Manganese(II) in Tetrahedral Halide Environment: Factors Governing Bright Green Luminescence.

Finding narrow-band light emitters for the visible spectral region remains an immense challenge. Such phosphors are in great demand for solid-state lighting and display application. In this context, green luminescence from tetrahedrally coordinated Mn(II) is an attractive research direction. While the oxide-ligand environment had been studied for decades, much less systematic efforts have been undertaken with regard to halide coordination, especially in the form of fully inorganic halide matrixes. In this study, we synthesized a series of hybrid organic-inorganic Mn(II) halides as well as a range of fully inorganic Zn halide hosts (chlorides, bromides, iodides) doped with Mn(II). In the latter, tetrahedral coordination is attained via substitutional doping owing to the tetrahedral symmetry of Zn sites. We find that the choice of the halide as well as subtle details of the crystal structure profoundly govern the photoluminescence peak positions (500-550 nm range) and emission line widths (40-60 nm) as well as radiative lifetimes (shorter for iodides) through the altered ligand-field effects and degrees of spin-orbit coupling. The photoluminescence quantum yields were as high as 70-90%. The major hurdle for the practical use of these compounds lies in their low absorption coefficients in the blue spectral regions.

Introduction

Light-emitting materials have a broad spectrum of applications including solid-state lighting, flat panel display technologies, optical data storage, radiation detection, and photovoltaics.[1−5] Narrow-band green-emitting phosphors with an emission peak around 520–535 nm are central for displays and solid-state lighting technologies.[6] In recent years, cadmium(II) chalcogenide (Se, S, or Te) nanocrystals (NCs), exhibiting photoluminescence (PL) full width at half maxima (fwhm) of <30 nm and PL quantum yields (QYs) of above 95%, have been used as phosphors in backlighting of displays.[7] Limited RoHS compliance and steadily decreasing public and commercial acceptance of heavy-metal-containing materials motivated commercial deployment of environmentally benign indium phosphide NCs[8] in spite of their considerably broader emission line width (38–40 nm) and lower stability. An emerging alternative are highly luminescent lead halide perovskite NCs (fwhm ≈ 20 nm at 520–530 nm PL peak),[9] whose practical potential still needs to be thoroughly examined. Thus far, the most successful commercial phosphors are still those based on emissive rare-earth or transition-metal ions embedded into a stable crystalline host.[6,10] In particular, β-SiAlON:Eu2+ (PL peak at 535 nm, fwhm ≈ 50 nm) is used in commercial LCDs,[11] while other Eu2+-based compounds comprise a subject of continued research efforts.[12,13]

Tetrahedrally coordinated Mn(II) exhibits green emission and is thus of interest for engineering solid-state phosphors. The emission of Mn(II) and Mn(IV) in octahedral (O) and tetrahedral (T) crystal fields has been studied for decades.[14,15] For instance, red-emissive K2SiF6:Mn4+ (octahedral coordination) has become a commercial red phosphor for white light-emitting diodes.[16] In a tetrahedral field (T) of oxide ligands,[17,18] green emission of Mn(II) is characterized by a fwhm below 45 nm and can be as narrow as 26 nm.[18] However, the main drawback of the oxide field is slow emission decay, typically 4–6 ms. The emission is due to intra-atomic transitions within the split energy states.[19,20] The magnitude of the splitting is quantified by B, C, and Δ parameters. The B and C parameters relate to the Coloumb (electron–electron) repulsion, and Δ quantifies the strength of the crystal field (CF). In addition, the so-called nephelauxetic parameter β = B′/B, where B and B′ are the parameters for a free d-metal ion and an ion in the complex, respectively, is often used to estimate the covalency of the metal–ligand bond. The nephelauxetic effect describes the influence of the ligand on the d-orbitals expansion upon ligation. Greater orbital expansion reduces the Coulomb repulsion. The variation of the Δ for different ligands is known as the spectrochemical series, whereas the nephelauxetic series reflects the magnitude of B parameter reduction compared to a free ion. Scheme and Table S1 illustrate the energy states splitting for Mn in T environment.[21,22] A spectrally narrow, i.e., 25–60 nm in fwhm, green emission peaked at 500–550 nm originates from the lowest 4T1 → 6A1 transition. The excitation spectra in the blue and near-ultraviolet (near-UV) features several bands, corresponding to two groups of transitions: 6A1 → 4G and 6A1 → 4D. In addition, interatomic Mn–Mn interaction can further modify the energy states. The Mn(II) emission in the octahedral environment is known to be in the red region, foremost due to the weaker ligand field, as compared to the T coordination. In addition, the distance between O-Mn(II) atoms had been reported to adjust the PL peak position in the orange-red region.[23] The d–d transition of Mn2+ in the centrosymmetric O is, however, forbidden by the Laporte selection rule, making the observed luminescence very weak. In noncentrosymmetric complexes the rule above is non-applicable, rendering a more efficient absorption and emission by Mn2+. In this study, we cast our attention toward Mn2+ green emission in a T halide environment (chloride, bromide, and iodide) to unveil the factors governing the PL characteristics.

