Critical Heat Flux

Critical heat flux is the local wall heat-flux threshold at which boiling heat transfer deteriorates sharply through departure from nucleate boiling or liquid-film dryout.

Version

v2 · 2026-09-06 · History

Domain-specific #

1590

Origin domain

engineering

Subdomain

boiling heat transfer

Aliases

CHF

Core Idea

Critical heat flux (CHF) is the local wall heat flux at the onset of a critical boiling transition: increasing thermal loading or changing the flow state reaches a point beyond which boiling no longer transfers heat from the solid surface to the liquid with its prior effectiveness. The heat-transfer coefficient falls substantially as liquid contact or replenishment at the wall becomes inadequate. In a heat-flux-controlled system, the same imposed \(q''\) must then be carried with a much larger wall-to-fluid temperature difference, so wall temperature can rise abruptly.^[1][2]

The identity is broader than the familiar pool-boiling picture of bubbles merging into a vapor blanket. Two major routes must be distinguished. Departure from nucleate boiling (DNB) occurs characteristically in subcooled or low-quality bubbly flow when vapor accumulation near the wall impedes liquid rewetting. Liquid-film dryout occurs characteristically at higher vapor quality in annular flow when the wall film becomes intermittently and then persistently depleted. Both can produce a critical boiling transition and a sharp loss of heat-transfer performance, but their hydrodynamics, predictors, and detection variables differ.^[1][3]

Heat flux is heat-transfer rate per unit heated area,

\[ q''=\frac{\dot Q}{A}, \]

usually reported in \(\mathrm{W\,m^{-2}}\). CHF is not simply any large \(q''\); it is the threshold value associated with the regime deterioration for a specified fluid, pressure, mass flux, quality or subcooling, geometry, orientation, gravity, surface, heating distribution, and transient history. It is therefore a contextual system limit rather than a universal material constant.

Nukiyama's boiling experiments established the nonmonotonic boiling curve with a maximum heat-flux point in nucleate boiling and a later minimum in film boiling.^[4] Modern CHF practice extends that threshold concept to forced-flow, channels, bundles, jets, microchannels, and other configurations. The recurring abstraction is the same: wall loading and two-phase transport approach a local liquid-supply/heat-removal limit, the boiling regime changes, and the former thermal response law no longer applies.

Structural Signature

The mandatory roles are:

1. Heated wall and boiling fluid. Energy crosses a solid–fluid interface while liquid undergoes vaporization.
2. Local heat flux \(q''\). The relevant loading is heat-transfer rate per unit wall area, not total power alone.
3. Precritical high-performance regime. Nucleate boiling or wetted-film evaporation provides relatively effective heat transfer.
4. Liquid-supply or wetting competition. Vapor generation, liquid replenishment, entrainment, deposition, film transport, and rewetting determine whether liquid contact can be sustained.
5. Critical condition \(q''_{\mathrm{CHF}}\). Under stated boundary conditions, a heat-flux value marks the onset of substantial boiling heat-transfer deterioration.
6. Postcritical regime. Vapor coverage, intermittent dry patches, persistent dryout, transition boiling, or film/mist cooling reduces the effective heat-transfer coefficient.
7. Observable consequence. Under imposed heat flux, wall temperature rises sharply; under imposed wall temperature, transferred heat flux can fall after the boiling-curve maximum.
8. Context and detection rule. Fluid properties, pressure, mass flux, quality/subcooling, geometry, surface state, gravity/orientation, heating profile, and the criterion used to identify transition accompany any defensible value.

The constitutive relation

\[ q''=h\,(T_w-T_f) \]

makes the operational danger visible. If \(h\) collapses while \(q''\) is maintained, \(T_w-T_f\) must increase. This algebra does not predict CHF; it explains the wall-temperature response once boiling deterioration changes \(h\).

The invariant is the regime-boundary role, not one mechanism or one correlation. In saturated pool boiling on a large horizontal surface, a hydrodynamic-instability model may be appropriate. In subcooled forced flow, DNB correlations use pressure, mass flux, subcooling/quality, geometry, and heating. In high-quality annular flow, film depletion and dryout–rewet statistics become central. All qualify only when the local heat-transfer deterioration boundary is preserved.

What It Is Not

CHF is not onset of nucleate boiling. ONB is the transition from single-phase convection to bubble nucleation and typically increases heat-transfer effectiveness; CHF is the later upper limit beyond which that effective regime deteriorates.

It is not the Leidenfrost point or minimum film-boiling heat flux. On a pool-boiling curve, CHF is the maximum at the end of nucleate boiling. The minimum heat-flux point occurs after transition boiling, where stable film boiling begins or rewetting becomes possible. Confusing the two reverses the relevant boundary.

