Skip to content

Nanothermometry in Living Cells - Physical Limits, Conceptual and Material Challenges

FieldValue
JournalAdvanced Functional Materials
Published2026-03-23
AuthorsTaras Plakhotnik
DOI10.1002/adfm.202528655
Citations2

Standard heat-transport estimates place the temperature rise at 100 nm from a cell’s heat source near 0.1 mK, not 1 K.

The estimate assumes 0.1 nW of cell heating, water’s thermal conductivity of 0.6 W m⁻¹ K⁻¹, and all heat confined within 100 nm.

Claims of 1 to 10 K intracellular differences therefore conflict with standard heat transport.

The perspective argues that measurement artifacts may explain many such reports, but it stops short of rejecting genuine cellular temperature gradients.

For a steady spherical source, energy conservation fixes total heat power through every surrounding surface.

Fourier’s law then links that power to thermal conductivity and radial temperature change.

Unknown intracellular geometry therefore cannot remove the modeled order-of-magnitude mismatch.

Statistical mechanics constrains thermal conductivity through fluctuation-dissipation theory.

Most liquids have room-temperature conductivity between 0.1 and 0.3 W m⁻¹ K⁻¹.

Water is exceptional because its hydrogen-bond network extends the effective correlation length.

Simulations of lipid bilayers and proteins in water give 0.1 to 0.6 W m⁻¹ K⁻¹.

The 2D model changes heat-transfer rate by about four when string stiffness changes by more than four orders of magnitude.

That result argues against force-field uncertainty as an explanation for orders-of-magnitude gradients.

A 50 nm cytoplasmic sphere has temperature fluctuations near 6 × 10⁻³ K with a 5 ns correlation time in water.

The paper also limits heat production through fuel diffusion rather than assuming unrestricted intracellular oxidation.

For glucose, the model uses an enthalpy of 2.8 MJ mol⁻¹ and an extracellular concentration of 27.8 mol m⁻³.

For oleic acid, it uses 11.2 MJ mol⁻¹ and 5 mol m⁻³.

The model also uses 0.11 W m⁻¹ K⁻¹ conductivity and 3 × 10⁻¹⁰ m² s⁻¹ glucose diffusivity.

The resulting upper limit is 0.2 K for glucose and 0.15 K for oleic acid.

The author treats these values as upper limits because membrane transport can reduce the calculated heat by up to three orders of magnitude.

Measured glucose uptake averages about 3 × 10⁻¹⁷ mol s⁻¹ per cell at 28 mM extracellular glucose.

Separate MCF-7 consumption measurements average 1.4 × 10⁻¹⁶ mol s⁻¹ per cell.

The 400 to 900 mitochondria in one HeLa cell also make the single-region fuel estimate too large.

This corrected fuel budget limits modeled heating well below several reported gradients.

Reported measurements span several mechanisms

Section titled “Reported measurements span several mechanisms”

The perspective reviews prior fluorescence, microthermistor, and ODMR measurements rather than reporting a new intracellular experiment.

Measurement methodCell and conditionReported resultRelevance to the paper
MTY fluorescence and enzyme kineticsHuman skin fibroblasts and HEK293Mitochondria near 50°C; ΔT of about 10 to 15 KConflicts with heat-transport limits
Ratiometric NaYF₄:20% Yb³⁺, 2% Er³⁺ nanoparticlesHeLa cellsΔT of 2 to 4 K with Ca²⁺ shock; 3 to 7 K with FCCPOptical and microenvironment effects remain
50 × 50 μm² microthermistorCOS-7 cellsFluctuations of 1.5°C at 37°C; 0.2°C at 45°C; 0.4°C at 25°CSignal depends on cell state
Single NV nanodiamond on a micropipetteHEK cells; medium at 36°CRegional means of 37.5 ± 0.8°C, 36.5 ± 0.8°C, and 35.9 ± 1.2°CAverage mitochondrial excess was about 1.5°C
Implanted NV nanodiamond ODMRAdherent cancer cells with FCCPΔT of 1.7 KODMR is less sensitive to photonic artifacts

MTY reports fluorescence intensity, while UCNP thermometry compares 525 and 545 nm emissions.

The single-particle experiment used one pre-calibrated nanodiamond at seven locations near mitochondria, nucleus, and cytoplasm.

ODMR reads the temperature-dependent shift of the NV ground-state splitting.

Fit uncertainty was about ±1.5°C at 95% confidence.

Regional sample deviations were about 1.8 K for mitochondria and nuclei, and 2.5 K for cytoplasm.

One mitochondrion-near reading was 35.2 ± 0.9°C, below a cytoplasmic reading of 36.7 ± 0.9°C.

That negative difference complicates a simple artifact explanation.

Calibration against external temperature shifts the cell’s thermal boundary condition.

It does not verify that local pH, ions, proteins, or refractive index remain unchanged during stimulation.

Many optical thermometers change signal by only about 1% per kelvin.

Effects negligible over a broad calibration range can match a 1 to 3 K intracellular difference.

