16 Semiconductors: Fabrication and Switching Devices
16.1 Start With the Decision
An integrated circuit is built in layers that form many controlled junctions. Those junctions become switches with real voltage and loss limits.
16.2 Route Overview
This is part 2 of 2. Review Semiconductors: Doping and Diode Behaviour for the preceding evidence.
16.3 Learning Objectives
- Outline the main stages of integrated-circuit fabrication.
- Compare MOSFET and BJT switching limits and losses.
16.4 Chapter Roadmap
- How ICs Are Actually Made
- Cross-Hub Connections
- IoT Transistor Selection
- MOSFET vs BJT Power Loss
- GPIO Voltage MOSFET Limits
- Deep dive: Re-deriving the chapter’s numbers
- Checkpoint: Choosing and Driving Transistors
- Knowledge Check
- Quiz 1: Semiconductor Switching Devices
- Quiz 2: Comprehensive Review
- For Kids: Meet the Sensor Squad!
- Flyback Diode for 12V Relay
- BJT vs MOSFET Checklist
- Checkpoint: Applied Diode and Transistor Design
- Concept Check: PN Junction Behavior
- Concept Relationships
- Try It Yourself
- Key Takeaway
- See Also
- Match the Concepts
- Order the Steps
- Common Pitfalls
- Diode Voltage Variation
- Silicon Diode Recovery Check
- LED Reverse Voltage Limits
- Diode Forward Voltage Varies
- Label the Diagram
- What’s Next
- Code Challenge
- Diode Clamp and Recovery
- Summary
Every junction and layer structure in this chapter — the diode’s single PN junction, the transistor’s two back-to-back junctions — can be built two ways: as a discrete part in its own package, wired to its neighbors with copper traces, or printed directly onto a shared piece of silicon alongside thousands or billions of other components. The second option is an integrated circuit (IC), and the printing process is a direct extension of the doping this chapter just covered.
Three steps turn a design into silicon:
- Lithography — the circuit design is drawn onto a photomask, then projected through it onto a photoresist-coated silicon wafer, exposing the pattern in UV light.
- Etching — the exposed pattern is “carved” into the wafer, opening windows down to the silicon where each component will form.
- Deposition — dopants and metal are deposited into those windows, forming the p-type and n-type regions for diodes and transistors and the metal traces that wire them together.
A diode becomes a single p-type/n-type boundary drawn as a shape instead of a soldered part. A transistor becomes the same n-p-n or p-n-p sandwich this section just described, printed with three doped regions instead of assembled from three physical layers. Resistors and capacitors get the same treatment: a resistor is a shaped strip of doped silicon with a controlled width and length, and a capacitor is a doped plate separated from a conductor by a thin insulating layer — the identical R and C behavior from the electricity chapters, just fabricated instead of soldered.
Why not make everything an IC? ICs “componentize” a circuit — a designer drops in one part instead of understanding or re-verifying every internal transistor — and cost per unit falls sharply once a design is mass-produced. But the photomask and fabrication setup for a new IC design carries a large fixed cost, so ICs only make economic sense at high production volumes. High-current or high-voltage circuits that exceed on-chip logic levels, and low-volume or novel designs with no existing chip to buy, still get built from discrete diodes, transistors, resistors, and capacitors — the same parts this chapter covers, just assembled by hand or machine instead of printed.
The same doped-semiconductor trick pays off beyond switching. Because doping controls how a semiconductor’s electrons behave, the identical fabrication process that builds a switching transistor also builds: LEDs, where excited electrons release their energy as light instead of heat; sensors, where a material’s resistivity shifts predictably with temperature, light, or gas exposure (thermistors, photoresistors, gas sensors); and heat-dissipating structures, since silicon’s high thermal conductivity carries waste heat away from the same junctions that are switching current.
Explore More Electronics Resources:
- Simulations Hub - Try interactive circuit simulators (TinkerCAD, Falstad) to build transistor switches and test diode circuits without physical components
- Videos Hub - Watch visual explanations of semiconductor physics, PN junction behavior, and transistor operation at the atomic level
- Quizzes Hub - Test your understanding of BJT vs MOSFET selection, power dissipation calculations, and circuit design decisions
- Knowledge Gaps Hub - Clarify common confusions about gate drive voltage, current vs voltage control, and thermal management
Why This Matters: Electronics fundamentals connect to nearly every IoT topic. Understanding transistors enables you to design sensor interfaces (signal conditioning), control actuators (motor drivers), optimize power consumption (load switching), and troubleshoot hardware issues (thermal problems, shoot-through).
