown in Fig. S11 and S12, ESI†). However, this modification is found to be more significant in NEWO_0.75, suggesting the higher sensing performance of NEWO_0.75 with respect to NEWO_order. Therefore, it may be concluded that the antisite defect related plays the most important role in CH₃COCH₃ sensing by NaEu(WO₄)₂ at room temperature, while conductivity changes mostly involve Eu 4f orbitals.

Fig. 8 Three possible layouts of the acetone molecule and the (0 0 4) plane surface of the NEWO_0.75 material: (a) vertical position, (b) parallel to x-axis and (c) parallel to y-axis position. PDOS plot of the (0 0 4) slab surface + acetone molecule: (d) O 2p_x orbitals, (e) O 2p_y orbitals, and (f) O 2p_z orbitals.

Table 4 Adsorption energy of acetone gas on the (0 0 4) plane surface of targeted nanomaterials

Target materials	Adsorption energy
	Vertical position	Parallel X position	Parallel Y position
NEWOorder	−1.20	+0.45	+0.42
NEWO0.75	−1.30	+0.19	+0.18

---

6. Conclusion

In summary, the present study reports the synthesis and characterization of the antisite defect containing NaEu(WO₄)₂ synthesized at various concentrations of Na₃cit. Herein, we have identified that the antisite defects that can be controlled by Na₃cit concentration during hydrothermal reaction have a significant effect on the structural deformation of the EuO₈ dodecahedron, covalence of the tungstate host and polarization around the Eu³⁺ ions. Due to the variation of the polarized field around Eu³⁺ ions, a significant mixing of the Eu 4f orbitals, followed by the relaxation of the f–f transition rule, was observed and because of this relaxation, NaEu(WO₄)₂ exhibits strong luminescence at 615 nm with colour coordinate (0.667, 0.333) and color purity ∼97.1% under near-UV excitation (392 nm), indicating its potency as a red phosphor material in LEDs. We identified the presence of s and associated antisite defects, and they act as the active sites for O₂ adsorption, which in consequence facilitates the CH₃COCH₃ sensing property. Most importantly, we have found the sensing property at room temperature, which is very rare but has tremendous technological importance due to the low power consumption. Our ab initio calculations show a deep understanding of the electronic structure of NaEu(WO₄)₂, wherein the VBM and CBM originate from Eu–O and W–O hybridizations, respectively, indicating a charge transfer optical band. Finally, it may be concluded that the prepared materials could be multifunctional materials that can work as both red phosphor and room-temperature acetone gas sensor through the tuning of the antisite defect with the control of the Na₃cit concentration.

Data availability

The data supporting the findings of this work are available within the article and its ESI.† Raw data that support the findings of this work are available from the corresponding authors (C. K. G.) upon reasonable request.

Conflicts of interest

There is no conflicts to declare.

Acknowledgements

One of the authors (K. R. S.) thanks UGC, Govt. of India (NTA Ref. no.: 191620049221) for financial support during execution of this work. The authors would also like to extend their appreciation to the Researchers Supporting Project number (RSPD2024R956), King Saud University, Riyadh, Saudi Arabia.

