Advanced Nanoscale Spectroscopy of 2D Materials Using Photothermal AFM-IR

When atoms and electrons are confined to an atomically flat plane, 2D materials take on distinct properties, such as exceptional strength and conductivity, not seen in their 3D counterparts. These properties can be leveraged to create new electronic device types and to advance the functionality of silicon-based chips.

To be able to customize material functionality, 2D materials can be stacked to create heterostructures.¹ High surface areas promote chemical reactivity, which is advantageous for catalysis, chemical/biological sensing, energy storage, and drug delivery.

Atomic force microscopy (AFM) becomes an effective tool when it comes to characterizing 2D materials, as AFM delivers nanoscale, high-resolution topographical, mechanical, and electrical property mapping. Photothermal AFM-IR (AFM-IR) introduces localized chemical identification with nanometer-scale spatial mapping by pairing infrared spectroscopy with AFM.

When combined, these functionalities coordinate both structure and chemistry, allowing for the in-depth study of intricate 2D material systems. This article provides an overview of how AFM-IR acts as a complementary method to scattering-type scanning near-field optical microscopy (s-SNOM).

This is achieved through FTIR-correlated spectra and chemical maps at the ∼10 nm scale, including direct absorption imaging of hBN phonon polaritons and stacking-specific contrast in graphene heterostructures.

From s-NOM to Photothermal AFM-IR

s-SNOM is a proven AFM-based method for recording chemical and optical measurements at the nanoscale. This is performed by focusing the incident light on the AFM tip apex, which serves as a local antenna. In s-SNOM, an interferometer detects scattered light transmitting the local complex refractive index (n, k).

Extracted optical constants give full access to the optical refractive index and absorption coefficient, which correlate to chemical composition and optical resonances. The localized light-matter interaction under the AFM tip supports <10 nm spatial resolution.

Extensive use of s-SNOM has been documented in the research of surface polaritons, which form when light (photons) interacts with the collective oscillations of charged particles, including electrons (plasmons) in conductors or the vibrational modes of atoms (phonons) in dielectrics.

s-SNOM makes it possible to visualize the real-space of confined surface polaritons on the 2D materials, exposing their wavelengths, propagation lengths, and interference patterns. For the study of 2D materials, photothermal AFM-IR is a relatively new technique in contrast to s-SNOM. Both techniques can be seen in the schematic diagram displayed in Figure 1.

Figure 1. Representations of the operation of (a) s-SNOM and (b) AFM-IR. Image Credit: Bruker Nano Surfaces and Metrology

AFM-IR focuses pulsed, tunable IR light onto the sample at the AFM tip location. When the IR wavelength matches an absorption band of the material, the AFM can measure cantilever oscillations driven by rapid local thermal expansion. The signal generated by photothermal AFM-IR directly measures sample absorption, and the resulting spectra are well aligned with those generated by traditional Fourier-transform infrared (FTIR) in transmission mode.

Early studies of AFM-IR performance for characterizing 2D materials were limited by comparatively weak signals due to their low thermal expansion coefficient and narrow dimensions. Advances in detection sensitivity (down to monolayers/single molecules) and spatial resolution (<5 nm) make AFM-IR a very effective nanoscale characterization tool for 2D materials.

AFM-IR offers a series of benefits over s-SNOM, including faster measurements, simpler interpretation, and displacement-based detection rather than referenced optical detection, making it well-suited to accelerating research on 2D materials.

Previous Studies of 2D Materials with AFM-IR

AFM-IR as a Complement to s-SNOM

AFM-IR has commonly been used to complement s-SNOM when studying 2D materials, including plasmon polariton in graphene monolayer, phonon polariton in hBN and MoO₃, and circular dichroism in 2D planar-chiral metamaterials.^2-6

For graphene monolayers, SPP interference fringes with full intensity at the periphery of the graphene wedge were displayed in the IR images recorded at 930 cm^-1 using both Tapping AFM-IR and s-SNOM.

