Influence of conductivity and work function of polyaniline based HIL on PLED device performance

 

Joerg R. Posdorfer^a, Bettina Werner^a, Bernhard Wessling^a, Susanne Heun^b, Heinrich Becker^b
aOrmecon GmbH, Ferdinand-Harten-Straße 7, 22949 Ammersbek, Germany^*
bCovion Organic Semiconductors GmbH, Industrial Park Hoechst, G865, 65926 Frankfurt, Germany^#

 ABSTRACT

Polyaniline (PAni) dispersions can be efficiently used as hole injection layers (HIL) for passive and active matrix display applications. In earlier work the influence of conductivity and work function of HILs spin coated from water based PAni/PSS dispersions on device performance had already been presented. Recent investigations on hole transport mechanism in polyaniline systems now show the necessity of a minimum conductivity and an optimum work function for hole injection. Electrochemical Impedance Spectroscopy measurements combined with luminescence investigations showed that the lateral conductivity in the PAni films must be >10^-6 S/cm. Otherwise, a decrease in maximum efficiency and an increase in driving voltage in dependence on coating thickness occurs. Work function investigations on water-free, highly conductive polyaniline dispersions emphasize the theory of an optimum range for hole injection from the anode into the light emitting polymer. The work function of highly conductive, non-aqueous PAni dispersion (0.1-5 S/cm) was determined by Scanning Kelvin Probe method to be 4.5 – 4.7 eV, which is outside of the optimum range at about 4.95 – 5.05 eV for polymeric light emitting diodes, resulting in poor efficiency values (max. 30 – 50 % compared to PAni/PSS standard).

 

Keywords: PLED, PAni, hole injection, conductivity, work function, brightness, efficiency

 

1. INTRODUCTION

Thin film display devices based on organic polymer light emitting diodes (PLEDs) have been available commercially since 1997, ten years after the first low voltage organic thin film electroluminescent devices had been made¹. The PLED technology is widely recognized as a technology for flat panel display products². Apart from their advantages in cost and manufacturing, PLEDs must be able to meet product lifetime specifications in order to become a reliable display/lightning technology. Therefore improvements are required for both, materials in terms of efficiency and operational lifetime and device geometry and preparation methods.

 

The data on polymer light emitting devices presented here were based on an indium-tin-oxide(ITO)|polyaniline/poly (styrenesulfonate)(PAni/PSS)|poly (p-phenylene vinylene) Super Yellow (SY)³|cathode structure for passive matrix applications. In passive matrix displays anode cross-talk loss is caused by the conducting polymer deposited by spin coating over the completely patterned substrate. The hole conducting polymer PAni/PSS therefore connects the anodes electrically with each other. This effect can be reduced by a high lateral resistance of the hole injection layer (HIL). With genarally high resistance (lateral as well as vertical) the driving voltage across the device pixel increases, thereby increasing the power consumption. A key challenge for best power conversion in PLED devices is to reach the built-in voltage which is due to the difference in the work functions of the electrodes before charge carriers are injected. Therefore it is necessary to adjust this difference to the internal band gap of the light emitting polymer^4,5. The main focus of this paper is the adjustment of the lateral conductivity in relation to the work function by varying the ratio of PAni to PSS in the aqueous dispersions. Investigations on the device performance of the PLEDs in terms of maximum efficiency and brightness, operating voltage and lifetime were carried out by Covion Semiconductor GmbH.

 

Furthermore, the possibility of spin coating a hole injecting material from a solvent other than water was tested. As PAni powder synthesized by Ormecon can be dispersed in different media and doped with different acids three non water based dispersions were tested as hole conducting polymers in PLED applications and the results are presented in this paper.

2. EXPERIMENTAL

Polyaniline was chemically synthesized by oxidation with ammonium peroxydisulfate, doped with different acids and dispersed in water, xylene or alcohol. The particle size distribution was measured by Laser Doppler technique with a Microtrac UPA 150 Particle Analyzer.

