Supramolecular interactions have emerged as a viable route towards controlling the properties in organic electronic devices [1, 2]. The latter are based on molecular components whose arrangement in space affects their function and operational efficiency. The reasons for this are multiple and include bulk properties (e.g., anisotropic variations in the dielectric constant due to molecular dipoles) as well as molecular properties resulting from non-uniform distribution of molecular orbitals. The intermolecular orbital overlap is a critical component in determining the charge-carrier mobility of a material and depends on the molecular components and their mutual orientation over long distances. Therefore, controlling the morphology of the active layer down to the molecular level is crucial for bridging the bottom-up molecular and top-down device fabrication scales.
Various strategies exist for directing the self-assembly of small molecules into larger, well-defined architectures. These include metal ion coordination, π-stacking and hydrophobic interactions, electrostatic forces, and hydrogen-bonding (H-B) interactions. Hydrogen-bonding interactions are particularly interesting as they can be combined into arrays that code for specific sequences, such as those found in natural biopolymers like DNA and RNA. Furthermore, unlike many metal coordination complexes, H-B units do not generally absorb in the visible region of the spectrum, making them well adapted for the fabrication of light-triggered devices and for spectroscopic investigations. It is remarkable to note that, even for those compounds that possess a high level of symmetry, such as fullerenes, supramolecular ordering can induce a high degree of anisotropy in the electronic transitions.
Numerous systems containing a photosensitive moiety have been designed to show a measurable response to light irradiation . This can include morphological changes [4, 5], which then affect the size, shape, and physical properties of aggregates, but also the release [6, 7] and capture [8, 9] of substrates.
Supramolecular control of fullerene electronic interactions
Fullerene C60 is well known to pack into a symmetric fcc lattice, in which each fullerene molecule is surrounded by 12 others [10, 11]. This high level of symmetry and the low reorganization energy of fullerene are important in determining the bulk electronic properties of fullerene-containing materials. To study the impact of geometrical effects on fullerene interactions, it is necessary to restrict π-orbital overlap to the fullerene surface, either by covalent or by supramolecular approach [12–14]. With this in mind, we prepared a fullerene-barbituric acid adduct (Compound 1, Fig. 1) that opened up access to a variety of well-defined supramolecular C60 architectures , based on the melamine-cyanuric acid H-B motif. Using H-B interactions between 1 and electron donor moieties such as vinylene-thienylenes possessing complementary H-B motifs (compound 2), it is possible to enhance ground-state electronic interactions to the extent that a strong electronic absorption band appears at ca. 450 nm . The latter is attributed to the formation of a through-space charge-transfer complex between the donor and acceptor units in the supramolecular architecture in view of the ultra-fast (k ~ 5 × 1012 s–1) electron transfer observed upon photoexcitation. Direct irradiation of degassed solutions of 1 in the presence of the complementary triaminotriazine unit leads to the efficient photodimerization of the fullerene cages in solution , thereby providing indirect proof for the formation of linear, H-B ribbons in which the fullerenes are in close contact.
The self-assembly properties inherent to barbituric acid derivatives confine its solubility to more polar environments which break the H-B network it forms in aprotic non-polar media. Unfortunately, fullerenes are already sparingly soluble in non-polar aromatic solvents, and hardly in polar protic media. Thus, in the absence of complementary H-B receptors, the Janus-like solvatophilicity of 1 requires the use of binary solvent systems for solubilization. This is largely incompatible with the fabrication of electronic devices by spin-coating or drop-casting as the sequential evaporation of solvents leads to preferential precipitation of one or some of the components. Therefore, we proceeded to introduce solubilizing groups that would not interfere with the molecular recognition properties of 1. After numerous trials, we were rewarded by compound 3 (Fig. 2) , which crystallizes into linear, H-B ribbons in which the di(t-butyl)benzyl group is oriented distal to the barbituric acid residue and does not interfere with its H-B properties. Despite the strong interest in developing H-B supramolecular fullerene-based materials, there is still no direct evidence as to whether the self-assembly properties of such materials are indeed controlled by the H-B interactions rather than by the strong aggregation properties of fullerene, dominated by entropy loss due to solvation. Using 3, we were able to provide experimental evidence for the effect of H-B onto the electronic properties and to establish that the presence of H-B units is compatible with n-type charge transport in OFET devices . The linear packing of the fullerenes in crystals of 3 is also responsible for a unique photophysical process, the light-induced control of its emission polarization. Indeed, the emission from crystalline samples of 3 originates from ground-state dimers, analogously to that of pristine C60 . This emission is thus red-shifted with respect to fluid solution and, unlike fullerene crystals, it is highly polarized. The polarization of the emission is attributed to the anisotropy of the fullerene electronic interactions in the solid, which favor emission from fullerenes that are in closest contact. In this respect, fullerene emission represents a convenient means of characterizing the directionality of fullerene–fullerene contacts.
