Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/IC
Evaluation of Ligands Effect on the Photophysical Properties of Copper Iodide Clusters Brendan Huitorel,† Hani El Moll,† Raquel Utrera-Melero,‡ Marie Cordier,§ Alexandre Fargues,∥ Alain Garcia,∥ Florian Massuyeau,‡ Charlotte Martineau-Corcos,⊥,# Franck Fayon,# Aydar Rakhmatullin,# Samia Kahlal,g Jean-Yves Saillard,g Thierry Gacoin,† and Sandrine Perruchas*,†,‡ †
Laboratoire de Physique de la Matière Condensée (PMC), CNRSEcole Polytechnique, 91128 Palaiseau Cedex, France Institut des Matériaux Jean Rouxel (IMN), Université de Nantes, CNRS, 2 rue de la Houssinière, BP 32229, 44322 Nantes cedex 3, France § Laboratoire de Chimie Moléculaire, CNRSEcole Polytechnique, 91128 Palaiseau Cedex, France ∥ Institut de Chimie de la Matière Condensée de Bordeaux (ICMCB)CNRS, 87 Avenue du Docteur A. Schweitzer, 33608 Pessac Cedex, France ⊥ MIM, Institut Lavoisier de Versailles (ILV), UMR CNRS 8180, Université de Versailles St-Quentin en Yvelines (UVSQ), 45, avenue des Etats-Unis, 78035 Versailles Cedex, France # CNRS, CEMHTI UPR 3079, Université d’Orléans, F-45071 Orléans, France g UMR-CNRS, 6226 “Institut des Sciences Chimiques de Rennes”, Université de Rennes 1, 35042 Rennes Cedex, France ‡
S Supporting Information *
ABSTRACT: Luminescent materials based on copper complexes are currently receiving increasing attention because of their rich photophysical properties, opening a wide field of applications. The copper iodide clusters formulated [Cu4I4L4] (L = ligand), are particularly relevant for the development of multifunctional materials based on their luminescence stimuli-responsive properties. In this context, controlling and modulating their photophysical properties is crucial and this can only be achieved by thorough understanding of the origin of the optical properties. We thus report here, the comparative study of a series of cubane copper iodide clusters coordinated by different phosphine ligands, with the goal of analyzing the effect of the ligands nature on the photoluminescence properties. The synthesis, structural, and photophysical characterizations along with theoretical investigations of copper iodide clusters with ligands presenting different electronic properties, are described. A method to simplify the analysis of the 31P solidstate NMR spectra is also reported. While clusters with electron-donating groups present classical luminescence properties, the cluster bearing strong electron-withdrawing substituents exhibits original behavior demonstrating a clear influence of the ligands properties. In particular, the electron-withdrawing character induces a decrease in energy of the unoccupied molecular orbitals, that consequently impacts the emission properties. The modification of the luminescence thermochromic properties of the clusters are supported by density functional theory (DFT) calculations. This study demonstrates that the control of the luminescence properties of these compounds can be achieved through modification of the coordinated ligands, nevertheless the role of the crystal packing should not be underestimated.
■
sensor,7−11 photocatalytic,12 and organic light-emitting diodes (OLEDs)13−15 applications. Among the copper iodide compounds, the polynuclear complexes constitute a particular class of compounds exhibiting luminescence properties that can be influenced by the metallophilic interactions occurring within their molecular structure. Because of the sensitivity of these cuprophilic interactions to environmental conditions, the development of stimuli-responsive materials based on such compounds is
INTRODUCTION Luminescent materials based on copper complexes are currently receiving increasing attention because of their rich photophysical properties opening a wide field of applications in lightening and detection technologies, for instance.1−4 This success is also assigned to the economic aspect of copper metal which is earth-abundant and less expensive compared to noble metals. The family of copper(I) halides is particularly attractive owing to the bright luminescence displayed in the visible range associated with an extraordinary wide structural diversity.5,6 Indeed, copper(I) iodide complexes have been recently reported as promising emissive materials with potential © XXXX American Chemical Society
Received: December 15, 2017
A
DOI: 10.1021/acs.inorgchem.7b03160 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 1. General representation of the [Cu4I4L4] (L = phosphine ligand) copper iodide clusters studied.
state NMR spectra is also reported. Photoluminescence properties of the clusters have been studied as a function of the temperature. To rationalize the different optical behavior of the clusters, density functional theory (DFT) calculations have been performed giving insights into the different luminescence thermochromism observed.