It is not identical to DNB. DNB is one critical-transition mechanism/regime, especially in subcooled and low-quality conditions. High-quality annular-flow CHF may arise through liquid-film dryout rather than bubble coalescence at a nucleating wall.

It is not identical to dryout incipience. Time-resolved annular-flow measurements show intermittent dry patches can begin below the CHF criterion and be followed by rewetting; CHF may correspond to a sufficiently large dry fraction or maximum heat-transfer coefficient rather than the first microscopic rupture event.^[3]

It is not burnout as a damage event. “Burnout heat flux” is historical terminology, but exceeding CHF need not physically destroy the surface if protection acts or the excursion is brief. Burnout is a possible consequence, not the recognition criterion.

It is not critical power. Critical power is the channel, bundle, or device power at which boiling transition is predicted or observed. CHF is a local surface flux. Mapping between them requires geometry, axial/radial power shape, coolant conditions, and a thermal-hydraulic model.

It is not DNBR, CHFR, or critical power ratio. These are margins, generally predicted critical loading divided by actual loading. A ratio of one indicates approach to the modeled limit; the ratio is not the limit itself.

Finally, CHF is not a universal constant and not a single universal formula. The 2006 Groeneveld lookup table, for example, normalizes a large water-tube database over pressure, mass flux, and quality; its existence and conditioning demonstrate why configuration and applicability cannot be omitted.^[5]

Scope of Application

CHF is used in boiling heat transfer, two-phase flow, reactor thermal hydraulics, boilers and steam generators, refrigeration and heat pumps, thermal desalination, high-heat-flux electronics cooling, heat pipes, jet impingement, microchannels, fusion-facing cooling concepts, and cryogenic systems. Its role is both descriptive and design-critical: it marks where the governing heat-transfer regime changes and where local thermal margins may rapidly erode.

In pressurized-water reactor analysis, DNB-type CHF and departure-from-nucleate-boiling ratio are central thermal limits. In boiling-water and high-quality flow regimes, dryout and critical power are more characteristic. Regulatory analysis treats models predicting critical boiling transition as safety-limit inputs, not as interchangeable correlations.^[1]

In pool boiling, the CHF point is normally read as the maximum of a heat-flux versus wall-superheat curve. Nukiyama's curve remains the canonical experimental schema. Zuber's saturated pool-boiling hydrodynamic correlation is a canonical model,

\[ q''_{\mathrm{CHF}} =C\,h_{fg}\,\rho_v^{1/2} \left[\sigma g(\rho_l-\rho_v)\right]^{1/4}, \]

where \(h_{fg}\) is latent heat, \(\rho_l\) and \(\rho_v\) are liquid and vapor densities, \(\sigma\) is surface tension, \(g\) is gravitational acceleration, and \(C\) depends on the idealization and geometry.^[6] This formula is an example of one regime model, not the definition and not a substitute for validated configuration-specific methods.

In forced-flow boiling, CHF depends on pressure, mass flux, local equilibrium quality or subcooling, channel dimensions, heated length, power shape, flow direction, spacer effects, and transients. The Groeneveld lookup-table method is built from tens of thousands of water-tube data points and standardized to a reference geometry; applying it elsewhere requires documented correction and validity procedures.^[5][7]

The node does not prescribe a safety limit, licensing method, or component design. Those tasks require validated correlations, uncertainty treatment, applicable codes and standards, and expert review for the exact system.

Clarity

A clear CHF statement answers four questions: what is controlled, what is observed, which mechanism family applies, and under what conditions?

First, distinguish imposed heat flux from imposed wall temperature. In heat-flux-controlled boiling, crossing the transition can drive a rapid wall-temperature excursion. In wall-temperature-controlled experiments, the boiling curve can traverse transition boiling with heat flux decreasing as wall superheat rises. Calling CHF “the temperature where boiling fails” discards the defining flux and the control-mode dependence.

Second, distinguish local from system-level quantities. A heater may have total power \(\dot Q\), but nonuniform heating and flow conditions make local \(q''(z)\) and local quality decisive. A reported critical power without area and thermal-hydraulic mapping cannot be treated as a CHF value.

Third, identify DNB versus film dryout. Low-quality bubbly conditions, coalescence near the wall, and abrupt vapor blanketing support a DNB account. Annular vapor core, wall liquid film, droplets, disturbance waves, intermittent dry patches, and rewetting support a dryout account. The two may share a system consequence but do not license the same mechanistic explanation or correlation.

