Design principles for achieving red emission in Eu²⁺/Eu³⁺ doped inorganic solids

Doping metal ions with luminescent behaviors into inorganic solids has been regarded as an effective as well as significant method for the design and discovery of new luminescent materials, and, therefore, doped inorganic phosphors contain two parts: host lattice and luminescent center.^1,2 Normally, small amounts of rare earth (RE) ions or transition metal ions can act as the doped luminescent centers, and the host lattice provides certain crystallographic sites for the doped activators. Among all types of phosphors, RE ion doped phosphors have attracted much interest due to their abundant energy levels and tremendous allowed energy level transitions. RE elements are comprised of 17 chemical elements including the 15 lanthanides (Ln, from lanthanum to lutetium), scandium, and yttrium.³ According to the transition characteristics, the luminescent mechanism of Ln doped phosphors can be classified into two types: f–d transition with broad band emissions and f–f transition with sharp line emission.⁴ In fact, direct excitation of Ln³⁺ ion is a relatively inefficient process due to the parity forbidden character of the f–f transitions.⁵ Nevertheless, some host materials (such as $WO42−$ and $MoO42−$ based materials) or other co-doping ions with a higher absorption coefficient than Ln³⁺ emitters can effectively transfer the absorbed energy to Ln³⁺ ions and thus enhance their luminescence.⁶ In contrast to the parity forbidden f–f transitions, the 4fⁿ–4fⁿ⁻¹5d¹ electrical–dipole allowed transitions possess high emissive probability and short decay times. Moreover, unlike the fixed sharp emission line within f–f type transition, the f–d transition is sensitive to the local crystal field and thus features much broader absorption and emission spectra.⁷ 

Enlightened by the factors described above, RE-doped phosphors have been applied in different optoelectronic devices.^8–10 The commonly used RE ions in phosphors are Eu²⁺/Eu³⁺, Ce³⁺, Tb³⁺, Gd³⁺, Yb³⁺, Dy³⁺, Sm³⁺, Tm³⁺, Er³⁺, Pr³⁺, and Nd³⁺. Particularly, Eu possesses both divalent and trivalent states, which can realize f–d transition (Eu²⁺) or f–f (Eu³⁺) transition in different hosts, respectively.⁴ As for the 4f–5d transition of Eu²⁺ ions, the electronics of 5d levels are exposed to the surface of electronic shell; thus, the excitation and emission spectra are strongly affected by the surrounding environment of Eu²⁺ ions, such as crystal symmetry, atom coordination, and crystal field strength.¹¹ As for the f–f transition of Eu³⁺ ions, the 4f orbital is located in the interior of the ion and is shielded from the surrounding environment by the filled 5s² and 5p⁶ orbitals.¹² Therefore, the host lattice has a small (but important) influence on the optical transitions within the 4fⁿ configuration.¹ Thus, the emission spectra of Eu³⁺/Eu²⁺ doped phosphor can be designed by modifying the chemical composition of the host or controlling local structures around activators (Fig. 1).¹³ Consequently, Eu²⁺ and Eu³⁺ doped phosphors have become a research hotspot and have been widely used in white LED lighting (WLED), backlight displays, plant lighting, biomedical detection regions, and so on.^14–17 

Thanks to the rapid development of LED chips with high efficiency and low cost, phosphor converted white LEDs (PC-WLED) based on blue InGaN chips have become the state-of-art lighting devices and prominent backlight source for liquid crystal displays (LCDs). On the one hand, commercial WLEDs are manufactured by combining a blue LED chip with yellow emitting Ce-doped yttrium aluminum garnet (YAG:Ce) phosphor.^18,19 However, the lack of a red component leads to a cool white light emission with a high correlated color temperature (CCT > 4000 K) and low color rendering index (CRI < 75) [Fig. 2(a)].²⁰ High-quality warm-white light demands a color rendering index of more than 80 and a low correlated color temperature of 2700–4000 K.²¹ To achieve this goal, two modified ways could be adopted: one is to add red phosphors into blue chip pumped YAG phosphor systems; the other is to use a blue LED pumped with mixtures of green and red phosphors, as shown in Figs. 2(b) and 2(c).²⁰ Obviously, either way, it is crucial to develop highly efficient red phosphors. On the other hand, the most commercial PC-WLED backlight for LCDs is fabricated by a combination of blue InGaN chip with green-emitting phosphor (β-SiAlON:Eu²⁺) and narrow-band red-emitting phosphor (K₂SiF₆:Mn⁴⁺) [Fig. 2(d)].^15,22,23 However, the unstable chemical property and high cost of K₂SiF₆:Mn⁴⁺ restrict its popular applications, and massive evaporation of HF gas during the sample preparation process pollutes the environment. Hence, a novel red phosphor with narrow-band emission is also urgently required.

