Direct, physical manipulation of matter at the atomic level is the heart of nanostructure science and technology. This requires very special capabilities in terms of tools and personnel. In the past, emphasis has been placed on specialized equipment (e.g., e-beam tools, plasma etchers, proximal probes, etc.), and on environmental control. The point taken here is that management philosophy is at least equally important (if not more important) in achieving project success. The discussion represents a perspective, as derived from experience as director of the Nanoelectronics Processing Facility at the US Naval Research Laboratory and interactions with other Nanofabrication Facilitiesi.
This paper presents soft-contact x-ray lithography exposure results at sub-40 nm length scales and shows that the process latitude for such exposures is extremely wide. For feature sizes as large as 70 nm and as small as 30 nm in PMMA resist, no statistically significant difference in printed linewidth is seen for development times up to 50 percent greater than the time required for clearing of features. Within this 50 percent development window, dense features as small as 45 nm and isolated features as small as 30 nm are within a +/- 10 percent CD variation.
Efficient diffractive optical element (DOE) spatial image separator devices that segment an image and redirect corresponding regions to independent photodetectors are presented. These devices constitute a compact single-element alternative to prism arrays or reflective wedge arrays. DOE's with micro features are fabricated suing conventional microlithography techniques amenable to low-cost production. The system application described is a high-speed RF phase measurement processor for the NRL channelized direction finding receiver. This paper describes the design, fabrication, analysis, and empirical measurements of a 3 segment and 15 segment diffractive spatial image separator.
Pattern placement errors are a problem in the manufacture of masks for proximity X-ray lithography. Many of these errors are attributable to long term drifts in beam position relative to external fiducials. To address this we have developed a technique based on through-the-membrane monitoring of the electron beam position. This uses a solid state detector with high bandwidth and gain mounted near the back surface of the target membrane. An accurately patterned overlayer on the detector provides the fiducial reference. The overlayer is designed to modulate the electron-hole pair current generated in the diode by absorbing the incident beam. Position information is obtained by analyzing the image created from recording the diode current during patterning. The phase in a Fourier transform of the data at the spatial frequency of the patterned absorber gives a measure of the position of the incident beam. Changes in the observed phase from one frame to the next can then be used to correct position errors of the beam in real time. We report results from tests of various components of this system. Early indications are that the system will be sufficiently fast and accurate for proximity X-ray mask pattern placement correction.
KEYWORDS: Point spread functions, Electron beam lithography, Scattering, Laser scattering, Backscatter, Analog electronics, Optical engineering, Data processing, Electron beams, Optimization (mathematics)
We have developed a technique based on use of a novel e-beam detector to enhance pattern placement accuracy of high-performance e-beam tools in the patterning of membrane masks. The detector consists of a reverse biased Schottky diode whose area coincides with that of the membrane, and which is placed immediately behind the membrane/-absorber/resist multilayer. An accurately patterned absorber overlayer covers the detector surface. The overlayer absorbs a fraction of the incident electrons, modulating the detector signal as the beam passes over the membrane. We have studied and modeled the performance of prototype detectors covered with 16 micrometers period gratings over an incident energy range of 5-50 keV. The combination of Schottky diode and patterned overlayer has been used to improve electron scattering models. We have observed excellent SNR with a 40 nC/cm2 dose at 50 kV, and spatial resolution better than 0.1 micrometers of a beam transmitted through a membrane. More accurate position information can be obtained by taking a Fourier transform of the diode current waveform created as the incident beam traverses several grating periods. In a poor operating environment, we have observed phase accuracy in 100 micrometers fields of +/- 4 degrees. We have developed an algorithm capable of processing pattern-based data sets at high speeds and low information storage requirements. This technique can be easily implemented with little overhead and little modification of existing systems.
In this paper we explore the pattern density dependent contrast which results from the non- Gaussian nature of the energy deposition profile in exposed resist. A long, spatially extended tail on the actual beam energy deposition profile degrades exposure contrast over that which is predicted by two-Gaussian models. As most background dose equalization schemes, such as GHOST, depend critically on the two-Gaussian model for their success, these schemes cannot correct for background dose variation caused by these long tails. This paper shows that this effect can lead to significant degradation of critical dimension control. A strategy to minimize this degradation based on test structure evaluation is proposed.