Upon continued irradiation (385 nm), the polarized emission from single crystals of 3 undergoes a switch in polarization direction. The new emission is still polarized, but now along a different direction to the initial one, with a rotation of ca. 60°. Interestingly, this switch in polarization is not accompanied by a change in the intensity, energy (i.e., emission spectrum), or lifetime of the emitting species. This is particularly surprising, as the difference in emission polarization is characteristic of a change in the orientation of the electronic transition dipole moment, which is intimately associated with the electronic (and therefore chemical) structure of the emitting chromophore. We have dubbed this behavior photopolism , in analogy to photochromic materials, which undergo a switch in color upon irradiation. To explain this, we propose that, during irradiation, the fullerene excimers may also undergo photoinduced [2 + 2] cyclodimerisation across the 6,6 C=C bonds. Sun and co-workers showed that the emission from C120 photodimers is very similar to that from fullerene C60 and does not bear resemblance to the fluorescence emission from the excimer-like emitting states observed in solid samples of fullerenes . This indicates that photodimerisation affects the electronic interactions between the fullerenes, raising the energy of the photodimer to a level similar to or above that of the locally excited fullerene moiety.
Close inspection of the crystal structure of 3 shows that each fullerene moiety is in close contact with four other fullerenes, two of which are located on each side along the molecular hydrogen-bonded ribbon while the other two belong to a proximal in-plane ribbon (Fig. 2). One may expect that, amongst the three possible excimer-like states that may be formed from these pairs of adjacent fullerenes, one would be lower in energy and act as an energy sink for the excitation energy, thus accounting for the initial polarisation of the fluorescence emission. Slow photoinduced dimerization of this excimer-like state will raise its energy level above that of the other possible excited dimer states, one of which (oriented 60° with respect to the original one) will be populated through exciton hopping in the crystal.
Energy transfer in self-assembled vesicles
Organic materials with tailored functional groups capable of inducing the formation of nanostructures prior to or during deposition represent a route to supramolecular devices via a “bottom-up” approach, in which an ordered structural arrangement with nanometer precision over a large scale is constructed by aggregation . However, the use of long alkyl chains yields aggregates in which the aromatic π-conjugated electroactive units are buried deep within an inert alkane matrix [22, 23]. If the alkyl chains are reduced or omitted, the solubility of the core units is too low to be of any practical use. Vesicles are possibly the most prevalent structure in nature, and artificial vesicles have been explored in view of controlling and compartementalizing chemical processes as well as enhancing energy transfer efficiency.
Bis-urea (biuret) H-B motifs are prone to forming an intramolecular H-B that favors a cyclic conformation exposing two H-B sites that are not self-complementary. These lead to the formation of extended sheets in the solid as evidenced by numerous X-ray crystal structures. By appending biuret H-B motifs to electro- and photoactive π-conjugated units, it is possible to direct their assembly into well-defined vesicle-like architectures in non-protic media (anhydrous THF, 0.1 mM, Fig. 3). Unlike the majority of examples of artificial vesicle, the process does not require the presence of biphasic water-containing solvent mixtures to induce self-assembly. This presents a distinct advantage, as water may be incompatible with the construction of devices for molecular electronics. The mechanism for the formation of vesicles in the absence of water is unclear, but small-angle neutron scattering (SANS) experiments indicate that spherical objects of the same dimensions (ca. 250 nm) as the vesicles observed by SEM are present in solution upon dissolution of 4 . From this, and the identification by TEM of tubular aggregates whose diameter is also similar to that of the vesicles, we propose that 4 forms H-B sheets in the solid whose initial exfoliation upon dissolution releases them into the solvent. There, hydrodynamic forces induce curvature of the sheets, which form the tubular aggregates observed by TEM. Stretching and constriction of the tubular aggregates then releases spherical aggregates whose size is similar to the diameter of the tubes.