Scheme 1

Energy States Splitting and Optical Transitions in Tetrahedrally Coordinated Mn2+ Ion

We pursue two approaches for realizing Mn(II) green emission in a T field of halide atoms: (a) structurally zero-dimensional hybrid organic–inorganic manganese(II) tetrahalides or (b) doping of hosts with T centers. The organic–inorganic manganese(II) tetrahalides (case a) comprise isolated MnX4 anions surrounded by the bulky organic cations. Although such hybrid compounds have been known for several decades,[20] they have recently regained popularity due to high PLQY of up to 90%,[24] spectrally narrow emission (fwhm ≈ 50 nm), and demonstration of electroluminescence with a high external quantum efficiency of up to 10%.[25,26] In some of these hybrids, triboluminescence (i.e., luminescence under mechanical stress) had been reported as well.[27] Case b is represented by either a fully inorganic matrix of binary zinc halides (ZnBr2, ZnI2) or ternary alkali zinc halides doped with Mn2+ (Figure ). Interestingly, despite the simplicity of the idea, ZnX2 had thus far not been reported as a host for Mn2+. For comparison, emission of Mn(II) in MgBr2 matrix stems from the octahedral coordination and hence is in the red region.[28] Both ZnBr2 and ZnI2 have the same structure comprising groups of four corner-sharing ZnX4 tetrahedra, extending in three dimensions. Although Mn(II) ions are larger than Zn2+ ions (0.66 vs 0.6 Å), they occupy exclusively T sites as they are not large enough for the stable coordination in O voids. Zero-dimensional alkali zinc halides, wherein ZnX4 units are disconnected, can be prepared in two common stoichiometries, A2ZnX4 and A3ZnX5 (A = Li, Na, K, Rb, Cs).[29−37] There are very few reports upon which one can try to build the comprehension of the factors governing the PL characteristics of these hybrids.[20,38] No reports on green emission from Mn in fully inorganic halide hosts could be identified. Furthermore, Mn doping of such matrices has remained rather unexplored; we found only one report on electron paramagnetic resonance measurements of manganese in Cs2ZnBr4 and Rb3ZnBr5[39] and a recent report on the synthesis and optical properties of Cs3MnBr5, which is isostructural to Cs3ZnBr5.[40]

Figure 1

Various compounds, studied in this work, featuring tetrahedral coordination of Mn(II). Mn ions are introduced either in the form of dopant (ZnX2, A2ZnX4, A3ZnX5) or as anion in organic–inorganic hybrids (Cat)2MnX4.

Experimental Section

The list of chemicals is available in the Supporting Information.

Synthesis of Organic–Inorganic Hybrids

Hybrid phosphors were obtained from MnX2 and organic cation halide precursor solutions in ethanol, methanol, or dimethylformamide. In addition, some compounds were obtained in a reaction of MnCO3 with organic cation in aqueous hydrohalic acids. For example, for the preparation of Bmpip2MnCl4, BmpipCl (0.4 mmol, 0.0767 g) and MnCl2 (0.2 mmol, 0.0252 g) were dissolved in 1.5 mL of MeOH with stirring. The solution was filtered through a 0.2 μm PTFE filter. For single-crystal growth, the crude solution was either slowly concentrated at 50 °C or placed into an antisolvent evaporation chamber with dichloromethane. In a typical synthesis from a hydrohalic acid, N-benzyl-N,N,N-trimethyl chloride (1 mmol, 0.1875 g) and MnCO3 (0.5 mmol, 0.0575 g) were dissolved in aqueous HCl (6 mmol, 0.53 mL) at 70 °C. Single crystals were grown by slow evaporation of the filtered precursor solution at room temperature.

