How Science Works | Thermodynamics — Heat, Work, Energy, Entropy and the Limits of Conversion

1. Why energy conservation is only the beginning

Imagine a proposal for a cyclic machine that absorbs heat from one reservoir and converts all of it into work, leaving no other change. Its energy account can be written consistently: the heat entering equals the work leaving. Yet energy conservation alone does not establish that the process is physically possible. Thermodynamics contains a further restriction on the direction and quality of energy conversion.

That distinction is one of the subject’s most useful intellectual habits. Before debating a mechanism in detail, ask whether it violates a general constraint. Then, after it passes that test, ask how it could actually proceed. A process may be permitted without being fast, practical or efficient. The relationship between the first and second laws is developed in Feynman’s Laws of Thermodynamics.

2. Draw the boundary before following the energy

A thermodynamic system is the region or material selected for analysis. A closed system exchanges energy but not matter with its surroundings. An open system can exchange both. An isolated system exchanges neither in the ideal description. These definitions determine which terms belong in an energy balance; they are not decorative labels added after the calculation.

A kettle can be analysed as water alone, water plus container, or the appliance together with part of the surrounding room. Each boundary produces a different account. Energy heating the container is an outward transfer from a water-only perspective but internal redistribution in a water-plus-container model. Scientific disagreement sometimes disappears when the two explanations are found to be counting different systems. See the system-based treatment in MIT’s thermodynamics notes.

3. State variables describe the present; paths describe the change

Temperature, pressure, volume and composition can describe a macroscopic equilibrium state. Internal energy and entropy are state functions: their changes between specified equilibrium states do not depend on which path joined those states. Heat and work are different. They describe energy crossing a boundary during a process and depend on how that process occurs.

Two routes can therefore take a system from the same initial state to the same final state while transferring different amounts of heat and work. The change in internal energy remains the same, but the division between transfer channels changes. This is why “How much heat is inside the object?” is not the standard thermodynamic question. The object has internal energy; heat describes transfer. See the first law of thermodynamics.

4. Temperature is not a measure of total stored energy

Temperature characterises thermal state and determines the direction of spontaneous heat transfer between systems in thermal contact. It is not interchangeable with internal energy. A small hot object and a large cooler object can contain very different total amounts of internal energy because quantity of material, composition and microscopic degrees of freedom matter.

A thermometer reaches a state related to the system being measured. Its calibration, thermal contact and response time affect the reading. A temperature number without a measurement location can also be misleading in a nonuniform system: the surface and interior need not share the same temperature. The role of thermal equilibrium and temperature is explained in The Laws of Thermodynamics.

5. Heat and work name different transfer channels

Heat is energy transfer driven by a temperature difference. Work includes other organised energy-transfer mechanisms, such as a moving boundary against a force or electrical work supplied to a system. The classification follows the chosen boundary and physical interaction, not simply the eventual appearance of a temperature rise.

For example, an electrically powered heater receives energy through an electrical-work channel before transferring energy thermally to nearby material. Calling every part of that process “heat entering” can obscure the mechanism. The energy may end up in similar internal degrees of freedom, but the transfer histories differ. The distinction is set out in OpenStax’s first-law treatment.

6. Choose a work sign convention and keep it

In this article, for a closed system with negligible changes in macroscopic kinetic and gravitational potential energy, use ΔU = Q − W. Here Q is positive for heat entering the system, and W is positive for work done by the system on its surroundings. Compression work done on the system therefore contributes with the opposite sign to expansion work done by it.

Other texts use an equally valid convention in which work done on the system is positive. The danger is mixing conventions halfway through an argument. The law has not changed because a symbol’s sign has changed; the definitions have. Before comparing answers, write the convention beside the equation. The convention used here matches OpenStax’s first-law presentation.

7. Heat capacity connects energy transfer to temperature change

When a material remains in one phase and its specific heat capacity is adequately approximated as constant over the interval, the thermal-energy change can be estimated using Q = mcΔT under the relevant conditions. The mass, material and temperature interval all matter. Equal energy inputs do not guarantee equal temperature rises in different samples.

