Akintunde I. Akinwande

A low-frequency (561 MHz), injected beam, and linear format crossed-field amplifier (CFA) using gated field emission arrays (GFEAs) has been experimentally studied and compared with simulation. The CFA uses a copper wire on Teflon meander line circuit with retardation of ~21. Eight silicon tip GFEA dies were used as the injected electron source to provide up to 160 mA. A segmented end-collector system (nine electrodes) was used to measure the spatial variation of the beam current with and without gain. A gain of ~5.5 dB was measured for a sole-circuit voltage of −2.9 kV, an injected beam current of ~160 mA, an applied magnetic field of 0.0125 T, a radio frequency (RF) input power of 15 W, and a sole-circuit gap of 2 cm. A CST particle in-cell model shows a high gain (~1–2 dB) than the experiment, but the gain variation versus injected current, voltage, and magnetic field matches well. Variation with RF input power shows a significant decrease in gain above 15 W in the experiment with the decrease seen in simulation observed after 25 W. Analysis of the end-collector current shows a rapid decrease after 12 W in the experiment and 25 W in the simulation. This result occurs because the highly cycloidal electrons are close to the CFA circuit and get collected on the circuit before providing amplification energy. This observation is confirmed in simulation, which shows that the current going to the circuit rapidly increases and the end-collector current rapidly decreases. This effect also accounts for the higher gain observed in simulation. These experiments provide a basis for using gated field emitters to study beam–wave interactions in microwave vacuum electron devices.

High power crossed-field devices such as crossed-field amplifiers (CFA) are advantageous in terms of power density and efficiency. Disadvantages are low gain and relatively high noise. Improving gain and noise characteristics would make the CFA more appealing for a variety of applications. In this work, we demonstrated operation of a linear format CFA using a meander line slow wave circuit. Operating frequency is 561 MHz. The circuit is 138 mm long and 72.5 mm wide with endhats and a segmented (9 element) end collector. Circuit retardation is 21. The circuit to sole gap is 20 mm. An injected beam configuration using silicon gate field emitter arrays was used.

We demonstrate the first steps in engineering the anode of field emitter arrays (FEAs) for optimal vacuum packaging by studying a parallel anode–cathode configuration. As part of our study, we report an unexpected gate-controlled negative differential resistance (NDR) region in the output characteristics of FEA-based triodes. The FEA triode consists of an FEA cathode and a silicon MEMS anode that are separated by insulated standoffs to form a vacuum channel. The FEA cathode is an array of high-aspect nanowires connected in series with gated emitter tips. Electrons extracted by the gate–emitter voltage undergo ballistic transport to the anode. It is generally assumed that a parallel anode–cathode triode structure would be ideal due to its geometrical compactness and symmetry. In this work, an on-chip integrated flat silicon anode was fabricated to characterize the parallel configuration for well-defined anode-to-emitter distances of ≤100 μm. The observation of NDR in the “triode” operation regime, which is space-charge limited, suggests that the parallel anode–cathode structure will not be ideal for the integration of triodes for some circuit applications because of unfavorable electrostatics in the vacuum channel between the cathode and the anode. In addition, we demonstrated that the performance of the triodes could be engineered to reduce the ON-resistance (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\textit{R}_{\biosc{on}}$</tex-math></inline-formula>) and increase the output current by varying the geometry and position of the anode.

Field Emitter Arrays (FEAs) have the potential to operate at high power, high frequencies, and endure harsh environments. However, vacuum packaging these devices poses a challenge due to the sensitivity of the emission phenomena to the surface properties of the cathode, such as the work function and the tip radius. Studying the effect of different residual gases on FEAs can enhance our understanding of the interaction between the emission surface and the environment and help estimate the permissible amount of residual gases within the package. In this study, the effect of N 2 exposure on 500 × 500 Silicon Field Emitter Arrays (Si-FEAs) was investigated. The device was exposed to 10 000 Langmuir (L) of N 2 at 10 −7 Torr. During the exposure, the anode current increased from 4.7 μA to 16 μA. However, this enhancement in current was temporary, and upon closing the leak valve, the anode current gradually returned to the pre-exposure level. No significant change in current was observed when the device was powered off during N 2 exposure. The extent of current enhancement showed a direct relationship with the partial pressure of N 2 . These results suggest that the presence of N 2 in a vacuum package does not degrade the performance of Si-FEAs. • Si-FEA emission current increased significantly in an N 2 environment. • The changes on the Si-FEA emission current happens only when the device is on and the changes are reversible. • O 2 partial pressure, compared to N 2 and Ar, limits Si-FEA lifetime.

