Article pubs.acs.org/Langmuir
Linker-Assisted Attachment of CdSe Quantum Dots to TiO2: Timeand Concentration-Dependent Adsorption, Agglomeration, and Sensitized Photocurrent Meghan E. Kern and David F. Watson* Department of Chemistry, University at Buffalo, The State University of New York, Buffalo, New York 14260-3000, United States S Supporting Information *
ABSTRACT: We have characterized the concentration and time dependences of the attachment of colloidal CdSe quantum dots (QDs) to 16-mercaptohexadanoic acid (MHDA)-functionalized nanocrystalline TiO2 thin films. The amount of QDs and the extent of their agglomeration on MHDA-functionalized TiO2 films were characterized by transmission- and reflectance-mode UV/vis absorption spectroscopy and scanning electron microscopy. Optically transparent films with spatially homogeneous coloration and minimal agglomeration of QDs were prepared from 2.2 and 5.0 μM toluene dispersions of QDs at short reaction times (98%) and sodium hydroxide (98.0%); (5) Solaronix: cis-bis(4,4′-dicarboxy-2,2′-bipyridine)dithiocyanatoruthenium(II) (N3 dye). Fluorine-doped tin oxide (FTO)-coated glass (12−14 Ω/square) was obtained from Pilkington. Methanol (MeOH), ethanol, acetonitrile, tetrahydrofuran (THF), toluene, and nitric acid were obtained from various sources and used as received. Synthesis of CdSe QDs. Nominally TOPO-capped CdSe QDs were synthesized from tri-n-octylphosphine selenide and cadmium acetate following the method of Peng and co-workers27,28 and as described previously.23 As-synthesized CdSe QDs were purified to remove excess TOPO and/or other capping ligands by dispersing the QDs into toluene and then flocculating with MeOH. After three such purification cycles, QDs were dispersed into toluene. The resulting dispersions of CdSe QDs exhibited broad first excitonic absorption bands with maxima at 528 nm (Figures S1 and S2 in Supporting Information), which did not shift as a function of concentration. On the basis of the sizing curves reported by Yu et al.,29 the average diameter of the QDs was approximately 2.7 nm. Synthesis of Nanocrystalline TiO2 Films. Nanocrystalline TiO2 films on glass substrates, which were used for concentration- and timedependent adsorption experiments, were synthesized by hydrolysis of titanium(IV) tetraisopropoxide as described previously.30,31 Projected surface areas of films were approximately 4 cm2. Our prior characterization revealed that films were 4.1 ± 0.9 μm thick and consisted of anatase TiO2 particles with average diameters of 36 ± 6 nm.30,31 Adsorption of MHDA to Nanocrystalline TiO2 Films. TiO2 films were immersed into THF solutions of MHDA (2 mM, 100 mL). (We previously reported that MHDA was more persistent on TiO2 when adsorbed from THF than when adsorbed from other solvents.23) TiO2 films were soaked at room temperature (22 °C) for approximately 8 h, then removed, and rinsed by immersing for 2−5 13294
dx.doi.org/10.1021/la503211k | Langmuir 2014, 30, 13293−13300
Langmuir
Article
measurements, samples were placed at the exit port, and all other ports were covered with a reflective material. This configuration was used to determine the absorptance (α) (one minus transmittance, or the fraction of photons absorbed) of samples as follows:
α=1−
IS I0
(1)
where IS and I0 are the measured irradiances in the presence and absence of a given sample at the exit port of the integrating sphere. Scanning electron microscopy (SEM) was performed with a Hitachi SU70 field emission instrument with an Oxford Inca SDD EDS detector. Photoelectrochemistry. Two-electrode, single-compartment photoelectrochemical cells were constructed and characterized as described previously.16 Short-circuit photocurrent action spectra were acquired for electrochemical cells with a QD-functionalized TiO2-onFTO working electrode and a PbS counter electrode, which was prepared following the method of Tachan et al.35 A hollow Teflon spacer separating the working and counter electrodes was filled with 0.5 mL of an aqueous polysulfide electrolyte,35 consisting of sodium sulfide (1 M), sulfur (0.1 M), and sodium hydroxide (0.1 M). Shortcircuit photocurrent action spectra were also acquired for dyesensitized solar cells consisting of N3-functionalized TiO2-on-FTO working electrodes and Pt mesh counter electrodes. The electrolyte for these cells consisted of iodine (0.05 M), lithium iodide (0.1 M), PMII (0.6 M), 4tBP (0.5 M), and guanidinium thiocyanate (0.1 M) in acetonitrile.36,37
Figure 1. Digital photographs of QD-functionalized TiO2 films on glass substrates prepared by immersing MHDA-functionalized TiO2 films into toluene dispersions of CdSe QDs of different concentrations for 2 h (a) or 22 h (b). Concentrations of dispersed QDs are indicated on each photograph. Two films were prepared from each concentration.
