Documente Academic
Documente Profesional
Documente Cultură
Encyclopedia of Physical Science and Technology EN004G-160 June 15, 2001 12:44
Crystallography
Jeffrey R. Deschamps
Judith L. Flippen-Anderson
Laboratory for the Structure of Matter, Naval Research Laboratory
I. Introduction
II. Evolution of Crystallography
III. Structure of a Crystal
IV. Steps in Crystal Structure Analysis
V. Comparison of X-Ray and Neutron Diffraction
VI. Results
Appendix I: Factors Affecting Intensities
Appendix II: Methods of Structure Solution
Appendix III: Methods of Refinement
121
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
122 Crystallography
Miller indices A set of integers with no common factors, Unit cell Parallelepiped bounded by three noncoplanar
inversely proportional to the intercepts with the crystal vectors a, b, c with angles , , that repeats by trans-
axes of a lattice plane. lation. If this unit is the smallest volume that meets
Orientation matrix A matrix relating the crystal axes to these criteria it is referred to as the primitive unit cell.
the instrument axes such that one can predict the values X-ray Electromagnetic radiation with wavelengths in the
of the instrument angles (2, , , and ) for a given range 0.01 to 1.0 nm (0.1 to 10 A). The wavelength
reflection of the crystal. range most commonly used in diffraction experiments
Patterson function A Fourier summation that uses the is between 0.71 and 1.54 A. Shorter wavelengths re-
squares of the structure factor magnitudes as coeffi- quire longer path-lengths (distance from crystal to de-
cients. The peaks in this map correspond to vectors tector) in order to resolve adjacent peaks in the diffrac-
between atoms. The peak height is related to the scat- tion pattern. Both the sample and the air along the beam
tering powers of the atoms at the two ends of the vector. path can significantly attenuate longer wavelengths.
The region around the origin gives information about
bonded distances.
Phase problem A central problem of crystallography. CRYSTALLOGRAPHY is a broadly encompassing dis-
The intensities of the different reflections allow deriva- cipline that involves a variety of fields of study. The pri-
tion of the amplitude of the structure factors but not mary concern of modern crystallography is the three-
their phases. The phases are required in order to calcu- dimensional arrangement of atoms in matter. Although the
late the electron density, which is a map showing the term most often refers to studies of crystalline solids (ei-
position of atoms in the unit cell. ther single crystals or crystalline powders) using X-ray or
Point group A group of symmetry operations that leave neutron diffraction, it encompasses a much broader range
unmoved at least one point within the object to which of methodologies.
they apply.
Polar space group Space group in which the origin is not
fixed by symmetry and hence must be defined (e.g., the I. INTRODUCTION
space group P21 ).
Reciprocal lattice A set of imaginary points constructed In the late 1660s, crystallography began as the study of the
in such a way that the direction of a vector from one macroscopic geometry of crystals. Crystals were grouped
point to another coincides with the direction of a normal into systems on the basis of the symmetry of their external
to the real space planes within the crystal. The sepa- shapes. Based on these observations, seven crystal systems
ration of those points (absolute value of the vector) is were identified: triclinic, monoclinic, trigonal, tetragonal,
equal to the reciprocal of the real inter-planar distance. hexagonal, orthorhombic, and cubic. It was theorized that
Space group Identical atom groups are usually symmetri- the observed crystal shapes could be built up by stacking
cally arranged within the crystal lattice. The symmetry minute balls. This idea was refined, and in the late 18th
relating the groups may be due to rotations, inversions, century it was thought that crystals were composed of ele-
mirror planes, or some other relational operation. The mentary building blocks (now referred to as the unit cell).
space group constitutes a mathematical shorthand de- This was supported by the orderly cleavage angles of cal-
scription of the symmetry operations required to pro- cite, which suggested a regular stacking of these elemen-
duce the unit cell. tary blocks. Further studies in the 1800s led to derivation
Special position A point left invariant by at least two of the 32 different point groups, the Bravais lattices, and
symmetry operations of the space group. the 230 space groups. All of these advances were made
Structure factor Fhkl , complex quantity corresponding before any direct observations of the arrangement of atoms
to the amplitude and phase of the diffraction maximum within a crystal were possible.
associated with the reciprocal lattice point hkl: In 1912, Von Laue reasoned that the arrangement of
N atoms in crystals could help him measure the wavelength
Fhkl = f j exp[2i(hx j + ky j + lz j )] of X-rays. Based on experiments with copper sulfate,
j=1 he demonstrated the ability of crystals to act as three-
where N is the number of atoms in the unit cell and dimensional diffraction gratings. Crystallography had en-
x j , y j , z j are the fractional coordinates of the jth atom. tered a new era: the analysis of the arrangement of atoms
Torsion angle If a group of four atoms (ABCD) is pro- in a crystal by careful analysis of the diffraction pattern of
jected onto a plane normal to the bond between B and that crystal. Crystal structures were now viewed as being
C, the angle between bonds connecting A and B, and built up from repeating units of an atomic pattern rather
C and D is the torsion angle. than the regular stacking of solid shapes.
