resonant tunneling interband zener wkb approximation rtd

# Resonant Tunneling Diodes and Band-to-Band Tunneling Physics in Nanodevices

## 1. Introduction: Tunneling as a Carrier Transport Mechanism

Tunneling is the quantum mechanical phenomenon where a particle penetrates a potential barrier despite having insufficient classical energy to surmount it. In semiconductor devices, tunneling becomes the dominant transport mechanism when:
- Feature sizes drop below ~10 nm
- Applied fields exceed ~100 kV/cm (breakdown regime)
- Dopant concentrations are ultra-high (>10¹⁹ cm⁻³)

Two primary tunneling mechanisms dominate modern devices:
1. Resonant tunneling: Elastic tunneling through aligned energy levels
2. Band-to-band tunneling (Zener tunneling): Interband transitions in strong electric fields

## 2. WKB Approximation for Tunneling Probability

The Wentzel-Kramers-Brillouin (WKB) approximation gives the tunneling probability through a potential barrier:

$$T \approx \exp\left(-2\int_{x_1}^{x_2} \kappa(x) dx ight)$$

where $\kappa(x) = \sqrt{2m(V(x) - E)}/\hbar$ is the imaginary wave vector in the classically forbidden region ($x_1$ and $x_2$ are the classical turning points).

For a rectangular barrier of height $V_0$, thickness $W$, and particle energy $E$:

$$T \approx \exp\left(-\frac{2\sqrt{2m(V_0 - E)}}{\hbar} W ight)$$

The exponential dependence is crucial: reducing barrier width by just 1 nm can increase transmission by orders of magnitude.

### Numerical Example

For an electron tunneling through a 3 eV barrier of thickness 5 nm:
$$\kappa = \sqrt{\frac{2 imes 9.109 imes 10^{-31} imes 3 imes 1.6 imes 10^{-19}}{(1.055 imes 10^{-34})^2}} \approx 1.7 imes 10^{10} ext{ m}^{-1}$$

$$T \approx \exp(-2 imes 1.7 imes 10^{10} imes 5 imes 10^{-9}) \approx 10^{-15}$$

Essentially zero transmission. However, for $W = 2$ nm:
$$T \approx 10^{-6}$$

This exponential sensitivity drives the scaling of tunneling devices: every nanometer of barrier thickness matters.

## 3. Resonant Tunneling Diodes (RTD): Structure and Operation

A resonant tunneling diode (RTD) consists of:
- Emitter region: N-doped (high carrier concentration)
- First barrier: Thin (~2 nm) AlGaAs layer
- Quantum well: Narrow (~5-10 nm) GaAs layer with discrete energy levels
- Second barrier: Second thin AlGaAs layer
- Collector region: N-doped GaAs

### Energy Alignment and Resonance Condition

At zero bias, the Fermi level in emitter and collector are equal. The well contains quantized states with energies:
$$E_n = \frac{\hbar^2 \pi^2 n^2}{2m_e^* L^2}, \quad n = 1, 2, 3, \ldots$$

When forward bias $V$ is applied, the energy band in the collector drops by $eV$. Resonance occurs when an emitter state aligns with well state:
$$E_{ ext{emitter}} + eV_{ ext{resonant}} = E_{ ext{well}}$$

At resonance, electrons tunnel elastically through both barriers with high probability (limited only by coupling strength). Off-resonance, tunneling suppresses exponentially.

## 4. Current-Voltage Characteristics and Peak-to-Valley Ratio

The current through an RTD exhibits characteristic negative differential resistance (NDR):

$$I(V) = \begin{cases} I_{ ext{peak}} \propto T_{ ext{resonant}} & ext{at resonance} \\ I_{ ext{valley}} \propto T_{ ext{off}} \propto \exp(-\kappa W) & ext{off-resonance} \end{cases}$$

The peak-to-valley ratio (PVCR):
$$ ext{PVCR} = \frac{I_{ ext{peak}}}{I_{ ext{valley}}}$$

is a key figure of merit. High PVCR (>5:1, ideally >20:1) indicates sharp resonances and well-isolated well states.