■
RESULTS AND DISCUSSION Concentration Dependence of the Linker-Assisted Attachment of QDs to TiO2. We hypothesized that the amount of QDs on linker-functionalized TiO2 and the extent of their agglomeration would depend on the adsorption time and the concentration of QDs. All of our experiments involved the attachment of nominally TOPO-capped CdSe QDs to TiO2 films functionalized with MHDA at saturation surface coverage, corresponding to average ΓMHDA of (1.4 ± 0.1) × 10−7 mol cm−2.23 We first investigated the concentration dependence of linker-assisted assembly and agglomeration. MHDA-functionalized TiO2 films were immersed into toluene dispersions of CdSe QDs of varying concentrations for 2 or 22 h. QDs adsorbed uniformly to MHDA-functionalized TiO2 films immersed in dispersions of QDs for 2 h regardless of the concentration of the QDs (Figure 1a). Alternatively, when films were immersed into dispersions of QDs for 22 h, the coverage and spatial homogeneity of QDs on TiO2 varied with concentration (Figure 1b). At low concentrations of QDs, coverages of QDs were relatively low, but films were transparent and colored uniformly. As the concentration of dispersed QDs increased, both the uniformity and transparency of the QD-coated TiO2 films decreased. At high concentrations, QDs also physisorbed to the glass substrate that was not coated with TiO2. QD-functionalized TiO2 films were characterized by UV/vis absorption spectra acquired in transmission and reflectance modes. The first excitonic absorption band of CdSe was unshifted upon immobilization onto TiO2 (Figure 2). In transmission-mode spectra of the films prepared from exposure to QDs for 22 h, the absorbance within the first excitonic absorption band of the CdSe QDs (490−575 nm) initially increased with the concentration of dispersed QDs, indicating an increase of the amount of QDs per projected surface area of TiO2 (ΓQD), and then saturated for concentrations of QDs greater than 14 μM (Figure 2b). However, the baselines of the absorption spectra became increasingly sloped at higher
concentrations of dispersed QDs and higher ΓQD (Figure 2b,c). These sloped baselines, with greater absorbance at shorter wavelengths, reveal that films with relatively high ΓQD scattered more light than films with relatively low ΓQD, consistent with their increased opacity (Figure 1b) and indicative of substantial agglomeration of QDs. We have previously described such films as being of low quality.23 When MHDA-functionalized TiO2 films were immersed for only 2 h into toluene dispersions of CdSe QDs, the QDs adsorbed at lower ΓQD but with less agglomeration, as evidenced by lower absorbances within the first excitonic absorption band of CdSe and baselines with slopes that were less severe and essentially independent of the concentration of QDs (Figure 2a,c). The absorption spectra are consistent with the lower but more uniform loading of QDs on TiO2 and the greater transparency of QD-functionalized films prepared at the shorter adsorption time (Figure 1a). We have previously described relatively transparent and spatially homogeneous QD-functionalized films that scatter light minimally as being of high quality.23 Diffuse-reflectance UV/vis spectra were acquired to minimize the contribution of light scattering to the baselines of absorption spectra. Diffuse-reflectance spectra of QD-functionalized films corresponded closely to those of dispersed QDs and did not exhibit interference fringes or sloping baselines (Figure 3a,b). Because the path length in diffuse-reflectance measurements was unknown, we report y-axes of diffusereflectance UV/vis spectra as apparent absorption (log(1/ R′∞)), where R′∞ is the intensity of scattered light divided by the intensity of incident light.38 The apparent absorption at 528 nm (log(1/R′∞)528), within the first excitonic absorption band of CdSe, provides a relative indication of the amount of QDs on TiO2. For both adsorption times (2 and 22 h), log(1/ R′∞)528 increased with the concentration of QDs and saturated 13295