Compared to 4, compounds 5 and 6 (Fig. 4) absorb and emit at longer wavelengths due to the presence of charge transfer character in the ground and excited state. They are highly emissive and, because they retain the rigid π-conjugated core and possess the H-B biuret motifs that induce the formation of vesicular aggregates, they are also expected to spontaneously aggregate into spherical architectures. This is indeed the case, as demonstrated by TEM and SEM analysis of drop-cast samples of 5 and 6 in anhydrous THF (0.1 mM). Confocal fluorescence microscopy was used to map the emission spectra of individual aggregates which, along with compound 4, define a gamut of possible colors that can be attained by this tri-component system. Comparison with that of a conventional LCD display shows that the red and blue components are already acceptable, whereas the green component (compound 5) should be shifted hypsochromically by ca. 10–20 nm to improve color definition. Nonetheless, this initial systems is a very promising start towards designing molecular-based emissive devices (OLEDs) whose definition is limited not by the resolution of the top-down vacuum deposition through a mask, but by a bottom-up approach in which the molecular components self-assembles into a small object of precise size and shape. Taking the systems formed by compounds 4–6 as an example, we may estimate the size of each color pixel to be twice the average diameter of an individual vesicle to allow for variations in size. Combining three such vesicles into a volumetric pixel (voxel) gives a dot of a minimum size of ca. 2 µm, which translates into a theoretical display resolution of well over 10 000 dpi.
An interesting aspect of the fabrication of highly emissive nano-spheres is the possibility of generating single point light sources . Indeed, the aggregates formed by 4–6 are smaller than λ / 2, i.e., the limit of diffraction for the light they generate. Furthermore, one is not limited to the spectral region of the emission envelope of a single species, as it should be possible to mix different compounds to produce an emission of a desired color. For this to work, the compounds must be freely miscible within the aggregates, without affecting its spherical morphology. To test this, we proceeded to add increasing amounts of 5 into a solution of 4 in THF (0.1 mM). The aggregates obtained possessed emission spectra (as determined using a calibrated confocal fluorescence microscope) that were intermediate between the emissions of the pure compounds. As expected, only a small amount of energy acceptor (0.5–2 %) is sufficient to observe efficient energy transfer, indicative of a good dispersion of the dopants in the aggregate. Further addition of 6 to the solutions of 4 and 5 led to a deviation of the color towards the red region of the spectrum, in agreement with the presence of additional emission from 6. With this approach, it is therefore possible to target a specific color and to determine a combination of primary emitters 4–6 that will combine into a single aggregate with the desired emission profile. As an example, D65 white light (open triangle) was obtained for a combination of 4 (0.1 mM in THF), 5 (0.20 mol%), and 6 (0.25 mol%).
Charge transport in self-assembled monolayers
As we have seen, the organization of electroactive components is a key parameter to control the properties of molecule-based devices. A key objective is to achieve spatial ordering over multiple length scales (e.g., over several microns with nanometer precision) as this will affect reproducibility and performance. Compared to drop-casting or other deposition techniques, surface grafting can benefit from a high degree of control as it does not rely on or permit changes in the morphology of the active layer after deposition. However, the obtention of dense, well packed monolayers using large, π-conjugated molecules is challenging and requires that the deposition conditions be optimized for every molecule/solvent/substrate combination (temperature, incubation time, etc). This can be long and requires significant amounts of material, which may not be easily available.
To circumvent the difficulties associated with the preparation of functional monolayers of extended π-aromatic structures, we explored an approach that is intermediate between the self-organized SAMs of thiols on gold and the covalent grafting of silanes on SiO2 . It relies on the reversible covalent cycloaddition reaction between fullerene and anthracene . This [4 + 2] Diels–Alder reaction is reversible at room temperature, and does not require or release external reagents. Because the reaction is reversible, it is possible to attain very high degrees of surface modification to produce well-packed monolayers. Although monolayers of fullerene have been previously reported, the extension towards modified C60 derivatives is synthetically challenging as it demands the preparation of multiply-substituted fulleroids [28, 29]. In this respect, although the Diels–Alder reaction is sensitive to steric effects, it is tolerant of different solvents and substitution of the fullerene component, thus allowing a wide variety of fullerene derivatives to be used. The use of anthracene as a fullerene-selective functional tether presents several advantages in that anthracene is generally unreactive towards common dieneophiles and other functional groups, and because anthracene monolayers are easily lithographed through photoinduced oxidation or photodimerization reactions.
