Cu–Fe–S Nanocrystals Exhibiting Tunable Localized Surface

Article pubs.acs.org/IC

Cu−Fe−S Nanocrystals Exhibiting Tunable Localized Surface Plasmon Resonance in the Visible to NIR Spectral Ranges Grzegorz Gabka,† Piotr Bujak,*,† Andrzej Ostrowski,† Waldemar Tomaszewski,† Wojciech Lisowski,‡ Janusz W. Sobczak,‡ and Adam Pron† †

Faculty of Chemistry, Warsaw University of Technology, Noakowskiego 3, 00-664 Warsaw, Poland Institute of Physical Chemistry, Polish Academy of Science, Kasprzaka 44/52, 01-224 Warsaw, Poland

‡

S Supporting Information *

ABSTRACT: Cu−Fe−S nanocrystals exhibiting a strong localized surface plasmon resonance (LSPR) effect were synthesized for the first time. The elaborated reproducible preparation procedure involved copper(II) oleate, iron(III) stearate, and sulfur powder dissolved in oleylamine (OLA) as precursors. The wavelength of the plasmonic resonance maximum could be tuned by changing the Cu/ Fe ratio in the resulting nanocrystals, being the most energetic for the 1:1 ratio (486 nm) and undergoing a bathochromic shift to ca. 1200 nm with an increase to 6:1. LSPR could also be observed in nanocrystals prepared from the same metal precursors and sulfur powder dissolved in 1-octadecene (ODE), provided that the sulfur precursor was taken in excess. Detailed analysis of the reaction mixture by chromatographic techniques, supplemented by mass spectrometry and 1H NMR spectroscopy enabled the identification of the true chemical nature of the sulfur precursor in S/OLA, namely, (C18H35NH3+)(C18H35NH−S8−), a reactive product of the reduction of elemental sulfur by the amine groups of OLA. In the case of the S/ODE precursor, the true precursors are much less reactive primary or secondary thioethers and dialkyl polysulfides.

■

INTRODUCTION Localized surface plasmon resonance (LSPR) is a well-known phenomenon characteristic of metal nanoparticles. There are many reports on fundamental studies of this phenomenon,1 and it has also found applications in various fields of science and technology, for example, in ultrasensitive spectroscopic detectors used in biodiagnostics,2 including those exploiting surface-enhanced Raman spectroscopy (SERS).3 More recently, LSPR was also discovered in nanocrystals of doped inorganic semiconductors.4 These new plasmonic nanomaterials are extremely interesting because their resonance conditions can be precisely tuned by changing the concentration of free charge carriers through chemical or electrochemical doping.5,6 Nonstoichiometric copper chalcogenides such as Cu 2 − x S, 7 − 1 1 Cu 2 − x Se, 1 0 , 1 2 Cu 2 − x Te, 1 3 − 1 5 Cu2−xSySe1−y,16−18 and Cu2−xTeyS1−y18 are, so far, the uncommon examples of nonmetallic plasmonic nanomaterials. LSPR has also been reported for ternary CuxInySz and CuxSnySz nanocrystals.19,20 Their plasmonic peaks are located in the nearinfrared (NIR) part of the spectrum, but the exact peak positions and intensities can be tuned by the nanocrystals’ composition12 and shape.7,9 In addition, the surface chargecarrier density can be controlled by selecting appropriate ligands. In particular, in the case of Cu2−xS and Cu2−xSe, uncharged ligands such as oleylamine stabilize the plasmon resonance, whereas anionic ligands such as carboxylates serve as © XXXX American Chemical Society

traps for holes. As a result, the plasmonic peak is bathochromically shifted, and its intensity decreases.10 It is widely believed that nonmetallic plasmonic nanocrystals belong to the groups of nanomaterials that should be the most intensively studied.21 Comin and Manna, in their comprehensive article devoted to semiconductor nanocrystals exhibiting LSPR phenomena, classified nonstoichiometric ternary nanocrystals, CuxFeySz, originating from CuFeS2, CuFeS3, CuFe2S3, and Cu5FeS 4 semiconductors, as potentially promising plasmonic nanomaterials.6 However, to date, no investigation of the LSPR phenomenon in Cu−Fe−S nanocrystals has been reported. Despite their interesting properties such as low band gaps (0.6−0.7 eV for the bulk) and promising Seebeck coefficients,22 CuFeS2 nanocrystals are among the least studied semiconducting colloidal nanoparticles in terms of the number of articles and patents.23 Few preparation methods are known that yield CuFeS2 nanocrystals of controlled composition,24,25 size,22 shape,26 and crystal structure.24,27 In this article, we report, for the first time, Cu−Fe−S nanocrystals exhibiting strong LSPR and also demonstrate that, by controlling the nanocrystal composition, the observed plasmonic peak can be shifted from the visible part of the Received: April 13, 2016

A

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Table 1. Precursor Molar Ratios, Compositions of Nanocrystals, and Reaction Yields Determined for Nanocrystals Obtained from Copper(II) Oleate, Iron(III) Stearate, and Sulfur Dissolved in Either OLA (A-1) or ODE (B-1, B-2) A-1 B-1 B-2 a

mass (mg)

C + O (wt %)

Cu + Fe + S (wt %)

Cu/Fe/S molar ratio of precursors

nanocrystal compositiona

yield (%)

233 395 370

32.0 58.0 56.0

68.0 42.0 44.0

1.0:1.0:2.0 1.0:1.0:1.6 1.0:1.0:2.4

Cu1.0Fe1.0S1.8 (S2.0) Cu1.0Fe0.6S0.7 (S1.4) Cu1.0Fe0.7S1.2 (S1.6)

