Entropy and the Second Law: Why Efficiency Has Hard Limits

What Entropy Adds Beyond the First Law

The first law helps you keep an energy “budget”: energy in, energy out, and what gets stored. But it does not tell you whether a proposed machine is actually possible, nor does it tell you why some energy conversions are inevitably wasteful even with perfect insulation and perfect machining. Entropy and the second law provide that missing constraint.

For makers, the second law shows up as hard limits on efficiency, unavoidable heat rejection, and the reason “100% efficient heat engine” or “self-cooling box that powers itself” designs fail. It also gives you a practical way to estimate the best-case performance of real devices (engines, turbines, compressors, refrigerators, heat pumps) before you build them.

Entropy: A Practical Definition You Can Use

Entropy is a property of a system that tracks how energy is distributed relative to temperature and how much of that energy is unavailable for conversion into useful work when interacting with an environment. In practice, entropy is most useful because it lets you compute the minimum possible losses for a process.

The key operational definition is based on reversible heat transfer:

dS = δQ_rev / T

Here, dS is the differential change in entropy, δQ_rev is an infinitesimal amount of heat transferred in a reversible way, and T is the absolute temperature at the boundary where the heat crosses (in kelvin). This is not saying real processes are reversible; it is saying entropy is defined so that if you imagine a reversible path between the same states, you can compute the entropy change.

Continue in our app.

You can listen to the audiobook with the screen off, receive a free certificate for this course, and also have access to 5,000 other free online courses.

Or continue reading below...

Download the app

Two immediate maker-relevant implications:

• Heat moved at a higher temperature carries less entropy per joule than heat moved at a lower temperature (because of the division by T).
• If you dump heat into a cold sink, you increase entropy a lot; if you dump the same heat into a hot sink, you increase entropy less.

Entropy Is a State Property (So You Can Use Tables/Models)

Entropy depends on the state (for example, pressure and temperature for many gases), not on the path taken to get there. That means you can compute ΔS from property data or equations of state without tracking every microscopic detail of the process. In real design work, this is what makes entropy usable: you can compare “ideal” and “actual” paths between the same inlet and outlet states.

The Second Law in Forms Makers Actually Apply

Clausius Statement (Heat Doesn’t Flow Uphill by Itself)

Heat will not spontaneously flow from a colder body to a hotter body. If you want heat to move from cold to hot (refrigeration, heat pumping), you must supply work. This is why a Peltier cooler, vapor-compression fridge, or DIY heat pump always consumes electrical power.

Kelvin–Planck Statement (No Perfect Heat Engine)

No cyclic device can take heat from a single thermal reservoir and convert it entirely into work. A heat engine must reject some heat to a colder reservoir. This is the core reason every engine needs a radiator, cooling fins, exhaust heat, or some other heat rejection path.

Entropy Balance (The Accounting Tool)

For a system, the entropy change equals entropy transferred plus entropy generated:

ΔS_system = ∫(δQ/T_boundary) + S_gen

S_gen is entropy generation due to irreversibilities (friction, mixing, finite temperature differences, electrical resistance, inelastic deformation, turbulence, throttling, chemical reactions, etc.). The second law requires:

S_gen ≥ 0

That inequality is the “hard limit” in equation form. If your proposed process requires S_gen < 0, it cannot happen as drawn.

Reversible vs Irreversible: The Maker’s Intuition

A reversible process is an idealization where the system and surroundings can be returned to their initial states with no net changes anywhere. In practice, reversibility means: no friction, no unrestrained expansion, no mixing of different compositions without separation work, no electrical resistance heating, and heat transfer only across infinitesimal temperature differences.

Real builds are irreversible. The value of the reversible model is that it sets the best possible performance. Your actual device will always be worse by an amount tied to S_gen.

Common Irreversibilities You Can Spot in Hardware

• Friction in bearings, seals, pistons, gears: mechanical work turns into internal energy, increasing entropy.
• Pressure drops in piping and valves: flow losses dissipate useful pressure potential into heat.
• Throttling (expansion through a valve/orifice): large entropy generation; you “spend” pressure without extracting work.
• Heat exchange with a big temperature difference: e.g., a very hot surface heating cool air quickly; fast heat transfer, but high entropy generation.
• Mixing: two streams mix and you cannot unmix them without work; entropy increases.

Why Efficiency Has Hard Limits: Carnot as the Ceiling

Consider any heat engine operating between a hot reservoir at T_H and a cold reservoir at T_C (kelvin). The maximum possible thermal efficiency is the Carnot efficiency:

η_max = 1 - (T_C / T_H)

This is not a “good design guideline”; it is a strict upper bound for any engine, regardless of working fluid, mechanism, or cleverness, as long as it is a cyclic engine exchanging heat with reservoirs.

Practical reading: to get higher maximum efficiency, you must raise the hot-side temperature, lower the cold-side temperature, or both. Everything else (better bearings, better timing, better insulation) helps you approach the bound but cannot exceed it.

