Application of [Pt(II)(Tetra- Tert -Butylsalophen)] Complex within Organic Devices: Deep Red Emission, Bistable Light-Emitting Diodes and Operational Stability

: This paper focuses on the Negative Differential Resistance (NDR) we observed on organic light-emitting diodes (OLEDs) using [Pt(II)(tetra- tert -butylSalophen)] as host, since this Pt(II) complex displays a deep-red emission ( λ max = 660 nm). Electrical characterizations of monolayer devices have shown that doping Tris-(8-hydroxyquinoline)aluminum (Alq 3 ) as matrix emissive layer with this complex, leads to the modulation of the charge transport properties highlighted by Negative Differential Resistance (NDR). Upon electrical driving stresses, the conductivity of active layer can be switched between two electrical states (ON and OFF) with a ﬁgure of merit higher than 10 3 . By adding an electron-blocking layer, we demonstrated that the NDR trend is closely related to negative charge accumulation within Alq 3 leading to the modiﬁcation of electronic properties in the vicinity of anode/active layer interface. The NDR phenomenon is interpreted in terms of space charge polarization (SCP) linked to charge trapping/untrapping mechanism as a consequence of the polarization/depolarization of the Pt(II) complex. Under electrical driving stresses, the performance of the devices which include the Pt(II) complex, are stabilized. A schematic model is proposed to depict the SCP responsible for NDR and decrease-resetting behaviors observed in these devices.


Introduction
Metal organic complexes have been investigated and have shown promising optoelectronic features for energy conversion such as electroluminescence or photoelectricity [1]. Phosphorescent complexes including d 6 or d 8 metal surrounded by π-conjugated ligands have been used for optimization of Organic Light-Emitting Diodes (OLEDs) [2,3]. Among these ligands, Schiff base ligands are attractive because their structure can be easily modified [4]. By using such phosphorescent dyes in host fluorescent matrix, both singlet and triplet excited states participate to light emission of electroluminescent devices [5,6] and a 100% internal quantum yield can be theoretically achieved.
Depending on the device structure and operating mode, the lifetimes of OLEDs produced in laboratories can reach ten thousand hours [7][8][9][10][11][12]. Nevertheless, degradation mechanisms occur and display through a modification of the emission color and/or a reduction of the driving current and the luminance. The effect of aging leads to a progressive increase in impedance during the operation of OLEDs and in order to maintain a constant current, the bias polarization has to be raised more and more [13]. However, it was observed that half lifetime (the time for which the decrease in luminance is 50% from the initial value) and the initial decay are closely related to the amount of non-emissive triplets in emissive layer [9,10]. Thus, among the degradation mechanisms that should be controlled, the management of the non-emissive excitons in the active layer seems to be a key factor in aging of devices. It was observed that doping an emissive layer with a metal complex affects the charge transport properties of emissive layer and device stability [7,14]. In this work, we focused on the study of [Pt(II)(tetra-tert-butylSalophen)] used as phosphorescent dye. In fact, this complex displays a deep-red emission (λ max = 660 nm). We present the optical and electrical features of Pt(II) complex based organic devices obtained by Physical Vacuum Deposition (PVD).
The paper is organized as follows: we first discuss the optical features in a host-guest system of [Pt(II)(tetra-tert-butylSalophen)] embedded within a Tris-(8-hydroxyquinoline)aluminum (Alq 3 ) layer [15]. Then, we analyze the electrical features of Pt(II) complex based devices. Finally, the stability measurements are studied and discussed in relation with the electrical features of the Pt(II) complex based layer. Space Charge Polarization (SCP) mechanism is proposed for the interpretation of the electrical characteristics of the devices.