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 123
As new types of scattering were discovered, they be- by eye, and Fourier transforms were performed with
came incorporated into the discipline, and crystallogra- BeeversLipson or PattersonTunell strips and summed
phy came to include structural studies of all classes of by hand. Three-dimensional electron-density maps were
substances. Absorption, diffraction, or other scattering plotted by hand, one section at a time, traced onto glass
methods are used to study crystals, powders, amorphous sheets, and stacked in frames for interpretation. As long
materials, surfaces, liquids, and gases. The physics and as all calculations were done by hand, small, flat molecules
methodologies associated with these techniques are also were the ones most amenable to study. Early computers
part of the science of crystallography. were difficult to program and had limited storage capacity,
Crystallographic studies play a vital role in materials but they made it possible to solve and refine crystal struc-
science, chemistry, pharmacology, mineralogy, polymer tures of molecules of moderate size in 6 months to a year.
science, and molecular biology. Accurate knowledge of The rapid advances in crystallography owe much to the
molecular structures is a prerequisite for rational drug de- development of computer-controlled diffractometers for
sign and structure-based functional studies. Crystallogra- data collection, high-speed computers for data analyses,
phy is the only method for determining the absolute and, most recently, powerful graphics devices for display-
configuration of a molecule. Absolute configuration is a ing structures with the ability to perform real-time manip-
critical property in biological systems, as changes in this ulations. Tasks that required months of effort can now be
may alter the response of the biologic system. accomplished in minutes with the aid of a computer.
A requirement for the high accuracy of crystallographic
structures is that a good crystal must be found, and this A. The Early Years
is often the rate-limiting step. Additionally, only limited
The first experiments in X-ray diffraction were recorded
information about the dynamic behavior of the molecule
on film. By 1913, W. H. Bragg had constructed the first
is available from a single diffraction experiment.
X-ray spectrometer to allow a more careful study of
In the past three decades, new developments in detec-
X-rays. This instrument also proved useful in studies
tors, increases in computer power, and powerful graph-
on crystals. Using measurements made with this X-ray
ics capabilities have contributed to a dramatic increase in
spectrometer, Braggs son determined the structures of
the number of materials characterized by crystallography.
fluorspar, cuprite, zincblende, iron pyrites, sodium nitrate,
Synchrotron sources offer the possibility of time-resolved
and the calcite group of minerals.
studies of physical, chemical, and biochemical processes
The first crystal structures reported were of substances
in the millisecond to nanosecond range; the ability to study
crystallizing in cubic space groups. The structure of
the nearest neighbors of cations present at parts per mil-
diamond was determined in 1913 by the Braggs, from
lion concentrations; and the possibility of recording small-
symmetry considerations using the observed intensities to
angle scattering data and powder data in seconds. Charge-
discriminate among possible structures. Their model
density studies have been made on numerous light atom
established the carboncarbon single bond distance of
structures and are beginning to provide new insights into and confirmed that bonds to carbon are directed
1.52 A
bonding of transition metals. Rietveld refinement is rev-
tetrahedrally. The younger Bragg combined symmetry ar-
olutionizing the study of powders and is being extended
guments with the notion that the scattering power of atoms
to fibers. Direct methods of structure solution are being
is related to their atomic weight to explain the structures
applied successfully to structures of over 1000 atoms. The
of the alkyl halides. These concepts were extended to ZnS,
Human Genome Project has created many opportunities
CaF2 , and FeS2 . The alkyl halide structures show anions
for crystallographic studies of biological macromolecules
surrounded by cations and cations surrounded by anions,
and resulted in intense activity in the areas of structural
demonstrating conclusively that discrete molecules
genomics and proteomics. Polarized neutrons are being
of the type NaCl do not exist in crystals of ionic
used to determine the spin structure of magnetic materials
materials. Pauling published the first intermetallic struc-
and to probe the surface structure of such materials. The
ture, Mg2 Sn, in 1923. In that same year, Dickerson and
review that follows attempts to describe the fundamentals
Raymond showed that hexamethylene tetramine consisted
of crystallography and capture the excitement and diver-
of discrete molecules, each having the same structure,
sity within the discipline. in a body-centered cubic
with CN distance of 1.44 A
lattice of edge 7.02 A.
II. EVOLUTION OF CRYSTALLOGRAPHY
1. Heavy-Atom Methods
It was not very long ago that X-ray diffraction data were
collected on photographic film, intensities of spots on The elder Bragg realized that the periodic pattern in the
the film (corresponding to data points) were measured electron-density distribution could be represented by a
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
124 Crystallography
Fourier summation. The coefficients of this summation The crystal structure can be regarded as alternating layers
became known as the structure factors. This allowed the of Zn and S atoms perpendicular to the {111} direction.
solution of structures where positions of atoms were not Looking at one layer of Zn atoms, we find it lies closer to
restricted to special positions. The determination of the one of the two adjacent layers of S atoms. If we assume
structure of diopside, CaMg(SiO3 )2 , in 1928 was the first that the short ZnS spacing is a bonding interaction and
example of the use of this method. the long Zn . . . S gap is a van der Waals contact, we do not
In 1934, Patterson showed that a Fourier series with |F|2 expect cleavage between bonded layers. This implies that
as coefficients could be summed without knowledge of the one of the {111} faces corresponds to a layer of S atoms
phases and would reveal interatomic vectors. Since the and the other to a layer of Zn atoms.
weight of a peak in the Patterson function is proportional It was fortuitous that the intensities in the Coster, Knol,
to the product Z i Z j of the atomic numbers of the atoms at and Prins study were measured using AuL radiation
the ends of the vector, vectors involving heavy atoms stand ( = 1.276 A).