Synthesis of Fully Inorganic Mn-Doped Hosts

Na2ZnBr4 and Li2ZnBr4 were prepared as follows. Dry ABr (A = Na or Li) and ZnBr2 mixed at a molar ratio of 2:1 was loaded in a Pyrex ampule and sealed under vacuum. The ampule was placed in a quartz tube and heated to the melting point (326 and 347 °C for Li2ZnBr4 and Na2ZnBr4, respectively) at a heating rate of 50 °C/h. Samples were kept at this temperature for 12 h and cooled to room temperature at a rate of 10 °C/h. For other ternary halides, samples were kept 10–20 °C above the melting point for 1–2 h and then cooled to room temperature at a rate of 50 °C/h.

Characterization

Powder X-ray diffraction (XRD) patterns were collected in transmission (Debye–Scherrer geometry) with a STADI P diffractometer (STOE Cie GmbH), equipped with a silicon strip MYTHEN 1K Detector (Fa. DECTRIS) with a curved Ge (111)-Monochromator (Cu Kα1 = 1.54056 Å). For the measurement, a grounded powder was placed between the adhesive tape. Single-crystal XRD measurements were conducted on Oxford Xcalibur S diffractometer equipped with a Sapphire 3 CCD detector and a molybdenum (Mo Kα = 0.71073 Å) sealed tube as an X-ray source. Crystals were tip mounted on a micromount with paraffin oil. The data were processed with Oxford Diffraction CrysAlis Pro software; structure solution and refinement were performed with SHELXS and SHELXL, respectively, imbedded in the Olex2 package.[41,42] The crystal structure of the synthesized compound was solved with direct methods, light elements (C, N) were located in the difference Fourier map, most of the positions of the cations were refined as rigid groups, and hydrogen atoms were placed at calculated positions.

Photoluminescence emission and excitation steady-state spectra were recorded with a FluoroMax4-Plus-P (Horiba Jobin Yvon) equipped with a 150 W Xe lamp. Absolute quantum yields of the powders were measured using a Quantaurus-QY (Hamamatsu) spectrometer with an integrating sphere in a quartz Petri dish. The relative uncertainty of the PLQY measurement is ±3%.

Photoluminescence decay was measured with a FluoroMax4-Plus-P equipped with a pulsed Xe lamp (3 μs). Decay curves were measured as a PL peak intensity over time. The temporal window size was 5 μs; each data point is accumulated over 100 lamp flashes.

Attenuation coefficient estimation was obtained from the transmission spectra of a single crystal through two parallel facets, recorded on a Jasco V500 spectrometer. The crystal was positioned into an opaque diaphragm with a hole of the crystal size. The diaphragms were used to calibrate baselines of the spectrometer according to the sample’s sizes. Reflection and scattering losses were taken into account by the measurement of specular and diffuse reflection spectra in the integrating sphere with an aluminum foil of the corresponding size as a reference. The spectral attenuation coefficient μλ of a crystal with dimension l was estimated from the optical density τ according to the equation

To analyze Mn–Br distances in the host lattice of ternary metal halides, density functional theory (DFT) was used as implemented in the Vienna Ab Initio Simulation Package (VASP) code. The projector augmented wave (PAW) potentials for atoms were used. For the generalized gradient approximation (GGA), the Perdew–Burke–Ernzerhof exchange-correlation functional (PBE) was used.[43−45] A 112-atom supercell (2 × 2 × 1) of each compound containing one Mn on Zn site was used for the calculations. Atomic positions were optimized while keeping the lattice parameters constant until the forces on atoms were smaller than 0.01 eV/Å. A gamma-centered automatic k-point mesh with 2 × 2 × 2 density was used.