For gases, heat capacity at constant pressure generally differs from heat capacity at constant volume because expansion can carry energy out as work. Near phase changes or across large temperature ranges, a single constant c can also be inadequate. The formula is valuable because its conditions are visible, not because it applies without qualification. See specific heat and calorimetry.

8. Worked example: why ideal heating time is not measured heating time

Constructed teaching example. Assume 0.50 kg of water warms by 20 K, with specific heat capacity approximated as 4,180 J/(kg K). Ignore the container, evaporation and temperature dependence of heat capacity. The required energy increase is 0.50 × 4,180 × 20 = 41,800 J, or 41.8 kJ.

If energy reaches the water at a constant 1,000 W, the ideal time is 41,800 ÷ 1,000 = 41.8 s. If a separate hypothetical model assumes only 80% of that input rate reaches the water during the interval, the effective rate is 800 W and the predicted time becomes 52.25 s.

Neither time is a measured performance claim for a real appliance. The 80% factor is an assumed teaching input, not a typical efficiency estimate. The calculation shows how an omitted transfer channel changes the prediction. A real experiment would need to include the container, changing heat loss and the actual power history. The model is based on calorimetric energy accounting.

9. Phase changes can absorb energy without the expected temperature rise

During a phase transition under suitable fixed-pressure conditions, energy can change the phase composition rather than simply raising temperature. Melting or vaporisation involves different molecular arrangements and interactions. A constant-temperature interval during heating is therefore not evidence that energy has stopped entering the system.

The correct model must track phase fraction as well as temperature. Applying mcΔT alone across melting or boiling can omit a major part of the energy requirement. Mixtures, changing pressure and nonequilibrium conditions introduce further qualifications, so the textbook temperature plateau is a useful idealisation rather than a universal shape for every heating curve. The necessary distinction is described in heat transfer and calorimetry.

10. Processes with similar names can have different energy accounts

An isothermal process keeps temperature constant. An adiabatic process has no heat transfer across the chosen boundary. An isochoric process keeps volume constant. An isobaric process keeps pressure constant. These conditions are not interchangeable. An expanding ideal gas can remain at constant temperature while receiving heat, or it can cool during an adiabatic expansion while doing work.

Likewise, a slow process is not automatically reversible. Friction, mixing or a finite temperature difference can generate irreversibility even when changes occur gradually. The label must describe what actually remains fixed or absent, not merely how the apparatus looks. See thermodynamic processes.

11. Expansion work depends on the route

When a gas pushes a moving boundary, the work depends on the opposing pressure and change in volume. For an appropriate quasistatic description, the work can be represented by the area under a pressure–volume curve. Different pressure histories between the same starting and ending volumes can therefore produce different amounts of work.

This provides a graphical way to understand path dependence. A final volume does not reveal how much work was performed unless the process is also specified. Nor should the gas’s equilibrium pressure be inserted blindly into a strongly nonequilibrium expansion. The pressure–volume representation is useful only when the corresponding states and boundary forces are meaningful. The distinctions are developed in OpenStax’s process analysis.

12. Entropy is a state quantity, not a synonym for mess

Entropy can be introduced thermodynamically through reversible heat transfer divided by temperature, and statistically through the accessible microscopic states compatible with a macroscopic description. “Disorder” is a limited analogy because ordinary visual untidiness is not a reliable thermodynamic measurement. A neatly arranged hot object and a visually messy cooler object cannot be compared by appearance alone.

For a real irreversible process, the entropy change between equilibrium endpoints can be evaluated using a hypothetical reversible path joining those endpoints. This does not claim that the actual process was reversible. It uses the fact that entropy is a state function. The distinction is important in the free-expansion example below. See entropy and its calculation.

13. The second law applies to the complete account

For an isolated system, total entropy does not decrease in a spontaneous macroscopic process. A part of that system can become more ordered or have lower entropy while exporting entropy to its surroundings. The correct test therefore concerns the combined account, not a selected local region that makes the desired story look simpler.

A refrigerator can lower the entropy of material inside its cold compartment while releasing more entropy through its warmer surroundings and energy supply. A growing organism is an open system exchanging matter and energy; it is not an isolated exception to thermodynamics. The rule is not that every local structure must become less organised every moment. The governing distinction is discussed in The Laws of Thermodynamics.

