Search

There is abundant experimental and theoretical evidence of the fact that the primary optical excitation in conjugated polymers is a singlet exciton with a binding energy of typically 0.5 eV. In the absence of excess photon energy, free or geminately bound charge carriers (polarons) have to be generated from relaxed excitons via a secondary dissociation process. This process requires further excitation of several tenths of eV in order to compensate for the exciton binding energy. Concomitantly, exciton breaking can occur either in a strong external electric field or via charge transfer to an electron or hole scavenger. Experimentally, these processes are revealed in field- and dopant-assisted photoluminescence quenching, photoinduced optical absorption, and photoconductivity. Experimental and theoretical aspects of these phenomena are reviewed in the present article. (© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

The effect of the morphology on charge‐carrier injection into methyl‐substituted ladder‐type poly(para‐phenylene) (Me‐LPPP) thin films deposited on a Au(111) substrate has been studied by scanning‐tunneling‐microscope‐based spectroscopy. We find that the charge‐carrier injection barrier as well as the single‐particle bandgap, Egsp, of the polymer show significant variations at different locations of the sample surface. Normally, we find that the values of Egspare larger than the optical absorption edge, the energy difference being attributed to the exciton binding energy. In some regions of the sample, however, Egspappears to be close to or below the optical absorption edge, pointing to the effect of aggregates within the polymer film which act as hole‐trapping centers with a depth of a few 100 meV. Density functional calculations are used to elucidate the dependence of the electronic states on the polymer packing density. Our results show that in this polymer morphological inhomogeneities strongly influence the charge carrier injection and transport properties. This points to a common behavior of materials exhibiting a tendency to form aggregates. In addition, the exciton binding energy of Me‐LPPP is determined to be approx. 0.85 eV. Moreover, the comparison between the charge‐injection energy gap and the photocurrent action spectrum indicates that the photoionization threshold is not directly related to the exciton binding energy.

Not Available

We report on the observation of delayed fluorescence (DF) and phosphorescence (P) from films and dilute frozen solutions of various conjugated polymers of the PPP‐type. The materials differ with respect to the rotational freedom along the polymer backbone. Upon pulsed optical excitation into the S1←S0transition of the materials, delayed emission occurs on a time scale of μs to ms in solid films at 80 K. The phosphorescence in dilute frozen solution decays monoexponentially with a radiative lifetime on the order of one second. The data analysis reveals that the DF is caused by recombination of geminate electron hole pairs rather than triplet‐triplet annihilation. This conclusion is supported by investigations of the response of the DF to an applied electric field.

We have investigated a multilayer organic light-emitting diode (OLED) with 1,3,5-tris(N,N-bis-(4,5-methoxy-phenyl)-aminophenyl)-benzene (TAPB) acting as hole transporting layer (HTL) and tris(8-hydroxy-quinolinolato) aluminium (Alq3) as electron transporting layer (ETL). Positive charge carriers in the HTL were detected optically as a function of the applied bias. Furthermore, we investigated the DC-characteristics of current and brightness as well as the onset behaviour of the electroluminescence (EL) as a function of the applied bias. An analytical model is developed to describe the observed carrier concentrations as well as the current–voltage characteristics and the transient EL measurements quantitatively.

AbstractStationary photoconduction in thin films of a cyano-phenylamine-substituted poly-phenylenevinylene, in both sandwich and gap arrangement, has been studied. In order to separate electrode from bulk effects a thin layer of SiOxcould be inserted between the indium-tinoxide (ITO) electrode and the film. Photocurrents were measured as a function of electric field, temperature (100 to 300K), photon energy (2.5 to 5 eV) and polarity of the applied bias. The results indicate that for ho < 3.2 eV the photocurrent is a superposition of bulk ionization involving relaxed singlet excitons and, depending on the direction of the electric field, exciton induced injection of predominantly electrons from the ITO. For ho > 3.2 eV intrinsic photogeneration from hot singlet excitons takes over.

The question how in conjugated polymers neutral optical excitations (excitons) dissociate into a pair of unbound charge carriers has been discussed. The analogous process in molecular crystals will be outlined briefly as well as experiments to delineate the bandgap in crystalline polydiacetylenes and the equivalent quantity in disordered conjugated polymers. Based upon studies on the prompt and delayed fluorescence in a ladder-type poly-para-phenylene under an applied electric field, it will be argued that, due to the spatial extension of the singlet exciton, on-chain excitons are virtually degenerate with off-chain geminate electron–hole pairs at adjacent chains. The exciton binding energy is by and large the coulombic energy of the pair. The question of intrinsic photogeneration of unbound charge carriers will be discussed.

