Direct methods are a powerful tool for determining crystal structures without prior knowledge of atomic positions. They use statistical relationships between structure factor magnitudes and phases, relying on normalized structure factors (E-values) to enhance these relationships.

Phase determination in direct methods is based on probability distributions and statistical methods. Techniques like the tangent formula, Karle recycling, and multi-solution approaches are used to solve and extend phases. These methods are crucial for unraveling complex crystal structures.

Phase Relationships and Probability

Fundamentals of Direct Methods

Direct methods determine crystal structures without prior knowledge of atomic positions
Utilize statistical relationships between structure factor magnitudes and phases
Rely on normalized structure factors (E-values) to enhance phase relationships
E-values calculated by dividing observed structure factors by expected average intensities
Triplet relationships form the basis for phase determination in direct methods
Triplet involves three reflections with indices that sum to zero (h1 + h2 + h3 = 0)

Probability Distributions in Phase Determination

Probability distributions describe likelihood of phase relationships being correct
Cochran distribution predicts probability of triplet phase sum being near 0 or 180 degrees
Tangent formula derives from probability considerations for phase relationships
Karle-Hauptman determinants assess reliability of phase predictions
Joint probability distributions account for multiple phase relationships simultaneously
Conditional probability used to refine phase estimates based on known information

Application of Statistical Methods

Sayre equation relates structure factors of related reflections
Hauptman-Karle inequality provides constraints on possible phase combinations
Wilson statistics help estimate overall scale factor and temperature factor
Likelihood functions assess agreement between observed and calculated structure factors
Maximum entropy methods optimize phase choices based on information theory principles

Fundamentals of Direct Methods, X-ray crystallography - wikidoc

Solving and Extending Phases

Phase Determination Techniques

Tangent formula refines phase estimates using weighted contributions from multiple relationships
Karle recycling iteratively applies tangent formula to improve phase accuracy
Multi-solution methods generate multiple phase sets to explore solution space
Random phases used as starting points for phase determination in some approaches
Shake-and-bake method combines reciprocal and direct space calculations
Dual-space methods alternate between real and reciprocal space refinement (charge flipping)

Phase extension gradually includes higher resolution data in phasing process
Density modification techniques improve phases (solvent flattening, histogram matching)
Electron density map interpretation guides phase improvement (atom picking, model building)
Maximum likelihood refinement optimizes agreement between observed and calculated data
Difference Fourier maps reveal missing or incorrectly placed atoms
Anomalous scattering data can provide initial phases or validate phase choices

Overcoming Challenges in Phase Determination

Space group determination crucial for correct application of phase relationships
Pseudo-symmetry can complicate phase determination and structure solution
Heavy atom methods provide initial phases for larger structures
Molecular replacement utilizes known similar structures to obtain initial phases
Fragment-based approaches build up structure from small, well-determined motifs
Machine learning algorithms increasingly used to predict phases and interpret maps