Sandwich‐Like Encapsulation of a Highly Luminescent Copper(I) Complex

A small molecular weight cationic copper(I) complex showing high luminescence quantum yield based on a thermally activated delayed fluorescence mechanism is immobilized between two 1 nm thin silicate layers. Partial ion exchange of the emitter into a synthetic layered silicate (fluorohectorite) yields an ordered heterostructure with two types of strictly alternating interlayers: a monolayer of the cationic emitter and a monolayer of hydrated Na+ cations. Osmotic swelling of the latter produces dispersions of double‐stacks in which the emitter monolayer is encapsulated between two silicate layers. The electrostatic attraction of the emitter interlayer with the oppositely charged silicate layers exerts electrostatic pressure on the emitter. Compared to crystalline salt, rigid confinement for the encapsulated emitter provides improved thermal stability and increased emission quantum yield at ambient temperature. The suspension of delaminated, micrometer‐sized double‐stacks of 3.9 nm thickness allows for easy solution processing of low‐cost optoelectronic devices, such as light‐emitting electrochemical cells and organic light‐emitting diodes.


DOI: 10.1002/adom.202100516
subclass of optoelectronic devices gene rating light from electricity employing electron-hole recombination in an organic semiconductor.
The vacuum processing technique of multilayer optoelectronic devices is com pared to solution processing unfavorable due to its high processing costs. Neverthe less, the solution deposition method still suffers from the requirement of orthog onal solvents [3] due to the possible dissolu tion of already deposited layers. [4] Such dis solution processes can result in undesired blending of the different layers causing unwanted decrease of performance. In addition, performance of optoelectronic devices is limited by the diffusion of the emitter molecules.
Thus, improvement of the stability and increase of brightness of optoelec tronic devices is highly desirable. One way of addressing this problem is to uti lize the crosslinking technique. [5] For solutionprocessed opto electronic devices, this technique enables to prevent the unde sired diffusion of the small emitter molecules and results in the enhancement of the structural stability, [4a,6] improving the device lifetime, [7] and enhancing the overall performance. [4b] On the other hand, the achievement of even higher performance can be hampered by the presence of byproducts produced by crosslinking technique. [8] A recently established method com prising the encapsulation of organic cations and transition metal complexes between negatively charged silicate nanolayers seems to be ideal for these purposes, as the encapsulation is not accompanied by the formation of any unwanted byproducts and at the same time the diffusion of the small emitter mole cules is greatly suppressed. [9] Moreover, this method allows the deposition of active emitter compounds as thin uniform layers with defined thickness. The synthesis approach is based on facile intercalation of the cationic emitter into layered sili cate stacks [9b] forming ordered heterostructures consisting of strictly alternating interlayers occupied with emitter molecules and with inorganic cations (Figure 1b). Similar heterostructures based on synthetic rectorite with alternating nonhydrated potas sium and hydrated sodium interlayers were prepared by Möller et al. [10] In water, these materials undergo osmotic swelling and delaminate into doublestacks of two silicate layers encap sulating the central layer of the intercalated emitter molecules (Figure 1c). [9c] The doublestacks show very high aspect ratios of more than 5000, with the thickness of the silicate sandwiched A small molecular weight cationic copper(I) complex showing high luminescence quantum yield based on a thermally activated delayed fluorescence mechanism is immobilized between two 1 nm thin silicate layers. Partial ion exchange of the emitter into a synthetic layered silicate (fluorohectorite) yields an ordered heterostructure with two types of strictly alternating interlayers: a monolayer of the cationic emitter and a monolayer of hydrated Na + cations. Osmotic swelling of the latter produces dispersions of doublestacks in which the emitter monolayer is encapsulated between two silicate layers. The electrostatic attraction of the emitter interlayer with the oppositely charged silicate layers exerts electrostatic pressure on the emitter. Compared to crystalline salt, rigid confinement for the encapsulated emitter provides improved thermal stability and increased emission quantum yield at ambient temperature. The suspension of delaminated, micrometer-sized double-stacks of 3.9 nm thickness allows for easy solution processing of low-cost optoelectronic devices, such as light-emitting electrochemical cells and organic light-emitting diodes.

