Using Nanoscale IR Spectroscopy to Characterize Semiconductor Materials and Processes

Submicron and nanoscale chemical identification of semiconductor materials, particularly organic ones, poses a considerable challenge in the analysis of devices and also in the process control and failure analysis of environments.

In view of this, semiconductor manufacturers use an array of methods to detect chemical components.

Energy dispersive X-ray spectroscopy (SEM/EDX) in combination with scanning electron microscopy is the industry standard in surface analysis and enables semi-quantitative elemental analysis with nanometer-scale spatial resolution. However, while this elemental analysis helps in gaining valuable chemical insights into surface contaminations and defects, it is rather limited when it comes to detecting organic materials.

Tapping AFM-IR spectra clearly identifying each chemical component of a block copolymer material.

Tapping AFM-IR spectra clearly identifying each chemical component of a block copolymer material.

Infrared, or IR, spectroscopy is a robust method developed for the chemical characterization of organic materials that otherwise cannot be easily detected through SEM/EDX. Conversely, Abbe diffraction laws limit conventional IR spectroscopic techniques to spatial resolutions between 3 and 10 μm, based on the type of method used.

The extensively used nanoscale imaging technique — atomic force microscopy or AFM — offers users with a high spatial resolution topographic map of the surface of a sample. To date, one major disadvantage of the AFM technique was its inability to chemically identify or characterize the material beneath the tip.

Resonance-Enhanced AFM-IR and Tapping AFM-IR

These limitations are overcome with nanoscale IR spectroscopy by integrating IR and AFM techniques to realize nanoscale FTIR spectroscopy through photothermal IR spectroscopy (PTIR). Nanoscale IR spectroscopy integrates the nanoscale capabilities of AFM with the accurate chemical identification of IR spectroscopy to chemically identify sample components with monolayer sensitivity and with less than 10 nm spatial resolution, breaking the diffraction limit by more than 100x.

AFM-PTIR absorption spectra are direct measurements of the absorption of a sample, independent of other intricate optical properties of the sample and tip. Therefore, AFM-PTIR spectra have been shown to correlate well with traditional bulk IR spectra.1

Nanoscale Organic Contaminants

Nanoscale chemical characterization capabilities on the nanoIR3 are demonstrated by preparing and analyzing contaminated silicon wafers using known materials that are commonly found in semiconductor fabrication settings. For every sample, high-resolution tapping mode AFM images were obtained to identify the contaminants, and this was followed by making respective nanoIR measurements.

nanoIR measurements

In Figure 1a, the AFM height image shows the thickness difference (20–100 nm) of the contaminant residue (a human skin tissue, in this case) on the wafer. After a point of interest is located in the AFM image, the probe is placed in the required location and the nanoIR spectra are acquired by scanning the laser wavelength, as shown in Figure 1b.

AFM height image (a) and corresponding AFM-IR spectra (b) of organic residue on a silicon wafer show clear Amide I and II bands, indicating human skin residue. The colored line spectra corresponded to colored measurement pixels in the AFM image.

Figure 1. AFM height image (a) and corresponding AFM-IR spectra (b) of organic residue on a silicon wafer show clear Amide I and II bands, indicating human skin residue. The colored line spectra corresponded to colored measurement pixels in the AFM image.

Spectra were subsequently collected at sites that had different sample thickness. As anticipated, the observed IR intensities varied with the thickness of the sample; yet, the overall signal-to-noise ratio is adequate enough to precisely detect the material, even at a thickness of 20 nm, indicating the exceptional sensitivity in the detection of thin samples.

Thanks to the new FASTspectra capabilities on the nanoIR3, spectra over the entire IR tuning range can be acquired faster, thereby enabling a reduction in spectral acquisition time by a factor of 10. This is achieved by the laser source sweeping over its range, whereas the cantilever oscillation amplitude is simultaneously determined with the change in wavelength, as shown in Figure 2.

AFM height image (a) and resulting nanoIR spectra (b) from a contaminant on a bare silicon wafer. The resulting match from the FTIR library identifies the contaminant as Poly(ethylene terephthalate).

Figure 2. AFM height image (a) and resulting nanoIR spectra (b) from a contaminant on a bare silicon wafer. The resulting match from the FTIR library identifies the contaminant as Poly(ethylene terephthalate).

In order to show the correlation with traditional FTIR spectra, the acquired spectra were evaluated against a standard FTIR database (KnowItAll, Bio-Rad Inc.). The ~30 nm tall contamination residue was positively identified as polyethylene terephthalate (PET), a polymer typically used in polyester fabrics, which can be seen in Figure 2c.