The K-absorption edge of Zn is 1.283 A.
out among light atomlight atom vectors. In 1936, Harker Thus, the measured intensities of Friedel pairs were not
showed that symmetry properties of the crystal caused the same. This was the first example of anomalous scatter-
vector density to accumulate on certain planes and lines ing. The difference due to anomalous scattering is greatest
(later known as the Harker section) in the Patterson func- when data are collected near an absorption edge of a heavy
tion. These two papers were the foundation of the heavy- atom in the structure. It was nearly 20 years after the ZnS
atom method for crystal structure solution. The method experiment that Bijvoet realized this principle could be
assumes that phases calculated from the heavy-atom posi- used to determine the absolute configuration of the sodium
tions will be sufficiently accurate that a Fourier synthesis rubidium salt of (+)-tartaric acid.
(using |F| as opposed to |F|2 ) will reveal the positions of
more atoms, thus allowing solution of the structure. Phase
B. Modern Crystallography
information from the new atoms could then be added to
the Fourier synthesis to locate more new atoms and so on, The early years can be characterized as a period in which
until the full structure was revealed. the size of a problem amenable to analysis was compu-
The method of isomorphous replacement was used first tationally and instrumentally limited. Data collection and
to solve the structure of the alums by Lipson and Beevers analysis were manual processes. This began to change in
in 1935. If a centrosymmetric light-atom structure and its the 1950s.
heavy-atom derivative differ only in the presence of the
heavy atom, then the differences in the intensity of equiv-
1. Data Collection
alent reflections can be used to determine the signs of
the structure factors. Robertson applied this method to the Progress in crystallographic data collection can be charted
phthalocyanines in 1936. These molecules form an iso- by examining early issues of Acta Crystallographica,
morphous series and crystallize in P21 /a. By comparing published by the International Union of Crystallography
intensities of nickel phthalocyanine and the unsubstituted (IUCr). Although a manual diffractometer was available
molecule, Robertson was able to assign phases to all but in 1913, the primary method for collecting intensity data
a few of the measured reflections from the h0l projection. for crystallographic studies relied on X-ray cameras and
Harker generalized the method to the noncentrosymmetric film. In 1948, about two-thirds of the crystal structures re-
case in 1956. Isomorphous replacement can be combined ported used a camera and film to collect the intensity data,
with anomalous dispersion to obtain phase information for and less than one-quarter used a combination of film and a
large molecules. diffractometer. Despite the existence of automated single-
crystal diffractometers, the situation was little changed by
1962.
2. Absolute Configuration
Use of the Cambridge Structure Database (CSD) allows
Friedels law states that the scattering from the front and a more systematic study of the evolution of data collec-
back sides of a plane, hkl, are the same. This means that the tion methods (Fig. 1) with the caveat that this database
measured intensities of the 111 and 111 reflections (and references only organic and organometallic structures. Al-
all other Friedel pairs) should be equivalent. However, though automated diffractometers were first available in
in a 1930 study on ZnS by Coster, Knol, and Prins, it about 1955 it was not until the mid-1960s that this new
was noted that the 111 reflection was not equivalent to technology made much of an impact. At that time, only
its Friedel mate. In the case of the ZnS crystal, the {111} a few hundred structures were being added to the CSD
faces are prominent but, even by visual examination, do each year. By the mid-1970s, over 1000 structures were
not appear identical. One face is shiny and the other is dull. added to the CSD each year and data for over 80% were
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 125
FIGURE 1 Changes in data-collection methods with time. Using structures entered in the Cambridge Structure
Database, changes in data collection strategies were determined by examining the fraction of structures reported
in one year for various data collection methods (photograph with visual estimation of intensities, photograph with
densitometer, diffractometer, or unknown). The large fraction listed as unknown between 1960 and 1965 is likely due
to changes occurring at that time in how data were collected and how data-collection methods were reported.
collected on a diffractometer. In 1997, over 15,000 struc- concerned the development of the mathematical aspects
tures were added to the CSD and virtually all were col- of crystal structure analysis. The first general procedures
lected on a diffractometer. The explosive growth of this for solving both centrosymmetric and noncentrosymmet-
database cannot be attributed solely to improvements in ric structures was developed in the early part of the 1960s.
data collection, but certainly the routine use of automated Use of this method grew with the power of computers and
instrumentation had a significant impact. computer programs. It is now the most widely used method
of solving crystal structures of moderate size. Efforts are
currently being made to apply direct methods to very large
2. Structure Solution structures such as proteins. For a more detail discussion
Data collection was not the only beneficiary of the post- of structure solution methods, see Section IV.H.
World War II progress in crystallography. No general
method existed for solving unknown structures without
3. Charge-Density Distribution
heavy atoms until the advent of direct methodsa means
of determining the values of phases from relationships The first papers to explore the difference between the
among the structure factor magnitudes associated with results of X-ray and neutron diffraction experiments ap-
those phases. The earliest structure solved by such meth- peared in the early 1970s. Systematic differences between
ods was decaborane by use of some inequalities derived by the positional and thermal parameters determined by the
Harker and Kasper. However, this technique was limited to two techniques were reported. These differences were at-
centrosymmetric structures. The major effort in the 1950s tributed to the difference in how neutrons and X-rays
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
126 Crystallography
interact with atoms. Neutrons are diffracted primarily systematically underestimate lone-pair peak heights and
from the nucleus; hence, neutron diffraction produces in- place hydrogen atoms too close to the atoms to which
formation about nuclear position. X-rays are diffracted by they are bonded. The experiments must be conducted at
electrons and therefore yield information about the dis- low temperature (i.e., 75 C or preferably less) for this
tribution of electrons in the molecule. It was logical to method to succeed.