therefore very attractive. In this context, the cubane clusters of formula [Cu4I4L4] (L = ligand), are particularly relevant.16 Their molecular structure which presents a cubane geometry formed by four copper and four iodine atoms occupying alternatively the corners of a distorted cube, comprises six Cu− Cu interactions. These compounds can actually display luminescence thermochromism,17,18 luminescence mechanochromism,19 luminescence rigidochromism,20 or luminescence vapo/solvatochromism21 properties, which are often in close relation with modification of cuprophilic interactions. The combination of multi-stimuli-responsive properties is particularly appealing to access original multifunctional materials toward sensory applications, which are currently the object of active research aiming at developing highly reactive technologies. Indeed, promising applicative perspectives have been recently reported with the successful use of [Cu4I4L4] clusters as phosphors22 or emitters23 in lighting devices. For the development of such applications, the possibility to control and modulate their photophysical properties is of prime importance and this can be achieve only by thorough understanding of the origin of the exhibited properties. Here, we describe the comparative study of a series of cubane copper iodide clusters of formula [Cu4I4L4] coordinated by different phosphine ligands, with the goal of analyzing the effect of the ligands nature on the photoluminescence properties. In particular, we want to determine in what extent this parameter can influence the energy of the emissive states responsible for the emission properties. As presented in Figure 1, the phosphine ligands studied are all based on the P(C6H4-R)3 skeleton with variation of the nature of the R group in para position. In particular, this group has been chosen in order to display different electronic properties. Therefore, the CF3 electron withdrawing group is compared to the OCH3 and CH 3 electron more donating ones. The prototypical triphenylphosphine (PPh3) derivatives18,24,25 have been also included in this series for comparison. In this study, we report the synthesis, structural and photophysical characterizations along with theoretical investigations of these copper iodide clusters, namely, C−CF3, COCH3, and C−CH3, as designed in Chart 1. The structural differences of the clusters have been analyzed by single crystal X-ray diffraction and solid-state NMR (63Cu, 31P, 13C and 1H) techniques. A method to simplify the analysis of the 31P solid-
■
EXPERIMENTAL SECTION
Synthesis. All manipulations were performed with standard air-free techniques using Schlenk equipment, unless otherwise noted. Solvents were distilled from appropriate drying agents and degassed prior to use. Copper(I) iodide and the phosphine ligands were purchased from Aldrich and used as received. [Cu4I4(PPh3)4] (C-PPh3) was synthesized according to literature procedures.18,24 [Cu4I4(P(C6H4−CF3)3)4] (C−CF3). To a suspension of CuI (0.097 g, 0.51 mmol) in 10 mL of dichloromethane was added L−CF3 (tris(4trifluoromethylphenyl)phosphine) (0.199 g, 0.42 mmol). The reaction mixture was stirred for 12 h. The white powder was filtered off and solubilized in hexane. The solution was concentrated and left for slow evaporation. Colorless crystals suitable for diffraction analysis were collected by filtration after few days. Yield = 87% (0.24 g, 0.091 mmol). 1H NMR (300 MHz, CD2Cl2): δ (ppm) = 7.64 (d, 24H, arom. CH, 3J = 8.7 Hz), 7.60 (d, 24H, arom. CH, 3J = 8.7 Hz). 31P NMR (121.5 MHz, CD2Cl2): δ (ppm) = −24.0 (br). Anal. calcd (% wt.) for C84H48F36Cu4I4P4: C, 38.41; H, 1.84. Found: C, 38.54; H, 1.96. [Cu4I4(P(C6H4−OCH3)3)4] (C−OCH3). To a suspension of CuI (0.097 g, 0.51 mmol) in 10 mL of toluene was added L−OCH3 (tris(4-methoxyphenyl)phosphine) (0.150 g, 0.42 mmol). The reaction mixture was stirred for 12 h at 110 °C. At warm temperature, the fine powder was filtered off and the filtrate was slowly cooled down to room temperature. Colorless crystals suitable for diffraction analysis were collected by filtration after several days. Yield = 70% (0.19 g, 0.074 mmol). 1H NMR (300 MHz, CD2Cl2): δ (ppm) = 7.48 (d, 24H, arom. CH, 3J = 8.7 Hz), 6.85 (d, 24H, arom. CH, 3J = 8.7 Hz), 3.80 (s, 36H, CH3). 31P NMR (121.5 MHz, CD2Cl2): δ (ppm) = −24.5 (br.). Anal. calcd (% wt.) for C84H84O12Cu4I4P4+4C7H8: C, 52.96; H, 4.60. Found: C, 52.47; H, 4.42. [Cu4I4(P(C6H4−CH3)3)4] (C−CH3). To a suspension of CuI (0.31 g, 1.64 mmol) in toluene (80 mL) was added L−CH3 (tris(4tolyl)phosphine) (0.5 g, 1.64 mmol). The reaction mixture was stirred for 12 h at 110 °C. The solution was filtered and after cooling down to room temperature the product was recovered as yellowish crystals. Yield = 81% (0.66 g, 0.33 mmol). 