Issue

Laterally resolved phase measurements at 45nm using photomask phase metrology

01/01/2008

EXECUTIVE OVERVIEW

As the lithography process moves toward the 45nm and 32nm nodes, phase control on the mask is becoming more important than ever. To ensure an accurate printing, both attenuated and alternating phase shift masks (PSMs) need precise control of the phase shift as a function of both pitch and target sizes. Simulations show that the phase shift in the image plane of a microlithography scanner is strongly affected by numerical aperture (NA), mask pitch, 3D mask effects, and polarization, especially if the feature sizes come close to the imaging wavelength. A new phase metrology system that overcomes the limitations of currently existing tools has been developed.

The new optical metrology tool, Phame, enables the industry to perform in-die phase measurements on alternating PSM (altPSM), attenuated PSM (attPSM), and CPL (chromeless phase lithography) masks down to 120nm half-pitch at the mask [1-3]. On-axis measurement results show the dependency of scanner phase on pitch and the impact of polarization on scanner phase. An off-axis measurement procedure has been developed and the first off-axis measurement results on 45nm node features will be presented.

The use of PSMs combined with high NA and specially adapted illumination conditions drives 193nm lithography down to the 45nm node and beyond. The industry faces the challenge that mask complexity increases steadily, mask specifications tighten, and process control becomes extremely important. Both altPSM and attPSM require accurate and precise phase control to ensure best CD printing performance at wafer lithography steps. For example, the International Technology Roadmap for Semiconductors specifies that the phase error of altPSMs should be ±1° in 2008.

Kirchhoff approximation and rigorous EMF simulations show that imaging effects and rigorous effects (3D mask effects) need to be considered when going down to the 45nm node. Using on-axis in-die phase measurements, we showed the dependence of scanner phase on pitch and the impact of polarization. The off-axis measurement procedure will be explained and the first off-axis phase results on 45nm node test features will be presented.

Simulation

To understand the behavior of the optically-relevant phase shift occurring in the image field of a scanner, we performed simulations both in a rigorous and a Kirchhoff regime. The simulation results over pitch showed that the scanner relevant phase shift of an ideal altPSM in a Kirchhoff simulation coincides with the etch depth equivalent (near field) phase shift for large features only when the size on the mask exceeds the illumination wavelength by an order of magnitude. For decreasing feature sizes, the effective scanner-relevant phase shift increasingly oscillates around the near field phase. These oscillation effects are caused by the loss of higher diffraction orders due to the limitations of the scanner NA [1].

An RCWA (rigorous coupled wave analysis)-based simulator was used to research the rigorous effects of an altPSM. First, the 3D topography of the mask feature was optimized to generate a 180° phase shift in the image field for a mask pitch of 500nm that was printing a pitch at a wafer level of 62.5nm. The optimized mask parameters were set to an etch-depth equivalent phase shift of 162°, an undercut of the quartz etched line of 12nm on each side, and a chrome line duty cycle of 0.67. These parameters were held constant for simulations of varying pitch.

Figure 1. Through-pitch behavior: a) Kirchhoff and rigorous simulation of scanner relevant phase for an altPSM of etch depth equivalent (near field) phase of 162° for varying pitch. The features were optimized to have a scanner equivalent phase of 180° at 62.5nm print pitch. And b) aperture images showing the imaging effect of capturing higher diffraction orders at approximately the pitch of a 180nm wafer level.Click here to enlarge image

To see the deviations of 3D mask effects from the phase oscillations caused by the imaging aperture, the through-pitch behavior in a Kirchhoff regime is plotted in Fig. 1 as well. For small features approaching the resolution limit of the imaging system (here mask NA 0.4), rigorous mask effects cause phase deviations up to 9°. Taking both effects into account, it is clearly seen that the etch depth equivalent phase differs by 16° from the scanner phase, which is quite large.

Additionally, it was shown that the largest process window is achieved if the scanner relevant phase shift is at 180° [1], which underlines the necessity of a lateral-resolving phase metrology tool measuring the scanner-relevant phase.