Process-Induced Defects2

Low-k α-SiOC:H/Cu interconnects serve as a valuable example with regards to the susceptibility of contamination at the time of manufacturing because of the sensitivity exhibited by low-k α-SiOC:H in response to subtle chemical changes caused during fabrication. Figure 3 shows the structure of the interconnect device developed for this experiment. The low-kα-SiOC:H’s average width differs between 1650 and 330 nm, which is well within the spatial resolution limits of traditional IR spectroscopy, testing the potential of nanoscale IR spectroscopy.

Schematic diagram of the cross-section of the Low-K SIOC:H—copper interconnect structures fabricated for this experiment.

Figure 3. Schematic diagram of the cross-section of the Low-K SIOC:H—copper interconnect structures fabricated for this experiment.

An AFM topography image of the low-kα-SiOH/Cu structure is shown in Figure 4a. The markers on the image indicate the locations for subsequent nanoIR measurements. Each nanoIR spectrum was acquired from a different part of the interconnect, including copper and both the exposed regions of the α-SiOC:H dielectric. In addition, spectra were collected from wide (1650 nm) and narrow (390 nm) regions of the α-SiOC:H interconnect domains.

(a) 10 µm X 5 µm AFM topography image illustrating the regions from which the AFM-IR spectra were collected. Note the color of the marker corresponds to the color of the individual spectra. (b) AFM-IR C-H stretch spectra from the patterned a-SiOC:H/Cu interconnect.

Figure 4. (a) 10 μm X 5 μm AFM topography image illustrating the regions from which the AFM-IR spectra were collected. Note the color of the marker corresponds to the color of the individual spectra. (b) AFM-IR C-H stretch spectra from the patterned α-SiOC:H/Cu interconnect.

In Figure 4b, a close examination of the spectra shows slight variations in the C-H stretch area of the wide and narrow regions of the α-SiOC:H dielectric. The asymmetric methyl CH3-stretching vibration at 2968 cm−1 showed reduced intensity in relation to the asymmetric methylene CH2-stretching vibration at 2924 cm−1 for the α-SiOC:H dielectric’s narrow regions. For the α-SiOC:H dielectric’s wide regions, the peak intensity ratio for the CH2/CH3 modes was closer to that noted by bulk FTIR as well as nanoIR of an unpatterned α-SiOC:H dielectric control sample, as shown in Figure 5.

(a) T-FTIR and AFM-IR of the C-H stretching band from 1 µm a-SiOC:H. (b) spectra of the asymmetric Si-O-Si stretch and symmetric Si(CH3)x deformation mode from both narrow (390 nm) and wide (1,650 nm) regions of the a-SiOC:H dielectric. Note: a T-FTIR spectrum of the unpatterned a-SiOC:H dielectric is also included for comparison.

Figure 5. (a) T-FTIR and AFM-IR of the C-H stretching band from 1 μm α-SiOC:H. (b) spectra of the asymmetric Si-O-Si stretch and symmetric Si(CH3)x deformation mode from both narrow (390 nm) and wide (1,650 nm) regions of the α-SiOC:H dielectric. Note: a T-FTIR spectrum of the unpatterned α-SiOC:H dielectric is also included for comparison.

Spectra were also acquired for the Si-O-Si stretching region to confirm that the variations noticed in the C-H spectral regions of wide against narrow α-SiOC:H regions are not caused by other phenomena, like optical effects.

Comparative variations were also seen in the nanoIR spectra of the Si-O-Si stretch from the narrow versus wide regions of the α-SiOC:H dielectric, analogous to the C-H stretch (see Figure 5b). The main observation is the relative variation in absorbance for the Si-O-Si cage mode at 1050 cm1, that is more perceptible in the narrow regions of the α-SiOC:H dielectric in relation to the wide regions of the α-SiOC:H dielectric.

The cage mode for the 320 nm wide patterned α-SiOC:H region, in relation to the bulk transmission FTIR spectra of the unpatterned α-SiOC:H dielectric, is also downshifted to lower wavenumbers and the Si-O-Si stretching mode network seems to have a more narrow line width.

