Article pubs.acs.org/JPCC
Probing Transient Electric Fields in Photoexcited Organic Semiconductor Thin Films and Interfaces by Time-Resolved Second Harmonic Generation Xiaoxi Wu, Heungman Park, and X.-Y. Zhu* Department of Chemistry, Columbia University, New York, New York 10027, United States ABSTRACT: We have probed photoinduced charge separation dynamics in organic semiconductor thin films and at their interfaces by femtosecond time-resolved second harmonic generation (TR-SHG), using the model systems of fullerene (C70) thin films and copper phthalocyanine (CuPc)/C70 interfaces. In neat C70 thin films, the formation of an internal electric field on a ∼10-ps time scale following photoexcitation is attributed to the photo-Dember effect, namely, charge separation resulting from a gradient in excitation density and the differential electron/hole mobility. When an ultrathin film of the electron donor CuPc was deposited on top of the C70 film, an additional interfacial charge separation channel occurring on the ultrafast time scale of ∼0.1 ps was observed. We discuss how SHG fields from different origins can interfere to give the overall transient response.
SHG from thin films of centrosymmetric materials, such as CuPc and C70, has been studied before.23−25 In the electric dipole approximation, SHG is forbidden in the bulk of a centrosymmetric material but allowed at the surface or interface where symmetry is broken. However, bulk SHG in a centrosymmetric material is allowed from other contributions, including magnetic dipole coupling (MDC), magnetic dipole (MD), electric quadruple coupling (EQC), and electric quadruple (EQ) mechanisms.26−28 The interference between bulk and surface SHG responses can result in thickness dependent oscillation in SHG intensity.26 Changes in effective second-order nonlinear susceptibility can result from electronic excitation of the molecules, leading to a pump-induced change in SHG signal from a bulk thin film.15 To delineate various contributions to transient SHG signals from photoexcited organic semiconductor interfaces, we have chosen the model systems of neat C70 and planar CuPc/C70 bilayer thin films. The CuPc/fullerene solar cell is the most extensively studied smallmolecule OPV system, with CuPc as electron donor and fullerene as acceptor.29−32 Preliminary experiments on this system focused on the interfacial EFISH contribution from photoinduced charge transfer.2,15 Herein, we report a new transient EFISH response in the neat C70 thin film rising from the generation of an internal electric field. We further demonstrate that this transient electric field perpendicular to the film originates from the photo-Dember effect,33 where photoexcited charge carriers separate due to the intrinsic gradient of excitation density and the unbalanced electron and hole motilities. This is illustrated in Figure 1 by the simulated charge density distributions across the C70 thin film at two
1. INTRODUCTION Understanding photoinduced charge separation in organic semiconductor thin films and at their interfaces is essential to the development of organic photovoltaics (OPVs).1,2 The “standard” model for OPVs assumes that charge separation occurs almost exclusively at the donor/acceptor (D/A) interface, where an energetic driving force is present between the exciton state in the donor material and charge-transfer states across the D/A interface.3,4 However, it is also known that photoinduced charge separation can occur in neat organic semiconductor materials.5−7 Previous studies on charge separation in OPVs have relied on bulk-sensitive spectroscopies, such as transient absorption, 8−10 time-resolved fluorescence,11 and time-resolved terahertz and microwave conductivity.12,13 Recently, we showed that time-resolved second harmonic generation (TR-SHG) can be a powerful technique in specifically probing charge separation at donor/ acceptor interfaces on the femtosecond time scale.2,14−16 The transient electric field resulting from charge separation adds an asymmetry component to the nonlinear optical response, leading to electric-field-induced second harmonic (EFISH) generation.17−22 EFISH can be viewed as a four-wave mixing nonlinear process where two fundamental optical fields at frequency ω mix with the quasi-direct-current field to generate SHG at 2ω. However, the analysis of transient EFISH responses in organic thin films and at their interfaces is not straightforward, as there can be multiple contributions to the overall SHG response with complex phase relationships. In addition to the EFISH signal from interfacial charge separation, transient SHG can also arise from photoexcitons of a bulk film, in addition to static SHG from the sample. The static and pump-induced SHG signals might have different phases that can give rise to constructive or destructive inferences. © 2014 American Chemical Society
Received: March 9, 2014 Revised: April 28, 2014 Published: May 5, 2014 10670
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
The Journal of Physical Chemistry C
Article
incident angles of 42° and 45°, respectively, from the surface normal. The spot diameters of the pump and probe beams on the surface were approximately 125 and 100 μm, corresponding to power densities of 100 μJ/cm2 and 1 mJ/cm2, respectively. The inset in Figure 1 shows the experimental setup. The polarization of the probe pulse was controlled with a half-wave plate, with the pump pulse fixed at p-polarization. The time delay between the pump and the probe pulses was controlled by a translation stage. We optimized the spatial overlap and determined time zero between the pump and probe pulses by maximizing the sum frequency generation signal from the sample surface. The cross-correlation determined by SFG was 110 fs (fwhm), corresponding to a laser pulse width of ∼70 fs. The reflected SHG light, almost entirely of p-polarization, was then spectrally filtered by a combination of color glass filters and a monochromator (Oriel Cornerstone 130), sent to a photomultiplier tube, with the signal amplified (Stanford Research Systems SR445A) and recorded on a two-channel gated photon counter (Stanford Research Systems). All experiments were performed in a N2-filled cryostat at room temperature.