Results and Discussion

Organic–Inorganic Hybrids

Organic–inorganic Mn(II) halides can be prepared from the corresponding aqueous hydrohalic acid solution of manganese halide and organic cation or from the solution of cation halide and manganese halide in polar organic solvents (methanol, ethanol, acetonitrile, N-methylformamide, N,N-demithylformamide, dimethyl sulfoxide). To display the trends in emission properties, we have selected three manganese(II) halide complexes with 1-benzyl-1-trimethylammonium (Bz(Me)3N)+ cation (Figure ). This cation was chosen for its bulkiness and anticipating that π–π stacking aids in crystallization. The cation bulkiness favors the crystallization of manganese in isolated MnX4 units. (Bz(Me)3N)2MnBr4 and (Bz(Me)3N)2MnI4 are isostructural and feature one asymmetric unit of MnX4, whereas (Bz(Me)3N)2MnCl4 comprises two inequivalent MnCl4 units. The PL excitation (PLE) spectra of all three compounds contain bands that are corresponding to the transitions described above (Scheme ). In the region between 300 and 500 nm there are two distinct groups of bands, corresponding to 6A1 → 4G (360, 375, and 390 nm for Cl, Br, I, respectively) and 6A1 → 4D (450, 460, and 475 nm) transitions. Emission spectra feature only one emission peak (Figure a–c). Generally, the emission and excitation wavelengths of Mn in such hybrids depend on the ligand field splitting: the energy splitting decreases from Cl to I, according to spectrochemical series.[46] The adequate assignment of PLE lines is still rather unrealistic.[47] To simplify the description of optical properties, we have adopted the difference between the strongest excitation peak around 450 nm (6A1 → 4T2) and the emission peak (4T1 → 6A1), E[4T2] – E[4T1] as a measure for the magnitude of the splitting. In this case, larger E[4T2] – E[4T1] values correspond to a stronger field. In addition, from lighter to heavier halide, spin–orbit coupling (SOC) gains importance. Figure d shows how SOC influences the decay time of the emission: τ is longest in chlorides (∼4 ms) and fastest in iodides (∼40 μs). In comparison, Mn2+ in a T oxide environment exhibits radiative lifetimes above 5 ms.[48] The PLQYs are highest when excited within the lowest energy PLE bands (around 450 nm) and decrease from chloride to iodide (Figure e). Spectral dependence of PLQY generally follows the PLE spectrum, as expected. However, optical absorption at PLE/PLQY minima is still significant (Figure ), pointing to the background absorption by nonemissive impurity species such as degradation products (molecular halides, trihalide anions, etc.). Organic–inorganic manganese tetrahalide hybrids can be prepared with a wide variety of cations (Table , Table S2). Optical properties of the resulting complexes cover the spectral range between 505 and 547 nm. All of these complexes are bright with PLQYs from 30% to 90% (at 450 nm excitation).

Figure 2

(a–c) PLE (empty) and PL (shaded, exc. 360 nm) of three hybrid compounds: (Bz(Me)3N)2MnX4 (X = Cl, Br, I). (d) Emission decay curves for (Bz(Me)3N)2MnX4 (X = Cl, Br, I). (e) PLQY dependence on the excitation wavelength for (Bz(Me)3N)2MnX4 (X = Cl, Br, I). (f) Photographs of (Bz(Me)3N)2MnX4 under visible light and UV excitation (365 nm).

Figure 8

Absorption spectra and spectral absorption coefficients of (a) Bz(Me)3N)2MnBr4 and (b) (Bz(Me)3N)2MnI4. (c) Photo of the (Bz(Me)3N)2MnI4 crystal under UV and daylight.

Table 1

Optical Properties (peak position, FWHM, PLQY) of the Organic–Inorganic Manganese(II) Tetrahalide Hybridsa

