Atomic-layer Slicing: Methods & Applications

• Atomic-layer slicing is a technique for manipulating and analyzing materials one atomic plane at a time, enabling depth-resolved studies and precise 2D assembly.
• It employs resonant x-ray reflectivity, CVD growth, and mechanical peeling to achieve sub-angstrom control of layered structures and interfaces.
• This methodology supports device engineering, heterostructure synthesis, and interfacial spectroscopy, with measurable metrics like layer thickness and Hall mobility.

Atomic-layer slicing refers to a class of methodologies—both experimental and theoretical—designed to enable depth-resolved characterization, manipulation, or assembly of materials at the level of individual atomic planes. The term encompasses both quantitative spectroscopic analysis of internal structure in layered compounds and device-relevant approaches for the mechanical removal or transfer of atomically thin layers. Key developments include resonant x-ray reflectivity with atomically resolved modeling for buried interfaces, and scalable chemical vapor deposition (CVD) plus mechanical transfer for atomically thin two-dimensional (2D) materials. These strategies permit unprecedented control and measurement at Ångström scales, impacting heterostructure synthesis, interfacial spectroscopy, and 2D device engineering (Zwiebler et al., 2015, Wang et al., 2018).

1. Theoretical and Computational Principles of Atomic-Layer Slicing

Atomic-layer slicing in the context of resonant soft x-ray reflectivity treats layered materials as stratified media, with optical properties (complex refractive index) varying discretely on the scale of atomic planes. This approach departs from conventional slab-based models that presume homogeneous optical constants throughout each material, a simplification that fails at resonance when dramatic spatial variations exist even within a single composition.

The framework models the refractive index as $$n(z, \omega) = 1 - \delta(z, \omega) + i\beta(z, \omega)$$, where $\delta$ and $\beta$ can vary from slice to slice. The reflectivity of a stack of $N$ atomic slices is computed recursively using the Parratt formalism:

• $k_0 = 2\pi/\lambda$ is the incident wavevector,
• $k_{z,j} = k_0\sqrt{n_j^2 - \cos^2\theta}$ inside slice $j$,
• $$r_{j, j+1} = (k_{z, j} - k_{z, j+1})/(k_{z, j} + k_{z, j+1})$$ is the Fresnel reflection coefficient.

The total reflectivity $R_0$ is computed by iterating the recursion from substrate ($j=N$) to surface ($\delta$0), accounting for the phase accumulation and multiple reflections at each atomic interface.

For thin slices ($\delta$1), the kinematic scattering approximation simplifies the interface term, revealing sensitivity to differences in form factor ($\delta$2) and areal density ($\delta$3) between adjacent atomic planes. Crucially, this atomic-slice model captures sub-unit-cell modulations that become prominent when the x-ray energy is tuned to an absorption edge (Zwiebler et al., 2015).

2. Experimental Implementation: Resonant Soft X-ray Reflectivity and CVD Transfer

2.1 Resonant Soft X-ray Reflectivity

Atomic-layer slicing analysis is applied to resonant soft x-ray reflectivity for resolving Å-level variations in both structure and electronic character across interfaces. Key steps include:

• Structural parameterization: Layered films are divided into atomic planes of established thickness from crystallography. Substrate, films, surface segregations, and contamination layers may each have distinct parameterizations.
• Optical constants on and off resonance: For Mn-based perovskites, resonant $\delta$4 for the Mn $\delta$5 edge are derived from TEY-XAS, followed by Kramers–Kronig transforms to obtain slice-specific $\delta$6.
• Interface mixing: Realistic intermixing is addressed by error-function smoothing, interpolating refractive index profiles between adjacent atomic slices, parameterized by roughness $\delta$7.
• Global fit: All reflectivity data (various $\delta$8, polarizations, energies) are fit using nonlinear least squares, having free parameters including interface positions, roughness, adsorbate densities, and local $\delta$9, $\beta$0. Implementation is feasible in frameworks such as ReMagX (Zwiebler et al., 2015).

2.2 CVD Growth and Delamination of 2D Materials

Atomic-layer slicing is also applied in the context of monolayer transfer via CVD growth and clean peeling. For hexagonal boron nitride (h-BN), sequential step growth (SSG) on platinum allows controlled nucleation and domain expansion to monolayer films ($\beta$1 mm domains), which can be mechanically delaminated by a polyvinyl acetate (PVA) carrier:

• Growth variables: Seeding at 1200 °C, $\beta$2 mbar for 3–5 min, followed by homogenization and lateral growth at lower borazine pressures.
• Transfer: Ambient O$\beta$3 intercalation reduces h-BN/Pt adhesion; PVA/h-BN stack can then be peeled, stamped onto a target substrate at 125–130 °C, and released in warm DI water. The process supports sequential pick-up for heterostructure assembly and repeatable substrate reuse (Wang et al., 2018).

