Frank Sinclair

Gate oxides in high current implanters: how do they survive?

Abstract The use of thinner gate oxides in ULSI, susceptible to catastrophic breakdown at lower voltages, presents an increasingly difficult problem for ion implanter manufacturers. The center of a high current ion beam can be modeled as a voltage source whose potential depends on the current density in the beam, the neutralization of the beam by electrons and the geometry of the process chamber. Many measurements using a variety of methods support the validity of these models. The peak voltages are almost always larger than the expected breakdown voltages of thin gate oxides and thus would predict zero yields. Studies of oxide failure in electrical tests have shown two distinct classes of effects: some devices fail promptly with characteristics consistent with a localized defect that leads to a reduced breakdown voltage, while other devices have reduced reliability that correlates with reduced charge to breakdown. The charge to breakdown test typically gives results in the range from 1 × 10 18 to 1 × 10 21 electrons/cm 2 , much greater than the total dose in an implant, so that a simple charge based model would predict 100% yields. This paper puts forward some physical models based on recent experimental results aimed at approaching the problem from an interdisciplinary perspective. First, we characterize the environment in an implanter using in situ measurements, and conclude that at scales larger than the beam it behaves like a pure current source, while at scales smaller than the beam it behaves like a voltage source with a surprisingly high current limit. Second we review some recent trends in electrical tests of gate oxides, and finally we put forward some hypotheses that could be tested in future experiments.

Charging measurement and control in high-current implanters

Abstract We present data on beam potential measurements in a high-current implanter. The beam potential measurements were made using a test structure on a silicon wafer with a time-resolved in situ data gathering system in the mechanically scanned implanter. Data comparing different conditions in the implanter show that the beam potential strongly influences the surface potential of the wafer when it is exposed to the ion beam. This voltage can be effectively controlled by the use of electron injection from an electron shower as currently practiced. Analysis of results in terms of a theoretical model suggests that charge-induced breakdown of very thin gate oxides in ion implantation will not become an insuperable obstacle in the foreseeable future.

In situ monitoring of wafer charging during ion implantation

Abstract Substantial progress has been made in reducing device charging during implantation, but further progress will require a greater understanding of the processes involved as well as of the environmental effects encountered on the wafer itself. To this end an in situ charge monitor has been developed allowing real time measurements of charging during implantation. The monitors are MOS capacitors ranging from 0.25 to 1 mm 2 in area with an oxide thickness of 1000 A. These monitors show that device charging is first negative, then positive and finally negative. Positive charging is due to the beam ions and the ejected secondary electrons, while negative charging is associated with the electron cloud surrounding the beam. The results obtained show the effects of electron flooding, floating the water electrically, and different wafer surfaces. Under appropriate conditions charging can be held below ± 10 V.

Novel ion implanters for the semiconductor industry

Rapid growth in implant applications in the fabrication of semiconductors has encouraged a dramatic increase in the range of energies, beam currents and ion species used. The challenges of a wider energy range, higher beam currents, continued reduction in contamination, improved angle integrity and larger substrates have motivated the development of many innovations. We review the trends in semiconductor implanter beamline designs, including linacs and tandems for high energy applications up to about 5 MeV, improvements in low energy beam transport down to 1 keV, electrostatic and magnetic collimating lenses with improved energy purity and electrostatically scanned, magnetically scanned and ribbon beam systems that simplify mechanical scanning requirements. Recent proposals also include possible dramatic simplifications that eliminate mass separation as well as plasma immersion that avoids the use of conventional electrode structures.

Ion Implant Beam Guide Material Evaluation

Shrinking critical dimensions of semiconductor devices make it essential to reduce particulate levels in ion implanters. Recent work indicates that the dominant particulate contamination contributed by a high current ion implanter is transported to the wafer by the ion beam itself.[1, 2] One important source of these particulates is the beam guide materials which erode through sputtering and spalling. This paper compares the particulate contamination generated by a variety of competing beam guide materials.

