MOS Transistors: Structure, Operation, and Derivations

Fundamentals of MOS Transistors

MOS (Metal-Oxide-Semiconductor) transistors are foundational electronic components utilized in modern digital and analog circuitry. They act as the primary switching and amplifying elements in personal devices, wireless communications, and microcontrollers, including smart phones, laptops, Wi-Fi and Bluetooth modems, and Arduino boards.

MOS transistors operate primarily as voltage-controlled switches that establish or break a conductive channel between two main terminals: the Source and the Drain.

MOS transistors are classified into two fundamental device types:

  • nMOS Transistors:

    • Function as a switch connecting the Source terminal to the Drain terminal.

    • Turn ON when the Gate-to-Source voltage (VGSV_{GS}) exceeds the threshold voltage (VthV_{th}), which typically exhibits a nominal value of around 1 V1\,\text{V}.

    • Positive Gate-to-Source voltages (VGS>0 VV_{GS} > 0\,\text{V}) activate device conduction.

  • pMOS Transistors:

    • Function as a switch connecting the Source terminal to the Drain terminal.

    • Turn ON when the Gate-to-Source voltage (VGSV_{GS}) drops below the threshold voltage (VthV_{th}), which typically exhibits a nominal value of around −1 V-1\,\text{V}.

    • Negative Gate-to-Source voltages (VGS<0 VV_{GS} < 0\,\text{V}) activate device conduction.

nMOS and pMOS switch operation and symbol basics

Physical Structure and Terminal Definitions

A MOSFET is inherently a four-terminal solid-state device comprising:

  1. Gate (GG): The control terminal positioned directly above the channel region.

  2. Drain (DD): The terminal where charge carriers exit the channel.

  3. Source (SS): The terminal where charge carriers enter the channel.

  4. Substrate / Body (BB): The semiconductor bulk substrate on which the transistor is fabricated.

In standard digital circuit implementations, the Source terminal and the Substrate terminal are internally or externally tied together to establish a baseline reference potential.

3D physical structure of an N-channel MOSFET
Structural Dimensions and Composition
  • Gate Material: Composed of highly conductive Polysilicon.

  • Dielectric Layer: A thin insulating layer of Silicon Dioxide (SiO2\text{SiO}_2) separates the conductive Gate electrode from the underlying semiconductor substrate, forming a physical parallel-plate capacitor geometry.

  • Channel Dimensions: The Gate region is defined by physical length (LL) along the path of current flow and width (WW) perpendicular to current flow.

  • Threshold Voltage (VTV_T): The minimum Gate-to-Source voltage required to induce an active conduction channel (inversion layer) beneath the dielectric layer. Threshold voltage is an intrinsic material and design parameter, typically ranging between 0.5 V0.5\,\text{V} and 3 V3\,\text{V}.

When VGS<VTV_{GS} < V_T, no conducting channel is present, resulting in a Drain-to-Source current of zero (IDS=0 AI_{DS} = 0\,\text{A}).

When a positive voltage (VGS>VTV_{GS} > V_T) is applied to the Gate terminal of an nMOS device, negative charge carriers (electrons) are attracted to the substrate surface, forming an inversion layer that creates a continuous conductive path between the Drain and Source diffusions.

Cross-sectional diagrams of nMOS and pMOS devices
Material Doping Profiles
  • nMOS Construction: Fabricated on a pp\text{-type} silicon bulk substrate with two heavily doped n+n^+ diffusion regions acting as the Source and Drain terminals.

  • pMOS Construction: Fabricated on an nn\text{-type} silicon bulk substrate with two heavily doped p+p^+ diffusion regions acting as the Source and Drain terminals.

Types and Circuit Symbols of MOS Transistors

MOS transistors are categorized based on channel existence at zero Gate bias:

  1. Enhancement Mode Transistors:

    • No physical channel exists between the Drain and Source at zero Gate bias (VGS=0 VV_{GS} = 0\,\text{V}).

    • The application of a suitable Gate-to-Source voltage induces and creates the conducting channel.

    • Includes nMOS Enhancement mode and pMOS Enhancement mode transistors.

  2. Depletion Mode Transistors:

    • A physical conductive channel naturally exists between the Drain and Source at zero Gate bias (VGS=0 VV_{GS} = 0\,\text{V}).

    • The application of a Gate-to-Source voltage depletes carrier concentration, reducing channel width and eventually turning the device off.

    • Includes nMOS Depletion mode and pMOS Depletion mode transistors.