Phase metrology

For large features, the new phase metrology system takes over the capability of currently existing interferometer-based tools. In addition, the high resolution capabilities realize phase shift measurements in any in-die feature of the active mask area for on- and off-axis applications for altPSM, attPSM, and CPL masks, capturing diffraction limitations, rigorous effects, and polarization effects.

The optical beam path of the new metrology system allows actinic phase measurements of 193nm photomasks with a mask side NA up to 0.4, which would be 1.6NA scanner equivalent at the wafer. This enables full compatibility to future 193nm immersion scanners down to the 32nm node.

Its 193nm laser combined with a low sigma illumination unit generates a coherent illumination (i.e. single source point) of the mask, which is handled face down. On-axis or off-axis illumination can be used depending on the PSM type.

Partial coherent illumination settings of a scanner can be sampled in consecutive measurements of adjustable intervals, allowing phase control under scanner-relevant illumination settings including polarization. The CCD-camera is in the same position as the wafer in the actual scanner. Phase information is obtained through dedicated phase manipulation by pupil filter and software algorithms. In addition to in-die phase shift, the tool also measures in-die transmission.

Off-axis measurement/high resolution phase

Off-axis phase measurement is realized by applying consecutive measurements of single source points according to the scanner relevant illumination setting (e.g., dipole illumination is the measurement of two opposite source points). Phame measures the scanner equivalent phase and amplitude in the image plane for each coherent source point. Phase shift value extraction requires symmetry of the diffraction spectrum. Comparing the diffraction spectrum-in particular, the 0th and 1st, -1st diffraction orders for an altPSM and an attPSM-the following applies. For an altPSM using on-axis illumination, the -1st and 1st diffraction order is captured in the scanner NA and ideal two-beam interference occurs. The diffraction spectrum lies perfectly symmetrical in the pupil. To achieve the same printing information with an attPSM or CPL mask using the same scanner NA, off-axis illumination is required.

Thus, for an attPSM the 1st, respectively -1st, and 0th diffraction orders (i.e., the combination is 1 and 0, or -1 and 0) are captured in the pupil. In general, the diffraction spectrum for off-axis illumination might be asymmetrical in the pupil depending on the feature pitch. This asymmetry causes large phase tilts in the image plane, making a phase shift value extraction impossible. Simulation and measurement results show that the symmetry of the diffraction spectrum depends on the pitch, and the symmetry determines the phase tilts in the image plane [2]. To get rid of the tilts, a symmetric diffraction spectrum is required to extract phase shift information. Zeiss has developed a new method for off-axis phase shift value extraction that is called high-resolution phase.

The diffraction spectrum of each single source point is measured in the new phase metrology system. Due to off-axis illumination, the 0th diffraction order for each source point is shifted in the pupil with respect to the optical axis. For each source point, the electrical field is determined in the phase metrology tool. The zero diffraction orders of each source point are algorithmically shifted in such a way that they fall together in the optical axis of a fictive pupil and the resulting electrical fields are coherently merged. By applying this procedure, a symmetric diffraction spectrum of the mask is obtained and at the same time, the original mask side NA is enlarged by the off-axis angle.

The symmetric diffraction spectrum that is generated by the procedure described above is now propagated into the image plane and a high-resolution phase image and a high-resolution intensity image is obtained. The off-axis high-resolution phase shift is sensitive to the diffraction spectrum and mask phase errors. It can be used for design verification, defect analysis, and process control.

Measurement results

The new phase metrology system has the capability of existing tools in that it can measure large reference features with high reproducibility. Figure 2 shows the on- and off-axis measurement results for a large reference feature. The phase shift result of 178° for on-axis and 180° for off-axis corresponds to the expected phase difference of 2.3° due to off-axis angle and the light path difference caused by the angled geometry. Additionally, excellent reproducibility values <0.15° (3σ) can be reported for both measurements.

Figure 2. Phase image of a large reference feature with on-axis (left) and off-axis (right) illumination.Click here to enlarge image

Figure 3 shows on-axis phase shift images of a contact hole array on an altPSM. Moreover, on-axis investigations of phase shift behavior on an altPSM over pitch and duty cycle had been performed showing significant phase variations [1].