For all investigations ITO patterned glass substrates of 30x30 mm² were used. The ITO was applied as 140 nm thick films with a surface resistance of 15 W/sq. The water and solvent based polyaniline dispersions were spin coated on flame pretreated surfaces (P-6708 D, Specialty Coating Systems). Deposition was carried out in a spin cycle at a dispense spin rate of 500 rpm and at a spin rate of 3000 rpm. The deposited films were thermally cured at 180 °C for 10 minutes and film thickness of 95 nm ± 5 nm was determined using a stylus profiler (Dektak 8000, Veeco Metrology Group).

Absorption spectra were measured using an Analytic Jena Specord S 100 UV/Vis spectrophotometer.

Conductivity measurements at Ormecon were performed by an Electrochemical Impedance Spectrometer operating in the Fourier transformation mode (FT-EIS model 6416B by G. Popkirov^6,7) with a probe set for high impedance measurements. By this, electrical noise artifacts were minimized. The lateral resistance of the PAni samples was measured as a two-point probe method using the ITO patterned substrates. The measurement was carried out under vacuum between two patterns using gold pins as contacts. A frequency-rich perturbation signal with a small amplitude was applied to the substrate at open circuit potential (E» 0 V) controlled by a potentiostat. A computer programmed sum of 42 sine waves distributed over 4 decades was used to synthesize the perturbation signal of a home-built signal generator. The peak-to-peak amplitude of the perturbation voltage was usually 25 mV. Data analysis was carried out using an equivalent circuit consisting of a series combination of a contact resistance R_c and a parallel RC element for film resistance and capacitance of the cell. Conductivity measurements carried out at Covion Organic Semiconductors were carried out on specially designed substrates with interdigital structures (IDS). Film thickness was in the order of 80 nm. The current was measured for voltages of -10 to +10 V and the resistance was calculated from the slope of the straight lines. As the polyaniline layers are very sensitive to water⁸, resistance was measured under vacuum to avoid the exposure of the films to humidity in all cases.

A scanning Kelvin probe (UBM Messtechnik GmbH) was used to determine the work functions of PAni layers on ITO. The measurements were performed under atmospheric conditions identical to the preparation process of PLED devices. As measurements were not performed in ultrahigh vacuum, all work functions given were related to gold as a reference material with a work function of 4.8 eV under these conditions⁴.

The light emitting devices were prepared by spin coating and curing the HIL PAni/PSS layer onto ITO patterned plates followed by spin coating of the light emitting polymer Super Yellow³. As the cathode material, Ba with a capping layer of Ag for lowering the cathode resistance was deposited by evaporation. The device configuration was thus ITO|80 nm PAni/PSS|80 nm SY|6 nm Ba|100 nm Ag. The devices were characterized by stepwise increasing the applied voltage and measuring the current through- and light output. The measuring equipment was a Keithley 230 voltage source, two Keithley 199 DMM’s to measure voltage as well as current through the LED, and a UDT 265 brightness sensor (a combination of a photo diode, a photometric filter and a lens in front) to measure the brightness. As a standard, the device performance was compared to the results for a similar device where PAni/PSS is replaced by PEDT/PSS (Baytron P 4083 EL, from H. C. Starck,).

 

3. RESULTS AND DISCUSSION

The polymerization of polyaniline yields the non conducting emeraldine base. After purification the emeraldine base can be transferred to the conductive emeraldine salt by protonation with acids. Two different approaches were followed to obtain materials with different conductivity:

1. doping the material with different amounts of the water soluble polymeric material PSS

2. protonating the emeraldine base with low molecular weight dopants for solvent based polyaniline dispersions.

 

3.1 Water based polyaniline dispersions

For water based dispersions with PSS as doping and blend material, an improved purification technology was set up. This resulted in dispersions with ammomium contents of 20 to 60 ppm and sulfate contents of 45 to 50 ppm as determined by HPLC. The conductivity was adjusted by varying the ratio of dopant and polymer at solid contents of about 3.5 weight-%. Six PAni dispersions with varying PSS content were investigated. The physical properties are summarized in Table 1.