extend these ideas and attempt to map the redistribution Other approaches include refinement of separate param-
of electron density that occurs on bonding. eters for core and valence electrons or the direct refinement
A number of different kinds of mapping have been of a deformation model. The major advantage of the di-
done. Subtracting core-electron densities from experimen- rect refinement methods is that they make no assumptions
tal electron densities (i.e., p(r)) should reveal details of about the (sin )/ dependence of bonding features.
the redistribution of valence electrons on bonding. The Errors in the deformation density maps arise from er-
term valence density is used to describe the difference rors in both the model and the data. Demands on diffrac-
function: tion methodology and interpretation are many times more
severe than those relevant to an average stereochemical
pvalence (r) = p(r) pi,core (r) investigation. The following considerations are extremely
atoms
important:
To define either of these functions the positions and ther-
mal parameters must be known. One way to approach this 1. The X-ray data set must be complete (all symmetry-
is to use neutrons to determine the positional and thermal related reflections measured) to a (sin )/ limit of about
parameters and to use these parameters in conjunction 1.3 A 1 . This implies the use of short-wavelength ra-
with X-ray, spherical-atom scattering factors to calculate diation such as MoK ( = 0.7107 A) or AgK ( =
structure factors for the promolecule using: 0.5612 A). The maximum value of (sin )/ is 0.65 for
CuK ( = 1.5418 A).
Fcalc,N = f i exp(2iH ri )Ti
atoms
2. The data must be corrected for absorption, extinc-
tion, and thermal diffuse scattering. Multiple reflections
Fcalc,N is the X-ray structure factor calculated from neu- must be either avoided or eliminated. If these conditions
tron positions. The deformation density or X N map then are met, internal agreements should be of the order of 2%;
corresponds to: i.e.,
1
X N
pdeformation (r) = (Fobs,X Fcalc,N ) exp(2iHr)
V H
F 2 (H) F 2 (H)
F 2 (H) = 0.02
H H
Since both Fobs and Fcalc contain the effects of thermal mo-
tion, this deformation density map is thermally smeared.
4. Rietveld Analysis
Its resolution is limited by the maximum (sin )/ value
of the data. There has been a renaissance in powder diffraction in re-
Given that there are few neutron-diffraction facilities cent years because Rietveld refinement allows determina-
in the world and that it is difficult to correct adequately tions of positional and thermal parameters from powder
for systematic effects in the two experiments (i.e., ab- data, even when the diffraction peaks are not well sep-
sorption, extinction, thermal diffuse scattering, multiple arated in the recorded pattern. Rietveld analysis is not a
reflections), it would be desirable to study bonding effects method of structure solution and can only be applied when
using exclusively X-ray data. There are several approaches cell dimensions and space group are known and when a
to this problem. reasonable model exists for the structure.
The X X formalism is similar to the X N formal- In a polycrystalline sample, information may be lost as
ism described above, except that the calculated values for a result of the random orientation of the crystallites. A
positional and thermal parameters are derived from re- more serious loss of information can result from the over-
finement of high-angle X-ray data [(sin )/ > 0.70 A 1 ]. lap of independent diffraction peaks in the powder pat-
For deformation density maps, neutral spherical atoms are tern. Using the total integrated intensities of the separate
subtracted from the observed density; in valence density groups of overlapping peaks in the least-squares refine-
maps, HartreeFock core-electron densities are used to ment of a structure leads to the loss of all the information
evaluate Fcalc . Comparison of X X and X N maps shows contained in the often-detailed profile of these composite
that they do indeed yield the same qualitative information. peaks. Rietveld developed a refinement method that uses
Bonding density shows in the middle of bonds, and lone- the profile intensities of the composite peaks instead of
pair density is in the correct location. However, X X maps the integrated quantities. This is a pattern-fitting method
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 127
of structure refinement and allows extraction of the max- precision of the resulting parameters. However, precision
imum amount of information contained in the powder comparable to single-crystal X-ray diffraction can now be
pattern. obtained from neutron diffraction at a fixed wavelength
A powder pattern is recorded in a step-scan mode with with Rietveld refinement. For X-rays from conventional
a step width of 0.02 to 0.03 2 . No attempt is made to sources, the precision of positional parameters is compa-
allocate observed intensity to individual reflections or to rable to the single-crystal case, but thermal parameters are
resolve overlapping reflections. Instead, the intensity of less reliable by a factor of two or three. Considerable ef-
the powder diffraction pattern is calculated as a stepwise fort is being expended to improve profile functions for the
function of the angle, 2 . Refinement allows calculation various X-ray and neutron sources.
of the shifts in the parameters that will improve the fit of
the calculated powder pattern to the observed one. The
5. Small-Angle Scattering
quantity minimized is
Scattering at small angles is derived from large structural
wi [yi (obs) yi (calc)]2 unitsunits whose dimension D is much larger than the
i
wavelength of the radiation used in the experiment. The
where yi (obs) is the observed intensity of the ith step of the acronyms SAXS and SANS refer to small-angle X-ray
profile, and yi (calc) includes the usual structural parame- scattering and neutron scattering, respectively.
ters (i.e., positional parameters xi , yi , z i ; thermal parame- Different sorts of small-angle experiments are typi-
ter Bi j ; and site-occupancy parameters p j ). However, the cal of the kind of material studied and yield a charac-
model must also include instrumental and sample parame- teristic type of pattern. Low-angle data from ordered or
ters: 20 (overall scale), overall temperature factor, profile semi-ordered systems give Bragg peaks at specific values
breadth (H2 = U tan 2 + V tan + W), profile asymme- of scattering vector. Examples include aligned structures
try, background, preferred orientation, lattice parameters, with long-range periodicity, such as two-dimensional bi-
and wavelength. The agreement factors most often quoted ological structures or samples such as opal that present
are long-range order.