When choosing between BJT and MOSFET for an IoT switching application, follow this 3-step decision process:
Step 1: Determine load current
| Load Current | Recommendation | Reason |
|---|---|---|
| < 100 mA | BJT (2N2222, BC547) | Simple, $0.02-0.05, no gate driver needed |
| 100 mA - 5 A | Logic-level MOSFET (IRLZ44N) | Low Rds(on) (~22mΩ), direct GPIO drive at 3.3V |
| > 5 A | MOSFET + gate driver (IR2110) | High-side switching, proper gate voltage |
Step 2: Check supply voltage compatibility
| MCU Voltage | BJT Base Resistor | MOSFET Gate Drive |
|---|---|---|
| 3.3V (ESP32, STM32) | R_base = (3.3V - 0.7V) / (I_load / hFE) | Must use logic-level MOSFET (Vgs_th < 2V) |
| 5V (Arduino Uno) | R_base = (5V - 0.7V) / (I_load / hFE) | Standard logic-level MOSFET (Vgs_th < 3V) works |
Step 3: Calculate power dissipation
- BJT: P = V_CE(sat) x I_C = 0.2V x I_load (typically 0.1-0.3W for small loads)
- MOSFET: P = I^2^ x Rds(on) (e.g., 2A^2^ x 0.022 ohm = 0.088W for IRLZ44N at Vgs=10V)
Worked example: ESP32 driving a 12V DC fan (0.5A)
- Option A (BJT 2N2222): R_base = (3.3 - 0.7) / (0.5/100) = 520 ohm. Power = 0.3V x 0.5A = 0.15W. Cost: $0.03. Works fine.
- Option B (MOSFET IRLZ44N): Direct GPIO to gate. Power = 0.5^2^ x 0.022 = 0.0055W at Vgs=10V (slightly higher at 3.3V Vgs). Cost: $0.65. Runs cooler.
- Verdict: BJT is cheaper and adequate at 0.5A. MOSFET becomes clearly superior above 1A where BJT dissipation exceeds 0.3W.
Before applying the rule, view Figure 16.1 and locate the labelled parts of this is the TO-220 package Option B in the worked example above ships in — three leads for gate, drain, and source, plus a metal tab that would bolt to a heatsink if losses climbed past the 0.0055W the IRLZ44N dissipates driving the 12V fan at 0.5A. Photo: Suyash Dwivedi, CC BY-SA 4.0.
Within Figure 16.1, This is the TO-220 package Option B in the worked example above ships in — three leads for gate names the first relevant element, and source identifies the second, and CC BY-SA 4.0 states the result or constraint. Their relationship shows this is the TO-220 package Option B in the worked example above ships in — three leads for gate, drain, and source, plus a metal tab that would bolt to a heatsink if losses climbed past the 0.0055W the IRLZ44N dissipates driving the 12V fan at 0.5A. Photo: Suyash Dwivedi, CC BY-SA 4.0. The surrounding section uses that labelled evidence to continue its technical argument.
Compare power dissipation between BJT and MOSFET switching for your load current. Adjust the sliders to see which device runs cooler.
The Misconception: Many beginners assume any MOSFET can be driven directly from microcontroller GPIO pins (3.3V or 5V), leading to circuits that “sort of work” but run hot, waste power, and fail prematurely.
The Reality with Real Numbers: Standard MOSFETs require Vgs = 10-12V to fully turn ON and achieve their datasheet specifications. When driven with insufficient gate voltage, they operate in the linear region with dramatically higher on-resistance.
Real-World Impact (IRF540N Example):
- Datasheet spec: Rds(on) = 77mΩ @ Vgs=10V
- With 5V gate drive: Rds(on) ≈ 200mΩ (2.6× higher!)
- Switching 2A load:
- Expected power: (2A)² × 0.077Ω = 0.31W (manageable)
- Actual power: (2A)² × 0.2Ω = 0.8W (hot!)
- TO-220 package without heatsink: 1W max → MOSFET runs at 80°C → premature failure
Measured in Real Projects: A student designed a smart fan controller using IRF540N driven from 5V Arduino. Initial testing worked fine for 10 minutes, but after 30 minutes the MOSFET reached 95°C and the fan speed dropped 40% due to voltage drop across the partially-on MOSFET. After switching to IRLZ44N (logic-level), temperature dropped to 35°C and fan ran at full speed.
