Solid-State Electronics 51 (2007) 123–129 www.elsevier.com/locate/sse
Electrical transport properties characterization of PVK (poly N-vinyl carbazole) for electroluminescent devices applications P. D’Angelo a, M. Barra a, A. Cassinese a, M.G. Maglione P. Vacca c, C. Minarini c, A. Rubino d
b,*
,
a
d
CNR-INFM Coherentia and Dipartimento di Scienze Fisiche, Universita` di Napoli ‘‘Federico II’’, Piazzale Tecchio 80, 80125 Napoli, Italy b STMicroelectronics, Front-End Technology and Manufacturing, Advanced R&D – Non Volatile Memories and Derivatives, Post Silicon Technology, Piazzale Enrico Fermi, 1 Localita` Granatello, 80055 Portici (NA), Italy c ENEA C.R. Portici, via Vecchio Macello snc, 80055 Portici (NA), Italy Dipartimento di Ing. dell’Informazione e Ing. Elettrica, Universita` degli Studi di Salerno, Via Ponte don Melillo 1, 84084 Fisciano (SA), Italy Received 18 May 2006; received in revised form 17 October 2006; accepted 8 November 2006 Available online 12 January 2007
The review of this paper was arranged by Prof. Y. Arakawa
Abstract Electrical transport properties of PVK (poly N-vinyl carbazole), a polymer exhibiting an intrinsic electroluminescence (EL) in blue region, have been investigated. In particular, resistivity and charge carrier mobility have been assessed by means of I–V measurements and by Field Effect (FET) doping, respectively. Dielectric constant has been also evaluated by means of C–V measurements performed on an ITO/PVK/Al single layer electroluminescent (EL) device. All electrical characterizations, carried out in different environmental conditions to deeply investigate PVK intrinsic conduction mechanisms, allowed to assess the good quality of spin coated PVK thin films. Electroluminescent measurements performed on ITO/PVK/Al and on ITO/PVK/Alq3/Al have been also performed. Finally, charge injection mechanisms from the electrodes to PVK have been investigated also as a function of the temperature. Experimental results have been compared with some well known theoretical models. 2006 Elsevier Ltd. All rights reserved. PACS: 73.61.Ph; 73.50.h; 73.21.Ac; 73.40.c Keywords: Organic materials; Electrical properties; Electroluminescent devices
1. Introduction The present strong interest about polymers in electronics is widely justified by remarkable advantages in manufacturing electro-optic devices, such as solar cells, PFETs (polymeric field effect transistors) and PLEDs (polymeric light emitting diodes), by these materials. Indeed, polymers can open the way to the development of innovative and low cost devices where flexible, large area and transparent sub-
*
Corresponding author. Tel.: +39 0817769305; fax: +39 0817769332. E-mail address:
[email protected] (M.G. Maglione).
0038-1101/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.sse.2006.11.008
strates can be used, thanks to the easy processing and deposition techniques [1]. In particular, today, many research efforts are focused on PLEDs large area applications [2], such as flat panel displays. In this field, one of the most interesting materials is poly N-vinyl carbazole (PVK). PVK is a hole transport material (hole mobility lh is much greater than electron mobility lel) exhibiting an emission spectrum that, owing to the properties of carbazole groups, covers the entire blue region [3]. For this reason, recently, PVK is being widely utilized in the fabrication of blue light emitting diodes [4,5] in the place of conjugated polymers, which present poor emission in blue region, because of their long p-conjugation and relatively
124
P. D’Angelo et al. / Solid-State Electronics 51 (2007) 123–129
Fig. 1. Molecular structure of PVK.
