Tracking the Energy Flow on Nanoscale via Sample-Transmitted Excitation Photoluminescence Spectroscopy Pavel Moroz,†,‡ Natalia Razgoniaeva,†,‡ Yufan He,†,§ Gregory Jensen,‡ Holly Eckard,§ H. Peter Lu,*,†,§ and Mikhail Zamkov*,†,‡ †
The Center for Photochemical Sciences, ‡Department of Physics, and §Department of Chemistry, Bowling Green State University, Bowling Green, Ohio 43403, United States S Supporting Information *
ABSTRACT: Tracking the energy flow in nanoscale materials is an important yet challenging goal. Experimental methods for probing the intermolecular energy transfer (ET) are often burdened by the spectral crosstalk between donor and acceptor species, which complicates unraveling their individual contributions. This issue is particularly prominent in inorganic nanoparticles and biological macromolecules featuring broad absorbing profiles. Here, we demonstrate a general spectroscopic strategy for measuring the ET efficiency between nanostructured or molecular dyes exhibiting a significant donor−acceptor spectral overlap. The reported approach is enabled through spectral shaping of the broadband excitation light with solutions of donor molecules, which inhibits the excitation of respective donor species in the sample. The resulting changes in the acceptor emission induced by the spectral modulation of the excitation beam are then used to determine the quantum efficiency and the rate of ET processes between arbitrary fluorophores (molecules, nanoparticles, polymers) with high accuracy. The feasibility of the reported method was demonstrated using a control donor−acceptor system utilizing a protein-bridged Cy3-Cy5 dye pair and subsequently applied for studying the energy flow in a CdSe560-CdSe600 binary nanocrystal film. KEYWORDS: energy transfer, nanocrystals, FRET, crosstalk, bleed through
M
The feasibility of FRET measurements tends to diminish with the increasing molecular size. Large fluorophores, such as inorganic nanocrystals (NCs), proteins, or synthetic macromolecules exhibit spectrally broad absorbing transitions which obscure individual contributions from donor and acceptor moieties.30,31 The resulting spectral crosstalk in the excitation region complicates the ET analysis, leading to large experimental uncertainties. This issue appears to be particularly concerning in light of the growing demand for ET studies in complex molecular assemblies and diverse nanoscale systems, fueled by the ongoing advances in nano- and biomaterial synthesis. Here we demonstrate a general strategy for measuring the ET efficiency in molecular or nanostructured systems featuring a significant donor−acceptor spectral overlap. The developed sample-transmitted excitation photoluminescence (STEP) technique is enabled through spectral shaping of the broadband
easurements of the energy transfer (ET) on the nanoscale are important in many areas of science. Cascade-like ET is the first step of the energy conversion during photosynthesis1 and is the primary process of the energy flow in nanostructured (excitonic) solids2−4 and organic crystals.5 Similar patterns of the energy diffusion are observed in living tissues and proteins, where photon energy is transmitted tens of angstroms away from single-site excitations.6,7 To gain insights into these processes, optical techniques have been steadily evolving as nonintrusive probes of energy dynamics on nanoscale. Among those, Förster resonance energy-transfer (FRET)-based measurements8−15 represent a particularly rewarding strategy that relies on short-range dipole−dipole interactions of donor and acceptor dyes for detecting small changes in the intermolecular separation.16 Such distance sensitivity of FRET has found widespread applications in the fields of biology and biochemistry,17−20 where the dynamics of biomolecules can be imaged with a subnanometer precision, leading to important insight into processes of DNA folding, protein conformation21−25 and interaction,26,27 and metabolic pathways.28,29 © 2017 American Chemical Society
Received: February 17, 2017 Accepted: March 21, 2017 Published: March 21, 2017 4191
DOI: 10.1021/acsnano.7b01141 ACS Nano 2017, 11, 4191−4197
Article
www.acsnano.org
Article