Silicon field emitter arrays (FEAs) are cold electron sources for devices such as x-ray sources, ion sources, high-power microwave amplifiers or multi-beam electron lithography. For each device, it is imperative to optimize the anode structure and device package to obtain reliable performance. It is generally assumed that a parallel configuration of the FEA cathode-anode structure is the most ideal due to the compactness and symmetry. However, little work has been done to study the behavior of the FEA devices with this structure and consequently there is no compact model for the device. In this work, we report an unexpected yet repeatable negative differential resistance (NDR) in the device output characteristics, suggesting a need for an optimal FEA cathode-anode configuration. A compact model for the FEA cathode-anode parallel-plate configuration is proposed for anode-to-emitter distances <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\leq 100\mu \mathrm{m}$</tex>.

We present a proof-of-concept inverter based on silicon field emitter arrays (Si FEAs) that could be fabricated as a vacuum integrated circuit (IC). A circuit model for Si FEAs is developed, and the voltage transfer characteristics of the FEA inverter are simulated. In addition, a sample of 30 Si FEAs is characterized to determine the variations in <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$a_{\text{FN}}$</tex> and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$b_{\text{FN}}$</tex> caused by the fabrication process. A Monte Carlo analysis is used to test the impact of variations on the FEA inverter performance.

Field emitter arrays (FEAs) have the potential to operate at high frequencies and in harsh environments. However, the vacuum packaging of these devices poses a challenge due to the sensitivity of the emission phenomena to the surface properties of the cathode. Studying the effect of different residual gases on FEAs can help to understand the interaction of the emission surface with the environment and identify the feasibility and requirements for vacuum packaging. In this work, the effect of N2 exposure on 150 × 150 gallium-nitride-field emitter arrays (GaN-FEAs) was studied. The GaN-FEA was first operated at 10−9 Torr with a 1000 V DC anode voltage and a 50 V DC gate voltage, where the anode current was 6 μA. The device was then exposed to 10 000 l N2 at 10−7 Torr, and the anode current increased by 2.7 times during N2 exposure. The increase in the current was not permanent, and the current gradually decreased to its pre-exposure level after the N2 source was cut off. The results of N2 exposure were compared to Ar and O2.

Modern day field emitters can fail due to several mechanisms that are not well understood. This paper presents experiments that aim to identify the mechanisms behind failure. Two types of devices, Silicon gated field emitter arrays (SiGFEAs) and Titanium Silicon Oxy Nitride (TiSiON) lateral field emitter devices were characterized experimentally. Si-GFEAs were tested for arc occurrence time for a fixed gate voltage of 50 V and a fixed collector voltage of 200 V. The emitter was grounded. Initial results from the temporal response experiment show that the emitter experiences arcing first. However, future experiments will provide an accurate identification of the arc initiating electrode. For the planar device, a diode was chosen and IV characterization was performed at 50 °C and 400 °C. Experiments showed that for an applied collector voltage of 10 V, the field emission current was ≈ 5.5 nA before the heat treatment and was ≈ 2.75 nA after the 400 °C heat treatment. This reduction in current could be attributed to the removal of water vapor by heat treatment resulting in the reduction in the surface leakage current.