low fluorescence quantum yields. Furthermore, more recently, bistable resistance behaviour has been observed in multi-layer structures based on PVK films, renewing the interest on this material also for non volatile organic memory applications [6]. Morphologically, PVK is an amorphous hole conduction polymer, constituted by linear chains of repeated molecular groups (H2C–HC)n, with pendant carbazole side groups [(C6H4)2NH], arranged randomly around the same chain (Fig. 1). Such groups act both as chromophores (EL) and hole conducting centres, so that they are mainly responsible for the electrical and electroluminescent properties of PVK. Similarly to what occurs for other polymers, the electrical conduction in PVK is ruled by both field assisted and temperature activated hopping processes, whereas luminescence occurs via radiative decay of a Frenkel exciton. The charge transport in PVK, just as in many other disordered high molecular weight materials, is affected by the presence of impurities that can drastically limit both bulk electrical conduction and charge carrier injection. In particular, the presence of impurities at electrode/polymer interface strongly affects charge injection mechanisms, inducing localized energy states with charge trapping-detrapping phenomena. In this contribution, the main electrical parameters, such as resistivity, charge carrier mobility and dielectric constant of PVK spin coated films have been estimated. By comparing the measured values with the data reported in literature, it has been possible to assess the quality of the obtained films, in order to manufacture prototype PLEDs and characterize their performances. Resistivity and charge injection mechanisms have been also analysed as a function of temperature. Such a characterization is necessary in order to optimise device operation, since optimal efficiency and device lifetime can be obtained only by a deep understanding of charge transport and injection restrictions.
Fig. 2. Schematic view of manufactured field effect devices.
It is worth to mention that gold contacts allow realizing a very good condition for holes injection (barrier height at Au/PVK interface is about 0.3 eV). Charge carrier mobility has been evaluated by fabricating appropriate field effect devices. In this case, PVK films have been deposited on 500 lm thick SrTiO3 (STO) substrates, used as dielectric barriers with er 300 at room temperature. In this case, beyond the top gold contacts (source and drain), a silver one (gate) is deposited on the other side of the STO substrate by sputtering (Fig. 2).1 Electroluminescent devices have been manufactured according to two different structures, both based on commercial glass/ITO substrates. The former structure is based on a single layer configuration, where a spin-coated 80 nm thick PVK single layer is sandwiched between an ITO (indium tin oxide) anode (holes injection) and an aluminium cathode (electrons injection), both deposited by thermal evaporation. The latter is based on a PVK/Alq3 bilayer, where Alq3, a small molecule organic material, acts as electron transport layer (ETL) and PVK serves as hole transport layer (HTL). In this case, 80 nm thick Alq3 film has been evaporated on PVK thin films with the same thickness. For all fabricated devices, after the spinning, PVK films have been annealed in vacuum at 80 C for 10 h. During the experimental investigation, the environmental conditions have been controlled by performing all the measurements by means of a cryogenic probe station, where four metallic probes are mounted on micrometric slide plates. In particular, field effect measurements and I–V characteristics have been carried out by a Keithley 486 picoammeter, whereas EL spectrum has been evaluated by using a CCD Jobin Yvon Mod. 3500 V. Finally, C–V measurements have been performed by means of an LCR meter Agilent 4248A.
2. Samples preparation and experimental setup For the resistivity measurements, PVK films (with a thickness of about 1 lm) have been spin coated over glass substrates at 4000 rpm for 30 s, from a 9.36 · 105 M chlorobenzene solution. Then, two gold contacts have been evaporated on polymer surface, in order to define exactly the size of the investigated region, with length of 100 or 200 lm and width of about 10 mm.
3. Experimental results and discussion As above mentioned, in order to estimate the resistivity value at room temperature, current–voltage (I–V) measure1 For interpretation of references in color, the reader is referred to the web version of this article.
P. D’Angelo et al. / Solid-State Electronics 51 (2007) 123–129
ments have been performed on PVK films spin coated on glass substrates. In such a way, a resistivity of about 109 X cm has been determined, in good agreement with literature data [7,8]. It is important to outline that a linear behavior has been observed only in low applied electric field regime (E < 105 V/cm), while the current voltage dependence increased more than linearly for higher electric fields. This effect is probably the sign of a possible space charge phenomenon, made more complicated by the dependence of charge mobility on the electric field. More investigation is required on this subject. For all the considered samples, the temperature dependence of the resistance has been also analysed to better understand the basic conduction mechanisms. In Fig. 3a, the resistance R as a function of the temperature T for a typical film is reported. As in other amorphous polymers, in PVK, the charge transport occurs by hopping between strongly localized energy states. The hopping phenomenon is thermally assisted and, in general, it is possible to define an unique activation energy D which rules the phenomenon according to the exponential law eD/KT (Ahrrenius behaviour), where K is the Boltzmann’s constant. In our case, as evidenced in the inset of Fig. 3a, where the logarithm of R is plotted as function of 1/T, a simple Ahrrenius behaviour is not valid. Hence, it is not possible to define an unique activation energy, but different D values are to be considered for different temperature ranges. To complete our study, PVK films resistance temperature dependence has been also compared with the results predicted by the Variable Hopping Range model, firstly introduced by Mott in order to describe hopping processes occurring at low temperatures in disordered insulators [9]. In VHR model, conductance in the presence of thermal assisted charge carrier tunneling, expected for amorphous materials with localized states (impurities), is expressed by the formula: " # 1=4 T0 G ¼ G0 exp ð1Þ T
Fig. 3a. Typical PVK film resistance as function of temperature. In the inset, the logarithm of resistance is plotted as a function of 1/T.