Materials and Methods
[Pt(II)(tetra-tert-butylsalophen)] was prepared as previously reported [15]. The devices were fabricated by successive vacuum depositions (<10 −5 mbar) of the organic layers and a cathode on a glass substrate covered by a semitransparent electrode. Patterned Indium Tin Oxide (ITO) coated glass substrates from Kintec (100 nm and sheet resistivity about 14-18 Ω/sq at room temperature) were used as a transparent anode. A 20-nm thick of Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate layer (PEDOT:PSS) (CLEVIOS PJET HKV2 charge GSD from Heraeus) was spin coated on the top of ITO glass substrates and was annealed at 100 • C for 10 min. PEDOT:PSS was used as hole injection layer. Thin layers of N,N -Di(1-naphthyl)-N,N -diphenyl-(1,1 -biphenyl)-4,4 -diamine (NPD) and Alq 3 (sublimated grade purchased from Kintec) were used as the hole (HTL) and electron (ETL) transporting layers respectively. A 150-nm thick calcium layer (99.99% from Aldrich) was used as the cathode. The doping of emissive layer with 5% of Pt(II) complex was performed by co-evaporation of Alq 3 as host material at a rate of 1 Å/s and the Pt(II) complex labeled [Pt] at a rate of 0.055 Å/s. The thickness of organic films was determined in situ by quartz-crystal monitor. The thickness and the homogeneity of the resulting films were controlled by profilometry (TENCOR P2 stylus profilometer) and AFM measurements (Dimension TM 3100 Atomic Force Microscope (Digital Instruments)). Monolayer device structures were: ITO/PEDOT:PSS/emissive layer (EML)/Ca. With Alq 3 doped with 5% [Pt] as EML. In bilayer devices, NPD was added between the ITO/PEDOT:PSS anode and EML. The active area of the devices was 18.3 mm 2 . Schematic and energy band diagrams with the energy levels of the OLEDs are illustrated in Figure 1a,b, respectively. The The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) levels were previously estimated from cyclic voltammetry measurements [15].
Characterizations of the OLEDs were obtained under inert atmosphere. Current-luminancevoltage (J-L-V) measurements were carried out by using current-voltage source (Keithley, model 2450). The electroluminescence (EL) spectra and the luminance of the OLEDs were measured by using a Spectroradiometer (JETI SpecBos 1201). For stability measurements, the current and the luminance of devices were recorded every 30 s at a constant driving voltage of 8 V. The stability of the following structures were studied: ITO/PEDOT/NPD(55 nm)/Alq 3 (85 nm)/Ca and ITO/PEDOT/ NPD(55 nm)/Alq 3 :5% [Pt] (85 nm)/Ca labeled as non-doped and doped devices respectively.

Host-Guest System
Based on C. Che
[Pt] exhibits absorption bands in the range of 400-600 nm attributed to metal-ligand-ligand charge transfer (MLLCT). After excitation at λ ex = 390 nm, an emission band was observed between 575 nm and 750 nm (λ max = 660 nm) ( Figure 2). Alq 3 emission spectrum overlaps with the absorption spectrum of [Pt] and should favor an efficient energy transfer from Alq 3 to [Pt].
The doping rate of the emissive layer plays a key role in the devices efficiency and color emission. It was demonstrated that the device efficiency increased with a concentration of platinum (II) dopant, up to 5 wt % or 6 wt % depending on the platinum(II) complex [16]. By considering a 6 wt % concentration of [Pt] (with Alq 3 number density~10 21 cm −3 [17]) and by assuming that the Pt-complexes do not aggregate, the number density of Pt-complex is estimated at about 6.10 19 [Pt]/cm 3 and the average distance between neighboring complexes at~3 nm. Thereby, the effective distance over which the energy transfers between complexes (estimated at 3 nm) would be higher than the radius of triplet-triplet annihilation (<1 nm through Dexter mechanism) [18][19][20]. Furthermore, the steric hindrance between adjacent molecules induced by the tert-butyl substituents of [Pt] tends to limit the formation of [Pt] aggregates and hence reduces aggregation-induced phosphorescence quenching in thin solid film [15,16]. Consequently, doping rate of [Pt] is fixed at 6 wt % to maximize the performance of the devices.

Host-Guest System
Based on C. Che    The doping rate of the emissive layer plays a key role in the efficiency and the color emission of devices. It was demonstrated that the device efficiency increased with a concentration of platinum (II) dopant, up to 5 wt % or 6 wt % depending on the platinum(II) complex [16]. By considering a 6 wt % concentration of [Pt] (with Alq3 number density ~10 21 cm −3 [17]) and by assuming that the Ptcomplexes do not aggregate, the number density of Pt-complex is estimated at about 6.10 19 [Pt]/cm 3 and the average distance between neighboring complexes at ~ 3 nm. Thereby, the effective distance over which the energy transfers between complexes (estimated at 3 nm) would be higher than the radius of triplet-triplet annihilation (<1 nm through Dexter mechanism) [18][19][20]. Furthermore, the steric hindrance between adjacent molecules induced by the tert-butyl substituents of [Pt] tends to limit the formation of [Pt] aggregates and hence reduces aggregation-induced phosphorescence quenching in thin solid film [15,16]. Consequently, doping rate of [Pt] is fixed at 6 wt % to maximize the performance of the devices.