1H NMR (300 MHz, CDCl3): δ (ppm) = 2.34 (s, 36H, CH3), 7.01−7.20 (m, 24H, Ph), 7.30−7.53 (m, 24H, Ph). Anal. calcd (% wt.) for C84H84Cu4I4P4 + C7H8: C, 52.76; H, 4.48. Found: C 53.01; H 4.55. Characterizations. 1H and 31P liquid NMR spectra were recorded at room temperature on a Bruker Avance II spectrometer operating at the radiofrequency of 300 MHz. 1H spectra were internally referenced from peaks of residual protons in deuterated solvents. A solution of H3PO4 (85 wt %) was used as an external standard for the 31P spectra. Elemental analyses (C, H) were performed by the Service de microanalyse de l’ICSNCNRS Gif-sur-Yvette. The solid state static 63Cu NMR spectra were recorded on a Bruker Avance III 750 MHz spectrometer (B0 = 17.6 T) using the wideband uniform rate and smooth truncation26 Carr−Purcell Meiboom−Gill27 (WURST-Q-CPMG)28 pulse sequence in static condition, with an
Chart 1. Designation of the Studied Clusters
B
DOI: 10.1021/acs.inorgchem.7b03160 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry interpulse delay of 150 μs. The WURST pulse had a length of 50 μs with an excitation bandwidth of 1 MHz. 1H spinal6429 decoupling was applied during acquisition. The recycle delay was 1 s. Number of transients was about 10 000 for each sample and the number of echoes 40. The chemical shifts were referenced to CuI at 0 ppm. Three spectra with different offsets were co-added to ensure the full excitation of the pattern. The samples were packed in 4 mm outer diameter rotors. The 1H−31P cross-polarization (CP) magic-angle spinning (MAS) and 31P−63Cu J-based heteronuclear multiplequantum coherence (J-HMQC)30 MAS NMR spectra were recorded at MAS frequency of 12.5 kHz. For the 1H−31P CP, the contact time was set to 3 ms, 128 to 1024 transients were recorded for each sample with 10 s recycling delay. The recoupling time in the J-HMQC experiment was set to 320 μs (four rotor periods), the number of transients was 1024 to 5120 and the recycling delay to 10 s. 1H−31P CP was applied prior to the magnetization transfer for enhanced sensitivity. 1H SPINAL-64 decoupling is applied during excitation/ reconversion periods and during the signal acquisition. A 63Cu 90° pulse was applied after the reconversion period to remove antiphase coherence.31 The 31P chemical shifts were referenced to a solution of H3PO4 at 0 ppm. The 1H MAS and 13C CPMAS NMR spectra were recorded on an Avance Bruker 500 spectrometer (B0 = 11.7 T). The samples were packed in 3.2 or 4 mm outer diameter rotors and spun at 20 (1H) and 10 kHz (13C). The 1H and 13C chemical shifts were referenced to TMS at 0 ppm. The 19F MAS (25 kHz) NMR spectrum was recorded at 470.3 MHz Larmor frequency, using a 1H−19F−X triple-resonance 2.5 mm probe.32 A Hahn-echo sequence was applied with π/2 pulse duration of 2.5 μs and an interpulse delay set to one rotor period. 1H SPINAL decoupling was applied during the signal acquisition. The recycle delay was 5 s and 16 transients were accumulated. 19F chemical shifts are referenced to NaF at −221 ppm. Luminescence spectra were recorded on a SPEX Fluorolog FL 212 spectrofluorimeter (Horiba Jobin Yvon). The excitation source is a 450 W xenon lamp, excitation spectra were corrected for the variation of the incident lamp flux, as well as emission spectra for the transmission of the monochromator and the response of the photomultiplier (Peltier cooled Hamamatsu R928P photomultiplier). Low temperature measurements were recorded with a liquid helium circulation cryostat SMC TBT Air Liquid model C102084. The absolute internal quantum yields (Φ) were measured by using the Fluoromax-4 integrating sphere. The error evaluation of the measured values is about 10%. Emission lifetimes (τ) were recorded with a time-resolved photoluminescence setup consisting of a frequency-tripled regenerative amplified femptosecond Ti:sapphire laser system from Spectra Physics to obtain λex = 266 nm and a C7700 streak camera from Hamamatsu coupled to an SP2300 imaging Acton spectrograph for the luminescence detection. Data were analyzed by exponential curve fitting using Origin software. Powder X-ray diffraction (PXRD) diagrams were recorded on an X’Pert Philips diffractometer (40 kV, 40 mA) with CuKα radiation (λ = 0.154056 nm). The calculated patterns were obtained from the single crystal data using the Mercury software. Single crystal X-ray diffraction analyses were realized