Introduction
Nowadays, the global energy market demand has triggered intense research in the development of lighting devices such as the lightemitting electrochemical cell [1] (LEEC) and organic lightemitting diode [2] (OLED). LEECs and OLEDs belong to a emitter being 3.9 nm and a lateral extension ≈ 20 µm. Virtually any kind of cationic emitter can be deposited into a device from a solution as a monomolecular layer. In this work, the synthesis of such doublestacks applying sodium fluorohectorite [11] (Na + hec) and as a representative for such emitters, a luminescent copper(I) complex [Cu(dppb) 2 ] + (dppb = 1,2bis(diphenylphosph ino)benzene, Figure 1) [12] is demonstrated.
[Cu(dppb) 2 ] + was chosen as a model compound because it is a benchmark cationic copper(I) complex showing thermally acti vated delayed fluorescence (TADF) at ambient temperature as will be shown below. The advantage of using TADF [13] materials for lighting applications is based on several aspects. First, they allow for 100% use of singlet and triplet excitons formed in the course of an electroluminescent excitation resulting, according to the singlet harvesting mechanism based on the molecular TADF effect, in very high quantum efficiencies. [14] Second, con centration quenching or selfquenching at high emitter con centrations, for instance, in crystalline samples, is highly sup pressed. [15] Third, emitters in optoelectronic devices, such as copper compounds, represent an inexpensive alternative for 3rd row transition metal complexes. Finally, [Cu(dppb) 2 ] + is almost nonemissive in solvents (see Table 1), and therefore, incorpora tion into the interlayer space of layered silicate allowed to study the enhancement of photoluminescence by the solid matrix.

Ordered Interstratification
The [Cu(dppb) 2 ] + /fluorohectorite sandwich was synthe sized according to the procedure illustrated in Figure 1. [9b,c] As inorganic host synthetic Na + hec with a composition of [Na 0.5 ] inter [Mg 2.5 Li 0.5 ] oct [Si 4 ] tet O 10 F 2 was applied, which was pre pared by melt synthesis according to a published procedure. [11] It is transparent and comes in micrometersized crystals ( Figure S1a, Supporting Information). The cation exchange capacity (CEC) of Na + hec hydrated at 43% relative humidity was determined, applying the [Cu(trien)] 2+ method, to be 1.18 mmol g -1 . [9b,16] For fast kinetics of the intercalation of [Cu(dppb) 2 ] + the one layer hydrate of Na + hec was first swollen in acetonitrile: water    Table S1, Supporting Information); c) Energy difference between the lowest excited singlet and triplet states. Roughly estimated from the peak maxima at T = 300 and 77 K, respectively; d) Measured in inert N 2 gas atmosphere. λ exc = 350 nm; e) The decay is not monoexponential, reflecting spectroscopic inhomogeneities of the measured material. The value given represents the main (short-lived) component; f) The decay transients are shown in Figure S4 in the Supporting Information. mixture (75:25 vol%) ( Figure 1a). Swelling in this solvent mix ture is limited to the crystalline regime and the basal spacing [17] expanded in this mixture from d 001 = 12.3 Å for the onelayer hydrate [11] to 30.7 Å for the crystalline swollen Na + hec ( Figure S1b, Supporting Information), as determined by means of small angle Xray scattering (SAXS). Enlarging the gap opening between adjacent silicate layers to ≈21 Å (subtracting the layer thickness of hectorite, 10 from 30.7 Å) facilitates the intercala tion of the bulky metalcomplex ions [Cu(dppb) 2 ] + of ≈10 Å. Partial cation exchange leads to segregated dye layers that are arranged in an ordered fashion [9b, 10,18] consisting of strictly alternating hydrated Na + and [Cu(dppb) 2 ] + interlayers as evi denced by the appearance of a superstructure reflection (d OI ) at a dspacing of ≈ 35 Å ( Figure S2, Supporting Information). As previously shown, these ordered heterostructures actually represent the thermodynamic equilibrium. [9b] The formation of such ordered structures is driven by pronounced differences in interlayer cation densities and different interlayer heights of the hydrated Na + and [Cu(dppb) 2 ] + interlayers. To obtain a strictly alternating heterostructure (ordered interstratification), the thrive on forming heterostructures is, however, not yet suffi cient. In addition, the probability of the two types of interlayers must be equal. Since the cation densities of the two interlayer types are different, and since the partition function of the dye between intercalated and dissolved state are dependent on the degree of exchange, the concentration of [Cu(dppb) 2 ] + yielding equal probabilities needed to be identified iteratively and was found to be 22% of the total CEC. At this level of partial ion exchange, the intensity of the 001 d OI at 35.0 Å was found to be maximal and the series of 00 d l OI reflections was found to be nicely rational with a coefficient of variation of 0.01.