cation	halide	peak position, nm	fwhm, nm	fwhm, meV	PLQY, %b	E[4T2] – E[4T1], meV
Et4N	Cl	518	53.1	246	75	36
n-Pr4N		512	49.8	236	81	35
Bz(n-Bu)3N		522	57.5	262	60	39
Ph(Me)3N		522	55.1	251	89	36
Ph4P		517	55.8	260	N/A	39
Et(Ph)3P		520	55.8	257	66	38
Bz(Me)3N		547	71.8	299	78	51
(PPh3)2=N		544	62.4	262	N/A	51
Et4N	Br	516	49.1	229	86	34
n-Pr4N		511	47.8	227	N/A	33
Bz(n-Bu)3N		520	53.5	246	68	36
Ph(Me)3N		520	49.2	226	76	35
Ph4P		516	48.3	225	N/A	34
Me(Ph)3P		507	40.5	196	74	30
Et(Ph)3P		510	43.6	208	45	32
Bz(Me)3N		516	48.7	227	63	34
K[crypt-222]		504	39.4	193	N/A	31
Et4N	I	535	52	226	62	32
n-Pr4N		531	54	238	N/A	31
Ph(Me)3N		540	55.3	236	N/A	31
Bz(Me)3N		537	49.8	215	33	30

The strength of the crystal field is reflected in the difference between optical bands E[4T2] and E[6A1].

PLQY measured with 450 nm excitation.

Fully Inorganic Alkali Zinc Halides Doped with Manganese

We have prepared and analyzed Mn-doped zinc halides and several alkali metal zinc halides (Figure ). They were synthesized by melting the mixtures of the respective metal halides. With ZnBr2 as a host, high MnBr2 loadings of up to 15 wt % can be achieved without any evidence of multiple phases in powder X-ray diffraction (PXRD, Figure S1). In a CsBr–ZnBr2 system, all compounds melt congruently, which results in the formation of a pure Cs2ZnBr4 phase (Figure S2a,b). In the case of K2ZnBr4 and Rb2ZnBr4, impurity peaks at small 2θ might remain unassigned and are attributed to a rather technical purity of the precursors (Figure S2c,d). In Li2ZnBr4 and Na2ZnBr4 we have observed identical groups of peaks that belong to unreacted ZnBr2 and ABr (Figure S2e,f) due to incongruent melting of these ternary phases.[49] Both Cs2ZnI4 and Cs3ZnI5 form as pure phases (Figure S3a,b). In the case of K2ZnI4 and Rb2ZnI4, unknown impurity phases can be seen (Figure S3c,d). Among olivine structures, only Li2ZnI4 was obtained as a pure phase (Figure S3e) and the formation of Na2ZnI4 has not been observed; the reaction mixture contained unreacted NaI and ZnI2 (Figure S3f). All of the doped phosphors exhibit bright characteristic Mn emission at room temperature, and the deeper color of the doped samples reflects the increase in Mn concentration (Figure a–c).

Figure 3

Various crystal structures of A2ZnBr4, where A is an alkali metal cation.

Figure 4

(a) Photograph of various samples under ambient and UV light. (b and c) Photograph of Cs2ZnBr4 and Cs3ZnBr5 doped with various Mn amounts (w/w %) under visible light and UV excitation (360 nm). (d and e) PL peak position and fwhm of Mn-centered emission in various fully inorganic metal halide hosts (exc. 450 nm). (f) PLQY (exc. 450 nm) dependence on the Mn concentration in Cs2ZnBr4 and Cs3ZnBr5 demonstrating fast saturation of PLQY values at Mn contents of above 1–2%.

Mn-containing ZnBr2 and ZnI2 exhibit a bright green emission that peaked at 515 and 545 nm, respectively; with a very similar line broadening (196 and 199 meV). Ternary halides as hosts allow for tuning the luminescence over a wider region. A2ZnBr4 crystallize in three distinct crystal structures, stable at room temperature, all with T symmetry around Zn. Na2ZnBr4 and Li2ZnBr4 crystallize in olivine-Mg2SiO4 orthorhombic structure type, whereas Cs2ZnBr4 and Rb2ZnBr4 are isostructural to the Cs2CuCl4 orthorhombic structure type. K2ZnBr4 is found in a lower symmetry monoclinic polymorph. In addition, for Cs, another ternary phase exists with Cs3ZnBr5 stoichiometry, which consists of alternating CsZnBr and CsBr layers (Figure ). In ternary iodides, Li2ZnI4 and Na2ZnI4 crystallize in olivine structure, whereas both K and Rb homologues exist in monoclinic structure. Only Cs2ZnI4 has the same structure as its bromide counterpart.

