Subpicosecond Photoinduced Hole Transfer from a CdS Quantum Dot

Jun 9, 2016 - This paper describes the enhancement of the rate of hole transfer from a photoexcited CdS quantum dot (QD), with radius R = 2.0 nm, to a...
0 downloads 0 Views 3MB Size
Subpicosecond Photoinduced Hole Transfer from a CdS Quantum Dot to a Molecular Acceptor Bound Through an ExcitonDelocalizing Ligand Shichen Lian, David J. Weinberg, Rachel D. Harris, Mohamad S. Kodaimati, and Emily A. Weiss* Department of Chemistry, Northwestern University, 2145 Sheridan Rd., Evanston, Illinois 60208-3113, United States S Supporting Information *

ABSTRACT: This paper describes the enhancement of the rate of hole transfer from a photoexcited CdS quantum dot (QD), with radius R = 2.0 nm, to a molecular acceptor, phenothiazine (PTZ), by linking the donor and acceptor through a phenyldithiocarbamate (PTC) linker, which is known to lower the confinement energy of the excitonic hole. Upon adsorption of PTC, the bandgap of the QD decreases due to delocalization of the exciton, primarily the excitonic hole, into interfacial states of mixed QD/PTC character. This delocalization enables hole transfer from the QD to PTZ in 97%.

Table 1. Number of Bound PTC-PTZ or BA-PTZ Ligands per QD Determined by NMR counting method (ii)a

counting method (i) sample CdS-OA CdS-PTCPTZ CdS-BAPTZ

total OA per QDb

displaced OA perQDc

bound PTZ ligands per QD (i)d

added phenyl rings per QDe

190 ± 3 190 ± 3 190 ± 3 190 ± 3 190 ± 3 190 ± 3 190 ± 3

0 94 ± 1 190 ± 3 190 ± 3 48 ± 1 57 ± 1 75 ± 1

0 47 ± 1 48 ± 1 57 ± 1 75 ± 1

0 110 ± 10 210 ± 30 300 ± 40 55 80 150

measured phenyl rings per QD

average

bound PTZ ligands per QD (i) = added − measured

0 0 60 ± 7 50 ± 10 100 ± 10 110 ± 10 170 ± 20 130 ± 40 only counted using method (i) because BA-PTZ is in fast exchange between free and bound

bound PTZ ligands per QD (λ)f 0 49 ± 6 110 ± 10 130 ± 40 48 ± 1 57 ± 1 75 ± 1

a Error bars for all quantities obtained with this method are calculated by the NMR calibration experiment described in Figure S5. bMeasured by integration of the 1H NMR signal from 5.9 to 5.4 ppm and averaged over all 7 listed samples; the error bar is the standard deviation of these 7 measurements. cError bars are propagated from the uncertainties in total OA per QD. dEquals [(displaced OA per QD)/2] for PTC-PTZ or (displaced OA per QD) for BA-PTZ. eFor BA-PTZ, equals the concentration of BA-PTZ measured by mass. For PTC-PTZ, equals the measured by integration of the 1H NMR signal from free PTC-PTZ molecules before addition of QDs. This measurement is necessary because some PTC-PTZ molecules degrade to PTZ-Ph-NH2 before addition of QDs, so the molecular weight of the mixture is not known a priori. fOnly the result from method (i) is used for the two highest coverages of PTC-PTZ because all the OAs have been displaced.