1/2 Scattering from polymers in dilute solution or from
wi [yi (obs) (1/c)yi (calc)]2 biological materials yields patterns that look rather like
Rweighted pattern =
wi [yi (obs)]2 the DebyeScherrer rings observed in wide-angle data
from powders. When scattering arises from the spheri-
and cal particles in a mono-disperse system, the pattern is a
Bessel function consisting of a succession of peaks of
|Ik (obs) Ik (calc)|
RBragg = diminishing magnitude that are broad relative to Bragg
Ik (obs)
peaks. Analysis of the pattern yields the radius of the
In RBragg , obs has quotation marks because Ik (obs) is particles.
calculated by partitioning the intensity. In some circumstances, multiple Bragg reflections give
To date, most of the papers published using this rise to scattering in the small-angle region. One of the
method have been neutron-diffraction studies from reac- advantages of the tunable sources of X-rays and neu-
tor sources. Advantages of neutron data include minimal trons is that multiple Bragg scattering can be avoided by
preferred orientation, no polarization, and neutron absorp- choosing wavelengths larger than the lattice spacing.
tion cross-sections smaller than X-ray values by a factor of Small-angle scattering can be applied to a wide variety
104 , scattering independent of , and for fixed-wavelength of materials. In polymer science, it has been used to in-
experiments the peak shape is simple. vestigate chain conformation in amorphous polymers, the
In X-ray experiments using radiation from conventional state of mixing in polymer blends, the compatibility ranges
sources, peak shape is complicated by both 1 2 split- of polymer blends, and the measurement of domain struc-
ting (at high angles) and by the fact that the peak shape ture and molecular conformation within those domains.
is neither Lorentzian nor Gaussian, but is better described In biological materials, examples include measurement of
by a convolution of these two functions called the Voigt the radius of gyration of proteins in solution; aggregation
function. The pseudo-Voigt function used in many pro- of chlorophyll into micelles; diffraction patterns of semi-
grams is an approximation of the Voigt function that can ordered materials such as muscle, collagen, etc.; and stud-
be evaluated much more quickly. ies of the shapes and constitution of viruses. Separation
Spallation neutron sources, time-of-flight experiments processes, such as those used in refining of metals or ex-
with reactor sources, and synchrotron sources all have spe- traction of tar sand, frequently involve micelle formation.
cial problems in defining peak shapes, and this limits the Small-angle scattering can give information regarding the
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
128 Crystallography
shape, size, and degree of polymerization of the aggre- 7. Implications of New X-Ray
gates. Studies of materials such as cements, zeolites, and and Neutron Sources
catalysts involve the measurement of size and distribution
of pores and measurements of specific surface. Synchrotron sources present a unique combination of
properties that are very attractive for X-ray scattering,
absorption, and diffraction experiments. The radiation
6. Extended X-Ray Absorption Fine Structure
produced has extreme brightness over a broad spectral
Extended X-ray absorption fine structure (EXAFS) is a range. Synchrotron sources are five to six orders of mag-
technique for studying the local environment of a specific nitude brighter than the bremsstrahlung (that part of the
atomic species in a complex matrix. Because the inter- X-ray spectrum caused by the slowing of electrons on
action of X-rays with the material under study is ab- impact with the target, also referred to as white radia-
sorption rather than diffraction, the technique can be ap- tion) available from a conventional rotating anode source.
plied to gases, liquids, and amorphous solids as well as to Monochromators currently available produce resolution
crystals. of 0.1 eV at 8 keV. The radiation is naturally collimated
In the experiment, the X-ray absorption coefficient is with a divergence of the order of 2104 radians, is plane
measured from slightly below to about 1000 eV above polarized, and has a precise time structure (subnanosec-
the absorption edge for the atomic species whose environ- ond pulses repeated every 0.5 to 1 sec). These properties
ment is to be studied. Through analysis of the fine structure allow experiments that simply cannot be done with con-
above the edge it is possible to determine the coordination ventional sources: EXAFS on dilute samples (parts per
of the atom. The actual position of the edge gives infor- million range), measurement of the magnitude and angu-
mation about the oxidation state of the absorbing atom, lar dependence of the real and imaginary components of
while structure at energies at or just below the edge pro- anomalous dispersion, and determination of the structure
vides information about bound states associated with the of a protein using only one derivative and three wave-
absorbing atom and hence about the symmetry of the en- lengths. (In this context, the use of anomalous dispersion
vironment. The technique can be applied to any atom with is formally equivalent to multiple isomorphous replace-
Z > 15 to yield a determination of radial distance to a pre- ment with the added feature that the isomorphism is ex-
cision of 0.02 to 0.01 A. For elements with high values act.) Determinations of cation-site distribution in minerals
of Z , it may be preferable to use the L-absorption edge and diffraction from monolayers on surfaces have many
rather than the K edge because higher fluxes are available; applications in such areas as catalysis or materials sci-
however, theory for the L edge is not yet well developed. ence. Perhaps the most exciting application is the ability
Extended X-ray absorption fine structure has applica- to do time-resolved studies of physical, chemical, and bi-
tions in many branches of science. In molecular biology, ological processes using small-angle scattering, powder
it has been used to study Ca+2 transport in membranes, diffraction, and other scattering techniques.