### Practical Values

Modern lattice-matched AlGaAs/GaAs RTDs achieve:
- PVCR > 10:1 at room temperature
- Peak current density > 10⁴ A/cm²
- Characteristic voltage span ~0.5 V

## 5. Device Physics: Scattering and Dephasing

Perfect resonance requires elastic tunneling—the electron's energy is conserved. However, several inelastic mechanisms broaden resonances:

### Broadening Mechanisms

1. Phonon scattering (~10-20 meV at room temperature): Electron couples to lattice vibrations, spreading energy
2. Impurity scattering: Dopant atoms in barriers scatter electrons
3. Interface roughness scattering: Fluctuating barrier height causes energy uncertainty
4. Lifetime broadening: Finite lifetime $ au$ of well state → energy uncertainty $\Delta E \sim \hbar/ au$

The quality factor $Q = E_n / \Delta E$ quantifies sharpness:
$$Q = \frac{E_n}{\Gamma}$$

where $\Gamma$ is the total broadening (full-width half-maximum, FWHM). High-Q wells ($Q > 100$) require ultraclean interfaces and low-temperature operation.

## 6. Band-to-Band Tunneling (Zener Tunneling)

When an electric field is applied parallel to a junction (vertical in a p-n diode), the band alignment changes spatially. Near the junction, the conduction band in the n-region can drop below the valence band in the p-region, creating a tunneling window.

The Zener tunneling rate (probability per unit time) is:

$$\Gamma_Z \propto E^2 \exp\left(-\frac{\pi E_g^{3/2}}{3\hbar eE} ight)$$

where $E$ is the electric field. The exponential dependence on $E_g$ (wider-bandgap materials tunnel less) and inverse dependence on $E$ (stronger fields enable tunneling) are key.

## 7. Python Implementation: RTD I-V Characteristics and Tunneling Probability

import numpy as np
import matplotlib.pyplot as plt
from scipy.integrate import quad
from scipy.optimize import fsolve

def wkb_transmission(E, V0, W, m_star=0.067):
    """
    WKB tunneling probability through rectangular barrier.
    
    T = exp(-2κW) where κ = sqrt(2m(V0-E))/ℏ
    
    Parameters:
    -----------
    E : float (eV)
        Particle energy
    V0 : float (eV)
        Barrier height
    W : float (m)
        Barrier thickness
    m_star : float
        Effective mass (in units of m_e)
    
    Returns:
    --------
    T : float
        Transmission coefficient
    """
    
    if E >= V0:
        return 1.0  # Classical transmission
    
    hbar = 1.055e-34  # J·s
    m_e = 9.109e-31  # kg
    m = m_star * m_e
    
    # Convert eV to J
    dE = (V0 - E) * 1.6e-19
    
    kappa = np.sqrt(2 * m * dE) / hbar
    exponent = -2 * kappa * W
    
    T = np.exp(max(exponent, -100))  # Avoid underflow
    return T

def fermi_dirac(E, mu, T=300):
    """
    Fermi-Dirac distribution.
    
    f(E) = 1 / (1 + exp((E-μ)/(k_B T)))
    """
    k_B = 1.381e-23  # J/K
    k_B_eV = 8.617e-5  # eV/K
    
    if T == 0:
        return 1.0 if E < mu else 0.0
    
    return 1.0 / (1.0 + np.exp((E - mu) / (k_B_eV * T)))

def rtd_current_density(V, T_res, E_well, Gamma_res, Gamma_off, sigma_N, T_temp=300):
    """
    Compute RTD current density.
    