These collective results support the notion that the major modification of the α-SiOC:H chemical structure takes place in the narrow regions of the α-SiOC:H/Cu interconnect structure in relation to the wide regions. Various analyses of low-kα-SiOC:H materials have demonstrated that these materials have a tendency to lose terminal methyl (CH3) groups at the time of plasma etching, CMP, and ashing steps used for fabricating inlaid Cu wiring. Within low-kα-SiOC:H dielectrics, the loss of such terminal organic groups leads to the creation of new chemical bonds and a composition more like SiO2.

In this respect, the decr No-fault scanner, sensor, and transducer repair eased the intensity of the CH3 mode in the narrow regions of the α-SiOC:H dielectric, as illustrated in Figure 4b, and is consistent with the modification or loss of terminal CH3 groups at the time of the plasma etch and ashing processes used for patterning the dielectric. Moreover, it is also consistent with the greater absorbance noticed for the Si-O-Si cage mode in the narrow regions of the α-SiOC:H dielectric.

Characterizing Advanced Semiconductor Structures

Over the past few decades, ongoing advancement in semiconductor process technology has resulted in the development of devices with ever-shrinking nanometer-scale domains. As depicted in Figure 6, directed self-assembly, or DSA, of block copolymers (BCPs) is one among the top candidates for state-of-the-art lithography, offering sub-14 nm nanostructures with controlled placement.3,4

Illustrating “lift-off” process to create epitaxial DSA patterns. The AFM image highlights the linear array of block copolymers.

Figure 6. Illustrating “lift-off” process to create epitaxial DSA patterns. The AFM image highlights the linear array of block copolymers.6

This innovation in the fabrication of nanoscale patterns using DSAs with 10–20 nm pitch promotes an increasing need for characterizing alignment, location, and size, and also material-specific identification with high spatial resolution.

New advancements in the AFM-IR method, like Tapping AFM-IR, drive the spatial resolution limit below 10 nm. As a result, this method is suitable for the chemical characterization of DSA components and defects for failure analysis (FA).5

Tapping AFM-IR was applied in this experiment to examine various block copolymers regularly used for fabricating directed self-assemblies on Si wafers. AFM topography and following nanoscale chemical analysis of polystyrene-poly(2-Vinyl Pyridine) block copolymer [PS-P2VP] are shown in Figure 7.

Chemical characterization of PS-P2VP block copolymer sample by Tapping AFM-IR; (a) Tapping AFM height image; (b) Tapping AFM-IR spectra clearly identifying each chemical component; (c) Tapping AFM-IR overlay image highlighting both components (PS@ 1492 and P2VP@ 1588); and (d) profile cross section highlighting the achievable spatial resolution, 10 nm. Sample courtesy of Dr Gilles Pecastaings and Antoine Segolene at University of Bordeaux.

Figure 7. Chemical characterization of PS-P2VP block copolymer sample by Tapping AFM-IR; (a) Tapping AFM height image; (b) Tapping AFM-IR spectra clearly identifying each chemical component; (c) Tapping AFM-IR overlay image highlighting both components (PS@ 1492 and P2VP@ 1588); and (d) profile cross section highlighting the achievable spatial resolution, 10 nm. Sample courtesy of Dr Gilles Pecastaings and Antoine Segolene at University of Bordeaux.

The high-resolution topography image reveals 3–4 nm tall lamellar features with a pitch of 50 nm. Following chemical analysis with Tapping AFM-IR measurements examines the chemical compositions in the nanopattern. Furthermore, AFM-IR spectra directly match with FTIR absorption bands and emphasize the components’ distinct chemical signature. Tapping AFM-IR imaging at absorption bands relevant to every component, as indicated in Figures 7(c), emphasizes the total distribution, with an observed spatial resolution of 10 nm. It was established that the taller features were polystyrene blocks (domains) and the matrix was P2PV.

Further measurements were carried out on DSAs with different functional molecules and nanopatterns. Chemical mapping of spherical DSAs consisting of PS-b-PMMA as well as PS-b-P4VP [polystyrene-poly-4-vinylpyridine block copolymer] is shown in Figure 8. The measurements produced excellent chemical specificity with high spatial resolution, approximately 4 nm, as illustrated in the Figure 8 inset.

Tapping mode height images with PS-b-PMMA (a) and PS-b- P4VP directed self-assemblies (c). The nanopatterns are spherical with 10–20 nm domains. Tapping AFM-IR image at 1730 cm-1 highlights PMMA beads embedded in the polystyrene matrix (b). The ratio image at 1492 cm-1/1598 cm-1 illustrates distribution of PS relative to P4VP in the patterned wafer (d). The inset in (b) shows the IR intensity variation along the white dashed line in the main panel. A spatial resolution of 4 nm is observed.

