Spectroscopic evidence for trap-dominated magnetic field effects in organic semiconductors

Polaron traps are ubiquitous in organic semiconductors and recent evidence suggests they might be crucial for the large observed magnetic field effects (MFEs) in organic semiconductors. Here we measure MFEs in polymer thin-film devices with engineered, radiative trap sites in order to spectroscopically investigate the influence of the traps. Surprisingly, the luminescence at the trap sites and the polymer backbone is found to have an opposite response to a magnetic field. All our results are compatible with a mechanism in which spin mixing at the traps can create the large MFEs observed on the backbone. This scenario is quantitatively confirmed by numerical drift-diffusion modeling with magnetic-field-dependent exciton densities at the traps. These insights solve an outstanding controversy within the research field.

Figure 1

(a) The chemical structure of the polymer components. The components $p$ and $q$ are the green and red dyes, respectively. The composition of the blue polymer (B) is $m$ = 50%, $n$ = 40%, $o$ = 10%, $p$ = 0%, and $q$ = 0%. For the blue-green copolymer (BG), $m$ = 50%, $n$ = 39.90%, $o$ = 10%, $p$ = 0.1%, and $q$ = 0%, and for the white-emitting copolymer (BGR), $m$ = 50%, $n$ = 39.88%, $o$ = 10%, $p$ = 0.1%, and $q$ = 0.02%. (b) Schematic representation of the band diagram of the white-emitting copolymer containing both dyes. Figure adapted from Ref. [22].

Figure 2

Electroluminescence (EL) and magnetoelectroluminescence (MEL) measurements performed on the blue-light-emitting polymer. (a) The normalized EL as a function of wavelength for a pristine device at currents corresponding to 2.8 and 5 V and a stressed device at 2.8 V. Two peaks are observed: one for bimolecular recombination at 460 nm and one for trap-assisted recombination at 640 nm. (b) The relative ratio of the EL at 640 nm with respect to 460 nm as a function of voltage and different amounts of electrical stressing. Results are shown for a pristine device (squares), after stressing at 6.5 V (circles), and after stressing at 7.5 V (triangles). The data points along the arrow were measured at the same current to show the increased ratio is not solely due to a voltage shift. Power-law fits are used as a guide to the eye (lines). (c) The MEL at a fixed magnetic field of 10 mT as a function of wavelength for the same conditions as shown in (a). (d) The MEL at 460 nm (open symbols) and 700 nm (solid symbols) as a function of voltage and electrical stress with the same conditions as shown in (b).

Figure 3

Electroluminescence (EL), magnetoelectroluminescence (MEL), and magnetoconductance (MC) measurements for the blue backbone copolymer (B), the copolymer with green dyes (BG), and the copolymer with green and red dyes (BGR). (a) The normalized EL of the three different devices is shown at 2.8 V. (b) The red dye emission (640 nm) in the BGR polymer shows a strong voltage dependence with respect to the backbone emission, while the green dye emission (520 nm) does not. (c) The MEL of the devices is shown at 2.8 V. (d) EL spectra for the BGR device with varying bias voltage. (e) MC and (f) MEL are shown as a function of voltage. Note that the MEL in (c) and (d) was measured at constant current, while the MEL in (e) was measured at constant voltage as it was measured with a photodiode together with the MC. The MC and MEL are shown for a fixed magnetic field strength of approximately 10 mT.

Figure 4

Schematic overview of the possible spin-selective steps in a trap-assisted recombination process. (i) The electron trapping itself can be spin selective, resulting in a magnetic-field-dependent trapping rate. (ii) Competition between electron-hole recombination and dissociation can be spin selective, resulting in a magnetic-field-dependent (effective) recombination rate. (iii) Spin-selective recombination of electron-hole precursor pairs can result in a magnetic-field-dependent singlet/triplet exciton ratio.

Figure 5

(a) Schematic overview of the spin-dependent trap-occupation mechanism described in the text, where a magnetic field will increase the T/S ratio on the traps. (b) The calculated ratio between SRH recombination ($R_{\text{SRH}}$) and Langevin recombination ($R_{\text{L}}$). (c) The calculated MC and (d) bimolecular MEL, where $P_{\text{S}}$ is reduced from 0.90 to 0.88. The MEL at low voltages cannot be measured in the experiments. Trap densities of $1\times {10}^{-4}$ (squares), $2\times {10}^{-4}$ (circles), and $4\times {10}^{-4}$ ${\mathrm{nm}}^{-3}$ (triangles) were used.

Figure 6

Calculations of the (a) MC and (b) MEL for magnetic-field-dependent recombination (squares) or trapping (circles). We used a relative reduction of 5% in $C_{\text{p}}$ and $C_{\text{n}}$.

Figure 7

Calculations of the (a) MC and (b) MEL for magnetic-field-dependent exciton formation, when an increasing number of a second type of trap is added. Unlike the initial “intrinsic” traps, this second type does not allow for exciton formation, resulting in a reduction of the MFEs. Added trap densities of 0 (squares), $1\times {10}^{-4}$ (circles), $2\times {10}^{-4}$ (upward triangles), and $4\times {10}^{-4}$ ${\mathrm{nm}}^{-3}$ (downward triangles) were used, with a change of $P_{\text{S}}$ from 0.90 to 0.88 and an intrinsic trap density of $2\times {10}^{-4}$ ${\mathrm{nm}}^{-3}$.