6374

DOI: 10.1021/acsnano.6b02814 ACS Nano 2016, 10, 6372−6382

Article

ACS Nano The complete ligand counting data and the portions of the NMR spectra used to obtain these data are in Table 1 and Figure 3, respectively. To estimate the number of bound PTC-PTZ per QD using method (i), we integrated the NMR signals of freely diffusing oleic acid (OA) (δ = 5.5 ppm, sharp as well as the signals from bound oleate (δ = 5.7 ppm, broad); these two signals are well-resolved spectrally in benzene-d646,47 (Figure 3A). The increase in free OA and the decrease in bound OA signal indicates that addition of PTC-PTZ displaces oleate ligands from the QD surface. Integration of the free oleate signal before and after addition of, for example, 110 equiv PTC-PTZ per QD yields 94 displaced oleate molecules per QD (Figure 3A, red). Given our recent result that another PTC derivative (4-hexylphenyldithiocarbamate) displaces two oleate ligands (in the form of Cd(OA)2) per bound PTC on CdS QDs,43 we estimate through method (i) that the average number of bound PTC-PTZ molecules per QD, λ, is 47 in this sample. To estimate λ for PTC-PTZ by method (ii), we compared the aromatic proton intensity of the same amount of added PTCPTZ with and without QDs. When performing this procedure, we must account for the fact that, in the absence of QDs, PTCPTZ degrades exclusively to its aniline derivative (PTZ-PhNH2), Chart 1 and Figure 3B (top), but in the presence of QDs, unbound PTC-PTZ also forms phenothiazine-phenyl-thiourea (PTZ-Ph-TU), Chart 1 and Figure 3B (bottom); see Figure S4 for MALDI data. These degradation products do not bind, or bind only weakly, to the QD and do not displace bound PTCPTZ.43 All of the aromatic signals in mixtures of PTC-PTZ with QDs correspond to freely diffusing or weakly binding degradation products of PTC-PTZ; the signals from bound PTC-PTZ are not detectable. We then calculate the number of bound PTC-PTZ per QD, λ, using eq 1: λ=

18 (IA − IB) × MIS a IIS × MQD

(1)

where IA and IB are the integrated 1H NMR intensities of all of the aromatic proton signals in a sample of PTC-PTZ alone and in a mixture of PTC-PTZ with QDs, respectively. IIS is the intensity of the signal from an internal standard hexamethylcyclotrisiloxane. MQD and MIS are the number of moles of QDs and the number of moles of internal standard, respectively, in the sample. The prefactor “18/a” accounts for the fact that the internal standard molecule has 18 protons, and the number of aromatic protons in the integrated region is a. Using this method, we measured that the number of bound PTC-PTZ per QD for the sample with an NMR spectrum shown in Figure 3B (with 110 equiv of the ligand added) is 50. The average of our estimates of λ from these two methods (47 and 50) is 49 PTC-PTZ ligands bound per QD upon addition of 110 equiv of the ligand. We performed this analysis for three samples with different equivalents of PTC-PTZ added, Table 1. All the native OA ligands have been displaced upon addition of PTC-PTZ molecules for the two highest coverages of PTC-PTZ, so λ is estimated only through method (ii) in these cases. Upon addition of QDs to BA-PTZ, which has a carboxylate binding group (like the native oleate) rather than dithiocarbamate binding group, the signals of all the aromatic protons of BAPTZ are broadened but distinguishable from the baseline noise, which indicates that BA-PTZ is in dynamic exchange on and off the QD surface (Figure 3C). Since there are not distinct populations of bound and free BA-PTZ on the NMR time scale, we found λ for the BA-PTZ/QD mixtures exclusively by method

Figure 3. (A) Vinyl proton regions of the 1H NMR spectra of oleatecoated CdS QDs (black) and those QDs after treatment with 110 equiv of PTC-PTZ (red) and after treatment with 55 equiv of BAPTZ (blue). (B) Top: Aromatic region of the spectrum of PTC-PTZ (2H, δ = 6.92 ppm, and 2H, δ = 6.86 ppm, both doublets), which has partially degraded into its aniline derivative PTZ-Ph-NH2 (2H, δ = 6.95 ppm, and 2H, δ = 6.82 ppm, both doublets). Bottom: NMR spectrum of CdS QDs treated with 110 equiv of PTC-PTZ. All of the protons corresponding to undegraded PTC-PTZ disappear and a new set of peaks, corresponding to PTZ-Ph-TU (4H, δ = 6.96, doublet), appear. (C) Top: Aromatic region of the NMR spectrum of BA-PTZ. Bottom: NMR spectrum of CdS QDs treated with 55 equiv of BA-PTZ. All spectra are scaled by the signal at 0.5 ppm from an internal standard, hexamethylcyclotrisiloxane. See Chart 1 for molecular structures and the Supporting Information for specific proton assignments.