Figure d and 4e summarizes the optical properties of the binary and ternary hosts doped with 2 wt % of Mn. Compounds with olivine structure (Li2ZnBr4, Na2ZnBr4, K2ZnI4) systematically show the emission peak shifted toward higher energies: 506–514 nm for bromides and 530 nm for iodide. For comparison, K, Rb, and Cs counterparts emit above 520 nm (bromides) and above 540 nm (iodides). Higher peak energies correspond to a stronger ligand field, which is corroborated by sharper PLE peaks. The narrowest fwhm is observed for olivine-type hosts (177 meV for Na2ZnBr4, Figure b). For monoclinic structures (K2ZnBr4, K2ZnI4, and Rb2ZnI4), the fwhm is is higher (195–210 meV). Cs2CuCl4-type Rb2ZnBr4, Cs2ZnBr4, and Cs2ZnI4 exhibit the emission with the highest fwhm ranging between 210 and 240 meV. Among all cesium-based compounds, the phase with excess cesium halide (Cs3ZnX5) shows narrower emission at a lower wavelength. To test the PLQY dependence on Mn concentration, we have chosen the compounds in the CsBr–ZnBr2 system. From 0.3 to 2 wt % of Mn, the PLQY increased almost seven times. We have also estimated the equilibrium Mn–halide distances in the host lattice (for the case of A2ZnBr4) using DFT calculations. The calculated distances are summarized in Table . Generally, metal–ligand distances are larger in olivine-type structures and shorter in orthorhombic structures.

Table 2

Equilibrium Mn–Br distances (Angstroms) Calculated by DFT in Five A2ZnBr4 Hosts

compound	d1	d2	d3	d4	daverage
Li2ZnBr4	2.378	2.386	2.380	2.378	2.380
Na2ZnBr4	2.383	2.381	2.383	2.384	2.383
K2ZnBr4	2.363	2.363	2.380	2.382	2.372
Rb2ZnBr4	2.373	2.377	2.355	2.355	2.365
Cs2ZnBr4	2.354	2.386	2.354	2.374	2.367

Limitations and Prospects of Mn Green Emission: Decay Rates, PL Peak Position, fwhm, and Absorption Coefficient

Assessment of the practical utility of a novel phosphor for applications such as LCD displays and lighting requires knowledge of the following optical parameters: emission decay rate, PL peak position, fwhm, absorption coefficient, and PLQY. As can be noticed from Figure a–c, the widths of the excitation and emission bands are notably different: characteristic Mn2+ PLE bands (fwhm = 10–30 nm) are narrower than the PL band (fwhm = 40–60 nm). This is often attributed to the difference in the chemical bonding in the ground and excited states. The observable variability in these parameters can be rationalized by considering energy level splittings of the Mn2+ in a tetrahedral halide crystal field, in particular, by focusing on the strength of the crystal field (Δ) and covalency (B) of the Mn–ligand bond as two primary factors. To illustrate the splitting of the energy states as a function of Δ, a Tanabe–Sugano (TS) correlation diagram is used (Figure ). The field strength increases from left to right. On the basis of this general theory of d-ions emission, the strength of the crystal field defines the position of the Mn2+ emission peak: in the weaker field, the 4T1 and 6A1 levels are separated stronger and hence emission occurs at higher energies (506–520 nm). As the crystal field increases, the emission peak shifts to lower energies (525–545 nm). The d-energy levels are also influenced by the electron–phonon coupling. A general intuition from the TS diagram would be higher slopes shall correspond to greater sensitivity to the electron–phonon coupling and hence higher homogeneous broadening of the transition. This argument fully explains the trends in organic–inorganic chlorides and bromides (Figure a). However, in the case of iodide, being the weakest ligand in the halide series, crystal field strength arguments fail to predict correctly the optical properties. Although the experimental splitting between the states still follows the TS diagram (the smallest values of E[4T2] – E[4T1] for iodides: 30–32 meV), the PL peak position shifts to the red and the fwhm is broader (Figure a). This can be explained by the Mn–I bond having a more covalent nature and iodine having a higher nephelauxetic effect. “Covalent” in this context refers to the degree of mixing between the d orbitals of Mn and orbitals of the ligands. As a result, all energy states in tetraiodomanganates, including those that do not depend on the crystal field splitting, e.g., 4A1, 4E(G), 4E(D), shift toward lower energies (B′ < B). This effect is responsible for tetrahedral Mn2+ orange emission (ca. 575–580 nm) in ZnS host.[50] In general, the ordering of the ligands in the spectrochemical series and nephelauxetic series is opposite: free ion < I– < Br– < Cl– < S2– < F– < O2– vs free ion < F– < O2– < Cl– < Br– < I– < S2–.[51] For the case of weak tetrahedral fields, Coloumb repulsion of d electrons has a higher contribution and the nephelauxetic effect is, therefore, more pronounced. In other words, a more ionic Mn–ligand bonding renders the states involved in the optical transitions more localized on Mn ions and thus less sensitive to the electron–phonon coupling. This can explain an observation of the narrowest spectral width for Mn2+T emission (fwhm = 17.5 nm) in the strongest crystal field of oxides.[52,53] Theoretically, a tetrahedral fluoride environment could result in an even narrower emission line width. However, finding a proper matrix poses a problem: due to the smallest ionic radii of F– among halides (1.33 vs 1.81 Å for Cl), structures with tetrahedral coordination are scarce. For example, ZnF2 has a rutile structure that features O coordination of Zn. The covalency argument also explains why Mn2+ in T oxide or Mn4+ in O fluoride field display longer decay times (few milliseconds):[54] higher mixing of states in a more covalent case helps to relax the forbidden nature of the transition.

