1. Introduction
Two imortant conditions for organic pigments, used in photovoltaic devices, are long range energy transfer and effective charge carrier separation. Phthalocyanines (Pc's) are promising candidates for the active part of organic solar cells, because they exhibit a characteristic structural self-organisation [1], which is reflected in an efficient energy migration in the form of exciton transport. At the same time, the electrical properties of these layers are determined by a p-type doping due to molecular oxygen from air. During the past, the importance of this oxygen doping has been recognized and the effects have been described in numerous papers, e.g [2-5]. However, uncertainty exists about the exact concentration of molecular oxygen inside these layers. As a consequence, the degree of doping is unknown. In this presentation, the results are shown of effusion measurements on thin films of zinc phthalocyanine (ZnPc), yielding the content of molecular oxygen inside the layers. We discuss characteristic properties of the oxygen diffusion process, like characteristic time scales and temperature dependence. From electrical measurements in dark conditions and under illumination, the effect of oxygen on film conductance and range of energy migration is investigated.

2. Experimental
The oxygen diffusion inside ZnPc thin films was investigated using gas effusion and conductivity measurements. ZnPc was evaporated in vacuum (p~10-5 Torr) by resistive heating of a molybdenum crucible. Before deposition, the substrates (silicon for the effusion samples and Corning glass for the conductivity samples) were thoroughly rinsed with isopropanol, deionized water and blown dry with nitrogen. The thickness of the films was measured with a step profiler after deposition and was in the range of 200 nanometer to a few micron.
The effusion measurements took place in a two chamber system, consisting of a sample chamber and a reference chamber, interconnected with a valve. The whole is connected to a turbo pump. A differential pressure gauge measures the pressure difference between the chambers, which builds up upon degassing a sample in the sample chamber at a low background pressure of typically 0.1 mbar. Due to the small dimensions of the system (chamber volumes are < 1 cm³), relatively small quantities of effused molecules can be detected. The detection limit corresponds to approximately 5*1015 effused particles. For the conductivity measurements, silver contacts were evaporated onto the ZnPc films. The samples were placed in a vacuum chamber, where at 10 Volt bias potential the dark current through the electrodes was measured as a function of time after pumping down from 1 bar to 0.1 mbar, approximately the same value of the pressure at the start of the effusion experiments. In order to determine the activation energy for electrical conduction, measurements were performed as a function of increasing as well as decreasing temperature, measured by a thermocouple.

3. Results and discussion
In figure 1, the effusion from a typical, 2.9µm thick ZnPc film is shown. The differential pressure dp on the vertical axis is converted to the number of effused molecules n through the relation dpV 3D nkT, where V represents the volume of the sample chamber minus the volume of the sample, k is the Boltzmann constant and T the absolute temperature. At room temperature, there is a significant effusion from an air saturated sample of more than 2 mbar, while no effusion occurs after exposure to nitrogen. This is an indirect proof of the detection of oxygen. After ca. 20 minutes the remaining oxygen inside the sample has reached an equilibrium distribution. Measurements at elevated temperatures point out that a significant part of the oxygen has remained inside the ZnPc film at room temperature and comes out at elevated temperatures.

Fig. 1: Pressure difference of reference chamber relative to sample chamber as a result of exposure of 2.9 µm ZnPc to ~0.1 mbar pressure.

In figure 1 it is seen that for the ZnPc layer, which had been degassed already two times consecutively at room temperature, leading to poor residual effusion signals, a large effusion takes place when increasing the temperature to 423 K. At around 360 K, saturation takes place, indicating that all oxygen has diffused out of the sample. Apparantly, a large part of the oxygen in the ZnPc is fixed to a certain degree and only comes out of the film in a temperature activated process. The ratio of this fixed oxygen to the mobile fraction effusing at room temperature is about 1:2. From conversion of dp to oxygen molecules, taking into account all effusion steps, it follows that the total concentration of O₂ is (1.7 ± 0.4) x10²⁰ cm^-3. This oxygen content corresponds to one O₂ per ten ZnPc units. The average distance between two oxygen molecules is 1.8 nm, which is much less than the range of energy migration inside ZnPc layers of ca. 30 nm [6,7]. As a result, it is concluded that the molecular oxygen inside Pc layers does not hinder energy migration, and that, regarding the property of oxygen being an efficient quencher of triplet excitons, the energy transfer takes place as singlet excitons.