In the last two decades, efforts have been made by researchers to explore red phosphors that can realize high-quality white light illumination or broaden color-gamut of LCDs. Among them, Eu²⁺/Eu³⁺-doped red phosphors have been widely investigated. The previous warm WLED was fabricated by packaging CaS:Eu²⁺ phosphor as the red component.^24–26 These Eu²⁺ doped sulfide phosphors have attractive luminescent properties, but also show serious drawbacks, such as low thermal stability and poor chemical instability. Alternatively, Eu³⁺ doped red phosphors including Y₂O₃:Eu³⁺ and Y₂O₂S:Eu³⁺ have been extensively used in lighting and television screens due to its narrow-band red emission around 610 nm.^27–29 Nevertheless, the weak absorption in the blue light region leads to a low external quantum efficiency and further limits their application. Recently, Eu²⁺ doped nitride phosphors, such as Sr₂Si₅N₈:Eu²⁺, Ca₂Si₅N₈:Eu²⁺, and CaAlSiN₃:Eu²⁺, have been commercialized in warm WLEDs.^30–32 Theses nitrides can be effectively excited by blue LEDs and show intense red emission due to the large polarizability of N³⁻ ligand and strong covalent of Eu–N bonds. Unfortunately, such warm LED devices have lower luminous efficiencies because of the broad emission band that extends to the deep red region beyond the human eye's long-wavelength sensitivity limit (∼700 nm). Moreover, their broad excitation spectra always overlay with the emission spectra of YAG, which induces cascade excitation and decreases the efficiency of WLEDs.³³ Though some nitride phosphors such as Sr(LiAl₃N₄):Eu²⁺ possess high luminescence efficiency and narrow-band emission with a full-width at half-maximum (FWHM) of 50 nm, the cascade excitation and poor chemical stability hinder its commercialization.³⁴ It is, therefore, necessary to discover new Eu²⁺/Eu³⁺-doped red phosphors to meet the needs of high-quality lighting and display applications.

Indeed, many researchers have recently focused their attention on the design principles to tune or improve the luminescent properties of phosphors.³⁵ For example, Liu's group adopted some useful ways, such as “Cation-Size-Mismatch effect” and “Neighboring-Cation Substitution effect,” to modify luminescent properties of the existed phosphors.^36–38 Kalaji et al. found that the distortion of the local crystal structure leads to a large redshift in Ca₂SiO₄:Ce.³⁹ In addition, Ye et al. reviewed the adjustment principles of excitation and emission spectra based on the crystal field theory.⁴⁰ Lin's group summarized the current methods of achieving spectral control from the perspective of local crystal structure changes.²⁰ Besides, our group proposed the mineral-inspired phosphor discovery strategy, which expounds the design principles of developing new phosphors based on the existed structural models.⁷ Furthermore, our group recently reviewed the structural strategies on manipulation of Eu²⁺ emission in silicates.⁸ However, few papers have summarized the design principles of Eu²⁺/Eu³⁺-activated red phosphors. To fill in this blank, herein we will summarize the design principles for Eu²⁺/Eu³⁺-doped red-emitting phosphors based on the dominant factors that contribute to the luminescent properties of Eu²⁺/Eu³⁺ ions. Moreover, the present challenges in red phosphors are highlighted, together with the prediction of future development trend.

The luminescence of the Eu²⁺ ion comes from the parity-allowed electronic transitions from the 4f⁶5d¹ excited state to the 4f⁷ ground state. In general, Eu²⁺-based phosphors feature broad absorption and emission spectra owing to the high sensitivity of 5d orbital to the surrounding environment. For a free Eu²⁺ ion, there is a large energy gap (∼4.2 eV, 34 000 cm⁻¹) between the lowest 5d excited state and the 4f ground state.⁷ Therefore, no visible emission could be observed for the free Eu²⁺ ion. When the Eu²⁺ ion is doped in a given crystalline material, the 5d energy level will be strongly affected by the crystal lattice environment, such as covalence, polarizability, crystal field strength, bond length, atom coordination numbers (CNs), and crystal symmetry.⁷ As a result, the 5d energy level is lowered and presented as redshift in emission spectrum including the centroid shift ɛ_c, crystal field splitting (ɛ_cfs), and Stokes shift (ΔS) [Fig. 3(a)].³ The emission energy of Eu²⁺-activated phosphors can be written as E_em = E_free − ɛ_c − ɛ_cfs − ΔS.⁴¹ Therefore, different emission spectra of Eu²⁺ ions varying from near ultraviolet (NUV) to near infrared (NIR) could be achieved by controlling the environment surrounding Eu²⁺ ions.