References

1. M. Yin, Y. Liu, L. Mei, M. S. Molokeev, Z. Huang and M. Fang, RSC Adv., 2015, 5, 73077–73082.
2. M. Laguna, N. O. Nuñez, A. I. Becerro and M. Ocaña, CrystEngComm, 2017, 19, 1590–1600.
3. K. G. Sharma and N. R. Singh, New J. Chem., 2013, 37, 2784–2791.
4. H. Zhu, C. Liang and W. Huang, Phys. B, 2020, 582, 411999.
5. X. Yu, Y. Jiang, X. Li, Z. Song, X. Zhang, H. Liu, B. Zhao, T. Ye, L. Duan and J. Fan, CrystEngComm, 2022, 24, 805–817.
6. Y. Ding, J. Liu, M. Zeng, X. Wang, J. Shi, W. Wang, Y. Miao and X. Yu, Dalton Trans., 2018, 47, 8697–8705.
7. V. Mahalingam and J. Thirumalai, RSC Adv., 2016, 6, 80390–80397.
8. R. Singh, A. King and B. B. Nayak, Optik, 2021, 247, 167870.
9. Y. Yang, C. Chen and Q. Wang, J. Alloys Compd., 2024, 175042.
10. V. Chauhan, P. Dixit, P. K. Pandey, S. Chaturvedi and P. C. Pandey, Methods Appl. Fluoresc., 2023, 12, 015002.
11. L. Kong, H. Sun, Y. Nie, Y. Yan, R. Wang, Q. Ding, S. Zhang, H. Yu and G. Luan, Molecules, 2023, 28, 2681.
12. J. Huang, J. Xu, H. Luo, X. Yu and Y. Li, Inorg. Chem., 2011, 50, 11487–11492.
13. S. Neeraj, N. Kijima and A. K. Cheetham, Chem. Phys. Lett., 2004, 387, 2–6.
14. N. Haldar, T. Mondal, A. Dutta, D. Sarkar, U. K. Ghorai and C. K. Ghosh, Appl. Phys. A, 2023, 129, 708.
15. N. Haldar, T. Mondal, T. Das, D. Sarkar, M. Pal and C. K. Ghosh, CrystEngComm, 2023, 25, 3514–3527.
16. A. K. Munirathnappa, A. K, A. K. Sinha and N. G. Sundaram, Cryst. Growth Des., 2018, 18, 253–263.
17. A. K. Munirathnappa, V. C. Petwal, J. Dwivedi and N. G. Sundaram, New J. Chem., 2018, 42, 2726–2732.
18. N. Haldar, T. Mondal, T. Das, D. Sarkar, M. Pal, A. H. Seikh and C. K. Ghosh, Mater. Adv., 2024, 5, 4480–4490.
19. S. Brahma, Y. W. Yeh, J. L. Huang and C. P. Liu, Appl. Surf. Sci., 2021, 564, 150351.
20. C. Li, H. Zhou, S. Yang, L. Wei, Z. Han, Y. Zhang and H. Pan, ACS Appl. Nano Mater., 2019, 2, 6144–6151.
21. F. A. Rabuffetti, S. P. Culver, L. Suescun and R. L. Brutchey, Inorg. Chem., 2014, 53, 1056–1061.
22. L. Li, Y. Su and G. Li, Appl. Phys. Lett., 2007, 90, 054105.
23. G. M. Kuz'Micheva, I. A. Kaurova, V. B. Rybakov, P. A. Eistrikh-Geller, E. V. Zharikov, D. A. Lis and K. A. Subbotin, CrystEngComm, 2016, 18, 2921–2928.
24. Z. Xu, C. Li, G. Li, R. Chai, C. Peng, D. Yang and J. Lin, J. Phys. Chem. C, 2010, 114, 2573–2582.
25. Y. Chen, S. W. Park, B. K. Moon, B. C. Choi, J. H. Jeong and C. Guo, CrystEngComm, 2013, 15, 8255–8261.
26. S. Suwanboon, W. Somraksa, P. Amornpitoksuk and C. Randorn, J. Alloys Compd., 2020, 832, 154963.
27. J. Guo, D. Zhou, Y. Li, T. Shao, Z. M. Qi, B. B. Jin and H. Wang, Dalton Trans., 2014, 43, 11888–11896.
28. R. F. Gonçalves, L. S. Cavalcante, I. C. Nogueira, E. Longo, M. J. Godinho, J. C. Sczancoski, V. R. Mastelaro, I. M. Pinatti, I. L. V. Rosa and A. P. A. Marques, CrystEngComm, 2015, 17, 1654–1666.
29. V. V. Popov, Y. V. Zubavichus, A. P. Menushenkov, A. A. Yastrebtsev, B. R. Gaynanov, S. G. Rudakov, A. A. Ivanov, F. E. Dubyago, R. D. Svetogorov, E. V. Khramov and N. A. Tsarenko, Crystals, 2022, 12, 892.
30. Z. Mo, Z. Deng, R. Guo, Q. Fu, C. Feng, P. Liu and Y. Sun, Mater. Sci. Eng., B, 2012, 177, 121–126.
31. H. Yang, D. Zhang, L. Shi and J. Fang, Acta Mater., 2008, 56, 955–967.
32. R. Gutkowski, C. Khare, F. Conzuelo, Y. U. Kayran, A. Ludwig and W. Schuhmann, Energy Environ. Sci., 2017, 10, 1213–1221.