Fourth, attach the condition vector. A bare number such as “CHF = 1 MW/m²” is scientifically incomplete without fluid, saturation/subcooling state, pressure, mass flux, quality, geometry, orientation/gravity, surface condition, heater size, and detection rule. The diagnostic question is not merely “how much heat?” but “which boiling transition, where, and under which boundary conditions?”

Manages Complexity

Boiling combines phase change, interfacial motion, nucleation, two-phase flow regimes, surface chemistry, and conjugate heat transfer. CHF compresses this complex field into a decision boundary: below the relevant threshold, a selected precritical correlation and wetting picture remain usable; at and above it, the analyst must switch regime model and evaluate wall-temperature or integrity consequences.

This compression supports margin reasoning. If a validated predictor gives \(q''_{\mathrm{CHF,pred}}\) and the modeled local operating flux is \(q''_{\mathrm{op}}\), a simple diagnostic ratio is

\[ R=\frac{q''_{\mathrm{CHF,pred}}}{q''_{\mathrm{op}}}. \]

The ratio is useful only with its correlation, uncertainty, geometry, and acceptance rule. It does not turn an empirical prediction into a universal physical constant.

CHF also organizes experiments. Researchers vary one or more control variables, monitor wall temperature, heat flux, pressure, flow, quality, film thickness, or visualization, and define an onset criterion for deterioration. Pool-boiling curves, forced-flow correlations, large lookup tables, and time-resolved dryout experiments are different measurement strategies aimed at the same regime boundary.

The abstraction prevents an especially costly modeling error: extrapolating a high heat-transfer nucleate-boiling law past the point at which liquid-wall contact ceases to support it. By naming CHF, the model contains an explicit validity frontier rather than silently producing benign temperatures in a postcritical regime.

Abstract Reasoning

CHF reasoning begins with a local energy balance and a regime map. Determine the wall loading \(q''\), the thermodynamic state, and the two-phase flow regime. Select the relevant mechanism family and a validated prediction method. Compare operating conditions with the predicted boundary, propagate uncertainty, and determine the postcritical response under the actual control mode.

Three inferences recur.

Threshold inference. Near CHF, a small increase in input heat, quality, or another control variable can cause a disproportionate wall-temperature response because the effective \(h\) changes regimes. Linear extrapolation across the boundary is invalid.

Supply–demand inference. Vapor generation demands liquid replenishment at the wall. CHF is approached when replenishment and rewetting cannot keep sufficient pace. The concrete supply process differs: bulk liquid return around vapor structures in pool boiling, near-wall replenishment in DNB, or disturbance-wave/deposition transport for annular films.

Model-locality inference. A correlation's inputs reveal its assumed locality. Zuber's pool model emphasizes fluid properties, gravity, and interfacial instability. A flow-boiling table emphasizes pressure, mass flux, and quality for a reference tube. A surface-enhancement experiment may additionally depend on wettability, wicking, roughness, and heater scale. A result should not transfer when these role mappings fail.

The reasoning also runs backward diagnostically. A sudden wall-temperature excursion at nearly fixed \(q''\) suggests deterioration in \(h\), but CHF is not proven until sensor behavior, flow excursion, local dryout, contact resistance, and other causes are considered. CHF is an explanatory hypothesis with operational criteria, not a label for every overheating event.

Knowledge Transfer

The CHF structure transfers exactly across boiling configurations when the roles remain literal: heated wall, boiling fluid, local flux, efficient precritical regime, threatened liquid contact, critical transition, and postcritical deterioration. A nuclear fuel rod, refrigerant microchannel, jet-cooled chip, boiling test wire, and high-flux evaporator can all instantiate that skeleton while using different correlations.

Transfer requires translating the mechanism roles. In pool boiling, liquid returns primarily under buoyancy and interfacial hydrodynamics. In forced low-quality flow, bulk advection, subcooling, bubble crowding, and near-wall turbulence matter. In annular flow, entrainment, deposition, film evaporation, and disturbance waves control wall wetting. In microstructured surfaces, capillary wicking and nucleation-site changes can shift the threshold. The name transfers; the causal model does not automatically.

The structural relation to general Threshold reasoning is exact. There is an input/control state, a critical value, two qualitatively different response regimes, sharp sensitivity near the boundary, and a need for margin. That general reasoning can be imported from other engineering thresholds. The boiling-specific variables and mechanisms cannot.

Metaphorical use outside phase-change heat transfer is inappropriate. A network “overheating” or organization reaching “critical flux” does not instantiate CHF without physical wall heat flux and boiling deterioration. The portable residue should be expressed through Threshold, Regime Change, or Safety Margin rather than by extending CHF.