The host-guest confinement methodology was used to modulate and to improve the electroluminescent properties of the devices (Figure 3). The monolayer device exhibited a deep red emission band attributed to the [Pt], whereas the green emission of Alq3 was vanished, implying an efficient energy transfer from the host to the guest. In our previous investigation [15], it was observed that during OLED polarization, some of charges were directly captured by the dye to form an exciton. HOMO/LUMO levels of [Pt] (estimated at −5.84 eV and −3.48 eV respectively from electrochemistry measurements) are within the electronic gap of Alq3 matrix (with HOMO level: −5.98 eV and LUMO level: −3.13 eV); involving that exciton formation, arising from charge transfer mechanism, would be confined on [Pt] ( Figure 3) by a charge transfer (CT) process. The emission process can be described Normalized absorption (in black) and emission spectra (in red) of [Pt(II)(tetra-tert-butylsalophen)] in chloroform; and normalized emission spectra of Alq 3 (in green) and of an Alq 3 :[Pt] (in orange) based OLEDs.
The host-guest confinement methodology was used to modulate and to improve the electroluminescent properties of the devices (Figure 3). The monolayer device exhibited a deep red emission band attributed to the [Pt], whereas the green emission of Alq 3 was vanished, implying an efficient energy transfer from the host to the guest. In our previous investigation [15], it was observed that during OLED polarization, some of charges were directly captured by the dye to form an exciton. HOMO/LUMO levels of [Pt] (estimated at −5.84 eV and −3.48 eV (Figure 1) respectively from electrochemistry measurements) are within the electronic gap of Alq 3 matrix (with HOMO level: −5.98 eV and LUMO level: −3.13 eV); involving that exciton formation, arising from charge transfer mechanism, would be confined on [Pt] ( Figure 3) by a charge transfer (CT) process. The emission process can be described by (i) migration of charge carriers (holes and electrons) within the host layer (ii) CT from the host to the guest (iii) exciton formation on [Pt] (iv) radiative recombination of triplet states.

Negative Differential Resistance
The effect of [Pt] on the charge transport in the Alq3 layer was investigated from J-V characteristics in a monolayer device ( Figure 4a). After performing sweeping voltage from 0 V to +14 V, the current abruptly increased in the range of 0 to 5 V then decreased in the range of 5 to 8 V. The peak of current is obtained for an applied voltage of 5 V. This shape of J-V is characteristic of negative differential resistance (NDR). It was only observed on [Pt] doped devices. By performing ten successive sweeping cycles (positive poling), the repeatable NDR behavior was observed with an increase in current peak. The intensity of this peak achieves a constant value after ten successive sweeping cycles (as shown in Figures S1-S3 of Supplementary Information). By performing negative poling (ten successive sweeping from 0 V to −14 V, as shown in Figures S2 and S3 of Supplementary Information), J-V characteristics recovered the initial shape, indicating that the electrical switching behavior is related to a reversible mechanism. At low voltage (below 8 V), the current could be switched between two electrical states: high-resistance state (OFF state) and low-resistance state (ON state). Thus, the current can be identified in three parts: (i) bistable area; (ii) Negative Differential Resistance (NDR) area and (iii) monostable area. An ON/OFF current ratio higher than 10 3 is achieved in the bistable area.
The origin of NDR features of devices could be assigned to the formation of electrically active defects introduced by the [Pt] within the Alq3 layer. Similar behavior was observed in devices based on organic layer doped by metal nano-clusters or fluorescent dyes [21][22][23][24][25]. It was proposed that charge accumulation within Alq3 layer would play a key role in the bistable trend and NDR properties [21]. At low voltage (<5 V), the switching conductance of the active layer can be due to localized electronic states leading to electron like-traps. The filling of these trapping states leads to the formation of Alq3 + -[Pt] − dipole resulting in space charge polarization (SCP) of active layer. The SCP can modify charge injection properties at the electrode/active layer interface; promoting conductive pathways or space charge field inhibition of charge injections [26].
