Skip to content

Nanoscale Size Effects and Surface Chemistry

CSIR NET nanoscience: dimensionality, surface-to-volume ratio, magic-number clusters, Laplace and Kelvin effects, melting-point depression, superparamagnetism.

  • 21 min read
  • Updated
On this page · 15 sections
  1. Prerequisites
  2. The Hook: The Same Gold, a Different Element
  3. What Counts as “Nano”
  4. Surface-to-Volume Ratio
  5. Curvature: Laplace, Kelvin and Ostwald–Freundlich
  6. Melting-Point Depression
  7. Magnetism at the Nanoscale: Superparamagnetism
  8. Summary of Size-Dependent Properties
  9. Memory Tricks
  10. Don’t Confuse
  11. Common Mistakes
  12. Practice
  13. Quick Revision Table
  14. Key Takeaways
  15. What’s Next

Prerequisites

The Hook: The Same Gold, a Different Element

What Counts as “Nano”

A nanomaterial has at least one external dimension, or an internal structure, in the range of roughly 1–100 nm. One nanometre is \( 10^{-9} \) m: about four gold atoms laid side by side, or the length of a few C–C bonds. The boundary at 100 nm is a convention used in definitions (the ISO and EU wordings both use it); it is not a physical threshold. The physics that matters is the comparison of the particle size with a characteristic length of the property being studied:

PropertyCharacteristic length it is compared with
Fraction of surface atomsatomic diameter (≈ 0.3 nm)
Electronic level spacing in semiconductorsexciton Bohr radius (a few nm)
Optical response of metalswavelength of light (hundreds of nm)
Ferromagnetic domain structuresingle-domain size (tens of nm)
Electron scattering in metalsmean free path (tens of nm)

A property becomes “nano” when the particle size falls below its characteristic length. That is why a 50 nm gold particle already shows a plasmon band, while its electronic structure is still that of bulk gold.

Classification by dimensionality

Nanostructures are classified by how many dimensions are not confined to the nanoscale, that is, how many directions the electrons can still move freely in.

graph LR
    A["3D: bulk
no confinement"] --> B["2D: nanosheet / quantum well
confined in 1 direction
graphene, MoS2 monolayer"] B --> C["1D: nanowire / nanotube
confined in 2 directions
carbon nanotube, Si nanowire"] C --> D["0D: nanoparticle / quantum dot
confined in 3 directions
CdSe dot, Au cluster, C60"]

Surface-to-Volume Ratio

Geometry

For a sphere of radius \( r \), area/volume \( = 4\pi r^2/(\tfrac{4}{3}\pi r^3) = 3/r \); for a cube of edge \( a \) it is \( 6/a \). Halve the size and the surface per unit volume doubles.

A useful exercise: cut a 1 cm cube into cubes of edge 1 nm. The number of pieces is \( (10^7)^3 = 10^{21} \), each with area \( 6 \times 10^{-14} \) cm², so the total area rises from 6 cm² to \( 6 \times 10^7 \) cm² = 6000 m², close to the area of a football pitch, from one sugar-cube volume of material.

Specific surface area

For non-porous spheres of diameter \( D \) and density \( \rho \),

$$\boxed{S = \frac{\text{area}}{\text{mass}} = \frac{\pi D^2}{\rho\,\pi D^3/6} = \frac{6}{\rho D}}$$
Gold spheres (\( \rho = 19.3 \) g cm⁻³)\( S \) / m² g⁻¹
\( D = 50 \) nm6.2
\( D = 5 \) nm62
\( D = 2 \) nm155

Watch the units: with \( \rho \) in g m⁻³ (\( 19.3 \times 10^6 \)) and \( D \) in m, \( S \) comes out in m² g⁻¹. Inverting the formula turns a BET surface area into an equivalent particle size; that is how a powder sold as “300 m² g⁻¹ silica” is read as roughly 9 nm primary particles (\( \rho = 2.2 \) g cm⁻³).

Counting surface atoms: magic-number clusters

Metal atoms pack around a central atom in closed shells. A cuboctahedral (fcc-like) or Mackay icosahedral cluster with \( n \) complete shells around one central atom contains

$$\boxed{N_{\text{total}} = \frac{10n^3 + 15n^2 + 11n + 3}{3}, \qquad N_{\text{surface}} = 10n^2 + 2}$$
Shells \( n \)Total atomsAtoms in outer shell% on the surface
1131292
2554276
31479263
430916252
556125245
7141549235

The totals 13, 55, 147, 309, 561 are the geometric magic numbers: clusters of exactly these sizes are unusually stable because every shell is complete. A 561-atom cluster of a close-packed metal is only about 3 nm across (eleven atomic diameters), yet nearly half of its atoms are at the surface. The 55-atom gold cluster stabilised by phosphine and chloride ligands (Schmid’s cluster) is the textbook example of a geometric magic number met in the laboratory.