 

Table 1: Characteristic data for water based PAni/PSS dispersions.

 

Sample	PAni/PSS composition	Solid content [wt.-%]	Viscosity [cP]	Particle size [nm]	Max. absorbance at ~ 770 nm
1	1 : 1.2	3.42	3.53	99	0.128
2	1 : 1.7	3.37	3.51	90	0.116
3	1 : 3.0	3.49	3.61	59	0.104
4	1 : 3.9	3.57	3.73	79	0.088
5	1 : 6.2	3.36	3.82	48	0.065
6	1 : 7.4	3.65	4.10	46	0.056

 

The ratio of PAni to PSS was in the range of 1:1 to 1:8. The particle size distribution was monitored by changes in the frequency spectrum of scattered laser light due to the Brownian motion of the particles. Particle sizes shown in Table 1 mean that 90% of the numbers of particles are below the given value. The resulting PAni dispersions contain particles with a size of 30 to 100 nm and are stable over time in terms of viscosity and particle size distribution. With an increase

Fig. 1: Absorbance at 770 nm of aqueous PAni/PSS dispersions measured for 95 nm thick layers on glass substrates.

in PSS contents, a decrease in particle size was observed. Due to the higher molecular weight of PSS, an increase in viscosity was observed. For the UV/Vis absorption spectra no shift of the absorption maxima were observed. A plot of the PSS content versus the absorbance of the absorption maximum in the visible region at about 770 nm results in a straight line as shown in Figure 1. From this calibration curve the composition of PAni to PSS can be calculated for aqueous dispersions.

 

Results of conductivity measurements are shown in Table 2. The values from EIS measurements with small AC perturbation signal are compared with those from higher perturbation DC measurements on IDS substrates. Good agreement was achieved between both techniques. The conductivity varies in the range of 10^-2 S/cm to 10^-7 S/cm. As shown in Figure 2, an increasing PSS contents results in a decrease in conductivity of the dried films. The PSS contents in aqueous dispersion required for a desired conductivity for either passive matrix display or active matrix display can be determined from Figure 2.

Table 2: Conductivity and work function for water based PAni/PSS dispersions.

 

Sample	Conductivity by EIS [S/cm]	Conductivity by IDS [S/cm]	Work function f [eV]
1	1.5 E-02	1.2 E-02	4.75
2	2.1 E-03	1.3 E-03	4.81
3	9.0 E-05	6.5 E-05	4.96
4	3.6 E-05	1.4 E-05	5.06
5	2.5 E-06	2.0 E-06	5.00
6	5.3 E-07	8.5 E-07	4.88

 

For strong electrolytes in diluted solutions the conductivity is linear to the concentration. Assuming that spin coated films of PAni/PSS are Ohmic conductors, one should expect a linear relationship between conductivity and rate of dilution, too. Figure 2, however, suggests that there must be another contribution that is responsible for the deviation from linear behavior. From Table 1 it is obvious that the particle size changes significantly upon diluting the PAni/PSS dispersions, which could lead to a conductivity changing difference in film morphology.

Fig. 2: Dependence of conductivity on PAni/PSS composition, O measured by EIS and D measured by IDS.

This was studied by varying the solid contents of a given PAni/PSS dispersion and measuring the resulting viscosity, particle size distribution and conductivity. Table 3 shows that the viscosity, the particle size and the film conductivity all increase with increasing solid contents. Since the doping level should not be affected by changing the solid contents of a dispersion with a constant PAni / PSS ratio, the change in conductivity must be due to another effect. A possible explanation is that the free carrier mobility is hindered in films made of smaller particles by the energy barriers formed at the individual interfaces. Similar results for particle size reduction and loss in conductivity were described for PEDT/PSS by tailoring the PSS contents⁹.

Table 3: Variation in solid contents for PAni/PSS dispersions with a composition of 1:2.