ACS Nano excitation light with solutions of donor fluorophores designed to selectively suppress the excitation of respective donor species in the sample. The resulting changes in the acceptor emission induced by the spectral modulation of the excitation beam are then converted to the quantum efficiency of the ET between two fluorophores, ED→A. The feasibility of the STEP spectroscopy was demonstrated using a “control” system featuring a pair of Cy3 (donor) and Cy5 (acceptor) dyes terminally attached to a cell-signaling protein, calmodulin (CaM). The resulting ET scaffold exhibited a well-defined donor−acceptor distance and a low Cy3-Cy5 spectral overlap needed for comparing the measured ET efficiency with standard fluorescence (FL) lifetime-based methods. The STEP spectroscopy was subsequently applied for measuring the energy diffusion in a binary nanoparticle film of 3.5 nm (donor) and 4.5 nm (acceptor) CdSe semiconductor NCs with excitonic transitions at λD = 560 nm and λA = 600 nm, respectively. A significant donor−acceptor spectral overlap in this case is generally preclusive from using traditional spectroscopic methods to discern energy and charge-transfer pathways of the exciton decay. Application of the STEP methodology was therefore uniquely poised for distinguishing the CdSe560-to-CdSe600 ET contribution into donor emission quenching, which allowed determining quantum efficiencies for charge and ET processes between the two NC sublattices. Overall, present measurements demonstrate that shaping of the excitation light with solutions of donor molecules represents a powerful strategy for resolving the donor−acceptor spectral crosstalk. For materials where ET is primarily due to the FRET mechanism, the present method should enable accurate measurements of interparticle distances, offering an analog of the spectroscopic ruler for large molecules and nanostructures. Meanwhile, the reliance of the STEP strategy on the acceptor emission makes it amenable for studying ET processes in systems with nonemissive donors, such as plasmonic nanoparticles or inorganic surfaces. Further advantages of STEP are related to its capacity for distinguishing between energy and charge-transfer processes that often coexist in donor−acceptor assemblies. Indeed, by tracking the transfer of excitons from donor to acceptor species, this method accurately accounts for possible quenching of the donor fluorescence caused by the charge transfer, thus complementing donor lifetime-based ET measurements: ED→A = 1 − (τDonor+Acceptor/τDonor). Finally, the ability to measure ET efficiency without the need for photon counting or optical filters should further enhance the expediency of STEP, making it a useful addition to a toolbox of FRET-based spectroscopy methods.
Figures 2b and 3c, gray circles) and is generally expected for molecular materials under low-power excitation conditions. The strategy for determining the ET efficiency is described below for a forward ET (e.g., Dexter, FRET) between two fluorophores, which is subsequently adapted to the case of multiple ET partners (see eqs 4, 5 and SE16 ). For ET processes involving two molecules, application of the donorbased excitation filter was found to be sufficient for obtaining the absolute ET efficiency, ED→A (D-only mode). However, systems with multiple donor moieties may require consecutive applications of donor(s) and acceptor(s) excitation filters (D− A mode, described in the Supporting Information (SI) section). Filtering the broadband excitation light with the solution of donor molecules causes the reduction in the number of the excited donor species in the sample. The resulting changes in the number of photons emitted by the acceptor nanoparticle, NFL A , due to “blocking” of the excitation light with the donor solution are best illustrated using the f = NFL A /NA ratio. With increasing donor filter (D-filter) optical density (OD), f gradually diminishes (Figure 1d), ultimately reaching the following asymptotic form: ⎛ N FL ⎞ ⎛ N ⎞ lim f = lim ⎜ A ⎟ = lim ⎜1 + E D → A D ⎟ × QYA = QYA ND → 0 ND → 0⎝ NA ⎠ ND → 0⎝ NA ⎠ (2)
As illustrated in Figure 1d, the evolution of f with increasing filter OD reveals two characteristic parameters: the “saturation” value, NFL A /NA → M2 = QYA (when ND/NA ≪ 1), and the
Figure 1. Illustration of the STEP approach for measuring the ET efficiency, ED→A, in a simulated donor−acceptor system featuring an 80% absorption overlap. (a) The number of photons emitted by the acceptor molecule (NFL A ) can be expressed as a function of the quantum efficiency for the D → A ET, ED→A: NFL A = QYA(NA + ED→AND), where ND and NA are the numbers of excited donor and acceptor molecules, respectively, and QYA is the fluorescence quantum yield of the fluorophore A (in the presence of D). (b) Gaussian absorption profiles used in the simulation along with the associated acceptor emission. (c) Schematic illustration of shaping the white light (WL) excitation beam with a donor filter solution, which is designed to suppress the donor-to-acceptor ET in a sample. (d) Simulated evolution of f = NFL A /NA with increasing Dfilter optical density. The asymptotic form of f reveals M1 and M2 parameters, which are used to calculate the ET efficiency from eq 3. N0A and N0D are the relative numbers of excited acceptor and donor nanoparticles prior to the application of the D-filter, N0D/N0A = /Abssample . Abssample D A
RESULTS AND DISCUSSION The STEP approach is based on the assumption that a linear relationship exists between the number of photons emitted by an acceptor fluorophore, NFL A , and the numbers of excited acceptor (A) and donor (D) molecules, NA and ND, respectively: NAFL = QYA(NA + E D → A ND)