    Parameters:
    -----------
    V : float (V)
        Applied bias
    T_res : float
        Resonant tunneling probability (~0.01 to 0.1)
    E_well : float (eV)
        Energy of well state
    Gamma_res, Gamma_off : float (meV)
        Resonance and off-resonance broadening
    sigma_N : float (eV·cm²)
        Density of states parameter
    T_temp : float (K)
        Temperature
    
    Returns:
    --------
    J : float (A/cm²)
        Current density
    """
    
    q_e = 1.602e-19  # C
    k_B = 8.617e-5  # eV/K
    hbar = 1.055e-34  # J·s
    m_e = 9.109e-31  # kg
    
    # Emitter Fermi level (reference)
    mu_emitter = 0.0  # Set to zero
    mu_collector = -V  # Shifted by applied bias
    
    # Resonance condition
    E_aligned = E_well + V  # Energy of well state relative to collector
    
    # Detuning from resonance
    delta_E = E_aligned - mu_emitter  # Detuning (meV)
    
    # Lorentzian resonance: sharp peak at delta_E = 0
    if abs(delta_E) < 10*Gamma_res/1000:  # Within 10 linewidths
        T_tunnel_res = T_res / (1 + (delta_E / (Gamma_res/1000))**2)
    else:
        T_tunnel_res = 0
    
    # Off-resonance tunneling (exponentially suppressed)
    T_tunnel_off = T_res * np.exp(-abs(delta_E) / 0.05)  # Exponential tail
    
    # Total tunneling probability (approximate)
    T_total = T_tunnel_res + 0.1 * T_tunnel_off
    
    # Current (rough estimate)
    # J ∝ T * (number of carriers) * (frequency)
    v_thermal = np.sqrt(3 * k_B * T_temp / (m_e / m_e))  # ~10^5 m/s at 300K
    n_2d = 1e12  # cm^-2 (2D electron density)
    
    J = q_e * T_total * n_2d * v_thermal * 1e-4  # Convert to A/cm²
    
    return J

def zener_tunneling_rate(E_field, E_g=1.42, m_e_star=0.067, m_h_star=0.45):
    """
    Band-to-band Zener tunneling rate.
    
    Γ_Z ∝ E² exp(-π E_g^(3/2) / (3ℏ eE))
    
    Parameters:
    -----------
    E_field : float (V/m)
        Electric field
    E_g : float (eV)
        Bandgap
    m_e_star, m_h_star : float
        Effective masses (in units of m_e)
    
    Returns:
    --------
    rate : float
        Tunneling rate (arbitrary units, logarithmic scale)
    """
    
    if E_field < 1e3:  # Field must be significant
        return 0
    
    hbar = 1.055e-34
    m_e = 9.109e-31
    q_e = 1.602e-19
    
    E_g_J = E_g * 1.6e-19
    
    # Reduced effective mass
    mu_r = (m_e_star * m_h_star) / (m_e_star + m_h_star) * m_e
    
    # Pre-exponential factor (proportional to E²)
    prefactor = E_field**2
    
    # Exponent
    exponent_arg = np.pi * (E_g_J)**(3/2) / (3 * hbar * q_e * E_field)
    if exponent_arg > 100:
        exponent = np.inf
    else:
        exponent = exponent_arg
    
    rate = prefactor * np.exp(-min(exponent, 100))  # Cap at 100 to avoid overflow
    
    return rate

# Simulation: RTD I-V curve
V_array = np.linspace(-0.5, 1.0, 500)

# RTD parameters (typical AlGaAs/GaAs)
T_res_prob = 0.05  # Resonant transmission ~5%
E_well = 0.15  # Well state at 150 meV
Gamma_res = 5  # Resonance linewidth (5 meV)
Gamma_off = 15  # Off-resonance broadening (15 meV)

J_rtd = np.array([rtd_current_density(V, T_res_prob, E_well, Gamma_res, Gamma_off, 1.0) 
                   for V in V_array])

# Band-to-band tunneling
E_field_array = np.abs(V_array) / 5e-9 * 1e9  # V/m (assuming 5 nm junction width)
rate_zener = np.array([zener_tunneling_rate(E) for E in E_field_array])