In contrast to the Eu²⁺ ion, the crystal field has a small effect on the emission of the Eu³⁺ ion because the partially filled 4f shell is shielded the from 5s²5p⁶ orbital.⁴² Thus, Eu³⁺-based phosphors are famous for their commonly line-shaped red emission originated from the forbidden ⁵D₀–⁷F_J transitions (J = 0, 1, 2, 3, 4), as shown in Fig. 3(b).^43,44 Actually, other high energy emissions were also observed in some Eu³⁺-doped materials, such as the transitions from ⁵D₁ (green), ⁵D₂ (green, blue), and ⁵D₃ (blue) level to ⁷F_J levels.¹ However, these transitions are often neglected because the higher ⁵D_1,2,3 emissions are quenched by cross relaxation (induced by short Eu³⁺–Eu³⁺ distance) when the doping concentration is high. In addition to the characteristic sharp excitation line originated from the transition of ⁷F_J → ⁵D_1,2,3, Eu³⁺ ion doping compounds, especially Eu³⁺-doped nitrides or oxides, also possess a broad charge transfer band (CTB) formed by the electronic transition from the 2p orbital of O²⁻ or N³⁻ to the 4f orbital of the Eu³⁺ ion.^45,46 Unlike the weak absorption of f–f transitions, the broad CTB makes commercialization become possible. For example, as one of the most promising red phosphors, Y₂O₃:Eu³⁺ has been widely applied to field emission display applications.²⁸ 

As mentioned above, the emission energy of the Eu²⁺ ion is mainly determined by three elements: centroid shift ɛ_c, crystal field splitting ɛ_cfs and Stokes shift ΔS. Therefore, realizing larger ɛ_c, ɛ_cfs, or ΔS values becomes the key factor to design new Eu²⁺-doped red phosphor. Some proposed design principles are summarized as follows.

The relatively short distance and strong covalency of Eu–N and Eu–S chemical bonds in nitride or sulfide solids result in strong crystal field strength and large nephelauxetic effect, then yielding large redshifts in both emission and excitation spectra. Thus, research studies on red-emitting phosphors previously focused on the Eu²⁺-doped nitride or sulfide phosphors, and many commercial red phosphors such as CaS:Eu²⁺, SrS:Eu²⁺, Sr₂Si₅N₈:Eu²⁺, and CaAlSiN₃:Eu²⁺ have been successfully synthesized.^24,26,31,32 As the representatives, SrS:Eu²⁺ and Sr₂Si₅N₈:Eu²⁺ can be effectively excited by blue light and exhibit extensive red emission peaking at 615 and 618 nm, respectively, as shown in Figs. 4(a) and 4(b). However, nitride or sulfide phosphors also have some serious drawbacks. Normally, sulfide phosphors are chemically unstable in the atmosphere due to their deliquescent behaviors. Therefore, it should continue to explore the emerging sulfide phosphors with new crystal structures and promising luminescent properties. Moreover, phosphor coating technology is urgently needed to improve the chemical stability of sulfide phosphors. With regard to nitride phosphors, they are usually synthesized under tough conditions such as high temperature and high pressure, which causes the energy loss. In addition, the particle size, morphology, and crystallinity of nitride phosphors are out of control. Thus, the discovery of novel Eu²⁺-doped nitride and sulfide red phosphors is still an endless mission for scientists.

For inorganic phosphors, the nearest neighbor anions around Eu²⁺ ions have a great effect on the 5d levels. This effect was described by centroid shift ɛ_c, which is related to the nephelauxetic effect. The nephelauxetic effect is directly linked to the covalency between Eu²⁺ ions and anion ligands in the host lattice.⁴⁷ For an isomorphous structure, the shorter ionic radius and the higher negative charge will lead to the stronger covalency, thus causing a larger centroid shift. The nephelauxetic effect for different anion tends to decrease in the sequence of Se²⁻ > S²⁻ > N³⁻ > Cl⁻ > O²⁻ > F⁻. Dorenbos surveyed the luminescence of Eu²⁺ in a variety of inorganic compounds and demonstrated that the ɛ_c values tend to follow the nephelauxetic effect sequence as fluorides < chlorides < bromides < iodides < oxides < nitrides < sulfides (Fig. 5).⁴⁸ Normally, with the strengthening of nephelauxetic effect or covalent strength, both the bonding and antibonding orbitals will move to lower energies, leading to redshift emission.

