How to Learn Scanning Tunneling Microscopy and Spectroscopy (STM/STS): From Quantum Tunneling to Atomic Images, Local Density of States and Ultrafast Single-Atom Experiments

## Wait, What? An STM “Height” Image Is Not Purely a Height Map STM is famous for images in which individual atoms appear as bright bumps. But the tip normally does not touch the surface. A quantum tunneling current crosses the vacuum gap, and that current depends both on tip–sample distance and on the electronic states available for tunneling. > **STM images a tunneling junction. Its contrast mixes geometry and electronic structure unless experiments deliberately separate them.** ## The One-Sentence Answer **Learn STM by tracing quantum tunneling → exponential current–distance dependence → feedback-controlled scanning → atomic-scale contrast, then add bias dependence, tip density of states, vibration/drift, band bending and spectroscopy before treating a bright atom, defect state or gap as a unique property of the sample.** # Beginner Layer — The Vacuum Gap Is a Quantum Barrier ## Stage 1: Tip and Sample Contain Electronic States Bring them close without touching. ## Stage 2: Quantum Wavefunctions Decay Into the Vacuum Barrier Electrons can tunnel through a classically forbidden gap. ## Stage 3: Tunneling Probability Falls Exponentially With Distance A simplified relation is **I ∝ V exp(−2κz)**. ## Stage 4: Tiny Vertical Motion Produces Large Current Change This is the source of STM’s extraordinary vertical sensitivity. # Bias Layer ## Stage 5: Apply a Voltage Between Tip and Sample Bias selects the energy window for tunneling. ## Stage 6: Bias Sign Selects Occupied or Unoccupied Sample States Under a Given Convention Always check the instrument convention. ## Stage 7: Changing Bias Can Change the Image Without Moving Atoms Electronic contrast is fundamental. # Tip Layer ## Stage 8: The Tip Is Ideally Atomically Sharp W and Pt–Ir are common. ## Stage 9: The Frontmost Tip States Dominate the Junction The entire wire does not need to be one atom wide. ## Stage 10: Tip Electronic Structure Is Part of Every Image A changed apex can radically change contrast. ## Stage 11: Double Tips Can Duplicate Features A beautiful repeating image can still be an artifact. # Constant-Current and Constant-Height Modes ## Stage 12: Constant-Current STM Uses Feedback to Adjust z The recorded z signal becomes the main image. ## Stage 13: The z Image Follows Approximately Constant Integrated Tunneling Probability It is not pure geometric topography. ## Stage 14: Feedback Bandwidth Limits Scan Speed Poor settings create overshoot and stripes. ## Stage 15: Constant-Height STM Holds z Nearly Fixed and Records Current It can be faster, but demands a very flat surface and increases crash risk. # Vibration, Piezo and Drift ## Stage 16: Atomic Resolution Requires Extreme Mechanical Stability Sub-ångström vibration is already significant. ## Stage 17: Piezoelectric Scanners Have Creep, Hysteresis and Cross-Coupling Raw coordinates are not perfectly linear. ## Stage 18: Known Atomic Lattices Can Provide Internal Calibration ## Stage 19: Thermal Drift Warps Long Scans Low-temperature systems improve stability but add cryogenic complexity. # Surface Preparation ## Stage 20: Atomic Resolution Usually Requires a Clean Surface UHV, cleavage, annealing or epitaxial growth may be used. ## Stage 21: Preparation Changes the Surface State Sputtering creates defects; annealing can reconstruct or segregate. # Atomic Contrast ## Stage 22: Surface Atoms Can Appear as Periodic Features But electronically inequivalent atoms may appear with different brightness. ## Stage 23: Bias Can Reverse or Reorganize Atomic Contrast The atoms did not necessarily move. ## Stage 24: Surface Reconstructions Can Differ From the Bulk STM made real-space surface reconstruction a central experimental object. # STS Layer — Local Spectroscopy ## Stage 25: Sweep Bias at One Position to Measure I(V) ## Stage 26: Differentiate to Obtain dI/dV Under suitable approximations, **dI/dV ∝ local density of states (LDOS)**. ## Stage 27: dI/dV Is Not the Sample DOS in Isolation Tip DOS, tunneling matrix elements, temperature and modulation broadening matter. ## Stage 28: Lock-In Detection Measures Differential Conductance Efficiently A small AC modulation improves sensitivity while broadening narrow features. # Semiconductor Layer ## Stage 29: STS Can Map Band Edges and Defect States ## Stage 30: The Tip Electric Field Can Bend Semiconductor Bands This is **tip-induced band bending**. ## Stage 31: Apparent Band Edge Can Depend on Junction Geometry and Doping The semiconductor canonical owns device band physics; STM owns this measurement artifact. # Superconductivity and QPI ## Stage 32: STS Can Resolve a Superconducting Gap Low temperature and high energy resolution are essential. ## Stage 33: Gap Shape Is Model Dependent Lifetime broadening, multiple bands and anisotropy affect fitting. ## Stage 34: Defects Can Generate Quasiparticle Interference Fourier transforms of dI/dV