Enhancement Mode versus Depletion Mode physical mechanisms
Schematic Symbols

Circuit representations capture either 4-terminal connections (GG, SS, DD, BB) or simplified 3-terminal connections (GG, SS, DD). When the fourth Body terminal (BB) is omitted from schematic diagrams, it is implicitly assumed to be tied to the appropriate voltage supply rail (GNDGND for nMOS, VDDV_{DD} for pMOS).

Circuit symbols for 3-terminal and 4-terminal NMOS and PMOS devices

Semiconductor Surface States and Channel Formation

Applying varying voltages to the Gate of a metal-oxide-semiconductor stack on a pp\text{-type} substrate drives three distinct surface states beneath the Gate oxide insulator:

1. Accumulation Mode
  • Bias Condition: Vg<0 VV_g < 0\,\text{V}

  • Physical Behavior: A negative voltage on the Gate deposits negative charges onto the Polysilicon electrode.

  • Carrier Dynamics: Mobile positively charged holes within the pp\text{-type} bulk substrate are electrostatically attracted to the surface directly underneath the SiO2\text{SiO}_2 dielectric layer, accumulating majority positive carriers near the interface.

MOS structure operating in Accumulation Mode
2. Depletion Mode
  • Bias Condition: 0 V<Vg<Vt0\,\text{V} < V_g < V_t

  • Physical Behavior: A small positive voltage on the Gate deposits positive charge onto the Polysilicon electrode.

  • Carrier Dynamics: Positively charged holes in the pp\text{-type} body are electrostatically repelled away from the region beneath the Gate, leaving behind fixed, uncompensated negative acceptor ions and forming a non-conductive depletion region free of mobile charge carriers.

MOS structure operating in Depletion Mode
3. Inversion Mode
  • Bias Condition: Vg>VtV_g > V_t

  • Physical Behavior: A higher positive potential exceeding the threshold voltage VtV_t is established on the Gate terminal.

  • Carrier Dynamics: Majority holes are repelled deeper into the substrate bulk, while minority free electrons present in the pp\text{-type} bulk are strongly attracted toward the surface underneath the Gate.

  • Inversion Layer: When the concentration of attracted minority electrons at the surface equals or exceeds the bulk acceptor concentration (ns≅NAn_s \cong N_A), an nn\text{-type} conductive layer called the inversion layer forms in the pp\text{-type} substrate. This inverted channel permits electron flow between n+n^+ diffusion terminals.

MOS structure operating in Inversion Mode
Physics of Threshold Voltage and Surface Potential

The threshold voltage VtV_t depends directly on the doping concentration of the body substrate and the thickness toxt_{ox} of the gate silicon dioxide dielectric layer.

The surface potential ψs\psi_s governing carrier distribution is expressed mathematically as:

ψs=KTqln⁡(nsnB)\psi_s = \frac{K T}{q} \ln\left(\frac{n_s}{n_B}\right)

where:

  • KK = Boltzmann constant

  • TT = Absolute temperature

  • qq = Electron charge magnitude

  • nsn_s = Electron concentration at the semiconductor surface

  • nBn_B = Majority carrier concentration in the bulk substrate

Beyond strong inversion, surface concentration scales logarithmically as ln⁡(nsnB)\ln\left(\frac{n_s}{n_B}\right), causing the surface potential to pin at approximately twice the Fermi potential when ns≅NAn_s \cong N_A.

Operating Regions and Mathematical Derivations

MOSFET operation is divided into three distinct functional operating regions depending on applied bias voltages VGSV_{GS} and VDSV_{DS}.

1. Cut-off Region
  • Operating Conditions: VGS<VTV_{GS} < V_T OR VGS>VTV_{GS} > V_T with VDS=0 VV_{DS} = 0\,\text{V}.

  • Physical State: If VGS<VTV_{GS} < V_T, no inversion layer forms. If VGS>VTV_{GS} > V_T but VDS=0 VV_{DS} = 0\,\text{V}, an inversion layer exists, but no potential gradient drives carrier drift along the channel length.

  • Current Flow: The ideal Drain-to-Source current is zero:

ID=0I_D = 0

  • Practical Non-ideality: Although approximated as zero for simplified functional analysis, tiny subthreshold leakage currents persist through OFF transistors. When multiplied across millions or billions of integrated devices on microchips, subthreshold leakage contributes measurably to total static power consumption.

2. Non-Saturation (Linear / Resistive) Region
  • Operating Conditions: VGS≥VthV_{GS} \ge V_{th} AND 0 V<VDS<Vsat=VGS−Vth0\,\text{V} < V_{DS} < V_{sat} = V_{GS} - V_{th}.