The ideal degree of partial ion exchange was further cross checked by elemental analysis (Table S1, Supporting Informa tion) and by monitoring the partition equilibria [9b] in solution (Figure 2). At levels of [Cu(dppb) 2 ] + where the dissolved dye is in equilibrium with the heterostructure, around 54% of the dye remained in solution. To achieve higher levels of dye being ion exchanged, the concentrations need to be signifi cantly higher. At this stage, a new equilibrium is established with 60% of the dye remaining in solution that is in equilib rium with fully exchanged interlayers. The inflection point between the two different equilibria was observed at 22% exchange. Similarly shaped partition functions have been observed with partial ion exchange of Na + hec using other organic cations. [9b] Thus, by exchanging 22% of hydrated Na + by [Cu(dppb) 2 ] + , the two types of interlayers have equal probability and are arranged strictly alternating (Figure 1b)
The SAXS pattern (Figure 3) for an aqueous 4 wt% suspen sion of [Cu(dppb) 2 ] + encapsulated between two silicate layers revealed strong form factor oscillations and scatter intensity scaling to q -2 typical for plateletlike objects.   for more details) with a total thickness of 3.9 nm, which is in decent agreement with the Xray data of the heterostructures before delamination. Atomic force microscopy (AFM) images of a dried suspension also confirmed the presence of encap sulated [Cu(dppb) 2 ] + with lateral extensions of several micro meters ( Figure S3, Supporting Information) and a thickness of about 4.3 nm, which is again in fair agreement with the SAXS and Xray results.

Photophysical and Thermal Properties
The obtained translucent colloidal dispersion showed intense yellowgreen luminescence of [Cu(dppb) 2 ] + encapsulated between two silicate layers (Figure 4a,b). The emission is uni formly intense in the whole volume, demonstrating a homog enous distribution of the encapsulated [Cu(dppb) 2 ] + . The proof of the emission originating from the [Cu(dppb) 2 ] + encapsulated between layered silicate platelets was shown by AFM combined with a confocal microscope measurement (Figure 4c,d). Table 1 compares emission properties of [Cu(dppb) 2 ]PF 6 dis solved in tetrahydrofuran (THF), of micrometersized powder form, and of a dried powder of [Cu(dppb) 2 ] + encapsulated between two silicate layers. The photoluminescence spectra of the latter are shown in Figure 5. In THF solution, [Cu(dppb) 2 ] PF 6 shows only very weak luminescence at 550 nm with the emission quantum yield of φ PL < 1%.
The relatively rigid crystalline environment of [Cu(dppb) 2 ] + in the [Cu(dppb) 2 ]PF 6 powder reduces vibrational quenching strongly as compared to the solution because molecular dis tortions in the excited molecules are significantly suppressed leading to an increase of φ PL to 57%. At the same time, the emission is blueshifted from 550 nm found for solution to 490 nm for the powder.
