Javan H.
Cook
a,
José
Santos
b,
Hameed A.
Al-Attar
a,
Martin R.
Bryce
*b and
Andrew P.
Monkman
*a
aDepartment of Physics, Durham University, Durham, DH1 3LE, UK. E-mail: a.p.monkman@durham.ac.uk
bDepartment of Chemistry, Durham University, Durham, DH1 3LE, UK. E-mail: m.r.bryce@durham.ac.uk
First published on 2nd September 2015
Two new deep blue/violet emitting alternating co-polymers, comprising readily-available carbazole (C) and fluorene (F) monomer units, have been synthesised and shown to produce extremely bright solution-processed polymer light-emitting diodes (PLEDs) with the structure ITO/PEDOT:PSS/polymer/TPBi/LiF/Al. The para-conjugated polymer, CF1, gave PLED devices with external quantum efficiency (EQE) values of ηext,max 1.4%, Lmax of 565 cd m−2 with CIEx,y (0.16, 0.07). The EQE was raised to ηext,max 2.1%, after the addition of a TAPC hole injection layer. For the isomeric meta-conjugated polymer, CF2, values of ηext,max 0.35%, Lmax of 16 cd m−2 with CIEx,y (0.18, 0.12) were obtained. The λELmax was 409 nm for both the CF1 and CF2 devices. The CF1 devices also possess low turn-on and low operating voltages for devices of such high brightness. Moreover, the CF1 emission is very stable from 10 cd m−2 up to peak brightness, with only a negligible shift in CIE coordinates. The combination of a simple co-polymer structure synthesised using readily-available monomer units, and high brightness and good colour stability from a simple device architecture, makes CF1 suitable for a wide range of applications requiring deep blue/violet emission.
OLEDs which emit from the deep blue/violet region of the spectrum, especially with Commission Internationale de l'Eclairage (CIE) coordinates that match the National Television System Committee (NTSC) standard blue CIEx,y (0.14, 0.08) have been the subject of increased investigation to meet the demands of high quality displays.12–21 Maximum external quantum efficiencies (EQEs) as high as 3–6% for emission peaks in the range 400–480 nm have been achieved for a range of small molecule based devices through rational molecular design.12,21 However, efficient deep blue polymeric emitters in simple device architectures remain under-developed.11,21 Polymeric systems which possess good solubility in organic solvents offer the advantages of rapid industrial-scale polymer LED (PLED) fabrication by solution processing techniques, such as ink-jet printing, spin-coating and roll-to-roll processing.22 The key challenge in the molecular design of saturated blue emitting polymers is that backbone conjugation should be restricted to short (oligomer) segments, while the charge carrier transport capabilities of the polymer must be retained.23
Poly(9,9-dialkylfluorene-2,7-diyl) derivatives are widely used as blue/deep blue emitting polymers because of their high charge-carrier mobilities, good thermal and electrochemical stability, high photoluminescence quantum yields (PLQY), and versatile chemical modification.24 Specific examples include: poly(3,6-silafluorene-co-2,7-fluorene),25 poly(fluorene-co-thiophene) host–guest systems26 and spiro-polyfluorenes.27 Incorporation of tetrafluoro-p-phenylene units into a polyfluorene backbone affords PLEDs with EQEs as high as 5% with CIE coordinates of (0.16, 0.05).28 It has recently been shown that the attachment of bulky side chains to a polyfluorene backbone increases intrachain torsion angles and isolates minimal conjugated segments.29–31 In an alternative approach, 9,9-diphenylfluorene units serve to disrupt conjugation in the polymer backbone leading to PLEDs with EQE values of ηext,max 3.9%, Lmax of 274 cd m−2 with CIEx,y coordinates (0.17, 0.07).32 It is also established that meta-linked aromatic units in the backbone can blue shift emission by reducing the effective conjugation length.33–35 Sergent et al. have recently synthesised regiorandom fluorene–carbazole copolymers [with emphasis on different length alkyl chains at the fluorene C(9) position]. Very preliminary electroluminescence data showed blue emission (EQE 1.32%; CIEx,y 0.16; 0.11).36
The aim of the present work is to study new readily-available polymers which yield efficient deep blue/violet PLEDs. This has been achieved in two new high triplet energy poly(carbazole-alt-fluorene) co-polymers CF1 and CF2 which function as both the charge carrier transporters and the fluorescent emitters simultaneously. In both polymers the carbazole unit (C) is linked through 3,6-positions, while the linkage to the fluorene unit (F) is varied (2,7 in CF1 – para-conjugated; 3,6 in CF2 – meta-conjugated) in order to modulate the conjugation along the backbone. The photophysics and PLED applications of these co-polymers are reported. In particular, CF1 devices emit deep blue/violet light with remarkably high brightness (Lmax of 565 cd m−2) when the luminosity function is taken into account. The luminosity function shows the perception of brightness to different wavelengths of light, as observed by the human eye and is the function by which OLED emission is normalised. Because the luminosity function peaks at 555 nm, blue and red emission is reduced in comparison to green emission, so more emission is required for these colours to achieve the same brightness. This becomes more of a problem the further the emission is from 555 nm, making the high brightness of CF1 at 409 nm especially notable.