binding of oxygen in hemoglobin, and other coordination One early example was the study by Larson of temper-
problems. It is an invaluable tool for the study of amor- ature and temperature gradients in silicon during pulsed-
phous substances such as glass, since the manufacture of laser annealing. In this example, the duration of the laser
glasses with particular mechanical and thermal properties pulse was 15 nsec and that of the synchrotron X-ray pulse
depends on structure. Catalysis has become extremely im- was 0.15 nsec. The laser bursts were synchronized so that
portant in energy development, resource utilization, pollu- the probing X-rays arrived at 20, 55, and 155 nsec after
tion abatement, refining of metals, etc. The chemical state the laser pulse. The experiments showed that the lattice
and atomic environment of an atomic species in a catalyst temperature of silicon reaches the melting point during
in situ while reduced with hydrogen, chemisorbed with the 15-nsec pulse and remains at the melting point during
oxygen, heated, quenched, etc., can be determined with the high reflectivity phase, after which time the tempera-
EXAFS. This allows the design of heterogeneous catalysts ture rapidly subsides. Temperature gradients at the liquid
that are tailored from precise knowledge of electronic and solid interface were measured for the first time and were
structural parameters. found to be in the range of 107 K/cm. Larson received the
Synchrotron sources have revolutionized EXAFS stud- Warren Award for Diffraction Physics for this pioneering
ies. The intensity of the source and high collimation make work in nanosecond, time-resolved X-ray diffraction.
it possible to collect the relevant data in about 20 minutes The new pulsed spallation sources (such as the Euro-
in a sample such as Cu metal with a resolution of 1 eV pean Spallation Neutron Source at Forschungszentrum
at 8.8 keV. To collect similar data with a rotating anode Julich GmbH, or the Spallation Neutron Source at Oak
source would take about 2 weeks, and the precision would Ridge National Laboratory) provide spectra substantially
be reduced by a factor of about 100. richer in short-wavelength neutrons than those available
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 129
from the reactor sources. Pulse duration and repetition are globin, for example, at 1.8 A resolution, the negative den-
source parameters, but the time structure can be exploited sity of the protons (which form about half of the scattering
by a variety of techniques. Essentially the same types of material in the cell) makes the polypeptide chain stand out
experiments are done at both the reactor and spallation clearly in the Fourier maps. Soaking the same crystal in
sources. The higher neutron fluxes available with the new D2 O has allowed identification of exchangeable protons.
sources allow experiments to be done on smaller samples Neutron- and X-ray-scattering experiments using non-
and/or in shorter times than was previously possible. Bragg scattering can be used to study the size, shape, and
aggregation of micelles. Contrast variation can be used to
8. Contribution of Diffraction to Molecular Biology study the internal structure of viruses.
The average scattering densities from the protein coat
Molecular biologists seek to unravel the mysteries of the
and the RNA interior of viruses are different. Each can be
cell by mapping gene location, function, and control. By
matched by a different H2 O/D2 O ratio. Thus, for virus
understanding all of these we gain insight into how and
particles in solution, the matched phase becomes invisi-
when genes are turned on and how these might be used
ble to the neutron beam and allows the radial distribution
to perform useful tasks. One goal of such studies is to al-
of scattering of the other component to be recorded. For
ter (i.e., reengineer) a natural biologic process to perform
spherical viruses, this allows measurement of the thick-
some other function. Examples of this include attempts
ness of the protein coat and the degree of interpenetra-
to modify metal-binding proteins such that a protein that
tion of the two phases. Enzymes and other proteins bind
was originally selective for calcium is selective for zinc
some substrates so well that detailed, atomic-level anal-
or copper. This reengineered metal-binding protein could
yses of the structure of the native protein can be com-
then be used to construct a sensor for zinc or copper. Other
pared with those of proteinsubstrate complexes, protein
examples include modifying enzymes for use in industrial
inhibitor complexes, and proteins with catalytic groups
processes rather than biologic processes. In order to ef-
bound to allosteric sites. Detailed comparisons between
fect these changes, the relationship between structure and
such structures have greatly enhanced our understanding
function must be understood.
of the mechanisms of biological catalysis.
The structure of a macromolecule can be known at
Molecular graphics programs are now available that can
a number of levels. Primary structure is the linear se-
display the full three-dimensional structure of a macro-
quence of building blocks (i.e., amino acids) from which
molecule and zoom in on any portion of it. It is possible
the protein is built. For example, the chain of human
to examine the active site and to attempt to fit known sub-
hemoglobin contains 144 amino acids, of which the first
strates and/or inhibitors into that site. The availability of
five have the sequence valine, histamine, leucine, threo-
coordinates for macromolecules in the Protein Data Bank
nine, and proline. The term secondary structure refers
(see Section VI.A.8) has allowed many fruitful applica-
to local interactions that determine the conformation of
tions of this type.