Considering the relatively large centroid shift and highly chemical stability of Eu²⁺-doped nitride phosphors, N³⁻ anion was often used as a promising candidate for the anion substitutions to improve the luminescence properties of oxide phosphors. Take Sr₂SiO₄:Eu²⁺ for example, the emission peaks shifted from 563 to 583 nm through partial co-substitution of [Sr²⁺–O²⁻] with [Lu³⁺–N³⁻] unit [Fig. 6(a)].⁴⁹ The reason is that N³⁻ has less electronegative and stronger polarization than O²⁻, which increases the strength of the covalent bond. By singly substituting O²⁻ with N³⁻, Zhao et al. successfully synthesized red-emitting phosphor Sr₂Si(N,O)₄:Eu²⁺.⁵⁰ Compared with the emission peak of Sr₂SiO₄:Eu²⁺ at 560 nm, the maximum emission peak at 617 nm in Sr₂Si(N,O)₄:Eu²⁺ phosphor is attributed to the formation of strong Eu²⁺–N³⁻ covalent bond [Fig. 6(b)]. Thus, introducing N³⁻ into Eu²⁺-doped oxide phosphors can be served as a strategy to adjust centroid shift and further realizes red emission.

For the free Eu²⁺ ion, the five 5d-orbitals have the same energy. When the Eu²⁺ ion is doped into a crystalline lattice, the 5d-orbitals of Eu²⁺ will be affected by the crystal field and the degenerate 5d levels will be split into groups with different energies.⁵¹ The degree of crystal field splitting can be described by ɛ_cfs, which is generally related to the type, size, and coordination numbers (CNs) of polyhedrons occupied by Eu²⁺. The crystal field splitting appears to behave as $εcfs=βpolyQRav−2$⁠.⁵² Herein, $βpolyQ$ is a constant depends on the type of polyhedron and R_av is related to the individual bond length and ionic radius of cations occupied by Eu²⁺ ion. The shorter bond length, smaller CNs, and higher distortion index of polyhedron will lead to stronger crystal field splitting and longer emission wavelength. Thus, red emission can be realized by adjusting crystal field splitting ɛ_cfs.

Typically, the CNs of transition metal elements and main-group elements are about 2–6, while the CNs of rare earth ions in inorganic hosts are greater than or equal to 6. If we can manipulate sites occupation of Eu²⁺ ion in a host with small CNs, red or even near infrared (NIR) emission will be realized. Based on this assumption, our group has established a new design principle on the discovery of red phosphors, i.e., site-selective occupation of Eu²⁺ at polyhedrons with small CNs to achieve blue light-excited red/NIR emission. For example, K₃YSi₂O₇ unit cell contains five different polyhedrons K1O₆, K2O₈, K3O₉, Y1O₆, and Y2O₆, and Eu²⁺ will selectively occupy both the highly distorted K2O₈ polyhedrons and Y2O₆ polyhedrons with small CNs.⁵³ Consequently, a large crystal field splitting in 5d energy levels is realized and the orange–red-emitting for K₃YSi₂O₇:Eu phosphor could be observed [Fig. 7(a)]. In addition, Rb₃YSi₂O₇ crystal belongs to the P6₃/mcm space group and contains three different polyhedrons. The selective occupation of Eu²⁺ in YO₆ and Rb2O₆ polyhedrons with small CN results in large crystal field splitting, and further leads to the broad red emission under blue light excitation, as shown in Fig. 7(b).⁵⁴ Moreover, we also successfully designed and synthesized the abnormal Eu²⁺-doped NIR phosphor K₃LuSi₂O₇:Eu²⁺ by substituting the large Rb⁺ and Y³⁺ ions in Rb₃YSi₂O₇:Eu²⁺ with small K⁺ and Lu³⁺ ions. This substitution further increases the crystal field splitting effect, and the first Eu²⁺-doped NIR phosphor that gives a broad band NIR emission (600–900 nm) under 460 nm blue light excitation was obtained [Fig. 7(c)].⁵⁵ The small CNs’ principles have been used to design other red/NIR-emitting phosphors. For example, Li et al. successfully achieved NIR emission in Bi³⁺-activated XAl₁₂O₁₉ (X = Ba, Sr, Ca) phosphor by constructing the selective site occupation of Bi³⁺ in AlO₄ polyhedrons.⁵⁶ Therefore, occupying polyhedrons with small CNs can be employed as an effective design principle for the discovery of red/NIR phosphor. Recently discovered Sr₃LnAl₂O_7.5 (Ln = Y or Lu) and Mg₂Al₄Si₅O₁₈:Eu²⁺ phosphors by our group could further support this, with red emission originating from Eu²⁺ at the six-coordinated YO₆ and LuO₆ polyhedrons and vacant channels, respectively.^57,58