125
with KBT0 20a3/g [10], where g is density of states (DOS) at Fermi level and a represent the decay of squared wave function from localized state. As shown in Fig. 3b, the experimental curves are well fitted by Mott model at low temperatures and, assuming a in the range of nanometers [10], the density of states for PVK has been estimated to be about 102 eV1 nm3. Charge carrier mobility has been investigated by field effect FET measurements. In our FET devices, by applying a gate voltage VG and exploiting STO high dielectric constant, the charge density in the polymer channel between drain and source can be controlled, thus modulating the corresponding current IDS for a fixed voltage VDS. In Fig. 4, some trans-characteristics (IDS curves as a function of VG) obtained for different VDS values, are reported. As shown, the majority charge carriers in PVK are holes, since IDS increases for negative VG. Furthermore, a mobility value of 4.8 · 109 cm2 V1 s1 has been estimated using the expression of trans-characteristics in saturation regime (defined for jVDSj P jVGj):
Fig. 3b. Comparison, at low temperatures, between the logarithm of experimental conductance and behavior predicted by VHR Mott’s model.
Fig. 4. Tran-characteristics for a field effect device based on PVK.
126
P. D’Angelo et al. / Solid-State Electronics 51 (2007) 123–129
I D ¼ I DSAT ¼
W C i lðV G V T Þ2 2L
ð2Þ
where W, L and Ci are the channel width, length and the gate capacitance (per unit area), respectively, and VT is the threshold voltage. This value of PVK mobility is in good agreement with those evaluated in literature by means of electrical criterions [8]. In this respect, it should be considered that the charge carrier mobility can be estimated by time-of-flight (ToF) experiments (where the electrical current is photo-induced) too. Usually, in this case, estimated mobility values are higher since the effects of interfaces and the mutual interaction between the charges are negligible. After the electrical characterization and the subsequent assessment of the good quality of the spin coated films, electroluminescent prototype devices have been fabricated and analyzed. Firstly, the electroluminescence spectrum of single layer ITO/PVK/Al devices has been measured. As shown in Fig. 5a, two peaks are clearly observable with the main one falling in the indigo region (411 nm) and the second one in the green region (480 nm). This measurement gives a direct confirmation of the material macromolecular size, since, as expected from literature data [5], the mean emission wavelength of PVK falls into the blue region. However, the brightness performances of this device are, as expected, poor. In this case, low luminescence is evidently related to the high energy barriers present at ITO/ PVK and PVK/Al interfaces (about 1.1 eV and 1.2 eV, respectively), which strongly limit the electrons and holes injection and the consequent recombination rate, directly proportional to the light power emission. It is clear that this condition greatly affects also the current regime flowing into device that, for this reason, can be defined as a ‘‘injection limited’’ [11]. The brightness performances have been noticeably improved in the second considered electroluminescent device, based on the structure ITO/PVK/Alq3/ Al. Indeed, in this case, the device architecture facilitates charge injection and transport, in particular as far as electrons are concerned. Furthermore, since the recombination processes occur in Alq3, the emitted light is green (Fig. 5b)
Fig. 5a. Electroluminescence spectrum of pure PVK.
Fig. 5b. Electroluminescence of PVK-Alq3 double layer device; the emission, in green region, is dominated by Alq3.