Fig. 2: Specific conductivity as a function of time of ZnPc layers with variable thickness, during exposure to 0.1 mbar air (starting at t=0) at 296 K.

The specific film conductivity ? was measured as a function of time when exposing ZnPc layers with different thicknesses to a stabilized pressure of ca. 0.1 mbar (see figure 2). As the concentration of free charge carriers is assumed to be proportional to the oxygen concentration in the films, the decrease in conductivity reflects the velocity of the oxygen out-diffusion process [2]. This is confirmed by the fact that the conductivity of ZnPc layers correlates with the effusion behaviour of films with the same thickness. In all films, conductivity dropped over an order of magnitude during the experiment. For thin ZnPc films (< 500 nm), saturation occurred within several minutes, whereas a 7 µm thick film showed electrical degradation during more than half an hour. The diffusion coefficient for O₂ in ZnPc DO₂ was calculated from the ratio of the square of film thickness to the saturation time of conductivity, and, taking the average for all layers, is (4 ± 2) x10^-11 cm²/s at room temperature. This implies oxygen saturation on a minute time scale for typical films used in photovoltaic devices. This explains the observation that after bringing ZnPc single junctions from ultra high vacuum to air, they exhibit a rectifying behaviour within 10 minutes [3]. In order to obtain the fraction of molecular oxygen contributing to electrical film conductance, the thermal activation energy of the ZnPc layers described above was determined under 1 bar air in the temperature range 277 - 355 K. The activation energy, averaged over a number of films, was 0.23 ± 0.02 eV, as determined from the Arrhenius dependence of the conductivity in the temperature range 277 - 325 K. At temperatures above 325 K, conductivity dropped for all films, which can be explained by effusion of oxygen due to thermal annealing, as observed in figure 1.
From the concentration of molecular oxygeninside ZnPc layers and the value for activation energy for electrical conduction, we estimate a high ionized oxygen concentration of ca. 2 x10¹⁶ electron acceptors per cm³. This leads to a space charge layer extending ~40 nm into the ZnPc film from the p-n interface in a typical organic p-n junction as described in [7] and [8]. In such a space charge layer, the electrical field strength might be high enough to result in an effective separation of charge carriers and succesive collection, correlating to the space charge ('box') model as described by Rostalski and Meissner [9] for organic solar cells. So in conclusion, we can say that the two conditions of long range energy transfer and effective charge carrier separation for achieving high efficiency organic solar cells are in principle present for devices incorporating ZnPc layers.

Acknowledgements
We are indebted to R. Heller and R. Giliamse for using the set-up for effusion measurements and for providing the figure of the effusion instrument. We thank the Dutch Agency for Energy and Environment (NOVEM) for financial support.

References

[1]	J. M. Kroon, R. B. M. Koehorst, M. van Dijk, G. M. Sanders and E. J. R. Sudhölter, J. Mater. Chem., 7(4), 1997, 615.
[2]	H. Yasunaga, K. Kojima, H. Yohda and K. Takeya, J. Phys. Soc. Jap., 37 (4), 1974, 1024.
[3]	M. Martin, J.96J. AndrE9 and J. Simon, J. Appl. Phys. 54(5), 1983, 2792.
[4]	H. Laurs and G. Heiland, Thin Solid Films, 149, 1987, 129.
[5]	J.-P. Meyer, D. Schlettwein, D. Wöhrle and N. I. Jaeger, Thin Solid Films, 258, 1995, 317.
[6]	H.R. Kerp and E.E. van Faassen, Physical Chemistry Chemical Physics, in press.
[7]	H.R. Kerp, H. Donker, R. Koehorst, T. Schaafsma and E. E. van Faassen, Chem. Phys. Lett., 298(4-6), 1998, 302.
[8]	H.R. Kerp and E. E. van Faassen, Energy Migration and Oxygen Doping in Phthalocyanine Layers, ECOS'98 conference, 3-5 Dec. 1998, Cadarache, France.
[9]	J. Rostalski, A. Bertram, s. Günster and D. Meissner, The Origin of Limited Photocurrent Efficiencies of Molecular Organic Solar Cells, ECOS'98 conference, 3-5 Dec. 1998, Cadarache, France.

[BACK] [MAIN] [FURTHER]