Upon laser excitation into the S1→S0transition, delayed luminescence from methyl-substituted ladder-type poly-para-phenylene has been observed. It varies linearly with light intensity, decays according to a power law, IDF∝t−0.8±0.1, and its spectrum consists of an intrinsic and a defect fluorescence band as well as a phosphorescence component. Since the delayed fluorescence part can be quenched by an electric field after excitation, but not the phosphorescence, it must originate from geminate electron–hole recombination rather than from triplet–triplet annihilation.

By employing time-resolved fluorescence spectroscopy it is shown that the emission spectrum of non-crystalline layers of anthracene, prepared by vacuum sublimation onto a 80 K substrate, consists of three distinct excimer bands. The band maxima are at 21 900 cm-1, 21 500 cm-1 and 19 600 cm-1, the corresponding lifetimes are 10 ns, 35 ns and 170 ns, respectively. The underlying molecular pair conformations, formed during sample preparation, must be metastable in the groundstate.

We report on two different approaches to modify the emission spectra of poly(phenyl-phenylenevinylene). By doping with different amounts of a laser dye, the fluorescence spectra shift to lower energies due to picosecond transfer of the excitations from conjugated segments of the polymer to the acceptor molecules. As an alternative approach we show that Fabry-Perot microcavities can be used to control and enhance the spontaneous emission of a conjugated polymer layer.

Temporal decay of delayed luminescence (DL) of PVK excited by 4.02 eV photons of an excimer laser has been studied in the time domain 10 μs to 20 s at 297 K and 80 K, respectively, and at selected emission wavelengths (388 nm, 418 nm, 500 nm). At 297 K, DL is controlled by the recombination of geminately bound electron-hole pairs because under the high excitation doses employed, triplet-triplet annihilation occurs on a faster time scale. At 80 K fusion among localized triplet states contributes to DL as well. In combination with literature data for different systems studied in the time domain 10 s to 104 s the present results confirm that geminate pair recombination in disordered organic solids obeys a hyperbolic decay law in accord with the predictions of both Monte Carlo simulation and localization theory for systems with built-in energetic disorder.

Pressure dependent reflectivity studies up to 15 kbar near 120 K show that the pressure shift of both components of the doublet near 16000 cm−1is significantly different leading to level crossing near 9 kbar. This demonstrates that both doublet components must be associated with the two different molecular conformations present in the low temperature phase. An abrupt change of dṽ/dpnear 0.6 kbar is tentatively attributed to a pressure-induced change in the conformation of the side groups.

The Monte Carlo technique has been employed to simulate hopping transport through a Gaussian energy distribution of hopping sites located on a cubic lattice having a nearest-neighbour distance of 10-7 cm. Dispersive transport is observed at temperatures where the width of the Gaussian s`> kT and for electric fields where eE x 10-7/s`< 1. The dispersive regime is observed over a considerable range in the average number of jumps and in this range both 〈x〉 and 〈x2〉 can be described by a mobility and a diffusion constant which are proportional to tα-1 and where α is a function of electric field and s`/kT.

Monte Carlo techniques have been employed to study the drift current behaviour in a hopping system where the energy of each site is arranged at random according to a Gaussian distribution with s`= 0·1 eV. We also introduced a deep-level trap at 0·4 eV below the centre of the Gaussian distribution. The initial current is dispersive and the trapping is propertional to exp (const tα) which is the time dependence for the number of new sites visited. With trap emptying, a steady current is achieved in the long-time limit which is governed by CT--1exp [-(ε∞ - εT)/kT], where ε∞ is the average energy of the hopping particles, εT is trap energy and CT is concentration of traps.

This paper re-examines literature data on the charge-carrier mobility in tri-phenyl methane/polycarbonate, triphenyl amine/lexan and PVK: TNF composite systems. In all cases the dependence of μ on both temperature T and electric field E, as well as its absolute magnitude, can be interpreted consistently in terms of hopping across a Gaussian density-of-states distribution of width σ. The exponential dependence of μ on T-2 and E predicted by computer simulation and analytic theory for transport in the long-time limit are recovered. The essential parameters that control μ are μ0= lim μ(T) which is the mobility of the hypothetical crystalline counterpart structure, and σ. The latter quantity increases with increasing polarity of the constituent molecules and is responsible for the variation of μ upon chemical substitution reported for the triphenyl methane polycarbonate system.

