Article pubs.acs.org/ac
The Chemical Environments of Oleate Species within Samples of Oleate-Coated PbS Quantum Dots Laura C. Cass, Michał Malicki, and Emily A. Weiss* Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208-3113, United States S Supporting Information *
ABSTRACT: A combination of FT-IR, 1H NMR, nuclear Overhauser effect (NOESY), and diffusion-ordered (DOSY) NMR spectroscopies shows that samples of oleate-coated PbS quantum dots (QDs) with core radii ranging from 1.6 to 2.4 nm, and purified by washing with acetone, contain two species of oleate characterized by the stretching frequencies of their carboxylate groups, the chemical shifts of their protons, and their diffusion coefficients. One of these oleate species exists primarily on the surfaces of the QDs and either chelates a Pb2+ ion or bridges two Pb2+ ions. The ratio of bridging oleates to chelating oleates on the surfaces of the QDs is approximately 1:1 for all sizes of the QDs we studied. The second oleate species in these samples bridges two Pb2+ ions within clusters or oligomers of lead oleate (with a hydrodynamic radius of ∼1.4 nm), which are byproducts of the QD synthesis. The concentration of these clusters increases with increasing size of the QDs because larger QDs are produced by increasing the concentration of the oleic acid ligand in the reaction mixture. The oleate molecules on the surfaces of the QDs and within the lead oleate clusters are in rapid exchange with each other. Additional washes with methanol progressively eliminate the contaminating clusters from the PbS QD samples. This work quantitatively characterizes the distribution of binding geometries at the inorganic/organic interface of the nanocrystals and demonstrates the utility of using organic ligands as probes for the composition of a colloidal QD sample as a function of the preparation procedure.
T
ties, such as the available surface area and interaction potentials for adsorption of molecular redox partners6,7 or other QDs (leading to QD aggregation) and optimal conditions for ligand exchange. It is, therefore, difficult to understand, for example, the dynamics of electron trapping in QDs or the formation of self-assembled QD lattices without quantitatively characterizing the degree and type of heterogeneity in this interfacial chemistry, present both within a single QD and within the ensemble of QDs. Two of the most powerful techniques for directly interrogating chemical bonding between organic molecules and the ions on the surface are Fourier transform infrared (FT-IR) spectroscopy, which identifies and quantifies the relative contributions of vibrational signatures of functional groups in different binding modes to ions, and nuclear magnetic resonance (NMR) spectroscopy, which probes the chemical environments of surface ligands. These techniques have been employed previously to construct binding isotherms for QDligand systems,8−11 to monitor the progress of ligand exchanges,12,13 to chemically identify a ligand in a sample of QDs,14,15 and to determine the number and types of ligand-
his paper describes the characterization of samples of colloidal PbS quantum dots (QDs), produced and purified with the most common methods,1 by the binding conformations of oleate, the sole surfactant for the QDs within the sample. Lead chalcogenide QDs exhibit (i) a tunable absorption spectrum between 800 and 1800 nm, (ii) high molar extinction coefficients (on the order of 105−106 M−1 cm−1), and (iii) high photoluminescence quantum efficiencies in organic and aqueous media.1−3 These properties make PbS QDs an attractive material for photovoltaic devices and biological sensors.1,2 Ensembles of colloidal QDs include a set of well-established structural and chemical inhomogeneities, including polydispersity in the size of their inorganic cores and distributions in the number and chemical structure of surface ligands, surface defects, and lattice defects. Routine characterization of QDs post-purification includes transmission electron microscopy and steady-state absorption and emission spectroscopy, in order to estimate the average size and shape (and corresponding dispersities) of the particles and their general level of surface passivation. The details of the interfacial chemistry of these particles (their least well-defined and least synthetically reproducible feature), however, determine important aspects of their electronic structure, such as the density of states available for nonradiative pathways of excitonic decay4,5 and the range of redox potentials of surface ions. Surface chemistry also determines important physical proper© 2013 American Chemical Society
Received: May 31, 2013 Accepted: June 20, 2013 Published: June 20, 2013 6974
dx.doi.org/10.1021/ac401623a | Anal. Chem. 2013, 85, 6974−6979
Analytical Chemistry
Article
Figure 1. FT-IR and 1H NMR spectra of PbS QDs of a series of sizes (radii shown in the legend) after purification with (A and B) acetone (C and D) acetone and an additional wash with methanol, and then dissolved in C6D6. Each NMR spectrum is multiplied by a constant such that the intensity of the oleate vinyl peak at 5.54 ppm is the same in every spectrum in B and such that the intensity of the oleate vinyl peak at 5.72 ppm is the same in every spectrum in D (in order to most clearly present the ratio of the two peaks). Each FT-IR trace is normalized to the intensity of the aliphatic stretch at 1468 cm−1 and displaced vertically for clarity. For all FT-IR spectra, we inserted a second break in the x axis between 1445 and 1458 cm−1, due to strong absorptions by C6D6 in this region. (E) Molar ratios of oleate molecules in a bridging bidentate structure and in a chelating bidentate structure, calculated from areas of corresponding peaks in the FT-IR spectra in Figure 1 (panels A and C) of QD samples washed with both methods. (F) Molar ratios of species A (5.54 ppm) to species B (5.72 ppm), calculated from areas of corresponding peaks in the NMR spectra in B and D, of QD samples washed with both methods. The error bars represent the standard deviation of the ratio found from three methods of integration of NMR signals (see the Supporting Information).