the polypeptide chain and the interchain hydrogen bond-
ing scheme. Extended chain, -helix, and -sheet
are terms used to describe secondary structure. Tertiary
III. STRUCTURE OF A CRYSTAL
structure is the three-dimensional arrangement of atoms
within the macromolecule, while quaternary structure de-
A. Choice of Unit Cell
scribes the arrangement and interaction of aggregates of
the macromolecules themselves. Diffraction techniques A crystal is a multifaceted solid, similar in appearance to
are widely used to study the secondary, tertiary, and some- an unpolished gemstone. Internally it consists of a basic
times quaternary structures of macromolecules. pattern, known as the repeating unit, of molecules that
Single crystals of macromolecules may be studied by repeats itself by translation, in three dimensions, to the
X-ray diffraction. A wide variety of techniques are edges of the crystal. By choosing one corner of the repeat-
now available for macromolecular structure solution. In ing unit to be the origin, one can use three translational
the past, heavy-atom multiple isomorphous replacement vectors, having both length and direction, to construct a
(MIR) was the most common method of structure solution; parallelepiped that contains the entire basic pattern. This
however, anomalous dispersion is becoming more com- parallelepiped is defined as the unit cell and one need only
mon as a method of structure solution. The relatively small determine the contents of the unit cell, the basic pattern,
protein, crambin, was an early example of using anoma- to know the structure of the entire crystal. The general
lous dispersion to effect structure solution of a macro- symbols for the unit-cell vectors are a, b, c and for their
molecule. In a few cases, neutron-diffraction studies have magnitudes a, b, c. The coordinate axes, or directions of
been carried out on single crystals of proteins. In myo- the sides of the unit cell, are referred to in general as the
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
130 Crystallography
Crystallography 131
+
fj = V j (uvw)
exp[2i(hu + kv + lw)] du dv dw
where j (uvw) is the electron density of the jth atom re-
ferred to x j , y j , z j as origin. Thus, if we know the position
of an atom, then we can calculate the phases. Conversely,
if we know the phases, then we can calculate the electron
density and hence the positions of the atoms. The central
problem in crystallography is that the phases are not ob-
served in the diffraction experiment. Solving a structure
consists of finding positions for the atoms or of finding
FIGURE 3 Braggs law. The distance AB = BC = d sin . Con- phases for the structure amplitudes. Details of the major
structive interference occurs if the path difference is a whole num- methods for structure solution are found in Appendix II.
ber of wavelengths. Thus, n = 2d sin .
The amplitude of the wave scattered by the plane hkl is IV. STEPS IN CRYSTAL
1 1 1 STRUCTURE ANALYSIS
Fhkl = V (x yz)
0 0 0
The steps outlined here would be typical of those used for
exp[2i(hx + ky + lz)] d x d y dz a moderately complex structure. Notes are also included
on alternative techniques that may be more applicable to
where (x yz) is the electron density at the point x, y, z
macromolecular crystallography.
in the unit cell. The quantity (hx + ky + lz) is the vector
product (H r).
By the properties of Fourier series, Fhkl is a Fourier A. Growing Crystals
coefficient of (x yz) so that:
All strategies for the growth of crystals for diffraction ex-
1 periments are aimed at bringing a concentrated solution of
(x yz) = Fhkl exp[2i(H r)] a homogeneous population of molecules very slowly to-
V h k l
ward a state of minimum solubility. The goal is to achieve
where the summations in h, k, and l each run from a limited degree of supersaturation, from which the system
to +. Note the change in sign of the exponent between can relax by formation of a crystalline precipitate. Many
the two expressions. Fhkl is the Fourier transform of the techniques developed for achieving these ends have been
electron density in the cell. The electron density is the described.
inverse Fourier transform of structure factors. Crystallization techniques used in routine synthetic
If the electron density is a superposition of N atomic methods tend not to produce crystals of the quality re-
densities, then the structure factor expression can be quired for structural work. Suitable crystals must be grown
rewritten as: slowly at near-equilibrium conditions. This implies low
N supersaturation ratios and small gradients. Generally, su-
Fhkl = f j exp[2i(hx j + ky j + lz j )] persaturation is achieved by changing the composition of
j=1 a solution containing the sample to be analyzed (and pos-
The summation is over all atoms in the cell and the scat- sibly other additives) or by altering the temperature. In
tering factor of the jth atom is either case, the concentration of the sample is driven be-
yond its saturation limit, whereby the sample is forced out
of solution and crystal formation may result.
When crystals are grown from solution, changing the
solvent may have a pronounced effect on their habit and
size. Properties that may influence crystal growth, such
as density, viscosity, dielectric constant, and solubility,
may be varied over a wide range by mixing two or more
solvents.
FIGURE 4 Millers index. The plane has intercepts a/3, b/4, c/2. A small surface-to-volume ratio is useful for slow evap-
The Millers index is 342. oration of solvent. For small samples, an NMR (nuclear
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
132 Crystallography
magnetic resonance) tube may be a suitable crystallization with no satellite growths or twinning, and on the order
vessel. of 0.1 mm on each edge. Fine-tuning conditions, adding
Very insoluble compounds may be crystallized using detergents or counterions to the precipitant, and seed-
a method known as reactant diffusion. In this method, ing droplets with microcrystals are all techniques used
reactants A and B are allowed to mix by diffusion; the to prepare large, single crystals from initial successful
very insoluble product C will crystallize in the zone of experiments.
mixing. Sparingly soluble compounds can sometimes be
crystallized from boiling solvent in a soxhlet extractor.