Figure 1. Net charge density as a function of distance from the top surface of a 120-nm-thick C70 thin film irradiated from the top. The simulation results at two different delay times (blue, 0.1 ps; red, 40 ps) are shown. The asymmetric charge distribution results from a gradient in excitation density (from top to bottom of the film) and differential mobility of photogenerated electrons and holes. Details of the simulation can be found in section 3.2. The inset illustrates the sample geometry and the SHG experimental setup.
different pump−probe delays (see section 3.2). On a short time scale (≤0.1 ps), the C70 film is almost charge neutral at all distances from the top surface. At longer times (40 ps), the photo-Dember effect leads to a net positive charge near the top surface and a net negative charge near the bottom surface. Such charge-separation processes result in a time-varying electric field across the C70 thin film and an internal EFISH contribution. We show that this transient EFISH contribution interferes with the static SHG constructively or destructively, depending on both film thickness and light polarization. We also show that the deposition of a thin CuPc layer on top of the C70 thin film leads to an additional interfacial EFISH contribution from photoinduced electron transfer from CuPc to C70. We emphasize the need for careful and quantitative analysis when applying the transient EFISH technique to probe interfacial charge separation.
3. RESULTS AND DISCUSSION 3.1. Time-Resolved SHG Profiles from Neat C70 and CuPc/C70 Bilayer Thin Films. We first focus on the photoexcitation of neat C70 films. Figure 2A shows timedependent SHG responses probed with p-polarized light following photoexcitation at 610 nm (hν = 2.03 eV) of C70 films with different thicknesses (from bottom to top, 40−120 nm). In all cases, we observed a pump-induced change in SHG
2. EXPERIMENTAL SECTION All thin-film samples were prepared on single-crystal sapphire (0001) surfaces (MTI Corporation). The sapphire substrate was rinsed with isopropanol and loaded in a vacuum chamber (base pressure = 10−6 Torr) for thermal evaporation. A C70 (Luminescent Technologies, >99%) thin film was deposited onto the sapphire surface at a rate of 0.01 nm/s to a total thickness of 40−160 nm as monitored in situ by a quartz crystal microbalance and calibrated with ellipsometry. To make a bilayer sample, we deposited a 5-nm CuPc (Sigma-Aldrich, >99%) thin film on top of each C70 film. The sample was transferred to a cryostat (Janis ST-100) in a N2-filled glovebox for SHG measurements. The morphology of the thermally evaporated C70 film was imaged by atomic force microscopy (AFM; XE-100, Park Systems) in tapping mode. Cantilevers with resonance frequencies of 204−497 kHz were used in the measurements. We carried out TR-SHG experiments using a Ti-sapphire laser system. A femtosecond regenerative amplifier (Coherent REGA, 250 kHz, λ = 810 nm) was pumped by a home-built oscillator to generate an 810-nm beam with a bandwidth of 40 nm. The REGA output was split into two beams: one used as the probe beam and the other used to pump a tunable optical parametric amplifier (Coherent Vis-OPA, 500−750 nm). The 610-nm pump beam from the OPA and the 810-nm probe beam from the REGA were focused on the sample surface with
Figure 2. TR-SHG profiles (dots) from (A) neat C70 thin films and (B) 5-nm CuPc deposited on C70 thin films. In both panels, the thicknesses of the C70 thin films are (from bottom to top) 40, 80, 120, and 160 nm, respectively. The experimental data were obtained with ppolarized fundamental light (ω) and p-polarized SHG (2ω). The red curves are fits to (A) eq 3 and (B) eq 5, with the fitting parameters summarized in Table 1 and Figure 4 10671
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
The Journal of Physical Chemistry C
Article
Table 1. Parameters Obtained from Global Fits to Time-Resolved SHG Profiles of Neat C70 Thin Films (τ and ϕ) and CuPc/C70 Bilayers (τ, ϕ, and θ) p-probe
s-probe
C70 thickness (nm)
τ (ps)
ϕ (rad)
θ (rad)
τ (ps)
ϕ (rad)
θ (rad)
40 80 120 160
12.1 ± 1.0 11.3 ± 0.7 6.3 ± 0.4 9.7 ± 1.2
−1.68 ± 0.01 1.45 ± 0.02 −1.70 ± 0.01 1.57 ± 0.01
3.10 ± 0.01
15.8 ± 1.7 10.9 ± 7.8 13.9 ± 3.2 9.0 ± 0.8
1.94 ± 0.03 −1.93 ± 0.27 1.71 ± 0.01 −1.46 ± 0.01
−2.11 ± 0.38
0 Edc(t ) ≈ Edc (1 − e−t / τ )
intensity, but the sign (positive or negative) of the change depended on the C70 film thickness. The similar time constants (∼10 ps) at different film thicknesses point to a common origin of the photoinduced dynamic process in C70. This process is slower than exciton− exciton annihilation, which features a time constant of 1 ps,34 but much faster than the nanosecond intersystem crossing process.35 The photon energy of 2.03 eV is known to excite intermolecular charge-transfer (CT) states,36,37 leading to photocarrier generation. The absorption coefficient of C70 at this photon energy is α = 4 × 104 cm−1,38 which, for a 160-nmthick film, corresponds to a 50% decrease in excitation density from the top surface to the bottom of the C70 thin film. Such a concentration gradient of photocarriers generated at time zero should be followed by charge redistribution due to drift diffusion. Because the electron mobility is 2 orders of magnitude higher than that of the hole in C70,39,40 the driftdiffusion process in the concentration gradient is dominated by movement of the electrons, leading to charge separation and the formation of an internal electric field perpendicular to the film surface. This process is the photo-Dember effect known