binding modes on the surface of a QD.16−18 Here, by comparing FT-IR, 1H NMR, nuclear Overhauser effect (NOESY) NMR, and diffusion-ordered (DOSY) NMR spectra of identical liquid-phase samples of PbS QDs, where the surfactant, oleate, serves as the structural probe, we obtain a detailed and quantitative picture of the inorganic−organic interface of a QD-ligand system in solution. We detect impurities that do not contribute in a straightforward way to optical spectra of the sample, and relate the properties of the samples to their synthesis and purification conditions. Our measurements reveal that within dispersions of PbS QDs purified by washing with acetone there are two oleate species present: one species that spends the majority of its time adsorbed to the QD surface by chelating to and bridging Pb2+ ions and one that spends the majority of its time bridging Pb2+ ions within lead oleate clusters (or oligomers) that coexist with
QDs in the sample. The two species of oleate molecules are in fast exchange with respect to the NMR timescale. The partitioning of oleate between the QD surfaces and the lead oleate clusters depends on the amount of oleic acid used in the synthesis, an important result given that varying the concentration of oleic acid is the preferred method for tuning the size of PbS and CdS QDs.1,15,19−21 One to two additional washes with methanol as the nonsolvent eliminates lead oleate clusters in the samples. Importantly, our study reveals that oleate ligands on the surfaces of PbS QDs of a range of radii have two approximately equally contributing adsorption modes to Pb2+, chelating bidentate and bridging bidentate, and that size-dependent heterogeneity in the oleate structure is a result of “contamination” of the PbS QD sample with lead oleate clusters or oligomers that can be removed or prevented. 6975
dx.doi.org/10.1021/ac401623a | Anal. Chem. 2013, 85, 6974−6979
Analytical Chemistry
Article
Scheme 1. Binding Structures of Deprotonated Carboxylates with Metal Ions (M)
■
EXPERIMENTAL SECTION We synthesized QDs of five different radii (R), between 1.6 and 2.4 nm, by varying the concentration of oleic acid in the reaction mixture within a method reported by Hines and Scholes; a larger concentration of oleic acid leads to larger QDs.1 We added different amounts of oleic acid (OA), 1.5, 3.0, 6.0, 12, and 20 mL, to a volume of 1-octadecene (ODE) that resulted in a total volume of the reaction mixture of 20 mL (there is no octadecene in the reaction mixture with 20 mL of oleic acid), and we deaerated the reaction mixture in a 50 mL three-neck round-bottom flask at room temperature by bubbling with nitrogen for 30 min. The addition of PbO (0.36 g) followed by heating to 150 °C with stirring under a N2 flow for 60 min produced a clear and colorless solution that was then cooled to 115 °C. Injection of 0.17 mL of hexamethyldisilathiane dissolved in 8 mL of ODE caused the solution to change from orange to brown within 3 s. After 10 min, we cooled the reaction to room temperature in an ice bath. Figure S7 of the Supporting Information shows the absorption spectra of the set of QD samples we used in this work. The lowest-energy excitonic peaks range from 940 nm for the QDs with R = 1.6 to 1373 nm for the QDs with R = 2.4 nm. We purified the QDs, using one of two methods, in order to determine the effect of this purification procedure on the final sample composition. In method 1, we added 80 mL of acetone to the reaction mixture and centrifuged the sample at 3500 rpm for 5 min to produce a dark brown pellet, which we redispersed in 10 mL of hexanes. We then added 10 mL acetone, centrifuged the sample at 3500 rpm for 5 min, collected the dark solid, washed it with acetone, and redispersed it in deuterated benzene (C6D6) (i.e., the solvent in which we performed the NMR and FTIR characterization). In method 2, we performed method 1 and subsequently added 10 mL of methanol to the purified product, centrifuged the sample at 3500 rpm for 5 min, and redispersed it in C6D6 for analysis. The Supporting Information contains details of the FT-IR and NMR experimental procedures.