33. J. V. B. Moura, C. Luz-Lima, G. S. Pinheiro and P. T. C. Freire, Spectrochim. Acta, Part A, 2019, 208, 229–235.
34. L. S. Cavalcante, J. C. Sczancoski, J. W. M. Espinosa, J. A. Varela, P. S. Pizani and E. Longo, J. Alloys Compd., 2009, 474, 195–200.
35. N. Dirany, E. McRae and M. Arab, CrystEngComm, 2017, 19, 5008–5021.
36. S. S. Bhat, D. Swain, C. Narayana, M. Feygenson, J. C. Neuefeind and N. G. Sundaram, Cryst. Growth Des., 2014, 14, 835–843.
37. A. K. Munirathnappa, D. Dwibedi, J. Hester, P. Barpanda, D. Swain, C. Narayana and N. G. Sundaram, J. Phys. Chem. C, 2018, 123, 1041–1049.
38. B. P. Singh, J. Singh and R. A. Singh, RSC Adv., 2014, 4, 32605–32621.
39. H. Sun, B. Zhang, Q. Zhao, Y. Liu, R. Wu, S. Zhang, Y. Li and A. Chang, J. Am. Ceram. Soc., 2022, 105, 3715–3724.
40. P. W. Atkins, T. L. Overton, J. P. Rourke, M. T. Weller and F. A. Armstrong, Inorg. Chem, Oxford University Press, New York, 5th edn, 2010, p. 16.
41. S. K. Gupta, M. Sahu, P. S. Ghosh, D. Tyagi, M. K. Saxena and R. M. Kadam, Dalton Trans., 2015, 44, 18957–18969.
42. X. Wang, Z. Zhao, Q. Wu, C. Wang, Q. Wang, L. Yanyan and Y. Wang, J. Mater. Chem. C, 2016, 4, 8795–8801.
43. P. Jena, S. K. Gupta, N. K. Verma, A. K. Singh and R. M. Kadam, New J. Chem., 2017, 41, 8947–8958.
44. B. Xu, X. Cao, G. Wang, Y. Li, Y. Wang and J. Su, Dalton Trans., 2014, 43, 11493–11501.
45. B. P. Singh, P. V. Ramakrishna, S. Singh, V. K. Sonu, S. Singh, P. Singh, A. Bahadur, R. A. Singh and S. B. Rai, RSC Adv., 2015, 5, 55977–55985.
46. X. Dong, Z. Fu, Y. Yu, S. Li and Z. Dai, Mater. Lett., 2012, 74, 140–142.
47. G. Rui, Q. Dong and L. Wei, Trans. Nonferrous Met. Soc. China, 2010, 20, 432–436.
48. L. S. Zhao, J. Liu, Z. C. Wu and S. P. Kuang, Spectrochim. Acta, Part A, 2012, 87, 228–231.
49. Z. Wang, J. Zhong, H. Jiang, J. Wang and H. Liang, Cryst. Growth Des., 2014, 14, 3767–3773.
50. A. K. Parchur and R. S. Ningthoujam, RSC Adv., 2012, 2, 10859–10868.
51. G. Nishimura and T. Kushida, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 9075.
52. S. K. Gupta, P. S. Ghosh, M. Sahu, K. Bhattacharyya, R. Tewari and V. Natarajan, RSC Adv., 2015, 5, 58832–58842.
53. V. Mahalingam and J. Thirumalai, J. Mater. Sci.: Mater. Electron., 2016, 27, 8884–8890.
54. L. Jinglei, Y. U. Yan, L. Lingjie, S. Cheng, L. I. Guanyu and L. I. Nianbing, J. Rare Earths, 2012, 30, 330–334.
55. M. Rai, G. Kaur, S. K. Singh and S. B. Rai, Dalton Trans., 2015, 44, 6184–6192.
56. B. Yang, Z. Yang, Y. Liu, F. Lu, P. Li, Y. Yang and X. Li, Ceram. Int., 2012, 38, 4895–4900.
57. H. Jing, C. Guo, N. Zhang, Z. Ren and J. Bai, ECS J. Solid State Sci. Technol., 2012, 2, R1.
58. J. Zhao, C. Guo, T. Li, X. Su, N. Zhang and J. Chen, Dyes Pigm., 2016, 132, 159–166.
59. N. Zhang, C. Guo, J. Zheng, X. Su and J. Zhao, J. Mater. Chem. C, 2014, 2, 3988–3994.
60. H. Li, R. Zhao, Y. Jia, W. Sun, J. Fu, L. Jiang, S. Zhang, R. Pang and C. Li, ACS Appl. Mater. Interfaces, 2014, 6, 3163–3169.
61. Y. Li, N. Li, P. Zhang, Z. Wei, Z. Wang, L. Zhao and W. Chen, Spectrochim. Acta, Part A, 2021, 248, 119247.
62. M. You, J. Xu, Z. Zhang and Y. Zhou, Ceram. Int., 2014, 40, 16189–16194.
63. X. Zhang, L. Zhou and M. Gong, Opt. Mater., 2013, 35, 993–997.
64. R. S. Yadav, R. K. Dutta, M. Kumar and A. C. Pandey, J. Lumin., 2009, 129, 1078–1082.
65. X. Zhang, Y. Chen, S. Zeng, L. Zhou, J. Shi and M. Gong, Ceram. Int., 2014, 40, 14537–14541.
