Dual interfacial H-bonding-enhanced deep-blue hybrid copper–iodide LEDs

Materials

All materials were used as received without further purification. Copper(I) iodide (≥99.999% (Cu basis), STREM CHEMICALS INC MS); copper(I) iodide (98%, Alfa Aesar); 1,4-diazabicyclo[2.2.2]octane (>98.0%(GC), TCI); hydroiodic acid (57% (w/w), BTC); ethanol (190 proof, VWR; 200 proof, anhydrous, KOPTEC); acetone (99.5%, VWR); N,N-dimethylformamide (≥99.8%, anhydrous, Alfa Aesar); ethyl acetate (anhydrous, 99.8%, Sigma-Aldrich); carbazole (≥95% (GC), Sigma-Aldrich); N-Bromosuccinimide (NBS, Oakwwood Chemical); acetyl chloride (>98.0%, TCI); sodium methoxide (≥98%, Thermo Scientific); sodium hydride (60%, dispersion in Paraffin Liquid, TCI); diethyl-2-bromoethyl-phosphonate (97%, Aldrich); bromotrimethylsilane (TMSBr, ≥97%, stabilized, BTC); nickel oxide (99.99%, 3.0"×0.125", indium bonding on Cu backing plate, MSE Supplies LLC); PMMA (M_w = 35000 Da, ACROS Organics); Ca(acac)₂ (calcium acetylacetonate, anhydrous, Sigma-Aldrich Inc); PEDOT:PSS (Clevios P VP AI 4083); TPBi (99.5+%, Sigma-Aldrich); Corannulene (>97.0%(GC), TCI).

Preparation of 1,4-diazabicyclo[2.2.2]octan-1-ium iodide (HdabcoI)

1,4-Diazabicyclo[2.2.2]octane (11.22 g, 0.1 mol) was dissolved in DI water (100 ml) under nitrogen (N2) protection, and hydroiodic acid aqueous solution (13.21 ml, 0.1 mol) was added dropwise at room temperature in dark condition. The reaction mixture was stirred for 3 hours. The solvent was then evaporated under reduced pressure. The crude product was purified by recrystallization in methanol. The yield was 97%.

Preparation of 1D-Cu₄I₈(Hdabco)₄ precursor solution

In an Ar-filled glovebox, CuI (88.4 mg, 0.465 mmol) and HdabcoI (112.6 mg, 0.232 mmol) were dissolved in DMF (2 mL) and stirred at 75 °C for overnight. This solution was filtered through a PVDF filter (pore size of 0.2 μm) and kept at 75 °C before being used as the precursor solution for the subsequent fabrication of CuI(Hda) thin films.

Crystal growth of 1D-Cu₄I₈(Hdabco)₄

The single crystals of 1D-Cu₄I₈(Hdabco)₄ [hereafter also referred to as CuI(Hda)]were grown using a facile low-temperature vapor-assist recrystallization method. 2 mL of the precursor solution was added to an open vial (4 mL) and was placed in a sealed 20 mL vial with 5 mL of ethyl acetate as anti-solvent. The system was kept at 50 °C for three days. Rod-shaped transparent single crystals were obtained with a yield of 62%.