The J-V characteristics carried out after positive poling show a transition in conduction mechanism from injection limited (OFF) to bulk limited (ON) suggesting the variation in the electricfield distribution between OFF and ON states. Due to the distribution of the LUMO levels (estimated at −3.3 eV and 3.48 eV for Alq3 and [Pt] respectively), the interface with the cathode is expected to be ohmic contact for electrons ( Figure 1b). Therefore, the asymmetrical J-V characteristics observed for OFF state could be explained by an energetic barrier at the interface (with the anode). That should result in injection-limited current in the device for negative applied voltages. After the positive poling, the charges are expected to be stored in the active layer due to the trap like behavior of [Pt] within Alq3 matrix. Consequently, the trapped charges within active layer should be compensated by charges from the metal electrode. A portion of the compensated charges could lead to band bending ( Figure 4b) and that would likely lower the energetic barrier for electrons at the interface with the anode [27,28]. As a result, the current in the device increases to form the ON state in both polarity of the applied voltage (between −5 V and +5 V). It can be assumed that the injection properties of Anode/Alq3:[Pt] would be modified by charges trapped within Alq3:[Pt] layer.

Negative Differential Resistance
The effect of [Pt] on the charge transport in the Alq 3 layer was investigated from J-V characteristics in a monolayer device (Figure 4a). After performing sweeping voltage from 0 V to +14 V, the current abruptly increased in the range of 0 to 5 V then decreased in the range of 5 to 8 V. The peak of current is obtained for an applied voltage of 5 V. This shape of J-V is characteristic of negative differential resistance (NDR). It was only observed on [Pt] doped devices. By performing ten successive sweeping cycles (positive poling), the repeatable NDR behavior was observed with an increase in current peak. The intensity of this peak achieves a constant value after ten successive sweeping cycles (as shown in Figures S1-S3 of Supplementary Information). By performing negative poling (ten successive sweeping from 0 V to −14 V, as shown in Figures S2 and S3 of Supplementary Information), J-V characteristics recovered the initial shape, indicating that the electrical switching behavior is related to a reversible mechanism. At low voltage (below 8 V), the current could be switched between two electrical states: high-resistance state (OFF state) and low-resistance state (ON state). Thus, the current can be identified in three parts: (i) bistable area; (ii) Negative Differential Resistance (NDR) area and (iii) monostable area. An ON/OFF current ratio higher than 10 3 is achieved in the bistable area.
The origin of NDR features of devices could be assigned to the formation of electrically active defects introduced by the [Pt] within the Alq 3 layer. Similar behavior was observed in devices based on organic layer doped by metal nano-clusters or fluorescent dyes [21][22][23][24][25]. It was proposed that charge accumulation within Alq 3 layer would play a key role in the bistable trend and NDR properties [21]. At low voltage (<5 V), the switching conductance of the active layer can be due to localized electronic states leading to electron like-traps. The filling of these trapping states leads to the formation of Alq 3 + -[Pt] − dipole resulting in space charge polarization (SCP) of active layer. The SCP can modify charge injection properties at the electrode/active layer interface; promoting conductive pathways or space charge field inhibition of charge injections [26]. The J-V characteristics carried out after positive poling show a transition in conduction mechanism from injection limited (OFF) to bulk limited (ON) suggesting a variation in the electric-field distribution between OFF and ON states. Due to the distribution of the LUMO levels (estimated at −3.13 eV and −3.48 eV for Alq 3 and [Pt] respectively), the interface with the cathode is expected to be as ohmic contact for electrons ( Figure 1b). Therefore, the asymmetrical J-V characteristics observed for OFF state could be explained by an energetic barrier at the interface (with the anode). That should result in injection-limited current in the device for negative applied voltages. After the positive poling, the charges are expected to be stored in the active layer due to the trap like behavior of [Pt] within Alq 3 matrix. Consequently, the trapped charges within active layer should be compensated by charges from the metal electrode. A portion of the compensated charges could lead to band bending ( Figure 4b) and that would likely lower the energetic barrier for electrons at the interface with the anode [27,28]. As a result, the current in the device increases to form the ON state in both polarity of the applied voltage   The monolayer devices displayed poor luminance performance (below 10 Cd/m 2 ) compared to bilayer structures with the NPD layer (inset of Figure 5). Bilayer devices with 10 nm and 55 nm thicknesses of NPD exhibited a luminance of 15 Cd/m 2 and 35 Cd/m 2 at 14 V respectively. The intensity of NDR is decreasing with the enhancement of NPD thickness ( Figure 5). For monolayer devices, the stored charges on trapping states lead to the modification of both charge injection and conduction. For