A rough rule for larger particles: the fraction of atoms in the surface layer of a sphere of diameter \( D \) made of atoms of diameter \( d \) is about \( 6d/D \) when \( D \gg d \) (surface shell volume \( \pi D^2 d \) over total volume \( \pi D^3/6 \)). For \( d = 0.29 \) nm and \( D = 20 \) nm this gives about 9%; the exact shell count is a little lower because the estimate overcounts curvature.

Why surface atoms change chemistry

A surface atom has fewer neighbours than a bulk atom: coordination number 12 inside an fcc metal, 9 on a (111) face, 7 at an edge and fewer at a corner. Each missing bond raises the energy of the atom, so surface atoms are:

  • more reactive (adsorption and catalysis happen there; the corner and edge atoms that dominate very small particles are often the most active sites);
  • less tightly held (lower melting point, higher vapour pressure, higher solubility);
  • the source of surface energy \( \gamma \), the excess free energy per unit area that every size-dependent formula below contains.

This is why heterogeneous catalysts are made as nanoparticles on high-area supports: the same mass of platinum or gold exposes far more active sites. The kinetics of the catalytic step itself are treated on Heterogeneous Catalysis.

Curvature: Laplace, Kelvin and Ostwald–Freundlich

Laplace pressure

A curved interface of radius \( r \) and surface tension \( \gamma \) holds the inside at a higher pressure than the outside:

$$\Delta P = \frac{2\gamma}{r}$$

For a water droplet (\( \gamma = 0.072 \) N m⁻¹) of radius 10 nm, \( \Delta P = 1.44 \times 10^7 \) Pa, about 144 bar. Nanoscale curvature produces very large pressures.

Kelvin equation

The extra pressure raises the chemical potential of the material inside by \( V_m\Delta P = 2\gamma V_m/r \), where \( V_m \) is the molar volume. Equating chemical potentials with the vapour:

$$\boxed{\ln\frac{p(r)}{p_\infty} = \frac{2\gamma V_m}{rRT}}$$
Water droplet at 298.15 K (\( V_m = 1.8 \times 10^{-5} \) m³ mol⁻¹)\( p(r)/p_\infty \)
\( r = 10 \) nm1.11
\( r = 1 \) nm2.85

Small droplets evaporate more readily than large ones; a concave meniscus (negative \( r \)) has a lower vapour pressure, which is why vapour condenses in narrow pores below its saturation pressure (capillary condensation, the origin of hysteresis in type IV isotherms).

Ostwald–Freundlich equation and Ostwald ripening

The same argument for a small solid particle in contact with its saturated solution gives

$$\ln\frac{S(r)}{S_\infty} = \frac{2\gamma V_m}{rRT}$$

where \( S \) is the solubility. With \( \gamma = 0.2 \) J m⁻², \( V_m = 3.0 \times 10^{-5} \) m³ mol⁻¹ and \( T = 298.15 \) K (illustrative values), a particle of radius 2 nm is about 11 times more soluble than the bulk solid; at 20 nm the factor is only 1.27.

The consequence is Ostwald ripening: in a dispersion of particles of different sizes, the solution is supersaturated with respect to the large particles and undersaturated with respect to the small ones. Small particles dissolve, large ones grow, the mean size rises and the size distribution broadens with time. Ripening is a central problem (and sometimes a tool) in nanoparticle synthesis; its role in size control is taken up on Synthesis, Stabilisation and Characterisation.

Melting-Point Depression

A small solid particle melts at a lower temperature than the bulk because its surface atoms are less tightly bound and because the solid–liquid interface it must maintain costs energy. The Gibbs–Thomson relation, in its simplest form for a spherical solid particle of radius \( r \) in its own melt, is

$$\boxed{\frac{T_m(r)}{T_m(\infty)} = 1 - \frac{2\gamma_{sl}V_m}{\Delta_{\text{fus}}H\,r} = 1 - \frac{\beta}{r}}$$

where \( \gamma_{sl} \) is the solid–liquid interfacial energy and \( \Delta_{\text{fus}}H \) the molar enthalpy of fusion. Two features matter for the exam:

  1. The depression scales as \( 1/r \): plotting \( T_m \) against \( 1/r \) gives a straight line whose intercept is the bulk melting point.
  2. It is negligible above a few tens of nanometres and dramatic below about 5 nm.