(i): integrating the proton signals of displaced native oleate species. For example, for the sample with the NMR spectrum 6375

DOI: 10.1021/acsnano.6b02814 ACS Nano 2016, 10, 6372−6382

Article

ACS Nano

Figure 4. (A) Left: TA spectra of 10 μM QDs with 48 ± 1 bound BA-PTZ per QD, at various delay times after photoexcitation of the QDs at 420 nm. Right: Kinetic traces extracted from TA spectra of BA-PTZ-coated CdS QDs with various coverages at 520 nm, the peak of the PTZ+• feature, which is marked by a dashed line in the spectrum. (B) Left: TA spectra of 10 μM QDs with 49 ± 6 bound PTC-PTZ per QD, at various delay times after photoexcitation of the QDs at 420 nm. Right: Kinetic traces extracted from TA spectra of PTC-PTZ-coated CdS QDs with various coverages at 520 nm, the peak of the PTZ+• feature. Inset: TA spectra of QDs with 49, 110, and 130 bound PTC-PTZ averaged from 300 to 600 fs. The contribution of the “shelf” feature (a broad band of photoinduced absorptions of the QD) has already been removed from these spectra and kinetics by the deconvolution procedure described in the text.

shown in Figure 3A (blue), 48 oleate ligands are displaced upon the addition of 55 equiv of BA-PTZ. We assume that each displaced oleate corresponds to one bound BA-PTZ because (i) they have the same binding group; (ii) fewer oleate are displaced than BA-PTZ added even though all of the BA-PTZ molecules are sampling the surface; and (iii) it is documented in the literature that oleate desorbs as OA in the presence of excess carboxylic acid upon a proton exchange.47,48 We performed this analysis for three samples with different equivalents of added BAPTZ (Table 1). Observation of Subpicosecond Photoinduced Hole Transfer in QD-PTC-PTZ Assemblies. Adsorption of ∼50 BAPTZ or PTC-PTZ ligands per QD quenches the photoluminescence (PL) of the QDs by 98% and 97%, respectively (Figure 2, inset). Addition of the same number of equivalents of PTC or BA ligands, with no attached PTZ to the QDs, does not quench the PL of QDs (Figure S6). Addition of 1000 equiv of PhPTZ, with no BA or PTC linker, quenches the PL of the QDs, but only by 70%. It is therefore the presence of PTZ, especially when bound through the BA or PTC linkers, that is primarily responsible for quenching the PL of the QDs. Both energy transfer and photoinduced electron transfer from the lowest

excitonic state of the QD to PTZ are energetically uphill, but, as shown in Figure 1C, photoinduced hole transfer from the QD to PTZ has a driving force of ∼ −1.2 eV, so we tentatively conclude that PTZ quenches the PL through hole transfer, as has been observed for CdSe QDs,21 CdS nanorods,20 and CsPbBr3 perovskite QDs.49 Our assignment of the PL quenching to hole transfer from the QD to PTZ is confirmed by the appearance of the characteristic absorption peak of the radical cation of PTZ (PTZ+•) in the transient absorption (TA) spectra of the complexes. The TA spectra of CdS-PTC-PTZ and CdS-BA-PTZ, upon photoexciting the QDs at 420 nm, have two overlapping features in the visible region (see the Supporting Information, Figures S7 and S9): (i) the absorption of PTZ+•, which peaks at 520 nm, and (ii) a broad photoinduced absorption feature (denoted the “shelf”) from ∼480 nm to ∼700 nm, which is mainly composed of the transitions of valence band holes and trapped holes of the QD.50 We are primarily interested in the dynamics of formation of the PTZ+• feature, because these dynamics are specific to hole transfer from the QD to PTZ. Conveniently, only the amplitude of the QD’s shelf feature, and not its shape, evolves in time, as has been reported previously for CdS QDs,11 CdSe QDs,21,51 and 6376

DOI: 10.1021/acsnano.6b02814 ACS Nano 2016, 10, 6372−6382

Article

ACS Nano

Table 2. Time Constants (τi) and Amplitudes (Ai) Extracted from Fits to the Deconvoluted Kinetics Traces Corresponding to PTZ+•a sample

average no. PTZ Bound per QD (λ)

ΔR (nm)

CdSPTCPTZ

49 110 130 not measured (1000 equiv added)e

0.10 0.15 0.19 0

CdSNH2PhPTZ CdS-BAPTZ

48 57 75

0 0 0

fraction of hT that occurs within IRFb 23% 35% 43% 0

0 0 0

τ1, ps (A1)c