Enhanced Field Emission from Diamond Coated Molybdenum Emitters

The field emission characteristics of individual Mo emitters were measured before and after diamond deposition, as well as before and after annealing. After coating without annealing, the emission current increased dramatically. Subsequent annealing resulted in further increases in emission and improved current stability. After coating, the energy distribution of the emitted electrons was found to shift with applied field. Annealing reduced the peak shift but resulted in the appearance of a second peak, coincident with the formation of Mo,C at the diamond-molybdenum interface. These changes in the emission spectra in conjunction with I-V data are discussed in terms of possible changes in emission mechanism. A nwnber of investigators have reported that coating silicon and metal substrates with several types of diamond have led to significant improvements in field electron emission 11-51. We have found that Mo emitters coated with diamond powders deposited by dielectrophoresis give enhancement in emission comparable to other methods [6]. The emission current was found to increase by more than an order of magnitude compared with the same uncoated emitter. Several possible mechanisms for this improvement have been suggested. Zhirnov et al. [4] suggested that emission from protrusions on the diamond surfaces enhances the local field strength. Huang et al. [8] suggest that electrons originated from a subband in the energy gap of the diamond. Geis et al. [3] proposed that a roughened metal substrate enhances the local field at the diamondmetal interface. Xu et al. suggested that electrons are transferred through graphitic channels within the diamond. Latham [7] proposed a MIV (Metal-Insulator-Vacuum) model for insulators in which he assumed a high defect density and high doping levels from impurities. He suggested that electrons could then tunnel into the conduction band of the insulator through a thin depletion layer and subsequently be heated by the penetrating electric field and transported through the barrier at the insulator-vacuum surface. Even if the diamondvacuum surface has a negative electron affinity, the diamondmetal interface remains a barrier to electron transport. Clearly it is essential to reduce the interfacial barrier if one is to achieve high electron emissivity. Two methods can be considered: minimizing the width of the depletion layer using highly doped diamond andor introducing defects at the metal interface e.g by fonning an interfacial carbide. No true n-type doping of diamond has yet been reported, but carbide-forming metals such as Mo, Ti, and Ta have been used as substrates for diamond coatings. Moazed et al. I101 reported an ohmic contact on MoJp-type diamond after annealing. Tachibana et al. [I I] and Gildenblat et al. [12] also suggest that formation of titanium Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1996515 JOURNAL DE PHYSIQUE IV carbide may alter the Schottky barrier to ohmic contact with diamond. Our earlier results showed that diamond powders deposited onto Mo and Si emitters, even without annealing, enhance electron emissivity substantially, suggesting that the barrier to electron tunneling can be reduced by formation of a physical interface alone. This result can be explained if one includes the effective mass of the electrons in the conduction band of diamond and the diamond is nitrogen doped [2,9]. However, it is clearly better to achieve a more permanent bond between the diamond particles and the Mo emitter, for example, by annealing the specimens. There are two main aims: no shape change of the Mo and the formation of a chemical bond. TEM results showed that Mo2C carbide formed on the diamond/Mo interface at annealing temperatures as low as 430 "C [13, 141. The results presented here relate the current-voltage (IV) characteristics and the field emission energy distribution from the same Mo field emitters before and after diamond coating, and then after heat treatment. In turn these data suggest a plausible mechanism for the enhanced emission. 2. EXPERIMENTAL PROCEDURE Mo emitters (with radii 60nm or smaller) were fabricated from 0.125mm wire by electrochemical etching in a concentrated KOH solution at 10V dc. After suitable cleaning, I-V characterization was conducted on pure Mo emitters. Then the emitters were coated with HPHT (High pressure, High temperature) diamond powder using dielectrophoresis [6]. I-V characterization was repeated after coating, and again after annealing on the same specimen for direct comparison. The diamond coated emitters were annealed for times up to 30 min. and temperatures as high as 600°C in a partial pressure of 100 Torr hydrogen. All emission measurements were carried out at 1 * 10" Torr vacuum and at room temperature with a tip-to-anode distance of 10 mm [2]. The morphology of the diamond coatings was determined by a scanning electron microscope. The electron energy distribution was obtained using a suitable applied voltage (200 700V) between the holder and an extraction gate with a 500pm opening and a tip-to-gate distance of 400pm. This holder was positioned at the entrance slit to a CLAM11 (Vacuum Generators Ltd.) hemispherical electron energy analyzer . Typical data collection conditions included a dwell time of looms, an energy step of 0.025eV. Three scans taken over each desired energy range. Before all measurements, a thermal desorption treatment was performed at 400°C for one hour. 3. RESULTS AND DISCUSSION 3.2 I-V cbruacterization The field emission characteristics of the Mo emitters were investigated before and after diamond deposition, and then after annealing. Fig.1 shows the I-V characterization of a Mo tip with a 150nm tip radius coated with a thin (100nm) coating of diamond particles. Afier coating the electron emissivity was found to increase in a manner comparable to the data presented earlier[l, 2,6]. The improved emission was attributed to a reduction in the "effective workfunction", i.e. a lower overall barrier to emission as a consequence of coating alone with no measurable change in the shape of the Mo tip [2,9]. Further annealing increased the electron emissivity an additional two orders of magnitude compared with the unannealed diamond coated tip. The tip morphology was not measurably changed by such annealing as deduced from electron microscopy (Fig. 2). It was also observed by High Resolution Transmission Electron Microscopy that M0,C regions were formed at the diamond-molybdenum interface during annealing [13,14]. Other investigators [lo, 11, 121 have suggested that carbide formation between diamond and transition metals may lower the Schottky barrier and even develop an ohmic contact. We suggest that these carbides may reduce the energy barrier between Mo and the conduction band of diamond. Fig. 1 Field emission characteristics of Mo field emitter Fig. 2 SEM micrograph of diamond coated Mo field before and after diamond coating, and after annealing. emitter after annealing. 3.2 Energy distribution Fig. 3 shows the results of energy distribution analysis from a pure Mo tip, after diamond coating, and after annealing treatment. The three emitter conditions gave widely different intensities, varying by over lo3 times at the same voltage (Vo = 450 volts) and their maxima are displaced one from another when plotted as a function of the applied bias. The dependence of this peak displacement with applied voltage is plotted in Figure 4 and shows strikingly differing behavior depending upon emitter condition. The peak shift of the energy spectrum maxima with respect to the Mo Fermi energy level for the diamond coated tip and annealed tip increased with the applied voltage while that of the Mo tip remained constant (Fig. 4). This peak shift represents the internal voltage drop which is comprised of the difference between the Fermi level in the metal substrate and the effective Fermi level at the emitting surface. This is consistent with the constant value of 4.5eV obtained as a function of a V, for pure Mo in (Figure 4). After diamond coating, the shift of the energy distribution maximum increased up to 7eV from that of pure Mo at 625 applied volts. It is proposed that the peak shift in the diamond coated Mo tip is related to the condition of the metaYdiamond contact and to the diamond bulk voltage drop (Fig. 5). However after annealing, the magnitude of the peak shift was considerably reduced. The annealing reduces the total voltage drop because the Moldiamond barrier is reduced through formation of Mo,C at this interface. JOURNAL DE PHYSIQUE IV Fig. 3 The field emission energy spectrum at applied voltage of 450V for Mo, MoDiamond and annealed Mo/Diamond emitter. m e spectrum fiom annealed emitter shows 100 times reduction and pure Mo shows 10 times enlargement from its real size. ( unit of x-axis is eV, e=electron charge, V,= applied bias, %= kinetic energy of electrons emitted h m sample) /1 Pure Mo * MoDia Annealed MOD.