The Solution: Always check the Vgs specification in datasheets:
- Logic-level MOSFETs: low Rds(on) specified at 4.5-5V, and preferably at 2.5-3.3V for ESP32-class GPIO; Vgs(th) only tells you where conduction begins
- Standard MOSFETs: Vgs(th) = 2-4V, but require 10V for rated Rds(on)
Cost of This Mistake:
- Wasted heat: 0.8W - 0.31W = 0.49W continuous loss
- Battery life: For a 2A load running 50% duty cycle, that’s 5.9Wh/day wasted → significant drain on any battery-powered system
- Component damage: Thermal stress reduces MOSFET lifetime from 100,000 hours to <10,000 hours
- Debugging time: Students typically spend 2-4 hours troubleshooting “hot MOSFET” before discovering gate voltage issue
Key Takeaway: When selecting MOSFETs for microcontroller projects, ALWAYS verify Rds(on) at YOUR gate voltage (not the datasheet headline spec). Add 50% safety margin for thermal calculations. Logic-level MOSFETs cost the same ($0.50) but save hours of debugging and prevent field failures.
| Check | Arithmetic | What it proves |
|---|---|---|
| Doping creates many more carriers | Silicon has 5 x 10^22 atoms/cm³. At 1 dopant per 10^6 atoms, 5 x 10^22 / 10^6 = 5 x 10^16 dopant carriers/cm³. Compared with 1.5 x 10^10 intrinsic carriers/cm³, the ratio is 5 x 10^16 / 1.5 x 10^10 = 3.33 x 10^6. | The “roughly 3 million times” conductivity improvement is the right order of magnitude. |
| Electron mobility is about three times hole mobility | The chapter gives electron mobility mu_n = 1400 cm²/(V·s) and hole mobility mu_p = 450 cm²/(V·s). 1400 / 450 = 3.11. | N-channel devices often switch faster because electron mobility is about 3.1x hole mobility, not because holes are absent. |
| A forward-biased rectifier must dissipate heat | The 1N4007 example gives 1 A and 0.9 V. P = V x I = 0.9 x 1 = 0.9 W. In reverse at 12 V and 0.5 µA, P = 12 x 0.5 x 10^-6 = 6 µW, or 0.006 mW. | Forward conduction is the thermal case; normal reverse leakage is negligible until breakdown is approached. |
| Schottky savings are a supply-budget calculation | Replacing a 0.7 V silicon drop with a 0.3 V Schottky drop saves 0.4 V. On a 5 V rail, 0.4 / 5 = 0.08, or 8%. | The “recovers 8% efficiency” claim is a voltage-drop fraction, so it applies when diode drop is a dominant series loss. |
| MOSFET switch loss scales differently from BJT loss | For the 12 V, 0.5 A fan example, BJT loss is 0.3 x 0.5 = 0.15 W. MOSFET loss is 0.5^2 x 0.022 = 0.0055 W. 0.15 / 0.0055 = 27.3. | Above small currents, I^2R MOSFET loss can be tens of times lower than BJT saturation loss when the gate is driven properly. |
Carry full precision through the calculation, then round for the learner-facing statement. That keeps the panel consistent with the chapter’s own warning: gate threshold, voltage drop, and current gain are selection inputs, not decorative datasheet trivia.
Checkpoint: Choosing and Driving Transistors
You now know:
- BJTs are current-controlled and need continuous base drive; MOSFETs are voltage-controlled with near-zero gate current — which is why MOSFETs win in battery-powered IoT designs.
- A 3.3 V GPIO needs a logic-level MOSFET. A standard MOSFET like the IRF540N specifies 77 mΩ at Vgs = 10 V but rises to about 200 mΩ with 5 V gate drive.
- Conduction loss is P = I² × Rds(on) — always verify Rds(on) at your gate voltage, not the datasheet headline spec.
16.5 Knowledge Check
Test your understanding with these questions. They put the selection rules above under field conditions — treat them as part of the chapter, not optional homework.
The Sensor Squad learns about one-way doors and super switches!