Applying the Monte Carlo technique, we have simulated the time decay of a reservoir of geminate electron - hole pairs for one- and three-dimensional lattices where the site energies are subject to a Gaussian distribution of width σ. As a result of disorder the recombination spectra shift towards longer times, broaden, and lose their power-law decay pattern. The average recombination time [τr] scales with exp (σ/T)2. This behaviour reflects dispersive motion of the carriers within their mutual Coulombic well. A relationship is proposed that allows one to predict [τr] once the dispersive parameter α for the transport process is known, independent of the type of distribution of hopping sites. Comparison is made with experimental data for a-Si: H and a-As2S3.

Decay of donor phosphorescence is used as a sensor for time-dependent diffusion of triplet excitons in an amorphous organic solid (benzophenone) where diagonal disorder of the hopping sites prevails. The exciton jump rate is found to follow a tα-1 law. Between 4·2 K and 100 K the dispersion parameter α increases from 0·25 to 0·55. The results are compared with computer simulations based on experimental data for the density-of-states distribution.

Field- and temperature-assisted dissociation of carrier pairs produced by the absorption of radiation has been simulated for perfectly isotropic (3D) and perfectly linear (1D) systems composed of a discrete array of hopping sites, as well as for systems with anisotropic hopping rates parallel and perpendicular to a preferred direction. For the 3D and 1D systems the exact solutions of the Onsager formalism are recovered. In neither case is Poole - Frenkel behaviour observed. The dissociation yield increases with increasing separation a of the hopping sites. The governing parameter is ΔU/kT, where ΔU, which is a function of a, is the difference in potential energy between the sites at which a carrier starts its random walk and the nearest-neighbour site in the direction of the field. Agreement with experimental data for organic solids indicates that the assumption of field-independent hopping rates implicitly contained in both simulations and analytic theory is justified. Anisotropy of the carrier transport coefficient has a marked effect on the zero-field dissociation yield. On the basis of the computer results, an analytical expression is proposed that relates φ(E = 0) and μ|/μ⊥. The effect of finite sample size on the zero-field yield is also treated quantitatively.

This paper deals with the question of how charge carriers in conjugated polymers and their oligomeric counterparts can be photoexcited and how they recombine. It is intimately tied to the question concerning the magnitude of the exciton binding energy ΔEexc. Studies of fluorescence quenching by large electric fields indicate that ΔEexcmust be about 0.4 eV. Geminate e–h pairs formed by field assisted dissociation of singlet excitons act as precursors for photoconduction. Another source of photogenerated charge carriers (holes) is dissociation of excited states at an electrode. It gives rise to a photocurrent action spectrum that is symbatic with absorption. Recombination of charge carriers is in accord with the Langevin formalism implying a mutual capture radius determined by the range of the coulombic potential.

An analytical treatment of quasi-equilibrium carrier transport in disordered materials at moderate electric fields is presented. The transport equation derived has the form of a conventional Fokker-Plank equation with drift, diffusion and field-assisted diffusion terms. General formulae are obtained for drift mobility as well as for the coefficients of conventional and field-assisted diffusion. The temperature dependences of these quantities are considered for different energy distributions of localized states.

By means of Monte Carlo simulation techniques the relaxation of a packet of charge carriers migrating in a Gaussian density of states of width σ has been studied. The relaxation time determines the transition from dispersive to non-dispersive transport seen in a time-of-flight signal. For a 10 μm thick sample under a bias field of 105 V cm-1 the critical disorder parameter σ/kT turns out to be 4·4. Consistency with experimental results from polymers is found.

A previously developed hopping concept for rationalizing charge transport in random organic photoconductors is further elaborated. It rests on the premise that random molecular packing gives rise to energetic disorder that can be accounted for in terms of a Gaussian-shaped density of hopping states (DOS). Computer simulations confirm that an array of randomly positioned dipoles generates a Gaussian DOS whose width is in excellent agreement with experiment. The existence of energetic disorder can explain in a quantitative manner the scaling behavior of the shape of time of flight signals with respect to sample length and electric field as well as the transition to dispersive transport at low temperatures.

The spectrally resolved decay of photoluminescence in PPV - type conjugated polymers has been measured employing fluorescence-up-conversion and Streak-camera techniques. The results are in account with the concept of stochastic excitation transport amoung an energetically disordered manifold of chromophores, identified as subunits of the polymer chain comprising several repeat units. They demonstrate that the primary excitations are of excitonic origin.

