Inverted-Mode STM: New Regimes

• Inverted-Mode STM is a set of techniques that invert conventional STM roles to achieve contrast reversal and explore new imaging frontiers.
• It leverages high-voltage field-emission and constant di/dz feedback to boost lithographic throughput and improve atomic-scale image accuracy.
• Molecule-mediated mechanosynthesis and rigorous quantitative models extend its application in precise nanofabrication and surface characterization.

Inverted-Mode Scanning Tunneling Microscopy (IM-STM) encompasses a collection of STM methodologies that invert conventional STM roles, imaging modalities, or control variables, enabling new regimes of spatial, chemical, or lithographic control unattainable through traditional approaches. IM-STM implementations include contrast inversion imaging on transition-metal surfaces, high-voltage field-emission lithography, feedback control on tunneling current sensitivity, and nanoscale mechanosynthesis via molecule-mediated probe-sample interactions. These modes are characterized by inversion of imaging contrast, probe/sample functional roles, or feedback/operation principles relative to canonical STM practice. The following sections delineate the principal variants of IM-STM, their theoretical frameworks, operational metrics, and applications in atomic-scale fabrication and characterization.

1. Contrast Inversion STM: Spin-Polarized and Orbital-Dependent Imaging

STM contrast inversion, commonly termed IM-STM in surface physics literature, refers to regimes in which the imaging contrast of surface atoms and hollow sites reverses with respect to conventional topography, causing atoms to appear as depressions rather than protrusions. Quantitative modeling is provided by three-dimensional WKB (3D-WKB) atom-superposition theories incorporating explicit orbital and spin-polarization dependence (Mándi et al., 2013):

• The tunneling current $I_{\rm TOT}(\mathbf{R}_{\rm TIP}, V)$ is decomposed into topographic ($I_{\rm TOPO}$) and spin-polarized ($I_{\rm MAGN}$) components, each integrated over a bias window.
• The orbital-resolved transmission $T_{\beta\gamma}(E, V, \mathbf{d}_a)$ encodes both WKB barrier decay and the angular symmetry of the involved tip and sample orbitals:

$$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$

• Corrugation inversion is linked to the dominance of off-axis ($m\ne0$) orbital channels (e.g., $d_{xz}, d_{yz}$) at low tip-sample distances or at specific bias voltages. As the tip is retracted, normal ($m=0$) channels regain dominance, reverting the contrast.
• Simulations for Fe(110) reveal that inversion contours $\Delta I(z, V) = 0$ in the $(z, V)$ plane are acutely sensitive to tip orbital composition and magnetization orientation. For Fe apex tips with parallel, perpendicular, or antiparallel alignment, inversion occurs at $I_{\rm TOPO}$0 Å, $I_{\rm TOPO}$1 Å, and $I_{\rm TOPO}$2 Å, respectively, at $I_{\rm TOPO}$3 V.

This regime is operationally accessed in constant-current spin-polarized STM (SP-STM) and requires recognition that apparent topography may invert due to orbital symmetry and spin-dependent selection rules, not simply geometric height variations (Mándi et al., 2013).

2. High-Voltage Field-Emission IM-STM: Accelerated Lithography via Fowler–Nordheim Injection

Inverted-mode STM, as high-voltage field-emission lithography (HVFE), exploits field emission at elevated tip-sample voltage ($I_{\rm TOPO}$4 V, e.g., up to 110 V) and large tip-sample gaps ($I_{\rm TOPO}$5 nm) to enlarge the effective patterning spot and increase patterning throughput by over two orders of magnitude relative to conventional tunneling mode (Rudolph et al., 2014):

• Electrons traverse the vacuum barrier in the Fowler–Nordheim regime, with the emission profile laterally broadened due to the large gap.
• The local field $I_{\rm TOPO}$6 creates a Gaussian-like emission spot (radius $I_{\rm TOPO}$7 nm).
• H-desorption cross-section saturates, and the patterning rate scales as $I_{\rm TOPO}$8, with $I_{\rm TOPO}$9 attaining its maximum for single-electron desorption.
• For device-scale features ($I_{\rm MAGN}$0m), electrical performance (contact resistance, carrier density, mobility) is unaffected compared to slow low-bias STM lithography; for nanoscale features, ultimate resolution is set by donor incorporation chemistry—not electron emission spot size—with a minimum width $I_{\rm MAGN}$110 nm.