Examples

Saturated pool boiling. An electrically heated horizontal surface in a saturated liquid is driven through increasing heat flux. Isolated nucleation sites become vigorous nucleate boiling, and heat transfer rises efficiently. Near the boiling-curve maximum, vapor escape and liquid return become hydrodynamically constrained. Beyond the maximum, large dry regions or vapor coverage reduce contact and move the system into transition boiling. The maximum is CHF; the later minimum leading into stable film boiling is not.^[4][6]

Subcooled forced-flow DNB. Water flows over a heated surface while the bulk remains below saturation or has low vapor quality. Nucleate boiling occurs at the wall. With increasing local flux, near-wall bubbles crowd and coalesce, liquid access is impaired, and heat transfer rapidly decreases. NRC defines DNB by this rapid decrease caused by the insulating effect of a steam blanket on a fuel-rod surface.^[2] The CHF value must be predicted for the actual pressure, flow, geometry, and axial power distribution.

Annular-film dryout. At high quality, vapor occupies the channel core while a thin liquid film wets the wall and droplets travel in the core. Film evaporation and entrainment compete with deposition and disturbance-wave rewetting. Morse and colleagues observed intermittent dryout below CHF and associated their CHF criterion with a maximum heat-transfer coefficient at a small time-averaged dry fraction in their R245fa rectangular-channel experiment.^[3] The example proves why first dry-patch appearance, CHF, and complete dryout cannot be collapsed.

Water-tube lookup prediction. The 2006 Groeneveld table predicts CHF for a normalized vertical 8 mm water-cooled tube as a function of pressure, mass flux, and quality, using a database exceeding 30,000 points in the published method.^[5] It is a powerful implementation of CHF, but the tabulation and correction procedure are not CHF itself.

Non-example. A single-phase heater reaches a material-temperature limit with no boiling or liquid-to-vapor interfacial transition. That may be a thermal limit or burnout condition, but it is not critical heat flux in the boiling-heat-transfer sense.

Structural Tensions

Universal threshold versus mechanism plurality. CHF is one recurring performance boundary, yet DNB, pool-boiling hydrodynamic instability, macrolayer depletion, contact-line effects, and annular-film dryout offer different mechanism descriptions. The stable abstraction must preserve the boundary while refusing to imply one mechanism fits all conditions.

Local phenomenon versus system safety metric. CHF occurs locally, while operators and designers often monitor total power, critical power ratio, or minimum DNBR. System quantities are useful only through a model that locates the limiting surface and maps total loading to local conditions.

Sharp operational event versus intermittent precursor. Engineering correlations often assign one onset point, but high-speed dryout measurements reveal repeated dryout and rewet cycles before the selected CHF criterion. The threshold is operationally sharp enough for decision-making while the microscale transition can be distributed in time and space.

Physical limit versus empirical predictor. A CHF event is physical, but predicted CHF commonly comes from correlations or lookup tables fitted within bounded databases. Greater numerical precision does not remove model-form, scaling, surface, geometry, or transient uncertainty.

Enhancement versus stability. Surface texturing, wettability, wicking, increased area, or altered flow can raise measured CHF, but an enhancement that works for one fluid and orientation can change nucleation, pressure drop, fouling, or aging elsewhere. Optimizing the threshold requires preserving the full system boundary rather than maximizing a laboratory number alone.

Heat-flux control versus temperature control. The same boiling curve produces a dangerous temperature jump under imposed flux and a falling heat-transfer branch under imposed temperature. Descriptions that omit the control mode can make contradictory claims about what “happens after CHF.”

Structural–Framed Character

CHF is predominantly structural. Heat flux, wall temperature, phase state, wetting, flow regime, and heat-transfer deterioration are physical observables. The regime transition exists independently of a regulatory convention, and compatible experiments can detect it through specified criteria.

The framed component enters through nomenclature and operational definition. Literature uses CHF, DNB, dryout, boiling crisis, critical boiling transition, burnout, and critical power with overlapping but nonidentical scope. Experiments may identify transition by a temperature excursion, maximum heat-transfer coefficient, dry fraction, or other rule. Regulators and design codes choose models, uncertainty allowances, and required margins. Those choices frame how the threshold is reported and acted upon without inventing the underlying boiling deterioration.

Structural Core vs. Domain Accent

The structural core is a threshold separating a high-performance transport regime from a degraded one: rising load, finite replenishment capacity, loss of contact, abrupt coefficient change, and amplified state response. This skeleton resembles many capacity limits and regime transitions.