87.0 91.0 89.0

Amounts of S2− anions necessary to fully compensate Cu+ and Fe3+ cations indicated in parentheses . equipped with a LYNXEYE position-sensitive detector using Cu Kα radiation (λ = 0.15418 nm). The data were collected in the Bragg− Brentano (θ/2θ) horizontal geometry (flat reflection mode) between 10° and 70° (2θ) in a continuous scan, using 0.04° steps at 960 s/step. The incident-beam path in the diffractometer was equipped with a 2.5° Soller slit and a 1.14° fixed divergence slit, whereas the path of the diffracted beam was equipped with a programmable antiscatter slit (fixed at 2.20°), a Ni β-filter, and a 2.5° Soller slit. The sample holder was rotated at an angular speed of 15 rpm. The data were collected under standard conditions (temperature and relative humidity). Transmission electron microscopy (TEM) analysis was performed on a Zeiss Libra 120 electron microscope operating at 120 kV. Elemental analysis was carried out with a multichannel Quantax 400 energy-dispersive X-ray spectroscopy (EDS) system with a 125 eV xFlash Detector 5010 (Bruker) using a 15 kV electron beam energy. For X-ray photoelectron spectroscopy (XPS) analysis, the nanocrystals were first dispersed in chloroform and then deposited on a Si(100) wafer and dried at room temperature. Survey and high-resolution (HR) XPS spectra were recorded using a PHI 5000 VersaProbe (ULVAC-PHI) spectrometer with monochromatic Al Kα radiation (hν = 1486.6 eV). The HR XPS spectra were collected with a hemispherical analyzer at a pass energy of 23.5 eV, an energy step size of 0.1 eV, and a photoelectron take-off angle of 45° with respect to the surface plane. CasaXPS software was used to evaluate the obtained XPS data. Deconvolution of the HR XPS spectra was performed using a Shirley background and a Gaussian peak shape with 30% Lorentzian character. The binding-energy (BE) scale of all detected spectra was referenced by setting the BE of C 1s to 284.8 eV. For quantification, the PHI Multipak sensitivity factors and determined transmission function of the spectrometer were used. UV−vis−NIR spectra were recorded using a Cary 5000 (Varian) spectrometer. 1H NMR spectra were recorded on a Varian Mercury (500 MHz) spectrometer and referenced with respect to tetramethylsilane (TMS) and solvents. MS [electron-impact (EI)] spectra were recorded on an AutoSpecPremier (Waters) mass spectrometer. GC/MS analyses were performed using a system equipped with a GC 7890A gas chromatograph coupled with a VL MSD 5975C mass detector, both from Agilent. An HP-1701 capillary column (30 m × 250 μm × 0.25 μm) was used, with helium as the carrier at 1 mL/min. The sample volume was 0.1 μL at a split value of 1:50, and the injector temperature was 200 °C. The temperature program was as follows: From an initial temperature of 40 °C held for 2 min, a linear temperature increase of 25 °C/min to 250 °C was applied. After reaching 250 °C, the temperature was held constant for 25 min. No additional peaks were recorded when the temperature range was extended to 300 °C using the same temperature program. The analyses were performed using electrospray ionization, and chromatograms were acquired in total ion chromatogram (TIC) mode at m/z 10−700. Chloroform of p.a. grade was used for preparing the solutions, which had a concentration of 30 mg/mL.

spectrum to the NIR. This discovery was made possible through the elaboration of new synthetic procedures. Moreover, using a combination of spectroscopic and chromatographic techniques, we were able to elucidate all transformations of the precursors and other reagents in the course of the reaction and to chemically characterize the nanocrystal surface by identifying the chemical nature of the capping ligands.

■

EXPERIMENTAL SECTION

Preparation of S/OLA Precursor. Sixty-four milligrams (2.0 mmol) of sulfur powder and 2.0 mL of oleylamine were loaded into a glass vial, which was then immersed in an ultrasonic bath. The mixture was sonicated at room temperature (for about 10 min) until a clear red solution was formed. Preparation of S/ODE Precursor. Six hundred forty milligrams (20 mmol) and 25 mL of 1-octadecene were loaded into a 50 mL three-neck flask. The mixture was heated to 200 °C under a nitrogen atmosphere and kept at this temperature for ca. 20 min to yield a clear red solution. It was then allowed to cool to room temperature. Synthesis of Cu−Fe−S Nanocrystals. All operations were carried out under constant dry flow of argon. Copper(II) oleate (96.0%) (0.653 g, 1.0 mmol) and iron(III) stearate (90.3%) (1.000 g, 1.0 mmol) were mixed with 15 mL of 1-octadecene in a three-neck flask. The mixture was heated to 160 °C until a homogeneous greenbrown solution was formed. Then, 2 mL of S/OLA precursor was quickly injected into the reaction solution. Upon injection, the color of the solution instantly changed to black. The temperature was raised to 240 °C, and the mixture was kept at this temperature for 60 min. After the mixture had cooled to room temperature, chloroform (10 mL) was added, and the reaction mixture was centrifuged; the isolated precipitate consisting of organic waste and agglomerated particles was separated. The supernatant was treated with 30 mL of ethanol, leading to the precipitation of the desired fraction of nanocrystals. The nanocrystals were separated by centrifugation (7000 rpm, 5 min), dried under a vacuum, and finally redispersed in chloroform. The amounts of the reagents used in all preparations are listed in Table S1 (Supporting Information). Ligand Recovery. A colloidal solution of Cu−Fe−S nanocrystals (200 mg in 10 mL of chloroform) and 10 mL of concentrated HCl were placed in a screw-capped ampule, which was immersed in an ultrasonic bath for 2 h (and shaken every 30 min). Water (20 mL) was added, and the resulting mixture was centrifuged to achieve phase separation; the remaining solids were discarded. The organic phase was collected, and the aqueous phase was extracted with 15 mL of chloroform. The combined organic layers were washed twice with water, evaporated, and dried under reduced pressure to afford a black oil. Preparation of Samples for Gas Chromatography/Mass Spectrometry (GC/MS) Analysis. Cu−Fe−S nanocrystals were synthesized as described above, except that 1,2-dichlorobenzene was used as the solvent. Then, nanocrystals were separated by precipitation with methanol and centrifugation. The supernatant was distilled under reduced pressure using a Venturi pump until no condensate was collected at 100 °C. Residual viscous oil was diluted with chloroform and injected onto the GC column. A blank sample was prepared in the same way but without the addition of any metal precursors. Characterization Methods. X-ray powder diffractograms were recorded at room temperature on a Bruker D8 Advance diffractometer

■

RESULTS AND DISCUSSION The general procedure for the preparation of Cu−Fe−S nanocrystals elaborated in this research can be outlined as follows: Sulfur precursor was injected at 160 °C into a mixture of copper(II) oleate and iron(III) stearate in 1-octadecene. The reaction mixture was then heated to 240 °C and kept at this temperature for 1 h. After the solution had cooled to room temperature and the noncolloidal part had been separated B

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 1. X-ray diffractograms, TEM images, and UV−vis−NIR spectra (in chloroform) of (a) A-1 (Cu1.0Fe1.0S1.8), (b) B-1 (Cu1.0Fe0.6S0.7), and (c) B-2 (Cu1.0Fe0.7S1.2).