A key attribute is that donor incorporation probability $I_{\rm MAGN}$2 imposes a hard threshold, leading to edge sharpness finer than the lithographic spot, governed by requirements such as three-adjacent-dimer desorption events for phosphorus donor placement.

Parameter	HVFE (IM-STM)	Conventional STM
Tip-sample voltage	$I_{\rm MAGN}$3 V (up to 110 V)	$I_{\rm MAGN}$4–$I_{\rm MAGN}$5 V
Tip-sample distance	$I_{\rm MAGN}$6 nm	$I_{\rm MAGN}$7 nm
Spot diameter	$I_{\rm MAGN}$8 nm	$I_{\rm MAGN}$9 nm
Areal patterning rate	$T_{\beta\gamma}(E, V, \mathbf{d}_a)$0 nm$T_{\beta\gamma}(E, V, \mathbf{d}_a)$1/s	$T_{\beta\gamma}(E, V, \mathbf{d}_a)$2 nm$T_{\beta\gamma}(E, V, \mathbf{d}_a)$3/s
Minimum feature size	$T_{\beta\gamma}(E, V, \mathbf{d}_a)$4 nm	$T_{\beta\gamma}(E, V, \mathbf{d}_a)$5 nm

Operation outside this process window (reduced bias, insufficient gap) degrades throughput and edge definition, while excessively large spots jeopardize feature definition at the nanometer scale (Rudolph et al., 2014).

3. Constant di/dz Feedback: Sensitivity-Optimized IM-STM for Imaging and Lithography

An alternative IM-STM implementation is constant $T_{\beta\gamma}(E, V, \mathbf{d}_a)$6 control, where feedback maintains not the current itself, but its distance sensitivity (Mishra et al., 2024). This mode enhances geometric contrast, suppresses current noise, and mitigates artifacts due to local barrier-height variations:

• The tip z-position is modulated with a small high-frequency (2 kHz) dither, and the resulting modulation in tunneling current is detected via a lock-in amplifier (LIA).
• The magnitude $T_{\beta\gamma}(E, V, \mathbf{d}_a)$7 provides a real-time measure of tunneling sensitivity.
• Feedback is closed on $T_{\beta\gamma}(E, V, \mathbf{d}_a)$8 rather than $T_{\beta\gamma}(E, V, \mathbf{d}_a)$9 using a PI controller, resulting in images sensitive to atomic-scale surface geometry rather than electronic structure variations alone.
• Hardware additions include summing modulation signals on the piezo, a chain of notch filters, and a LIA. Calibration sets modulation amplitude $$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$0 nm, scan speed 80–100 nm/s, and typical setpoints of $$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$1–$$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$2 nA/nm.

Empirically, $$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$3-mode STM on Si(100)-2×1:H resolves individual dimers with $$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$4–$$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$5 contrast enhancement versus conventional current (I) mode, reduces tip wear, and yields sharper patterns in hydrogen-depassivation lithography (Mishra et al., 2024).

4. Molecular-Probe and Mechanosynthetic Inverted-Mode STM

Recent advances have redefined IM-STM as a platform for molecule-mediated probe characterization and mechanosynthesis, inverting the conventional paradigm where the tip acts as the sole functional entity (Barrera et al., 30 Dec 2025, Cowie et al., 26 May 2026):

• Surface-bound, tall, rigid molecules (e.g., EAOGe-C$$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$6I) act as localized STM probes, forming reflected-probe images (RPIs) of an atomically clean and crystalline probe apex as the tip scans beneath the molecule.
• Mechanical approach/retraction of the probe in controlled sub-ångström increments can induce chemical reactions—H abstraction, C$$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$7 donation—at zero bias.
• Imaging and reactivity characteristics are uniquely tied to the molecular headgroup’s chemical termination, enabling in situ, single-image verification of probe structure and post-reaction state.

Mechanosynthetic operations demonstrated include:

• H abstraction from the Si(100) probe apex by activated ethynyl radicals tethered to the surface (Barrera et al., 30 Dec 2025).
• C$$T_{\beta\gamma}(E, V, \mathbf{d}_a) = \exp\bigl[-2\,\kappa(U,V)\,d_a\bigr]\; \chi_\beta^2(\vartheta_a,\varphi_a)\; \chi_\gamma^2(\pi-\vartheta_a,\pi+\varphi_a)$$8 donation to surface-prepatterned Si dangling bond (DB) pairs for the assembly of polyyne fragments, with sequential C–C coupling enabling sub-ångström site selectivity and >90% yields (Cowie et al., 26 May 2026).