66. H. Liu, Y. Hao, H. Wang, J. Zhao, P. Huang and B. Xu, J. Lumin., 2011, 131, 2422–2426.
67. H. Ullah, Z. H. Yamani, A. Qurashi, J. Iqbal and K. Safeen, J. Mater. Sci.: Mater. Electron., 2020, 31, 17474–17481.
68. P. P. Ortega, B. Hangai, H. Moreno, L. S. R. Rocha, M. A. Ramírez, M. A. Ponce, E. Longo and A. Z. Simões, J. Alloys Compd., 2021, 888, 161517.
69. B. Sanches de Lima, P. R. Martinez-Alanis, F. Guell, W. A. dos Santos Silva, M. I. Bernardi, N. L. Marana, E. Longo, J. R. Sambrano and V. R. Mastelaro, ACS Appl. Electron. Mater., 2021, 3, 1447–1457.
70. Q. Zhang, G. Xie, M. Duan, Y. Liu, Y. Cai, M. Xu, K. Zhao, H. Tai, Y. Jiang and Y. Su, ACS Appl. Nano Mater., 2023, 6, 17445–17456.
71. R. T. Parayil, B. Bhagat, S. K. Gupta, K. Mukherjee and M. Mohapatra, Phys. Chem. Chem. Phys., 2024, 26, 7424–7434.
72. R. T. Parayil, S. K. Gupta, R. Rohilla, J. Prakash, K. Sudarshan and M. Mohapatra, ACS Appl. Electron. Mater., 2023, 5, 5151–5163.
73. T. Das, S. Mojumder, S. Chakraborty, D. Saha and M. Pal, Appl. Surf. Sci., 2022, 602, 154340.
74. J. Wu, Z. Zheng, H. Chi, J. Jiang, L. Zhu and Z. Ye, ACS Appl. Mater. Interfaces, 2024, 16, 9126–9136.
75. K. Yuan, C. Y. Wang, L. Y. Zhu, Q. Cao, J. H. Yang, X. X. Li, W. Huang, Y. Y. Wang, H. L. Lu and D. W. Zhang, ACS Appl. Mater. Interfaces, 2020, 12, 14095–14104.
76. T. Das, S. Mojumder, D. Saha and M. Pal, Sens. Actuators, B, 2024, 406, 135358.
77. Y. Chen, H. Li, D. Huang, X. Wang, Y. Wang, W. Wang, M. Yi, Q. Cheng, Y. Song and G. Han, Mater. Sci. Semicond. Process., 2022, 148, 106807.
78. S. Zhang, M. Yang, K. Liang, A. Turak, B. Zhang, D. Meng, C. Wang, F. Qu, W. Cheng and M. Yang, Sens. Actuators, B, 2019, 290, 59–67.
79. Z. Wen, H. Ren, D. Li, X. Lu, S. W. Joo and J. Huang, Sens. Actuators, B, 2023, 379, 133287.
80. D. Chen, J. Xu, Z. Xie and G. Shen, ACS Appl. Mater. Interfaces, 2011, 3, 2112–2117.
81. J. Lu, C. Xu, L. Cheng, N. Jia, J. Huang and C. Li, Mater. Sci. Semicond. Process., 2019, 101, 214–222.
82. W. Ge, S. Jiao, Z. Chang, X. He and Y. Li, ACS Appl. Mater. Interfaces, 2020, 12, 13200–13207.
83. T. Zhou, T. Zhang, J. Deng, R. Zhang, Z. Lou and L. Wang, Sens. Actuators, B, 2017, 242, 369–377.
84. L. Ma, S. Y. Ma, X. F. Shen, T. T. Wang, X. H. Jiang, Q. Chen, Z. Qiang, H. M. Yang and H. Chen, Sens. Actuators, B, 2018, 255, 2546–2554.
85. X. Sun, H. Ji, X. Li, S. Cai and C. Zheng, Mater. Lett., 2014, 120, 287–291.
86. P. A. Murade, V. S. Sangawar, G. N. Chaudhari, V. D. Kapse and A. U. Bajpeyee, Curr. Appl. Phys., 2011, 11, 451–456.
87. H. Avireddy, H. Kannan, P. Shankar, G. K. Mani, A. J. Kulandaisamy and J. B. B. Rayappan, Mater. Chem. Phys., 2018, 212, 394–402.
88. H. Y. Lee, J. H. Bang, S. M. Majhi, A. Mirzaei, K. Y. Shin, D. J. Yu, W. Oum, S. Kang, M. L. Lee, S. S. Kim and H. W. Kim, Sens. Actuators, B, 2022, 359, 131550.
89. S. Zhang, P. Song, J. Zhang, H. Yan, J. Li, Z. Yang and Q. Wang, Sens. Actuators, B, 2017, 242, 983–993.
90. G. Feng, Y. Che, S. Wang, S. Wang, J. Hu, J. Xiao, C. Song and L. Jiang, Sens. Actuators, B, 2022, 367, 132087.
91. C. N. Wang, M. X. Peng, L. J. Yue, X. Y. Yang and Y. H. Zhang, Sens. Actuators, A, 2022, 342, 113650.
92. S. Neogi and R. Ghosh, Sens. Actuators, B, 2024, 415, 135980.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ma00617h

---