maps constrain scattering vectors and dispersion. ## Stage 35: QPI Is Not Direct Band Imaging Spin, orbital texture and impurity potential select allowed scattering. # Spin and Inelastic Tunneling ## Stage 36: Magnetic Tips Make Tunneling Spin Sensitive Spin-polarized STM can image magnetic structures at atomic scale. ## Stage 37: Magnetic Contrast Can Mix With Topographic/Electronic Contrast Field or tip-magnetization reversal is a strong test. ## Stage 38: Electrons Can Lose Energy to Local Vibrations or Spins Inelastic tunneling features reveal molecular and magnetic excitations. # Atomic Manipulation ## Stage 39: The Tip Can Intentionally Move Atoms Voltage pulses, forces or currents can act as nanoscale actuators. ## Stage 40: Observation and Intervention Become One Instrument The microscope is not inherently passive. # 2026 Frontier — ESR and Ultrafast STM ## Stage 41: Radiofrequency Driving Can Produce ESR-STM Single atomic or molecular spins can be driven while tunneling current reads out the response. ## Stage 42: 2026 Work Extends ESR-STM to Delocalized Electronic States This pushes STM from imaging toward local quantum-state control. ## Stage 43: Terahertz Fields Can Drive Tunneling on Sub-Picosecond Timescales THz-STM combines atomic spatial resolution with ultrafast temporal control. ## Stage 44: Ultrafast STM Adds a New Calibration Problem The junction field depends on waveform, tip enhancement and geometry. # Autonomous / ML Layer ## Stage 45: STM Is Slow and Tip Quality Is Unstable Automated systems can detect bad tips, drift and contaminated regions. ## Stage 46: Neural Denoising Can Invent Atomic-Looking Structure Atomic periodicity is particularly dangerous for hallucinated detail. ## Stage 47: Raw Current, Scan Direction, Bias and Setpoint Must Remain Available Processed images need an auditable path back to the measured junction signal. # Professional Layer ## Stage 48: Separate Four Objects 1. physical surface; 2. tip electronic structure; 3. tunneling junction; 4. recorded image/spectrum. ## Stage 49: Professional STM/STS Is a Tip–Barrier–LDOS Inverse Problem > **Which atom, defect, gap or quantum state remains identifiable after tip structure, bias, work function, feedback, drift, band bending, tunneling matrix elements and processing are all allowed to shape the same atomic-scale contrast?** # Evidence: What Makes an STM/STS Claim Strong? Stronger evidence combines repeat images after tip conditioning, forward/backward scans, bias and setpoint series, lattice calibration, drift correction, spectra at multiple positions, field/temperature perturbations and comparison with DFT/ARPES/TEM/transport. # Misconceptions Worth Hunting – STM physically drags a needle across atoms. – Bright regions are purely taller. – Constant-current z is pure geometry. – dI/dV equals sample DOS exactly. – Lock-in modulation improves noise with no resolution cost. – STS band edges are immune to tip-induced band bending. – Spin-polarized contrast is automatically magnetic. – AI denoising can safely reconstruct weak atomic lattices without raw-data tests. # Transfer Check Atomic contrast reverses when sample bias changes sign. Did the atoms move? **Not necessarily. Different states can dominate tunneling.** A semiconductor apparent gap shifts when the tip moves closer. Did the bulk gap change? **Not necessarily. Tip-induced band bending changed.** Two independent tips reproduce the same defect-state energy. Is the assignment stronger? **Yes.** # Model Limits STM requires a conductive tunneling path and is strongly surface sensitive. The image is a property of the whole junction, not the sample alone. Professional STM keeps **tip state + bias + current setpoint + environment + piezo calibration + drift + feedback + local electronic states + tunneling model + repeatability** visible together. # Teaching Guide Teach in this order: **quantum barrier → tunneling current → exponential gap → bias → tip → constant-current/height → vibration/piezo/drift → atomic electronic contrast → I(V) → dI/dV → band bending → superconducting gap → QPI → spin polarization → inelastic tunneling → atomic manipulation → ESR → THz ultrafast → autonomous scanning → validation.** # Connect This to the eduKate Learning Estate – https://edukatesengkang.com/2026/08/29/how-to-learn-quantum-measurement-superposition-tunnelling-state-reasoning/ – https://edukatesengkang.com/2026/08/28/how-to-learn-semiconductors-transistors-energy-bands-modern-electronics/ – https://edukatesengkang.com/2026/08/29/how-to-learn-superconductivity-quantum-materials/ – https://edukatesengkang.com/2026/08/30/how-to-learn-spintronics-magnetic-memory/ # The Quiet Ending The beginner asks, “Is that bump an atom?” The developing microscopist asks, “Which tunneling states made it bright?” The advanced learner asks, “Could the tip, bias or electric field create the same contrast?” And the professional asks: > **Which atomic-scale claim survives after the entire tip–barrier–sample junction is treated as the measuring instrument?**