Multiple Twist Implants: Channeling Avoidance with Full Symmetry

In the processing of submicron CMOS devices, the source and drain regions are typically implanted using a self-aligned high dose implant to form a shallow highly doped junction. The requirement for a shallow junction forces the use of an implant angle that is typically 7° from the normal orientation so that channeling of the projectile ions down the symmetry direction does not occur. This tilt angle has been found to give asymmetrical device characteristics because of the penetration of dopants under the gate on one side and the extended shadow on the other. The solution to this problem lies in dividing the dose between a multiple sequence of implants, all at the same tilt angle, but with different azimuthal or twist angles. In this way device symmetry is preserved while all the implanted ions enter the crystal far enough from the symmetry directions that there is not significant channeling. These processes have been tested with an Eaton implanter, and we present data on the effect of different tilt angles on the channeling as measured by SIMS spectra and spreading resistance measurement. We also show sheet resistance, uniformity and throughput data to demonstrate the impact of these procedures on production schedules and process qualification. This shows that the implanter has the necessary angle positioning capability combined with a throughput advantage.

Initial performance results from the NV1002 high energy ion implanter

Abstract The Eaton NV1002 is a high energy ion implanter with beam current capability greater than 1 mA. Acceleration to energies between 80 and 2000 keV is achieved with a variable phase linear accelerator (linac). The first production NV1002 is being tested, and the initial results are reported. Currents of 1–2 mA can be generated over an energy range of 40–1000 keV for boron, phosphorus, and arsenic. Useful currents are available at energies as low as 10 keV, and using doubly charged ions, as high as 2 MeV. Considerations for a commercial implanter such as ease of operation are discussed. Analyses of implanted wafers are presented, demonstrating uniformity, correct implant depth profiling, and freedom from contamination.

Simulation of the geometrical characteristics of a mechanically scanned high current implanter

Abstract An understanding of the geometrical properties of the implantation process is essential for the optimization of dose control algorithms and for control of channeling effects. The EATON mechanically scanned implanters use a closed loop dose measurement system to control the scan speed in the slow direction. The input for this is provided by a measurement of the beam through a rectangular slot in the disk. Analytical calculations of the approximations involved in this approach become prohibitively complicated when one wishes to consider all possible beam shapes, angular spreads and disk geometries. For this reason, we have implemented a computer program which uses a simple approach based on a Monte Carlo simulation of the implant process. The results show that the basis of the control algorithm is sound, leading to deviations of less than 0.3% across a 150 mm wafer for all plausible beam geometries. Another feature of this approach is that it allows easy calculation of the angle with which the ion beam penetrates the wafer or other geometrical considerations.

Shallow SIMOX Technology (SST): A Double Mechanically Scanned Approach

Oxygen implantation into silicon at 2.2-3.0E17cm -2 doses and energies 30-40keV has been successfully utilized to produce shallow silicon-on-insulator (SOI) layers using the separation by implantation of oxygen (SIMOX) technique. The SST films were analyzed by cross section transmission electron microscopy (XTEM). High quality superficial silicon and buried oxide layers have been obtained with very sharp interfaces. No defects were observed in the superficial silicon by XTEM and by plan view TEM. Good breakdown voltage of the buried oxide layer has been obtained.

Charging measurement and control in high current implanters

Data on beam potential measurements and gate oxide yield studies a high current implanter are presented. The beam potential measurements were made using a test structure on a silicon wafer with a time resolved in situ data gathering system in the mechanically scanned implanter. The yield studies used specially designed test devices with oxide breakdown measurements before and after a high dose implant. Data for different conditions in the implanter show that the voltage of the beam is forced on the surface exposed to the ion beam. This voltage can be effectively controlled by the use of electron injection from an electron shower as currently practiced. Preliminary results from the yield studies show very good yields over a wide range of conditions. Analysis of the results in terms of a theoretical model suggests that charge induced breakdown of very thin gate oxides in ion implantation will not become an insuperable obstacle in the foreseeable future.<<ETX>>