3. Spectroscopic and Structural Extraction at Atomic Resolution

To extract quantitative spectroscopic depth profiles, layer-specific dielectric functions are refined using Kramers–Kronig constrained variational fitting. Changes to $\beta$4 in target layers are modeled as weighted sums of triangular basis functions, constrained via the Kramers-Kronig relationships. This allows retrieval of $\beta$5 for selected slices and determination of site-resolved resonant scattering factors.

Application to LaSrMnO$\beta$6/NdGaO$\beta$7 heterostructures demonstrates direct sensitivity to:

• Interface and surface terminations: The fit distinguishes MnO$\beta$8-terminated surfaces ($\beta$9 choice) with minimized $N$0.
• Depth-resolved electronic changes: The topmost MnO$N$1 slice exhibits altered Mn $N$2 branching ratios, indicative of partial Mn$N$3 $N$4 Mn$N$5 reduction or ligand symmetry changes confined to the surface unit cell (Zwiebler et al., 2015).

4. Quantitative Performance Metrics and Device Outcomes

Atomic-layer slicing by CVD-transfer produces monolayer devices with reproducible, high-quality interfaces and precise control:

• h-BN monolayer thickness by AFM is $N$60.4 nm; monocrystalline domains $N$7 mm.
• Raman E$N$8 mode FWHM is $N$913 cm$k_0 = 2\pi/\lambda$0 for monolayers; peak area scales with layer number.
• Devices assembled with all-CVD graphene/h-BN heterostructures show Hall mobility $k_0 = 2\pi/\lambda$1 cm$k_0 = 2\pi/\lambda$2/V s at RT and residual carrier density $k_0 = 2\pi/\lambda$3 cm$k_0 = 2\pi/\lambda$4.
• Substrate reuse for multiple SSG/transfer cycles without degradation of growth or transfer quality (Wang et al., 2018).

Metric	Typical Value	Source
h-BN monolayer thickness	0.4 nm	AFM (Wang et al., 2018)
Monolayer domain size	$k_0 = 2\pi/\lambda$50.5 mm lateral	SEM (Wang et al., 2018)
Hall mobility (Gr/h-BN FETs)	$k_0 = 2\pi/\lambda$6 cm$k_0 = 2\pi/\lambda$7/V s	(Wang et al., 2018)

5. Mechanisms and Physical Underpinnings

The effectiveness of atomic-layer slicing analysis stems from both intrinsic materials physics and careful control over experimental variables:

• In reflectivity, large variations in local optical constants near absorption edges produce pronounced slice-resolved contrast.
• The error-function parameterization for roughness enables discrimination between atomically sharp and smeared (slab-like) interfaces, with the slicing model reverting to a homogeneous slab as $k_0 = 2\pi/\lambda$8 becomes much larger than the interplanar spacing (Zwiebler et al., 2015).
• In mechanical peeling, weak physisorbed interaction (Pt–h-BN) and further O$k_0 = 2\pi/\lambda$9-induced interfacial decoupling enable the adhesion energy hierarchy required for clean layer removal (PVA/h-BN $k_{z,j} = k_0\sqrt{n_j^2 - \cos^2\theta}$0 Pt/h-BN). Substrate grain growth to (111) texture and suppression of multilayer formation are critical for monolayer selectivity (Wang et al., 2018).

6. Applicability, Strengths, and Limitations

Atomic-layer slicing as an analysis and fabrication paradigm offers atomic-scale sensitivity, non-destructive (in reflectivity) or residue-free (in CVD transfer) layer discrimination, and direct access to buried interfaces or heterostructure assemblies. Applicability spans oxide heterostructures, magnetic multilayers, and 2D material stacks, particularly when orbital, valence, or density changes occur on the scale of atomic planes.

Limitations include the need for high material quality (epitaxy, uniform thickness), reliable optical constants from XAS or data tables, and computational complexity in high-slice-number fits leading to overfitting risks. For interface disorder far exceeding interplanar spacing, the model's utility diminishes (Zwiebler et al., 2015). In mechanical peeling, performance is limited by substrate quality, adhesion control, and precision in persistence of single-layer domains (Wang et al., 2018).

In summary, atomic-layer slicing provides an advanced methodological foundation for both depth-resolved measurement and controlled assembly of allotropes, oxides, and 2D materials, establishing a basis for future explorations of interfacial phenomena, ultrathin device engineering, and quantum materials design (Zwiebler et al., 2015, Wang et al., 2018).