[12]. However, a detailed discussion of the performances of this kind of device is out of the scope of this paper and will be presented elsewhere. All fabricated ITO/PVK/Al single layer devices have been characterized by current–voltage (I–V) measurements, also as a function of temperature and in controlled environmental conditions. A typical I–V curve for a single layer electroluminescent device is displayed in Fig. 6. Since, as above mentioned, in this case the current is ‘‘injection limited’’, the measured current density–voltage JV curve has been compared with the results predicted by some well known theoretical models, usually used to describe the charge injection in metals and in semiconductors. In this regard, Richardson–Schottky [13] model (R–S), based on the thermoionic emission, has been considered. In this model, current density depends on temperature and barrier height UB, according to the following expression: UB qV J ¼ A T 2 exp q exp ð3Þ nK B T K BT
Fig. 6. Current–voltage curves for an ITO/PVK/Al device.
P. D’Angelo et al. / Solid-State Electronics 51 (2007) 123–129
127
Fig. 8. Fowler–Nordheim plot in high field region. Fig. 7. Richardson–Schottky model for charge injection.
where A* is Richardson’s constant (12 · 105 A m2 K2]), T is the sample temperature, q is charge of carriers, UB is injection barrier height and n is the ‘‘ideality factor’’ (between 1 and 2 for silicon). By using the fit parameters shown in Fig. 7, a barrier height UB of about 0.84 eV and an ideality factor n greater than 10, apparently without a physical significance, have been extracted. Similar results are reported in literature [14] for evaporated PVK thin film devices. Another model considered in order to describe injection limited currents is the Fowler–Nordheim (F–N) one, which describes the charge injection in the semiconductor as a tunneling process through a triangular energy barrier [15]. In this case, current density is given by the expression: 2 C V BU3=2 J¼ exp ð4Þ UB d V =d where V is the applied voltage, d is the material thickness, C = 2.2q3/8ph and B = 8p(2m*)1/2/2.96hq, where m* is charge carrier effective mass and h is Planck’s constant. If this model is valid, the logarithm of the ratio between the measured J and (V/d)2 should depend linearly on (d/V) (Fowler Nordheim plot). As shown in Fig. 8, in our case, this occurs only for high values of (V/d), so that it is possible to conclude that the tunnelling could affect the injection mechanism only in the high field region, considering however that the breakdown field has been measured to be about 3.5 · 108 V/m. Using FN model, a barrier height of about 0.52 eV has been evaluated. The difference with the ideal barrier height value (ITO–PVK = 1.1 eV) could stem from many adopted approximations (i.e., the use of electron mass instead of hole effective mass). More in general, the found discrepancies between experimental results and the theoretical predictions remark the necessity to adopt more specific models to describe the charge injection in disordered organic systems. In this regard, some interesting numerical approaches have been recently introduced where the hopping phenomena from metals to organics
have been modelled taking into account Gaussian distributions to describe the energetically disordered localized states inside the organic films [16,17]. Furthermore, a capacitance measurement has been also carried out by a LCR meter, applying a sinusoidal signal of 1 V amplitude and 120 Hz frequency to a single layer device and superimposing a VBIAS, ranging from 0 to 25 V. In this way (Fig. 9), the measured capacitance at low VBIAS values, where the device acts simply by a parallel plates capacitor, allowed to estimate PVK dielectric constant (about 3). Increasing VBIAS, this kind of measurement showed a typical maximum that occurs at the bias voltage producing the injection of both majority and minority charge carriers and, consequently, the recombination phenomenon followed by light emission. This means that this type of measurement offers a precise electrical criterion to determine the optical device switching, even without using photo-sensible detectors. In our device, this condition is satisfied at about 20 V [18]. This value is higher than equivalent results reported in literature and, in our opinion, this is to be attributed to a not optimized quality of interface
Fig. 9. Typical CV measurement performed on ITO/PVK/Al device.
128
P. D’Angelo et al. / Solid-State Electronics 51 (2007) 123–129
4. Conclusions
Fig. 10. ITO/PVK/Al device I–V characteristics at different temperatures.
