Photogenerated Intrinsic Free Carriers in Small-molecule Organic Semiconductors Visualized by Ultrafast Spectroscopy

Confirmation of direct photogeneration of intrinsic delocalized free carriers in small-molecule organic semiconductors has been a long-sought but unsolved issue, which is of fundamental significance to its application in photo-electric devices. Although the excitonic description of photoexcitation in these materials has been widely accepted, this concept is challenged by recently reported phenomena. Here we report observation of direct delocalized free carrier generation upon interband photoexcitation in highly crystalline zinc phthalocyanine films prepared by the weak epitaxy growth method using ultrafast spectroscopy. Transient absorption spectra spanning the visible to mid-infrared region revealed the existence of short-lived free electrons and holes with a diffusion length estimated to cross at least 11 molecules along the π−π stacking direction that subsequently localize to form charge transfer excitons. The interband transition was evidenced by ultraviolet-visible absorption, photoluminescence and electroluminescence spectroscopy. Our results suggest that delocalized free carriers photogeneration can also be achieved in organic semiconductors when the molecules are packed properly.

Sample preparation. ZnPc was purchased from Sigma-Aldrich Co. and F 16 CuPc from Luminescence Technology Co. The inducing layer material 2,5-bis(4-biphenylyl)-bithiophene (BP2T) was synthesized with a previously described method 1 . All materials were purified twice prior to use by thermal-gradient train sublimation.
Fused quartz substrates were ultrasonicated sequentially in a NaOH solution, deionized water and alcohol for 30 min in each process, and then dried at 120 C for 30 min in an oven to eliminate the influence of moisture. During weak epitaxy growth (WEG) of ZnPc films, a 5 nm layer of BP2T was thermally evaporated on fused quartz substrates at a substrate temperature of 155 C. A 40 nm layer of ZnPc was then deposited onto the BP2T film with a substrate temperature maintained at 150 C. All the organic thin films were deposited in a vacuum chamber at 10 -4 -10 -5 Pa and the deposition rate was about 1 nm/min. Film thicknesses were monitored by a quartz-crystal microbalance during deposition. The fabrication condition for pristine ZnPc films was identical to that for WEG ZnPc films, except for the absence of the BP2T inducing layer.
For the samples measured in the mid-infrared (IR) region, we used 0.5 mm thick CaF 2 plates as replacement substrates for fused quartz, which were cleaned with acetone and alcohol.
For the sample prepared for electroluminescence (EL) measurements, indium-tin oxide (ITO)-coated glass with a sheet resistance of 15 / was used as the substrate. 3 Before deposition, the substrates were cleaned with detergent, and then ultrasonicated sequentially in acetone, alcohol and deionized water and dried in pure N 2 . Subsequently, poly (3,4- To acquire SAED of WEG ZnPc films, a 40-nm thick ZnPc film was vacuum deposited on 5-nm thick amorphous SiN using WEG method. The WEG ZnPc film was mounted on a holder with heating function and the temperature-dependent SAED was measured by a TEM with a thermionic gun (JEOL, JEM 2010). During the measurement, the intensities of electron beams were controlled as low as possible (spotsize: 5, current density: ~0.0 pA/cm 2 ) to minimize possible radiation damage to the sample. Spectroscopic characterization. Ultraviolet (UV)-visible absorption spectra of the ZnPc films were measured with a U-3010 spectrophotometer (Hitachi). When the temperature-dependent UV-visible difference spectra were measured, the films were attached to a metallic water jacket as the sample holder, and the temperature was controlled with a circulating water bath (HAAKE DC30-K20, Thermo Scientific).
Steady-state IR absorption spectra were measured in vacuum using a Fourier transform infrared (FTIR) spectrometer (VERTEX 70v, Bruker). Each FTIR spectrum was recorded at a resolution of 4 cm -1 and 20 scans were signal-averaged.
Photoluminescence (PL) spectra of the ZnPc films were acquired on a Raman microscopic spectrometer (LabRAM HR800, HORIBA Jobin Yvon) in PL configuration due to the extremely low photoluminescence intensity. A cryostat (TMS 94, Linkam Scientific Instruments) was used to control the temperature of the sample when the temperature-dependent PL spectra were measured. The excitations used were 532 nm and 785 nm lasers, and the size of the excitation spot on the sample was about 1 m.
For EL measurements, the bottom negative electrode (ITO) and top positive electrode (Au) were connected to a source meter (Keithley 2601A) using soft Au probes.
The sample was first subjected to a 0 V~2.5 V~0 V process. Then a durative -3 V voltage was applied to the sample. The EL spectra of the sample were collected with a low-noise CCD (Princeton Instrument ProEM: 512B) coupled spectrometer (Princeton Instrument Acton 2300) with an integration time of 1s. 5 To obtain standard spectral fingerprints for cations in the WEG ZnPc film, a sample with CaF 2 as the substrate was exposed to I 2 vapor for 30 s to produce ZnPc cations by oxidation with I 2 . UV-visible and FTIR spectra were measured after and before I 2 exposure from which the doping-induced difference absorption spectra in the visible region and IR regions were derived. and a repetition rate of 1 kHz from an amplified Ti:Sapphire laser (Hurricane, Spectra Physics) was split into two beams by an optical wedge. The more intense beam passed through a delay line and was focused onto the sample as the pump beam, while the weaker one was focused into a sapphire plate to generate a white light continuum as the probe beam. A filter was used to remove the 800 nm residual in the white light continuum, which was then split into two probe beams denoted as the signal and reference beam by a 1:1 broad-band beam splitter. The signal beam was focused at the same spot as the pump on the sample. The polarization between the pump and probe was set to the magic angle (54.7°) with an 800 nm half-wave plate. Reference light and transmitted signal light were collected by optical fibers and sent through an imaging spectrograph (SpectraPro-150, 6 Acton Research) to be acquired by a CCD (128HB, Acton Research). The as-acquired data were further smoothed, and corrected of the chirp effect using the chirped white light continuum as a reference, which was measured by the optical Kerr gating effect employing a quartz plate as the Kerr medium. The excitation wavelength was 800 nm for both pristine ZnPc and WEG ZnPc films, with excitation energies of ~400 Jcm -2 /pulse and ~300 Jcm -2 /pulse, respectively. Both measurements were conducted in the ambient environment.
