Role of electron injection in polyfluorene-based light emitting diodes containing PEDOT:PSS

We report electromodulation (EM) studies of polyfluorene-based light-emitting diodes containing poly(3,4-ethylene-dioxythiophene)-poly(styrene-sulfonate) (PEDOT:PSS), in which the barrier to hole injection is large $(\sim 0.7\mathrm{eV})$. Measurements are reported on devices fabricated with aluminium and barium cathodes to provide respectively poor and efficient electron injection into the active layer. The Al devices exhibit low currents, indicating low rates of electron and hole injection, whereas the Ba devices exhibit high currents and high electroluminescence efficiencies, implying efficient injection of both electrons and holes despite the large hole injection barrier. The Al devices show conventional field-induced EM behavior consistent with the Stark effect (SE). The Ba devices show conventional SE behavior for low applied biases but, above turn-on, the (field-induced) SE features vanish, indicating suppression of the internal field, and are replaced by charge-induced bleaching and absorption features. The behavior of the devices is attributed to the presence of electron traps close to the PEDOT:PSS/organic interface. The experimental findings are consistent with earlier findings by Murata et al., Van Woudenbergh et al., Poplavskyy et al., and Lane et al.

Figure 1

(a) Schematic diagram showing the HOMO and LUMO energy levels in a trap-rich LED containing a high density of trapped electrons close to the anode; the charge accumulation at the anode causes local enhancement of the electric field, leading to an increased rate of hole injection, better balance between electron and hole currents, and higher overall device efficiencies. (b) Energy level diagram for an ITO/PEDOT:PSS/5BTF8/cathode device, where the cathode material is chosen to be either Al (Ref. 29) or Ba (Ref. 29) in order to achieve poor or efficient electron injection. The work function of the ITO/PEDOT:PSS anode is taken from Ref. 30.

Figure 2

Current-voltage-luminance measurements for devices based on $\mathrm{ITO}∕\mathrm{PEDOT}:\mathrm{PSS}∕5\mathrm{BTF}8∕\mathrm{Ba}$ (circles) and $\mathrm{ITO}∕\mathrm{PEDOT}:\mathrm{PSS}∕5\mathrm{BTF}8∕\mathrm{Al}$ (squares). The symmetrical $\mathit{\text{I-V}}$ characteristics at low applied biases indicate the presence of conducting filaments in the active layer. The Ba device shows a high sharp onset in current at 2.2 V whereas the Al device shows a much weaker threshold. The inset shows the EL intensity from the two devices in arbitrary units; as shown, the limit of detection (L/O/D) coincides with the $x$-axis.

Figure 3

First harmonic electromodulation spectra for the Al and Ba devices, using 1.4 V and 0.7 V RMS modulation biases respectively, and a variety of applied dc offsets. The EM spectrum of the Al device is dominated by the electroabsorption (EA) response of the polymer at all applied biases. The EM response of the Ba device is also dominated by electroabsorption in reverse bias but, when the device is operating in forward bias, the EA features are replaced by excited state bleaching and absorption due to injected charge.

Figure 4

The frequency dependence of the EM signals for (a) Al and (b) Ba devices. The EM response of the Al device, measured at 414 nm, shows no frequency dependence in the range 0.1–5 kHz in either reverse (circles) or forward (triangles) bias, indicating a fast response that is consistent with electroabsorption. The EM response of the Ba device in reverse bias is also consistent with EA (circles) but the forward bias EM signal measured at 375 nm (triangles, squares) shows a strong frequency dependence which fits well to a monomolecular decay of lifetime $41\mu \mathrm{s}$ and is consistent with trapped charge. The curves labeled $I_x\left(\omega \right)$ and $I_y\left(\omega \right)$ correspond respectively to the in-phase and quadrature components of the forward bias EM response of the Ba device.

Figure 5

The dc bias dependence of the first harmonic electromodulation signal at 414 nm for (i) Al and (ii) Ba devices; the dashed line indicates the expected signal in the absence of field redistribution.

Figure 6

The variation of the weighting coefficients $\alpha$ and $\beta$ with applied dc bias, where $\alpha$ indicates the field-induced modulation and $\beta$ indicates the charge-induced modulation. Above 2.6 V, $\alpha$ is approximately zero, indicating suppression of the electric field. The inset shows a typical fit to Eq. (4) obtained at 3.6 V.