Diamond vacuum field emission devices

By W.P. Kang, J.L. Davidson, A. Wisitsora, Y.M. Wong, R. Takalkara, K. Holmes, D.V. Kerns

Diamond & Related Materials 13 (2004) 1944– 1948

Purpose of the study:

This article reports the development of (a) vertical and (b) lateral diamond vacuum field emission devices and testing their field emission characteristics.

Methodes:

These diamond field emission devices, diode and triode, were fabricated using a self-aligning gate formation technique from silicon-on-insulator wafers. The SOI wafer is comprised of a 15um thick Si active layer, 1um thick SiO2 layer (BOX) and 525um thick Si handle.

For vertical devises, 0.2um thick SiO2 layer was grown on the wafer surfaces. Inverted pyramidal cavities were then formed on the silicon active layer by photolithographic patterning and anisotropic etching of Si using KOH solution. The square patterns are sized such that complete inverted pyramidal cavities are formed within the Si active layer. Next, a SiO2 layer was grown on the active Si layer to form the gate dielectric, which also produces a well-sharpened apex on the inverted pyramidal SiO2 layer. Diamond was then deposited in the mold by plasma enhanced chemical vapor deposition technique (PECVD). Next, the backside of the silicon was etched away and stopped at the embedded SiO2 layer. Finally, the SiO2 layer was etched and the sharpened diamond pyramidal apexes exposed. The remaining SiO2 and Si form the dielectric spacer and the gate, respectively. For the diode configuration the SiO2 spacer and the remaining spacer were also etched to completely expose the diamond pyramids.

For lateral devices, 1um thick SiO2 layer was first grown onto the SOI wafer. Conventional photolithography was then performed to pattern the anode and cathode structures onto the SiO2 layer. The exposed SiO2 was etched away using BOE exposing the Si below. Next, electrically conductive diamond was preferentially grown on Si using biasenhanced PECVD, by introducing trimethyl boron (TMB) gas in the plasma mixture for boron doping. The unwanted diamond that grew on SiO2 was lifted-off by etching the SiO2 using an HF.The patterned diamond layer was then used as a masking layer to etch Si to get the required final structure.

The fabricated diamond emission diodes and triodes were tested for electron emission under high vacuum. The emission current was recorded as a function of applied voltages. Fowler–Nordheim (F–N) equation was used to analyze the diamond field emission data.

Key findings:

1. A diamond field emission diode operable at high emission current over 0.1 A in an indented anode vertical configuration has been achieved.

2. A diamond field emission triode with excellent transistor characteristics of high DC voltage gain and large AC voltage amplification is achieved with high DC gain of ~800 and large AC output voltage of ~100 V p–p.

3. A lateral diamond field emitter (cathode–anode spacing less than 2 um) with the lowest turn-on voltage (~5 V) and high emission current (6 uA) has been realized.

4. The low turn-on voltage (field ~3 V/um) and high emission characteristics are the best of reported lateral field emitter structures.

Keywords: Diamond; Field emission; High current; DC gain; Lateral field emitter

Special Reliability Features for Hf-Based High-k Gate Dielectrics

T. P. Ma, Fellow, IEEE, Huiming M. Bu, X. W. Wang, Liyang Y. Song, W. He, Miaomiao Wang, H.-H. Tseng,
and P. J. Tobin

IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY, VOL. 5, NO. 1, MARCH 2005 (Invited Paper)

Purpose of the study:

HF-Based gate dielectrics are extensively investigated as alternatives to for future CMOS technology, but before any of these dielectrics can be selected by the semiconductor industry, its reliability must be satisfactorily demonstrated. This paper reviews some its recent reliability results.

Methods:

PolySi, TaSiN, or TiN gate MOSFETs were fabricated following a standard CMOS process on bulk silicon. ALD HfO2 of 4.4 nm in physical thickness deposited at 300 C with precursor, was used as the gate dielectric. Some of the MOSFETs had a silicon nitride cap layer of 0.5 nm thick between and gate for studying the effects of polySi/ reactions. The typical I–V and C–V characteristics for nMOSFETs are presented to show that the devices under study exhibit normal electrical characteristics. To measure the trapping-induced instability, constant voltage stress as well as pulse stress was applied to the polySi gate of MOSFETs in inversion. The threshold voltage shift, the midgap voltage shift and transconductance degradation caused by charge trapping were continuously monitored during the stress. Both pulsed voltage stress and dc constant voltage stress have been applied. Between each stressing period, there is a sensing period, in which the threshold voltage is measured.The threshold voltage shift is obtained from comparison with the initial one. Special attention is also paid to minimize the detrapping effect by using automatic program control to minimize the time interval between stressing and sensing. The density of the trapped charges, was derived from the midgap voltage shift. The interface trap density was obtained from ac conductance measurement. Voltage-dependent TDDB (time-dependent dielectric breakdown) measurements were used to determine the device lifetime. IETS spectra were taken on MOS capacitors by measuring the second harmonic signals with a standard lock-in method at liquid helium temperature (4.2 K). The modulation voltage of the excitation signal for the IETS measurements was 2 mV. A dual temperature (4.2 K, 77 K) technique was used to remove the elastic tunneling background.