As mentioned in Sec. III C, phosphor is composed of “host + activator.” Normally, there are more than one type of polyhedrons in host lattices. Eu²⁺ ions can distribute in different polyhedrons by controlling the doping concentration, and thus not only the luminescence efficiency can be optimized but also the emission peak can be tuned. For example, Sato et al. reported a new dicalcium silicate phosphor Ca_2−xEu_xSiO₄, which realized tunable emission from 580 to 650 nm by increasing doping concentration.⁵⁹ When the doping level is low (x < 0.2), Ca_2−xEu_xSiO₄:Eu²⁺ shows extensively yellow emission peaking at 580 nm, by virtue of the large bond length caused by the occupation of Eu²⁺ ions in Ca1nO₇ sites. When x > 0.2, the emission color gradually transforms to red because more Eu²⁺ ions tend to occupy Ca2nO₆ sites with shorter bond length [Fig. 8(a)]. The substitution of Eu²⁺ ions in the small Ca2n sites causes strong crystal field splitting and further leads to the red emission. The similar phenomenon has also observed in (SrBa)₂SiO₄:xEu²⁺ phosphor. XRD refinement results revealed that the Eu²⁺ ion is distributed in two different crystallographic sites and Eu²⁺ located at the nine-coordination sites at low doping concentration. As shown in Fig. 8(b), with increasing concentration of Eu²⁺, the ratio of the red emission with lower energy increases, whereas the high energy ones decrease, resulting in redshifts of photoluminescence (PL) and photoluminescence excitation (PLE) spectra.⁶⁰ However, very few phosphors possess this property, and high doping concentrations will lead to luminescence quenching. Thus, this design principle has certain limitations and deep research studies are needed in the future.

The Stokes shift is defined as the energy difference between excitation and emission spectra, which is related to the parabola offset that is induced by lattice relaxation at the excited states.⁶¹ According to the previous work of Gettinger et al.,⁶² the Stokes shift has an inversely proportional correlation with the rigidity of the host lattice, whereas Daicho et al. believe that ΔS is related to the symmetry of Eu-doping sites, and the higher distorted condition will result in the larger Stokes shift due to a significant reorganization of the atoms around the luminescent center.⁶³ For example, K₂Ca(PO₄)F:Eu²⁺ has a large ΔS value about 1.38 eV, which is large enough to convert NUV light to red luminescence without the absorption at the region of blue and green light [Fig. 9(a)].⁶³ For the sites of Eu²⁺ in K₂Ca(PO₄)F:Eu²⁺, the 5d₁ orbital can hardly extend toward the F⁻ ion side because that F⁻ is more electronegative than the O²⁻ ion. The O–Eu–O angle increased due to the rotation of the PO₄ tetrahedra to alleviate the electrostatic repulsion with the 5d₁ electron [Fig. 9(b)]. As a result, 5d₁ electrons localized in the asymmetrical space and created a large ΔS. This method should be used to adjust the small ΔS in Eu²⁺-activated nitride phosphors.

Narrow-band emitters play a key role in WLED backlights that are made by the combination of blue emitting LED chip with narrow green- and red-emitting phosphors. Nowadays, β-SiAlON:Eu²⁺ phosphor has been commercialized in LCDs owing to its excellent luminescence properties of narrow green emission band (FWHM 55 nm) and high quantum efficiency.⁹ However, the narrow-band red-emitting phosphors activated by the Eu²⁺ ion are still in vacancy. Up to now, only a few Eu²⁺ doped nitride/oxynitride phosphors with the UCr₄C₄-type structure exhibit narrow-band red emission. The first narrow-band red-emitting phosphor Sr[LiAl₃N₄]:Eu²⁺ possesses a triclinic crystal structure with UCr₄C₄-type, which contains a highly rigid network formed by order corner- and edge-sharing AlN₄ and LiN₄ tetrahedra [Fig. 10(a)].³⁴ Sr²⁺ sites are filled in ring channels with high symmetric cubic polyhedrons. The occupation of Eu²⁺ in Sr²⁺ sites with highly symmetric cubic polyhedrons and rigid network results in abnormal narrow-band red emission (FWHM 50 nm). Soon afterward, Schnick and co-workers discovered other UCr₄C₄-type narrow red-emitting phosphors, such as Ca[LiAl₃N₄]:Eu²⁺ (FWHM = 60 nm) and Sr[Mg₃SiN₄]:Eu²⁺ (FWHM = 43 nm).^64,65 Recently, Hoerder et al. reported a new red-emitting phosphor Sr[Li₂Al₂O₂N₂]:Eu²⁺, which shows a narrow-band emission peaking at 614 nm with a FWHM of 48 nm [Fig. 10(b)].⁶⁶ It is worth noting that the crystal structure of Sr[Li₂Al₂O₂N₂] also belongs to the ordered variant of UCr₄C₄-type. Therefore, wide attention should be paid to UCr₄C₄-type nitrides in future research on Eu²⁺ activated narrow-band red-emitting phosphors.

