Source Feed Materials Research Articles

Silicon damage control through the use of aromatic hydrocarbon ion implantation

The continuous shrinking of transistor devices calls for precise doping profile control. Generally, monomer carbon (C+) is implanted at doses too low to adequately form an amorphous layer in silicon. Previous experiments used a molecular carbon species combined with dedicated vaporizer-based hardware, but this experiment was carried out using a commercial implanter and source with a liquid precursor material. The use of higher mass molecular carbon ions, such as C7H7+, impart significantly more damage to the silicon than the monomer, and can easily result in a well-defined amorphous layer. There are many species of molecular carbon that can be implanted, but in order to utilize a hot cathode plasma (HCP) ion source some additional criteria must be met. The source feed material should fragment with high-yield of molecular carbon (Cx+, x = 2–8) ions, which could then be extracted at beam currents sufficient to meet production needs using standard ion implant techniques. In this work, we surveyed several candidates for forming the molecular carbon species that could meet the requirements outlined above, reported the performance in plasma formation, and investigated materials properties of the Si after implantation.

Additional peaks in mass spectra due to charge exchange events and dissociation of molecular ions during extraction

It is well-known that charge exchange events and dissociation of molecular ions after extraction and prior to mass analysis result in changes to a so-called apparent mass of the ion and increase the possibility of energetic contamination in ion implantation. It was found that if charge exchange events and dissociation of molecular ions already take place during extraction, additional peaks occur in a mass spectrum. They were observed for all of the used source feed materials, such as AsH3, BF3, and argon. In this paper, the origin of these peaks is discussed and a model that calculates the apparent mass as a function of the extraction voltage is presented.

Liquid alloy ion sources for Pd and Cu

For generation of Cu and Pd ions in liquid metal ion sources Cu-Ga, Cu-Ge, Pd-Ge, Pd-Si and Pd-Sn alloys have been tested as source feed materials. The range of coexistence of solid and liquid phases and the eutectic temperature of the investigated alloys were determined by differential thermal analysis. A stable operation of liquid metal ion sources could be obtained with Cu63.5Ge36.5 (at.%) and Pd36Ge64 (at.%) alloys. The recorded mass spectrum of a Cu63.5Ge36.5 liquid metal ion source has shown a high current of Cu+ ions. The high current of Pd+ ions in the mass spectrum of the Pd36Ge64 liquid metal ion source shows that this alloy is a suitable source material for Pd ion generation. An investigation of the ion flux ratio, which depends on the total source current, has revealed a maximum intensity of doubly charged ions in the low-current region. The fraction of doubly charged Cu and Pd ions is less than 1%.

Mass spectra of alloy liquid metal ion sources (abstract)

Liquid metal ion sources are widely used in focused ion beam systems with applications such as maskless ion implantation and microfabrication. For these applications a variety of technologically important ions are required. Alloys with a eutectic composition as source feed materials are generally used and the desired ions are selected by means of a mass separating system. This paper presents the results of mass spectrometric investigations on Pd-Ge, Cu-Ge, Au-Si-Be, Au-Si-Sb, Au-Ge-Sb, and Pt-Sb-P-Au-Si alloys using tungsten as needle material. The suitable atomic alloy composition was deduced from the corresponding phase diagrams and the temperature range of melting of the employed alloys was determined by differential thermal analysis. Relative intensities of the main ion species in dependence on the total source current were investigated.

Energy contamination in ion implantation

Extending the effective energy range of an implanter is possible by implanting multiple-charged species. A problem when using these species is energy contamination. The energy contamination using P 3+, B 2+ and B + has been investigated on commercially available high-current implanters. The major cause of energy contamination is charge exchange reactions due to (the residual) gas inside an implanter. Careful control and design of the vacuum system are therefore important. Energy contamination can also be caused by sputtering/reflection from slits and lining of the implanter. Ionisation of residual gas is probably the origin of fluorine contamination when using BF 3 as source feed material. The fluorine contamination again stresses the need to control the amount and composition of the residual gas in an implanter.