Anthracene monolayers on SiO2 were obtained from the corresponding amine-terminated monolayers prepared using well-established procedures (Fig. 5). These were characterized by AFM and spectroscopic ellipsometry prior to incubation with fullerene solution. The methodology for preparing the anthracene monolayers is not important, nor is the substrate used, provided the anthracenes can engage in the Diels–Alder reaction with fullerene. The excess of fullerene in solution ensures complete transformation of the surface-accessible anthracenes and allows correction of surface defects owing to the reversible nature of the reaction. Once the substrate is removed from solution, the reaction stops and the surface does not evolve further. This combination of thermodynamic reversibility coupled with a relatively slow kinetic regime allows rinsing and manipulation of the sample without particular difficulties. The anthracene- and fullerene-terminated monolayers were characterized using typical instrumental techniques which confirmed the formation of a single molecular layer of fullerene upon derivatization of the anthracene monolayers .
To test the applicability of the monolayers in electronic devices, we sought to probe their charge transport properties as the charge carrier mobility of the active layer is one of the most important parameters in determining the properties of a molecular electronic device. However, this is a major challenge as it requires to precisely place metal contacts onto a single molecule layer without damaging it while ensuring good electrical conductivity. For this reason, although the electrical properties of self-assembled monolayers have been studied both experimentally and theoretically, most studies have focused on the electrical conduction along the molecular axis, while measuring intermolecular (lateral) conduction has only been achieved indirectly. Several approaches have been proposed to study the lateral electrical transport in SAMs. Field-effect transistor devices in which the semiconductor channel is a single sheet of molecules formed on the gate dielectric (SAMFETs) are the most common approach and allow the charge carrier mobility to be extracted from the field-dependence of the source-drain current . Conducting atomic force microscopy (C-AFM), is particularly well adapted to probe lateral electrical transport in monolayers as the combination of AFM imaging and C-AFM electrical characterization enables simultaneous investigation of the structure and function of molecular assemblies.
We therefore set out to employ C-AFM to measure the distance- and voltage-dependence of the current between a fixed electrode and a mobile electrode (the AFM tip) placed in direct contact with the monolayer. However, achieving a good contact between a monolayer and the electrodes is critical for electrical measurements. The evaporation of a metal electrode subsequently to the monolayer formation ensures optimum contact, but results in the degradation of the monolayer near the electrode from contamination by metal particles (shadow-effect). The latter can be prevented by growing the monolayer after the deposition of the electrode, but this leads to imperfect monolayer coverage in the close vicinity of the electrode edge resulting from poor local wetting and limits the composition of the monolayer to those whose formation is compatible with both substrate and electrode. We hence developed an alternative approach in which the electrode is evaporated onto the monolayer-coated substrates through a soft polydimethylsiloxane (PDMS) shadow mask placed directly onto the monolayer . Once the fixed metal electrode (Au) is deposited, the PDMS is gently peeled off to reveal the undisturbed monolayer underneath (Fig. 6). Using this technique, sharp electrode edges could be obtained reproducibly, without deterioration of the monolayer or shadow-effects.
The electrical properties of three different monolayers were compared: amide-terminated monolayers (7); anthracene-terminated monolayers (8) and fullerene-terminated monolayers (9). In each case, several samples were tested, and each sample was tested in several places. The advantage of the C-AFM technique is that the location of the metal electrode can be determined with accuracy using the topology feature. Then, the conducting tip is used to gather electrical information [current–voltage curves, I(V)] at different locations on the sample. The use of a gold fixed electrode and a gold-covered tip further ensured that there are no internal electric fields induced by the presence of two interfaces with different workfunction potentials. In the case of 7, the current flowing between the tip and the fixed electrode is below the detection limit of our instrument (1 pA) for all distances until the tip contacts the electrode, at which point the current exceeds the upper limit of the current preamplifier (20 nA). This indicates a completely insulating behavior for lateral charge transport, and is not surprising in view of previous reports on the insulating properties of alkylsilane and alkanethiol SAMs.