# Plotting
fig, axes = plt.subplots(2, 2, figsize=(13, 10))

# Panel A: RTD I-V characteristic
axes[0, 0].plot(V_array, J_rtd*1e3, 'b-', linewidth=2.5)
# Mark peak and valley
J_max = np.max(J_rtd)
J_min = np.min(J_rtd[J_rtd > 0.1*np.max(J_rtd)])
V_peak = V_array[np.argmax(J_rtd)]
axes[0, 0].scatter([V_peak], [J_max*1e3], color='red', s=100, marker='o', label='Peak')
# Approximate valley
V_valley_idx = np.where((V_array > V_peak) & (J_rtd < J_min*1.5))[0]
if len(V_valley_idx) > 0:
    V_valley = V_array[V_valley_idx[0]]
    axes[0, 0].scatter([V_valley], [J_min*1e3], color='orange', s=100, marker='s', label='Valley')
axes[0, 0].set_xlabel('Bias Voltage (V)', fontsize=11)
axes[0, 0].set_ylabel('Current Density (mA/cm²)', fontsize=11)
axes[0, 0].set_title('RTD I-V Characteristic (NDR)', fontsize=12, fontweight='bold')
axes[0, 0].legend(fontsize=10)
axes[0, 0].grid(True, alpha=0.3)

# Panel B: WKB transmission vs energy
E_scan = np.linspace(0, 0.3, 200)
T_barrier1 = np.array([wkb_transmission(E, 0.3, 2e-9, 0.067) for E in E_scan])
axes[0, 1].semilogy(E_scan*1000, T_barrier1, 'b-', linewidth=2)
axes[0, 1].axvline(x=150, color='r', linestyle='--', alpha=0.7, label='Well level (150 meV)')
axes[0, 1].set_xlabel('Energy (meV)', fontsize=11)
axes[0, 1].set_ylabel('Transmission (log scale)', fontsize=11)
axes[0, 1].set_title('WKB Tunneling Probability (2nm barrier)', fontsize=12, fontweight='bold')
axes[0, 1].legend(fontsize=10)
axes[0, 1].grid(True, alpha=0.3, which='both')

# Panel C: Barrier thickness effect
W_array = np.linspace(1e-9, 5e-9, 100)
T_W = np.array([wkb_transmission(0.15, 0.3, W, 0.067) for W in W_array])
axes[1, 0].semilogy(W_array*1e9, T_W, 'g-', linewidth=2.5)
axes[1, 0].axhline(y=1e-3, color='k', linestyle=':', alpha=0.5, label='Practical limit (~10⁻³)')
axes[1, 0].set_xlabel('Barrier Thickness (nm)', fontsize=11)
axes[1, 0].set_ylabel('Transmission (log scale)', fontsize=11)
axes[1, 0].set_title('Exponential Sensitivity to Barrier Width', fontsize=12, fontweight='bold')
axes[1, 0].legend(fontsize=10)
axes[1, 0].grid(True, alpha=0.3, which='both')

# Panel D: Zener tunneling rate
axes[1, 1].semilogy(E_field_array/1e5, rate_zener + 1e-20, 'purple', linewidth=2.5)
axes[1, 1].set_xlabel('Electric Field (× 10⁵ V/m)', fontsize=11)
axes[1, 1].set_ylabel('Zener Tunneling Rate (log scale)', fontsize=11)
axes[1, 1].set_title('Band-to-Band Zener Tunneling', fontsize=12, fontweight='bold')
axes[1, 1].grid(True, alpha=0.3, which='both')

plt.tight_layout()
plt.savefig('rtd_tunneling_physics.png', dpi=150, bbox_inches='tight')
plt.show()

print("=== RTD and Tunneling Physics Analysis Complete ===")
print(f"Peak current density: {np.max(J_rtd)*1e3:.2e} mA/cm²")
print(f"Estimated PVCR: ~{np.max(J_rtd)/np.min(J_rtd[J_rtd > 0.1*np.max(J_rtd)]):.1f}")
print(f"WKB transmission @ 150 meV, 2 nm barrier: {wkb_transmission(0.15, 0.3, 2e-9, 0.067):.2e}")