This paper focuses on the process latitude for achieving 100-500 nm dimensions in SAL-601 e-beam resist. Si wafers were coated with 140 nm of resist and lithographically exposed with a JEOL JBX-5DII e-beam lithography system, operated at 50 kV and a probe standard deviation, (sigma) , of approximately 10 nm. Post exposure bake (PEB) conditions were 105 degrees C for 1, 3, and 10 min. and 110 degrees C for 1 min. The measured backscatter coefficient was 0.51, in agreement with out Monte Carlo simulation results. Resist thickness measurements showed a factor of 2 variation in sensitivity with PEB conditions but no change in contrast. The measured linespread functions for isolated lines could be fit to single Gaussians with (sigma) of 27-33 nm, depending of the PEB conditions. The measured widths of the linespread functions were more than a factor of 2 larger than predicted by the Monte Carlo simulation. The PEB condition of 105 degrees C for 1 min. showed slightly better dose latitude than the other PEB conditions for fabricating gaps between pads. Still the process latitude became acutely narrow in patterning 100 nm gaps. For 20 micrometers pads the dose latitude was less than 3 percent. There was no single dose that could produce a gap of less than 250 nm in a tower pattern with pad widths of 0.5 to 20 micrometers . The results are compared to calculated dose profiles that incorporate both measured linespread functions and those generated by a Monte Carlo code.
Proc. SPIE. 2621, 15th Annual BACUS Symposium on Photomask Technology and Management
KEYWORDS: Electron beam lithography, Point spread functions, Backscatter, Scattering, Laser scattering, Data processing, Microelectronics, Integrated circuits, Optimization (mathematics), Digital electronics
As the critical dimensions required for masks and e-beam direct write become ever smaller, the correction of proximity effects becomes more necessary. Furthermore, the problem is beset by the fact that only a positive energy dose can be applied with the e-beam. We discuss here approaches such as chopping and dose shifting which have been proposed to meet the positivity requirement. An alternative approach is to treat proximity correction as an optimization problem. Two such methods, local area dose correction and optimization using a regularizer proportional to the informational entropy of the solution, are compared. A notable feature of the regularized proximity correction is the ability to correct for forward scattering by the generation of a 'firewall' set back from the edge of a feature. As the forward scattering width increases, the firewall is set back further from the feature edge. The regularized optimization algorithm is computationally time consuming using conventional techniques. However, the algorithm lends itself to a microelectronics integrated circuit coprocessor implementation which could perform the optimization much faster than even the fastest work stations. Scaling the circuit to larger number of pixels is best approached with a hybrid serial/parallel digital architecture which would correct for proximity effects over 108 pixels about one hour. This time can be reduced by simply adding additional coprocessors.
Standard matrix inversion methods of e-beam proximity correction are compared with a variety of pseudoinverse approaches based on gradient descent. It is shown that the gradient descent methods can be modified using 'regularizers' (terms added to the cost function minimized during gradient descent). This modification solves the 'negative dose' problem in a mathematically sound way. Different techniques are contrasted using a weighted error measure approach. It is shown that the regularization approach leads to the highest quality images. In some cases, ignoring negative doses yields results which are worse than employing an uncorrected dose file.
A process for etching fine features in tungsten (100 nm linewidth or less) to produce patterned absorbers has been developed. The pattern is first defined in a chrome etch mask on the tungsten absorber layer using e-beam lithography and s then transferred into the tungsten by reactive-ion-etching. H2 is mixed with SF6 to passivate the sidewalls of the tungsten features because SF6 alone causes severe undercutting of the features. Control of undercutting is the key challenge in reactive ion etching of tungsten. With an optimum mixture of 20% H2 and 80% SF6, plus substrate cooling to -25 degree(s)C, undercutting can be controlled for 250 nm geometries. Increased undercutting has been observed at the endpoint of the etching process, the chromium etch stop layer. This is demonstrated through a computer model. The endpoint can be controlled through laser endpoint detection. For sub 250 nm geometries, additional sidewall passivation is accomplished with an intermittent etch process, thereby allowing the etching of high aspect ratio 100 nm features in 650 nm thick tungsten layers.
The results of the NRL program focuses on high resolution, high aspect ratio, patterning of W are summarized. The work investigates three parallel approaches: reactive ion etching (RIE), electron cyclotron resonance (ECR) etching, and chemically assisted ion beam etching (CAIBE). Key issues that are analyzed for each process are the etch mask, anisotropy, selectivity, etch stop, compatibility with high resolution (sub-250 nm) lithographic patterning of W, and applicability to membranes. In the first two methods, prevention of sidewall undercutting was the key issue. Here the effort focuses on sidewall passivation and substrate cooling. RIE is a commonly utilized fabrication tool and the process has been developed to etch 100 nm lines. ECR is a relatively new process and there are more degrees of freedom than RIE. Both SF6 chemistry and CBrF3 chemistry have been investigated. Methods to minimize the mask erosion are described and a comparison of Cl2 chemistry to SF6 chemistry is made. The results on the three dry etching techniques are described and contrasted.
