Microsphere-Assisted Spectroscopic Reflectometry (MASR) - A Tool for Nanoscale Characterization in Semiconductor Devices

Introduction

The semiconductor industry's relentless pursuit of device miniaturization continues to push the boundaries of materials, architectures, and manufacturing processes. 3D integration and aggressive dimensional scaling have significantly boosted the capabilities of memory and logic devices. However, this increased complexity has also posed new challenges for the optical metrology techniques relied upon to monitor and control semiconductor fabrication.

Established spectroscopic methods like reflectometry and ellipsometry have long been widely adopted for measuring the critical dimensions (CDs) and 3D structures of semiconductor features. These optical techniques offer attractive advantages, including high measurement speed, low cost, and minimal sample damage. However, to keep pace with the ever-shrinking size of device features, these metrology tools need to drastically reduce their probe beam spot size - ideally to dimensions smaller than 1/10th the size of a typical DRAM memory cell (around 30x40 μm²). This would enable CD monitoring and control at the individual memory element level, rather than just across larger die regions.

To address this challenge, recent research has introduced a novel spectroscopic measurement system that leverages the super-resolution effect of microspheres. This approach, termed MASR (Microsphere-Assisted Spectroscopic Reflectometry), has been demonstrated to shrink the probe beam diameter down to around 210 nm while maintaining a high signal-to-noise ratio - a key requirement for practical implementation. Through simulations and experiments, the feasibility of the MASR technique has been established, suggesting great promise in tackling the current bottleneck of tracking critical dimension variations at the individual memory cell level in advanced logic and memory devices.

Experimental set-up of MASR

The MASR system uses a broadband light source directed towards the sample through a microsphere. The reflected light from thin film layers interferes and is captured by an optical fiber, fed into a spectrometer.

INTINS offers the QEPro spectrometer series suitable for this technique. The QEPro is a next-generation scientific-grade spectrometer with high sensitivity and low stray light performance, making it ideal for a wide range of low light level applications. Its enhanced thermal design promotes excellent spectrometer wavelength stability, and the QEPro's triggering functions deliver accurate timing and synchronization between the spectrometer and other devices. The Hamamatsu FFT-CCD back-thinned detector at the core of the instrument is distinguished by high quantum efficiency (up to 90% maximum) and low etalon characteristics. Its design significantly improves the signal-to-noise ratio (>1000:1) and signal processing speed. The QEPro is capable of low light level detection and long integration times, ranging from 8 milliseconds to 15 minutes, with virtually no spectral distortion.

Figure 1. QEPro spectrometer.

INTINS also provides complementary fiber optic accessories and the HL-2000 broadband light source, covering 360-2400 nm. The HL-2000 features an integrated fan for stability, a filter holder, and in some models, a shutter and long-lifetime bulb, further expanding the capabilities of the MASR system.

Figure 2. HL-2000 light source and optical fiber.

Method and Results

The MASR system leveraging the "photonic nanojet" effect of microspheres to achieve sub-diffraction-limit spot sizes down to 210 nm, while maintaining acceptable signal-to-noise ratios. This allows MASR to measure critical dimensions and detect defects in extremely small areas like memory cell corners that are inaccessible to conventional SR systems. The MASR system's key advantage is enhanced SNR for spectral measurements. Conventional high-magnification optics suffer SNR degradation due to decreased optical power at the detector. However, the MASR's photonic nanojet effect concentrates both incident and reflected light, minimizing SNR loss at high magnifications. This enables high-resolution spectral analysis without the typical SNR tradeoff. Overall, MASR has significant potential to address metrology and inspection challenges in advanced semiconductor manufacturing.

One of the typical practical applications of MASR in the semiconductor industry is surveyed as follows. The DRAM sub-word line driver (SWD) area, featuring structures under 200 nm, was imaged using the MASR system. This region, smaller than the optical resolution limit of conventional microscopy, was calculated to have an approximate 280 nm resolution. However, the MASR technique was able to clearly resolve 57 nm and 146 nm lines that were previously undetectable (Fig. 3). The SWD area is critical for controlling parameters like gate oxide thickness and dent depth, which impact dielectric properties. Conventional ellipsometers struggle to directly measure this sub-100 nm region due to their large spot sizes. In contrast, the MASR system enables super-resolution imaging and spectral analysis within this critical DRAM feature area.

The MASR system was used to evaluate the spectral reflectance of a DRAM cell block, shown in Fig. 5a. Measuring the edges and corners of cell blocks is increasingly important, but challenging for conventional spectroscopy. Defects often occur at cell block edges during DRAM fabrication, making locality control critical. MASR can non-destructively measure spectral reflectance in these edge areas, unlike ellipsometers or imaging spectrometers.

Figure 3. Images of the SWD area in the DRAM imaged by a ×100 with 0.9 N.A. and b MASR. c Corresponding SEM image. 57 nm CD indicated by yellow arrows and 146 nm CD indicated by red arrows.

Figure 4. a ×20 image of DRAM cell array. b Principal component analysis map of the reflectance obtained by MASR. The red dots (#1 to #5) refer to the center positions of the cell block, while the black dots (#6 to #10) refer to the edge area from the outside of the cell block. c Spectral reflectance for positions #1 to #5 and d #6 to #10 in b.

As depicted in Fig. 5a, b, the reflectance at the center and edge of the cell block was compared for five positions by MASR. The distance between each position was 0.5 μm, which was considered appropriate for observing reflectance changes in the edge area, as DRAM edge defects often occur within 2 μm from the edge.

The central spectral reflectance shown in Fig. 5c varied slightly between positions, indicating small structural dimensional changes. Conversely, the reflectance at the edge shown in Fig. 5d significantly varied between positions, indicating substantial changes in the structure and the occurrence of either defects or imperfect structures.

The spectral map rapidly collapsed near the edge area of the cell block in Fig. 5b, demonstrating the MASR system's ability to monitor the critical 2 μm edge region, which cannot be measured using conventional spectrum systems. Microsphere super-resolution has the potential to be used in the semiconductor industry, which requires the measurement of many steps and structures.

Conclusion

              In summary, the MASR system enabled super-resolution imaging and spectral characterization of critical sub-200 nm features within the DRAM sub-word line driver (SWD) area. This region is challenging to measure directly using conventional techniques like ellipsometry due to the small feature sizes and localized nature of the structures. The MASR approach overcame the optical resolution limit, clearly resolving 57 nm and 146 nm lines that were previously undetectable. Additionally, the MASR system's photonic nanojet effect enhanced the SNR of spectral measurements at high magnifications, providing a powerful tool for monitoring edge effects and other localized phenomena in advanced semiconductor devices.