Temperature Terry had a problem. “Bella, what happens if someone plugs your battery in backwards?” the battery looked worried. “That could fry Max and all our friends!”
the microcontroller smiled. “Don’t worry — we have a special protector called a DIODE. Think of it like a one-way door. Electricity can flow through in one direction, but if it tries to go backwards, the door slams shut!”
“How does that work?” asked the LED. Max explained: “Remember how we learned about N-type and P-type semiconductors? When you stick them together, you get a PN junction. One side has extra electrons, the other side has holes where electrons are missing. When electricity pushes the right way, electrons jump across to fill the holes and current flows. But push the wrong way, and a big empty gap forms that blocks everything!”
“Wait,” said Lila, “my name has LED in it — Light Emitting Diode! Am I a diode too?” “Yes!” said Max. “You’re a special diode that gives off light when electrons jump across the junction. Different materials make different colors — that’s why you can glow red, green, or blue!”
Then Sammy asked the big question: “But what about transistors? How do they work?” Max got really excited. “A transistor is like TWO PN junctions sandwiched together — NPN or PNP. It’s a super switch! I send a tiny whisper of electricity to the middle layer called the base, and it controls a HUGE flow of electricity through the other two layers. It’s like a kid whispering to a giant who then opens or closes a massive gate!”
Bella clapped. “So that’s how you control the big fan motor even though you’re so small!” Max nodded proudly. “One tiny signal from me, and the transistor lets all of Bella’s power flow through to spin the motor. That’s the magic of semiconductors!”
16.5.1 Key Words for Kids
| Word | What It Means |
|---|---|
| Diode | A one-way door for electricity — lets current flow one way only |
| LED | Light Emitting Diode — a diode that glows with light |
| Transistor | A super switch controlled by a tiny signal |
| N-type | Semiconductor with extra electrons (negative charges) |
| P-type | Semiconductor with missing electrons called holes (positive charges) |
| PN Junction | Where N-type meets P-type — the magic boundary that makes diodes work |
Scenario: You are building an ESP32-controlled smart home hub that switches a 12V electromagnetic relay (coil: 120mA, inductance: 150mH). The relay controls a 120V AC lamp. What flyback diode should you use?
Step 1: Calculate the inductive voltage spike (worst case)
When the relay coil current is cut, the inductor generates: V = -L × (di/dt)
Worst-case scenario: transistor switches off in 100ns (modern MOSFETs are fast!)
V_spike = -(150 × 10⁻³ H) × (0.12A / 100 × 10⁻⁹ s) = -180,000V
Without protection, this idealized spike would instantly destroy the transistor (rated for 60V max). In a real circuit, parasitic capacitance and arcing limit the spike to hundreds of volts rather than 180kV — still far beyond the transistor’s rating, which is why the diode is non-negotiable.
Step 2: Select diode voltage rating
The flyback diode must withstand at least the supply voltage in reverse bias: V_reverse > 12V
Add 2× safety margin: 12V × 2 = 24V minimum
Common diodes meeting this:
- 1N4148 (100V, 300mA fast-switching signal diode)
- 1N4007 (1000V, 1A general-purpose rectifier)
- 1N5819 Schottky (40V, 1A, fast recovery)
Step 3: Select diode current rating
The flyback diode must handle the relay coil current: I_diode ≥ 120mA
All three diodes exceed this (300mA to 1A rating).
Step 4: Check reverse recovery time
When the relay switches ON again, the diode must stop conducting quickly to avoid shorting the power supply.
| Diode | Reverse Recovery Time | Suitability |
|---|---|---|
| 1N4148 | 4 ns (very fast) | Excellent for high-frequency switching |
| 1N4007 | 30 µs (slow) | Acceptable for relays (slow mechanical switching) |
| 1N5819 Schottky | <10 ns (very fast) | Best for motor PWM and high-frequency applications |
For a relay switching at <1 Hz, even the slow 1N4007 is adequate.
Step 5: Calculate power dissipation in the diode
When the relay switches OFF, the collapsing magnetic field forces current through the flyback diode.
Energy stored in inductor: E = ½ × L × I² = 0.5 × 0.15H × (0.12A)² = 1.08 mJ
This energy dissipates through the diode over several milliseconds. Average power is negligible for low-frequency relay switching.
Decision: Use 1N4007 — it is cheap ($0.02), over-specified for voltage (1000V >> 12V), handles the current easily, and the slow recovery time does not matter for relay applications.