Gold is the classic experimental case: electron-diffraction studies of supported gold particles showed the melting point falling steeply below a few nanometres. The precise values depend on the support and the model used, so learn the trend and the \( 1/r \) form rather than a specific number. The same lowering of cohesive energy explains why nanoparticle powders sinter at temperatures far below the bulk melting point, which is both useful (low-temperature conductive inks) and a nuisance (catalyst particles coarsen in use).

Magnetism at the Nanoscale: Superparamagnetism

A bulk ferromagnet such as iron or magnetite breaks up into domains to reduce its magnetostatic energy; the walls between domains cost energy. Below a critical size (typically tens of nanometres, material-dependent) a particle is cheaper to keep as a single domain: all its spins point together, and the particle behaves like one giant magnetic moment of perhaps thousands of Bohr magnetons.

The moment prefers to lie along an easy axis. Reversing it requires crossing an energy barrier \( KV \), where \( K \) is the magnetic anisotropy energy density and \( V \) the particle volume. The Néel relaxation time is

$$\boxed{\tau = \tau_0\exp\!\left(\frac{KV}{k_BT}\right)}, \qquad \tau_0 \approx 10^{-9}\ \text{s (typical)}$$
  • If \( \tau \) is short compared with the measurement time \( \tau_m \), the moment flips back and forth during the measurement, the time-averaged magnetisation in zero field is zero, and there is no hysteresis: zero remanence and zero coercivity. The particle is superparamagnetic. In a field it magnetises strongly (far more strongly than a paramagnet, because each “spin” is a giant moment) following a Langevin curve.
  • If \( \tau > \tau_m \), the moment appears frozen and the particle shows hysteresis: it is blocked.

The crossover defines the blocking temperature:

$$T_B = \frac{KV}{k_B\ln(\tau_m/\tau_0)}$$

For a DC magnetometer measurement (\( \tau_m \approx 100 \) s) and \( \tau_0 = 10^{-9} \) s, \( \ln(10^{11}) = 25.3 \), so \( T_B \approx KV/25k_B \).

The same particle can be superparamagnetic in one experiment and blocked in another: Mössbauer spectroscopy (measurement time of order \( 10^{-8} \) s) sees a frozen moment at temperatures where a magnetometer sees none. Superparamagnetic iron oxide nanoparticles are used as MRI contrast agents and for magnetic separation precisely because they magnetise strongly in a field and do not stay magnetised (and so do not clump) when the field is removed.

graph TD
    A["Bulk ferromagnet
multi-domain, hysteresis"] -->|"size below single-domain limit"| B["Single-domain particle
maximum coercivity"] B -->|"size reduced further,
KV comparable to kT"| C["Superparamagnet
no remanence, no coercivity
above T_B"]

Summary of Size-Dependent Properties

PropertyTrend as size decreasesPhysical reason
Fraction of surface atomsincreases (\( \propto 1/D \))geometry
Specific surface areaincreases (\( 6/\rho D \))geometry
Melting pointdecreases (\( \propto 1/r \))lower cohesive energy, interface cost
Vapour pressure, solubilityincrease (exponentially in \( 1/r \))Kelvin / Ostwald–Freundlich
Catalytic activity per gramusually increasesmore, and lower-coordinated, surface sites
Coercivity of a ferromagnetrises, then falls to zerosingle domain, then superparamagnetism
Band gap of a semiconductorincreasesquantum confinement (next page)
Metal plasmon colourchanges with size and shapesurface plasmon resonance (next page)

Memory Tricks

Don’t Confuse

ThisvsThat
Geometric magic numbers 13, 55, 147, 309 (atoms, closed shells of a close-packed cluster)Electronic magic numbers 2, 8, 20, 40, 58, 92 (valence electrons, jellium shells)
Kelvin equation: vapour pressure over a curved liquid surfaceOstwald–Freundlich: solubility of a small solid particle
Ostwald ripening: growth of large particles at the expense of small ones through solutionAggregation/coagulation: particles stick together on contact
Superparamagnet: giant moment, strong response, no hysteresis above \( T_B \)Paramagnet: independent atomic moments, weak response
Blocking temperature \( T_B \) (depends on measurement time)Curie temperature \( T_C \) (loss of ferromagnetic order itself)
Specific surface area \( 6/\rho D \) from diameterSurface-to-volume ratio \( 3/r \) from radius

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper.