Moreover, [Cu(dppb) 2 ]PF 6 powder displays distinctly dif ferent emission properties at 77 K compared to ambient temperature. In particular, the emission decay times with τ(77 K) = 1.9 ms and τ(300 K) = 3 µs differ by orders of mag nitude. However, it is more appropriate to compare the radia tive rates k r = φ PL /τ or the radiative decay times. Using the data shown in Table 1, it was found for T = 77 K k r (T 1 →S 1 phospho rescence) = 5.3 × 10 2 s -1 (τ r = 1.9 ms), representing a relatively slow decay process compared to other Cu(I) complexes. [14] For T = 300 K, the radiative emission rate is determined to 1.9 × 10 5 s -1 (τ r = 5.3 µs). Accordingly, the radiative process is by a factor of almost 360 times faster than at T = 77 K. This drastic increase is ascribed to the thermal activation process from the T 1 state to the S 1 state combined with a spinallowed S 1 →S 0 fluorescence decay path, the TADF emission. Further more, it is interesting to obtain an information about the energy gap ΔE(S 1 -T 1 ). Frequently, it can be determined from a fitting procedure of the temperature dependence of τ(T). [13,14,19] This, however, requires fast thermal equilibration between all states involved. Although this is mostly given for Cu(I) com plexes because of very fast intersystem crossing (ISC), being of the order of 10 ps. [20] As a consequence, also upISC (reverse ISC) is fast and thus, fast equilibration between the lowest excited states results. This leads to a monoexponential decay behavior of the emission. In several cases, however, as found for the compounds investigated, distinct inhomogeneities or even different sites are responsible for the emission. Therefore, decay curves with different decay components occur that are   related to the different sites. In this situation, the fitting proce dure is not successful for determining a defined energy gap. [21] On the other hand, with temperature increase, a blueshift of the emission peak maxima from 510 nm (77 K) to 490 nm (300 K) is observed (Table 1). This shift largely corresponds to the activation energy of this TADF process and therefore, may be taken for a rough estimate of the energy gap. Thus, we find ΔE(S 1 -T 1 ) ≈ 800 cm -1 (≈100 meV). To summarize, the data pre sented allow us to safely conclude that the emission behavior for T > 77 K becomes dominated by the TADF process up to ambient temperature.
The emission spectra and decay behavior of a dried powder of [Cu(dppb) 2 ] + encapsulated between two silicate layers are com parable to those described for the powder sample of [Cu(dppb) 2 ] PF 6 , indicating similar electronic structures of the emissive spe cies and similar emission mechanisms. Differences, as marked in Table 1, may be ascribed to the dissimilar environment of the emitting centers. The rigid confinement of the encapsulated [Cu(dppb) 2 ] + also causes a blueshift to λ em = 520 nm as compared to the solution of [Cu(dppb) 2 ]PF 6 (λ em = 550 nm). For the encap sulated [Cu(dppb) 2 ] + , the quantum yield of φ PL (300 K) = 65% was found to be even higher than for [Cu(dppb) 2 ]PF 6 powder, indicating that the confinement of [Cu(dppb) 2 ] + encapsulated between oppositely charged silicate layers suppresses nonradia tive relaxations to the electronic ground state more efficiently than an ordinary ionic crystalline environment. This correlates well with the longer TADF decay time of τ ≈ 7 µs at ambient temperature, a typical value for Cu(I) complexes. [14] Importantly, the yellowgreen emission was also observed in aqueous suspen sion of encapsulated [Cu(dppb) 2 ] + (Figure 4).
For completeness, it is remarked that strong emission at a high concentration of the emitter molecules (within the [Cu(dppb) 2 ] + interlayers) is not selfevident. For most Cu(I) complexes flattening distortions [22] of the molecular structure in the S 1 and T 1 excited states of chargetransfer character results in a loss of resonance condition required for efficient energy transfer between adjacent molecules. [15c] Therefore, traps of excited Cu(I) complexes are built and consequently concen tration quenching or energy transfer to solidphase defects (quenching traps) are not effective. Obviously, even the electro static pressure exerted by the oppositely charged rigid silicate layers does not prevent the important selftrapping effect. [23] This is in contrast to many conventional emitters, e.g., phos phorescent iridium complexes, where concentrationquenching at higher concentrations is very efficient. [24] Interestingly, the encapsulation of [Cu(dppb) 2 ] + between two silicate layers additionally enhances the thermal stability of the complex cation. Thermogravimetric analysis (TGA) revealed thermal stability of the [Cu(dppb) 2 ]PF 6 crystals up to 335 °C. Above this temperature, a fast mass loss was observed ( Figure S5, Supporting Information). The TGA curve recorded for the encapsulated [Cu(dppb) 2 ] + powder, using the same conditions, displays a massloss onset not below 415 °C, thus, pushing the thermal stability by 80 °C. Please note that the initial mass loss commencing at 120 °C is related to the first dehydration step of Na + interlayers followed by second dehydration step at 190 °C. Improved stability of the emitter materials applied in optoelec tronic devices should enhance their processing and operational stability, resulting in increased device lifetime.