Compared with Eu²⁺ doped red phosphors, Eu³⁺ activated red phosphors have better color purity due to their sharp line emission originating from f–f transition of Eu³⁺ ions.⁶⁷ However, there are some defects for Eu³⁺ activated phosphors. Take the commercial Y₂O₃:Eu³⁺ red phosphor for example, it can be only excited by UV light and shows red emission peaking at 610 nm.⁶⁸ Nevertheless, the PLE spectra of Y₂O₃:Eu³⁺ cannot be matched well with the blue or NUV chips. Therefore, the biggest challenge for Eu³⁺ doped red phosphors is adjusting the position of CTB. In addition, the ⁵D₀ → ⁷F₁ (590 nm) and ⁵D₀ → ⁷F₂ (610 nm) transitions are insensitive and sensitive to chemical environment, respectively, and the ⁵D₀ → ⁷F₂ transition will be dominant when Eu³⁺ occupies an non-symmetric site. Thus, high color purity can be achieved by controlling the sites occupation of Eu³⁺ ions in an unsymmetric site. Furthermore, the luminescent efficiencies of Eu³⁺ activated phosphors are still low and need to further optimize. Considering that the above problems existed in Eu³⁺ doped phosphors, four design principles to improve luminescence properties of Eu³⁺ are summarized.

Charge transfer (CT) describes the process in which Eu³⁺ ions obtain electrons from the coordination anions like O²⁻ or N³⁻ ions, and the energy of CT can be interpreted by Jorgensen's empirical formula⁶⁹ 

where $χopt(x)$ and $χopt(m)$ are the optical electronegativity of the anion and metal ion, respectively. Based on this formula, Dorenbos presented a compendium of CT energy of Eu³⁺ in different types of host materials, as shown in Fig. 11(a). Obviously, the CT energy changes with the type of anions and tends to decrease in a sequence: fluorides > oxides > nitrides > sulfides. Moreover, for the materials with same type of anions, the CT energy decreases gradually with the increase of average distance to the surrounding anions [Fig. 11(b)]. Therefore, developing Eu³⁺ doped nitride or sulfide phosphors seems to have higher probability of obtaining low CT energy, and the occupation of Eu³⁺ in a larger crystalline site tends to realize lower CT energy. Take α-M₃B₂N₄ (M = Ca, Sr):Eu³⁺ for example, both of them show low CT energy (300–400 nm) compared with oxide phosphor (250–310 nm), which is due to the relatively high optical electronegativity of N³⁻ than O²⁻ [Fig. 12(a)].⁷⁰ In addition, the excitation wavelength of α-Ca₃B₂N₄:Eu³⁺ is shorter than that of α-Sr₃B₂N₄:Eu³⁺. This is because the electron cloud of Sr²⁺–N³⁻ group is smaller than that of Ca²⁺–N³⁻ group, and thus Sr²⁺–N³⁻ group needs lower energy compared with the Ca²⁺–N³⁻ group [Fig. 12(b)]. The broad CTB peaking at the NUV region was also observed in other nitrides phosphors such as LiCaAlN₂:Eu³⁺.⁷¹ These results indicate that the CTB of Eu³⁺-activated nitride or sulfide phosphors has more probability to match with the commercial NUV LEDs.

Except the CTB formed by Eu–N and Eu–O groups, other CTBs formed by Mo–O, V–O, and W–O in Eu³⁺-doped molybdate, vanadate, and tungstate phosphors have also attracted increasing interest owing to their broad and intense absorption in the UV/NUV region.^72–74 Importantly, the captured energy by Mo–O, V–O, and W–O CTB can be efficiently transferred to the Eu³⁺ ion.⁷⁵ As shown in Fig. 12(c), MgWO₄:Eu³⁺ and MgMoO₄:Eu³⁺ phosphor possess broad W⁴⁺–O²⁻ and Mo⁴⁺–O²⁻ CTB blended with the Eu³⁺–O²⁻ CTB.⁷⁶ However, their CTBs are still located in the UV region and are insufficient to be effectively pumped by NUV or blue LED chips. Recently, researchers found that the position of CTB in MoO_x-based phosphors is related to the coordination number x, and the higher value of x contributes to the longer excitation wavelength.⁷⁷ By comparing the position of CTB in Figs. 12(c) and 12(d), it can be seen that the CTBs of A₂MoO₆:Eu³⁺ (A = Y, Gd) phosphors shift obviously to the longer wavelength than that of MgWO₄:Eu³⁺ and MgMoO₄:Eu³⁺ phosphors.⁷⁶ This is attributed to the higher coordination number of transition metal ion resulting in a longer and weaker Mo⁶⁺–O²⁻ bond, and thus the energy of the Mo⁶⁺–O²⁻ bond extends the CTB edge from 330 nm (for MoO₄ groups) to 420 nm (for MoO₆ groups). Therefore, synthesizing Eu³⁺-doped molybdate phosphors with a high coordination number (MoO₆ groups) is an effective way to reduce CTB energy.