bilayer devices, NPD acts as electron blocking layer due to its lower value of LUMO level [29] and Alq 3 as electron transport layer [30]. It can be inferred that the thin layer of NPD limits electron extraction from Alq 3 to the anode and promotes hole injection. Consequently, the amount of stored negative charges is reduced due to hole injection leading to the decrease in NDR peak. That is confirmed by the enhancement in luminance of devices with NPD layer (i.e., recombination of trapped electrons with holes). The monolayer devices displayed poor luminance performance (below 10 Cd/m 2 ) compared to bilayer structures with the NPD layer (in the inset of Figure 5). Bilayer devices with 10 nm and 55 nm thicknesses of NPD exhibited a luminance of 15 Cd/m 2 and 35 Cd/m 2 at 14 V respectively. The intensity of NDR is decreasing with the enhancement of NPD thickness ( Figure 5). For monolayer devices, the stored charges on trapping states lead to the modification of both charge injection and conduction. For bilayer devices, NPD acts as electron blocking layer due to its lower value of LUMO level [29] and Alq3 as electron transport layer [30]. It can be inferred that the thin layer of NPD limits electron extraction from Alq3 to the anode and promotes hole injection. Consequently, the amount of stored negative charges is reduced due to hole injection leading to the decrease in NDR peak. That is confirmed by the enhancement in luminance of devices with NPD layer (i.e., recombination of trapped electrons with holes). The NDR mechanism in organic devices is still discussed and depends on the device structures and/or the nature of the contacts [15,21,31]. Further electrical analyses are needed to identify clearly the main physical mechanism responsible for the NDR characteristic of [Pt] based devices. Nonetheless, the measurements reported in this work suggests that the NDR behavior in Alq3:5% [Pt] could be related to electron accumulation within Alq3 layer. Preliminary measurements have been carried out in order to evaluate retention time of space-charges ( Figure S4 of Supplementary  Information). It is observed that there is no obvious degradation for both the JON and JOFF in the retention time of 1 × 10 4 s, indicating the potential application in resistive memory device. Although their luminance performance should be improved, the electrical switching behavior of this type of OLEDs could likely open the development of potential applications such as light memory (which would combine the dual functions of the organic light emitting diodes and the organic bistable memory devices).

Stability Investigations
The degradation behavior under constant current or voltage driving can be influenced by several independent degradation mechanisms [14,32]. The lifetime curve could be interpreted from the analytical models of decays, which are related to several different sets of physical mechanisms. The NDR mechanism in organic devices is still discussed and depends on the device structures and/or the nature of the contacts [15,21,31]. Further electrical analyses are needed to identify clearly the main physical mechanism responsible for the NDR characteristic of [Pt] based devices. Nonetheless, the measurements reported in this work suggests that the NDR behavior in Alq 3 :5% [Pt] could be related to electron accumulation within Alq 3 layer. Preliminary measurements have been carried out in order to evaluate retention time of space-charges ( Figure S4 of Supplementary Information). It is observed that there is no obvious degradation for both the J ON and J OFF in the retention time of 1 × 10 4 s, indicating the potential application in resistive memory device. Although their luminance performance should be improved, the electrical switching behavior of this type of OLEDs could likely open the development of potential applications such as light memory (which would combine the dual functions of the organic light emitting diodes and the organic bistable memory devices).

Stability Investigations
The degradation behavior under constant current or voltage driving can be influenced by several independent degradation mechanisms [14,32]. The lifetime curve could be interpreted from the analytical models of decays, which are related to several different sets of physical mechanisms. During OLED operation, the current and the luminance tend to decrease exponentially and this suggests that two main degradation mechanisms may occur: • A fast initial decrease (with a rate of the order of few hours) which is one of the characteristics of space charge effects as trapping-untrapping/dipole reorientation [14,32,33] • A slow decrease (with a rate of the order of few hundred hours), which can be related to electro-chemical mechanisms (likely molecule dissociation) and/or chemical species diffusion within the OLED layers [14,34].
The lifetime measurements were performed in order to evaluate the effect of [Pt] on devices stability under a constant driving voltage. The study of degradation was performed on doped and non-doped bilayers OLEDs ( Figure 6). We observed: (i) a rapid decay in the initial stage of operation that may be related to SCP mechanism; (ii) Long-time decay related to electro-chemical mechanisms (likely molecule dissociation) and/or chemical species diffusion. We noticed that J-V characteristic measured on doped devices exhibited decrease-resetting zones.