  • Physical State: An inversion layer exists along the full channel length from Source to Drain. Applying a non-zero voltage difference VDSV_{DS} creates a uniform horizontal electric field EE across the channel, causing current IdsI_{ds} to flow from Drain to Source.

  • Derivation of Linear Current Equation:

The current IdsI_{ds} represents total mobile channel charge per unit drift transition time:

Ids=Qchannelt=Charge flowing through the channelTime taken to drift across the channel— (1)I_{ds} = \frac{Q_{channel}}{t} = \frac{\text{Charge flowing through the channel}}{\text{Time taken to drift across the channel}} \quad \text{--- (1)}

The electron drift time tt across channel length LL under electric field E=VdsLE = \frac{V_{ds}}{L} with electron mobility μn\mu_n is:

t=LchannelVelocity of electron=LμnE=LμnVdsL=L2μnVds— (2)t = \frac{L_{channel}}{\text{Velocity of electron}} = \frac{L}{\mu_n E} = \frac{L}{\mu_n \frac{V_{ds}}{L}} = \frac{L^2}{\mu_n V_{ds}} \quad \text{--- (2)}

At any point xx along the channel (0≤x≤L0 \le x \le L), local channel voltage is V(x)V(x), making local Gate-to-channel potential equal to VGS−V(x)V_{GS} - V(x). Assuming local voltage exceeds VTV_T everywhere, the channel charge is computed using parallel-plate capacitance rules.

The effective channel charge QchannelQ_{channel} is driven by Gate capacitance CgC_g multiplied by average overdrive potential along the channel. Because channel potential ranges linearly from 0 V0\,\text{V} at the Source to VdsV_{ds} at the Drain, average channel potential drop is 12Vds\frac{1}{2} V_{ds}:

Qchannel=CgV— (3)Q_{channel} = C_g V \quad \text{--- (3)}

Qchannel=Cg(Vgs−Vt−12Vds)— (4)Q_{channel} = C_g \left( V_{gs} - V_t - \frac{1}{2} V_{ds} \right) \quad \text{--- (4)}

where Vgs−VtV_{gs} - V_t is defined as the effective voltage (or overdrive voltage).

Substituting equation (2) and equation (4) into equation (1):

Ids=Cg(Vgs−Vt−12Vds)L2μnVds— (5)I_{ds} = \frac{C_g \left( V_{gs} - V_t - \frac{1}{2} V_{ds} \right)}{\frac{L^2}{\mu_n V_{ds}}} \quad \text{--- (5)}

The parallel-plate Gate capacitance CgC_g relates dielectric permittivity εox\varepsilon_{ox}, Gate oxide thickness toxt_{ox}, and surface area A=W×LA = W \times L:

Cg=εAd=εoxWLtox— (6)C_g = \frac{\varepsilon A}{d} = \frac{\varepsilon_{ox} W L}{t_{ox}} \quad \text{--- (6)}

Substituting equation (6) into equation (5):

Ids=εoxWLtox(Vgs−Vt−12Vds)L2μnVds— (7)I_{ds} = \frac{\frac{\varepsilon_{ox} W L}{t_{ox}} \left( V_{gs} - V_t - \frac{1}{2} V_{ds} \right)}{\frac{L^2}{\mu_n V_{ds}}} \quad \text{--- (7)}

Defining fundamental device structural parameters:

  • Aspect Ratio: WL\frac{W}{L}

  • Process Transconductance Parameter (knk_n):

kn=μnεoxtoxk_n = \frac{\mu_n \varepsilon_{ox}}{t_{ox}}

  • MOS Transistor Gain Factor (βn\beta_n or β\beta):

βn=kn⋅WL=μnCoxWL\beta_n = k_n \cdot \frac{W}{L} = \mu_n C_{ox} \frac{W}{L}

Simplifying equation (7) using device gain factor β\beta yields the complete non-saturation current formula:

Ids=β(Vgs−Vt−12Vds)Vds— (8)I_{ds} = \beta \left( V_{gs} - V_t - \frac{1}{2} V_{ds} \right) V_{ds} \quad \text{--- (8)}

valid for VGS≥VthV_{GS} \ge V_{th} and VDS<Vsat=VGS−VthV_{DS} < V_{sat} = V_{GS} - V_{th}.

3. Saturation Region
  • Operating Conditions: VGS>VTV_{GS} > V_T AND VDS≥Vsat=VGS−VTV_{DS} \ge V_{sat} = V_{GS} - V_T (or VGS−VDS≤VTV_{GS} - V_{DS} \le V_T).