by mounting the crystals on Kapton loop using paraton oil. All data were collected at 150 K on a Nonius Kappa CCD diffractometer using Mo Kα (λ = 0.71073 Å) X-ray source and a graphite monochromator. The cell parameters were initially determined using more than 50 reflections. Experimental details are described in Table S1. The crystal structures were solved with SIR 9733 and SHELXT-201434 and refined with SHELXL-201435 by full-matrix least-squares using anisotropic thermal displacement parameters for all non-hydrogen atoms. All the hydrogen atoms were placed in geometrically calculated positions. Details of crystal data and structure refinements are summarized in Table S1. The CCDC references are 1582204, 1582205, and 1582206 for C− CH3, C−OCH3, and C−CF3, respectively.
double-ξ polarized basis set, namely the LANL2DZ set,38 augmented with polarization functions on all atoms, that is, a p orbital with exponent 0.8 for H, a d orbital with exponents 0.8, 0.55, and 0.309 for C, P, and I, respectively, and a f orbital with exponent 0.8 on Cu. Spin-unrestricted calculations were performed in the case of triplet states. Vibrational frequency calculations have been performed on all the optimized structures in order to ascertain they are minima on the potential energy surface. The ground state optimized geometries of the four compounds were found to be of T symmetry. The T1 and T2 triplet states were optimized assuming C1 symmetry. The emission wavelengths were calculated as the difference between the energy of the optimized T1 or T2 triplet state and the energy of the singlet ground state assuming the same (unrelaxed) geometry as that of the corresponding triplet state. The compositions of the molecular orbitals were calculated using the AOMix program.39 The UV−visible transitions were calculated by means of timedependent DFT (TD-DFT) calculations40 at the same level of theory.
■
RESULTS AND DISCUSSIONS Synthesis and X-ray Diffraction Characterization. The phosphine ligands used for the clusters synthesis bear groups of different nature at the para position of the phenyl moieties. These groups present increasing electron withdrawing character according to the following order: OCH3, CH3, H, and CF3. Note that the effect of the variation of nitrogen-based ligands on the luminescence properties of copper iodide clusters has been investigated,41 but to our knowledge no systematic study has been reported. [Cu4I4(P(C6H4−CF3)3)4] (C−CF3) was synthesized by reacting CuI with the trifluoromethylphenylphosphine ligand in dichloromethane (details in Experimental Section). [Cu 4 I 4 (P(C 6 H 4 −OCH 3 ) 3 ) 4 ] (C−OCH 3 ) and [Cu4I4(P(C6H4−CH3)3)4] (C−CH3) were synthesized in toluene by mixing CuI with the corresponding phosphine ligand. Note that the synthesis of C−CH3 has been already reported using an acetonitrile/chloroform solvent mixture.42 All compounds were obtained as colorless crystals except C−CH3 whose crystals are slightly yellow (Experimental Section). The crystal structures were determined by single crystal X-ray diffraction analysis at 150 K. The crystal structure of C−CH3 has been already reported at 193 K.42 The two crystalline polymorphs with the triphenylphosphine ligand formulated [Cu4I4(P(C6H5)3)4], namely C-PPh3Y and C-PPh3G, for the yellow and the green emissive one respectively, have been synthesized according to previous reports and their X-ray diffraction crystal structures described here are those already reported.18,24 Molecules of toluene are included in the structure of C− OCH3 and C−CH3 leading to the formulas [Cu4I4(P(C6H4− OCH3)3)4]·4(C7H8) and [Cu4I4(P(C6H4−CH3)3)4]·C7H8, respectively, which are in agreement with elemental analysis. Liquid NMR (1H and 31P) and powder X-ray diffraction analysis (spectra in SI), also confirm the purity of the samples. C−CF3 crystallizes in the triclinic P1̅, C−OCH3 in the tetragonal I41/a and C−CH3 in the trigonal R3̅ space groups. The two known polymorphs C-PPh3 crystallize in the monoclinic P21/c (C−PPh3Y) and cubic I43̅ d (C−PPh3G) space groups. The unit cell contents of C−CF3, C−OCH3, and C−CH3 are shown in Figures S1. All three structures can be described as assembly of columns of clusters. For C−OCH3, the columns of clusters run along the c axis and for C−CF3 and
■
COMPUTATIONAL DETAILS DFT calculations have been carried out with the program Gaussian03,36 using the PBE037 functional and a standard C
DOI: 10.1021/acs.inorgchem.7b03160 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 2. Molecular structures of the studied copper iodide clusters [Cu4I4L4].