Device fabrications

Patterned ITO glasses (20 mm × 15 mm) were sonicated sequentially in detergent-deionized water solution, deionized water, ethanol, acetone and isopropanol for 15 min each, then dried with compressed N₂. The ITO glasses were then transferred into an Ar-plasma sputter (Denton Explorer) in cleanroom. The substrates were heated to 250 °C and NiO_x was deposited at a constant power of 120 W. After cooling to room temperature, the substrate was then transferred to an Ar-filled glove box and heated to 100 °C. The CuI(Hda)EML (90 nm) was fabricated by spin-coating the precursor solution (60 μL) at 4,000 r.p.m. for 60 s on 100 °C substrates. After spin-coating for 20 s, 100 μl ethyl acetate was dropped onto the film. The colorless film was kept in glovebox at room temperature for 24 hours for slow recrystallization. The electron transport layer was then fabricated by spin-coating 0.2 wt% of Ca(acac)₂ in methoxyethanol and annealed at 80°C for 15 min. The as-fabricated film samples were attached to deposition mask and transferred to an E-beam evaporator (Nexdep System, Angstrom Engineering Inc). After the chamber was pumped down to 1.8×10^-8 Torr, 1 nm LiF and 60 nm of Al were deposited sequentially. For PMMA-capped devices, 60 μL of PMMA solution in chloroform (1 mg/mL) was spin-coated on at 2,000 r.p.m for 45 s, followed by annealing at 100°C before the electron transport layer was deposited. For all devices with SAM (2PACz, Br2PACz, MeO2PACz and Ac2PACz) functionalized NiO_x HTL, 100 μL SAM solution (0.1 mmol/mL) in anhydrous ethanol was dropped onto the as-made ITO-NiO_x substrate at room temperature in Ar-filled glove box for 30s, then spin coating at 3,000 r.p.m for 60 s, then washed 2 times with anhydrous ethanol (100 μL each) while spinning at 6,000 r.p.m for 60s to remove unbonded SAM molecules, followed by annealing at 100°C for 30 min, which could directly follow the same EML deposition method stated above. 

For reference devices, PEDOT: PSS was spin-coated on the ITO glass at 4,000 r.p.m. for 60 s and annealed at 150 °C for 30 min. Corannulene thin film was fabricated by spin coating a filtered (0.2 μm PTFE) corannulene solution in chloroform (5 mg/mL) at 1,500 r.p.m for 60 s. TPbi was deposited in a high vacuum thermal evaporator. 

For single carrier devices, hole-only device with a structure of ITO/NiO_x(w/o SAM-functionalization)/CuI(Hda)/MoO₃/Au was applied, while electron-only device of ITO/corannulene/ CuI(Hda)/Ca(acac)₂/LiF/Al was adopted. MoO₃ (40 nm) and Au (40nm) were deposited sequentially by thermal evaporation in the same Nexdep system. CuI(Hda)thin film samples for other structural and optical measurements were fabricated as described above without deposition of other layers, on either ITO or quartz substrates.

Determination of single crystal structure of 1D-Cu₄I₈(Hdabco)₄

Single crystal structure analysis was carried out by single crystal X-ray diffraction method. A high-quality single crystalof 1D-Cu₄I₈(Hdabco)₄ was selected and mounted on MicroMesh (MiTeGen) with paraton oil. The data were collected on a single crystal X-ray diffractometer (Bruker D8 VENTURE) equipped with Mo micro-focus X-ray sources (λ = 0.71073 Å) at 298 K. Using Olex2, the structure was solved with the ShelXT structure solution program using Intrinsic Phasing and refined with the ShelXL refinement package using Least Squares minimization. The hydrogen atoms on carbon atoms were located at geometrically calculated positions and refined by riding. The refinement results are summarized in Table S1. Crystallographic data for the crystal structure in CIF format have been deposited in the Cambridge Crystallographic Data Centre (CCDC) under deposition number CCDC-2261492.The data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, U.K.)

Surface morphology characterizations of CuI(Hda) thin film 

The AFM study was done using an Asylum Research Cypher ES Atomic Force Microscope. The measurements were carried out under ambient conditions in the sealed AFM enclosure to reduce noise at a set sample temperature of 25 °C. Topography images were acquired by operating the AFM in tapping mode. Silicon cantilevers, HQ-300-Au (Asylum Research) and TAP300 (Ted Pella) were used each with a nominal spring constant of k = 40 Nm⁻¹ and nominal tip radius of r < 10 nm. A digital resolution of 512 lines × 512 points with a scan area of 5 µm × 5 µm and a scanning rate of 1.6 Hz were used. The oscillation frequency of the probe was set at or near resonance at 253 kHz. The spring constant of the tip was determined to be 25.26 Nm⁻¹ using thermal tune method. Kelvin probe force microscope was carried out with the EFM mode. Both samples were measured by a conductive tip with a work function of 5.1 eV.

SEM experiments were performed on a Zeiss-Sigma Field-emission scanning electron microscope (FE-SEM), while EDS data were collected on an Oxford-XmaX80 detector coupled to the FE-SEM. The elemental maps were obtained at an operating voltage of 5 kV.