less than 0.8% of oleate molecules in the samples are protonated. See Figure S6 (panels A and D) of the Supporting Information for details. The binding modes of deprotonated carboxylate molecules in the presence of metal cations are well-established (see Scheme 1). These structures include (i) a chelating bidentate, in which the charge of the carboxylate is shared across the oxygen atoms, which both form covalent bonds with a metal cation, (ii) a bridging bidentate, in which each oxygen of the carboxylate covalently binds to a different metal cation, (iii) an ionic, in which one negatively charged oxygen of the carboxylate and a metal cation form an ionic bond, and (iv) a covalent monodentate, in which the negatively charged oxygen of the carboxylate forms a covalent bond with a metal ion.22−24 In the FT-IR spectra of the oleate-coated QDs washed with both methods, there are three clear peaks, at 1553, 1525, and 1403, that do not correspond to aliphatic stretches and, based on studies of metal carboxylates,23−25 correspond to carboxylate (O−C−O) stretching modes. Previous studies show that asymmetric O−C−O stretches have energies of 1540−1660 cm−1 and symmetric O−C−O stretches have energies of 1360−1440 cm−1,22,24−28 so we can immediately assign the 1553 cm−1 and 1525 cm−1 peaks to asymmetric O−C−O stretches of two different oleate species, that is, oleate molecules with different binding structures (Scheme 1), and the 1403 cm−1 peak to a symmetric O−C−O stretch of an oleate species. We see below that this peak probably encompasses two overlapping symmetric peaks. We note that free, protonated oleic acid does not have a carboxylate mode, so it is not responsible for any of the signals we discuss in this section. In order to aid in our assignment of these peaks to one of the oleate species in Scheme 1, we acquired a set of FT-IR spectra of molecular lead oleate titrated with excess oleic acid, in which the molar ratio Pb2+:ligand ranges from 0:1 to 0.38:1 (see Figure S1 of the Supporting Information). We do not achieve a ratio Pb2+:ligand of 0.5:1, even with pure lead oleate, because it has an oleic acid impurity that we account for in interpreting the spectra. The only carboxylate peaks that are present in the spectrum of the sample with the lowest ratio of Pb2+ to ligand (Pb2+:ligand = 0.1:1) are those at 1404 and 1524 cm−1. We therefore conclude that the 1404 cm−1 and 1524 cm−1 peaks are a symmetric/asymmetric pair for the same oleate species. As the ratio Pb2+:ligand increases, a peak at 1401 cm−1, which grows on top of the peak at 1404 cm−1, and a peak at 1549 cm−1 both appear and increase in intensity. As the 1401 cm−1 and 1549 cm−1 peaks grow, the area of the 1524 cm−1 peak does not change, so the 1401 cm−1 peak is correlated exclusively with the 1549 cm−1 peak. We therefore assign the 1401 cm−1 and 1549 cm−1 peaks as a symmetric/asymmetric pair for a second oleate species. The splitting in energy between the symmetric and asymmetric O−C−O stretching features (Δ) of a given oleate
■
RESULTS AND DISCUSSION Figure 1 (panels A and B) show the carbonyl and carboxylate stretching region of the FT-IR spectra and the vinyl proton region of the 1H NMR spectra, respectively, of the PbS QD samples in C6D6, where the QDs were washed with acetone only (method 1). Figure 1 (panels C and D) show these spectra from the samples washed with acetone plus methanol (method 2) (see the complete spectra in Figure S8 of the Supporting Information). There are no signals in these spectra from the sulfur precursor, hexamethyldisilathiane sulfide, and only minor signals (indicating