previously in inorganic semiconductors.33,41,42 As describe in section 3.2, we numerically simulated the photo-Dember effect and found that the simulated internal electric field rises with a time constant similar to that of the bulk EFISH. Thus, the timedependent change in second harmonic signal (ΔSHG) can be attributed to bulk EFISH from the transient electric field across the C70 thin film. With the above interpretation, we can also understand the change in the sign of ΔSHG with different film thickness. It is known that the phase of SHG optical field from a bulk film, such as the transient bulk EFISH contribution described here, depends on the film thickness.26,27 However, another major contribution to the static SHG response is surface SHG, whose phase does not depend on the film thickness. As a result, there should be constructive or destructive interference between static SHG and photoinduced bulk EFISH, depending on the relative phase shifts. The total SHG intensity can be described by
We normalized I2ω(t ≥ 0) to the time-independent static SHG signal I0 and took the difference ΔI 2ω =
I 2ω(t ≥ 0) − I0 = [A1(1 − e−t / τ )]2 I0
+ 2A1(1 − e−t / τ ) cos ϕ
(3)
where A1 = |χ(3)/χ(2)|·|E0dc|. With the above interpretation of SHG from C70 thin films, we now turn to EFISH from photoinduced charge separation at the CuPc/C70 interface. In the experiment, we deposited an ultrathin CuPc film (5 nm) on top of the C70 film with varying thickness (40−160 nm). Our previous measurements established an interfacial EFISH contribution resulting from ultrafast (∼130 fs) electron transfer from photoexcited CuPc to C70.2 On the time scale of tens of picoseconds in Figure 2, this EFISH contribution should lead to a nearly instantaneous change in SHG signal. This is indeed observed in Figure 2B for 5-nm CuPc on C70 with thickness L = 40, 80, 120, and 160 nm (from bottom to top), respectively. A comparison of Figure 2A (neat C70) with those in Figure 2B reveals that the slow dynamics (∼10 ps) attributed to EFISH from bulk C70 layer is clearly present in the SHG profiles from CuPc/C70. For each C70 film thickness, the presence of CuPc adds an additional ultrafast (∼0.1 ps) and positive component to the SHG profile. The additional ultrafast component in each SHG profile in Figure 2B is consistent with the known charge-separation time of ∼130 fs at the CuPc/C70 interface.2 Thus, eq 1 can be modified to represent the total SHG response from the CuPc/ C70 samples (3) int I 2ω ∝ |χ (2) + χ (3) Edc(t )e iϕ + χint Edc (t )e iθ |2
(4)
where the first two terms are identical to those in eq 1. The third term is the interfacial EFISH contribution due to charge int separation at the CuPc/C70 interface; χ(3) int and Edc (t) are the effective third-order nonlinear susceptibility and electric field, respectively, at the CuPc/C70 interface; and θ is the relative phase between the interfacial EFISH and the static SHG. Because the interfacial EFISH signal rises on the ultrafast time scale of ∼130 fs, with little decay on the 35-ps time scale in Figure 2B, we can approximate it by a simple step function. After normalization of I2ω(t ≥ 0) to the static SHG signal I0 before the arrival of the pump laser pulse and take the difference, the pump-induced change in SHG intensity is given by
I 2ω ∝ |χ (2) + χ (3) Edc(t )e iϕ|2 = |χ (2) |2 + |χ (3) Edc(t )|2 + 2|χ (2) χ (3) Edc(t )| cos ϕ
(2)
(1)
where I2ω is the total SHG intensity; χ(2) and χ(3) are the effective second- and third-order nonlinear susceptibilities, respectively; t is the pump−probe delay time; and ϕ is the relative phase between static SHG and transient EFISH. Here, the first term in eq 1 is the static SHG signal, I0 = I2ω (t < 0), before photoexcitation. Edc is the time-dependent internal dc electric field from the separation of photocarriers and can be approximated by an exponential function with time constant τ (see simulation below)
ΔI 2ω = [A1(1 − et / τ )]2 + 2A1(1 − et / τ ) cos ϕ + (A 2 )2 + 2A 2 cos θ + 2A1(1 − et / τ )A 2 cos(ϕ − θ ) 10672
(5)
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
The Journal of Physical Chemistry C
Article
(2) int where A1, ϕ, and θ were defined earlier and A2 = |χ(3) int /χ |·|Edc | int with |Edc | (for t > 0) treated as a constant in the step-function approximation. Because A1, ϕ, and τ in eqs 3 and 5 should be the same for each C70 thickness with or without a CuPc overlayer, while the interfacial SHG contribution should not depend on C70 thickness, we fit time-dependent SHG responses from neat C70 films (data in Figure 2A) and from CuPc/C70 bilayer films (data in Figure 2B) simultaneously with global fitting. We link A1, ϕ, and τ for each C70 thickness and link θ for all thicknesses. The fitting curves are plotted in Figure 2 (red curves), and the values of the fitting parameters are summarized in Table 1 and Figure 4. The differences in relative phase ϕ between static SHG and dynamic EFISH in the C70 thin film satisfactorily account for the changing signs of pump-induced SHG signal with C70 thickness. For all thicknesses, the EFISH from the CuPc/C70 interface interferes with static SHG with a phase shift θ = 3.10 ± 0.01 rad. Similar results were also obtained for the same samples probed with s-polarization (ω) and detected with p-polarization (2ω). Figure 3 compares TR-SHG responses from neat C70 and
Figure 4. Relative phase shift (ϕ) of transient bulk EFISH from static SHG as a function of C70 film thickness with p- (red circles) and s(blue squares) probe polarization. The red and blue curves show approximate fits to sine waves.