Molecule design criteria require a rigid, upright geometry, reactive headgroups amenable to activation (UV or bias-induced), and minimal internal degrees of freedom. The operational protocol involves precise z-retraction, open-loop control (STM feedback off), and immediate imaging confirmation.

5. Quantitative Performance Metrics and Representative Applications

The application domains and performance regimes of IM-STM modalities are summarized as follows:

Lithographic Throughput and Feature Size (Rudolph et al., 2014)

• HVFE-STM: Areal rate %%%%49$I_{\rm TOPO}$050%%%% nm$m\ne0$1/s, micron-scale pads with $m\ne0$210 k$m\ne0$3, and $m\ne0$410 nm resolution for donor-based active regions.
• Edge sharpness set by donor chemistry, independent of emission spot.

Mechanosynthesis (Barrera et al., 30 Dec 2025, Cowie et al., 26 May 2026)

• H abstraction: $m\ne0$5% first H abstraction, $m\ne0$6% for second in double-DB patterning.
• C$m\ne0$7 and C–C bond formation: Yields range from 84–97% depending on pattern and chain length, with spatial error primarily limited by mechanical precision and pre-patterned DB accuracy.
• Real-time, high-contrast chemical state differentiation via RPI height and conductance.

Imaging and Feedback (Mishra et al., 2024)

• $m\ne0$8 imaging: Peak-to-valley geometric contrast 2–3$m\ne0$9 higher than constant current mode, with a typical signal bandwidth of 200 Hz, scan sizes 16–48 nm, and noise floor comparable to or lower than conventional STM.

6. Theoretical and Practical Limitations

Each IM-STM modality imposes critical operational and interpretive boundaries:

• 3D-WKB modeling neglects multi-atom tip effects, assumes zero temperature and elasticity, and often uses DFT-derived orbitals fixed for all biases (Mándi et al., 2013).
• HVFE lithography requires careful calibration of donor incorporation probability $d_{xz}, d_{yz}$0 and may introduce stray desorption islands below 10 nm scales (Rudolph et al., 2014).
• Mechanosynthetic approaches demand rigorous probe and substrate preparation—H passivation, probe apex annealing—to avoid off-target reactions and to ensure chemical state differentiability (Barrera et al., 30 Dec 2025, Cowie et al., 26 May 2026).
• Constant $d_{xz}, d_{yz}$1 imaging necessitates precise analog/digital hardware modifications and may be limited by LIA bandwidth and feedback loop stability (Mishra et al., 2024).

Experimental protocols recommend validation by independent post-reaction imaging, systematic variation (e.g., tip orbital composition or bias, for contrast inversion), and parallel topographic and spectroscopic measurements.

7. Extensions, Generalizations, and Future Directions

IM-STM architecture is extensible to broader regimes and application spaces:

• Mechanosynthetic protocols can, in principle, be generalized to other functional moieties (C$d_{xz}, d_{yz}$2, halogen radicals, group-V donors) using analogous upright molecules and activation schemes (Barrera et al., 30 Dec 2025).
• Probe materials (Ge, III–V semiconductors, metals) can be prepared to present clean, atomically defined apexes for RPI characterization, via in situ annealing or field evaporation.
• Two-dimensional materials can be engineered onto probe tips for quantum defect fabrication or optoelectronic studies.
• Multiplexed, DSP-based $d_{xz}, d_{yz}$3 control may enable parallel STM imaging or high-throughput atomic manipulation (Mishra et al., 2024).

In summary, inverted-mode STM constitutes a class of STM techniques that invert the structure, function, or control variable of the conventional STM paradigm, with demonstrated utility in imaging contrast engineering, accelerated patterning, atomically precise mechanosynthesis, and feedback optimization. Its implementations have produced quantum device-defining lithography, robust atomic manipulation protocols, and new avenues for probe and molecule co-characterization, as supported by empirical results and theoretical models (Mándi et al., 2013, Rudolph et al., 2014, Barrera et al., 30 Dec 2025, Mishra et al., 2024, Cowie et al., 26 May 2026).