Structural characterizations of CuI(Hda) thin film

PXRD analysis was carried out on both powder and thin film samples using a Rigaku Ultima-IV diffractometer with Cu Kα radiation (λ = 1.5406 Å). Specifically, the PXRD pattern of the thin film sample was taken on powders scratched and collected from 15 thin film samples fabricated on ITO. The data were collected at room temperature in a 2θ range of 3-50° with a scan speed of 0.5 °/min under the operating power of 44 kV/40 mA.

Grazing-incident wide-angle X-ray scattering pattern with an incident angle of 0.16° were obtained at Advanced Light Source beamline 7.3.3 in LBNL. 

Characterizations of photophysical properties

Optical absorption spectra of the CuI(Hda)thin film and powders were recorded at room temperature on a Shimadzu UV-3600 UV–vis–NIR spectrometer, using transmittance and reflectance mode, respectively. The diffuse reflectance data were converted to Kubelka–Munk function, α/S = (1 – R)²/2R (α is absorption coefficient, S is scattering coefficient and R is reflectance), and used to estimate the optical bandgap. The scattering coefficient (S) was treated as a constant as the average particle size of the samples used in the measurements was significantly larger than 5 μm. 

Room temperature PL measurements of a CuI(Hda) thin film on quartz substrate were performed on a Horiba Duetta fluorescence spectrophotometer. The PLQY measurements were carried out at room temperature on a C11347 absolute quantum yield measurement system (Hamamatsu Photonics) with 150 W xenon monochromatic light source and 3.3 in. integrating sphere. Powder samples for PLQY measurements were prepared by spreading fine powder samples evenly on the bottom of a quartz sample holder. Sodium salicylate was chosen as the standard with reported PLQY value of 60% at an excitation energy of 285 nm. Film samples were directly set on the PTFE base in the integrating sphere, using an uncoated quartz as blank to extract the optical absorption of the substrate.

Temperature dependent PL spectra and TRPL decay profiles were also recorded ona 90 nm thin film sample on quartz with a home-built time-correlated single photon counting instrument using the 285 nm pulsed light from a frequency doubled femtosecond solid state laser (Maitai-Spectra Physics, 100 fs pulse, 10 kHz repetition rate), a Janis cryostat model V500, and an optical detection system consisting of a single photon counting avalanche photodiode (PMD50, Picoquant, 45 ps response time), a time analyzer (TimeHarp 260 nano, Picoquant Germany) and an Ocean Optics FL65000 fiber optics spectrometer. The PL signals emitted by the sample were collected by a 50 mm biconvex lens and split by a 50/50 non polarizing beam splitter cube between the photodiode and PL fiber spectrometer. PL signals were acquired using an average power of 0.55 mW, with decays recorded in at least 1,000 channels using a 355 nm long-pass filter (Semrock). PL decays were individually fit model with the Fluofit Picoquant software using a biexponent fit model with lifetime contributions calculated as averaged amplitudes.

Femtosecond and nanosecond transient absorption measurements were performed using a pump-probe system based on a Pharos (Light Conversion) regenerative amplifier (285 nm pump delivered by an Orpheus OPA, 350fs pulses at 1 kHz repetition rate) and a Helios Fire spectrometer/microscope (Ultrafast Systems) with the white light generated by a Ti:Sapphire crystal. TA spectra and decays were analyzed by the Surface Xplorer software from Ultrafast Systems.

Polarized photoluminescence measurements were carried out in a Horiba Fluorolog-3 spectrofluorometer using a 290 nm excitation light and a 360 nm long-pass filter in the collection path. The spectrofluorometer was equipped with a 450 W xenon lamp source; a double monochromator was used on the excitation side and the PL emission was collected through an iHR 320 emission monochromator and a Hamamatsu R928P photomultiplier tube (PMT) detector. Excitation and emission slits widths were set to 3 nm (bandpass). The single crystal of CuI(Hda) was positioned with its (100) (011) and (0-11) faces perpendicular to the excitation and collection directions. Polarizers were added to the excitation and collection paths to analyze the polarization of the emitted light from the crystal. Polarization bias in the equipment was quantified using an amorphous conjugated polymer sample to give an instrument correction factor. Subsequently, the polarized PL spectra from CuI(Hda) crystals were corrected for polarization bias in the measurement using the instrument correction factor.