An analysis of vacuum effects on ion implanter performance

The generation of an ion beam and its impact into photoresist-masked wafers will have an adverse effect on the vacuum of an ion implanter. This is particularly significant when doping with boron using BF 3 or BCl 3 as the source feed material. As well as affecting high voltage performance, poor vacuum will allow charge exchange and partial neutralization of the ion beam which could cause energy contamination as well as dosimetry errors. The nature of these effects depends not only on the system design but also on the implanted species and its charge state and energy. This paper discusses these effects on ion implanter performance and shows how a matrix-based general vacuum model may be used to investigate vacuum integrity. A time-dependent vacuum model that simulates the effect of pressure on dosimetry is also described.

An improved ion source for ion implantation

Modifications to a standard Freeman ion source were made which increased 11B+ scanned wafer current by a minimum factor of 1.4, and 11B2+, 31P2+, and 75As2+ scanned wafer currents by a factor of 4.2. Order of magnitude increases in 31P3+ and 75As3+ scanned wafer currents were also measured in the extracted ion beam. Ion source feed materials were BF3 gas and elemental phosphorus and arsenic. Filament lifetime using BF3 gas has been observed to be in excess of 50 mA h. The apparent increased ionization efficiency was verified by a measured increase in electron density and temperature. In addition, at high extraction current densities (∼30 mA cm−2), the modifications resulted in a symmetric extracted ion beam and uniform filament erosion characteristics. Thus, it was concluded that the plasma uniformity has increased.

The convergent effect of the annealing temperatures of electron irradiated defects in FZ silicon grown in hydrogen : Qin Guogang and Hua Zonglu. Solid St. Commun.53 (11) 975 (1985)

A third generation industrial high current ion implanter, Series III, has been produced for the uniform implantation of milliampere-beams with total acceleration voltages up to 200 kV. Magnetic analysis of the beam with initial acceleration voltages variable up to 40 kV is carried out in the accelerator terminal followed by post-acceleration across a single gap up to a maximum total acceleration voltage of 200 kV into a grounded processor chamber.The ion source is a Freeman type capable of delivering milliampere beams of most elements. The source feed materials for the three main ion species required by the semiconductor industry, i.e. B, P and As, can all be pumped by surfaces with temperatures down to liquid nitrogen temperature. Consequently large liquid nitrogen cryopump/traps are used as the primary pumping system with diffusion pumps having a secondary role for non-cryopumpable gases. Three stages of differential pumping are provided in the accelerator together with a 12″ pumping system for the processor.The large processor takes a maximum load of 108 2″ silicon wafers or 54 3″ silicon wafers. Mechanical scanning is used because of the need to maintain space charge neutralisation in the beam which rules out electrostatic scanning and also because of the need to distribute the heat dissipation over a large number of wafers in order to prevent an excessive temperature rise in the wafers during high dose implants (>1015ions/cm2). The carousel which carries the wafers is shaped so that the wafers pass through the beam in a straight line and at a constant speed. A stepper motor driven vane unit in the terminal controls the ion beam current. In order to achieve high implantation uniformities the processor dose control system controls the ratio ion current/carousel speed and therefore any small non-uniformities in the carousel rotation speed can be compensated by the vane unit.The various functions in the terminal are controlled from ground potential using fibre optics. The system is highly automated and unskilled operators can be used.

Single wafer high rate RIE employing magnetron discharge : Steve Schultheis. Solid St. Technol. 233 (April 1985)

