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?

最先端半導体プロセス構造と式モデル

Posted at

────────────────────────────
【1】FinFET(Tri-Gate構造)

世代: 22 nm → 16 nm → 7 nm
構造: 三面ゲート(Top + 2 Sidewall)

主式:
Id = (1/2) * μ * Cox * (Weff / L) * (Vgs - Vth)^2 * (1 + λ * Vds)
Weff = 2 * Hfin + Tfin
gm = 2 * Id / (Vgs - Vth)
ro = 1 / (λ * Id)

特性:

  • DIBL抑制、高 gm/Id
  • 動作電圧 0.6〜0.9 V

────────────────────────────
【2】GAA Nanosheet / Nanowire FET

世代: 5 nm → 3 nm
構造: ゲートがチャネル全周を包囲

主式:
Id = (1/2) * μ * Cox * (Weff / L) * (Vgs - Vth)^2 * (1 + λ * Vds)
Weff = n * (2 * Hsheet + Wsheet)
S ≈ 60 mV/dec (理想値)

特性:

  • 完全デプレッションチャネル
  • 動作電圧 0.4〜0.7 V

────────────────────────────
【3】CFET(Complementary FET)

世代: 約 2 nm 以降
構造: NFET と PFET を垂直積層

主式:
Weff_total = Weff_N + Weff_P
Id_total = Id_N + Id_P

特性:

  • 面積約50%削減
  • 対称電流経路により Vth 差最小化

────────────────────────────
【4】Forksheet FET(分離ゲート構造)

世代: 約 2 nm(imec報告)
構造: N/Pデバイス間に絶縁壁を配置

主式:
Id = (1/2) * μ * Cox_eff * (Weff / L) * (Vgs - Vth)^2
Cox_eff: ゲート遮蔽の影響を含む有効酸化膜容量

特性:

  • ゲート間干渉を低減
  • CFETより実装容易

────────────────────────────
【5】2-D FET(原子層トランジスタ)

材料: MoS2, WS2, WSe2, Graphene 等
厚さ: 1 nm 以下

主式:
1 / Ctotal = 1 / Cox + 1 / Cq
Id = μ * Ctotal * (W / L) * (Vgs - Vth) * Vds

特性:

  • 短チャネル効果が極小
  • 動作電圧 0.2〜0.5 V

────────────────────────────
【6】NCFET(Negative Capacitance FET)

原理: 強誘電体層により負容量を生成

主式:
Ceff = (Cox * Cf) / (Cox + Cf)
gm_eff ∝ Ceff
Id = (1/2) * μ * Ceff * (Weff / L) * (Vgs - Vth)^2

条件: Cf < 0 のとき |Ceff| > Cox
特性: S < 60 mV/dec、VDD ≈ 0.3 V

────────────────────────────
【7】TFET(Tunnel FET)

原理: バンド間トンネル電流を利用
主式:
Id ∝ exp(-B / Efield)

特性:

  • サブ60 mV/dec
  • 低漏れ・低電力
  • 高速動作には不向き

────────────────────────────
【8】CNTFET(Carbon-Nanotube FET)

構造: カーボンナノチューブをチャネルに使用

主式:
Id = (4 * q^2 / h) * M * (Vgs - Vth)
M: 伝導チャネル数(CNT本数)

特性:

  • μ ≈ 10^4 cm^2/Vs
  • 動作電圧 0.2〜0.4 V
  • 量子伝導型

────────────────────────────
【9】比較要約

FinFET: Weff = 2Hfin + Tfin
GAA: Weff = n(2H + W)
CFET: Id_total = IdN + IdP
Forksheet:Cox_eff を考慮
NCFET: Ceff = (Cox * Cf)/(Cox + Cf)
TFET: Id ∝ exp(-B/Efield)
CNTFET: Id = (4q²/h)M(Vgs−Vth)
2D-FET: 1/Ctotal = 1/Cox + 1/Cq

────────────────────────────
【10】統一表現式(Generalized FET Equation)

Id = (1/2) * μ * Ceff * (Weff / L) * (Vgs - Vth)^2 * (1 + λ * Vds)
gm = 2 * Id / (Vgs - Vth)
ro = 1 / (λ * Id)
Ceff = 1 / (1/Cox + 1/Cq + 1/Cf)
Weff = 幾何構造依存 (Fin, Sheet, Tube, Layer)



import numpy as np

# ==================================================
# 定数 / Constants
# ==================================================
q = 1.602e-19     # 電子電荷 [C]
k = 1.381e-23     # ボルツマン定数 [J/K]
T = 300           # 温度 [K]
h = 6.626e-34     # プランク定数 [J·s]
eps0 = 8.854e-12  # 真空誘電率 [F/m]