 

Solid content [wt.-%]	Viscosity [cP]	Particle size [nm]	Film thickness [nm]	Conductivity [S/cm]	Work function [eV]
1.82	2.37	55	39	3.5E-04	5.03
2.67	3.02	65	67	5.9E-04	5.03
3.70	3.72	80	106	9.0E-04	5.02
4.78	4.69	93	158	1.4E-03	5.03

 

Data for the work functions of the polyaniline dispersions are summarized in Table 2. Due to the fact that the work function of gold is not well-defined under the conditions of the measurements, only relative data were obtained. It was confirmed that the work function of the reference did not change during the time of the investigations. For the most conductive dispersion, the lowest value for the work function was obtained. In order to improve the device performance, not only the conductivity has to be considered, but the related work function as well.

 

Table 4: Electroluminescence data for devices with the configuration ITO|PAni/PSS|SY|Ba|Ag.

 

Sample	Max. efficiency [cd/A]	E @ 100 cd/m² [V]	Max. brightness [cd/m2]
1	9.6	3.2	13970
2	9.9	3.2	12645
3	10.1	3.3	10630
4	10.2	3.2	5140
5	9.7	3.2	3065
6	7.9	3.9	360
Standard	10.7	3.2	-

 

Electroluminescence data for devices with the configuration ITO|PAni/PSS|SY|Ba|Ag are given in Table 4 for the samples 1 to 6 as HILs. Only sample 6 with a conductivity of 5.3E-07 S/cm showed a significant increase in driving voltage. Figure 3 shows the maximum brightness of the devices for the six dispersions used as hole injection layers.

 

Fig. 3: Dependence of maximum brightness on conductivity of hole injection layer in PLED devices

(o = EIS, D = IDS).

There is a linear decrease in maximum brightness with decreasing conductivity. For values lower than 10^-6 S/cm only poor brightness is observed due to the voltage drop in the device.

 

As shown in Figure 4, a maximum of efficiency was observed for sample 4. Therefore the conductivity cannot be the only parameter being responsible for a good device performance. The work functions of the six samples seem to have a very similar influence on device efficiency. So there must be a correlation between the work function of the hole injection layer and the efficiency. The maximum brightness is controlled by the conductivity. For passive matrix displays a PAni/PSS composition of 1:4 should be used.

Fig. 4: Dependence of efficiency (ÿ ) and work function (D) on conductivity of HIL.

For active matrix displays sample 2 with a PAni/PSS composition of 1:2 represents the optimum. Various PAni/PSS ratios were used by Monkman et al. for an improvement in device performance¹⁰. Here again the more conductive layers formed from dispersions with a composition of 1:1 and 1:2 were more suitable to PLED applications than for 1:5 and 1:10 showing high offset barriers.

For very high PSS contents, the devices became very unstable. DC-measurements beyond 300 cd/m² were not possible for Super Yellow on PAni layers of 80nm thickness. In order to figure out which process might be involved, pulsed IVL curves of Super Yellow devices with different PAni/PSS ratios were measured with increasing pulse width to witness the change from the still relatively good AC- to the very unstable DC-curves. Some results are shown in Figure 5. The PAni/PSS dispersions were compared to the standard in experiments with pulse widths of 100, 500 and 5000 µs and the data for maximum brightness are summarized in Table 5. Again for the samples with lower conductivity a higher efficiency and a lower maximum brightness were observed compared to the highly conductive probes with slightly lower efficiency but significant better brightness.

Table 5: Results for standard Super Yellow devices (80nm Pani/PSS|80nm SuperYellow|6 nm Ba|100 nm Ag).

 

Pani/PSS sample	(100 µs pulses, 8V)	Max. brightness [cd/m2] (500 µs pulses, 8V)	(5000 µs pulses, 8V)
1:1.2	33400	32200	36300
1:1.7	34300	33600	36800
1:3.0	25500	25200	27100
1:6.2	11100	10100	3500
1:7.4	4350	2270	300
Standard	13400	13700	13300

 

The DC-unstable polyaniline dispersions with PSS contents > 4.5 did indeed show the same instability for long pulses, indicating that the duration of the applied field plays an important role.