Organic light‐emitting diodes based on molecularly doped polymer blends have been fabricated and characterized using donor–acceptor substituted stilbenoids as active dye components for hole transport, electron transport and emission. A Comparison of the electrochemically generated luminescence (ECL), the photoluminescence (PL) and the electroluminescence (EL) of the organic light‐emitting devices (OLEDs) is given.

Previous data on triplet exciton diffusion in high concentrations (>0·l) of naphthalene in durene melts and carrier drift in molecularly doped polymers are re-examined from a percolation point of view. This is a departure from the view that hopping sites in these disordered media can be represented as an ideal gas with the average distance between sites being a continuous function of the concentration and this average distance determining the diffusion rate. Instead, the correlated motion imposed by the random media, determines the rate of diffusion. From numerical modelling the rate of diffusion and drift is found to be proportional to (C - Co)1.8 where C0 is the threshold concentration for percolation. Comparison with experiment shows excellent agreement.

Fluorescence emission from 2-methyltetrahydrofuran glasses containing two different polydiacetylenes has been studied over the range 4–250 K. The spectra are dependent on the thermal history of the glass and reveal the presence of different polymer chain species. The relaxation and transfer of excitation energy within these species has been studied by both steady state and time-resolved measurements. Localisation of the excitation energy in the tail of an inhomogeneously broadened excited state is observed at 4 K. The results are discussed in terms of current models for polydiacetylene chains in solution and the relationship between the species observed in glasses and possible precursor species in the initial solution is described.

A model of dispersive hopping transport is presented. Carrier transport under non-equilibrium conditions is shown to have characteristic features of an adiabatic process. The time dependences of transport characteristics are governed by the time dependence of the density of 'deep' localized states. An adiabatic equation of dispersive transport is derived. The general results are applied to systems having an exponential distribution of localized states.

Using the Monte Carlo technique, the time dependence of the dissociation of a Coulombically bound electron-hole pair has been simulated for both a regular cubic lattice and a one-dimensional linear chain as a function of applied electric field. In contrast to pair recombination, the dissociation kinetics turn out to be strongly affected by the stochastic character of carrier diffusion in the vicinity of the maximum potential. The results are compared with the predictions of an analytical treatment and discussed in terms of experimental realization.

Fluorescence spectra of single and polycrystalline tetracene have been recorded in the temperature range 4.2 to 300 K. They exhibit a 0-0 emission band centered at 18715 ± 50 cm-1independent from temperature and modification. The results are discussed in relevance to absorption data indicating a phase transition.

Due to the low oxidation potential of poly‐(2,5‐thienylenevinylene) (PTV) an ohmic contact for hole injection is established at an indium tin oxide (ITO)/PTV interface. It gives rise to space‐charge‐limited conduction in an ITO/PTV/Al diode. The ohmic nature of the ITO/PTV contact can be exploited to increase hole injection into multilayer organic light emitting diodes in which internal charge accumulation would otherwise lower majority carrier injection by virtue of electric field screening at the anode.

Low-temperature hopping transport in one-dimensional systems is considered under the condition of strong electric fields. Both the upper and the lower tails of the energy distributions of localized carriers can be described by Boltzmann-type functions of energy characterized by an effective temperature that depends on the electrical field and the spatial distribution of neighbouring hopping sites.

We present a simple analytical approach for determining both the mean transport energy of carriers hopping within an exponential distribution of tail states and their mean transit time in the dispersive transport regime. Good agreement with computer simulations is found.

The distribution of nearest neighbours in dilute hopping systems is shown to be determined completely by positional disorder. As a result, there is no interplay between the energy and distance as far as the easiest jump is concerned and the distribution in energy of hopping neighbours parallels the density-of-state distribution function. As a result the dependence of the charge carrier mobility on concentration and temperature can be factorized into terms dependent on concentration and temperature only. These dependences are calculated for the case of a completely random spatial distribution of localized states. The temperature dependence of the transport coefficients in dilute systems is shown to coincide with that in systems with a high concentration of hopping sites.

A model of dispersive hopping recombination is presented. Carrier recombination under a non-equilibrium conditions is shown to have characteristic features of an adiabatic process. Time dependencies of carrier density and transient photocurrent are governed by the time dependence of the density of "deep" localized states. An adiabatic equation of dispersive recombination is derived. The general results are applied to the low-temperature recombination and to systems having an exponential distribution of localized states.