Numerical Example: Why Room-Temperature Engines Are Disappointing

Suppose you try to build a small engine using a hot source at 150°C and ambient air at 25°C as the sink. Convert to kelvin: T_H = 423 K, T_C = 298 K.

η_max = 1 - 298/423 ≈ 0.296

Even a perfect engine could only convert about 29.6% of the heat input into work; the rest must be rejected as heat. Real small engines might achieve far less due to friction, heat losses, and non-ideal processes.

Step-by-Step: Estimating the Absolute Best Efficiency of a Heat Engine

• Step 1: Identify the hottest temperature at which heat is actually added to the working fluid (not the flame temperature, but the effective hot-side boundary temperature).
• Step 2: Identify the coldest temperature at which heat is rejected (often near ambient, but higher if your radiator is small or airflow is poor).
• Step 3: Convert both to kelvin: T(K) = T(°C) + 273.15.
• Step 4: Compute η_max = 1 - T_C/T_H.
• Step 5: Treat this as a ceiling; set realistic targets well below it based on expected irreversibilities (pressure drops, finite heat exchanger size, mechanical friction).

Hard Limits for Refrigerators and Heat Pumps (COP Limits)

For devices that move heat “uphill” (from cold to hot), the performance metric is the coefficient of performance (COP), not efficiency. For a refrigerator:

COP_R = Q_C / W_in

For a heat pump (heating mode):

COP_HP = Q_H / W_in

The maximum possible COPs (Carnot limits) are:

COP_R,max = T_C / (T_H - T_C)

COP_HP,max = T_H / (T_H - T_C)

These equations explain a common maker surprise: a heat pump can deliver several units of heat per unit of electrical work (COP > 1) without violating physics, because it is moving heat rather than creating it from work. But the second law still limits how large the COP can be, and that limit gets worse as the temperature lift (T_H - T_C) increases.

Numerical Example: Why Deep Freezing Is Hard

Say you want a freezer at -20°C in a 25°C room. T_C = 253 K, T_H = 298 K, so T_H - T_C = 45 K.

COP_R,max = 253/45 ≈ 5.62

That is the absolute best-case COP. Real systems might be 1.5–3 depending on size and design. If you instead want -80°C, T_C = 193 K, then:

COP_R,max = 193/(298-193) = 193/105 ≈ 1.84

The theoretical ceiling collapses as the temperature lift grows, which is why ultra-low freezers are power-hungry and use multi-stage cycles.

Step-by-Step: Estimating Best-Case COP for a Cooling Build

• Step 1: Decide the cold-side temperature you truly need at the cooled object (not just the evaporator coil temperature; include thermal resistances).
• Step 2: Estimate the hot-side temperature at the condenser/radiator surface during operation (often above ambient due to finite heat exchanger size).
• Step 3: Convert to kelvin.
• Step 4: Compute COP_R,max or COP_HP,max.
• Step 5: Use the result to sanity-check power supply sizing: W_in ≥ Q_C / COP_R,max for cooling, or W_in ≥ Q_H / COP_HP,max for heating.

Entropy Generation Explains “Where the Efficiency Went”

When a real device underperforms, the second law points to entropy generation sources. A useful design mindset is: reduce S_gen by reducing irreversibilities, and you move closer to the Carnot ceiling.

Finite Temperature Difference in Heat Exchangers

Heat exchangers are everywhere: radiators, condensers, evaporators, intercoolers, CPU coolers, water blocks, recuperators. Heat transfer across a finite temperature difference generates entropy. If you force heat through a small exchanger, you need a large temperature difference to get the required heat flow, which increases entropy generation and reduces achievable efficiency or COP.

Practical implication: bigger heat exchangers (more area, better airflow, better contact) can improve performance not just by “better cooling,” but by reducing the temperature lift your cycle must operate across.

Pressure Drop and Flow Losses

Fans, pumps, and compressors spend work to overcome pressure drops. Those drops are largely dissipative: they convert organized mechanical energy into disorganized internal energy. In a refrigeration loop, for example, extra pressure drop on the suction side forces the compressor to operate with a larger pressure ratio, increasing work input and reducing COP.

Practical implication: smooth tubing runs, appropriately sized lines, gentle bends, and low-restriction filters can have a measurable thermodynamic payoff.

Throttling vs Expansion Work

Many cooling systems use a throttling device (capillary tube, TXV, or orifice) to drop pressure. Throttling is simple and cheap, but it is highly irreversible. If you could replace throttling with an expander that extracts work while dropping pressure, you would reduce entropy generation and improve COP. In small systems, expanders are often not worth the complexity, but the second law explains why throttling-heavy cycles have a performance penalty.

Entropy as a Diagnostic Tool: Isentropic Benchmarks

In many machines, an “ideal” reference process is isentropic (constant entropy). For adiabatic devices (no heat transfer intended), the reversible limit is isentropic behavior. Real devices deviate due to irreversibilities, and that deviation is quantified with efficiencies.