Figure 5

Phenomenological model of the Tanabe–Sugano diagram for a d5 (Mn2+) ion. Colored areas show relative locations of tetrahedral and octahedral Mn2+ fields.

Figure 6

(a) Dependence of the fwhm on the crystal field strength (represented by the difference between optical bands E[4T2] and E[6A1]) in organic–inorganic hybrid Mn(II) chlorides, bromides, and iodides. (b) Dependence of the fwhm (shaded circles) and PL (open circles) on the equilibrium Mn–Br bond distance in fully inorganic A2ZnBr4 compounds.

The nephelauxetic effect is also present in the compounds with the same halide composition (e.g., A2ZnBr4). It has been previously demonstrated from ab initio calculations for a d3-metal impurity in A2NaBX6 hosts (A = K, Cs, B = Sc, Y) that the B parameter and Δ depend on the metal–ligand distance, which in turn reflects the covalency of the bond.[55] For A2ZnBr4:Mn compounds, the fwhm and PL peak positions appear to correlate with the average equilibrium Mn–Br distance (from Table , Figure b).

The PL peak position and energy states splitting relate also to the distance between manganese atoms.[21,23] This can be exemplified by comparing the Mn–Mn distance distribution in two compounds: Bmpip2MnCl4 and (Bz(Me)3N)2MnCl4 (Figure a). In (Bz(Me)3N)2MnCl4, the smallest Mn–Mn distance is 8.5 Å, whereas for Bmpip2 MnCl4 it is 9.1 Å (Tables S3 and S4, CCDC 1936177 and CCDC 1936170). The emission of Bmpip2MnCl4 is centered around 513 nm, and the E[4T2] – E[4T1] splitting is 38 meV. In comparison, a much larger splitting of 51.3 meV in the case of (Bz(Me)3N)2MnCl4 could originate from the smaller Mn–Mn interatomic distance (Figure b).

Figure 7

(a) Mn–Mn distance distribution in two tetrachloride compounds, Bz(Me)3N 2MnCl4 and Bmpip2MnCl4, calculated from crystal structure files measured with single-crystal X-ray diffraction (CCDC 1936177 and CCDC 1936170). (b) PL spectra (exc. 360 nm) of Bz(Me)3N 2MnCl4 and Bmpip2MnCl4. (c) Emission decay for tetraalkylphoshonium, tetraalkylammonium tetrabromomanganates(II) (green), and their bromoalkyl derivatives (blue). (d) Chemical formulas of the cations used for comparison.

In addition, the larger fwhm of the 547 nm peak can be attributed to two symmetrically inequivalent MnCl4 units present in the structure.