## 8. Applications of RTDs: Oscillators and High-Speed Electronics

RTDs exhibit negative resistance (dI/dV < 0) in the resonant region—unusual in passive devices. This enables:

1. Self-sustained oscillations (RTD oscillator): No external oscillator needed; intrinsic negative resistance drives L-C circuit oscillation
2. Frequency tuning via bias voltage (0.5–10 THz in advanced devices)
3. Compact THz sources for sensing and imaging

Modern RTD oscillators achieve:
- Frequency: up to 5 THz
- Output power: milliwatts (room temperature)
- Tuning range: 20–30% of center frequency

## 9. Interband Tunneling in Tunnel FETs (TFETs)

Tunnel Field-Effect Transistors exploit band-to-band tunneling for steep subthreshold swing:

$$S = \frac{dV_{ ext{gs}}}{d(\log I_d)} = \frac{k_B T}{q_e} \ln(10) \left(1 + \frac{C_d}{C_{ ext{ox}}} ight)$$

Conventional MOSFETs are limited by thermionic emission to $S \geq 60$ mV/dec at room temperature. TFETs achieve $S < 40$ mV/dec by using tunneling (voltage-independent tunneling rate).

## 10. Tunneling-Assisted Leakage in Ultra-Thin Devices

As devices scale below 3 nm, band-to-band tunneling becomes significant leakage current:

$$I_{ ext{leak}} \propto W_n W_p E^2 \exp\left(-\frac{2\pi E_g^{3/2}}{3\hbar eE} ight)$$

where $W_n, W_p$ are depletion widths. Strategies to mitigate:
- High-κ dielectrics: Reduce peak field for same voltage
- Pocket doping: Create graded potential near junction
- Source/drain engineering: Wider bandgap materials (SiGe, SiC)

## 11. Scattering Time and Phase Coherence

In resonant tunneling, the dephasing rate determines linewidth:

$$\Gamma = \hbar / au_\phi$$

Main dephasing sources at room temperature:
- Electron-phonon scattering: ~10–20 meV (strongest)
- Electron-electron interactions: ~1–5 meV
- Interface roughness: ~0.5–2 meV

At cryogenic temperatures (<77 K), phonon scattering drops exponentially; PVCR improves dramatically.

## 12. Comparison: Resonant vs. Quasiparticle Tunneling

  • Resonant tunneling: Energy-conserving, sharp resonances, high sensitivity to alignment
  • Off-resonance quasiparticle tunneling: Continuous background, exponentially suppressed
  • Band-to-band tunneling: Field-enhanced, broadband, enables new device physics

## 13. Advanced Structures: Superlattices and Resonator Arrays

Superlattice tunneling diodes stack many RTD stages in series, creating miniature filtering. Coupled resonator arrays enable:
- Narrowband filtering (THz meta-devices)
- Harmonic generation
- Parametric amplification

## 14. Challenges and Limitations

  • Temperature sensitivity: Phonon scattering degrades PVCR rapidly above 100 K
  • Fabrication precision: Barrier thickness must be controlled to ~0.1 nm
  • Series resistance: Parasitic RC time constant limits frequency
  • Heating: High current density (>10 kA/cm²) causes self-heating, degrading performance

## 15. Future Directions: Quantum Optics and Quantum Computing

Emerging applications:
- Single-photon sources: Resonant tunneling as quantum dot excitation mechanism
- Parametric amplifiers: Nonlinear tunneling for quantum-limited amplification
- Quantum dot networks: Coupled tunneling between multiple quantum states
- Topological tunneling: Edge states in photonic crystals and quantum materials

Go deeper with CFSGPT

Get AI-powered deep-dives, save terms, and run advanced simulations — free account.

Create Free Account