The electrical behavior of substrates 8 and 9 contrasts sharply with those of 7. In the case of substrate 8, the current above the threshold limit is detected well before the tip comes in contact with the fixed electrode (Fig. 8). The resistance of the sample is thus found to evolve smoothly from 1014 Ω at a distance of ~ 0.75 µm (corresponding to the detection limit of the instrument) to the saturating value of 106 Ω at the electrode. Between these two extremes, the resistance of the sample decreases monotonically until the tip makes contact with the electrode, at which point the current is limited by the instrumental response. This behavior is clearly indicative of a better medium for lateral charge transport in the anthracene-terminated monolayer and a similarly conductive behavior is also found for the fullerene-terminated monolayer (9). In this case, however, a current above the detection limit is observed at a much greater tip-electrode distance than in the case of sample 8 (~ 1.50 vs. ~ 0.75 µm for 9 and 8, respectively, Fig. 7).
To understand the charge transport properties of the monolayer, it is necessary to examine the I(V) behavior at various tip–electrode distances. For both samples 8 and 9, a quadratic dependence is observed, which is characteristic of a space-charge limited current (SCLC). This can be understood by the accumulation of charges in the vicinity of the electrodes, which then screens the electric potential across the device. There is, however, an additional element that does not fit into this model, which is the current–distance dependence at a fixed potential. The SCLC model predicts a dependence of I(d) that is proportional to d–2 in the case of thin layer devices (eq. 1) .
where I is the current, V the applied voltage, ε and µ are the permittivity of the material and the charge carrier mobility (assumed to be constant), L and W are the length and width of the contacts, and ξ is a coefficient that depends on the shape of the contacts.
In disagreement with the SCLC model, we find an exponential dependence of the current–distance dependence for both samples 8 and 9. Moreover, we note that although the fullerene monolayers are intrinsically ~ 104-times more conducting than the anthracene monolayers, slope of the linear fits is actually quite similar in both cases. This slope can be understood in terms of a β value that is often used to characterize tunnelling junctions. A value of β = 0.016 and 0.013 nm–1 is found for 8 and 9, respectively (at V = 2 V, Fig. 8).
To test whether the contrast in the distance vs. voltage dependencies of the current [I(d) and I(V), respectively] originates in the material or the implementation of a mono-molecular layer as the charge transport region, we investigated a thicker (6 nm thick) anthracene-terminated layer (substrate 10). The results are shown in Fig. 8b, and it can be seen that the same behavior is observed. Namely, an exponential dependence of I(d), while the I(V) curves are quadratic. For this sample, however, a value of β = 0.006 nm–1 is obtained, significantly lower (i.e., better charge transport) than that of substrate 8. Therefore, we can infer that for thick molecular layers (< 6 nm), the thickness of the layer affects the bulk conductivity of the sample, but does not change the fundamental charge transport mechanisms. A possible explanation for the similar value of β is that the fullerene substrates are prepared from the anthracene substrates by cycloaddition of fullerene onto the anthracene residues. Therefore, any patterns or breaks in the conductivity in the anthracene samples would be transferred onto the coverage in the fullerene samples.
It is evident from the above results that supramolecular interactions can be used to design and control how photo- and electroactive compounds organize into well-defined architectures. This strongly affects their properties in that intermolecular electronic interactions depend on the distance and orientation of the chromophores. Controlling this is an important aspect to designing molecular components for electronic devices possessing optimal efficiency, durability, and functionality. Benefitting from self-assembly can lead to more facile manufacturing processes, especially when approaching the resolution limit of conventional lithography. Another aspect is the observation of new phenomena (e.g., photopolism) that arise from the spatial organization of photoactive units. This suggests that it is possible to advance the capabilities of currently available materials by exploring their organization in space.
This work was supported by the ANR (ANR-08-BLAN-0161 and ANR-09-BLAN-0387) and the Region Aquitaine. Financial support from the the National Science Council (Taiwan) is gratefully acknowledged. Part of this work was supported by the LabEx AMADEus (ANR-10-LABX-0042-AMADEUS through grant ANR-10-IDEX-0003-02).