comments. The method used is based on combining mass balance with EDS analysis and does not necessitate knowledge of the composition of the organic shell. After careful drying of the sample, its mass was recorded, and the electron dispersion X-ray spectrum was measured. Based on the EDS data (see Figure S1, Supporting Information) and the total mass of the sample, the masses of Cu, Fe, and S in the inorganic core and the mass of the organic shell can be calculated. Dividing the mass of Cu in the nanocrystals by its mass in the precursor solution leads to the reaction yield. Independently of the sulfur precursor used (S/OLA or S/ODE), high reaction yields were obtained (∼90%). Sample A-1 was obtained with the S/OLA sulfur precursor using the stoichiometric Cu/Fe/S ratio of 1.0:1.0:2.0. In the resulting nanocrystals, the composition of the inorganic core was close to stoichiometric with a Cu/Fe molar ratio of 1, clearly indicating equal reactivity of the sulfur precursor toward the precursors of copper and iron. The large contribution of the inorganic core to the total mass of the nanocrystal should also be noted (68 wt %). Use of S/ODE as the sulfur precursor led to different results. First, S/ODE was found to be more reactive toward the copper precursor than toward the iron precursor, as evidenced by the

through centrifugation, the obtained nanocrystals were precipitated from the supernatant with an excess of methanol. Two different solutions were used as sulfur precursors: sulfur dissolved in either oleylamine (S/OLA) or 1-octadecene (S/ ODE). It turned out that these two solutions differed in their reactivity toward the metal precursors, which had a profound effect on the composition, crystal structure, and type of surfacial ligands in the resulting nanocrystals. The results are summarized in Table 1, where the masses of the obtained nanocrystals are collected together with the contributions of organic ligands and inorganic cores to the total mass (in weight percentages). The content of the organic part varies because it is dependent on the surface stoichiometry and the composition of the reaction medium. It is known that high-boiling surfactants used in preparation procedures are difficult to remove even after repeated washing.28,29 Moreover, they can also influence the surface structure.30−33 This is also the case in the present study, as evidenced by complementary XPS and NMR investigations (vide infra). In the fifth and sixth columns of Table 1, the precursor molar ratio in the reaction mixture and the molar ratio of the elements in the nanocrystal core are compared. In the last column, the percentage reaction yields are listed. The method used for their determination requires some C

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

the reaction temperature to 180 °C (because of the lower solvent boiling point) but allowed possible ambiguity concerning the origin of unsaturated hydrocarbons detected by 1H NMR spectroscopy in the recovered organic fraction of the nanocrystals to be avoided (vide supra). The principal characteristics of the nanocrystals prepared under these conditions (chemical composition, crystal structure, UV−vis− NIR spectra) were essentially the same as those recorded for nanocrystals obtained using 1-octadecene. Figure 2 shows 1H

chemical analysis of samples B-1 and B-2. The nanocrystal cores in both cases were enriched in Cu relative to the Cu/Fe ratio in the reaction mixture (see Table 1). Another characteristic feature of the nanocrystals obtained with S/ ODE is a deficit of sulfur: The contents of S2− were lower than what would be necessary to fully compensate for the charge of Cu+ and Fe3+ cations (see the values in parentheses in column 6 of Table 1). This effect was observed independently of the Cu/ S ratio in the reaction mixture; that is, it was observed both in B-2, where the sulfur precursor was taken in excess (with respect to the CuFeS2 stoichiometry), and in B-1, where it was taken in deficit (Table 1). This means that the nanocrystal surface is enriched in cations that have to be compensated by ligands of anionic type (vide supra). This deficit is smaller for B-2 (S1.2 vs S1.6, i.e., 25%) than for B-1 (S0.7 vs S1.4, i.e., 50%); however, it still persists. To summarize this section of the article, S/OLA shows equal reactivity toward copper(II) oleate and iron(III) stearate. Its use as a precursor results in a higher degree of sulfur to sulfide anion conversion as compared to the case of S/ODE. The latter is more reactive toward the copper precursor, yielding Cuenriched nanocrystals. X-ray diffractograms of samples A-1, B-1, and B-2 are presented in Figure 1. The diffraction pattern of A-1 (Cu1.0Fe1.0S1.8) nanocrystals is characteristic of the chalcopyrite structure of CuFeS2 (JCPDS 37-0471). The obtained nanocrystals range in size from 5 to 15 nm, predominantly exhibiting polygonal shape. In the UV−vis−NIR spectrum of a colloidal solution of these nanocrystals in chloroform, a distinct peak with a clear maximum at 486 nm can be distinguished, originating from the localized surface plasmon resonance (LSPR). Much smaller (6.8 ± 0.8 nm), sphere- or cube-shaped B-1 (Cu1.0Fe0.6S0.7) nanocrystals do not have the chalcopyrite structure (Figure 1). Their broadened X-ray reflections cannot be attributed to the structure of stoichiometric chalcocite (Cu2S) (JCPDS 26-1116). The observed diffraction pattern shows, however, a close similarity to those of regular Cu5FeS4 (JCPDS 34-0135) and regular Cu1.8S (JCPDS 47-1748). No plasmonic peak can be found either in the visible part of the UV−vis−NIR spectrum of their colloidal dispersion nor in its near-infrared (NIR) part, that is, in the spectral range where plasmonic absorption is expected for Cu1.8S nanocrystals.7−11 B-2 (Cu1.0Fe0.7S1.2) nanocrystals, showing higher contents of Fe and S as compared to the B-1 nanocrystals, are much larger and irregular in shape. Their diffraction pattern corresponds to the chalcopyrite structure of CuFeS2 (JCPDS 37-0471). In the UV−vis−NIR spectrum, a clear plasmonic absorption is observed with a peak at 494 nm. It is known that LSPR is very sensitive to the type of environment and, in particular, to the nanocrystal’s dielectric properties. For this reason, we undertook an analysis of the organic shells in samples A-1 and B-1, namely, both nanocrystals in which the plasmonic peak is very strong and nanocrystals that do not exhibit the plasmon resonance effect. For this purpose, we used a procedure, recently developed in our group, that consists of selective dissolution of the inorganic core with simultaneous recovery of the surfacial organic ligands. This procedure was previously applied with success to the identification of surfacial ligands in ternary alloyed Cu−In− Zn−S nanocrystals34 and Cu2ZnSnS4.35 In the preparation of nanocrystals for this analysis, 1-octadecene was replaced by 1,2dichlorobenzene. The use of this solvent required us to lower

Figure 2. 1H NMR spectra of the organic fractions of A-1 (Cu1.0Fe1.0S1.8) and B-1 (Cu1.0Fe0.6S0.7). Spectra of 1-octadecene, oleylamine, oleic acid, and stearic acid are shown for comparison.