The luminance and the operating current behavior of a device were recorded in the dark at constant voltage of 8 V and at ambient temperature. The driving voltage was fixed above the turn-on voltage (estimated to be in the range 4 to 5 V) and below the voltage at which the decrease in quantum efficiency occurs (estimated at 10 V). In high driving condition, the active layer was probably completely filled with excited triplet states, favoring triplet-triplet annihilation leading to roll-off of phosphorescent device. During OLED operation, the current and the luminance tend to decrease exponentially and this suggests that two main degradation mechanisms may occur:  A fast initial decrease (with a rate of the order of few hours) which is one of the characteristics of space charge effects as trapping-untrapping/dipole reorientation [14,32,33]  A slow decrease (with a rate of the order of few hundred hours), which can be related to electrochemical mechanisms (likely molecule dissociation) and/or chemical species diffusion within the OLED layers [14,34].
The lifetime measurements were performed in order to evaluate the effect of [Pt] on devices stability under a constant driving voltage. The study of degradation was performed on doped and non-doped bilayers OLEDs ( Figure 6). We observed: (i) a rapid decay in the initial stage of operation that may be related to SCP mechanism; (ii) Long-time decay related to electro-chemical mechanisms (likely molecule dissociation) and/or chemical species diffusion. We noticed that J-V characteristic measured on doped devices exhibited decrease-resetting zones.
The luminance and the operating current behavior of a device were recorded in the dark at constant voltage of 8 V and at ambient temperature. The driving voltage was fixed above the turn-on voltage (estimated to be in the range 4 to 5 V) and below the voltage at which the decrease in quantum efficiency occurs (estimated at 10 V). In high driving condition, the active layer was probably completely filled with excited triplet states, favoring triplet-triplet annihilation leading to roll-off of phosphorescent device. The stretched exponential decay (SED), given by Equation (1), can be used to fit the lifetime curve obtained under different driving conditions [14,30,33]. (1) in which τ is the decay time, and β the dispersion factor.
As the parameter β reflects the different initial luminance drop behavior for different device structures, i.e., materials and layer design, β is nearly constant for different current densities measured on the same structure. Some investigations performed on disordered semiconducting materials suggest that the analyze of the decay rate could provide information on density of trap states and trap release rates [35][36][37][38]. In this model based on hopping mechanism, it was suggested that β might depend on the spatial distribution of localized states within the band gap of the semiconductor. The exponential decrease might be explained by the reduction of the number of emissive centers over time, caused by the flow of charge carriers [32,39]. SED analysis is a convenient tool for describing relaxation processes in disorder molecular systems allowing a qualitative analysis of OLED aging. Although a direct relation with physical parameters cannot be clearly established The stretched exponential decay (SED), given by Equation (1), can be used to fit the lifetime curve obtained under different driving conditions [14,30,33].
in which τ is the decay time, and β the dispersion factor. As the parameter β reflects the different initial luminance drop behavior for different device structures, i.e., materials and layer design, β is nearly constant for different current densities measured on the same structure. Some investigations performed on disordered semiconducting materials suggest that the analyze of the decay rate could provide information on density of trap states and trap release rates [35][36][37][38]. In this model based on hopping mechanism, it was suggested that β might depend on the spatial distribution of localized states within the band gap of the semiconductor. The exponential decrease might be explained by the reduction of the number of emissive centers over time, caused by the flow of charge carriers [32,39]. SED analysis is a convenient tool for describing relaxation processes in disorder molecular systems allowing a qualitative analysis of OLED aging. Although a direct relation with physical parameters cannot be clearly established from this analysis, the decay rate could nonetheless give information on the dynamic of physical mechanisms involved in OLED aging.

Analysis of Characteristics Measured on Non-Doped Devices
The stretched exponential fit traced the normalized decay curves accurately. The parameters are summarized in Table 1. On non-doped device, values for β = 0.59 and 1.1 with a decay time of τ = 8.5 h and 500 h were applied to the fast and long-time decays, leading to a very good agreement of the SED fit with the measured data ( Figure 6). The values of the parameter β are in agreement with those obtained in Alq 3 based diodes [32,40]. We can assume that the charge transport would become dispersive during the operating time and is likely due to the creation of defects in organic devices [8]. The fast decay (occurring at the initial stage) could be attributed to the creation of the internal electrical field (due to SCP) likely caused by the presence of traps, diffusion of mobile ions, or reverse orientation of the dipoles [14,32]. It was suggested that the long-time exponential decay of the luminance, results from a degradation model governed by electrochemical reactions [14,32]. Consequently, it can be assumed that, during the long-time decay, the degradation mechanism would be related to the formation of unstable cationic Alq 3 species [32,41].