with a constant phase shift θ between interfacial EFISH and static surface SHG. As shown in Table 1, the phase shift (ϕ) of bulk EFISH in C70 relative to static SHG depends strongly on the C70 film thickness. This can be understood from the fact that the phase of the transient EFISH from bulk C70 depends on film thickness, whereas the static surface SHG (likely a major contribution to the static SHG signal probed here) is insensitive to film thickness. This leads to thickness-dependent interference between the transient and static SHG components, namely, oscillatory behavior in ϕ versus C70 film thickness. Similar oscillatory behavior has been observed for static SHG from C60 thin films as a function of film thickness and quantitatively explained by the interference between surface and bulk SHG.26,27 In the present case on transient SHG from C70 and CuPc/C70, we interpret the thickness-dependent pump-induced SHG signal as resulting from interference between a transient third-order (χ(3)) response and static second-order responses (χ(2)). The involvement of both second- and third-order responses adds much more complexity to the problem (as compared to second-order processes26,43), and we were not able to obtain tractable formulas for the interference effect. In Figure 4, we plot the limited number (four) of data points for ϕ versus C70 film thickness and show qualitatively that the thickness dependence can be approximated by a simple sine oscillatory function, with repeating periods of λ ≈ 80 nm, for s- and p-polarized probes. These λ values are similar to the repeating length scales reported and quantitatively analyzed by Koopmans et al. for the thicknessdependent interference between surface electric-dipole-allowed SHG and bulk magnetic-dipole-allowed SHG from C60 thin films.26,27 3.2. Numerical Simulation of Photoinduced Charge Separation in C70 Thin Films. To support the mechanism of transient bulk EFISH generation by the photo-Dember effect, we carryied out a numerical simulation of the charge-separation process. The simulation was done for C70 thin films only, not C70/CuPc bilayers; the addition of a thin CuPc layer on top of each C70 thin film adds an ultrafast (∼130 fs) interfacial EFISH contribution, with no effect on the slower photo-Dember effect in the bulk C70 thin film. In the photo-Dember effect, the transient charge separation and resulting electric field across the C70 film come from the excitation density gradient and the differential electron/hole mobility in C70 thin films. We numerically solved the time-dependent electron and hole drift-diffusion equations, in conjunction with Poisson’s equation,44 using finite-element analysis in MATLAB. Each C70 thin film was treated as a slab of thickness L, with z = 0
Figure 3. TR-SHG profiles (dots) from (A) neat C70 thin films and (B) 5-nm CuPc deposited on C70 thin films. In both panels, the thicknesses of the C70 thin films are (from bottom to top) 40, 80, 120, and 160 nm, respectively. The experimental data were obtained with spolarized fundamental light (ω) and p-polarized SHG (2ω). The red curves are fits to (A) eq 3 and (B) eq 5, with the fitting parameters summarized in Table 1 and Figure 4
5-nm CuPc on C70 (from bottom to top, 40−160 nm). Fitting parameters ϕ and τ are also summarized in Table 1 and Figure 4. The differences in relative phase ϕ between static SHG and dynamic EFISH in the C70 thin film again account for the changing signs of the pump-induced SHG signal. We also show in Figure 3 global fits (red curves) to eqs 3 and 5 using the linked parameters (A1, τ, and ϕ) for bulk EFISH from C70 and 10673
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
The Journal of Physical Chemistry C
Article
n(z , t ) = 0 ⎫ ⎪ p(z , t ) = 0 ⎪ ⎪ dn ⎪ = 0⎪ Jn = dt ⎬ ⎪ dp = 0⎪ Jp = dt ⎪ ⎪ dV =0 ⎪ ⎭ dz
representing the surface and z = L representing the bottom of the film. In the one-dimensional approximation, Poisson’s equation is ∂ 2V (z , t ) ρ (z , t ) =− ε ∂z 2
(6)
where V(z,t) is the electric field and ε is the permittivity. The local net charge density, ρ(z,t), is ρ(z , t ) = q[p(z , t ) − n(z , t )]
(7)
∂n ∂ ⎛⎜ ∂V ∂n ⎞⎟ + μn n − Dn = G (z , t ) ∂t ∂z ⎝ ∂z ∂z ⎠ (8)
where we assume the Einstein equation relating carrier diffusivity Dn,p to mobility μn,p Dn,p =
(12)
These boundary conditions are necessary to ensure that there is no net charge flux leaving the C70 film at either the vacuum side or the sapphire side. Both electrons and hole are confined within the film. We used literature values39,46 for electron and hole mobilities in C70: μn = (2.5−5.5) × 10−3 cm2 V−1 s−1 and μp = 1 × 10−5 cm2 V−1 s−1. The absorption coefficient of C70 is 4 × 104 cm−1 at 610 nm.38 The pump laser pulse energy density on the surface of the C70 thin film is 100 μJ cm−2. The relative permittivity of C70 is ε = 4.84,47 and the calculated reflection coefficient at the surface is r = 0.2. Two snapshots of charge distributions at t = 0.1 and 40 ps for L = 120 nm and μn = 5 × 10−3 cm2 V−1 s−1 are shown in Figure 1. Figure 5 shows simulated time-dependent electric fields for four C70 film thicknesses. In panel A, we used the same electron