A 925 MHz, slow wave meander line crossed-field amplifier (CF A) structure was designed and developed using CST Microwave Studio. The structure has a phase velocity retardation of 24. The high frequency solver shows that the structure has a S11 of −40 dB at 925 MHz. For the PIC simulation, an already developed crossed field structure was used which includes a sole, a segmented collector, two end hats, and an injected electron source. PIC simulation shows that for a sole voltage of −3k V, a magnetic field of 150 G, an injected current of 150 mA, and an RF input power of 15 W, the gain of the structure was 10 dB. Also, experimental studies with a previously developed CF A structure are ongoing. This CF A structure was designed to operate at 600 MHz. An injected beam current of up to 60 mA using gated field emission arrays in an 8 die configuration has been tested. However, no gain was observed so far. Further experiments will be carried out using a higher injected beam current.

Field emitter arrays (FEAs) have the potential to operate at high frequencies and in harsh environments. However, they have been shown to degrade under oxidizing environments. Studying the effect of O2 on FEAs can help to understand the degradation mechanisms, identify the requirements for vacuum packaging, and estimate the lifetime of the device. In this work, the effect of O2 exposure on 100 × 100 gallium-nitride-field emitter arrays (GaN-FEAs) was studied. The GaN-FEAs were operated at 6 × 10−10 Torr with a 1000 V DC anode voltage and a 50 V DC gate voltage, where the anode current was 1 μA and the gate current was ≤4 nA. The devices were exposed to 10−7, 10−6, and 10−5 Torr of O2 for 100 000 L. The anode current dropped by 50% after 300 L and 98% after 100 000 L. It was observed that the degradation depends on the exposure dose, rather than pressure. The devices mostly degrade when they are ON, confirmed by exposing the device to O2 when the gate voltage was off, and also by the relation between the degradation and duty cycle when pulsing the gate. The results of O2 exposure were compared to Ar exposure to determine whether sputtering and changes in the surface geometry were the primary cause of degradation. The results suggest that changes in the work function and surface chemistry are the cause of emission degradation of GaN-FEA induced by O2.

Vacuum transistors (VTs) are promising candidates in electronics due to their fast response and ability to function in harsh environments. In this study, several oscillator and logic gate circuit simulations using VTs are demonstrated. Silicon-gated field emitter arrays (Si-GFEAs) with 1000 × 1000 arrays were used experimentally to create a VT model. First, transfer and output characteristics sweeps were measured, and based on those data, an LTspice vacuum transistor (VT) model was developed. Then, the model was used to develop Wein and Ring oscillator circuits. The circuits were analytically simulated using LTspice, where the collector bias voltage was 200 V DC, and the gate bias voltage was 30–40 V DC. The Wein oscillator circuit produced a frequency of 102 kHz with a magnitude of 26 Vpp. The Ring oscillator produced a frequency of 1.14 MHz with a magnitude of 4 Vpp. Furthermore, two logic circuits, NOR and NAND gates, were also demonstrated using LTspice modeling. These simulation results illustrate the feasibility of integrating VTs into functional integrated circuits and provide a design approach for future on-chip vacuum transistors applied in logic or radio-frequency (RF) devices.

Field-emission-based vacuum transistors have been proposed as promising candidates for future high-power and harsh-environment electronic devices. However, the lack of an integrated anode is still an issue for vertical field-emission vacuum transistors for some applications such as radiation-hard vacuum-electronic-based circuits. In this work, an anode-integration technology enabled by tilted metal deposition is proposed and experimentally demonstrated on GaN gated field emitter arrays (FEAs). Full transistor prototypes with a 10 3 on-off ratio in anode current are demonstrated. This process is compatible with gated FEAs of various materials, the vacuum channel can be sealed during fabrication, and the vacuum channel length can be controlled via multiple process parameters.

An experimental setup has been developed to perform experiments on a planar, crossed electric and magnetic field or crossed-field (CF) device. The structure, which is 15 cm long and 10 cm wide with an anode-to-sole gap of 2 cm, can measure electron beam perturbation as a function of injected beam current and magnetic field. The applied maximum anode to sole voltage is 3 kV and the applied maximum magnetic field is 0.02 T. A beam-measurement system, which consists of an anode with eight segmented sections and nine segmented end collectors, is incorporated. Eight silicon-gated field emitter arrays (Si-GFEAs) are used for the electron source. For the experiment, 1.5 mA of injected current at 50-V pulse was used. Experimental results without an applied magnetic field and with a magnetic field with different tilts are compared with simulation results and 1-D theory. The experimental planar crossed-field configurations demonstrate electron stability thresholds in current density and magnetic field tilt that agree with theory and simulation.