NMR spectra of the organic fractions of A-1 and B-1, recovered after the dissolution of the inorganic core. The 1H NMR spectra of 1-octadecene, oleylamine, oleic, and stearic acids are also included for comparison. All spectra were recorded with CDCl3 as the solvent. In the spectrum of the organic part of A-1 (Cu1.0Fe1.0S1.8), nearly stoichiometric nanocrystals obtained with S/OLA signals characteristic of olefin groups can be distinguished. In particular, two multiplets at 4.93−5.01 and 5.80−5.88 ppm can be ascribed to vinyl protons (CHCH2), whereas the multiplet at 5.37−5.41 ppm can be ascribed to vinylene protons. A clear signal at 1.95−2.04 ppm corresponds to the protons of methylene groups adjacent to vinylene groups. No lines corresponding to protons of methylene groups adjacent to other functional groups can be found at 2.35 ppm ( CH2COOH), 2.52 ppm (CH2SH), or 2.67 ppm ( CH2NH2), which are recorded when carboxylic acids, thiols, or amines, respectively, are primary ligands.35 Because the reaction was carried out in 1,2-dichlorobenzene, the recovered terminal alkenes had to be formed in situ during the dissolution of the inorganic core, as stabilization of colloidal nanocrystals D

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

713.9 eV, tentatively ascribed to Fe(II) in the form of FeS and FeSO4.40 In the case of A-1, the content of sulfur at the nanocrystal surface, determined by XPS, is essentially stoichiometric [Cu1.0Fe2.5S4.2 (S4.2)]; the same is observed in the bulk by EDS [Cu1.0Fe1.0S1.8 (S2.0)]. In B-1, both EDS [Cu1.0Fe0.6S0.7 (S1.4)] and XPS [Cu1.0Fe10.7S10.2 (S16.5)] show a deficit of sulfur with respect to stoichiometry. In Figure 3, HR XPS S 2p

by alkenes in air is unlikely. In the simplest explanation, the presence of aliphatic ligands, attached to the nanocrystal surface through covalent bonds with surfacial sulfur atoms, can be postulated. These aliphatic groups are being transformed into alkenes during the dissolution of the inorganic core through the disruption of the Salkane bond. The presence of alkyl ligands covalently attached to surfacial sulfur atoms was postulated for Cu2S nanocrystals, depending on the preparation conditions.36 We have also detected ligands of this type in Cu2ZnSnS4 of the kesterite structure.35 The organic fraction of strongly nonstoichiometric B-1 (Cu1.0Fe0.6S0.7) nanocrystals gives a different 1H NMR spectrum. In addition to signals characteristic of alkenes, present in the spectrum of A-1 (Cu1.0Fe1.0S1.8), an additional peak is recorded at 2.35 ppm (CH2COOH) that is characteristic of carboxylate anions originating from copper and iron precursors used in the preparation of nanocrystals. To summarize this section of the article, nearly stoichiometric A-1 (Cu1.0Fe1.0S1.8) nanocrystals, prepared with the S/ OLA precursor, are stabilized by long aliphatic groups attached to the surfacial sulfur atoms through a covalent bond. In this case, the neutral nature of the ligand does not affect the chargecarrier density whose electromagnetic field-induced oscillations yield the LSPR effect, as evidenced by a strong peak at ∼500 nm in the UV−vis−NIR spectrum. In the case of nonstoichiometric B-1 (Cu1.0Fe0.6S0.7) nanocrystals, obtained using S/ODE, the nanocrystal surface is strongly enriched in cations. Thus, the excess positive charge is compensated by carboxylic ligands originating from the metal precursors. Their presence at the nanocrystal surface results in the trapping of positive charge carriers, leading to the quenching of LSPR. The deficit of surfacial sulfur atoms can be substantially reduced by using an excess of S/ODE (with respect to the CuFeS2 stoichiometry), as was done in the preparation of B-2 (Cu1.0Fe0.7S1.2) nanocrystals. As a result, fewer anionic ligands are necessary to compensate the extra positive charge, and in these nanocrystals, a clear plasmonic peak is observed, although less pronounced than in the case of A-1 (Cu1.0Fe1.0S1.8) nanocrystals. Detailed XPS studies were carried out for samples A-1 and B1. Their survey spectra are presented in Figure S2, and the HR Cu 2p, Fe 2p, C 1s, and O 1s spectra are collected in Figures S3 and S4 (Supporting Information). A distinct difference between the surfacial and bulk compositions can be noticed by comparison of EDS and XPS results. In the case of A-1, EDS yields a Cu/Fe ratio of 1.0, whereas the Cu/Fe ratio determined from XPS is 0.4. In its HR XPS Cu 2p spectrum, a clear doublet at 932.0 eV (Cu 2p3/2) and 952.0 eV (Cu 2p1/2), attributable to Cu(I), can be observed.37 Its deconvolution gives rise to an additional peak of much lower intensity, yielding XPS parameters of 933.5 eV (Cu 2p3/2) and 954.4 eV (Cu 2p1/2), typical of Cu(II) in the CuO phase.38 Nanocrystals of B-1 contain only a small amount of copper on their surface (Cu/Fe = 0.1). A weak peak at 932.0 eV (Cu 2p3/2) in its HR XPS Cu 2p spectrum can be ascribed to the presence of Cu(I).37 In both samples A-1 and B-1, the nanocrystal surface is enriched in iron. In their HR XPS Fe 2p spectra, clear doublets at 709.6−710.2 eV (Fe 2p3/2) and 722.9−723.5 eV (Fe 2p1/2) can unequivocally be ascribed to the structure of chalcopyrite Cu(I)Fe(III)S2.39 However, the deconvolution of the broadened lower-binding-energy peak reveals the presence of an additional spectral line at 713.0−

Figure 3. High-resolution S 2p XPS spectra of (a) A-1 and (b) B-1 Cu−Fe−S nanocrystals.