,where q is the fundamental electron charge and p and n are the concentrations of holes and electrons, respectively. The chargecarrier transport process is described by the well-known driftdiffusion equations
∂p ∂p ⎞ ∂⎛ ∂V ⎜μp p ⎟ = G (z , t ) + − Dp ∂t ∂z ⎝ ∂z ∂z ⎠
for z = 0, L ; t > 0
k bTμn,p q
(9)
In eq 8, G(z,t) is the electron−hole pair-generation rate obtained from the spatial-temporal profile of the absorbed femtosecond laser pulse. We can approximate the pump pulse by a Gaussian, and the generation rate can be written as G(z , t ) = αf (1 − r )e−αz
1 W 2π
et
2
/2W 2
(10)
where α is the absorption coefficient of the C70 film, f is the photon flux (number of photons per pulse per unit area), r is the reflection coefficient at the N2 gas/C70 solid interface with 45° incidence, and W is the full width at half-maximum (fwhm = 100 fs) of the laser pulse obtained from pump−probe crosscorrelation measurements. Because the lifetimes of photogenerated charge carriers in C70 film are on the nanosecond time scale,45 we neglected bulk recombination within the 40-ps time window probed here. For simplicity, we assumed that all absorbed photons lead to charge-carrier generation. In reality, the quantum yield of carrier generation is less than 1, but including this factor would change only the magnitude, not the time constant, of electric field formation. In the finite-element simulation, we set the boundary conditions as follows. Before the arrival of the pump-laser pulse (t < 0), there are no net free electrons or holes in the C70 film. The initial conditions are
Figure 5. Simulated electric fields for four C70 film thicknesses: L = 40 nm (red), 80 nm (blue), 120 nm (green), and 160 nm (orange). (A) Constant electron mobility, μn = 5 × 10−3 cm2 V−1 s−1, for all thicknesses. (B) Varying electron mobilities of μn = 4 × 10−3, 5 × 10−3, 6 × 10−3, and 7 × 10−3 cm2 V−1 s−1 for L = 40, 80, 120, and 160 nm, respectively. The dashed curves in panel A show exponential fits with time constants of t = 13, 14, 15, and 16 ps for C70 film thicknesses of 40, 80, 120, and 160 nm, respectively.
mobility of μn = 5 × 10−3 cm2 V−1 s−1 in the simulation. We found that the simulation results agreed semiquantitatively with the experimental dynamics obtained from time-dependent SHG profiles. The rise of the electric field can be well described by a single exponential function, eq 2, with time constants of τ = 13−16 ps for film thickness of 40−160 nm (dashed curve in Figure 5A). For comparison, the slow part of each experimental SHG profile attributed to transient bulk EFISH generation can be well described by eq 2, with time constant constants of 6−16 ps (Table 1). Thus, the simulation provides strong support for the interpretation that the observed dynamics of pump-induced SHG from C70 thin films come from bulk EFISH, a result of
n(z , t < 0) = 0 (0 ≤ z ≤ L) p(z , t < 0) = 0 (0 ≤ z ≤ L)
(11)
The boundary conditions for the charge density (n, p), flux (Jn, Jp), and electric field (−dV/dz) are 10674
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
The Journal of Physical Chemistry C
Article
more unambiguously followed and, thus, the dynamic process of interest to be isolated. This development is underway in our laboratory.
charge separation due to the presence of an excitation density gradient (across the film) and differential electron/hole mobility. The rise time of the electric field depends on electron mobility. The variation in time constants (with a factor of 2) of the bulk EFISH contribution from sample to sample (Table 1) can be attributed to changes in charge-carrier mobility in fullerene thin films depending on morphology and crystallinity.48,49 Thicker films can give better crystallinity and, thus, higher charge-carrier mobility. The higher mobility can lead to a faster rise in the electric field, as demonstrated by the simulations in Figure 5B, where the electron mobility increased from μn = 4 × 10−3 cm2 V−1 s−1 for 40-nm thickness to 7 × 10−3 cm2 V−1 s−1 at 160 nm. In our organic thin-film deposition system, the sample was not intentionally cooled. Because of the longer deposition time required for thicker films, the sample temperature rises slowly with time, resulting in different morphologies for different film thickness. This was verified by atomic force microscopy (AFM) imaging, as illustrated by the phase images (Ø) of two C70 films with thickness of 60 nm (Figure 6A) and 160 nm (Figure 6B). The average grain size
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by National Science Foundation Grant DMR-1125845. We thank Dr. A. E. Jailaubekov and Mr. Cory Nelson for technical help with SHG measurements.