Compressors and Turbines: Why “Isentropic Efficiency” Exists

For a compressor, the ideal benchmark is an isentropic compression from inlet state to the same outlet pressure. The real compressor requires more work because entropy increases. For a turbine, the real turbine produces less work than the isentropic ideal because entropy increases.

Even if you do not compute full property tables, the concept is practical: if your compressor runs hotter than expected for a given pressure ratio, or your turbine produces less shaft power than expected, entropy generation (losses) is the underlying reason.

Step-by-Step: Using an Isentropic Benchmark in a Build (Conceptual)

• Step 1: Measure or estimate inlet pressure and temperature, and outlet pressure.
• Step 2: Use a property model (or manufacturer map) to estimate the ideal isentropic outlet temperature for that pressure ratio.
• Step 3: Compare to the actual outlet temperature. A higher actual outlet temperature in compression indicates more entropy generation and higher work input.
• Step 4: Use the gap to prioritize fixes: reduce leakage, improve cooling between stages, reduce pressure drops, improve motor and drive efficiency, or operate closer to the compressor’s efficient region.

Entropy and “Available Work”: Why Some Heat Is Useless for Power

Not all energy is equally useful. A hot object contains internal energy, but the fraction you can convert into work depends on the temperature difference to the environment. As the object cools toward ambient, the remaining energy becomes less and less able to do work. Entropy is the bookkeeper that captures this degradation.

A practical way to think about it: if you have a waste heat stream only slightly above ambient, it may contain many watts of heat but very little potential to produce mechanical or electrical power. You might still use it effectively for space heating or preheating, but expecting significant shaft power from low-grade heat runs into the second law limit.

Example: Waste Heat Recovery Reality Check

Suppose you have a steady waste heat source at 60°C and ambient at 20°C. T_H = 333 K, T_C = 293 K.

η_max = 1 - 293/333 ≈ 0.120

Even a perfect engine could only convert 12% of that heat into work. Real small-scale converters would be far below that. This is why many waste-heat projects are better framed as heat re-use (water preheat, drying, warming enclosures) rather than electricity generation.

Design Levers That Respect the Second Law

The second law does not just say “no”; it tells you what to change if you want better performance. The levers below all reduce entropy generation or improve the temperature levels in your favor.

Raise the Effective Hot-Side Temperature (Safely)

• Improve heat transfer into the working fluid so the cycle can accept heat at a higher average temperature (better burner-to-boiler coupling, better hot-side exchanger design).
• Reduce hot-side thermal resistance so you do not need a huge temperature drop from source to working fluid.
• Use materials and geometries that tolerate higher temperatures without creep, oxidation, or seal failure.

Lower the Effective Cold-Side Temperature (or Improve Heat Rejection)

• Increase radiator/condenser area, airflow, or coolant flow to reduce the hot-side rejection temperature.
• Keep fins clean, avoid recirculating hot exhaust air into the intake, and ensure good thermal contact.
• In enclosures, manage airflow paths so the sink is truly “ambient,” not a warmed pocket of air.

Reduce Irreversibilities

• Mechanical: align shafts, choose appropriate bearings, reduce seal drag, use proper lubrication.
• Fluid: avoid sharp bends, undersized tubing, and unnecessary restrictions; reduce leaks and bypass flows.
• Thermal: use larger heat exchangers to reduce temperature differences; add insulation where it prevents unwanted heat leaks that force larger temperature lifts.
• Electrical: reduce resistive losses in wiring and motors; keep power electronics cool to reduce conduction losses.

Second-Law Sanity Checks for Common Maker Claims

“It Runs on Ambient Heat”

A device that claims to extract heat from a single ambient reservoir and turn it into continuous work is a perpetual motion machine of the second kind. The Kelvin–Planck statement forbids it. To get work from heat, you need a temperature difference: a hot source and a cold sink. If both are “ambient,” there is no driving gradient.

“100% Efficient Engine”

Any cyclic engine must reject heat. If a design claims all input heat becomes work with no heat rejection, it violates the second law. In practice, even if you could minimize losses, you still need a sink at T_C, and Carnot sets the ceiling below 100% unless T_C = 0 K, which is unattainable.

“A Refrigerator That Cools Without Power”

Cooling below ambient without work input requires another resource: a colder reservoir (ice, night sky radiative cooling under certain conditions), stored chemical potential (evaporation of water, phase-change materials), or stored pressure (compressed gas). If none exists, the Clausius statement forbids spontaneous heat flow from cold to hot.

Hands-On Entropy Thinking: A Simple Mixing Demonstration

Entropy is often introduced with abstract language, but you can build intuition with mixing and heat spreading. If you mix hot and cold water in an insulated container, the final temperature becomes uniform. You can easily get from “separate hot and cold” to “mixed warm,” but you cannot get back to “separate hot and cold” without doing work (for example, running a heat pump). The spontaneous direction corresponds to entropy generation.

In machines, the same idea appears as: pressure equalizes through leaks, temperature equalizes through unwanted conduction, and velocity profiles become turbulent and dissipate. Each of these is a “mixing” of an organized gradient into a more uniform state, and each costs you potential to do useful work.