Figure 3. Phase image and phase profile for a 500nm contact hole array, altPSM.Click here to enlarge image

Figure 4 shows phase measurement results on an altPSM through pitch for changing polarization direction of the illuminating light. For pitches above 300nm at wafer, the phase shift approaches 180°. Below pitches of 300nm at wafer level, strong phase shift deviation over pitch up to 10° can be observed. Additionally, the impact of polarization on phase shift was investigated [3]. For small print pitches of 100nm at the wafer level, phase differences up to 7° between TM and TE polarization have been measured.

Figure 4. Phase measurements through pitch for changing polarization direction of the illuminating light, altPSMClick here to enlarge image

Off-axis measurements were done on a 6% MoSi mask containing 45nm (wafer level) lines and spaces of varying pitch from 1:1-1:3 combined with an isolated line. The high-resolution phase shift values for the dense lines and the isolated line were extracted. For the dense lines, a variation of phase shift up to 10° over pitch was observed.

Figure 5. High resolution phase shift over pitch comparing dense lines/spaces vs. iso lines, attPSMClick here to enlarge image

Figure 5 shows that for pitch 1:1 and 1:2 the phase shift is close to 180°, whereas the phase shift drops down to 170° for the pitch 1:3. Comparing the phase shift values of dense lines with the phase shift values of the isolated line, a strong deviation of up to 40° was found, especially when the isolated line is combined with the dense lines of pitch 1:1. This effect is decreasing as pitch is increasing. In real production features, OPC would be used to account for those effects.

Conclusion

The extension of optical lithography to the 45nm node and beyond goes along with increased mask complexity and tightening of specifications. The proper use of PSM becomes more and more important and the phase shift needs to be quantified exactly in order to achieve accurate CD printing results during wafer processing. The methods currently available run into limitations because they are not able to consider diffraction limitations caused by scanner NA and mask pitch, as well as 3D mask effects. In the transition to the 45nm node and beyond, these effects play an important role and need to be considered. The new phase metrology system captures diffraction limitations, rigorous effects (i.e., a failure of the Kirchhoff approximation), and polarization effects. The new phase metrology system measures the phase shift in any in-die feature of the active mask area for on- and off-axis applications.

Beside the large feature measurement capability and on-axis in-die measurement capability, Zeiss has developed a high-resolution phase concept for off-axis illumination. High-resolution phase shift measurements on 45nm (wafer level) MoSi test features showed strong variations of phase shift over pitch. Additionally, significant variations in phase shift up to 40° were observed for dense lines vs. isolated lines. This effect decreases with increasing pitch.

The high resolution phase is sensitive to the diffraction spectrum and mask phase errors. Thus, the phase metrology tool can be used for process control as well as for R&D processes especially in terms of design and OPC verification.

Acknowledgments

The authors would like to thank Steffen Weissenberg and Mario Laengle for their support in the experiments and result discussion, and Thomas Scherübl, Frank Stietz, and Oliver Kienzle for their support and contribution during this project. Phame is a registered trademark of Carl Zeiss SMS GmbH.

References

K.M. Lee, M. Tavassoli, M. Lau, S. Perlitz, U. Buttgereit, T. Scherübl, “Study of Rigorous Effects and Polarization on Phase Shifting Masks Through Simulations and in-die Phase Measurement,” Proc. SPIE, SPIE 6518, 65181Y, 2007.
U. Buttgereit, D. Seidel, S. Perlitz, K. M. Lee, M. Tavassoli, “Laterally Resolved Off-axis Phase Measurements on 45nm Node Production Features Using Phame,” Paper 6730-118, Photomask Technology, 2007.
S. Perlitz, U. Buttgereit, T. Scherübl, D. Seidel, K.M. Lee, M. Tavassoli, “Novel Solution for In-Die Phase Control under Scanner Equivalent Optical Settings for 45nm Node and Below,” Proc. SPIE, Vol. 6607.66070Z, 2007.