■
REFERENCES
(1) Bakulin, A. A.; Rao, A.; Pavelyev, V. G.; van Loosdrecht, P. H. M.; Pshenichnikov, M. S.; Niedzialek, D.; Cornil, J. J.; Beljonne, D.; Friend, R. H. The Role of Driving Energy and Delocalized States for Charge Separation in Organic Semiconductors. Science 2012, 335, 1340−1344. (2) Jailaubekov, A. E.; Willard, A. P.; Tritsch, J. R.; Chan, W.-L.; Sai, N.; Gearba, R.; Kaake, L. G.; Williams, K. J.; Leung, K.; Rossky, P. J.; Zhu, X.-Y. Hot Charge-Transfer Excitons Set the Time Limit for Charge Separation at Donor/Acceptor Interfaces in Organic Photovoltaics. Nat. Mater. 2012, 12, 66−73. (3) Clarke, T. M.; Durrant, J. R. Charge Photogeneration in Organic Solar Cells. Chem. Rev. 2010, 110, 6736−6767. (4) Brédas, J.-L.; Norton, J. E.; Cornil, J.; Coropceanu, V. Molecular Understanding of Organic Solar Cells: The Challenges. Acc. Chem. Res. 2009, 42, 1691−1699. (5) Kazaoui, S.; Minami, N.; Tanabe, Y.; Byrne, H. J.; Eilmes, A.; Petelenz, P. Comprehensive Analysis of Intermolecular ChargeTransfer Excited States in C60 and C70 Films. Phys. Rev. B 1998, 58, 7689−7700. (6) Wei, X.; Jeglinski, S.; Paredes, O.; Vardeny, Z. V. V; Moses, D.; Srdanov, V. I. I.; Stucky, G. D. D.; Khemani, K. C. C.; Wudl, F. Studies of Photoexcitations in C70 by Optically Detected Magnetic Resonance. Solid State Commun. 1993, 85, 455−459. (7) Paquin, F.; Latini, G.; Sakowicz, M.; Karsenti, P.-L.; Wang, L.; Beljonne, D.; Stingelin, N.; Silva, C. Charge Separation in Semicrystalline Polymeric Semiconductors by Photoexcitation: Is the Mechanism Intrinsic or Extrinsic? Phys. Rev. Lett. 2011, 106, 197401. (8) Gélinas, S.; Rao, A.; Kumar, A.; Smith, S. L.; Chin, A. W.; Clark, J.; van der Poll, T. S.; Bazan, G. C.; Friend, R. H. Ultrafast Long-Range Charge Separation in Organic Semiconductor Photovoltaic Diodes. Science 2013, 343, 512−516. (9) Grancini, G.; Maiuri, M.; Fazzi, D.; Petrozza, A.; Egelhaaf, H.-J.; Brida, D.; Cerullo, G.; Lanzani, G. Hot Exciton Dissociation in Polymer Solar Cells. Nat. Mater. 2013, 12, 29−33. (10) Kaake, L. G.; Moses, D.; Heeger, A. J. Coherence and Uncertainty in Nanostructured Organic Photovoltaics. J. Phys. Chem. Lett. 2013, 4, 2264−2268. (11) Chen, K.; Barker, A. J.; Reish, M. E.; Gordon, K. C.; Hodgkiss, J. M. Broadband Ultrafast Photoluminescence Spectroscopy Resolves Charge Photogeneration via Delocalized Hot Excitons in Polymer:Fullerene Photovoltaic Blends. J. Am. Chem. Soc. 2013, 135, 18502− 18512. (12) Ai, X.; Beard, M. C.; Knutsen, K. P.; Shaheen, S. E.; Rumbles, G.; Ellingson, R. J. Photoinduced Charge Carrier Generation in a Poly(3-hexylthiophene) and Methanofullerene Bulk Heterojunction Investigated by Time-Resolved Terahertz Spectroscopy. J. Phys. Chem. B 2006, 110, 25462−25471. (13) Savenije, T. J.; Kroeze, J. E.; Yang, X.; Loos, J. The Effect of Thermal Treatment on the Morphology and Charge Carrier Dynamics in a Polythiophene−Fullerene Bulk Heterojunction. Adv. Funct. Mater. 2005, 15, 1260−1266.
Figure 6. AFM phase (Ø) images of C70 thin films with thicknesses of (A) 60 and (B) 160 nm. The scale bar is 200 nm in each image.
increased from ∼40 to ∼80 nm when the film thickness was increased from 60 to 160 nm, suggesting improved crystallinity and, thus, higher charge-carrier mobility in thicker films.