During, the operation of diodes, Joule effect resulting from charge flow can lead to thermal degradation of OLEDs [42]. In addition to the effect of oxygen and the role of the temperature in the degradation process of organic devices, the charge flow during the device operation plays a role in the aging mechanisms. By applying a voltage, an internal polarization field is induced, which tends to oppose to the external field, which causes a decrease of the drive current. Several models are proposed to explain the mechanisms that cause polarization field. The first one is the trapping model [43] in which traps can be present and/or created in the organic layers. The second model is based on the permanent dipoles: the external field induces the reorientation of the permanent dipoles and can lead to the increase and/or decrease of internal field in the device [14,33]. The third model is the redistribution of mobile ions in active layers under the external bias: the diffusion of mobile ions leads to electrochemical processes at the electrode/organic interface [44]. Other investigations infer that the charge transport in the active layer of devices causes degradation of the organic film [45,46]. Parker et al. suggested that the gradual degradation of organic film is related to the passage of electrons through polymer film [45]. They assumed that the microscopic origin would be some form of redox instability in the materials. However, investigations on Alq 3 based diodes suggest that the degradation of these devices is primarily caused by the injection of holes and the subsequent formation of the unstable cationic Alq 3 species in vicinity of the hole transport layer [14,44]. It should be noted that the lifetime measurements were performed under electrical stresses and inert atmosphere. Oxidation of the organic materials can thus be excluded. SED suggests that the charge transport would become dispersive during the operating time. The OLED degradation could be explained in term of dipole formation and/or trap-controlled mechanism leading to the annihilation of the emission centers.

Doping Effect on the Device Stability
The stability measurements showed that [Pt] within the emissive layer would stabilize OLEDs upon operation (decrease of the degradation rate at τ = 1500 h) whereas the higher decrease rate of the initial decay (τ = 8.5 h) assumes that the effect of space charge on charge transport would be more pronounced in doped devices than in non-doped devices.
A decrease-resetting behavior of current was observed (Figures 6 and 7a) upon electrical stress. Three portions can be identified: (i) the current decreases from 88.5 A to 85 A with a rate of 0.130 h −1 between 7 h and 15 h; (ii) the current reaches a minimum value of 84 A after 16 h; and (iii) the current increases with a rate of 0.138 h −1 between 17 h and 25 h. This behavior would be related to the modulation of SCP within the active layer due to the polarization/ depolarization mechanisms of [Pt] [14,47]. This relaxation mechanism could be explained by the internal electronic field model based on diffusion of ionic species and dipole reorientation/charge trapping [14]. The NDR behavior observed in Alq 3 :[Pt] based devices is substantiated with polarization/depolarization assumption. A possible explanation is the formation of preferred current pathways in devices induced by the orientation of [Pt] − -Alq 3 + dipoles. Similar decrease-resetting behavior of current has already been observed in Pt based OLED by H. von Seggern et al. [9]. It was explained by the formation of preferred current pathways in devices leading to hot spots under electrical driving measurement. It can be assumed that similar effect might occur in Alq 3 :[Pt] based devices. However, the origin of this possible conductive pathway is not clearly established. The electron pathways within Alq 3 layer would lead to hot spots resulting in local increases of temperature as a consequence the higher electron conductivity. On Figure 7b, the lifetime curves measured at 30 • C and 80 • C highlight the influence of the temperature on the decrease-resetting characteristic, suggesting that it was related to thermally activated mechanisms.
The following stages, depicted in Figure  layer and at the vicinity of NPD/Alq 3 interface) would induce internal field (E int ) opposite to applied field, which tends to reduce electron mobility within Alq 3 layer. (ii) Formation of conductive pathways-The preferred orientation of [Pt] − -Alq 3 + dipoles would tend to create preferential electron pathways within Alq 3 layer favoring local electron charge transport. (iii) Thermal energy release-The current resetting could be explained by electron-phonon and/or electron-electron scattering in conductive pathways [48][49][50]. The local heating would induce thermal assisted mechanisms: depolarization of [Pt] − -Alq 3 + dipoles (i.e., thermally activated relaxation of electron captured by [Pt]) and thermal assisted injection of electron from the Alq 3 layer to the NPD layer.