We present Monte Carlo simulation results for time-independent drift and spatial spreading of a packet of charge carriers executing a random walk within a manifold of hopping sites under a bias field. The shape of the density-of-states profile is assumed to be Gaussian (width s`). The ratio eD/μkT, eaual to unity for ordinary diffusion, is found to increase dramatically with increasing disorder and bias field. For s`/kT=3.5 and fields of order 106Vcm-1, eD/μk T reaches values of order 103. The effect accounts for the anomalous spreading of time-of-flight profiles often seen in photocarrier transit time studies on various polymeric systems. The role of disorder in charge-carrier transport is emphasized.

The time dependence of recombining geminate pairs in a hopping system has been simulated as a function of the initial pair separation r0, the anisotropy of the diffusion constant, the dimensionality, the concentration of hopping sites and the final-step recombination velocity. It is found that the decay of the recombination rate R follows a power law, R ∼ t-x, with x ≥ 3/2. Both the exponent x and the initial onset of the recombination rate depend on r0 and the dimensionality. In the time domain of experimental relevance, significant deviations from the asymptotic form of the analytic solution of Hong and Noolandi are found. The introduction of anisotropy in the carrier mobility yields a multiplicity of decay rates. On the other hand, if the final recombination step is hindered by an energy barrier, the recombination kinetics ultimately become exponential.

The transport properties within a cubic lattice consisting of 70 × 30 × 30 hopping sites subject to a Gaussian energy distribution s`=0.1 eV wide, have been simulated by a Monte Carlo technique. Transit pulse shapes, the time evolution of the energy of the hopping carrier and number of new sites visited as a function of time are analysed. It is found that within the time regime of dispersive transport the energy loss per carrier per jump is inversely proportional to the previous number of new sites visited. This is consistent with a tα-1 behaviour of the jump rate. After a time τe, dynamic equilibrium is established leading to Gaussian transport. For disordered molecular solids in which hopping transport dominates and which are characterized by s`/kT < 4, i.e. α > 0.5, τe is found to be less than 10-8 s. This value would prevent detection of the dispersive regime employing conventional electrical circuitry. The mean carrier displacement at t=τe is typically ten lattice planes. In the long-time limit, In μ ∼ -(s`/T)2 is predicted.

The energy relaxation of optical excitations in poly(p-phenyl-phenylenevinylene) (PPPV) and PPPV blended in polycarbonate (PC) is investigated by means of picosecond time-resolved luminescence spectroscopy. In PPPV a pronounced dependence on the photon energy of the excitation laser is found. For excitation into the tail states of the inhomogeneously broadened density of states distribution, the spectral shape of the luminescence remains almost constant with time. However, a transient red-shift is observed for excitation at higher energies. Experiments on the diluted system PPPV / PC allow us to distinguish between intra-and inter-chain processes. The experimental results can be explained with a hopping model.

Motivated by the recent work of Santos Lemus and Hirsch and of Pai, Yanus and Stolka on the concentration dependence of hole transport in selected molecularly doped polymers, a Monte Carlo computer study has been conducted. It predicts that transport should in principle exhibit percolative features without, however, showing a well-defined percolation threshold ("pseudo-percolation"). A threshold concentration c0 for pseudo-percolation will be defined. In a van der Waals solid with exponential wavefunction overlap it decreases with decreasing wavefunction localization because of non-nearest-neighbour jumps. A relation between c0 and the effective coordination number of the lattice, defined via overlap, is established. For c < c0 the carrier mobility remains finite and follows approximately a power law, μ∝cm, in accord with experiment. The exponent m increases with increasing localization of the wavefunction which depends on the energetic separation between the transport levels of dopant and host.

In order to study the bulk-conductivity in organic liquids measurements with blocking quartzelectrodes were made. When applying a constant field, a bulk-current through the dielectric liquid can be observed, which is decreasing exponentially

The activation energy of the d. c. conductance of organic liquids lies between 0.04 and 0.45 eV in the lower region of temperature of their liquid state. A comparison of these values with the static dielectric constant shows, that the activation energy may be regarded as a pure COULOMB energy: E2= e2/2 ε r . The characteristic distance r has the approximate value of 8.5 Å for hydrocarbons. It decreases for halogen- and nitro-derivates. Formerly it was found that the conductivity of mixtures obeys the law σM= σAC· σB1-C. This can easily be explained assuming εM= c εA+ (1 — c) εB. In the case of rather different ε values or of homologuous compounds forming complexes, σ increases. This is identical with a kink in the log σ (c) -curve.