(1)
where ED→A is the quantum efficiency for D → A ET, and QYA is the emission quantum yield of the fluorophore A in the presence of the fluorophore D (as measured in the donor− acceptor assembly). The linearity of eq 1 was confirmed experimentally in this work using acceptor-only samples (see 4192
DOI: 10.1021/acsnano.7b01141 ACS Nano 2017, 11, 4191−4197
Article
ACS Nano relative reduction of this ratio, Δf → M1 = QYA × ED→A × (N0D/ N0A), where N0A and N0D are the numbers of excited acceptor and donor dyes in the absence of D-filter. If M1 and M2 are determined experimentally, one can calculate the quantum efficiency of the ET as follows: E D → A = (M1/M 2) × (N A0 /ND0)
N0A
circles in Figure 2b). Such acceptor-only measurements provided an estimate of the experimental error due to
(3)
N0D
where and are the relative numbers of excited acceptor and donor nanoparticles prior to the application of the D-filter, given by the probability of the photon absorbance by each specie in the investigated sample. As a result, the absolute ET efficiency, ED→A, is obtained without photon counting. Indeed, for a given photon wavelength, the actual number of molecules A undergoing excitation in the sample is proportional to its absorption value: NA = xA × Abssample , where xA is a constant A under a single-photon excitation regime (low power), which is proportional to the area of the excitation beam and its intensity. Similar strategy allows expressing the number of excited donor molecules as ND = xD × Abssample . Since both donor and D acceptor dyes occupy the same excitation volume in the homogeneous sample and irradiated with the same excitation beam, the unknown volume constants for D and A species must be identical: xA = xD ≡ x. Consequently, N0D/N0A = Abssample / D . Abssample A The strategy for extracting ED→A from STEP measurements is illustrated in Figure 1 using a simulated donor−acceptor system. By modeling Gaussian-shape donor and acceptor absorption profiles (80% spectral overlap) and the broadband excitation spectrum, nWL(λ), we obtain the relative numbers of acceptor and donor excitons in the sample, NA and ND, using eq SE4. Next, the expected emission intensity NFL A is simulated using eq 1 (for a given ED→A parameter) and normalized to obtain the ratio f = NFL A /NA. According to Figure 1d, the increase in the D-filter optical density results in a gradual reduction of f, leading to the saturation at a constant value (M2) when the filter concentration reaches 200 μM (assuming an extinction coefficient of ε = 1 × 106). D−A systems with a larger spectral overlap (>80%) will generally require a higher filter concentration to achieve asymptotic behavior (see Figure SF4). Finally, M1 and M2 parameters are retrieved from Figure 1d and used in eq 3 for calculating ED→A. Measuring the ET Efficiency in Cy3-CaM-Cy5 Assemblies. In our first test, the STEP technique was used to measure the ET efficiency across a pair of Cy3 (donor) and Cy5 (acceptor) dyes brought together using a calmodulin (CaM) protein linker. Since the intermolecular separation for these scaffolds is generally well controlled via a particular conformation of CaM, the Cy3-CaM-Cy5 system offered a suitable framework for assessing the accuracy of the developed approach. In the present study, an unfolded conformation of CaM with associated Cy3-Cy5 intermolecular distance of ≈60 Å was used.29,32 The Cy3-CaM-Cy5 conjugates were prepared in aqueous solutions using standard protocols.22−24 The sample was excited with a white light source allowing for the emission detection at a 90° angle (see the Supporting Information for details of measurement, Figure SF1). To determine the ET efficiency, the excitation beam was spectrally filtered using a solution of Cy3 molecules. The resulting Cy5 fluorescence count was normalized by NA (calculated from eq SE4) to obtain the value of fexp for each filter OD. In the absence of the Cy3 donor, the emission of isolated Cy5 molecules was found to be linear with NCy5: NFL Cy5 = QYCy5NCy5, evidenced by a nearly constant value of fexp (gray
Figure 2. STEP measurements of the ET efficiency between Cy3 and Cy5 dyes terminally attached to a CaM protein. (a) Absorption profiles of Cy3 (Donor) and Cy5 (Acceptor) dyes along with the emission of Cy5 (gray line). The excitation white light (WL) profile prior to the application of D-filter is indicated by a green shaded area. The inset shows a molecular graphics model of Cy3 and Cy5 attached to a CaM protein. Dye separation ≈6 nm.32 (b) Filtering the excitation light with a solution of Cy3 (donors) molecules results in the reduction of the fexp for Cy3-CaM-Cy5 samples but yields a constant fexp value in the case of isolated Cy5 molecules (no ET). The experimental data for Cy3-CaM-Cy5 assemblies is fitted with a parametric model curve, f theor(ED→A = 36.5%).
nonlinearity of eq 1, which was found to be