spectra of A-1 and B-1 are compared. Deconvolution of the A-1 spectrum indicates the coexistence of three types of chemically nonequivalent sulfur atoms. The most intense peaks at 161.2 and 162.1 eV correspond to surfacial S2− ions that are integral parts of the crystal lattice. Peaks at 162.4 and 164.5 eV can be attributed to sulfur atoms of aliphatic thiol ligands,36 whereas peaks in the range of 168.8−170.5 eV originate from the presence of S(VI) in the form of sulfates.41 Deconvolution of the spectrum of B-1 indicates that only two types of sulfur atoms coexist. Peaks at 161.3 and 162.5 eV correspond to S2− in the crystal lattice. More intense peaks at 163.1 and 164.3 eV must be attributed to sulfur in aliphatic thiols bound to the nanocrystal surface. No peaks originating from the oxidized forms of sulfur can be observed in the spectrum of B-1. The HR XPS C 1s spectra of A-1 and B-1 are consistent with the S 2p spectra. A strong peak at 284.8 eV can be deconvoluted into a strong component at 286.0 eV, corresponding to methylene (methyl) carbons, and a weaker peak at 288.5 eV ascribed to carboxylic carbons.42 The higher intensity of these peaks in the case of B-1 is fully consistent E

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 4. X-ray diffractograms of Cu−Fe−S nanocrystals of varying Cu/Fe/S ratios and UV−vis−NIR spectra of their dispersions in chloroform.

with the 1H NMR data obtained for the organic components of the nanocrystals (vide supra). Moreover, deconvolution of the peaks in the HR XPS O 1s spectra of A-1 and B-1 clearly indicates the dominant presence of carboxylic oxygens (peak at ∼532.0 eV).43 Lower-intensity peaks at 530.0 eV correspond to oxide-type oxygens, such as CuO.38 To summarize this section of the article, in sample A-1, for which the localized surface plasmon resonance is observed, the nanocrystals surface is enriched in Fe(III) cations, and the majority of sulfur is present in the form of S2− anions that are parts of the crystal lattice; some of them, however, are oxidized to higher oxidation degrees [S(VI)]. The surface of B-1 nanocrystals, which do not show the LSPR effect, is distinctly different. First, almost no copper can be found on the surface. Second, the content of S2− is diminished as compared to the case of sample A-1. This means that the cation-enriched surface must be compensated by anionic ligands. This conclusion is confirmed by HR XPS C 1s and O 1s spectra, which show a higher content of carboxylic groups in the case of B-1 nanocrystals, consistent with the NMR findings. XPS data also indicate that, in both cases, aliphatic thiol ligands are present. As already mentioned, the use of S/OLA as a precursor resulted in the most pronounced plasmonic peak in nearly stoichiometric Cu1.0Fe1.0S1.8 (S2.0) nanocrystals. We were tempted to verify how the composition of Cu−Fe−S nanocrystals influences the shape and the position of their plasmonic peak. A series of samples with varying molar ratios of copper and iron precursors were prepared: Cu/Fe = 1.0, 2.0, 4.0, and 5.0 (samples A-1−A-4, respectively). Based on the EDS spectra (see Figure S5, Supporting Information), the compositions of the obtained nanocrystals were as follows: A-1, Cu 1.0 Fe 1.0 S 1.8 (S 2.0 ); A-2, Cu 1.0 Fe 0.5 S 0.9 (S 1.25 ); A-3, Cu1.0Fe0.3S0.9 (S0.95); and A-4, Cu1.0Fe0.1S0.6 (S0.65). The UV− vis−NIR spectra and powder diffractograms of these samples are presented in Figure 4. As the Cu/Fe ratio was increased from 1:1 through 2:1 to 3.3:1, the intensity of the plasmonic peak decreased, and its maximum was bathochromically shifted from 486 nm (A-1) to 512 nm (A-2) and 547 nm (A-3). For the nanocrystals with the highest Cu/Fe ratio (10:1), the plasmonic peak was shifted to the NIR part of the spectrum (maximum at ∼1200 nm) and resembled that reported for nonstoichiometric binary Cu1.8S nanocrystals.7−11 The diffractograms of A-1 and A-2 are characteristic of the chalcopyrite (CuFeS2) structure. In the

case of sample A-2, in addition to the dominant reflection at 29.7° (112), characteristic of the chalcopyrite structure, a peak of low intensity, ascribed to the Cu5FeS4 phase, is present at 28.2°. The diffraction pattern of A-3 resembles that reported for the regular structure of Cu5FeS4 (JCPDS 34-0135) with an intense Bragg reflection at 47.8°. This reflection is shifted to slightly lower angles (46.8°) in the diffractogram of A-4 because of the transformation of the regular structure of Cu5FeS4 into the regular structure of Cu1.8S. It should be noted that this structural transformation is accompanied by a dramatic change in the character of the plasmonic peak (see Figure 4). The above-described synthetic studies unequivocally showed that the type of sulfur precursor plays a key role in obtaining Cu−Fe−S nanocrystals exhibiting the localized surface plasmon resonance effect. For this reason, we undertook the task of performing a detailed characterization of sulfur solutions in oleylamine (S/OLA) and in 1-octadecene (S/ODE), as well as their transformations in the reaction media. Application of S/OLA solution is a popular method of inducing nucleation in the synthesis of many sulfide-type nanocrystals. For example, Panthani et al.44 prepared CuInS2 nanocrystals by injecting OLA into a reaction mixture containing sulfur. In particular, small-sized nanocrystals could be obtained with higher ratios of OLA to precursors. This was caused by a larger number of seeds with a concomitant dramatic decrease in the precursor concentration. The analysis of sulfur solutions in amines was already reported,45 but for a model system involving sulfur and octylamine rather than sulfur and technical-grade oleyamine, as is typically used for the preparation of sulfide-type nanocrystals. In Figure S6 (Supporting Information) is presented a representative mass spectrum (EI+, electron-impact) of a solution consisting of 64 mg of sulfur (2.0 mmol) dissolved in 2 mL of OLA (70%, 4.2 mmol). A clear peak at m/z = 267.2 can be noticed, corresponding to the oleylamine (C18H35NH2) molecular ion. The remaining peaks can be attributed to its fragmentation. In Figure 5, the 1H NMR spectrum of S/OLA dissolved in CDCl3 is compared with the spectra of technicalgrade oleylamine and oleylamine hydrochloride (OLA-HCl). It should be noted that the chemical shifts of the lines corresponding to the protons of the methylene group adjacent to the amine group differ in all three cases. In pure oleylamine, this signal is located at 2.67 ppm. In oleylamine hydrochloride, this line is shifted to 2.95 ppm as a result of weaker screening caused by the protonation of the amine group. The spectral F