 

Fig. 5: Efficiency curves in pulsed measurements (100 µs) for Super Yellow devices with different conductivities of the HIL,

driven up to a maximum of 8 V.

3.2 Solvent based polyaniline dispersions

Water is very hard to remove from the hole injection layer and affects the functionality of the cathode. Therefore, a HIL which can be spin coated from another solvent is desirable but on the other hand the light emitting polymer solution must not dissolve the hole injecting material underneath. With water and toluene this is the case, other solvents must meet this requirement as well. Xylene and 2-propanol were used to disperse polyaniline doped with a number of low molecular weight dopants. The solutions were spun onto ITO substrates and cured for 10 min at 180 °C. As the light emitting polymer would be spin coated from toluene, toluene was spun on top of the layers to test the solvent resistance of the HIL. The thickness was measured by UV/Vis absorption and is given in Table 6. It can be seen that the layers were not completely insoluble in toluene.

 

Table 6: Influence of toluene on layer thickness of solvent based systems.

 

Sample	Solvent	Layer thickness after spin coating [nm]	Layer thickness after rinsing with toluene [nm]
Pani / pTS	Xylene	106	99
Pani / pTS + additive	Xylene	102	75
Pani / DBS	Xylene	104	96
Pani / DBS	2-Propanol	92	82

 

pTS = para-Toluenesulfonic acid DBS = Dodecylbenzenesulfonic acid

Some solvent based dispersions were tested as HIL in PLED devices with Super Yellow as the light emitting polymer. The characteristic data for HIL are given in Table 7. The efficiencies of all the devices were significantly lower than for the water based PAni/PSS (sample 2) due to inhomogeneities in the layers and leakage paths. Again both work function and conductivity contributed to the device performance. The DNNS sample with a work function comparable to the water based PAni/PSS samples had a very low conductivity which led to very low efficiency. For the highly conductive samples an increase in work function caused an increase in efficiency. A bad device performance for the small molecular weight dopants camphorsulfonic acid (CSA) and 2-acrylamido-2-methyl-1-propanesulfonic acid (AMPSA) compared to polystyrene sulfonated acid was observed by Monkman et al.¹⁰ also. The poor device stability was attributed to the relatively high mobility of the small counter ions and the lower work functions of the PAni/CSA and PAni/AMPSA layers.

Table 7: Characteristic data for some solvent based polyaniline dispersions.

 

Sample	Solvent	Dielectric constant	Conductivity [S/cm]	Work function [eV]
PAni / DNNS	Xylene	2.4	5 E-05	4.88
PAni / pTS	Xylene	2.4	1.0	4.56
PAni / DBS	2-Propanol	20.1	5 E-04	4.78
PAni / pTS + additive	Xylene	2.4	5	4.38
PAni / PSS	Water	78.5	1.4 E-03	4.96

 

DNNS = Dinonylnaphthalenesulfonic acid

To increase the work function of the HIL to values in the range of 5 eV, different solvents were investigated and a correlation between dielectric constant and work function was obtained. The data are given in Table 8. From the non polar solvent xylene to the polar solvent water, an increase in work function from 4.53 to 4.96 eV occurs. Tests on the improvement of the devices have to be performed.

Table 8: Work function and dielectric constant for highly conductive solvent based polyaniline dispersions.

 

Sample	Solvent	Conductivity [S/cm]	Work function [eV]	Dielectric constant
PAni / pTS	Xylene	5	4.53	2.4
PAni / pTS	Ethyl acetate	0.08	4.55	6.0
PAni / pTS	Octanol	0.1	4.59	10.3
PAni / pTS	tert-Butanol	0.7	4.61	10.9
PAni / pTS	Benzyl alcohol	3	4.74	13.1
PAni / pTS	2-Propanol	2	4.61	20.1
PAni / pTS	Ethanol	10	4.67	24.3
PAni / pTS	Methanol	1	4.69	32.6
PAni / PSS	Water	0.001	4.96	78.5

 

As shown in Figure 6 a linear dependence of work function on the dielectric constant of the solvent of highly conductive polyaniline dispersions was observed.