A third generation industrial high current ion implanter, Series III, has been produced for the uniform implantation of milliampere-beams with total acceleration voltages up to 200 kV. Magnetic analysis of the beam with initial acceleration voltages variable up to 40 kV is carried out in the accelerator terminal followed by post-acceleration across a single gap up to a maximum total acceleration voltage of 200 kV into a grounded processor chamber.The ion source is a Freeman type capable of delivering milliampere beams of most elements. The source feed materials for the three main ion species required by the semiconductor industry, i.e. B, P and As, can all be pumped by surfaces with temperatures down to liquid nitrogen temperature. Consequently large liquid nitrogen cryopump/traps are used as the primary pumping system with diffusion pumps having a secondary role for non-cryopumpable gases. Three stages of differential pumping are provided in the accelerator together with a 12″ pumping system for the processor.The large processor takes a maximum load of 108 2″ silicon wafers or 54 3″ silicon wafers. Mechanical scanning is used because of the need to maintain space charge neutralisation in the beam which rules out electrostatic scanning and also because of the need to distribute the heat dissipation over a large number of wafers in order to prevent an excessive temperature rise in the wafers during high dose implants (>1015ions/cm2). The carousel which carries the wafers is shaped so that the wafers pass through the beam in a straight line and at a constant speed. A stepper motor driven vane unit in the terminal controls the ion beam current. In order to achieve high implantation uniformities the processor dose control system controls the ratio ion current/carousel speed and therefore any small non-uniformities in the carousel rotation speed can be compensated by the vane unit.The various functions in the terminal are controlled from ground potential using fibre optics. The system is highly automated and unskilled operators can be used.

A compilation of mass spectra from liquid metal sources

Since our first report at LEIB-1 on the basic operating characteristics of liquid metal ion sources (LMIS) many groups throughout the world have taken up research and application of this source type. Ion spectra from LMIS are uniquely different from those of all other ion source types, perhaps the most notable feature being the very high relatively abundance of M n+ in many cases. The range of ion species available from LMIS was until recently somewhat restricted but has now been substantially extended by use of alloys as source feed material. Even elements that sublime in the pure state and thus cannot be used as source feedstock, can be alloyed in a suitable host matrix to produce useful high brightness ion beams. Ion spectra obtained from such alloys, eutectics or other matrices contain a wealth of useful ion species ranging from Li + to U 4+ which find diverse application in nuclear, atomic and surface physics. The authors have compiled and collated both published and previously unpublished spectra from international research groups together with source feed materials.

Microwave ion source for high current metal beams

A high current, metal ion source has been developed for application to materials modification in metals and insulators. In a microwave discharge type ion source, even highly reactive materials such as metal halides, as well as oxygen, can be used for source feed materials. The microwave ion source originally developed for conventional semiconductor fabrication has been modified to provide several mA of mass-separated metal ions. So far, ions of Al +, Ga +, Ti +, Hf + and Sc + have been obtained.

Magnetron ion source for antimony ions

A magnetron ion source is used for the production of an antimony ion beam. The maximum ion current is 500 μA. The intensity of the ion beam is stable and long-lasting without recharging the ion source feed material.

The design philosophy for a 200 kV industrial high current ion implanter

A third generation industrial high current ion implanter, Series III, has been produced for the uniform implantation of milliampere-beams with total acceleration voltages up to 200 kV. Magnetic analysis of the beam with initial acceleration voltages variable up to 40 kV is carried out in the accelerator terminal followed by post-acceleration across a single gap up to a maximum total acceleration voltage of 200 kV into a grounded processor chamber. The ion source is a Freeman type capable of delivering milliampere beams of most elements. The source feed materials for the three main ion species required by the semiconductor industry, i.e. B, P and As, can all be pumped by surfaces with temperatures down to liquid nitrogen temperature. Consequently large liquid nitrogen cryopump/traps are used as the primary pumping system with diffusion pumps having a secondary role for non-cryopumpable gases. Three stages of differential pumping are provided in the accelerator together with a 12″ pumping system for the processor. The large processor takes a maximum load of 108 2″ silicon wafers or 54 3″ silicon wafers. Mechanical scanning is used because of the need to maintain space charge neutralisation in the beam which rules out electrostatic scanning and also because of the need to distribute the heat dissipation over a large number of wafers in order to prevent an excessive temperature rise in the wafers during high dose implants (>10 15 ions/ cm 2). The carousel which carries the wafers is shaped so that the wafers pass through the beam in a straight line and at a constant speed. A stepper motor driven vane unit in the terminal controls the ion beam current. In order to achieve high implantation uniformities the processor dose control system controls the ratio ion current/carousel speed and therefore any small non-uniformities in the carousel rotation speed can be compensated by the vane unit. The various functions in the terminal are controlled from ground potential using fibre optics. The system is highly automated and unskilled operators can be used.