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Scheme 1. Reagent Conversion in the Reaction Mixtures Used for the Preparation of Cu−Fe−S Nanocrystals

Figure 5. 1H NMR spectra of oleylamine (OLA), oleylamine hydrochloride (OLA-HCl), and sulfur dissolved in oleylamine (S/ OLA).

lines of S/OLA are significantly broadened, with an upfield shift of the corresponding methylene group line to 2.44 ppm, indicating stronger screening. Very similar differences in the chemical shifts are observed when secondary amines are compared with primary ones. The observed shift can be attributed to the oxidation of the amine group with concomitant reduction of sulfur. As a consequence, after deprotonation of the oxidized amine group (C18H35NH2+), a N−S bond is formed, to yield C18H35NH−S8−H (or C18H35NH−S8−). The proposed mechanism is in agreement with previous studies of sulfur solutions in primary amines45,46 and involves a redox process associated with an acid−base one (in the Brönsted sense), as shown by the equation

and small amounts of oleanitrile (Figure S8a,c, Supporting Information). In the third sample, metal precursors were additionally introduced [copper(II) oleate and iron(III) stearate]. Apart from this change, all procedures were identical to those used in the case of the second sample. After the separation of the Cu−Fe−S nanocrystals, increased amounts of oleanitrile were detected, with oleylamine essentially nonexistent (Figure S8b,c, Supporting Information). Finally, a reaction mixture was analyzed in which 1-dodecanethiol (DDT) was added, in addition to metal precursors. Didodecyl disulfide was detected after the separation of the formed nanocrystals, clear confirmation of the oxidation of DDT (see Figure S9a,b, Supporting Information). To summarize this section of the article, the obtained results unequivocally indicate that, in S/OLA solutions as well as in the reaction mixture used for the preparation of nanocrystals, oleylamine serves as a reducing agent. At room temperature, it reduces elemental sulfur, whereas in the reaction mixture at >160 °C, it additionally reduces copper(II) to copper(I), being almost entirely converted into oleanitrile. This product does not bond with the nanocrystal surface and is quantitatively retained in the supernatant after centrifugation. Additionally introduced 1-dodecanethiol (DDT) serves as a stronger reducing agent because it is converted into didodecyl disulfide, its oxidation product, whereas only traces of oleanitrile are detected in the supernatant. This finding shows that no thiol is formed in the S/OLA solution because, under the conditions of nanocrystal formation, its oxidation product (dialkyl dilsulfide) had to be observed, which was not the case. Finally, no alkenes can be detected in the recorded chromatograms, which are clearly visible in the 1H NMR spectra of the organic part of Cu−Fe−S nanocrystals, recovered after the dissolution of their inorganic cores (vide supra). This finding additionally corroborates the conclusion that the nanocrystals prepared with S/OLA precursors are stabilized by alkyl groups directly attached to surfacial sulfur atoms and that alkenes are formed during the dissolution of the inorganic core as a result of the breaking of CS bonds. As already mentioned, S/ODE solutions were used as a second type of sulfur precursors in the synthesis of Cu−Fe−S

2C18H35NH 2 + S8 → (C18H35NH3+)(C18H35NH−S8−) (1)

The GC/MS technique was used to identify individual chemicals formed in the reaction mixture during the preparation of Cu−Fe−S nanocrystals from copper(II) oleate, iron(III) stearate, and S/OLA in 1,2-dichlorobenzene. In particular, the solutions remaining after the separation of the nanocrystals were studied. In Scheme 1, the results of these analyses are summarized, whereas all corresponding chromatograms and mass spectra are collected in the Supporting Information. In the first sample, that is, the initial S/OLA solution before its injection into the reaction mixture, in addition to oleylamine (C18H35NH2), 1-octadecanethiol (C18H37SH) was detected (see Figure S7a,b, Supporting Information). Previous (NMR) investigations of S/OLA (vide supra) did not show any evidence for the presence of thiol because its concentration was too low as compared to the sensitivity of the standard NMR spectrometer or because it formed only at elevated temperature in GC injection chamber. Its identification by GC/MS indicated that, in addition to the redox process involving the amine group and sulfur, consecutive or parallel reactions must take place resulting in the formation of the detected thiol. The second sample was prepared by injecting S/OLA solution into 1,2-dichlorobenzene at 160 °C (without the presence of metal precursors). The resulting mixture was also heated to 180 °C and kept at this temperature for an additional 1 h, after which it was cooling and the same procedures as in the separation of nanocrystals were applied. The solvents were evaporated, and the residue was analyzed, showing the presence of oleylamine G

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry nanocrystals. They were first applied in the preparation of CdS nanocrystals in 2002 by Yu and Peng47 and were more recently used in the synthesis of Cu2ZnSnS4 nanocrystals of the kesterite or wurtzite structure.48 In the latter case, the authors prepared two types of S/ODE solutions, namely, yellow and orange, by mixing 128 mg (4.0 mmol) of sulfur in 10 mL (31.2 mmol) of ODE and heating the mixture at 160 °C in an argon atmosphere for 1 and 2 h, respectively. In the research described here, three types of S/ODE solutions were prepared by dissolving 640 mg (20.0 mmol) of sulfur in 25 mL (78.1 mmol) of ODE. S-ODE-1 (yellow) and S-ODE-2 (orange) were prepared in the same manner as described in the literature48 (vide supra). The red solution (SODE-3) used for the preparation of Cu−Fe−S described here was obtained by heating the mixture of S and ODE at 200 °C for 20 min. Preparation of S-ODE-1 and S-ODE-2 requires strict adherence to the heating-temperature and heating-time regimes. At 200 °C, S-ODE-3 is always obtained, in a very short time. Images of all three solutions are presented in Figure 6.

Figure 7. EI-MS spectrum of S-ODE-3 solution.

Figure 6. Yellow (S-ODE-1), orange (S-ODE-2), and red (S-ODE-3) S-ODE precursor solutions. Figure 8. 1H NMR spectra recorded in CDCl3 of: a) S-ODE-3 conversion products; b) didodec-1-yl sulfide; c) dioctadec-2-yl sulfide.