14. Reversibility is a limiting comparison, not normal operation

A reversible process is an ideal limit in which the system and surroundings can be restored without leaving other changes. Real friction, finite-gradient heat transfer, uncontrolled expansion and mixing generally produce entropy. Reversible models are useful because they establish bounds against which actual processes can be compared.

The ideal should not be mistaken for a machine that can deliver arbitrary useful power while retaining every reversible assumption. Approaching a reversible heat-transfer step typically requires very small temperature differences and correspondingly careful control. Thermodynamics identifies the upper bound; engineering must also confront rate, size, cost and material constraints. The ideal comparison is developed through the Carnot cycle.

15. Heat engines require both an input and a return path

A cyclic heat engine receives heat from a hotter reservoir, produces net work and rejects heat to a colder reservoir. Because the working system returns to its initial state after a complete cycle, its net change in internal energy is zero. Energy conservation then gives W = QH − QC when the heat magnitudes are defined as positive quantities.

The thermal efficiency is W/QH. For an engine operating only between two reservoirs at absolute temperatures TH and TC, the reversible upper bound is 1 − TC/TH. Temperatures must be in kelvin. Substituting Celsius values into that ratio changes the physical meaning and can produce nonsense. See Carnot efficiency.

16. Worked example: an engine that passes both laws

Constructed teaching example. A hypothetical cyclic engine operates between ideal reservoirs at 600 K and 300 K. During one cycle it absorbs 1,200 J from the hot reservoir and produces 450 J of work. The first law requires it to reject 750 J to the cold reservoir.

Its efficiency is 450 ÷ 1,200 = 37.5%. The Carnot limit for these two temperatures is 1 − 300/600 = 50%. The proposed efficiency is below that bound. Passing this comparison does not prove that a particular mechanism can deliver it, but it avoids this obvious second-law violation.

The hot reservoir’s entropy change is −1,200/600 = −2.0 J/K. The cold reservoir gains 750/300 = 2.5 J/K. The working system returns to its original state, so its entropy change over the cycle is zero. The total increase is therefore 0.50 J/K. This original example shows how energy conservation, efficiency and entropy provide complementary checks using the two-reservoir engine model and entropy accounting.

17. Refrigerators move heat rather than create cold

A refrigerator uses work input to transfer heat from a colder region to a warmer one. Over a cycle, the heat rejected to the warmer region equals the heat removed from the colder region plus the work supplied. Describing the device as a producer of “cold” can hide this energy balance and the essential warm-side output.

The coefficient of performance, or COP, compares the desired heat transfer with the work input. For a refrigerator it is QC/W; for a heat pump delivering warmth it is QH/W. COP can exceed one because the device is moving existing thermal energy in addition to receiving work. It is not an efficiency greater than 100% in the sense of creating more energy than enters. See refrigerators and heat pumps.

18. Worked example: why a COP of three is not a violation

Constructed teaching example. A hypothetical refrigerator removes 2.4 kJ from a cold compartment while receiving 0.8 kJ of work. It must reject 3.2 kJ to its warm surroundings. Its refrigeration COP is 2.4 ÷ 0.8 = 3.

The apparently larger useful transfer is possible because the 2.4 kJ was already present in the cold region. The work enables its transfer; it does not have to supply every joule rejected on the warm side. If both sides of this ideal device were placed inside the same isolated room and all exchanges stayed there, the room’s net energy gain from the external power supply would be the work input.

This example illustrates a boundary test. A calculation can seem paradoxical when only the cold compartment is counted, then become straightforward when the warm side and power input are included. It applies the energy relationships in the refrigerator model, not a performance claim for any appliance.

19. Worked example: entropy can rise when no heat enters

Idealised thought experiment. Consider one mole of ideal gas expanding freely into an evacuated part of an insulated rigid enclosure until its accessible volume doubles. In this model there is no heat transfer across the enclosure and no external boundary work. The first law gives ΔU = 0. For an ideal gas, internal energy depends on temperature, so the initial and final equilibrium temperatures are the same.