One of the concerns related to the forbidden nature of the transition is the slower emission decay as compared to, for instance, semiconductor quantum dot phosphors. We have found that tetrabromomangates(II), a bromoalkyl derivative of the cation, expedites decay rates (four times for the case shown in Figure c and 7d). As it has been already discussed, the radiative lifetime decreases from Cl to I. To control the decay time beyond, we have compared (MePh3P)2MnBr4 and (n-PrMe3N)2MnBr4 with corresponding analogous compounds where cations are brominated: (BrMePh3P)2MnBr4 and (n-PrBrMe3N)2MnBr4. In both cases, we observed a decrease in the emission decay time for the brominated cation from hundreds of microseconds (380 and 260 μs for phosphonium and ammonium, respectively) to less than 100 μs (79 and 72 μs). This can be attributed to the proximity of the cation bromine atom to the Mn. In nonhalogenated cations, the shortest Mn–Br distance is about 2.5 Å (first coordination sphere of Mn), whereas in the halogenated cation, due to the organic cation bromine atom, there is a second Mn–Br distance at 4–5 Å (Table S5, (n-PrBrMe3N)2MnBr4, CCDC 1936114). In the previous study on the exciton decay in Cat2MnX4, no difference in the decay rate was found for cations comprising heavy atoms, e.g., Ph4As+.[38] This could be due to the greater distance between Mn and As: 6.2 Å. This discovery motivates future studies on the structural engineering of faster emission decay by introducing heavy atoms (e.g., iodine, bromine) in the second coordination sphere of Mn.

Another limitation that comes from the forbidden nature of the optical transitions is a relatively low absorption coefficient. To estimate the absorption, we have chosen two organic–inorganic hybrid compounds, for which crystals of sufficient quality and size (4–7 mm) could be isolated: (Bz(Me)3N)2MnBr4 and (Bz(Me)3N)2MnI4. Light absorption was deduced from the transmission spectra by subtracting the reflectivity. By dividing the spectral optical depth τλ (at 450 nm for bromide and 475 nm for iodide) by the dimension of the crystal, we obtained the directional attenuation coefficient μλ (Figure a and 8b): 0.5 cm–1 for bromide and 55 cm–1 (128 cm–1·M–1) for iodide. A higher absorption coefficient correlates well with the faster emission in iodides. However, these values are still very low when compared to the organic dyes (e.g., fluorescein, λ500 = 92 230 cm–1·M–1)[56] or direct band-gap semiconductors (e.g., InP, λ400 = 30 861 cm–1).[57]

Conclusions

We prepared and summarized the properties of various halides with tetrahedral coordination of Mn(II): organic–inorganic manganese(II) tetrahalide hybrids and Mn(II)-doped fully inorganic ternary halides. Thus far, research efforts around Mn-based green emitters for display and lighting applications have focused on oxide matrixes. A concerted effort is urgently needed for unveiling the potential Mn(II) in solid-state halides. The tetrahedral halide ligand environment, in particular, iodide and bromide, brings about its advantages over oxides: faster emission decay of 0.05–0.5 ms (vs 4–7 ms in oxides or oxynitrides), larger PLQY of up to 90%, and facile low-temperature synthesis (vs 1000–1500 °C for oxides). Emission and excitation spectra can be fine tuned by the crystal field splitting parameters. The latter are adjustable by both the halide composition as well as the distance between manganese atoms. The emission decay is accelerated by four times in organic–inorganic hybrids with the additional halide atom in the second coordination sphere of Mn. We propose that the two effects, namely, crystal field strength and nephelauxetic effect, compete in the Cl, Br, and I series and have opposite trends: the PL broadening and red shift in the case of Cl originates from strong crystal field splitting, whereas the same effect in I is attributed to the higher covalency of the Mn–I bond. The entirety of the experimental results on the structure–property relationship allows drawing the following conclusions as to how engineerable the PL characteristics can be. In particular, weaker fields and less covalent Mn–ligand bonds cause emission that is narrower yet slower (5–10 ms) and less efficient. The major practical hurdle for this class of phopshors concerns low absorption in the blue region due to the forbidden nature of the transition (Laporte rule). Overcoming this limitation requires new materials design strategies, which may include sensitization or enhancement of the SOC effect.