  • Pinch-off Mechanism: As VDSV_{DS} increases, local Gate-to-channel voltage at the Drain end (VGS−VDSV_{GS} - V_{DS}) declines. When VDSV_{DS} reaches or exceeds VGS−VTV_{GS} - V_T, local Gate-to-channel potential at the Drain drops below threshold voltage VTV_T.

  • The induced mobile charge concentration at the Drain edge drops to zero, causing the inversion layer channel to physically disappear or pinch off.

Channel pinch-off occurrence in Saturation Region
  • Saturation Current Behavior: Further increases in VDSV_{DS} beyond VsatV_{sat} shift the pinch-off point slightly toward the Source, but potential difference across the active inversion channel length remains fixed at Vsat=VGS−VTV_{sat} = V_{GS} - V_T. Consequently, drift current IdsI_{ds} stops increasing and saturates at a constant value.

  • Derivation of Saturation Current Formula:

To determine saturation current, substitute Vds=Vsat=Vgs−VtV_{ds} = V_{sat} = V_{gs} - V_t into non-saturation current equation (8):

Ids=β(Vgs−Vt−12(Vgs−Vt))(Vgs−Vt)I_{ds} = \beta \left( V_{gs} - V_t - \frac{1}{2} (V_{gs} - V_t) \right) (V_{gs} - V_t)

Ids=β(Vgs−Vt2)(Vgs−Vt)I_{ds} = \beta \left( \frac{V_{gs} - V_t}{2} \right) (V_{gs} - V_t)

Ids=β[Vgs−Vt]22— (9)I_{ds} = \beta \frac{[V_{gs} - V_t]^2}{2} \quad \text{--- (9)}

Expanding gain factor β=μnCoxWL\beta = \mu_n C_{ox} \frac{W}{L} yields standard saturation current equations:

Ids=ID=12μnCoxWL[Vgs−Vt]2=β([Vgs−Vt]22)I_{ds} = I_D = \frac{1}{2} \mu_n C_{ox} \frac{W}{L} [V_{gs} - V_t]^2 = \beta \left( \frac{[V_{gs} - V_t]^2}{2} \right)

Output Characteristic Curves

The relation between Drain current IDI_D and Drain-to-Source voltage VDSV_{DS} across different Gate-to-Source bias levels (VGS=1.0 VV_{GS} = 1.0\,\text{V}, 1.5 V1.5\,\text{V}, 2.0 V2.0\,\text{V}, 2.5 V2.5\,\text{V}) highlights operating boundaries.

Drain Current versus Drain-Source Voltage Output Characteristic Curves

Key features displayed on output characteristics:

  • Resistive / Linear Boundary: At low VDSV_{DS} (VDS<VGS−VTV_{DS} < V_{GS} - V_T), current increases linearly with increasing VDSV_{DS}.

  • Parabolic Saturation Boundary Line: Defined by locus points satisfying VDS=VGS−VTV_{DS} = V_{GS} - V_T, demarcating linear and saturation operational regimes.

  • Quadratic Dependence: In saturation (VDS≥VGS−VTV_{DS} \ge V_{GS} - V_T), current curves flatten out into horizontal lines whose vertical spacing exhibits a square-law (quadratic) dependence on Gate overdrive voltage (V_{GS} - V_T)^2$.\n\n# Summary of Operating Equations\n\n\begin{array}{|c|c|c|} \hline \mathbf{Operating\,Region} & \mathbf{Voltage\,Conditions} & \mathbf{Drain\,Current\,(I_D)\,Equation} \ \hline \text{Cut-off Region} & V_{GS} < V_T \quad \text{or} \quad V_{DS} = 0\,\text{V} & I_D = 0 \ \hline \text{Linear / Non-saturation} & V_{GS} \ge V_T \quad \text{and} \quad 0\,\text{V} < V_{DS} < (V_{GS} - V_T) & I_D = \mu_n C_{ox} \frac{W}{L} \left( V_{gs} - V_t - \frac{1}{2} V_{ds} \right) V_{ds} = \beta \left( V_{gs} - V_t - \frac{1}{2} V_{ds} \right) V_{ds} \ \hline \text{Saturation Region} & V_{GS} > V_T \quad \text{and} \quad V_{DS} \ge (V_{GS} - V_T) & I_{ds} = I_D = \frac{1}{2} \mu_n C_{ox} \frac{W}{L} [V_{gs} - V_t]^2 = \beta \left( \frac{[V_{gs} - V_t]^2}{2} \right) \ \hline \end{array}$$