Level 1: Foundation

  1. How many atoms are in the outer shell of a 147-atom cuboctahedral cluster, and what percentage of atoms is that?
  2. Write the Kelvin equation and state whether a concave meniscus has a higher or lower vapour pressure than a flat surface.
  3. What two features of a magnetisation curve identify a superparamagnetic sample above its blocking temperature?
Answers
  1. \( 10n^2 + 2 \) with \( n = 3 \): 92 atoms, 92/147 = 63%.
  2. \( \ln(p/p_\infty) = 2\gamma V_m/(rRT) \). For a concave meniscus \( r \) is negative, so the vapour pressure is lower; this drives capillary condensation.
  3. Zero remanence and zero coercivity (no hysteresis loop), with a large, saturating (Langevin-like) response to the applied field.

Level 2: Part B Style

Practice Q1. The specific surface area (m² g⁻¹) of non-porous spherical silver nanoparticles of diameter 10 nm (density 10.5 g cm⁻³) is closest to:

(a) 5.7 (b) 29 (c) 57 (d) 114

Answer: (c) 57

\( S = 6/(\rho D) = 6/(10.5 \times 10^6\ \text{g m}^{-3} \times 10 \times 10^{-9}\ \text{m}) = 6/0.105 = 57.1 \) m² g⁻¹.

Option (b) uses \( 3/(\rho D) \) (radius formula with the diameter); option (d) uses the radius in \( 6/(\rho D) \).

Practice Q2. In a 309-atom closed-shell cuboctahedral cluster, the number of atoms that are not on the surface is:

(a) 147 (b) 162 (c) 55 (d) 92

Answer: (a) 147

\( n = 4 \): surface atoms \( = 10(16) + 2 = 162 \); interior atoms \( = 309 - 162 = 147 \), which is itself the previous magic number (the interior of an \( n \)-shell cluster is the \( (n-1) \)-shell cluster). Option (b) is the surface count.

Practice Q3. Which statement about a dispersion of nanoparticles of the same solid with a range of sizes, left in contact with its saturated solution, is correct?

(a) All particles dissolve at the same rate because the solution is saturated with respect to the bulk solid. (b) The smallest particles grow because their surface energy is highest. (c) The smallest particles dissolve and the largest grow, so the mean size increases with time. (d) The size distribution narrows with time because large particles dissolve.

Answer: (c)

By the Ostwald–Freundlich equation the solubility rises as \( r \) falls. The solution is supersaturated with respect to large particles and undersaturated with respect to small ones: this is Ostwald ripening. The mean size increases and the distribution generally broadens.

Practice Q4. For single-domain magnetic nanoparticles measured with the same instrument, the blocking temperature of particles of diameter 12 nm compared with particles of diameter 6 nm (same material) is:

(a) the same (b) 2 times higher (c) 4 times higher (d) 8 times higher

Answer: (d) 8 times higher

\( T_B = KV/[k_B\ln(\tau_m/\tau_0)] \) with \( V \propto d^3 \): \( (12/6)^3 = 8 \).

Level 3: Part C Style

Practice Q5. Water at 298.15 K has \( \gamma = 0.072 \) N m⁻¹ and \( V_m = 1.8 \times 10^{-5} \) m³ mol⁻¹. Consider:

  1. The vapour pressure of a droplet of radius 1 nm is about 2.8 times that over a flat surface.
  2. The Laplace pressure inside a droplet of radius 10 nm is about 144 bar.
  3. In a cylindrical pore that water wets completely, condensation occurs at a relative pressure above 1.
  4. Increasing the temperature at fixed \( r \), with \( \gamma \) and \( V_m \) held constant, reduces the ratio \( p(r)/p_\infty \).

Which are correct?