 

 

Fig. 6: Dependence of work function on dielectric constant of solvent.

As water is the solvent with the highest dielectric constant it is not possible to reach higher work functions than 5 eV by changing the solvent only.

Another known phenomenon is the evolution of sulfate in PSS solutions and dispersions. The presence of sulfate can be responsible for the degradation of light emitting polymers and the decrease in life time of PLEDs. We are looking for another doping material for polyaniline to investigate the effects of the counter ions.

4. CONCLUSIONS

The synthesis of aniline leads to nanosized polyaniline emeraldine base. With PSS as dopant, chemically synthesized polyaniline powder was dispersed in water. The dispersion systems could be easily spin coated onto ITO to serve as hole injection layer in PLED devices. By variation of the PAni/PSS composition and variation of the solid contents from 2 to 5 weight %, the conductivity could be adjusted in the range of six orders of magnitude from 10^-7 to 10^-1 S/cm. With an increasing PSS concentration the conductivity decreased. This is the result of two effects, the dilution with the electrical inert material PSS and the decrease in particle size distribution. Work functions for the water based PAni/PSS dispersions are in the range of 4.75 eV to 5.05 eV. The device performance showed a linear increase in maximum brightness with increasing conductivity. For passive matrix display applications, a maximum in efficiency for Super Yellow as light emitting polymer was received in the range of 5E-06 to 5E-05 S/cm. These dispersions gave layers with a maximum work function of 5 to 5.05 eV. For active matrix display applications with the highly conductive PAni/PSS dispersions, higher currents and efficiencies were achieved for Super Yellow compared to the standard PEDT/PSS system. An explanation for this is the work function which is slightly higher for PAni. Water based PAni/PSS dispersions have a higher work function than organic solvent based polyaniline dispersions. Both current densities and efficiencies are significantly lower than for the water based PAni/PSS HILs. A higher work function of the hole injecting material gives a higher efficiency of yellow light emitting diodes. A linear correlation between work function and dielectric constant of the solvent was obtained. We are currently investigating the possibility of increasing the work function of non-water based PAni dispersions to replace PSS in the water based dispersions.

REFERENCES

1. C. W. Tang and S. A. Van Slyke, Appl. Phys. Lett. , 51, 913-915, 1987.

2. Intertech OLED 2000, San Diego, California, Nov. 3-5, 1999.

3. H. Becker, H. Spreitzer, W. Kreuder, E. Kluge, H. Schenk, I. Parker and Y. Cao, Adv. Mater., 12, p. 42 , 2000.

4. T. M. Brown, J. S. Kim, R. H. Friend, F. Cacialli, R. Daik and W. J. Feast, Appl. Phys. Lett., 75, 1679 - 1681, 1999.

5. G. G. Malliaris, J. R. Salem, P. J. Brock and C. Scott, Phys. Rev. B, 58, 13411 - 13414, 1998.

6. G. S. Popkirov and R. N. Schindler, Rev. Sci. Instrum. , 63, 5366 - 5372, 1992.

7. G. S. Popkirov and R. N. Schindler, Electrochim. Acta, 38, 861 - 867, 1993.

8. J. Posdorfer, B. Werner, B. Wessling, H. Becker, S. Heun, and H. Vestweber, Proc. SPIE , 4800, 115 - 122, 2002.

9. A. Elschner, F. Jonas, S. Kirchmeyer and K. Wussow, Asia Display, Nagoya, Japan, IDW 2001, 1427 - 1430.

10. R. W. T. Higgins, N. A. Zaidi and A. P. Monkman, Adv. Funct. Mater. , 11, 407 - 412, 2001.

Home