Robert John Socha

Printability of Phase-Shift Defects Using A Perturbational Model

A perturbational model which describes the printability of defects in phase-shift masks (PSM) as a function of focus position is presented and used along with the SPLAT to generate design data on tolerable defect size as a function of defect phase. Defects in a clear field mask with phases other than 180 degrees have the greatest tendency to print when out of focus. From a reformulation of the perturbational model, the tendency to print when out of focus can be explained by the combination of the phase contributed by defocus in the defect image with the phase of the light passing through the defect. To assess the printability of defects, the impact of defects near features in thin chrome attenuating PSM and embedded attenuating PSM were evaluated by calculating the image linewidth variation.

Models for characterizing phase-shift defects in optical projection printing

An algebraic model is developed for characterizing the printability, inspection, and repair of phase-shift defects in optical projection printing. Phase-shift defects are particularly difficult to characterize because of the many parameters associated with the exposure tool and with the attenuating phase shift mask (PSM) pattern. Furthermore, the parameters change during inspection of the attenuating PSM because the mask is examined under illumination conditions which differ from the exposure illumination. An algebraic model which encompasses this large set of variables is derived by considering the electric fields under the mask to be a combination of the electric fields from the feature and defect. These fields are then combined according to the mutual coherence function for the mask illumination. A notable difficulty is the relative phase shift due to defocus between large and small features. The model is shown to be valid for defects up to 0.35 /spl lambda//NA by comparison to SPLAT. Experimental verification is made for defects impacting a 6% transmitting PSM for 0.35-/spl mu/m features at i-line. The reliability of the model is illustrated by giving rules of thumb for defect printing in attenuating PSMs. >

Filter design methodology for defect detection in wafer inspection

A methodology for the design of a filter to image the bottom of a contact is developed through simulation. Without a filter, the topography of the hole causes normally incident light to refract out of the hole. By conditioning the light with a filter, however, light is launched into the hole that is capable of propagating down the hole and imaging the bottom. When using this filter, the reflected aerial image differs by as much as 10% when the contact hole is under etched by 0.05 μm. Although, this initial success shows that the filter is capable of imaging the bottom of the hole, the introduction of process variations such as a larger hole diameter or a thinner oxide cause large deviations in the aerial image.

Analysis of sub-wavelength sized OPC features

Abstract Rigorous 3-D electromagnetic analysis is used to examine the actual near field characteristics and assess the resulting effects on imaging of light passing through serifs and hammer head line-ends in optical proximity correction (OPC) of photomasks. The simulation is carried out using the finite-difference time-domain simulator TEMPEST using up to 5 million nodes and 150M of memory. OPC features which are contiguous and share sides with open features transmit very effectively, whereas OPC features which are isolated by chrome stringers or are cut-off by nearly touching chrome corners are much less effective and require rigorous electromagnetic analysis.

Role of illumination and thin film layers on the printability of defects

The roles of projection printer illumination and resist/substrate layers in the printability of defects are explored in a physical based approach. In this approach the mutual coherence is first calculated, and then the impact of a defect on a feature is calculated using the image perturbation approach. The concerns are that the use of modified illumination and high NA divergence and reflection of the fields in a thin-film might affect the tendency of a defect to print. Results of this extension for various illuminations, thick resists and antireflection coatings are compared with rigorous thin-film SPLAT simulations. The mutual coherence function (MCF) was found to be strongly dependent of the type of illumination and on the mask geometry; however, the MCF was found to be weakly dependent on the presence of substrate material. This indicates that certain types of illumination sources are more defect tolerant and that the substrate is insignificant in preventing defect printing.

Effect of the partial coherence on reflective notching

This paper explores through numerical simulation and experiment the role of partial coherent illumination on reflective notching in optical projection printing over a nonplanar substrate. The intensity in the photoresist when patterning a gate line over a LOCOS generated active area well covered in polysilicon was simulated with the rigorous electromagnetic simulation program TEMPEST-PCD which includes partial coherence through a decomposition method. The simulation results are compared to experimental SEM pictures of developed photoresist lines for patterning a gate over a trench with varying moat widths when using a DUV 248 nm stepper with a NA of 0.5 for a (sigma) of 0.3 and 0.6. Variations in linewidth of both plus and minus 20% are observed due to the interactions with the underlying topography. The simulations give physical insight into trends seen in the experimental critical dimension (CD) measurements. Simulations indicate that most of the effects observed are due to lateral specular reflection from the underlying structure aiding in the dissolution of the resist near the feature. These trends seen in simulation and in experiment indicate that the topography has a similar impact on the CD regardless of the coherence.

Effects of wafer topography on the formation of polysilicon gates

The effects of the active area topography on the critical dimension of a polysilicon gate are examined using the rigorous electromagnetic simulation program TEMPEST and SEM pictures of developed photoresist lines. The recessed nature of the active area contributes uneven exposure and hence a nonuniform critical dimension of the polysilicon gate as it reaches the end of the active area and as it traverses over the field oxide. This gate resist formation process over the active area was simulated and patterned in experiments of 0.35 micrometers features using an i-line system with an NA of 0.5 and a sigma of 0.6. The dominate effect of linewidth narrowing near the end of the active region is shown by simulation to track the three dimensional geometrical reflections of light illuminating the bowl shaped region near the end of the active area where the polysilicon covers the field oxide transition. The addition of a TiN anti-reflective coating (ARC) layer and the addition of dye to the photoresist were investigated by studying the energy deposited into the photoresist using TEMPEST as well as SEM pictures of photoresist lines. In both experiment and TEMPEST simulation, a TiN (ARC) layer was found to alleviate the problem while dyed photoresist showed little improvement.

Impact of optical thin film effects on CD control in DUV lithography

We examine the effects of optical thin films on CD control in DUV lithography. A number of different substrates typical of device lithography levels were fabricated. Film stack configurations were chosen for their optical properties as well as process integration compatibility. All wafers were then exposed with an open field to measure swing amplitude, followed by a resolution test pattern to measure proximity effect and dose latitude. The resolution test pattern consisted of 250 nm coded features varying in pitch from 500 to 1750 nm. Results will be presented for optimum CD control over four types of film stacks as a function of resist thickness. In general, it is always best to operate at an extremum of the swing curve in resist thickness. However, the underlying film stack effects the CD control differently at Emin versus Emax. We present the optimum exposure conditions for a number of film stacks, resist thickness, and types of illumination (conventional and annular). These results were corroborated by optical imaging software enhanced to incorporate resist thickness and thin film effects. A synopsis of the role of substrate optical properties on CD control will be presented.