TRMC and DMC measurements

Time-resolved microwave conductivity (TRMC) and dark microwave conductivity (DMC) measurements were performed using a system that has already been thoroughly described³⁶. Thin films were mounted as usual by depositing material on a 25 × 11 × 1 mm UV fused silica substrate, and measurements were conducted in our standard rectangular microwave cavity with a sensitivity factor of K=23,000. Single-crystal measurements required explicit electromagnetic modeling of the cavity response for both orientations of the individual crystal studied. In addition, the complex dielectric constant of the double-sided tape used to mount these samples to the quartz substrate was also explicitly included. Separate effective sensitivity factors³⁷ were calculated for each crystal orientation. These were K_e = -159 for the perpendicular orientation of the crystal and K_e = -368 for the parallel orientation. Care was taken in these measurements to ensure that there was no variation in the optical excitation intensity received by the crystal when its orientation was changed.

Excitation light was provided by a Spectra-Physics Quanta-Ray laser operating at 30 Hz repetition rate, and pumping an OPO (GWU Premi-Scan), which in turn pumps a doubler (GWU UV-Scan) to generate the final 300 nm pump beam. The spot diameter and delivered power were measured at the sample position before and after measurements to ensure there was no unacceptable drift in laser output over the course of the experiment. Specific excitation intensities are reported in each figure where transient data appears.

SCLC measurements

SCLC measurements were done on single carrier devices of CuI(Hda) with a Keithley 2400 source meter.

Photoemission measurements

To obtain band-edge alignment information experimentally, thin film samples were transferred from the glove box to a purged glove bag attached to a Thermo ESCALAB 250xi system, in a sealed container. Core levels are measured with an Al-Kα line (hν = 1486.7 eV) with an energy resolution of 0.6 eV, while the valence band is measured with a 40.8eV photon energy and the work function with a 21.2 eV photon energy, both with an energy resolution better than 0.1 eV. The distance between the VB edge and the VL was measured in UPS by applying a -10V bias to the sample and is given as hν - (SECO edge - VB edge).  All photoemission spectra were referenced to the Fermi level of a clean metallic surface in contact with the measured samples. REELS was performed in the ESCALAB 250xi using a 30 eV electron source of 0.6 eV full width half maximum in order to measure the optical absorption onset.

Imaging of the LED device

Imaging of the cross-section of the LED devices was done on a Carl Zeiss Orion Plus Helium Ion Microscope (Carl Zeiss Microscopy, Peabody, MA) operating at 30 KeV acceleration voltage with a beam current of about 1 pA. Electron flood gun was not used for charge neutralization. The vacuum reading in the analysis chamber during imaging was 2 x 10^-7 torr.

Performance evaluation of the LED devices

A Keithley 2400 apparatus was used to measure J–V curve of the as-fabricated LED devices from 0 to 9 V with a step voltage of 0.1 V; at the same time, luminance was collected using a luminance meter (Konica Minolta, CS-200). Electroluminescence data were recorded concurrently with a home-made fiber-coupled spectrometer (GlacierTM X, BWTEK Inc) with an integrating sphere (IS200-4, THORLABS).  Other parameters used to characterize LEDs were all calculated from the L–J–V and electroluminescence measurements under the assumption that the emission of the LED exhibits a Lambertian pattern. The operational lifetime (T₅₀) was conducted using the same set-up, but under a constant current density condition in ambient air. 

DFT calculations

The spin-polarized periodic DFT calculations were performed using plane-wave VASP code³⁸ with projector-augmented wave (PAW) method to treat core-valance electron interaction. We employed a kinetic energy of 520 eV in planewave basis sets. The initial configuration of 1D-Cu₄I₈(Hdabco)₄ was imported from the SCXRD refined structure. To balance the accuracy and efficiency, geometry optimization and standard self-consistent field calculation (SCFC) were performed with the Perdew−Burke−Ernzerhof (PBE) functional³⁹ within the generalized gradient approximation (GGA) and DFT-D3M method for long-range corrections. A 2×3×2 Monkorst-Pack k-point grid was used to sample Brillouim zones. All atoms were allowed to relax until the changes in force and energy were less than 0.01 eV/A and 10^-6 eV, respectively. The final structure was in good agreement with that of experimentally determined, with the difference in the lattice constants less than 1% and was used for further electronic property calculations. To obtain more accurate density of state (DOS) and band structure HSE06 functional was used, with a 20% Hartree–Fock exact exchange.