(a) 1 and 2 only (b) 1, 2 and 4 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4

Answer: (b) 1, 2 and 4 only
  1. \( \ln(p/p_\infty) = 2(0.072)(1.8 \times 10^{-5})/(10^{-9} \times 8.314 \times 298.15) = 2.592 \times 10^{-6}/2.479 \times 10^{-6} = 1.046 \), so \( p/p_\infty = 2.85 \). Correct.
  2. \( \Delta P = 2(0.072)/10^{-8} = 1.44 \times 10^7 \) Pa = 144 bar. Correct.
  3. A wetted pore has a concave meniscus (negative \( r \)), so condensation occurs below \( p/p_\infty = 1 \) (capillary condensation). Wrong.
  4. \( T \) is in the denominator of the exponent; the ratio falls towards 1. Correct.

Practice Q6. A series of gold particles is studied. The melting points measured for radii of 2.0 nm and 4.0 nm are 1003 K and 1170 K. Assuming the Gibbs–Thomson form \( T_m(r) = T_m(\infty)(1 - \beta/r) \), the bulk melting point and \( \beta \) are closest to:

(a) 1337 K and 0.50 nm (b) 1300 K and 0.45 nm (c) 1337 K and 0.25 nm (d) 1250 K and 0.50 nm

Answer: (a) 1337 K and 0.50 nm

\( T_m \) is linear in \( 1/r \): \( T_m = T_\infty - T_\infty\beta(1/r) \).

Slope \( = (1170 - 1003)/(0.25 - 0.50) = 167/(-0.25) = -668 \) K nm. Intercept: \( T_\infty = 1170 + 668 \times 0.25 = 1337 \) K. \( \beta = 668/1337 = 0.50 \) nm.

These numbers are constructed for the exercise; the intercept equals the known bulk melting point of gold, as it should.

Practice Q7. Spherical magnetite-like particles have \( K = 1.0 \times 10^4 \) J m⁻³ and \( \tau_0 = 10^{-9} \) s. A magnetometer has a measurement time of 100 s. Which statements are correct?

  1. Particles of diameter 20 nm have a blocking temperature of about 120 K.
  2. At 300 K the 20 nm particles show a hysteresis loop in this magnetometer.
  3. In a measurement with \( \tau_m = 10^{-8} \) s, the same 20 nm particles would appear blocked at 300 K.
  4. Doubling the diameter to 40 nm would raise \( T_B \) to about 960 K for the magnetometer measurement.

(a) 1 and 3 only (b) 1, 3 and 4 only (c) 2 and 4 only (d) 1, 2 and 3 only

Answer: (b) 1, 3 and 4 only
  1. \( V = (\pi/6)(20 \times 10^{-9})^3 = 4.19 \times 10^{-24} \) m³; \( KV = 4.19 \times 10^{-20} \) J; \( T_B = KV/(k_B \times 25.33) = 120 \) K. Correct.
  2. 300 K is above \( T_B \): superparamagnetic, no hysteresis. Wrong.
  3. At 300 K, \( KV/k_BT = 10.1 \), so \( \tau = 10^{-9}e^{10.1} = 2.5 \times 10^{-5} \) s, longer than \( 10^{-8} \) s: the moment looks frozen. Correct.
  4. \( T_B \propto d^3 \): \( 120 \times 8 = 960 \) K. Correct as a model result (in a real material \( T_B \) cannot exceed the Curie temperature, and large particles may no longer be single-domain).

Quick Revision Table

QuantityFormula / valueNote
Nanoscale1–100 nm in at least one dimensionconvention, not physics
Surface/volumesphere \( 3/r \); cube \( 6/a \)doubles when size halves
Specific surface area\( 6/(\rho D) \)\( 6000/(\rho[\text{g cm}^{-3}]D[\text{nm}]) \) m² g⁻¹
Magic clusters\( N = (10n^3 + 15n^2 + 11n + 3)/3 \); surface \( 10n^2 + 2 \)13, 55, 147, 309, 561
Laplace\( \Delta P = 2\gamma/r \)144 bar for water, \( r = 10 \) nm
Kelvin\( \ln(p/p_\infty) = 2\gamma V_m/rRT \)2.85 for water, \( r = 1 \) nm
Ostwald–Freundlich\( \ln(S/S_\infty) = 2\gamma V_m/rRT \)drives Ostwald ripening
Gibbs–Thomson\( T_m(r) = T_m(\infty)(1 - \beta/r) \)linear in \( 1/r \)
Néel relaxation\( \tau = \tau_0e^{KV/k_BT} \)\( \tau_0 \approx 10^{-9} \) s
Blocking temperature\( T_B = KV/[k_B\ln(\tau_m/\tau_0)] \approx KV/25k_B \)\( \propto d^3 \)

Key Takeaways

What’s Next