Energy transfer (ET) among lanthanide ions has been used to develop efficient phosphors for lighting, lasers, optical markers, and solar cells. It is necessary to explore sensitizers for Eu³⁺ to realize appropriate CTB that overlaps with the commercial InGaN chips. A group of possible transitions leading to effective luminescence under near-UV excitation are the allowed 4fⁿ → 4fⁿ⁻¹5d¹ transitions in Eu²⁺ or Ce³⁺. However, Eu²⁺/Ce³⁺ is unlikely to directly sensitize Eu³⁺ luminescence in virtue of the metal–metal charge transfer (MMCT) quenching.⁷⁸ One way to simultaneously avoid the MMCT quenching and sensitize Eu³⁺ emission is adding intermediate ions that can act as a bridge for energy transfer between Eu²⁺/Ce³⁺ and Eu³⁺. Blasse first used Gd³⁺ as an intermediate in (Y,Gd)F₃:Ce³⁺, Eu³⁺ phosphor and realized Ce³⁺ → (Gd³⁺)n → Eu³⁺ energy transfer.⁷⁹ Replacing Gd³⁺ by Tb³⁺, YBO₃:Ce, Tb, Eu phosphor exhibits strong absorption bands in the UV and NUV regions produced by the Ce³⁺ → (Tb³⁺)n → Eu³⁺ ET scheme.⁷⁸ Some other novel red phosphors, such as Na₂Y₂B₂O₇:Ce³⁺, Tb³⁺, Eu³⁺, Y₁₀Al₂Si₃O₁₈N₄:Ce³⁺, Tb³⁺, Eu³⁺, and Y₂SiO₅:Ce³⁺, Tb³⁺, Eu³⁺ have been synthesized by using this strategy.^80–82 Take Y₂SiO₅:Ce³⁺, Tb³⁺, Eu³⁺ for example, Y₂SiO₅:Ce³⁺ possesses several broad absorption bands at 270, 300, and 355 nm [Fig. 13(a)]. When Tb³⁺ was employed as the dopant, the corresponding PLE spectra monitored at 545 nm are similar with that of Y₂SiO₅:Ce³⁺, which evidences the existence of energy transfer between Ce³⁺ and Tb³⁺. Moreover, Eu³⁺ is effectively sensitized by Ce³⁺ → (Tb³⁺)n → Eu³⁺ energy transfer [Fig. 13(b)].⁸² As a result, not only the luminescence of Eu³⁺ was enhanced but also the excitation wavelength was tuned to 360 nm attributed to the allowed 4f–5d absorption of Ce³⁺ ions.

Though Eu³⁺ shows luminescence in most host materials, the luminescence intensity is still low and needs to be further improved. Particularly for the phosphors in which Eu³⁺ occupies the divalent crystalline sites, the doping concentration is restricted for the sake of charge unbalance. Alkali metal ions Li⁺, Na⁺, and K⁺ with distinct ionic radii and low oxidation states were often used as charge compensation ions to modify the local site symmetry and improve the Eu³⁺ luminescence efficiency.^83–86 For example, the substitution of Eu³⁺ at Mg²⁺ sites in MgAl₂O₄:Eu³⁺ nanophosphor gives rise to a net positive charge in the system due to different valence states of Eu³⁺ and Mg²⁺. Hence, some Eu³⁺ ions are incapable of entering the lattice sites and tend to aggregate together, which results in low efficiency. On this condition, charge compensation becomes important. Hence, Li⁺ ions were introduced to achieve an electrically neutral environment. Obviously, the excitation intensity of MgAl₂O₄:Eu³⁺ were enhanced with the charge balance mechanism of $2V′′Mg+Eu3++Li+→EuMg+Li′Mg$⁠, as shown in Fig. 14(a).⁸⁷ Moreover, by co-doping Na⁺ into CaWO₄:Eu³⁺ phosphor, the quantum efficiency of CaWO₄:Eu³⁺, Na⁺ reaches 92% [Fig. 14(b)].⁸⁸ The charge balance of Eu³⁺ substitution in Ca²⁺ sites can be realized via paths: 2Ca²⁺ → Eu³⁺ + Na⁺. Similar luminescence enhancements were also observed in other Eu³⁺ activated phosphors, such as Li⁺ doped Y₂O₃:Eu³⁺ and (Y,Gd)₂O₃:Eu³⁺ phosphors.⁸⁹ However, there is no charge balance problem in these systems and thus Li⁺ may only act as a flux in these phosphors.