4. SUMMARY We have applied femtosecond TR-SHG to probe photoinduced charge-separation dynamics in organic thin films and at their interfaces. Using the model systems of C70 thin films and CuPc/C70 interfaces, we tracked the transient electric fields resulting from charge-separation dynamics that give rise to EFISH contributions to SHG signal. In neat C70 thin films, we found that photoexcitation leads to a dynamic process on a ∼10-ps time scale, which we attribute to the photo-Dember effect, namely, charge separation resulting from the excitation density gradient (from the top to the bottom of the thin film) and the differential electron/hole mobility. This interpretation was supported by numerical simulation of the drift-diffusion process of photocarriers. When an ultrathin film of the electron donor CuPc was deposited on top of the C70 thin film, we observed an additional EFISH contribution due to interfacial charge-separation channel occurring on the ultrafast time scale of ∼0.1 ps. The transient EFISH signal can interfere with static SHG signal, leading to a time-dependent increase or decrease in SHG signal, depending on the relative phase shifts of the optical fields. To further improve the TR-SHG technique in probing charge-separation dynamics in thin films and at interfaces, it is necessary to implement the phase-resolved detection of the transient SHG signal. Such phase information might allow the particular transient SHG contribution to be 10675
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
The Journal of Physical Chemistry C
Article
Picosecond Time-Resolved Spectroscopy in the Visible and Near-IR Region. J. Phys. Chem. 1996, 100, 10518−10522. (36) Kazaoui, S.; Minami, N. Intermolecular Charge Transfer Excitons in C70 as Compared with C60 Films. Synth. Met. 1997, 86, 2345−2346. (37) Kazaoui, S.; Minami, N.; Tanabe, Y.; Byrne, H. J.; Eilmes, A.; Petelenz, P. Comprehensive Analysis of Intermolecular ChargeTransfer Excited States in C60 and C70 Films. Phys. Rev. B 1998, 58, 7689−7700. (38) Zhou, W.; Xie, S.; Qian, S.; Zhou, T.; Zhao, R.; Wang, G.; Qian, L.; Li, W. Optical Absorption Spectra of C70 Thin Films. J. Appl. Phys. 1996, 80, 459−463. (39) Haddon, R. C. C70 Thin Film Transistors. J. Am. Chem. Soc. 1996, 118, 3041−3042. (40) Haddock, J. N.; Zhang, X.; Domercq, B.; Kippelen, B. Fullerene Based n-Type Organic Thin-Film Transistors. Org. Electron. 2005, 6, 182−187. (41) Reklaitis, A. Terahertz Emission from InAs Induced by PhotoDember Effect: Hydrodynamic Analysis and Monte Carlo Simulations. J. Appl. Phys. 2010, 108, 053102. (42) Gurevich, Y. G.; Meriuts, A. V. Dember Effect: Problems and Solutions. Phys. Lett. A 2013, 377, 2673−2675. (43) Wilk, D.; Johannsmann, D.; Stanners, C.; Shen, Y. R. SecondHarmonic Generation from C60 Thin Films at 1.064 μm. Phys. Rev. B 1995, 51, 10057−10067. (44) Tisdale, W. A.; Muntwiler, M.; Norris, D. J.; Aydil, E. S.; Zhu, X.-Y. Electron Dynamics at the ZnO (101̅0) Surface. J. Phys. Chem. C 2008, 112, 14682−14692. (45) Lee, C. H.; Yu, G.; Kraabel, B.; Moses, D.; Srdanov, V. I. Effects of Oxygen on the Photocarrier Dynamics in a C60 Film: Studies of Transient and Steady-State Photoconductivity. Phys. Rev. B 1994, 49, 10572−10576. (46) Anthopoulos, T. D.; de Leeuw, D. M.; Cantatore, E.; van’t Hof, P.; Alma, J.; Hummelen, J. C. Solution Processible Organic Transistors and Circuits Based on a C70 Methanofullerene. J. Appl. Phys. 2005, 98, 54503−54506. (47) Shinar, J., Vardeny, Z. V., Katafi, Z. H., Eds. Optical and Electronic Properties of Fullerenes and Fullerene-Based Materials; CRC Press: New York, 1999. (48) Singh, T. B.; Sariciftci, N. S.; Yang, H.; Yang, L.; Plochberger, B.; Sitter, H. Correlation of Crystalline and Structural Properties of C60 Thin Films Grown at Various Temperature with Charge Carrier Mobility. Appl. Phys. Lett. 2007, 90, 213512. (49) Kwiatkowski, J. J.; Frost, J. M.; Nelson, J. The Effect of Morphology on Electron Field-Effect Mobility in Disordered C60 Thin Films. Nano Lett. 2009, 9, 1085−1090.