Conclusion
In conclusion, highly luminescent layered TADF nanomaterials that may be processed in suspension were synthesized by com bining lowcost synthetic silicates and copper(I) complexes. The encapsulation of the TADF emitter improved thermal stability and increased luminescence quantum yields at ambient tem perature. The synthesis route involves, in particular, delami nation of regularly ordered interstratifications (alternating heterostructures) yielding dispersible doublestacks of densely packed emitter interlayers encapsulated between large and thin transparent silicate layers. The encapsulation immobi lizes even low molecular weight emitters, and therefore, this approach enables the application of a myriad of luminescent compounds suitable for optoelectronic devices preventing their diffusion across the device, potentially resulting in enhanced stability, improved device lifetime, and better performance. The presented approach can be extended to diverse osmoti cally swellable layered inorganic materials, such as zeolites, titanates, niobates, or perovskites, [25] paving the way to novel nanomaterials for optoelectronic applications with improved quantum efficiencies.

Experimental Section
Synthesis of Na + -hec and [Cu(dppb) 2 ]PF 6 : Na + -hec with the composition [Na 0.5 ] inter [Mg 2.5 Li 0.5 ] oct [Si 4 ] tet O 10 F 2 was synthesized by melt synthesis in a closed molybdenum crucible according to a published procedure. [26] After the synthesis, the material was annealed for 6 weeks at 1045 °C to improve intracrystalline reactivity, charge homogeneity, and phase purity as described. [11] [Cu(dppb) 2 ]PF 6 was prepared from [Cu(CH 3 CN) 4 ]PF 6 and 1,2-bis(diphenylphosphino)-benzene (dppb) according to published procedure. [12b,27] Synthesis of [Cu(dppb) 2 ] + Intercalation Compounds: Interstratified samples were prepared in centrifuge tubes sealed with a septum in an Argon atmosphere. Mixing at 40 °C was performed using a temperaturecontrolled oven equipped with a self-made overhead shaker. During the preparation of intercalated compounds, the amount of solvent was adjusted concomitantly with the variation of the complex amount in order to keep the complex concentration for all samples constant at 1 mmol dm -3 . For instance, Na + -hec hydrated at 43% relative humidity (10.73 mg) was predispersed in 75 vol% acetonitrile (2.12 cm 3 ) for 30 min at room temperature using an overhead shaker. After adding of complex solution (4.21 cm 3 in 75 vol% acetonitrile, c = 1.5 mmol dm -3 ) the dispersion was mixed at 40 °C for 1 h. The solid was separated by sedimentation followed by centrifugation at 1270 × g for 5 min. The amount of complex remaining in supernatant was determined photometrically using Agilent CARY 300 UV-vis spectrophotometer. The exchanged amount of [Cu(dppb) 2 ] + in intercalated samples was quantified by determining the carbon content using CHN elemental analysis (Elementar vario EL III). Prior to analysis, the samples were thoroughly washed with 75 vol% acetonitrile mixture and two times with pure acetonitrile in order to remove the remaining salts. For weighting, all samples were heated at 110 °C for 2 d and transferred into a glovebox with an argon atmosphere. During the CHN analysis, the combustion temperature was set to 1150 °C using a combustion tube filled with tungsten(VI)-oxide-granules.
After Powder X-Ray Diffraction: Diffractograms were recorded in Bragg-Brentano geometry using Panalytical XPERT-PRO diffractometer. Textured films were prepared by drying a few drops of the suspension on microscope slides (Menzel Glass). The glass slides were dried at 80 °C and equilibrated at room temperature for 12 h at 43% relative humidity (using saturated K 2 CO 3 solution).