C−CH3 along the b axis and the orientation of the cluster changes from one chain to the other for the latter cluster. The molecular structures of the clusters are depicted in Figure 2. All clusters present the classical cubane structure formed by four copper atoms and four iodine atoms which occupy alternatively the corners of a distorted cube. The phosphine ligands are coordinated to each copper atom by the phosphorus atom. Therefore, all the copper atoms present a pseudotetrahedral PCuI3 geometric environment. Selected bond lengths and angles of the clusters are listed in Table 1. The ligands present a classical geometry for phosphine-based copper iodide clusters with similar Cu−P bond values. The Cu−I bond distances and the I−Cu−I angle mean values for all the clusters are comparable and within the range of reported values for this type of clusters coordinated by phosphine ligands. The main difference between the molecular structures are the Cu−Cu distances. All the clusters have mean Cu−Cu distances around 3 Å but C−OCH3 presents particularly long distances with one at 3.3 Å. This results in a Cu4 tetrahedron with a larger volume and more distorted for this cluster (Table 1). These long Cu−Cu distances are much greater than the sum of the van der Waals radii of 2.80 Å,43 which imply weak or no cuprophilic interaction (d10−d10 bonding), in this case.44 The two previously studied crystalline polymorphs of C−PPh324 present also different cuprophilic interactions with shorter Cu−Cu distances for C−PPh3Y (mean value = 2.901(1) Å) compared with those of C−PPh3G (mean value = 2.996(2) Å), the longer distances being around 3.1 Å for both of them (Table S2). Solid-State NMR Characterizations. The five different clusters were further analyzed by 1H, 13C, 31P, and 63Cu solidstate NMR spectroscopy, a local probe that may provide indications about the structural features responsible for their distinct photophysical properties. In particular, the NMR spectra of C−PPh3Y, already reported,18 are compared with the new data of its crystalline polymorph C−PPh3G, allowing a straightforward analysis of their structural differences. The 1H and 13C MAS NMR spectra (Figure S7) are in agreement with triphenyl-based ligands. For CF3, the 19F MAS spectrum has been additionally recorded (Figure S12). It can be deconvoluted using seven strongly overlapping contributions (Table S6), with relative intensities that correspond to the 36 independent fluor atoms in the structure. The 63Cu static NMR spectra of all the studied clusters are gathered in Figure 3. The 63Cu nucleus has a large quadrupolar moment, giving rise to particularly broad NMR spectra, even at high magnetic field (17.6 and 18.8 T for C-PPh3Y), in which the individual components overlap and are not resolved. We propose deconvolutions of the central transition patterns of CPPh3G, C−CH3, and C−OCH3 as they only have two
independent copper sites each (Figure S9). For C−CF3, the determination of the chemical shift and quadrupolar parameters of the four overlapping contributions (four inequivalent Cu sites), which would require recording spectra at several different magnetic fields, was not attempted here. The deconvolution of the C−PPh3Y spectrum was previously reported.19d The 63Cu NMR parameters are listed in Table S4. In overall, the studied compounds have 63Cu chemical shifts δiso (−70 to 450 ppm), quadrupolar coupling constants CQ (∼10 MHz for C-PPh3Y, 20−25 MHz for the others) and asymmetry parameters ηQ (