Refractive-index differences alter emission lifetimes and ratiometric line intensities without changing temperature.

Mitochondria have an index near 1.41, compared with 1.36 in surrounding cytoplasm.

The paper estimates a 0.25°C apparent rise for an index change from 1.39 to 1.43.

Cited experiments produced apparent changes of several kelvin, including about 50°C near a gold mirror.

ODMR is less exposed because its microwave wavelength is about 10 cm, far above micron-scale optical variations.

Rigid host particles can reduce dielectric shifts by limiting environmental access.

Focused light and electric fields can mimic heat

Section titled “Focused light and electric fields can mimic heat”

Focused illumination can move probes and solutes, so optical force and thermophoresis can alter the measured location.

The reported cellular Soret coefficient reaches 0.01 K⁻¹, corresponding to about 1% concentration change per kelvin.

Infrared heating can also redistribute intracellular components and raise local refractive index.

MIP-ODT resolves microsecond expansion, while fluorescent nanoprobes on second timescales can retain mass-displacement effects.

Chemical stimulation can cause osmotic swelling.

Extracellular glucose from 4 to 20 mM increases cell volume by about 12%.

Geometry and refractive-index changes can therefore strengthen optical bias.

Electric fields at charged probe surfaces can reach 1 MV cm⁻¹ and shift NV transitions.

A 1 MV cm⁻¹ axial field can mimic a 4.7 K temperature change.

Surface passivation with a dipeptide or an 18 nm silica shell suppresses fast spurious shifts.

The coating increases nanodiamond diameter by about 30 nm.

Surface termination can also change the NV⁻ population by about 15% at 360 kV cm⁻¹.

Diamond quantum sensors offer a cross-check

Section titled “Diamond quantum sensors offer a cross-check”

ODMR thermometry uses the temperature dependence of the NV ground-state zero-field splitting.

The splitting is about 2.87 GHz, and its thermal coefficient is about 74 kHz K⁻¹.

Absolute measurement therefore needs per-particle calibration because nanodiamond batches vary in size, shape, and response.

NV centers emit a narrow zero-phonon line near 638 nm, with a phonon sideband near 680 nm.

Silicon-vacancy and nickel-related centers offer distinct optical mechanisms for cross-checking within one nanocrystal.

Nickel-related centers have a zero-phonon line near 882 nm.

The paper presents this multi-center approach as a route toward internal consistency, not an established standard.

Typical intracellular nanodiamonds are about 100 nm or larger after functionalization.

They may act as heat sources when hotter than the cell average.

Any physical explanation must therefore include the crystal and its immediate surroundings.

Current fabrication uses CVD and HPHT routes, but particle morphology and thermal response remain broadly distributed.

Bottom-up HPHT growth also has difficulty producing a high SiV center concentration.

Current evidence does not establish a reference thermometer

Section titled “Current evidence does not establish a reference thermometer”

Neither the experiments nor the theory is internally consistent enough for a definitive conclusion.

The author rejects the idea that large gradients are easy consequences of unknown conductivity, but accepts that systematic errors and incomplete modeling remain.

New fundamental heat-transport mechanisms appear unlikely because biological matter should obey established physical principles.

An accepted intracellular reference thermometer does not exist.

Engineers should therefore separate raw signal response from a calibrated temperature claim.

Useful cross-checks include orthogonal readouts, per-particle calibration, and controlled variation of intracellular conditions.

Optical studies need controls for refractive index, osmolarity, probe motion, and illumination gradients.

NV studies need electric-field control, surface passivation, and explicit heat-source treatment for the nanodiamond.

Current probes do not yet establish whether reported 1 to 10 K gradients are thermal.

Any sensor selection should therefore require cross-validation before assigning a temperature to a subcellular region.

View Original Abstract

ABSTRACT Reports of large temperature inhomogeneities inside single cells have sparked intense debate over the past decade. Initial claims were challenged as inconsistent with basic heat transport, yet the number of such reports continues to grow. Here, we present a concise, physics‐led assessment of intracellular nanothermometry, grounded in experimental data from cellular biology and studies reporting intracellular temperature contrast, thermal diffusivity, and thermal conductivity. This perspective provides a stricter and more detailed treatment of physical models, including a microscopic approach to thermal conductivity and molecular dynamics simulations. We also review potential sources of biased measurements in cells, including photonic effects, osmotic changes, optical trapping by tightly focused beams (tweezer effect), and electric‐field shifts. Diamond color centers are highlighted as promising candidates for reference thermometers. Nitrogen‐vacancy (NV) centers enable optical readout through microwave‐driven resonance that is largely insensitive to optical inhomogeneities and allows in situ cross‐checks using multiple centers (NV, SiV, and others) embedded in the same nanocrystal. Their sensitivity to electric fields can be reduced by surface passivation, improving the reliability of intracellular temperature measurements. However, challenges in materials science, fabrication, and, most importantly, quantitative theoretical modeling remain to be addressed.