S-ODE-1 is rather unstable, and partial sulfur precipitation is always observed during storage. S-ODE-2 and S-ODE-3 are stable and can be used for the preparation of Cu−Fe−S nanocrystals even after one month of storage. Detailed GC/MS analyses were carried out for pure ODE, S-ODE-2, and S-ODE3 (see Figure S10a−c, Supporting Information). In the chromatogram of S-ODE-2, in addition to 1-octadecene, its conversion products can be detected. The contribution of these 1-octadecene conversion products is even more pronounced in the chromatogram of S-ODE-3, which was used for the preparation of Cu−Fe−S nanocrystals. In Figure 7, the mass spectrum (EI+, electron-impact) of S-ODE-3 is presented. In this spectrum, peaks corresponding to radical cations of aliphatic sulfides of the types C 18 H 37 −S−C 18 H 37 and C18H37−S-C18H36−S−C18H37 are clearly observed at m/z 538.3 and 822.6, respectively. Their presence is further corroborated by the peaks at m/z 285.1 and 313.2 attributed to the fragmentation of the above radical cations.49 The concentration of these conversion products in S-ODE-3, however, is too low to be detected by 1H NMR spectroscopy, and the corresponding spectrum shows only the presence of 1octadecene (see Figure S11, Supporting Information). Separation of the 1-octadecene conversion products on a chromatographic column (with hexane as the eluent) enabled us to record their NMR spectra (Figure 8a). Preliminary analysis of the spectrum confirmed the presence of a mixture of aliphatic sulfides, consistent with the GC/MS data. Their unequivocal identification, however, required the synthesis of appropriate

standards, namely, primary didodec-1-yl sulfide and secondary dioctadec-2-yl sulfide, whose spectra are shown in Figure 8b,c. The spectrum of the conversion products of S-ODE-3 (Figure 8a) can be considered as a superposition of the spectra of these two compounds. In particular, the triplet at 2.54 ppm is diagnostic of dioctadec-1-yl sulfide (P-1), whereas the multiplet at 2.78−2.85 ppm corresponds to dioctadec-2-yl sulfide (P-2). The P-1/P-2 ratio, as determined from the NMR spectrum, is equal to 1:5. Thus, the detailed analysis of red S-ODE-3 solution clearly indicates that alkyl sulfides are the real sulfur precursors in the reaction mixture. A comparison of the detailed analyses of the S/OLA and S/ ODE precursors is very instructive. In the first case, an ionic compound, resulting from the redox reaction between sulfur and amine group, is the real precursor (C18H35NH3+)(C18H35NH−S8−). This precursor shows high reactivity toward copper(II) oleate and iron(III) stearate, which results in its high degrees of conversion. Thioethers and dialkyl polysulfidesthe true sulfur precursors in S/ODE solutionsare less reactive, and their degrees of conversion are much smaller. For this reason, the resulting Cu−Fe−S nanocrystals show a cationenriched surface whose charge requires compensation by anionic-type (carboxylic) ligands. H

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

■

(15) Li, H.; Brescia, R.; Povia, M.; Prato, M.; Bertoni, G.; Manna, L.; Moreels, I. J. Am. Chem. Soc. 2013, 135, 12270−12278. (16) Xu, J.; Tang, Y.-B.; Chen, X.; Luan, C.-Y.; Zhang, W.-F.; Zapien, J. A.; Zhang, W.-J.; Kwong, H.-L.; Meng, X.-M.; Lee, S.-T.; Lee, C.-S. Adv. Funct. Mater. 2010, 20, 4190−4195. (17) Wang, J.-J.; Xue, D.-J.; Guo, Y.-G.; Hu, J.-S.; Wan, L.-J. J. Am. Chem. Soc. 2011, 133, 18558−18561. (18) Saldanha, P. L.; Brescia, R.; Prato, M.; Li, H.; Povia, M.; Manna, L.; Lesnyak, V. Chem. Mater. 2014, 26, 1442−1449. (19) Niezgoda, J. S.; Harrison, M. A.; McBride, J. R.; Rosenthal, S. J. Chem. Mater. 2012, 24, 3294−3298. (20) Liu, X.; Wang, X.; Swihart, M. T. Chem. Mater. 2015, 27, 1342− 1348. (21) Kovalenko, M. V.; Manna, L.; Cabot, A.; Hens, Z.; Talapin, D. V.; Kagan, C. R.; Klimov, V. I.; Rogach, A. L.; Reiss, P.; Milliron, D. J.; Guyot-Sionnest, P.; Konstantatos, G.; Parak, W. J.; Hyeon, T.; Korgel, B. A.; Murray, C. B.; Heiss, W. ACS Nano 2015, 9, 1012−1057. (22) Liang, D.; Ma, R.; Jiao, S.; Pang, G.; Feng, S. Nanoscale 2012, 4, 6265−6268. (23) Aldakov, D.; Lefrançois, A.; Reiss, P. J. Mater. Chem. C 2013, 1, 3756−3776. (24) Wiltrout, A. M.; Freymeyer, N. J.; Machani, T.; Rossi, D. P.; Plass, K. E. J. Mater. Chem. 2011, 21, 19286−19292. (25) Kumar, P.; Gusain, M.; Kumar, P. S.; Uma, S.; Nagarajan, R. RSC Adv. 2014, 4, 52633−52636. (26) Wang, Y.-H. A.; Bao, N.; Gupta, A. Solid State Sci. 2010, 12, 387−390. (27) Kumar, P.; Uma, S.; Nagarajan, R. Chem. Commun. 2013, 49, 7316−7318. (28) Mourdikoudis, S.; Liz-Marzán, L. M. Chem. Mater. 2013, 25, 1465−1476. (29) Wang, W.; Jiang, J.; Ding, T.; Wang, C.; Zuo, J.; Yang, Q. ACS Appl. Mater. Interfaces 2015, 7, 2235−2241. (30) Wadia, C.; Wu, Y.; Gul, S.; Volkman, S. K.; Guo, J.; Alivisatos, A. P. Chem. Mater. 2009, 21, 2568−2570. (31) Schliehe, C.; Juarez, B. H.; Pelletier, M.; Jander, S.; Greshnykh, D.; Nagel, M.; Meyer, A.; Foerster, S.; Kornowski, A.; Klinke, C.; Weller, H. Science 2010, 329, 550−553. (32) Shao, G.; Chen, G.; Yang, W.; Ding, T.; Zuo, J.; Yang, Q. Langmuir 2014, 30, 2863−2872. (33) Chiu, C.-Y.; Chen, C.-K.; Chang, C.-W.; Jeng, U.-S.; Tan, C.-S.; Yang, C.-W.; Chen, L.-J.; Yen, T.-J.; Huang, M. H. J. Am. Chem. Soc. 2015, 137, 2265−2275. (34) Gabka, G.; Bujak, P.; Gryszel, M.; Kotwica, K.; Pron, A. J. Phys. Chem. C 2015, 119, 9656−9664. (35) Gabka, G.; Bujak, P.; Gryszel, M.; Ostrowski, A.; Malinowska, K.; Zukowska, G. Z.; Agnese, F.; Pron, A.; Reiss, P. Chem. Commun. 2015, 51, 12985−12988. (36) Turo, M. J.; Macdonald, J. E. ACS Nano 2014, 8, 10205−10213. (37) Kuzuya, T.; Itoh, K.; Ichidate, M.; Wakamatsu, T.; Fukunaka, Y.; Sumiyama, K. Electrochim. Acta 2007, 53, 213−217. (38) Deroubaix, G.; Marcus, P. Surf. Interface Anal. 1992, 18, 39−46. (39) Tossell, J. A.; Urch, D. S.; Vaughan, D. J.; Wiech, G. J. Chem. Phys. 1982, 77, 77−82. (40) Siriwardane, R. V.; Cook, J. M. J. Colloid Interface Sci. 1985, 108, 414−422. (41) Konstantatos, G.; Levina, L.; Fischer, A.; Sargent, E. H. Nano Lett. 2008, 8, 1446−1450. (42) Virieux, H.; Le Troedec, M.; Cros-Gagneux, A.; Ojo, W.-S.; Delpech, F.; Nayral, C.; Martinez, H.; Chaudret, B. J. Am. Chem. Soc. 2012, 134, 19701−19708. (43) Gabka, G.; Bujak, P.; Giedyk, K.; Kotwica, K.; Ostrowski, A.; Malinowska, K.; Lisowski, W.; Sobczak, J. W.; Pron, A. Phys. Chem. Chem. Phys. 2014, 16, 23082−23088. (44) Panthani, M. G.; Akhavan, V.; Goodfellow, B.; Schmidtke, J. P.; Dunn, L.; Dodabalapur, A.; Barbara, P. F.; Korgel, B. A. J. Am. Chem. Soc. 2008, 130, 16770−16777. (45) Thomson, J. W.; Nagashima, K.; Macdonald, P. M.; Ozin, G. A. J. Am. Chem. Soc. 2011, 133, 5036−5041.