It would be wrong to conclude that entropy is unchanged merely because Q = 0 for the actual process. The expansion is irreversible. To calculate the state-function change, use an imagined reversible isothermal path between the same endpoints: ΔS = nR ln(V2/V1) = R ln 2 ≈ 5.76 J/K.

The reversible comparison path is a calculation device; it is not the history the gas actually followed. The example distinguishes heat transfer from entropy production and illustrates why δQ/T must not be used indiscriminately along an irreversible path. The conceptual basis is covered in OpenStax’s entropy section.

20. Free energy gives a condition-specific direction test

Different thermodynamic potentials are useful under different constraints. Gibbs free energy is particularly useful for systems considered at constant temperature and pressure under appropriate work conditions. Helmholtz free energy is useful under constant temperature and volume. Neither is a new kind of conserved energy that replaces the first law.

A negative free-energy change can indicate a permitted direction under the specified conditions, but it does not determine how quickly a process proceeds. A large activation barrier can make an energetically favourable transformation extremely slow. The distinction connects this physics branch to the existing Physical Chemistry guide; the thermodynamic framework is developed in MIT’s thermodynamics material.

21. Open systems carry energy with matter

In an open system, incoming and outgoing streams carry energy as well as mass. Flow energy, internal energy, kinetic energy and gravitational potential energy can all matter, depending on the device and chosen boundary. Enthalpy is useful because it combines internal energy with the pressure–volume contribution associated with flowing material.

A turbine, compressor, heat exchanger or living organism cannot be analysed reliably by forgetting the material streams. “The temperature stayed constant” does not imply that no energy passed through. A steady operating state can have large continuous throughputs even while stored quantities remain nearly unchanged. The distinction between stored state and continuing flow is developed in MIT’s system and control-volume treatment.

22. Heat-transfer rate is a separate question

Thermodynamics can constrain total energy changes without specifying how quickly heat moves. Rate depends on mechanisms such as conduction, convection and radiation, together with geometry, materials, temperature differences and surrounding motion. Two systems can have the same thermodynamic endpoints but reach them over very different times.

This is why a heating-time calculation needs both an energy requirement and a power-transfer model. It also explains why better insulation changes the time history without altering the first law. The balance tells us where energy can go; the transport model tells us how rapidly it goes there. The calorimetric treatment in OpenStax provides the energy foundation; the rate model is an additional layer.

23. Useful energy depends on the surroundings

Energy is conserved, but its capacity to deliver a chosen form of useful work can decline. A temperature difference can support an engine; once the relevant system and surroundings reach thermal equilibrium, that particular opportunity is gone. The energy has not disappeared. The difference in state that made the conversion possible has been reduced.

The concept of exergy formalises useful-work potential relative to a specified environment and permitted processes. This dependence on a reference environment is essential: useful-work potential is not simply an intrinsic label attached to a quantity of energy. It is a relationship between system and surroundings. The link to reversible limits can be developed from the Carnot comparison and the broader MIT thermodynamics notes.

24. A good calorimetry experiment audits the missing terms

If measured heating takes longer than an ideal calculation predicts, several explanations are possible: some energy heats the container, some leaves to the environment, the actual power differs from the assumed value, or the temperature probe does not represent the whole sample. The discrepancy alone does not identify which explanation is correct.

A useful test records the relevant masses, locations, calibration and time histories, then varies a condition that separates the competing models. Repeating one measurement can quantify random scatter, but a consistent omitted container heat capacity will remain a consistent bias. The measurement discipline is set out in NIST Technical Note 1297, while the physical energy model is in calorimetry.

25. What would count against the proposed explanation?

For the hypothetical engine, measurements of all heat and work transfers should satisfy the energy balance within uncertainty. If they do not, first inspect the system boundary, unmeasured storage changes and instrumentation. An apparent excess output is not automatically a refutation of conservation; it can be evidence that an input or transient store was missed.

At the same time, scientific discipline does not mean dismissing every discrepancy. Persistent, independently reproduced differences should be investigated with better controls and alternative methods. The explanation must state in advance which observations it predicts and which would require revision. This is the application of energy conservation together with transparent uncertainty reporting, not a licence to treat an unexplained measurement as established new physics.

