The Radial Distribution Function, often denoted g(r) and frequently abbreviated as RDF, is a cornerstone concept in physical chemistry, materials science and soft matter physics. It describes how particle density varies as a function of distance from a reference particle, offering a window into the microscopic arrangement of matter. This guide provides a clear, practical, and UK-focused exploration of the Radial Distribution Function, covering theory, computation, interpretation in different phases, and real-world applications.
What is the Radial Distribution Function?
Definition and intuition
The Radial Distribution Function g(r) is a statistical measure of structure. In a homogeneous, isotropic system, it represents the probability of finding a particle at a distance r from a reference particle, relative to the probability expected for a completely random distribution at the same density. When g(r) equals 1, particles are distributed as in an ideal gas; when g(r) exceeds 1, there is an enhanced probability of finding neighbours at that separation. Peaks in g(r) reveal preferred interparticle distances governed by bonding, packing, and interatomic forces.
Radial Distribution Function vs Structure Factor
In scattering experiments, the measurable quantity is often the structure factor S(q). The Radial Distribution Function and S(q) are connected by a Fourier transform. Conceptually, g(r) focuses on real-space arrangements, while S(q) captures reciprocal-space information. Together, they provide a complete picture of the spatial order in a system. For liquids, the RDF tends to become flat at large r, approaching unity, whereas crystalline materials exhibit sharp, long-range peaks in g(r).
Mathematical Foundations
Formal definition
For a system with N particles in a volume V and number density ρ = N/V, the radial distribution function is defined by the ensemble average:
- g(r) = (V / N^2) ⟨ Σ_{i≠j} δ(r − |r_i − r_j|) ⟩ / (4π r^2)
Equivalently, g(r) describes how the average local density at a distance r from a reference particle compares with the bulk density ρ. In practice, g(r) is computed from observed pair distances in snapshots of a simulation or from experimental data via appropriate processing of S(q).
Normalization and units
Normalisation ensures that, at very large r where correlations vanish, g(r) tends toward unity for a homogeneous fluid. The characteristic unit for r is length (often angstroms, Å). In multi-component systems, one distinguishes partial radial distribution functions g_{AB}(r) describing correlations between species A and B, each normalised with the respective number densities.
Calculating the Radial Distribution Function from Simulations
From Molecular Dynamics
In molecular dynamics (MD), the RDF is estimated by sampling interparticle distances across time. A typical workflow includes:
- Choose a bin width Δr and a maximum distance r_max that covers all relevant correlations.
- For each frame, compute all pairwise distances |r_i − r_j| (with i ≠ j) and accumulate them into a histogram with bins of width Δr.
- Normalise the histogram by the spherical shell volume 4π r^2 Δr and by the average density ρ, averaging over frames for improved statistics.
The resulting g(r) highlights short-range structure (first few Å for liquids like water) and decays toward unity as r increases beyond the correlation length.
From Monte Carlo
Monte Carlo (MC) simulations similarly provide configurations from which the RDF can be computed. Because MC samples are equiprobable under the chosen ensemble, the same binning and normalisation steps apply. In practice, MC RDFs are often used to assess static structures and equilibrium properties rather than dynamic behaviour.
From experiments: from S(q) to g(r)
Experimentally, RDFs are frequently obtained indirectly via scattering data. The measured intensity relates to S(q), which can be Fourier transformed to yield the pair distribution function, a real-space analogue of the radial distribution function. In practise, careful background subtraction, resolution corrections, and Fourier transform windows are required to obtain physically meaningful g(r) from S(q).
Interpreting the RDF across Phases
Liquids
In liquids, the Radial Distribution Function shows a pronounced first peak at a distance corresponding to the most probable bond length or contact separation. Subsequent damped oscillations reflect short-range order without long-range periodicity. The height and sharpness of the first peak provide insight into packing density and bonding strength, while the rate of damping indicates how quickly order decays with distance.
Crystalline solids
Crystalline materials exhibit sharp, well-defined peaks in g(r) at specific r values corresponding to lattice spacings. The degree of peak sharpness, their splitting, and their persistence reveal crystal symmetry, defect concentration, and phase transitions. In ideal crystals, the RDF contains delta-function-like spikes; in real materials, thermal motion broadens peaks and introduces small shoulders.
Amorphous materials and glasses
Amorphous materials, including many glasses, display broader, less regular peak structures in g(r) compared with crystals, yet retain pronounced short-range order. The first peak still marks nearest-neighbour separations, while the absence of long-range order leads to a smoother approach to unity and more gradual damping of oscillations.
Partial Radial Distribution Functions
Heterogeneous systems
In multi-component systems, such as alloys, ionic liquids, or aqueous solutions, partial Radial Distribution Functions g_{AB}(r) disentangle the relationships between specific species. Summing over all pairs with appropriate weighting recovers the total RDF, but g_{AB}(r) provides richer information about chemical interactions, coordination environments, and preferential bonding.
Examples: water and salts
For water, g_OO(r) reveals strong O–O correlations tied to the hydrogen-bond network, with characteristic first- and second-shell separations. In saline solutions, g_NaCl(r) and g_ClNa(r) reflect ion pairing tendencies, which influence properties like conductivity and viscosity. Partial RDFs are essential for understanding local structure in complex mixtures.