CONCLUSIONS We demonstrate, for the first time, that it is possible to synthesize Cu−Fe−S nanocrystals exhibiting a strong localized surface plasmon resonance (LSPR) effect, which has never been observed to date in this family of semiconductor nanoparticles. The position of the plasmonic absorption peak can be tuned from the visible to the near-infrared range by adjusting the nanocrystal composition, with the most energetic absorption being observed for nearly stoichiometric CuFeS2 nanocrystals. Extended analysis of the reaction mixture enabled us to identify the true chemical constitution of the sulfur precursors and, as a consequence, to elaborate new synthetic procedures ensuring high reproducibility of the optical properties of the nanocrystals.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b00912. Additional experimental procedures, energy-dispersive spectra, MS spectra, and GC chromatograms (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS G.G., P.B., and A.P. acknowledge financial support from National Centre of Science in Poland (NCN, Grant 2015/ 17/B/ST4/03837).

■

REFERENCES

(1) Jones, M. R.; Osberg, K. D.; Macfarlane, R. J.; Langille, M. R.; Mirkin, C. A. Chem. Rev. 2011, 111, 3736−3827. (2) Chinen, A. B.; Guan, C. M.; Ferrer, J. R.; Barnaby, S. N.; Merkel, T. J.; Mirkin, C. A. Chem. Rev. 2015, 115, 10530−10574. (3) Guerrini, L.; Graham, D. Chem. Soc. Rev. 2012, 41, 7085−7107. (4) Luther, J. M.; Jain, P. K.; Ewers, T.; Alivisatos, A. P. Nat. Mater. 2011, 10, 361−366. (5) Mattox, T. M.; Ye, X.; Manthiram, K.; Schuck, P. J.; Alivisatos, A. P.; Urban, J. J. Adv. Mater. 2015, 27, 5830−5837. (6) Comin, A.; Manna, L. Chem. Soc. Rev. 2014, 43, 3957−3975. (7) Hsu, S.-W.; On, K.; Tao, A. R. J. Am. Chem. Soc. 2011, 133, 19072−19075. (8) Hsu, S.-W.; Bryks, W.; Tao, A. R. Chem. Mater. 2012, 24, 3765− 3771. (9) Kruszynska, M.; Borchert, H.; Bachmatiuk, A.; Rümmeli, M. H.; Büchner, B.; Parisi, J.; Kolny-Olesiak, J. ACS Nano 2012, 6, 5889− 5896. (10) Liu, X.; Wang, X.; Zhou, B.; Law, W.-C.; Cartwright, A. N.; Swihart, M. T. Adv. Funct. Mater. 2013, 23, 1256−1264. (11) Liu, L.; Zhong, H.; Bai, Z.; Zhang, T.; Fu, W.; Shi, L.; Xie, H.; Deng, L.; Zou, B. Chem. Mater. 2013, 25, 4828−4834. (12) Dorfs, D.; Härtling, T.; Miszta, K.; Bigall, N. C.; Kim, M. R.; Genovese, A.; Falqui, A.; Povia, M.; Manna, L. J. Am. Chem. Soc. 2011, 133, 11175−11180. (13) Yang, H.-J.; Chen, C.-Y.; Yuan, F.-W.; Tuan, H.-Y. J. Phys. Chem. C 2013, 117, 21955−21964. (14) Li, W.; Zamani, R.; Rivera Gil, P.; Pelaz, B.; Ibáñez, M.; Cadavid, D.; Shavel, A.; Alvarez-Puebla, R. A.; Parak, W. J.; Arbiol, J.; Cabot, A. J. Am. Chem. Soc. 2013, 135, 7098−7101. I

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry (46) Davis, R. E.; Nakshbendi, H. F. J. Am. Chem. Soc. 1962, 84, 2085−2090. (47) Yu, W. W.; Peng, X. Angew. Chem., Int. Ed. 2002, 41, 2368− 2371. (48) Zou, Y.; Su, X.; Jiang, J. J. Am. Chem. Soc. 2013, 135, 18377− 18384. (49) Pretsch, E.; Bü h lmann, P.; Badertscher, M. Structure Determination of Organic Compounds; Springer-Verlag: Berlin, 2009.

J

DOI: 10.1021/acs.inorgchem.6b00912 Inorg. Chem. XXXX, XXX, XXX−XXX