Effects of gases on field emission performance were measured using silicon-gated field emitter arrays. Gas was injected into a vacuum chamber with a 1000 × 1000 tip array, which was driven by a DC gate and collector voltages. The collector voltage was fixed at 200 V while the gate voltage was swept to 40 V. For the gas exposure study, N2, He, and Ar were used. The sets of partial pressures, 5 × 10−6, 5 × 10−5, and 5 × 10−4 Torr, were used for the experiment. It was observed that N2 had the least effect and Ar had the worst effect on emission current performance. The degradation of collector current at 5 × 10−4 Torr pressure for Ar was ≈65% where for the N2, at the same level of pressure, the degradation was ≈41%. However, further experiments with high purity Ar gas showed that it was the water vapor present in the gas itself that was the primary cause of reduction in emission current and not the gas itself. The results expressed in reduction in emission current versus Langmuir exposure versus the current clearly showed the effect of water vapor. After the vacuum was recovered, the work function again restored partially to its original value. After ultraviolet light cleaning, the emission current was restored completely to the original state.

The impact of Oxygen (O<inf>2</inf>) exposure on Silicon Field Emitter Arrays (Si-FEA) was studied. A 50×50 array of Silicon field emitters was tested at 1000V DC anode and 45V DC gate voltage in 6×10⁻¹⁰ Torr before 10–⁷ Torr partial pressure of O<inf>2</inf> was introduced into the chamber. The results indicate that the anode current degradation rate is approximately 0.1 percent per Langmuir of O<inf>2</inf> exposure. This study can provide guidelines for the vacuum packaging requirements of Si-FEAs.

Gated field emitter arrays (GFEAs) can fail due to various mechanisms which are not well understood. In this paper, several proposed failure mechanisms are investigated using simulation and experiment. The modelling performed using CST considers an ion bombardment zone to calculate the locations and number of ions that hit the emitter tip apex. As the starting location of the ions moves away from the tip, the fraction hitting the tip apex increases until <tex>$5\mu \mathrm{m}$</tex> from the tip and then decreases until only ions born directly above the tip impact. Electrical measurements of arcs show that arcs only occur during forward bias with emission rather than in reverse bias indicating the mechanism is not surface breakdown.

We are developing a Crossed-Field Amplifier (CFA) experiment utilizing silicon Gate Field Emitter Arrays (GFEA) [1]. These devices have 9M emitter tips, but we are currently developing a driver system using arrays of 1000 x 1000 emitter tips [1]. The driver provides a 1 ms duration square pulse up to 60 V to drive the GFEAs in a repetitive or single shot mode. The driver circuit communicates wirelessly to a desktop computer through the “XBee” ISM 2.4 GHz module from where the pulse is triggered. In the experiment the driver system will float at -3 kV [2]. In a vacuum chamber at a pressure of 10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−8</sup> Torr the GFEAs are tested with the driver circuit. A collector spaced 1 mm above the arrays is kept at 200 V. The driver circuit has been demonstrated for the smaller arrays, and new experiments are underway with the larger arrays which consist of 36 sub-arrays (each array of 500 x 500 emitter tips). The experimental purpose is to use the driver circuit to pulse the test structure utilizing eight 9M tip array die to generate an emission current of 150 mA capable to produce gain in the Crossed-Field Amplifier (CFA) designed at BSU [3]. Design of this experimental CFA at a frequency of 900 MHz will be discussed.