D. M. Bassani, L. Jonusauskaite, A. Lavie-Cambot, N. D. McClenaghan, J. L. Pozzo, D. Ray, G. Vives. Coord. Chem. Rev.254, 2429 (2010).Google Scholar
D. Faure, J. Gravier, T. Labrot, B. Desbat, R. Oda, D. M. Bassani. Chem. Commun. 1167 (2005).Google Scholar
Y. Molard, D. M. Bassani, J. P. Desvergne, N. Moran, J. H. R. Tucker. J. Org. Chem.71, 8523 (2006).Google Scholar
Y. Molard, D. M. Bassani, J. P. Desvergne, P. N. Horton, M. B. Hursthouse, J. H. R. Tucker. Angew. Chem. Int. Ed.44, 1072 (2005).Google Scholar
Y. V. Pol, R. Suau, E. Perez-Inestrosa, D. M. Bassani. Chem. Commun. 1270 (2004).Google Scholar
V. Darcos, K. Griffith, X. Sallenave, J. P. Desvergne, C. Guyard-Duhayon, B. Hasenknopf, D. M. Bassani. Photochem. Photobiol. Sci.2, 1152 (2003).Google Scholar
S. Z. Liu, Y. J. Lu, M. M. Kappes, J. A. Ibers. Science254, 408 (1991).Google Scholar
H. B. Burgi, E. Blanc, D. Schwarzenbach, S. Z. Liu, Y. J. Lu, M. M. Kappes, J. A. Ibers. Angew. Chem., Int. Ed. Engl.31, 640 (1992).Google Scholar
J. J. Gonzalez, S. Gonzalez, E. M. Priego, C. P. Luo, D. M. Guldi, J. de Mendoza, N. Martin. Chem. Commun. 163 (2001).Google Scholar
N. D. McClenaghan, C. Absalon, D. M. Bassani. J. Am. Chem. Soc.125, 13004 (2003).Google Scholar
N. D. McClenaghan, Z. Grote, K. Darriet, M. Zimine, R. M. Williams, L. De Cola, D. M. Bassani. Org. Lett.7, 807 (2005).Google Scholar
C. C. Chu, G. Raffy, D. Ray, A. Del Guerzo, B. Kauffmann, G. Wantz, L. Hirsch, D. M. Bassani. J. Am. Chem. Soc.132, 12717 (2010).Google Scholar
P. M. Pippenger, R. D. Averitt, V. O. Papanyan, P. Nordlander, N. J. Halas. J. Phys. Chem.100, 2854 (1996).Google Scholar
B. Ma, J. E. Riggs, Y.-P. Sun. J. Phys. Chem. B102, 5999 (1998).Google Scholar
J. H. Ryu, D. J. Hong, M. Lee. Chem. Commun. 1043 (2008).Google Scholar
F. J. M. Hoeben, I. O. Shklyarevskiy, M. J. Pouderoijen, H. Engelkamp, L. Schenning, P. C. M. Christianen, J. C. Maan, E. W. Meijer. Angew. Chem., Int. Ed.45, 1232 (2006).Google Scholar
A. Ajayaghosh, C. Vijayakumar, V. K. Praveen, S. S. Babu, R. Varghese. J. Am. Chem. Soc.128, 7174 (2006).Google Scholar
S. K. P. Velu, M. H. Yan, K. P. Tseng, K. T. Wong, D. M. Bassani, P. Terech. Macromolecules46, 1591 (2013).Google Scholar
A. Ajayaghosh, V. K. Praveen, C. Vijayakumar. Chem. Soc. Rev.37, 109 (2008).Google Scholar
T. Hasobe, H. Imahori, S. Fukuzumi, P. V. Kamat. J. Phys. Chem. B107, 12105 (2003).Google Scholar
E. C. P. Smits, S. G. J. Mathijssen, P. A. van Hal, S. Setayesh, T. C. T. Geuns, K. A. H. A. Mutsaers, E. Cantatore, H. J. Wondergem, O. Werzer, R. Resel, M. Kemerink, S. Kirchmeyer, A. M. Muzafarov, S. A. Ponomarenko, B. de Boer, P. W. M. Blom, D. M. de Leeuw. Nature455, 956 (2008).Google Scholar
G. V. Dubacheva, M. Devynck, G. Raffy, L. Hirsch, A. Del Guerzo, D. M. Bassani. Small10, 454 (2014). DOI: 10.1002/smll.201300502 (2013).Google Scholar
A. A. Grinberg, S. Luryi, M. R. Pinto, N. L. Schryer. IEEE Trans. Electron Devices36, 1162 (1989).Google Scholar
About the article
Published Online: 2014-02-14
Published in Print: 2014-04-17