Key findings:

1) The operating lifetime extracted from time-dependent-dielectric-breakdown (TDDB) is too optimistic, and the actual device lifetime is limited by the trapping-induced threshold voltage shift.

2) nMOSFETs are much more prone to trapping-induced than their pMOSFETs counterparts under normal operating conditions,due to much more electron traps than hole traps in HfO2-based gate dielectrics.

3) Metal gate yields improved reliability compared to polySi gate.

Laser irradiation effect on electron field emission properties of carbon nanotubes

Hsiu-Fung Chenga, Yun-Shuo Hsieha, Yi-Chun Chena, I-Nan Linb

Diamond and Related Materials 13 (2004) 1004–1007

Purpose:

To study the effect of laser irradiation on the electron field emission properties of CNTs. The effect of laser irradiation on the field enhancement factor was also studied. This behavior was also compared with the temperature dependence of field emission for the CNTs, so as to understand the possible mechanism involved.

Methods:

Carbon nanotubes (CNTs) used in this study were multiwall and were produced by thermal chemical vapor deposition process. The electron field emission properties were carried out in an ultra-high vacuum chamber. A platinum sphere was used as anode and placed at 50 mm from the CNTs-coated substrates.The electrical potential applied to the nanotubes emitter relative to the Pt sphere was typically from 0 to 1100 V. The laser fluence used is 10 mW, which is markedly smaller than the thresholdfor destructing the CNTs. Moreover, to facilitate the comparison, the electron field emission properties of the CNTs were also measured by heating the samples. The current density–electric field (J–E) characteristics of the CNTs were analyzed using Fowler–Nordheim (F–N) theory.

Key findings:

1. CNTs are uniformly thin (approx.40 nm), although catalyst-clusters were occasionally observed in the CNTs.

2. With laser irradiation, the emission current density at electric field 2 V/um was enhanced from 0.75 to 14.0 mA/cm2.

3. The field enhancement factor was hardly altered.

4. Electron field emission properties of CNTs measured at elevated temperature implied that the mechanism modifying the electron field emission properties of CNTs was not the thermal effect.

Keywords: Nanotubes, Field emission, Laser

Effects of catalyst pre-treatment on the growth of single-walled carbon nanotubes by microwave CVD

K. Bartsch *, B. Arnold, R. Kaltofen, C. Ta¨schner, J. Thomas, A. Leonhardt

ScienceDirect, Carbon 45 (2007) 543–552

Purpose of the study:

1. To test the ability to synthesize SWCNTs by using MPCVD

2. Effect of different catalyst and pre-treatment time on it.

Methods:

A microwave CVD system equipped with a vertical quartz tube reactor was used to synthesize SWCNTs. Additional to that, a negative dc bias of 400 V was applied to the substrate (oxidized silicon). The substrates were directly heated to 975-1175K by the microwave plasma. The catalyst layers were prepared by magnetron sputtering using different target such as Mo, Fe, Al-Fe-Mo. Prior to the synthesis, the samples were pre-treated by hydrogen plasma at about 973 K for 10 min or were annealed in air at 823 K for 30 min. Then a gas mixture of CH4–H2 or C2H4–H2 was used to synthesize SWCNTs. The samples were investigated by scanning electron microscopy (SEM), analytical transmission electron microscopy (TEM) including energy dispersive X-ray spectroscopy (EDX) and electron energy loss spectroscopy (EELS), as well as by Raman spectroscopy.

Key findings:

1. A significant difference in SWCNTs’ orientation and structure were observed under with and without applied bias. Better alignment was achieved with applied potential.

2. It was also observed that optimum temperature is required to grow CNTs. Higher temperature etches the tubes a lot faster, resulting poor tube structure.

3. The catalyst film thickness plays another important role.

4. The appropriate pre-treatment of the substrates is essential for the deposition of single-walled carbon nanotubes.

Keywords:

MPCVD: Microwave plasma enhanced chemical vapor deposition

SWCNTs: Single walled carbon nano tubes