26. The statistical view connects the two scales

Macroscopic thermodynamics does not normally track every molecule. Statistical mechanics explains how collective molecular states support quantities such as temperature, entropy and heat capacity. This is a bridge between microscopic dynamics and macroscopic regularities, not a claim that each individual molecule has the temperature of the whole sample in the ordinary thermodynamic sense.

The bridge also clarifies why a macroscopic direction can be reliable even when microscopic motion is complicated. Many molecular arrangements can correspond to one coarse-grained state, and the accessible populations of those arrangements matter. For this connection, use the entropy discussion in The Laws of Thermodynamics and the broader molecular treatment in Physical Chemistry.

27. The CivDJ reading lens: follow storage, transfer and return

For this series, the CivDJ lens is an editorial discipline: define the system, identify its state, track transfers, enforce constraints and inspect the return. It is not an additional thermodynamic law. The method is useful because it stops an explanation at the point where a missing boundary or uncounted output would otherwise disappear into fluent language.

In a refrigerator, the cold compartment is not the whole system. In an engine, rejected heat is not an irrelevant leftover. In a warming sample, the container is not automatically negligible. The complete route matters. The reader should be able to follow each quantity into, through and out of the chosen account without changing the definition midway.

28. Learning route: from warming objects to constrained cycles

First encounter. Distinguish an object’s temperature from the amount of material and the energy needed to change its thermal state. Describe what is warmer, what is cooler and which direction heat transfers. Use observations to challenge the idea that “hotter” and “contains more energy” always mean the same thing.

Secondary progression. Introduce heat capacity, phase change, energy conservation and power. Require the student to list the system boundary and neglected transfers before calculating. The water-heating example is valuable because it separates an ideal energy requirement from an assumed rate and an actual measurement.

Higher-resolution progression. Add pressure–volume work, state functions, entropy, cycles, Carnot limits, refrigeration COP and free energy. The goal is to recognise which constraint applies and which extra information is needed for a rate or mechanism. This is a proposed instructional route, not a statement of any specific school’s current syllabus.

29. Five checkpoints, with the boundary made explicit

Can temperature stay constant while energy enters? Yes. During an appropriate phase transition, or an isothermal process with work transfer, incoming energy need not produce a temperature rise. The state and process conditions determine the account.

Does adiabatic mean constant temperature? No. It means no heat transfer across the chosen boundary. Work can still change internal energy and temperature, as explained in thermodynamic processes.

Can local entropy decrease? Yes, provided the full system-and-surroundings account remains consistent with the second law. A selected local decrease is not sufficient evidence of a violation.

Does a COP greater than one imply energy creation? No. COP compares useful heat moved with work input. It is not the ratio of all energy outputs to all energy inputs. The warm-side output includes both the removed heat and the supplied work.

Does a favourable free-energy change guarantee an immediate reaction? No. Thermodynamics constrains direction and equilibrium; kinetics determines the accessible rates and barriers. Keeping these questions separate is central to understanding both physics and chemistry.

30. The final lesson: count everything that the claim depends on

Thermodynamics is powerful because it can reject impossible accounts before every microscopic detail is known. It asks the reader to distinguish a stock from a flow, a state from a path, a local observation from a complete system, and a permitted change from a fast one. Those distinctions travel far beyond engines and refrigerators.

A strong explanation states the boundary, temperatures, material conditions, sign convention and neglected terms. It then checks energy, direction and uncertainty separately. The subject becomes scientific literacy when a reader can ask not only “Where did the energy go?” but also “What changes elsewhere make this local result possible?”

Sources and deeper study

The numerical examples here are original hypothetical calculations, not measured device specifications. Primary teaching sources include The Feynman Lectures, Volume I, Chapter 44: The Laws of Thermodynamics; MIT’s thermodynamics notes; and the linked OpenStax University Physics Volume 2 sections on calorimetry, the first law, processes, refrigerators, the Carnot cycle and entropy. Use NIST Technical Note 1297 for measurement uncertainty and NIST Special Publication 330 for SI units.

Continue into Physical Chemistry for reaction free energy and molecular states, Atmospheric Science for heat, moisture and circulation, and Oceanography for heat storage and transport. The common return is How Science Works.