In preceeding papers a model for the generation of charge carriers in organic liquids was developed. The activation energy of the dc-conductivity specific for one compound was said to be the sum of a resonance and a COULOMB part: E1= E+½3 ET(the factor ½ is due to bimolecular recombination) . Measurements with aza-compounds show that E1is decreasing linearily with increasing calculated π-electron-density at the N-atom. This fact argues for electrons to be the majority chargecarriers. E′ is the energy necessary to localize one electron at the N-atom for a short time, ETis the further separation-energy. With charge-transfer-complexes formed by acridine with aromatic hydrocarbons, E′ increases proportional to the electron-affinity of the hydrocarbons. For energetic reasons it can be assumed that the separated electron moves as a solvatised one through the liquid. An extrapolation shows that the proposed model holds for unsubstituted aromatic hydrocarbons too, but not for aliphatic ones. This is confirmed by experiments. Furthermore, arguments will be developed to distinguish between the conductivity specific for one compound and the conductivity due to impurities. In certain cases the impurity concentration can be estimated from the σ(1/T) curve.

Intense irradiation of pure molten naphthalene, pyrene, and anthracene produces a photocurrent if an external dc voltage is applied. From the spectral response curve, which resembles closely the optical absorption curve in the region of the forbidden S0–T1transition, we conclude that the photocurrent is due to excitation of the lowest triplet state. The excited state can produce charge carriers either when hitting a gold anode or by bimolecular annihilation within the bulk of the liquid. Evidence for cation formation at the anode is presented by observation of space-charge-limited currents and by the absence of this effect if the anode consists of nickel coated with an evaporated SiO2layer. From the rise and decay time of the photocurrent the lifetime of the excited states can be determined. The values are 200 ± 100 ms for naphthalene and 40 ± 10 ms for pyrene. This lifetime decreases with decreasing spacing of the electrodes, indicating diffusion of the excited state to the electrodes where deactivation occurs. The diffusion coefficient of the excited state in naphthalene is (6 ± 2) × 10−2cm2/s. This value and the estimated quantum yield for charge injection rules out any energy transfer involving molecular diffusion.

Non-destructive electric breakdown through mylar films is associated with emission of an intense light flash. Within an active volume of about 4×10-8cm3, L ≧ 109photons are generated during a current pulse. L is proportional to the number of transported charge carriers (electrons) and independent from sample thickness. The emission spectrum is a superposition of ordinary mylar fluorescence centered at 370 nm and impurity bands at longer wavelengths. Interpretation is given in terms of acceleration of injected electrons within the positive cathodic spacecharge zone giving rise to both impact excitation and ionization, predominantly involving guest molecules.

Bei höheren Temperaturen mißt man für organische Lösungen eine Aktivierungsenergie E1der Gleichstromleitfähigkeit, die charakteristisch für den gelösten Stoff ist. Es wird festgestellt, daß eine Substitution am Benzolring von Nitrobenzol und Pyridin eine Erhöhung von E1zur Folge hat. Der Effekt kann nicht durch eine Substituenten-bedingte Ladungsverschiebung innerhalb des Moleküls erklärt werden. Dagegen zeigt eine Abschätzung, daß die zusätzliche Aktivierungsenergie betragsmäßig genau mit der zusätzlichen potentiellen Energie übereinstimmt, die ein abzulösendes Elektron im Feld des Substituenten-Dipols besitzt. Dieser Befund bestätigt das früher aufgestellte Modell der Ladungsträgererzeugung, das als Ort der Elektronenablösung die Sauerstoffatome der NO2-Gruppe bei Nitrobenzolen bzw. das Stickstoffatom bei Azaverbindungen postuliert.

At higher temperatures organic liquids show an intrinsic conductivity with an activation energy E1which is characteristic for each substance. With carbonyl compounds a linear relation between E1and the wave number v of the infrared vibration of the C=O group is observed. Therefore a linear decrease of E1with increasing portion of the polar valence-band structure C+⃝—C-⃝within the carbonyl-bond can be assumed. Within one molecule this polar structure forms a “short-living ion-pair” which can be excited thermally out of the resonance state by the energy E′. If ECoulis the COULOMB energy necessary to separate the electron from the “ion-pair”, the total energy required to form a charge-carrier is: E1= E′ + ½ ECoul. Charge-transfer-complexes, always present as impurities, enable a similar formation of charge-carriers. In this case it holds E′=0 and one obtains the activation energy for the low temperature conductance of organic liquids: E2=e2/2 ε r. The model explains the activation energy without involving relatively high optical molecular levels.