This confirms a correlation of the s-SNOM SPP mode with AFM-IR tapping, as fringe spacing and intensity agree across both methods.² Phonon polaritons have been observed and analyzed in flakes as thin as 4 nm in hBN.⁴

AFM-IR for Unique Property Characterization

Compared to techniques that are purely dependent on optical near-field imaging, AFM-IR can measure properties that are more difficult to assess, such as heat dissipation mechanisms in graphene, non-radiative states in hBN, photothermal effects in thicker hBN, functionalized graphene, and MXene/graphene oxide hybrids:

• For graphene monolayers, AFM-IR spectra taken from a graphene flake and SiO₂ substrate display robust hybridization between graphene plasmon and SiO₂ phonon modes in the 900–1200 cm^-1 range, compatible with the photothermal expansion signal.²
• For hBN frustums (256 nm thick) of varying aspect ratios, AFM-IR spectra show a broad range of peaks, including some strong ones above 1550 cm^-1 that were assigned to non-radiative higher order modes, with a wide range of angular and radial momenta.³
• AFM-IR also exposes large photothermal expansions in nanostructured hBN bound to the height-to-width aspect ratio, an effect associated with the large anisotropy of the thermal expansion coefficients of hBN as well as the applied nanostructuring.⁷
• There have been several studies on functionalized graphene materials. In one study, three different functional groups were covalently attached to the graphene basal plane in a dense and well-defined pattern. The AFM-IR results, displayed in Figure 2, make clear the distinctions between patterned regions and the underlying graphene layer, as well as the differences among the various functionalized patterns.^8-11
• AFM-IR has also been used in the study of the hydrogen bonds that form between MXene and graphene oxide nanosheets, which demonstrably increased nanocomposite stability by reducing the overall energy of the system.¹²

Figure 2. AFM-IR spectra and images of graphene with covalent chemical patterning. (a) Spectra on (purple) and off (green) pattern for sample 1. (b) and (c) ratio IR image of 970:1100 cm^-1 and 1730:1100 cm^-1 of sample 1. (d) AFM-IR spectra on two different patterns for sample 2. (e) and (f) ratio IR image of 970:1100 cm^-1 and 1730:1100 cm^-1 of sample 2. Scan size: 20x20 μm. Sample courtesy Steven De Feyter (KU Leuven). Image Credit: Bruker Nano Surfaces and Metrology

Case Studies

AFM-IR Study of Phonon Polaritons in hBN

SPPs and phonon polaritons (PhPs) in 2D materials demonstrate excellent spatial confinement, paving the way for the development of enhanced light–matter interaction, super lenses, subwavelength metamaterials, and other novel photonic devices. Lower PhP loss is observed in hBN compared to graphene’s SPPs, supporting longer polariton lifetimes and propagation lengths, particularly in isotopically enriched or suspended hBN.

Tapping AFM-IR measurements for a thin hBN flake on a Si/SiO₂ substrate over the upper Reststrahlen band (1360–1600 cm^-1) clearly demonstrates a correlation between the phonon polaritons and the wavelengths being excited. Results are consistent with theory and prior s-SNOM observations, thereby confirming the suitability of AFM-IR for quantitative phonon-polariton studies.¹³

The topographic and Tapping IR data of a flat 75 nm thick hBN flake on Si/SiO₂ are displayed in Figure 3. These images were measured on a Dimension IconIR^® system, and a series of Tapping IR images were acquired from 1400 to 1580 cm^-1 in 20 cm^-1 steps, with three representative images shown at 1480, 1520, and 1560 cm^-1. The fringe patterns, which are a result of the interference of surface polariton waves, were observed clearly across all three IR images.

Figure 3. Topography and IR absorption images of a 75 nm thick hBN flake on Si/SiO₂. (a) Topography and (b-d) IR images at 1480, 1520, and 1560 cm^-1, obtained with Tapping AFM-IR. Scan size: 10x10 μm. Image Credit: Bruker Nano Surfaces and Metrology

By applying the theoretical models described by Equation 1 in Dai et al. and the parameters for h¹⁰BN in Table S4 of Giles et al., it is possible to calculate the predicted momentum of hBN polaritons as a function of excitation wavelengths in the upper Reststrahlen band. The dashed curve refers to the theoretical results, while the general trend shows the experimental data.^13,14

Figure 4. (a) Comparison of measured momentum of the polariton waves (points) with theoretical values (dashed curve). The gray circles and orange squares are results measured on IconIR and nanoIR3 instruments, respectively. (b) Spatio-spectral imaging of phonon polaritons in hBN. Image Credit: Bruker Nano Surfaces and Metrology

An array of 210 spectra was acquired using Tapping AFM-IR across the hBN/SiO₂ boundary and into the hBN flake, at a spacing of 12.5 nm and with the array perpendicular to the hBN edge.

The IR signal intensity of the normalized spectra is plotted as a function of distance from the hBN/SiO₂ boundary in Figure 4b. The resulting spatio-spectral image demonstrates polaritonic features in the same pattern as seen in s-SNOM imaging, which suggests AFM-IR is an appropriate technique for the study of phonon polaritons in hBN.¹³

AFM-IR Study of Stacking Order in Multilayer Graphene 

Multilayer graphene can exhibit different stacking orders, such as Bernal (ABAB), rhombohedral (ABCA or ACBA), or intermediate (ABCB). Stacking order has a significant impact on the material’s electronic properties; therefore, it is crucial to identify stacking order to ensure effective device fabrication.