The geometry optimizations, HOMO-LUMO energies and electrostatic potentials of the SAMs were calculated in the Gaussian 09 package⁴⁰ at the B3LYP/ def2TZVP level with DFT-D3 for the van der Waals correction⁴¹. A frequency calculation was always performed after geometry optimization of all SAM molecules to confirm that the calculations resulted in a true minimum. Multiwfn was used for ESP visualization⁴².

The surface adsorption calculations^43,44 were carried out in CP2K code⁴⁵. All calculations employed a mixed Gaussian and planewave basis sets. Core electrons were represented with norm-conserving Goedecker-Teter-Hutter pseudopotentials^46-48, and the valence electron wavefunction was expanded in a double-zeta basis set with polarization functions⁴⁹ along with an auxiliary plane wave basis set with an energy cutoff of 400 eV. The generalized gradient approximation exchange-correlation functional of PBE⁵⁰ was used. Each configuration was optimized with the Broyden-Fletcher-Goldfarb-Shanno (BGFS) algorithm with SCF convergence criteria of 1.0×10^-6 au. The van der Waals correction of Grimme’s DFT-D3 model was also adopted. 

The adsorption energy between the adsorbate (Ac2PACz) and the CuI(Hda) (200) surface or NiO (111) surface can be calculated using the following Eqation (1):

Additional References

36        Reid, O. G. et al. Quantitative analysis of time-resolved microwave conductivity data. Journal of Physics D: Applied Physics 50, 493002 (2017).

37        Zhang, F. et al. Enhanced Charge Transport in 2D Perovskites via Fluorination of Organic Cation. Journal of the American Chemical Society 141, 5972-5979, doi:10.1021/jacs.9b00972 (2019).

38        Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Physical Review B 54, 11169-11186, doi:10.1103/PhysRevB.54.11169 (1996).

39        Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized Gradient Approximation Made Simple. Physical Review Letters 77, 3865-3868, doi:10.1103/PhysRevLett.77.3865 (1996).

40        Gaussian 16 Rev. C.01 (Wallingford, CT, 2016).

41        Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J Chem Phys 132, 154104, doi:10.1063/1.3382344 (2010).

42        Lu, T. & Chen, F. Multiwfn: A multifunctional wavefunction analyzer. Journal of computational chemistry 33, 580-592 (2012).

43        Hohenberg, P. & Kohn, W. Inhomogeneous Electron Gas. Physical Review 136, B864-B871, doi:10.1103/PhysRev.136.B864 (1964).

44        Kohn, W. & Sham, L. J. Self-Consistent Equations Including Exchange and Correlation Effects. Physical Review 140, A1133-A1138, doi:10.1103/PhysRev.140.A1133 (1965).

45        VandeVondele, J. et al. Quickstep: Fast and accurate density functional calculations using a mixed Gaussian and plane waves approach. Comput. Phys. Commun. 167, 103-128, doi:http://dx.doi.org/10.1016/j.cpc.2004.12.014 (2005).

46        Goedecker, S., Teter, M. & Hutter, J. Separable Dual-Space Gaussian Pseudopotentials. Phys. Rev. B 54, 1703-1710, doi:10.1103/PhysRevB.54.1703 (1996).

47        Hartwigsen, C., Goedecker, S. & Hutter, J. Relativistic Separable Dual-Space Gaussian Pseudopotentials from H to Rn. Phys. Rev. B 58, 3641-3662, doi:10.1103/PhysRevB.58.3641 (1998).

48        Krack, M. & Parrinello, M. All-electron ab-initio Molecular Dynamics. Phys. Chem. Chem. Phys. 2, 2105-2112, doi:10.1039/b001167n (2000).

49        VandeVondele, J. & Hutter, J. Gaussian Basis Sets for Accurate Calculations on Molecular Systems in Gas and Condensed Phases. J. Chem. Phys. 127, 114105, doi:10.1063/1.2770708 (2007).

50        Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865 (1996).