# ==================================================
# パラメータ設定 / Parameters
# ==================================================
mu = 300e-4          # 移動度 [m^2/Vs]
L = 20e-9            # チャネル長 [m]
Vgs = 0.8            # ゲート電圧 [V]
Vth = 0.3            # しきい値電圧 [V]
Vds = 0.5            # ドレイン電圧 [V]
lambda_ch = 0.05     # チャネル長変調係数 [1/V]
Cox = 1e-2           # 酸化膜容量 [F/m^2]
Cq = 2e-2            # 量子容量 [F/m^2]
Cf = -5e-3           # 負容量(NCFET用)[F/m^2]
Hfin = 10e-9         # フィン高さ [m]
Tfin = 6e-9          # フィン厚 [m]
n_sheet = 3          # GAAチャネル層数
Hsheet = 5e-9        # ナノシート高さ [m]
Wsheet = 10e-9       # ナノシート幅 [m]
M = 2                # CNT チャネル数

# ==================================================
# 有効幅計算 / Effective Width
# ==================================================
Weff_FinFET = 2 * Hfin + Tfin
Weff_GAA = n_sheet * (2 * Hsheet + Wsheet)
Weff_C = Weff_FinFET + Weff_GAA  # CFET 合成例
Weff_CNT = M * 1e-9  # CNT換算仮値 [m]

# ==================================================
# 有効容量 / Effective Capacitance
# ==================================================
Ceff_FET = Cox
Ceff_GAA = 1 / (1/Cox + 1/Cq)
Ceff_NCFET = (Cox * Cf) / (Cox + Cf)  # Cf<0 の場合 |Ceff|>Cox
Ceff_2D = 1 / (1/Cox + 1/Cq)
Ceff_TFET = Cox  # 近似的扱い

# ==================================================
# 一般式(General FET Equation)
# ==================================================
def Id(mu, Ceff, Weff, L, Vgs, Vth, Vds, lam):
    return 0.5 * mu * Ceff * (Weff/L) * (Vgs - Vth)**2 * (1 + lam * Vds)

def gm(Id, Vgs, Vth):
    return 2 * Id / (Vgs - Vth)

def ro(Id, lam):
    return 1 / (lam * Id)

# ==================================================
# 各構造ごとのドレイン電流計算 / Drain Current per Architecture
# ==================================================
Id_FinFET = Id(mu, Ceff_FET, Weff_FinFET, L, Vgs, Vth, Vds, lambda_ch)
Id_GAA = Id(mu, Ceff_GAA, Weff_GAA, L, Vgs, Vth, Vds, lambda_ch)
Id_CFET = Id_FinFET + Id_GAA
Id_NCFET = Id(mu, Ceff_NCFET, Weff_FinFET, L, Vgs, Vth, Vds, lambda_ch)
Id_TFET = np.exp(-1e9 / (Vds / L)) * 1e-6  # トンネル近似モデル
Id_CNTFET = (4 * q**2 / h) * M * (Vgs - Vth)

# ==================================================
# 出力 / Output
# ==================================================
fet_list = {
    "FinFET": Id_FinFET,
    "GAA": Id_GAA,
    "CFET": Id_CFET,
    "NCFET": Id_NCFET,
    "TFET": Id_TFET,
    "CNTFET": Id_CNTFET
}

print("=== Advanced FET Models (Analytical Evaluation) ===")
for name, current in fet_list.items():
    gm_val = gm(current, Vgs, Vth)
    ro_val = ro(current, lambda_ch)
    print(f"\n{name}")
    print(f"  Id  = {current:.3e} [A]")
    print(f"  gm  = {gm_val:.3e} [S]")
    print(f"  ro  = {ro_val:.3e} [Ω]")

# ==================================================
# 補足:温度依存・スケーリング指標
# ==================================================
S = (np.log(10) * k * T / q) * (1 + 0.1)  # サブスレッショルド係数近似
print(f"\nSubthreshold slope S ≈ {S*1e3:.1f} mV/dec (ideal ≈ 60 mV/dec)")
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?