For motor drivers or high-frequency solenoid valves, upgrade to 1N5819 Schottky for faster recovery and lower forward voltage drop (0.3V vs 0.7V).
To understand why the next claim holds, inspect Figure 16.2. Its labels describe the 1N4007 chosen above for this 12V relay is exactly this part class — a cheap, axial-leaded general-purpose rectifier — installed in reverse bias across the coil, where its slow ~30 microsecond recovery time never matters because the relay switches under 1 Hz. Photo: oomlout, CC BY-SA 2.0.
Compare the labels The 1N4007 chosen above for this 12V relay is exactly this part class — a cheap and where its slow ~30 microsecond recovery time never matters because the relay switches under 1 Hz. Photo: oomlout in Figure 16.2; the additional label axial-leaded general-purpose rectifier — installed in reverse bias across the coil resolves their engineering meaning. This is how the figure shows the 1N4007 chosen above for this 12V relay is exactly this part class — a cheap, axial-leaded general-purpose rectifier — installed in reverse bias across the coil, where its slow ~30 microsecond recovery time never matters because the relay switches under 1 Hz. Photo: oomlout, CC BY-SA 2.0, which the surrounding narrative uses as its next premise.
Verification: After installing the 1N4007, measure the voltage across the transistor with an oscilloscope when switching the relay OFF. You should see voltage clamped to ~12.7V (12V supply + 0.7V diode drop) instead of the 180kV spike.
Use this checklist alongside the detailed Transistor Selection Decision Framework above:
| Question | If YES → | If NO → |
|---|---|---|
| 1. Is load current < 100mA? | BJT (2N2222) works fine, costs $0.02 | Go to Q2 |
| 2. Is the GPIO voltage 5V or higher? | BJT or MOSFET both work | Go to Q3 |
| 3. Is GPIO voltage 3.3V? | Must use logic-level MOSFET (Vgs(th) <2V) | Cannot drive standard MOSFET |
| 4. Is switching frequency >1kHz (PWM)? | MOSFET (BJTs waste power driving base continuously) | BJT acceptable |
| 5. Is battery life critical? | MOSFET (zero gate current) | BJT acceptable |
| 6. Is load current >5A? | MOSFET with proper heatsink | Use relay or SSR instead |
Checkpoint: Applied Diode and Transistor Design
Between the quizzes and the worked examples above, you now know:
- Any inductive load (relay, motor, solenoid) needs a flyback diode: the collapsing field generates V = -L × (dI/dt), and the diode clamps the spike to about 0.7 V above the supply (~12.7 V on a 12 V rail).
- A cheap 1N4007 ($0.02) is fine for slow relay switching; upgrade to a Schottky like the 1N5819 for motor PWM and high-frequency switching.
- The quick selection rule: BJT below 100 mA, logic-level MOSFET from 100 mA to 5 A, MOSFET plus gate driver above 5 A.
16.6 Concept Check: PN Junction Behavior
Test your understanding before proceeding:
16.7 Concept Relationships
This chapter connects to other electronics topics:
| Related Concept | Chapter Link | Relationship |
|---|---|---|
| Transistor Operation | Transistor Selection Guide | Transistors are two PN junctions sharing common layer |
| Voltage Regulation | Power Management | Zener diodes provide voltage reference |
| Flyback Protection | Motor Control Circuits | Diodes clamp inductive voltage spikes |
| Rectification | Power Supplies | Diode bridges convert AC to DC |
16.8 Try It Yourself
You have been reading numbers all chapter — 0.7 V silicon, 0.3 V Schottky, color-dependent LED drops. Now put a multimeter on real parts and watch those same numbers appear on your bench.
Challenge: Measure Diode Forward Voltage Drop
Objective: Verify diode forward voltage varies by type and current, understanding real vs ideal behavior.
Materials Needed:
- Multimeter with diode test mode
- 1N4007 silicon diode
- 1N5819 Schottky diode
- LED (any color)
- 9V battery, 1kΩ resistor, breadboard
Steps:
- Use multimeter diode mode to measure forward voltage of each diode (no circuit needed)
- Build circuit: 9V battery → 1kΩ resistor → diode → ground
- Measure voltage across diode with multimeter
- Calculate current: I = (9V - Vdiode) / 1kΩ
- Compare measured Vf at load current vs diode test mode
Solution Example:
- 1N4007 diode test: 0.65V (low current ~1mA)
- 1N4007 in circuit at 8mA: 0.72V (increases with current)
- 1N5819 Schottky: 0.28V (lower Vf, better efficiency)
- Red LED: 1.8V (forward voltage depends on semiconductor bandgap)
- Blue LED: 3.1V (wider bandgap = higher voltage)
Expected Observation: Forward voltage increases with current due to semiconductor resistance. Schottky diodes have 0.4V lower drop than silicon, saving power. LEDs show color-dependent forward voltage related to photon energy.