![]() | ||
Fig. 1 Structures of the new deep blue/violet emitting carbazole-alt-fluorene copolymers studied in this work. |
![]() | ||
Fig. 2 Normalised PL emission spectra for polymers CF1 and CF2 (a) in ethyl acetate; (b) in cyclohexane; (c) in thin film. The insets show a magnification of the λmax region. |
Polymer | Solvent/film | λ absmax/nm | λ PLmax/nm | PLQY, ΦPLa |
E
onsetT![]() |
---|---|---|---|---|---|
a PLQY is the photoluminescence quantum yield. b E T is the triplet energy determined for solid state samples at 16 K. | |||||
CF1 | Ethyl acetate | 348 | 399 | ||
Cyclohexane | 344 | 394 | |||
Film | 350 | 421 | 0.20 | 2.34 | |
CF2 | Ethyl acetate | 300 | 400 | ||
Cyclohexane | 304 | 399 | |||
Film | 287 | 401 | 0.17 | 2.34 |
From the photophysical data in Table 2 it can be observed that both CF1 and CF2 emissions peak at ca. 400 nm in solution with negligible solvent polarity effect (Table 2, Fig. 2a and b). Nevertheless, PL from thin films shows a different trend: while CF1 (para-conjugated) shows the typical aggregation induced red-shift (20 nm) in its emission, CF2's (meta-conjugated) λPLmax remains basically unaltered (Table 2 and Fig. 2c).
V on /V | Brt/cd m−2 | EQE/% | Dev eff/cd A−1 | Brtmax/cd m−2 | EQEmax/% | Dev effmax/cd A−1 | Lummax/lm W−1 | CIEx,ye | CIEx,yf | |
---|---|---|---|---|---|---|---|---|---|---|
a Device architecture: ITO/PEDOT:PSS/polymer/TPBi/LiF/Al. b V on is the turn-on voltage, defined here as the voltage at which the device reaches a brightness of 10 cd m−2. c A comparison current density of 10 mA cm−2 was selected. d Data in brackets are for devices with an additional TAPC layer between PEDOT:PSS and CF1 (see text for details). e CIE coordinates at the turn-on voltage (10 cd m−2). f CIE coordinates at the maximum brightness. CIE diagrams are shown in the ESI. | ||||||||||
CF1 | 3.40 | 232c | 1.10c | 0.51c | 565 (492)d | 1.4 (2.1)d | 0.65 | 0.63 | 0.16, 0.07 | 0.16, 0.07 |
CF2 | 4.50 | 16c | 0.05c | 0.04c | 16 | 0.35 | 0.25 | 0.99 | 0.18, 0.12 | 0.20, 0.15 |
No significant difference is observed in the film PLQY for polymers CF1 and CF2, with values of ΦPL 0.20 and 0.17, respectively. The observed triplet level for CF1 and CF2 (EonsetT 2.34 eV) corresponds well to the reported values for poly(9,9-dialkylfluorene)42 and poly(N-alkylcarbazole) derivatives43 (ca. 2.3 eV) and related polymers with limited backbone π-conjugation.29,30,32
The device results are shown in Fig. 3 and 4. Devices containing CF1 display deep blue/violet EL with λmax of 409 nm and maximum EQE of 1.4% and CIEx,y coordinates [turn on (0.16, 0.07) and peak (0.16, 0.07)] exhibiting excellent stability under device operation. CF2 also showed deep blue/violet emission, again with λmax 409 nm, but with a substantially reduced EQE of 0.35% and less impressive colour coordinates [turn on (0.18, 0.12) and peak (0.20, 0.15)] with reduced stability. The CIE coordinates for CF1 are considerably deeper in the blue than data recently reported for other fluorene–carbazole co-polymer PLEDs [(0.16, 0.11) and (0.18, 0.14)].36 This excellent colour stability of the CF1 emission is in contrast to most other deep blue polymers32 for which there is a shift to less blue emission at maximum brightness (as seen here for CF2). CIE diagrams can be found in the ESI.† The results for CF1 compare very favourably with pure PFO devices,28 having a deeper blue λmax (409 nm compared to 420 nm), superior CIE coordinates ((0.16, 0.07) compared to (0.16, 0.11)) and higher EQE and device efficiencies (1.10% and 0.51 cd A−1 compared to 0.56% and 0.20 cd A−1). The emission is also deeper blue than reported recently in very preliminary EL data for regiorandom fluorene–carbazole copolymers (CIEx,y 0.16; 0.11),36 demonstrating the benefits of the well-defined alternating structure of CF1.