Standard techniques (e.g., angle-resolved photoemission, infrared spectroscopy, and Raman spectroscopy) struggle to identify stacking orders due to inadequate resolution or limited relative contrast, hindering characterization of stacking orders in multilayer graphene. Working with Prof. Andrea Young’s group (Department of Physics, UC Santa Barbara), photothermal AFM-IR was combined with scanning microwave impedance microscopy (sMIM) to address potential limitations.¹⁵

Tapping IR imaging showed strong contrast between domains in a trilayer and a tetralayer graphene sample (Figure 5). Domains where the stacking orders differed were identified via cross-validation of AFM-IR, sMIM, and coherent Raman spectroscopy. These domains exhibit distinct electrical properties, as validated by sMIM. Further differentiation was revealed by AFM-IR spectra obtained from three distinct domains, with specific IR spectral fingerprints including the characteristic resonance at 1576 cm^-1.

Figure 5. (a) Topography and (b) Tapping AFM-IR images at 1576 cm^-1 of a multilayer graphene sample on SiO₂. (c) Tapping IR spectra collected at three locations with markers of the same colors in the IR image. Scan size: 30x30 μm. Sample courtesy: Andrea Young, UCSB. Image Credit: Bruker Nano Surfaces and Metrology

What’s more, high-resolution AFM-IR imaging displays narrow domain walls separating various regions of multilayer graphene. The domain walls are related to the shear strain that occurs at transitions between domains across the graphene flake.

The topographic and IR absorption data are shown in the images of Figure 6 at 1576, 1260, and 1128 cm^-1, with a number of stacking orders identified. Furthermore, a clear domain wall separating the two rhombohedral (ABCA and ACBA) phases is visible in the IR images at 1260 and 1128 cm-1, and a comprehensive analysis determined the domain wall width at the 10 nm scale.

Figure 6. (a) Topography and (b-d) IR images at 1576, 1260, and 1128 cm^-1, with (c) and (d) showing domain walls between ACBA and ABCA. Scan size: 3.5x3.5 μm. Sample courtesy: Andrea Young, UCSB. Image Credit: Bruker Nano Surfaces and Metrology

AFM-IR is also ideal for imaging the graphene layers found in van der Waals heterostructures. The images in Figure 7 present topographic and IR data from trilayer flakes featuring both rhombohedral (ABC) and Bernal (ABA) stacked regions, beneath 5−10 nm layers of hBN.

Several bright dots seen in the topographic image are the result of air bubbles trapped between the graphene trilayer and the outer hBN layer. Aside from that, the graphene flakes demonstrated a smooth profile.

Greater detail is shown in the IR images: at 1576 cm^-1, the signal is boosted for the ABC trilayer compared with the ABA trilayer, akin to the contrast for the uncapsulated graphene trilayer in Figure 5. Inverse contrast was detected in the IR image at 1128 cm^-1 between the ABC and ABA trilayers.

Figure 7. Subsurface imaging of trilayer graphene encapsulated by hBN. (a) Topography and (b,c) IR images at 1576 and 1128 cm^-1. Scan size: 20x20 μm. Sample courtesy: Andrea Young, UCSB. Image Credit: Bruker Nano Surfaces and Metrology

These results verify that AFM-IR, paired with correlative techniques (such as sMIM in this study), is an appropriate method for studying graphene stacking orders, identifying nanoscale walls between different domains, and imaging buried graphene layers within van der Waals heterostructures.

AFM-IR Correlates Structure and Chemistry in 2D Materials at the Nanoscale

It is proven that photothermal AFM-IR can now be considered a powerful technique for the characterization of 2D materials, including those on plasmon polaritons in graphene and phonon polaritons in hBN, as presented in this article.

AFM-IR delivers results that complement s-SNOM, while also adding a series of unique benefits. Unlike s-SNOM, AFM-IR can readily produce data on non-radiative “dark” states, heat transfer mechanisms, and the chemistry of functionalized graphene.

With its exceptional detection sensitivity, spatial resolution, measurement speed, and user-friendliness, AFM-IR can meet the increased demand for a more comprehensive understanding of 2D materials while leading the search for new materials.