(ii) In the near IR region. The experimental setup for measuring transient absorption spectra in the near-IR region was different to that used for the visible region. Briefly, the fundamental output of a pulsed laser with a central wavelength of 800 nm, a pulse duration of 120 fs (FWHM) and a repetition rate of 333 kHz from an amplified Ti:Sapphire laser (Spitfire, Spectra Physics) was split into two beams. One beam passed into an OPA (OPA-800 CF-1, Spectra Physics) to generate the excitation light while the other beam was focused into a sapphire plate to generate white light continuum as the probe beam. The polarization between the pump and probe was set to the magic angle (54.7°) and both were focused onto the sample at the same spot. Before focusing, the probe beam was filtered by a long-pass filter so that only the near-IR light could pass.
The excitation beam traveled through a delay line and was modulated by a mechanical chopper at a frequency of half of the laser repetition rate. The transmitted probe light was collected by an optical fiber and sent through an imaging spectrograph (SpectraPro 2300i, Princeton Instruments) to be acquired by an InGaAs array detector (OMA-V, Princeton Instruments). For both pristine ZnPc and WEG ZnPc films, the excitation wavelength was 617 nm and the excitation energy was ~50 Jcm -2 /pulse. The measurements were conducted in the ambient environment.
(iii) In the mid-IR region. Transient absorption spectra in mid-IR region were acquired with the following experimental setup. The output of an 800 nm centered pulsed laser with a duration of 35 fs (FWHM) at a repetition rate of 1 kHz from an amplified Ti:Sapphire laser (Spitfire Ace, Spectra Physics) was split into two beams by a 800 nm beam splitter. One of the beams passed through a delay line and was modulated by a mechanical chopper at a frequency of half of the laser repetition rate, which was employed as the excitation beam. The other beam was used to generate a broadband mid-IR source as the mid-IR probe via four-wave mixing through filamentation in air 3 .
During this process, the fundamental beam was compressed with a telescope and sent through a beta barium borate (BBO) crystal to generate the second harmonic generation (SHG). After SHG, the fundamental and second harmonic beams were separated with a dichroic mirror for individual adjustment of their relative time delays and polarization.
The fundamental beam first passed through a short delay line such that its time delay could be finely adjusted, then passed through a half-wave plate that rotated its light polarization by 90°, and finally recombined with the second harmonic beam by another dichroic mirror. The recombined beams were focused simultaneously by an aluminum 8 concave mirror into the air, where a bright filament formed, and a broad-band IR light was generated through the nonlinear optical processes in the filament 3 . A black screen was set some distance behind the filament, which was vertically adjusted such that it blocked most of the visible light while allowing enough mid-IR light as the probe to pass over the top edge of the block. Another aluminum concave mirror collected the mid-IR light and focused it onto the sample. Transmitted mid-IR light was collected, sent into an imaging spectrometer (iHR 320, HORIBA Jobin Yvon) and acquired by a 64-channel MCT array detector of Femtosecond Pulse Acquisition System (FPAS-0144, Infrared Systems Development). Both the spectral coverage and the peak wavelength of the IR probe generated from the filament are sensitive to the optical alignment of the mid-IR generation. To optimize the signal-to-noise ratio, we divided the spectrum range from 2941 to 9447 nm into three spectral segments for independent measurement by tuning the peak wavelength into the three spectral windows individually: from 2941 to 4202 nm, 3701 to 8298 nm (CO 2 absorption interference occurring in this range) and 7352 to 9447 nm, with a partial spectral overlap between the adjacent segments. Due to the limited detection range of the spectrometer, three spectral segments were further divided into slices for spectral acquisition. Finally, the full spectrum was obtained by merging all the individually measured spectral slices, where the overlapped spectral regions between two adjacent slices were used for data scaling. The excitation wavelength was 800 nm and 9 with an excitation energy of ~1200 J cm -2 /pulse. The measurements were conducted in the ambient environment.
Band structure calculation. WEG ZnPc films have been characterized to be in the metastable -phase (monoclinic, space group C2/n, which is referred to as -phase in some recent papers since -phase was redetermined to be of a monoclinic P -1 space group crystal structure 4 ) by X-ray diffraction (XRD) and selected area electron diffraction (SAED) 5 analyses. Because no detailed structural data of -ZnPc is available, generally the structural data of isomorphic CuPc 6 is used as the first approximation. This approach has been previously applied in calculating XRD patterns of ZnPc films 7 . Specifically, the Cu atom in -CuPc is replaced by a Zn atom, and the positions of the other atoms in the unit cell remain the same as those in -CuPc. The crystal structure of an -ZnPc primitive cell is shown in Supplementary Fig. 1a. The band structure of the -ZnPc crystal was calculated using density functional theory (DFT) in DMol 3 program 8 . We performed an all-electron calculation with a double numeric polarized (DNP) basis set 8 , generalized gradient approximation (GGA) 9 and the Becke-Lee-Yang-Parr (BLYP) exchange-correlation potential 10 . The k-point mesh was 441, and a Brillouin zone path was generated in the Monkhorst-Pack scheme 11