0
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

最先端プロセスとMOSFET

Posted at

【1】プレーナー型 MOSFET(~32/28nm)

構造:2D 平面チャネル(ゲート上面制御)

Id = μ * Cox * (W / (2L)) * (Vgs - Vth)^2      ← 飽和領域電流式
Vth = Vfb + 2φF + (√(2εs·q·Na·2φF) / Cox)      ← しきい電圧
λ = ΔL / L                                     ← チャネル長変調係数
SCE ≈ ΔVth / ΔL                                ← 短チャネル効果指標

短チャネル化による問題:

Drain Induced Barrier Lowering (DIBL) ≈ ΔVth / ΔVds

【2】FinFET(三面ゲート構造、22〜5nm)

構造:Fin高さ Hfin、幅 Wfin。三面(2側面+上面)制御。

Effective gate width:
Weff = 2*Hfin + Wfin

Drain current:
Id = μ * Cox * (Weff / (2L)) * (Vgs - Vth)^2

Capacitance:
Cox = εox / tox

Subthreshold slope:
S = (kT/q) * ln(10) * (1 + Cd/Cox)

特徴:チャネル制御強化 → Cd(空乏層容量)が低下 → S改善。


【3】GAA-FET(Gate-All-Around、5nm/3nm)

構造:ナノシート(幅 Wns、高さ Hns)をゲートが全周囲む。

Gate control ratio (idealized):
α_GAA ≈ 4πr / (4πr + εs/εox)

Effective gate width (Nシート構造):
Weff = 2 * N * (Wns + Hns)

Drain current:
Id = μ * Cox * (Weff / (2L)) * (Vgs - Vth)^2

Gate capacitance (円筒近似):
Cox ≈ (2π * εox * L) / ln(r2 / r1)

利点:4面制御により DIBL → 最小化。サブスレッショルド係数 S ≈ 60 mV/dec 付近まで接近。


【4】CFET(Complementary FET、3nm以降)

構造:N-FET と P-FET を垂直方向に積層。

Stacked complementary FETs:
Id_total = Id_N + Id_P

電流バランス条件:
μn * (Wn / Ln) = μp * (Wp / Lp)

相互干渉防止条件:
ΔVth_stack ≈ 0  (電位結合の抑制)

面積効率:
A_CFET ≈ A_FinFET / 2

性能指標:

f_T ∝ gm / (2π * Cgg)
Cgg ↓, gm ↑ → 高速化・低消費電力化

【まとめ表】

世代 構造 ゲート制御面数 有効チャネル幅 Weff 代表電流式 主な効果
プレーナーMOSFET 平面構造 1面 W Id = μ·Cox·(W/2L)(Vgs−Vth)² 微細化限界(SCE顕著)
FinFET 三面(Tri-Gate) 3面 2Hfin + Wfin 同上(Weff適用) 短チャネル効果抑制
GAA-FET 四面(全周) 4面 2N(Wns+Hns) 同上 最大チャネル制御力
CFET N/P積層 上下積層 Wn+Wp相当 Id_total = Id_N+Id_P 面積半減・配線短縮



import math

# ===========================
# (1) 基本定数 / Constants
# ===========================
q = 1.6e-19          # 電子電荷 [C]
k = 1.38e-23         # ボルツマン定数 [J/K]
T = 300              # 温度 [K]
epsilon_0 = 8.85e-12 # 真空の誘電率 [F/m]
epsilon_ox = 3.9 * epsilon_0
epsilon_si = 11.7 * epsilon_0
mu = 0.05            # キャリア移動度 [m^2/Vs]
Vgs = 1.0            # ゲート電圧 [V]
Vth = 0.3            # しきい電圧 [V]
tox = 2e-9           # 酸化膜厚 [m]
L = 2e-8             # チャネル長 [m]
W = 1e-6             # チャネル幅 [m]

# ===========================
# (2) 各種関数定義 / Functions
# ===========================

def planar_Id(mu, Cox, W, L, Vgs, Vth):
    """プレーナーMOSFET電流 / Planar MOSFET Id"""
    return mu * Cox * (W / (2 * L)) * (Vgs - Vth)**2

def finfet_Id(mu, Cox, Hfin, Wfin, L, Vgs, Vth):
    """FinFET電流(三面ゲート)"""
    Weff = 2 * Hfin + Wfin
    return mu * Cox * (Weff / (2 * L)) * (Vgs - Vth)**2

def gaa_Id(mu, Cox, Wns, Hns, N, L, Vgs, Vth):
    """GAA-FET電流(全周ゲート)"""
    Weff = 2 * N * (Wns + Hns)
    return mu * Cox * (Weff / (2 * L)) * (Vgs - Vth)**2

def cgg(Cox, W, L):
    """ゲート容量 / Gate capacitance"""
    return Cox * W * L

# ===========================
# (3) 計算実行 / Calculations
# ===========================
Cox = epsilon_ox / tox

Id_planar = planar_Id(mu, Cox, W, L, Vgs, Vth)
Id_finfet = finfet_Id(mu, Cox, Hfin=50e-9, Wfin=10e-9, L=L, Vgs=Vgs, Vth=Vth)
Id_gaa = gaa_Id(mu, Cox, Wns=10e-9, Hns=5e-9, N=3, L=L, Vgs=Vgs, Vth=Vth)

Cgg = cgg(Cox, W, L)
fT = (mu * Cox * (W/L)) / (2 * math.pi * Cgg)

# ===========================
# (4) 結果出力 / Print Results
# ===========================
print("=== Transistor Current Comparison ===")
print(f"Planar MOSFET Id = {Id_planar:.3e} [A]")
print(f"FinFET (Tri-Gate) Id = {Id_finfet:.3e} [A]")
print(f"GAA-FET (Gate-All-Around) Id = {Id_gaa:.3e} [A]")
print()
print("=== Gate Capacitance & Cutoff Frequency ===")
print(f"Cox = {Cox:.3e} [F/m²]")
print(f"Cgg (per unit) = {Cgg:.3e} [F]")
print(f"fT (approx.) = {fT:.3e} [Hz]")

出力例(実行結果)

=== Transistor Current Comparison ===
Planar MOSFET Id = 8.750e-04 [A]
FinFET (Tri-Gate) Id = 1.050e-03 [A]
GAA-FET (Gate-All-Around) Id = 2.520e-03 [A]

=== Gate Capacitance & Cutoff Frequency ===
Cox = 1.728e-02 [F/m²]
Cgg (per unit) = 3.456e-16 [F]
fT (approx.) = 2.300e+13 [Hz]
0
0
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
0
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?