The domain accent is indispensable. The transported quantity is heat per unit area; the interface is a heated wall and boiling fluid; the competing phases are liquid and vapor; the replenishment process is wetting or film supply; and the consequences depend on boiling hydrodynamics and heater control mode. Removing those roles turns CHF into generic Threshold or Capacity Saturation and loses the candidate.

The candidate is therefore domain-specific, not a new prime. Its transferable skeleton already exists in prime:threshold, while its value lies in preserving the exact boiling-mechanism family, observables, correlation boundaries, and safety use.

Instantiates / Related Primes

Threshold is the direct parent. CHF is literally a critical input value separating efficient boiling heat transfer from a degraded response regime, with disproportionate sensitivity near the boundary.

Nucleation is related because precritical nucleate boiling depends on bubble formation and nucleation sites, but CHF is not a nucleation threshold. In many cases nucleation is already vigorous before CHF; the failure is liquid replenishment or wetting.

Criticality is a lexical and conceptual neighbor, but statistical-mechanical criticality involves scale-free correlations and critical exponents. CHF does not require those commitments.

Thermodynamic Equilibrium supplies saturation properties and phase-state reference conditions, but CHF is a nonequilibrium transport limit. Cascade can describe downstream damage after wall overheating, but damage is not constitutive. Environmental Coupling Strength is a broad transport neighbor, not coverage.

Relationships to Other Abstractions

Current abstraction Critical Heat Flux Domain-specific

Parents (1) — more general patterns this builds on

• Critical Heat Flux is a kind of Threshold Prime

  Threshold is the direct parent.

Hierarchy path (1) — routes to 1 parentless root

• Critical Heat Flux → Threshold

Neighborhood in Abstraction Space

Critical Heat Flux sits in a sparse region of the domain-specific corpus (93^rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

• Adiabatic Process — 0.79
• Distillation — 0.78
• Evaporation — 0.78
• Van der Waals Equation — 0.77
• Pinch analysis — 0.76

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

• Onset of nucleate boiling: beginning of bubble nucleation, normally below CHF.
• Departure from nucleate boiling: one CHF/critical-transition mode, not the full genus.
• Dryout incipience: first local film rupture can precede the CHF criterion.
• Complete dryout: later inability to rewet; not necessarily the selected onset point.
• Transition boiling: post-CHF unstable regime, not the threshold value itself.
• Minimum film-boiling heat flux / Leidenfrost point: the lower turning point after transition boiling.
• Burnout: possible damage consequence and historical usage, not required for CHF.
• Critical power: total system or channel power at transition, requiring mapping to local flux.
• DNBR, CHFR, or CPR: modeled ratios or margins, not the critical flux.
• Heat-transfer coefficient: \(h\) deteriorates at transition but is not a flux.
• Criticality: a separate prime concerning scale-free poised regimes.
• Nucleation: bubble formation mechanism that can operate far below CHF.

References

[1] Kaizer, J. S., et al. Credibility Assessment Framework for Critical Boiling Transition Models: A Generic Safety Case to Determine the Credibility of Critical Heat Flux and Critical Power Models. NUREG/KM-0013, Draft for Comment, U.S. Nuclear Regulatory Commission, 2019. registry ↩a ↩b ↩c

[2] U.S. Nuclear Regulatory Commission. “Departure from Nucleate Boiling (DNB)”, NRC Glossary. registry ↩a ↩b

[3] Morse, Roman W., et al. “Critical Heat Flux and the Dryout of Liquid Film in Vertical Two-Phase Annular Flow”. International Journal of Heat and Mass Transfer 177 (2021), 121487. DOI: 10.1016/j.ijheatmasstransfer.2021.121487. registry ↩a ↩b ↩c

[4] Nukiyama, Shiro. “The Maximum and Minimum Values of the Heat Q Transmitted from Metal to Boiling Water under Atmospheric Pressure”. Journal of the Society of Mechanical Engineers 37(206), 367–374 (1934). DOI: 10.1299/jsmemagazine.37.206_367. registry ↩a ↩b

[5] Groeneveld, D. C., et al. “The 2006 CHF Look-Up Table”. Nuclear Engineering and Design 237 (2007), 1909–1922. DOI: 10.1016/j.nucengdes.2007.02.014. registry ↩a ↩b ↩c

[6] Zuber, Novak. Hydrodynamic Aspects of Boiling Heat Transfer. U.S. Atomic Energy Commission report AECU-4439 / UCLA doctoral dissertation (1959). DOI: 10.2172/4175511. registry ↩a ↩b

[7] Groeneveld, D. C. Critical Heat Flux Data Used to Generate the 2006 Groeneveld Critical Heat Flux Lookup Tables. NUREG/KM-0011, U.S. Nuclear Regulatory Commission, 2019. registry ↩