Effects Of High Temperature Bakeout on Gated Sillicon Field Emitter Arrays

Abstract

Nano scale vacuum channel transistors (NVCTs) demonstrate the possibility of high-power density and operation in harsh environments. However, characterization studies on present state-of-the-art devices <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> , under high temperature conditions are not complete. In this work, Gated field emission arrays (GFEAs), fabricated using a novel method2, are a strong candidate as an electron source for vertical NVCTs. These devices have been characterized at high temperature (400º C). Experiments were carried out on <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1000\times 1000$</tex> arrays of GFEAs <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> in a high vacuum chamber, using an in-house developed test jig consisting of a stainless-steel collector along with a tungsten probe pin to connect to the gate pad. A Molybdenum heater chuck for high temperature test was also developed. A fixed 100 V DC and 0–40 V DC sweep voltage were applied to collector and gate, respectively. I-V curves show that for gate voltage of 40 V, the collector current was <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\approxeq 80\mu \mathrm{A}$</tex> where a large gate leakage current of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\approx 6\ \text{mA}$</tex> was observed. This low collector and large gate current are partially because of the large collector to array gap <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(&gt; 2\ \text{mm})$</tex> ; however, the low emission current and high gate current are also the result of gas adsorbates. To study the performance, the GFEAs were baked at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$400^{\circ}\ \mathrm{C}$</tex> , and in-situ I-V characterizations were carried out over a range of temperatures. It was found that at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$400^{\circ}\ \mathrm{C}$</tex> , the collector current increased by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&gt; 10\mathrm{x}$</tex> while the gate leakage current decreased by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&gt; 10\mathrm{x}$</tex> , and even after cooling down to room temperature, this result remained unchanged. The F-N plots comparing the collector current, before and after bake out, show the improvement. This improvement can be attributed to the desorption of water vapor during bakeout. To confirm this phenomenon, the GFEA die were taken out of the vacuum chamber, kept in room air for <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$4\sim 5$</tex> days, and then tested again. After each test cycle, the high gate current and low emission current returned until being baked again.