Small-Angle X-Ray Scattering: A translucent 4 wt% gel in water was prepared by concentrating the aqueous suspension of [Cu(dppb) 2 ] + encapsulated between two silicate layers using a centrifuge (Hereus Multifuge). Centrifugation was performed two times at 14 090 × g for 5 min. During the measurement, the gel was placed in a 1 mm glass capillary (Hilgenberg, code 4007610, Germany). SAXS data were obtained by using the Double Ganesha AIR system (SAXSLAB, Denmark). As an X-ray source, a rotating anode (copper, MicoMax 007HF, Rigaku Corporation, Japan) was used, providing a microfocused beam. The data were recorded by a position-sensitive detector (PILATUS 300K, Dectris). In order to cover the range of scattering vectors between 0.004 and 1.3 Å -1 , different detector positions had to be used. Circularly averaged data were normalized to the incident beam, sample thickness, and measurement time prior to the subtraction of the used solvent. The data analysis was performed using the Scatter software (version 2.5), allowing 2D modeling of scattering intensities with respect to the layer distance, mean deviations from the ideal positions, and mean sizes of crystal domains. [28] Further calculations were done using the SASfit software. [29] SAXS Modeling: The SAXS intensities were fitted applying a model of stacked hamburgers. [9c] For further calculation of the SAXS pattern thickness of 0.8 nm was used for the silicate layer, which was experimentally derived from the intensity minima of the form factor oscillation of completely delaminated Na + -hec SAXS patterns. The obtained value is in good agreement with the platelet thickness derived by SAXS measurements of delaminated natural montmorillonites. [30] The thickness of the [Cu(dppb) 2 ] + layer was found to be 2.3 nm.
Atomic Force Microscopy: The height profile of double-stacks was obtained by using MFP3DTM Atomic Force Microscope (Asylum Research, Santa Barbara, California) equipped with silicon cantilevers (silicon tip, type NSC15/AlBS, Rmash, Tallinn, Estonia). The samples were prepared by dropping a few drops of diluted suspension (0.01 mg dm -3 ) onto a silicon wafer followed by slow evaporation of water under ambient conditions.
A combination of AFM with luminescence measurements was possible by using a VistaScope AFM (Molecular Vista) equipped with silicon cantilevers (silicon tip ATEC-NC, Nanosensors). The samples for these measurements were prepared by dropping a few drops of diluted suspension (0.01 mg dm -3 ) onto a glass coverslip followed by slow evaporation of water under ambient conditions. The optical part of the AFM instrument was equipped with a confocal microscope using a UPLSAPO oil immersion lens (Olympus, 100× magnification, numerical aperture = 1.4). The sample was excited using a frequency-doubled MIRA Laser source with λ = 404 nm (TiSa-Oscillator, 76 MHz, Coherent). The luminescence was detected by a single-photon avalanche diode (Micro Photon Devices).
Photophysical Characterizations: Photoluminescence spectra were measured with a Horiba Jobin Yvon Fluorolog 3 steady-state fluorescence spectrometer additionally modified, allowing to measure emission decay times. As an excitation source, a PicoQuant LDH-P-C-375 pulsed diode laser (λ exc = 375 nm, full width at half maximum 100 ps) was used. The emission signal was detected with a cooled photomultiplier attached to a FAST ComTec multichannel scalar PCI card with a time resolution of 250 ps. Photoluminescence quantum yields were determined with a Hamamatsu C9920-02 system equipped with a Spectralon integrating sphere.
Thermogravimetric Analysis: To speed up the kinetics of the thermal decomposition, the sample surface was increased by lyophilization applying a Christ Alpha 1-4 freeze-dryer (Martin Christ). The thermal stability of samples was afterward examined using a thermogravimetric analyzer STA 449 C (Netzsch) in the temperature range of 25-750 °C at a heating rate of 10 °C min -1 under a synthetic air atmosphere with a composition of 20.5 vol% O 2 and 79.5 vol% N 2 .

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.