PVK films have been deposited by spin coating techniques and electrically characterized by current–voltage and field effect measurements. The samples show resistivity, holes mobility and dielectric constant of 109 X cm, 4.8 · 109 cm2 V1 s1 and about 3, respectively. All these values are in good agreement with the data reported in the literature. The resistive measurements as a function of temperature are well described by VRH model in low temperature range. Light emitting diodes, based on a single layer of PVK, have been manufactured and tested as a function of temperature; in particular, I–V and impedance measurements have been performed to identify the switch on voltage of the analyzed devices. The experimental results have been compared with some theoretical models, such as Richardson–Schottky and Fowler–Nordheim ones. This analysis has shown that these models underestimate the barrier height at polymer/metal interfaces so that new models are necessary to correctly describe the charge injection in semiconducting polymers. Acknowledgements The technical support of A. Maggio and S. Marrazzo is gratefully acknowledged. References
Fig. 11. Comparison between measured and theoretical Richardson– Schottky model.
between ITO and PVK and the related formation of charge trap states [14]. In order to complete our study, another set of I–V measurements have been performed as a function of temperature. The results in the temperature range from 290 K to 350 K are reported in Fig. 10. All the measurements have been fitted by the Richardson–Schottky model, in order to investigate more deeply the temperature dependence of the charge injection mechanism. [19]. In particular, in Fig. 11, the estimated Richardson–Schottky prefactors J0(T) (normalized at room temperature value) are compared with those obtained from the theoretical model using UB values of 1.1 eV (i.e., expected barrier height value) and of 0.84 eV (i.e., barrier value calculated by the same model at room temperature). Also in this case, it is possible to note a discrepancy between theory and measurements, which basically present a weaker temperature dependence than that predicted by thermoionic emission model. This confirms once again the inability of traditional models to exactly explain the obtained experimental results.
[1] Gustafsson G et al. Flexible light-emitting diodes made from soluble conducting polymers. Nature 1992;357:477–9. [2] Burroughes JH et al. Light-emitting-diodes based on conjugated polymers. Nature 1990;347:539–41. [3] Pope M et al. Electronic processes in organic crystals and polymers. 2nd ed. Oxford University Press; 1993. [4] D’Almeida K et al. Carbazole based electroluminescent devices obtained by vacuum evaporation. Synth Met 2001;122:127–9. [5] Kido J et al. Bright blue electroluminescence from poly(N-vinylcarbazole). Appl Phys Lett 1993;63:2627–9. [6] Sheng Lai Yi et al. Bistable resistance switching of poly(N-vinylcarbazole) films for nonvolatile memory applications. Appl Phys Lett 2005;87:122101–3. [7] Safoula G et al. Electrical conductivity of halogen doped poly (N-vinylcarbazole) thin films. Eur Polym J 2001;37:843–8. [8] Pope M, Swenberg CE. Electronic processes in nanocrystalline materials. Oxford: Clarendon; 1971. [9] Mott NF. Conduction in non-crystalline materials. III. Localized states in a pseudogap and near extremities of conduction and valence bands. Philos Mag 1969;19:835. [10] Bo¨ttger H et al. Hopping conduction in solids. Berlin: AkademieVerlag; 1985. [11] Kao KC. Double injection in solids with non-ohmic contacts. 1. Solids without defects. J Phys D 1984;17:1433–48. [12] Michelotti F et al. Alq3/PVK heterojunction electroluminescent devices. Synth Met 2000;111–112:105–8. [13] Barth S et al. Current injection from a metal to a disordered hopping system. III. Comparison between experiment and Monte Carlo simulation. Phys Rev B 1999;60:8791. [14] Ouro Djobo S et al. PVK deposited by evaporation for light emitting diodes thin films. Synth Met 2001;122:131–3. [15] Parker ID. Carrier tunneling and device characteristics in polymer light-emitting diodes. J Appl Phys 1994;75:1656–66.
P. D’Angelo et al. / Solid-State Electronics 51 (2007) 123–129 [16] Arkhipov V et al. Current injection from a metal to a disordered hopping system. II. Comparison between analytic theory and simulation. Phys Rev B 1999;59:7514–20. [17] Arkipov V et al. Charge injection versus space-charge-limited current in organic light-emitting diodes. Appl Phys Lett 2003;83:5074–6.
129
[18] Shrotriya V et al. Capacitance–voltage characterization of polymer light-emitting diodes. J Appl Phys 2005;97:054054. [19] Van Wourdenbergh T et al. Temperature dependence of the charge injection in poly-dialkoxy-p-phenylene vinylene. Appl Phys Lett 2001;79:1697–9.