Coordination Numbers and Local Structure
Computing coordination numbers
The coordination number (CN) around a reference particle is obtained by integrating the RDF up to a chosen cut-off distance r_c:
- CN = 4π ρ ∫_0^{r_c} g(r) r^2 dr
The choice of r_c is guided by the first minimum after the first peak in g(r) or by specific physical criteria. CN provides a practical metric of local packing and bonding environments, shaping our understanding of phase behaviour and reactivity.
Relation to local chemistry
CN values offer insights into preferred coordination geometries (e.g., tetrahedral versus octahedral in oxides), which in turn influence mechanical properties, diffusion, and reaction pathways. In metallic systems, deviations from ideal coordination can signal defects, grain boundaries, or amorphous character.
RDF in Practice: Tips and Best Practices
Choosing bin width and range
Bin width Δr should balance resolution with statistical noise. Too small a Δr yields noisy RDFs; too large a Δr blurs important features. A common rule is to start with Δr in the range 0.01–0.05 nm (0.1–0.5 Å) and adjust based on the system size and desired detail. The maximum range r_max should at least span several coordination shells, typically 1–2 nm for liquids and up to a few nanometres for extended solids, subject to computational constraints.
Finite-size effects and boundary conditions
Periodic boundary conditions mitigate finite-size effects but can still impose artefacts at distances approaching half the box length. When computing g(r), ensure the simulation cell is large enough to capture the first few coordination shells without spurious wrap-around effects. For highly structured systems, consider multiple independent simulations to assess robustness.
Error analysis and smoothing
Statistical uncertainty in g(r) stems from finite sampling. Block averaging and bootstrapping can estimate errors. Smoothing with careful kernel methods may improve visual clarity but should not obscure real features. Always report uncertainty alongside key features such as peak positions and coordination numbers.
Case Studies and Real-World Applications
Water and the hydrogen-bond network
In liquid water, the RDF between oxygen atoms, g_OO(r), exhibits a prominent first peak around 2.8 Å, reflecting the most probable O–O separation in the hydrogen-bond network. The corresponding O–H RDF, g_OH(r), has a sharp peak near 1.0 Å associated with the covalent O–H bond length. The positions and amplitudes of these features illuminate the balance between hydrogen bonding and thermal motion, and track structural changes with temperature and pressure.
Metallic systems and glasses
RDFs in metals typically show shorter-range order with less pronounced oscillations than crystals yet clearer short-range order than simple liquids. In metallic glasses, RDFs reveal broad, damped peaks indicating disordered packing. Analyzing g(r) alongside the partial functions g_{AA}(r) helps identify preferred atomic pairings and short-range motifs that govern mechanical properties such as hardness and ductility.
Colloids and soft matter
In colloidal suspensions and soft matter, RDFs inform on crowding, phase separation, and gelation. For example, a rising first peak with increasing concentration signals enhanced local order, while changes in the second peak can indicate evolving hexagonal or random packings. Partially resolved RDFs in multicomponent systems reveal fractionation or depletion effects at the particle–solvent interface.
Advanced Topics: Beyond the Standard RDF
Three-body correlations and higher-order functions
The Radial Distribution Function captures pair correlations. For a complete structural description, higher-order correlations such as the three-body distribution provide information about angular relationships and bonding motifs. Three-body correlations are especially important in materials with directional bonding, such as water networks or covalent glasses.
Discrete versus continuous RDF interpretation
In simulations with rigid, well-defined particles, g(r) reflects discrete interaction distances. In coarse-grained models, the effective RDF can blur these features because of smoothing and effective potentials. Understanding the limitations of a given model is essential when interpreting RDFs and comparing to experimental data.
Temperature, pressure and phase transitions
RDFs respond to thermodynamic conditions. Increasing temperature generally broadens peaks and reduces their amplitudes, while pressure can shift peak positions due to compression. Near phase transitions, features in g(r) may evolve abruptly, signalling changes in local order and coordination.
Practical Workflow: Step-by-Step for Researchers
1) Define system and objectives
Clarify whether you study a pure substance, a mixture, a solution, or a solid. Decide whether you need total RDF or partial RDFs for specific species. Establish a target accuracy and a plan for statistical convergence.
2) Prepare data and choose parameters
Obtain representative configurations (frames) from MD or MC. Choose Δr and r_max based on expected interparticle distances. For ionic systems, consider separate RDFs for cation–anion pairs.
3) Compute and normalise
Compute pairwise distances, bin them, and apply the standard normalisation to obtain g(r). For mixtures, compute each g_{AB}(r) and the total g(r) as a weighted sum if needed.
4) Analyse and validate
Identify peak positions, compare with known bond lengths, and calculate coordination numbers. Validate by cross-checking with experimental data or established simulations. Document uncertainties and discuss potential sources of error.
Conclusion: The Radial Distribution Function as a Diagnostic Tool
The Radial Distribution Function is more than a mathematical construct; it is a practical diagnostic for understanding structure at the atomic and molecular level. By distilling complex arrangements into a function g(r), researchers can quantify local order, compare phases, and connect microscopic arrangements to macroscopic properties. Whether you are exploring the subtle hydrogen-bonding network in water, the short-range order in metallic glasses, or the packing in colloidal suspensions, the Radial Distribution Function remains an essential tool in the modern scientific toolkit.
As research progresses and computational power expands, the Radial Distribution Function continues to evolve in tandem with techniques for extracting and interpreting g(r) from real-world data. Mastery of RDF analysis — including partial RDFs, coordination numbers, and the relationship to S(q) — empowers scientists to unlock deeper insights into the structure and behaviour of matter across disciplines.