Doping creates N-type and P-type semiconductors that, when combined into PN junctions, form diodes (one-way current flow) and transistors (electronically controlled switches). MOSFETs are preferred for most IoT applications because they are voltage-controlled with nearly zero gate current, offer lower power losses, and are simpler to drive from microcontroller GPIO pins.
16.9 See Also
Diode Applications:
Follow these connections in order: Flyback Diode Protection - Protect transistors from inductive spikes; Voltage Regulators - Zener diode voltage references; LED Driver Circuits - Current limiting for LEDs.
Advanced Semiconductor Physics:
Follow these connections in order: Band Gap Theory - Why forward voltage varies by material; Temperature Effects - Forward voltage decreases ~2mV/°C; Diode Clamp and Recovery Contracts - Review forward-voltage drift, GPIO clamp current, Schottky protection choices, and reverse-recovery limits; Transistor Selection Guide - Switching speed considerations for BJT vs MOSFET.
16.10 Match the Concepts
Two quick games — matching and ordering — lock in the vocabulary and the LED design procedure before we look at the most common mistakes.
16.11 Order the Steps
Common Pitfalls
These four mistakes account for most diode-related debugging sessions. Skim the titles now, and open any that touch your current project.
A silicon diode’s forward voltage is approximately 0.6-0.7 V, but this varies: increasing current increases forward voltage; increasing temperature decreases forward voltage by approximately 2 mV/C. In precision rectifier or voltage reference applications, account for these variations. A diode used as a rough 0.7 V reference drifts significantly with temperature.
Standard silicon rectifier diodes (1N4001-1N4007) have reverse recovery times of 1-4 microseconds. At high switching frequencies (>50 kHz), they remain partly conductive during the reverse transition, causing power loss and heating. Use Schottky diodes (nanosecond recovery) or ultrafast diodes for switching power supplies and high-frequency rectifier circuits.
LEDs have very low reverse breakdown voltage (typically 5-30 V). In AC circuits or H-bridge motor drivers, LEDs can be momentarily reverse-biased. Exceeding the reverse voltage instantly destroys the LED junction. Add a small signal diode in reverse parallel with the LED to clamp reverse voltage, or ensure the circuit never applies reverse voltage to the LED.
Forward voltage varies significantly by diode type: germanium 0.2-0.3 V, Schottky 0.2-0.4 V, silicon PN 0.6-0.7 V, LED 1.8-3.5 V depending on color. Using the generic 0.7 V value for a Schottky or LED in a circuit calculation introduces significant errors. Always look up the actual forward voltage for the specific diode type at the expected operating current.
16.12 What’s Next
| Direction | Chapter | Why Read It |
|---|---|---|
| Previous | Electricity Applications | Ohm’s Law and power calculations that underpin diode and transistor circuit design |
| Next | Transistor Selection Guide | Systematically choose BJT vs MOSFET for every IoT load-switching scenario |
| Related | Electricity Fundamentals | Voltage, current, and resistance basics that explain forward voltage and diode conduction |
| Applied | DC Motor Control | See flyback diodes and H-bridge transistor circuits in real actuator designs |
16.13 Diode Clamp and Recovery
The main chapter above stays focused on semiconductors, doping, PN junctions, diodes, and transistor switching choices. For the deeper design contract behind diode forward-voltage approximation, GPIO clamp-current sizing, Schottky reverse-polarity tradeoffs, temperature drift, and reverse-recovery limits, continue to Diode Clamp and Recovery Contracts.
16.14 Summary
This chapter shows how doping creates p-type and n-type semiconductor behavior and how diodes use junctions to control current direction. It connects diode voltage drops, LEDs, flyback protection, and MOSFET behavior to practical IoT switching circuits.
16.15 Continue Your Route
This final part closes the route from How ICs Are Actually Made through Summary. Return to Semiconductors: Doping and Diode Behaviour or continue from the electronics module index.