dx.doi.org/10.1021/la503211k | Langmuir 2014, 30, 13293−13300
Langmuir
Article
Figure 2. (a, b) Transmission-mode UV/vis absorption spectra of QD-functionalized TiO2 films prepared by immersing MHDA-functionalized TiO2 films into toluene dispersions of CdSe QDs at varying concentrations for 2 h (a) or 22 h (b). Concentrations of dispersed QDs (units = μM) are listed underneath each spectrum. Absorption spectra are offset; dashed horizontal lines from 600 to 675 nm correspond to absorbance of 0 for each spectrum. Error bars represent ±1 standard deviation relative to the average absorbance of two films at 528 nm. (c) Absolute value of slope of the baseline of absorbance spectra (625−765 nm) of QD-functionalized TiO2 films as a function of the concentration of QDs in dispersions from which they were adsorbed for 2 or 22 h. Error bars represent ±1 standard deviation relative to the average slope of two films.
Figure 3. (a, b) Reflectance-mode UV/vis absorption spectra of QD-functionalized TiO2 films prepared by immersing MHDA-functionalized TiO2 films into toluene dispersions of CdSe QDs at varying concentrations for 2 h (a) or 22 h (b). Concentrations of dispersed QDs (units = μM) are listed above each spectrum. Spectra are offset; dashed horizontal lines from 575 to 675 nm correspond to log(1/R′∞) of 0 for each spectrum. (c) Value of log(1/R′∞)528 of QD-functionalized TiO2 films as a function of the concentration of QDs in dispersions from which they were adsorbed for 2 or 22 h. In each graph, error bars represent ±1 standard deviation relative to the average log(1/R′∞)528 of two films.
when the concentration of dispersed QDs was approximately 10 μM (Figure 3c). The maximum value of log(1/R′∞)528 was approximately 3-fold greater when QDs were adsorbed to TiO2 for 22 h than 2 h. However, the additional loading of QDs probably resulted primarily from agglomeration, as indicated by the opacity of the films, the heterogeneous coverage of QDs, and the sloped baselines of transmission-mode UV/vis spectra. SEM images were acquired to characterize more directly the agglomeration of QDs on TiO2. For this experiment, MHDAfunctionalized TiO2 films were immersed into 22 μM toluene dispersions of CdSe QDs for 2 and 22 h. A top-down secondary-electron (SE) image of the film prepared from 22 h exposure to QDs revealed the presence of agglomerates, ranging in size from 1 to 5 μm, at coverages of approximately 106 agglomerates per square centimeter of projected surface area of TiO2 (Figure 4a). These agglomerates probably contributed to the opacity and the scattering-induced raised and sloping baselines of transmission-mode absorption spectra of films prepared under these conditions (Figures 1b and 2b,c). In a cross-sectional backscattered-electron (BSE) image of the same film (Figure 4b), the agglomerates exhibited higher backscattered electron intensities than the surrounding regions of the TiO2 surface, suggesting that the agglomerates consisted of CdSe, which has higher molecular weight and backscatter
Figure 4. Top-down SE images and cross-sectional BSE images of QD-functionalized TiO2 films prepared by immersion of MHDAfunctionalized TiO2 films into 22 μM toluene dispersions of CdSe QDs for 22 h (a, b) or 2 h (c, d).
coefficient than the TiO2 substrate. The BSE image also revealed a thin (