![]() | ||
Fig. 3 Plots of (a) J–V curves. (b) Luminance vs. J. (c) EQE vs. J. (d) Device efficiency vs. J for the polymers CF1 and CF2. Inset to (b) shows the low turn-on voltages for the two devices in a plot of luminance vs. V. The device structure is stated in Table 2, footnote a. |
![]() | ||
Fig. 4 Normalised EL emission spectra for polymers CF1 and CF2 devices. Inset shows a magnification of the λmax region. The device structure is stated in Table 2, footnote a. |
Polymer CF1 produced an excellent maximum brightness of 565 cd m−2 for a device with these CIE coordinates. As stated in the introduction, this deep blue spectral region is heavily affected by the luminosity function (the eye sensitivity here is very low) making such a high value particularly notable. The low turn-on voltage (3.40 V) and low potential operating voltage makes this polymer well suited for a wide range of practical applications, including displays1–4 and medical devices: for the latter “blue light therapy” has a range of beneficial effects.45CF2 has a substantially reduced maximum brightness of 16 cd m−2. Also, despite the emission showing a bluer onset compared to CF1 and the same λmax of 409 nm for both devices, the CIE coordinates of CF2 are less blue than CF1 due to broadened emission into the green region which skews the coordinates (Fig. 4).
The meta-linkage through the 3,6-positions of fluorene in CF2 limits the backbone conjugation; this in turn reduces the conductivity which explains the decrease in the current density passing through the CF2 device, compared to CF1. This reduces both the maximum brightness and EQE of the CF2 device, whilst causing no substantial blue shift in EL emission.
To further investigate CF1, devices were fabricated with an additional hole-injection and electron-blocking layer of 1,1-bis[(di-4-tolylamino)phenyl]cyclohexane (TAPC).46 The primary purpose of this layer was to boost device efficiency by preventing electrons from reaching the anode and trapping them in the emissive layer, utilising TAPC's poor electron mobility. Two TAPC thickness were trialled to give the structure: glass|ITO (150 nm)|PEDOT:PSS HIL 1.5 (70 nm)|TAPC (60 nm or 90 nm)|LEP (20 nm)|TPBi (20 nm)|LiF (1 nm)|Al (100 nm). Whilst these TAPC thicknesses are both high relative to the LEP thickness, it was assumed that a portion would be washed away during deposition of the LEP layer leaving a more reasonable thickness behind that could be assessed by ellipsometry. The EQE was raised to ηext,max 2.1%. There was, however, a 25% increase in turn-on voltage and a 13% decrease in the maximum brightness. These were initially believed to be the result of increased device thickness and potential absorption by the TAPC layer, respectively. However, ellipsometry of the multilayer structures returns a combined thickness for the two layers between 19–20 nm. This indicates that the majority of the TAPC layer is removed when the LEP layer is deposited. Despite this, the observed changes in device properties fit with those expected for the addition of an electron blocking layer and so further investigation would be required to determine whether either a very thin film of TAPC remains, or whether the two layers have blended.
The devices were characterised in a calibrated Labsphere LMS-100 integrating sphere, connected to a USB 4000 CCD spectrometer supplied by a 30 μm UV/Vis fibre optic cable, under steady state conditions. Layer thicknesses were measured using a J. A. Woolam VASE Ellipsometer after having been spin coated onto Si/SiO2 substrates. The non-uniformity of the organic layer thicknesses across the samples leads to a 5–10% error in device efficiencies: all measurements were averages from at least four devices.
Footnote |
† Electronic supplementary information (ESI) available: Synthesis and characterisation of CF1 and CF2; absorption spectra; CIE diagrams. See DOI: 10.1039/c5tc02162f |
This journal is © The Royal Society of Chemistry 2015 |