Figure 8. Tapping mode height images with PS-b-PMMA (a) and PS-b- P4VP directed self-assemblies (c). The nanopatterns are spherical with 10–20 nm domains. Tapping AFM-IR image at 1730 cm−1 highlights PMMA beads embedded in the polystyrene matrix (b). The ratio image at 1492 cm−1/1598 cm−1 illustrates distribution of PS relative to P4VP in the patterned wafer (d). The inset in (b) shows the IR intensity variation along the white dashed line in the main panel. A spatial resolution of 4 nm is observed.

Conclusion

This article shows how the Tapping AFM-IR was effectively used for differentiating the chemical footprints of a number of nanoscale lithographic patterns containing intricate molecular assemblies with an incredible spatial resolution of 4 nm. In addition, nanoscale IR spectroscopy excels in the characterization of nanoscale defects caused during the fabrication processes of low-k interconnects.

When compared to thick low-k regions, considerable chemical differences were seen for narrow low-k regions. Since a wide spectral range is available for nanoIR, these variations were seen in the CH as well as Si-O-Si stretching regions, where traditional IR spectroscopy, with its limited spatial resolution, cannot expose these defects.

Finally, nanoIR spectra collected from typical examples of nanoscale contamination were easily detected as biological material from the presence of the amide absorption bands. In addition, any spectra of unknown materials produced using nanoscale IR spectroscopy can be easily searched in a traditional FTIR database for easy identification.

From these results, it is clear that nanoscale IR spectroscopy through AFM-IR demonstrates excellent potential as a characterization tool for sophisticated semiconductor materials and processes.

References

  1. Dazzi, A. and Prater, C. B., “AFM-IR: Technology and Applications in Nanoscale Infrared Spectroscopy and Chemical Imaging”, Chem. Rev., 117 (2017), pp. 5146-5173.
  2. Lo, M.; Dazzi, A.; Marcott, C.; and King, S., “Nanoscale Chemical- Mechanical Characterization of Nanoelectronic Low-k Dielectric/Cu Interconnects” ECS Journal of Solid State Science and Technology 2015, 5 (4), 3018-3024.
  3. Yang, G. W. Wu, G. P. Chen, X. X. Xiong, S. S. Arges, C. G. Ji, S. X. Nealey, P. F. Lu, X. B. Darensbourg, D. J. and Xu, Z. K.” “Directed Self-Assembly of Polystyrene-b-poly(propylene carbonate) on Chemical Patterns via Thermal Annealing for Next Generation Lithography”, Nano Lett., 17 (2017), pp. 1233-1239
  4. Bates, F. S. and Fredrickson, G. H., “Block Copolymer Thermodynamics: Theory and Experiment”, Annu. Rev. Phys. Chem., 41 (1990), pp. 525-557
  5. Roy, Anirban et al, ISTFA conference, “Latest Advancements in Nanoscale IR Spectroscopy for Characterization and Failure Analysis of Electronic Devices” 2017.
  6. Figure modified from figure 3, Seong, S.J., et al., “Directed self-assembly of block copolymers for next generation nanolithography”. doi10.1016/j.mattod.2013.11.002.

This information has been sourced, reviewed and adapted from materials provided by Bruker Nano Surfaces.

For more information on this source, please visit Bruker Nano Surfaces.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Bruker Nano Surfaces and Metrology. (2023, April 10). Using Nanoscale IR Spectroscopy to Characterize Semiconductor Materials and Processes. AZoNano. Retrieved on December 09, 2024 from https://www.azonano.com/article.aspx?ArticleID=5138.

  • MLA

    Bruker Nano Surfaces and Metrology. "Using Nanoscale IR Spectroscopy to Characterize Semiconductor Materials and Processes". AZoNano. 09 December 2024. <https://www.azonano.com/article.aspx?ArticleID=5138>.

  • Chicago

    Bruker Nano Surfaces and Metrology. "Using Nanoscale IR Spectroscopy to Characterize Semiconductor Materials and Processes". AZoNano. https://www.azonano.com/article.aspx?ArticleID=5138. (accessed December 09, 2024).

  • Harvard

    Bruker Nano Surfaces and Metrology. 2023. Using Nanoscale IR Spectroscopy to Characterize Semiconductor Materials and Processes. AZoNano, viewed 09 December 2024, https://www.azonano.com/article.aspx?ArticleID=5138.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.