To exploit new architectures for vacuum transistors (VT) that combine the positive attributes of semiconductor electronic devices (i.e. high gain, low noise, and microfabrication) with the positive attributes of vacuum electronic devices (harsh environments<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> compatible), cold cathodes that operate at high current density and low operating voltages are being studied. State-of-the-art approaches such as those reported in<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> exhibit most of these attributes. However, few studies are reported so far which describe these devices in terms of electronic circuits. In this work, Si-GFEA die with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1000\times 1000$</tex> arrays were used to create various oscillator circuits. First, a VT model<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> was established using the transfer and output characterization data from a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1000\times 1000$</tex> array. A Keysight B2902A source measurement unit (SMU) was used for the experiments. Then the model was used to develop a Colpitts, a Wein and a Ring oscillator circuit. The Colpitts oscillator circuit was experimentally tested<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> where the collector voltage was kept at 210 V DC and the gate voltage was kept at 50V DC. From the Wein oscillator simulation, an oscillation frequency of ~ 102 kHz was observed with a maximum amplitude of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{V}_{\text{PP}}\approx 26\mathrm{V}$</tex> for a capacitance of 206 pF. Further, a Ring oscillator was also simulated and an oscillation frequency of ~ 1.1 MHz was observed for a similar capacitance of 206 pF. The oscillation amplitude and frequency can be controlled using a biasing resistor and capacitors. This performances with shift in frequency for shift in capacitance at high temperature can be utilized as a temperature sensor. Further, plan for experimentally realizing the Wein oscillator and the Ring oscillator is going on.

The applications of Silicon gated field emitter arrays (Si-GFEAs) have always been limited by failures (arcs) which have been attributed to a number of reasons: (a) tip runaway due to ion space charge-initiated field emission, (b) ion bombardment, (c) thermal runaway, and (d) surface breakdown [1]. The resulting cathodic arcs can be measured experimentally [2], and the resulting damage observed. We have been studying the arc initiation of silicon tips [3] both with experiment and simulation. Experimentally, we are measuring the electrical signature of the breakdown under a variety of conditions (array size, UV cleaning, forward and reverse bias) and clearly see that arcs occur almost exclusively under forward bias when emission occurs. The approximate arc initiation timing for an arc to occur was 10–20 mins. We have also been studying the trajectories of ions that might be generated from electron impact ionization to quantify the fraction of ions that actually bombard the tip. The silicon gated emitter structure is modeled, and nitrogen ions are distributed across a rectangular box and an increasing distance away from the tip structure. The simulation results predict a zone in which ions born in this zone will impinge at the emitter tip apex. Using a dummy simulation electrode, it has been found that the maximum number of ions hit the tip apex when the ions start approximately <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$6\ \mu \mathrm{m}$</tex> away from the emitter tip while the minimum occurs when the ions are close to the dummy collector. This is expected as the ion energy will prevent it from being focused to the tip apex. Future simulations will be performed in which electron impact ionization will be modeled to see the effects on ion focusing.

To achieve high-performance printed electronic devices, scalable and cost-effective printing of high-quality metallic electrodes with narrow gaps, such as for transistors with short channel lengths, is desirable. Here, we demonstrate short channel (<10 μm) transistors, using thin (<100–200 nm) electrodes fabricated by flexographic printing with nanoporous stamps, with single-wall carbon nanotubes (SWCNTs) as the network semiconductor. The nanoporous stamps comprise polymer-coated vertically aligned carbon nanotubes and facilitate control of the printed ink thickness in the 50–200 nm range. The measured on–off ratio and mobility meet or exceed those of previously reported SWCNT network transistors fabricated by alternative printing methods.

Silicon gated field emitter arrays have been used as a vacuum transistor to demonstrate a 152 kHz Colpitts oscillator. The transfer and output characteristics of the 1000 × 1000 silicon arrays were measured using a collector placed ≈ 1 mm away with a gate voltage up to 40 V and a collector voltage up to 200 V. The data were used to establish an LTspice transistor model based on a field emission tip model and a collector current model that fit the characteristics. Then, the LTspice model was used to design a low frequency Colpitts oscillator. Furthermore, experiments were carried out to successfully demonstrate the oscillation. Oscillation frequency was 152 kHz with a peak to peak voltage of 25 V for a tip to ground series resistance value of 10 kΩ at 50 V on the gate and 210 V on the collector. Further, the oscillator was also tested at 50, 100, 200, 300, and 400 °C. It was observed that frequency shifts for each temperature which is due to the change in the overall capacitance of the test setup. This type of device could be used as a temperature sensor in harsh environments.