References and Further Reading

1. Pham, P.V., et al. (2022). 2D Heterostructures for Ubiquitous Electronics and Optoelectronics: Principles, Opportunities, and Challenges. Chemical Reviews, 122(6), pp.6514–6613. DOI: 10.1021/acs.chemrev.1c00735. https://pubs.acs.org/doi/10.1021/acs.chemrev.1c00735.
2. Menges, F., et al. (2021). Substrate-enhanced photothermal nano-imaging of surface polaritons in monolayer graphene. APL Photonics, (online) 6(4). DOI: 10.1063/5.0044738. https://pubs.aip.org/aip/app/article/6/4/041301/123665/Substrate-enhanced-photothermal-nano-imaging-of.
3. Brown, L.V., et al. (2018). Nanoscale Mapping and Spectroscopy of Nonradiative Hyperbolic Modes in Hexagonal Boron Nitride Nanostructures. Nano Letters, 18(3), pp.1628–1636. DOI: 10.1021/acs.nanolett.7b04476. https://pubs.acs.org/doi/10.1021/acs.nanolett.7b04476.
4. Ciano, C. et al. (2018) 'Observation of phonon-polaritons in thin flakes of hexagonal boron nitride on gold,' Applied Physics Letters, 112(15). DOI: 10.1063/1.5024518. https://pubs.aip.org/aip/apl/article-abstract/112/15/153101/35592/Observation-of-phonon-polaritons-in-thin-flakes-of?redirectedFrom=fulltext.
5. Schwartz, J.J., et al. (2023). Mid-Infrared, Near-Infrared, and Visible Nanospectroscopy of Hydrogen-Intercalated MoO₃. The Journal of Physical Chemistry C, 127(34), pp.17002–17013. DOI: 10.1021/acs.jpcc.3c05114. https://pubs.acs.org/doi/10.1021/acs.jpcc.3c05114.
6. Khanikaev, A.B., et al. (2016). Experimental demonstration of the microscopic origin of circular dichroism in two-dimensional metamaterials. Nature Communications, 7(1). DOI: 10.1038/ncomms12045. https://www.nature.com/articles/ncomms12045.
7. López, J.J., et al. (2018). Large Photothermal Effect in Sub-40 nm h-BN Nanostructures Patterned Via High-Resolution Ion Beam. Small, 14(22). DOI: 10.1002/smll.201800072. https://onlinelibrary.wiley.com/doi/10.1002/smll.201800072.
8. Cian Bartlam, et al. (2018). Nanoscale infrared identification and mapping of chemical functional groups on graphene. Carbon, 139, pp.317–324. DOI: 10.1016/j.carbon.2018.06.061. https://www.sciencedirect.com/science/article/pii/S0008622318306286.
9. Liu, Z., et al. (2018). Direct observation of oxygen configuration on individual graphene oxide sheets. Carbon, 127, pp.141–148. DOI: 10.1016/j.carbon.2017.10.100. https://www.sciencedirect.com/science/article/abs/pii/S000862231731093X.
10. Rodríguez González, M.C., et al. (2021). Multicomponent Covalent Chemical Patterning of Graphene. ACS Nano, 15(6), pp.10618–10627. DOI: 10.1021/acsnano.1c03373. https://pubs.acs.org/doi/10.1021/acsnano.1c03373.
11. Kumagai, R., et al. (2025). Nanoscale chemical characterization of functionalized graphene by heterodyne AFM-IR and chemical force microscopy. Nanoscale, 17(29), pp.17016–17023. DOI: 10.1039/d5nr01862e. https://pubs.rsc.org/en/content/articlelanding/2025/nr/d5nr01862e.
12. Yang, J., et al. (2024). Water-induced strong isotropic MXene-bridged graphene sheets for electrochemical energy storage. Science, 383(6684), pp.771–777. DOI: 10.1126/science.adj3549. https://www.science.org/doi/10.1126/science.adj3549.
13. Dai, S., et al. (2014). Tunable Phonon Polaritons in Atomically Thin van der Waals Crystals of Boron Nitride. Science, 343(6175), pp.1125–1129. DOI: 10.1126/science.1246833. https://www.science.org/doi/10.1126/science.1246833.
14. Giles, A.J., et al. (2018). Ultralow-loss polaritons in isotopically pure boron nitride. Nature Materials, 17(2), pp.134–139. DOI: 10.1038/nmat5047. https://www.nature.com/articles/nmat5047.
15. Holleis, L., et al. (2025). Nanoscale Infrared and Microwave Imaging of Stacking Faults in Multilayer Graphene. Nano Letters, 25(33), pp.12487–12494. DOI: 10.1021/acs.nanolett.5c02301. https://pubs.acs.org/doi/10.1021/acs.nanolett.5c02301.

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

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