(14) Tisdale, W. A.; Williams, K. J.; Timp, B. A.; Norris, D. J.; Aydil, E. S.; Zhu, X.-Y. Hot-Electron Transfer from Semiconductor Nanocrystals. Science 2010, 328, 1543−1547. (15) Kaake, L. G.; Jailaubekov, A.; Williams, K. J.; Zhu, X.-Y. Probing Ultrafast Charge Separation at Organic Donor/Acceptor Interfaces by a Femtosecond Electric Field Meter. Appl. Phys. Lett. 2011, 99, 83303−83307. (16) Park, H.; Gutierrez, M.; Wu, X.; Kim, W.; Zhu, X.-Y. Optical Probe of Charge Separation at Organic/Inorganic Semiconductor Interfaces. J. Phys. Chem. C 2013, 117, 10974−10979. (17) Lee, C.; Chang, R.; Bloembergen, N. Nonlinear Electroreflectance in Silicon and Silver. Phys. Rev. Lett. 1967, 18, 167−170. (18) Zhao, X.; Ong, S.; Wang, H.; Eisenthal, K. B. New Method for Determination of Surface pKa Using Second Harmonic Generation. Chem. Phys. Lett. 1993, 214, 203−207. (19) Qi, J.; Yeganeh, M.; Koltover, I.; Yodh, A.; Theis, W. DepletionElectric-Field-Induced Changes in Second-Harmonic Generation from GaAs. Phys. Rev. Lett. 1993, 71, 633−636. (20) Aktsipetrov, O. A.; Fedyanin, A. A.; Golovkina, V. N.; Murzina, T. V. Optical Second-Harmonic Generation Induced by a dc Electric Field at the Si−SiO2 Interface. Opt. Lett. 1994, 19, 1450−1452. (21) Bloch, J.; Mihaychuk, J.; van Driel, H. Electron Photoinjection from Silicon to Ultrathin SiO2 Films via Ambient Oxygen. Phys. Rev. Lett. 1996, 77, 920−923. (22) Glinka, Y. D.; Shahbazyan, T. V.; Perakis, I. E.; Tolk, N. H.; Liu, X.; Sasaki, Y.; Furdyna, J. K. Ultrafast Dynamics of Interfacial Electric Fields in Semiconductor Heterostructures Monitored by Pump−Probe Second-Harmonic Generation. Appl. Phys. Lett. 2002, 81, 3717. (23) Hoshi, H.; Yamada, T.; Ishikawa, K.; Takezoe, H.; Fukuda, A. Second-Harmonic Generation in Centrosymmetric Molecular Films: Analysis under Anisotropic Conditions. Phys. Rev. B 1995, 52, 12355. (24) Koopmans, B.; Anema, A.; Jonkman, H. T.; Sawatzky, G. A.; van der Woude, F. Resonant-Optical-Second-Harmonic Generation from Thin C60 Films. Phys. Rev. B 1993, 48, 2759−2764. (25) Kumagai, K.; Mizutani, G.; Tsukioka, H.; Yamauchi, T.; Ushioda, S. Second-Harmonic Generation in Thin Films of Copper Phthalocyanine. Phys. Rev. B 1993, 48, 14488−14495. (26) Koopmans, B.; Janner, A.-M. M.; Jonkman, H. T.; Sawatzky, G. A.; van der Woude, F. Strong Bulk Magnetic Dipole Induced SecondHarmonic Generation from C60. Phys. Rev. Lett. 1993, 71, 3569−3572. (27) Koopmans, B.; Janner, A. M.; Wierenga, H. A.; Rasing, T.; Sawatzky, G. A.; Woude, F.; van der Woude, F. Separation of Interface and Bulk Contributions in Second-Harmonic Generation from Magnetic and Non-Magnetic Multilayers. Appl. Phys. A 1995, 60, 103−111. (28) Yamada, T.; Hoshi, H.; Ishikawa, K.; Takezoe, H.; Fukuda, A. Origin of Second-Harmonic Generation in Vacuum-Evaporated Copper Phthalocyanine Film. Jpn. J. Appl. Phys. 1995, 34, L299−L302. (29) Xue, J.; Rand, B. P.; Uchida, S.; Forrest, S. R. A Hybrid PlanarMixed Molecular Heterojunction Photovoltaic Cell. Adv. Mater. 2005, 17, 66−71. (30) Akaike, K.; Kanai, K.; Ouchi, Y.; Seki, K. Impact of GroundState Charge Transfer and Polarization Energy Change on Energy Band Offsets at Donor/Acceptor Interface in Organic Photovoltaics. Adv. Funct. Mater. 2010, 20, 715−721. (31) Tang, J. X.; Zhou, Y. C.; Liu, Z. T.; Lee, C. S.; Lee, S. T. Interfacial Electronic Structures in an Organic Double-Heterostructure Photovoltaic Cell. Appl. Phys. Lett. 2008, 93, 043512. (32) Dutton, G. J.; Robey, S. W. Exciton Dynamics at CuPc/C60 Interfaces: Energy Dependence of Exciton Dissociation. J. Phys. Chem. C 2012, 116, 19173−19181. (33) Dember, H. Photoelectric E.M.F. in Cuprous-Oxide Crystals. Phys. Z. 1931, 32, 554. (34) Cho, H. S.; Yang, S. I.; Kim, S. K.; Shin, E.; Kim, D. Comparative Study of Exciton Relaxation Dynamics in C60 and C70 Solid. J. Phys. Chem. B 1999, 103, 6504−6508. (35) Watanabe, A.; Ito, O.; Watanabe, M.; Saito, H.; Koishi, M. Excited States of C70 and the Intersystem Crossing Process Studied by 10676
dx.doi.org/10.1021/jp502381j | J. Phys. Chem. C 2014, 118, 10670−10676