The phenomena responsible for device aging could not be clearly attributed to a single process. The analysis of the lifetime decay of the devices can give a qualitative interpretation of physical mechanisms: formation of SCP and electrochemical reactions. Further investigations are required to elucidate clearly the origin of decrease-resetting behavior of the lifetime curve of [Pt] based devices. Nevertheless, this specific behavior could be related to SCP effects responsible for NDR mechanisms in [Pt] embedded Alq 3 layer. Although the origin of NDR is not well understood in organic devices, our study has shown that this mechanism would be closely related to negative charge accumulation leading to the modulation of the electronic properties at anode/active layer interface; thus, a schematic model to comprehensively depict the SCP mechanisms responsible for NDR and decrease-resetting behavior observed in [Pt] based devices, was proposed.

Conclusions
[Pt(II)(di-tert-butylsalophen)] was used as dopant in Alq3 matrix and produced deep red OLEDs (λmax = 660 nm). Monolayer based OLEDs showed a NDR characteristic with a figure of merit higher than 10 3 , promoting them as potential candidates in memory device applications. NDR mechanism is interpreted in term of space charge polarization linked to charge trapping/untrapping mechanism. By adding electron-blocking layer of NPD, NDR can be amplitude-modulated up to its annihilation. The results suggest that NDR mechanism is closely related to trapped negative charges that would modify the electronic properties at the interface between the emissive layer and the anode. The stability of OLEDs was evaluated under electrical stress. The SED model can be used to fit the decay lifetime by assuming at least two degradation mechanisms likely related to a SCP effect and an electrochemical process. The rapid decay observed in the initial stage is more pronounced in [Pt] based devices, which would be due to space charge effects (trapping-detrapping/dipole reorientation) induced by [Pt] embedded within Alq3 layer. Doping active layer with [Pt] decreased the rate of the long-term decay. Therefore, the results suggest that [Pt] stabilize Alq3 layer under electrical stress. Moreover, relaxation phenomena specific to [Pt] based devices were observed and might be related to the physical process governing the NDR mechanisms. Thus, the specific characteristic can be interpreted in terms of internal electric field. This study can help to elucidate NDR mechanisms in organic molecular devices.

Conclusions
[Pt(II)(di-tert-butylsalophen)] was used as dopant in Alq 3 matrix and produced deep red OLEDs (λmax = 660 nm). Monolayer based OLEDs showed a NDR characteristic with a figure of merit higher than 10 3 , promoting them as potential candidates in memory device applications. NDR mechanism is interpreted in terms of space charge polarization linked to charge trapping/untrapping mechanism. By adding electron-blocking layer of NPD, NDR can be amplitude-modulated up to its annihilation. These results suggest that NDR mechanism is closely related to trapped negative charges that would modify the electronic properties at the interface between the emissive layer and the anode. The stability of OLEDs was evaluated under electrical stress. The SED model can be used to fit the decay lifetime by assuming at least two degradation mechanisms likely related to a SCP effect and an electrochemical process. The rapid decay observed in the initial stage is more pronounced in [Pt] based devices, which would be due to space charge effects (trapping-detrapping/dipole reorientation) induced by [Pt] embedded within Alq 3 layer. Doping active layer with [Pt] decreased the rate of the long-term decay. Therefore, the results suggest that [Pt] stabilize Alq 3 layer under electrical stress. Moreover, relaxation phenomena specific to [Pt] based devices were observed and might be related to the physical process governing the NDR mechanisms. Thus, the specific characteristic can be interpreted in terms of internal electric field. This study can help to elucidate NDR mechanisms in organic molecular devices.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3417/8/5/762/s1, Figure S1: Evolution of NDR peak after several positive sweeping cycles, Figure S2: Evolution of NDR peak after several negative sweeping cycles, Figure S3: Variation the current density versus the number of sweeping voltage cycles, Figure S4: Retention time of both ON and OFF states measured at 5 V.
Author Contributions: M.L. and B.B. conceived and designed the experiments; B.B. and A.C. performed the experiments; C.R., I.S. and F.A. analyzed and interpreted the data; B.B. and M.L. contributed materials/analysis tools; C.R. and I.S. wrote the paper. G.Z. contributed to the redaction of the paper.