A physical method of reducing feature size and proximity effects in sub-quarter-micrometer e-beam lithography is described. A thin layer (50 to 300 nm) of silicon nitride deposited on a semiconductor substrate, prior to resist deposition, has been found to enhance the resist resolution. The samples were patterned with a 50-keV, 15-nm-diam probe generated by a JEOL JBX-5Dll e-beam lithography system. Point spread function measurements in 60-nm-thick SAL-601 on Si are shown to illustrate the resolution enhancement in the nanolithographic regime (sub-100 nm). The technique has been applied to lithography on 400-nm-thick W films, such as would be used in x-ray mask fabrication. The 200 nm of SAL-601 was spun onto W film samples, which were half-coated with 200 nm of silicon nitride. Identical lithographic patterns were written on each half of the sample. On examination of the samples after postexposure processing and development, reduced feature sizes and proximity effects were seen on the sample half with the silicon nitride intermediary layer. For example, in a field effect transistor (FET) type pattern, with a coded 500-nm gap between the source and drain pads, the gate could only be successfully resolved when the intermediary nitride layer was present. Monte Carlo simulations were performed on a CM-200 connection machine. The results show a large number of fast secondary electrons are generated within a 100-nm radius of the incident electron beam. The implications of fast secondary electrons on resolution in e-beam lithography are discussed. The total number of fast secondary electrons entering the resist is reduced by the silicon nitride layer. Simulations compare the thin-layer technique to a bilayer resist technique, used to improve resolution at larger dimensions.
A physical method of reducing feature size and proximity effects in sub-quarter micron e- beam lithography is described. A thin layer (50 - 300 nm) of silicon nitride deposited on a semiconductor substrate, prior to resist deposition, has been found to enhance the resist resolution. The samples were patterned with a 50 keV, 15 nm diameter probe generated by a JEOL JBX-5DII e-beam lithography system. Point spread function measurements in 60 nm thick SAL-601 on Si are shown to illustrate the resolution enhancement in the nanolithographic regime (sub-100 nm). The technique has been applied to lithography on 400 nm thick W films, such as would be used in x-ray mask fabrication. 200 nm of SAL-601 was spun onto W film samples, which were half coated with 200 nm of silicon nitride. Identical lithographic patterns were written on each half of the sample. On examination of the samples after post exposure processing and development, reduced feature sizes and proximity effects were seen on the sample half with the silicon nitride intermediary layer. Monte Carlo simulations were performed on a CM-200 Connection Machine. The results show a large number of fast secondary electrons are generated within a 50 nm radius of the incident electron beam. The implications of fast secondary electrons on resolution in e-beam lithography is discussed. The total number of fast secondary electrons entering the resist is greatly reduced by the silicon nitride layer. Simulations compare the thin layer technique to a bilayer resist technique, used to improve resolution at larger dimensions.
A steered beam lithography will represent an essential part of the technology to meet the future need for ultra-high resolution mask making and direct write. Conventional high voltage e-beam lithography is being developed to meet these challenges. However, there are a number of physical limitations (proximity effects, resist sensitivity) which must be overcome. To do so will prove to be extremely expensive if in fact these problems can be overcome. There are significant advantages in going to extremely low energies in e-beam lithography. Proximity effects are eliminated although the electron-optics become more exacting. The need to focus a low energy e-beam can be eliminated by maintaining a sharp tip close to a surface as in a scanning tunneling microscope (STM). We have demonstrated technologically useful lithography with the STM operated with 4 - 50 V between tip and sample. Patterns have been defined in e-beam resists and by selective oxidation of semiconductor substrates under the action of the STM tip. In both cases the pattern can be transferred into the substrate with a dry etch. Sub 50 nm resolution is routine on a variety of substrates. A viable lithographic technology has been demonstrated in the research laboratory. However, several key issues must be addressed to develop a technologically viable lithography system compatible with existing microfabrication practice. These issues include: registration using the imaging properties of the STM for alignment, pattern accuracy and throughput. Advances in STM speed are described and suggestions made for improving lithographic performance with multiple sharp tips (each with an independent servo loop). The potential pay-off is high as a low voltage lithography tool will involve significantly less capital investment (and support cost) than the next generations of high voltage e-beam lithography tools.