In general, the electron dipole transition of ⁵D₀–⁷F₂ is very sensitive to the environment, whereas the magnetic dipole ⁵D₀–⁷F₁ transition is irrelevant to the environment. Thus, Eu³⁺ is often used to probe the local structure changes in the crystal lattice.⁹⁰ When Eu³⁺ occupies an unsymmetrical site, the dominant emission located at around 620 nm could be obtained with high color purity. On the contrary, symmetric site occupation of Eu³⁺ allows dominant emission peaked around 592 nm with low color purity. For example, the occupation of Eu³⁺ at symmetric sites in α-Sr₃B₂N₄ crystals leads to the sharp emission line at 593 nm [Fig. 15(a)], while Eu³⁺ ions occupy the unsymmetric Ca site in α-Ca₃B₂N₄ achieving red emission with high color purity at 613 nm [Fig. 15(b)].⁷⁰ Furthermore, researchers also found that Ba₂La_2/3TeO₆:Eu³⁺ exhibits orange–red color with chromaticity coordinates of (0.629, 0.369), ascribed to Eu³⁺ ions occupying the symmetric octahedral LaO₆ sites [Fig. 15(c)].⁹⁰ On the contrary, the unsymmetric YO₈ sites occupied by Eu³⁺ ions in CaYAl₃O₇ host lattice leads to a high color purity red emission with chromaticity coordinates of (0.6303, 0.3607), as exhibited in Fig. 15(d).⁹¹ From the perspective of discovering new Eu³⁺-doped red phosphors, it is necessary to select host materials with unsymmetric crystalline sites.

Researchers have made significant efforts to improve and develop Eu²⁺ and Eu³⁺ activated red phosphors to meet the various demands in high-quality LED lighting and high-definition displays. In this perspective, we mainly focus on the design principles of Eu²⁺/Eu³⁺ activated red-emitting phosphors. As for Eu²⁺ ion, the 5d–4f luminescence in the solids depends greatly on the surrounding environments. Based on this characteristic, we summarized six types of design principles for Eu²⁺ to realize red emission in a host lattice, including the synthesis of new Eu²⁺-doped nitride or sulfide phosphors, anionic substitution achieves large centroid shift, Eu²⁺ occupies polyhedrons with small CNs to obtain large ɛ_cfs, doping concentration controls the distribution of Eu²⁺ ions, mixed ligands induce large ΔS, and emerging narrow-band red emitters in Eu²⁺ doped UCr₄C₄-type nitrides. As for the Eu³⁺ ion, the main challenge is the weak absorption in NUV and blue light region due to the forbidden f–f transition. Thus, we proposed the design principles to adjust the charge transfer band and realize low energy excitation, including synthesis of Eu³⁺-doped nitride, sulfide, or molybdate phosphors to reduce CTB energy and doping sensitizer ions to realize low energy excitation. In addition, design principles of charge compensation and unsymmetrical sites occupation are proposed to improve the luminous efficiency and optimize the color purity of Eu³⁺ doped phosphors.

These design principles herein will prompt the development and optimization of Eu²⁺ and Eu³⁺ doped red phosphors for future studies. Looking forward, much effort will be required in further work, including but are not limited to the following: (1) although many design principles have been proposed to improve the luminescence properties of Eu²⁺ and Eu³⁺, there is still a lack of general rules applicable to all phosphor systems. Thus, more concerns should be focused on the deep understanding of structure-related luminescence mechanisms. (2) Currently, the main commercial Eu²⁺ doped phosphors are nitrides or sulfides. However, these phosphors suffer from some disadvantages, such as high cost and poor chemical stability. Therefore, much effort should be made to develop of Eu²⁺ activated oxide phosphors with a new design viewpoint. (3) The main problem of Eu³⁺ activated phosphor is the weak absorption from f–f transitions. Compared with the CT energy Eu³⁺–O²⁻, Eu³⁺–N³⁻ and Eu³⁺–S²⁻ have lower CT energy due to large covalency of N³⁻ and S²⁻. Hence, the research on Eu³⁺ doped nitride or sulfide phosphors should be concerned and encouraged to obtain low CT energy.

The authors acknowledge the support received from the National Natural Science Foundations of China (Grant Nos. 51972118 and 51961145101), International Cooperation Project of National Key Research and Development Program of China (No. 2021YEF0105700), Guangzhou Science & Technology Project (No. 202007020005), the State Key Laboratory of Luminescent Materials and Devices (No. Skllmd-2021-09), and the Local Innovative and Research Teams Project of Guangdong Pearl River Talents Program (No. 2017BT01X137).

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