B. Microscopic Examination
Crude product is placed in the thimble and the reservoir
is seeded. The crystal chosen for analysis must have a uniform in-
Sublimation is effective for some classes of compounds. ternal structure and be of an appropriate size and shape.
Vacuum sublimation reduces the temperature required and The first criterion implies that the substance is purethe
so increases the range of compounds for which it is suit- crystal contains no voids or inclusions and is not bent,
able. If the temperature gradient is too large, only micro- cracked, distorted, or composed of crystallites. It must be
crystals will be formed. Large crystals grow at the expense a single crystal, and ideally it should not be twinned (a
of small ones only when the process is carried out slowly. twinned crystal has two or more different orientations of
Vapor diffusion is a method that works well with mil- the lattice growing together).
ligram quantities. The solute is dissolved in a solvent in The size of crystal required is determined by the con-
which it is relatively soluble. A small container of this so- ditions of the experiment. For X-ray diffraction, approx-
lution is placed inside a closed beaker with a second sol- imately 0.1 mm is preferred; for neutrons, an order of
vent in which the solute is only sparingly soluble. The two magnitude larger is appropriate. A crystal with roughly
solvents must be miscible in one another and the second equal dimensions and well-defined edges is ideal; how-
solvent should be the more volatile. Suitable solvent pairs ever, many crystals grow as plates or needles. Although a
include ethanol/ether, benzene/ligroin, and water/ethanol. crystal with a highly asymmetric shape is far from ideal,
Diffusion of one liquid into another is also effective. The useful structural information can often be obtained from
solute is dissolved in the solvent in which it is more such a crystal. In some cases, a more uniformly shaped
soluble. Crystals form at the interface between the two fragment may be cut from a larger crystal with a sharp
solvents. razor blade.
A form of vapor diffusion is commonly used for the Examination with a binocular microscope allows a rapid
growth of protein crystals. For proteins (and other macro- screening of crystals. A few that appear suitable should
molecules) the solution properties modified to achieve su- then be examined more closely through crossed polariz-
persaturation include increasing the concentration of an ers. As the crystals are rotated, they should either appear
additive (e.g., a precipitant), decreasing total solution vol- uniformly dark in all orientations or they should be bright
ume, changing the solution pH, and/or changing the tem- and extinguish (appear uniformly dark) every 90 of ro-
perature. Although a wide variety of experimental setups tation. An unsuitable crystal may show dark and light re-
have been used in protein crystallization, the most com- gions simultaneously, or regions that do not extinguish, or
mon technique is hanging-drop microvapor diffusion different regions that display different colors.
(HDMVD). In HDMVD, a droplet of 4 to 20 l con-
taining protein and precipitant is suspended from a glass
C. Mounting a Crystal
coverslip which is sealed above a reservoir of a solution
at a higher precipitant concentration. Since the droplet is The usual way of mounting a crystal for X-ray diffraction
at a lower precipitant concentration than the reservoir, the is to glue it to the end of a glass fiber that is mounted in a
net migration of water vapor occurs from the droplet to brass or aluminum pin. For photographic work, it is neces-
the reservoir, resulting in a decrease in drop volume. The sary to align a real lattice vector (Weissenberg technique)
decrease in drop volume results in increased precipitant or a reciprocal lattice vector (precession technique) so that
and protein concentrations, which should drive the protein it is perpendicular to the X-ray beam. On a diffractometer,
out of solution. Crystals form if conditions are favorable. it is necessary for the crystal to be well centered in the
More often, protein precipitates as an amorphous solid X-ray beam, but alignment is neither necessary nor advis-
along the bottom of the droplet. able since the aligned position is the one that often leads
Once initial conditions are found where crystals have to overlapping reflections.
formed, additional experiments often must be performed Materials that are air or moisture sensitive may be sta-
to perfect crystal growth in order to produce X-ray- bilized by covering the crystal with a light coating of oil or
diffraction-quality crystals. The crystals must be single, the mounting glue (provided that the crystal is not soluble
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 133
F. Data Collection
It is always advisable to collect symmetry-equivalent
data (reflection intensities). The degree of agreement be-
tween equivalent reflections allows assessment of crys-
tal quality, absolute configuration, stability of the count-
ing chain, suitability of absorption correction, and other
systematic effects. For charge density studies, a com-
plete dataset should be collected to a resolution of at
least sin / = 1.3 A 1 with an internal agreement of
about 2%. For determination of absolute configuration, ef-
fects of decomposition, absorption, and extinction errors
are minimized if hkl and hkl are measured consecutively
at +2 and 2 . This may be done for a few dozen of
the most sensitive reflections or for the whole dataset. For
FIGURE 5 Polaroid rotation photograph.
large datasets, it is common practice to collect data in
shells.
in the solvent of the glue). Very sensitive substances may
have to be mounted in capillary tubes under an inert atmo- G. Data Reduction
sphere, with all mounting operations carried out in a dry The process of deriving structure amplitudes |F hkl | from
box. the observed intensities I hkl is known as data reduction. A
number of geometrical factors influence the intensities ob-
D. Preliminary Photographs served in a diffraction experiment. The most important of
Most diffractometers are equipped with a Polaroid cam- these are Lorentz, polarization, absorption, and extinction
era for a quick check of crystal quality. Alternatively a corrections.
rotation photograph can be acquired with an area detector. The first three corrections are normally applied to the
The spots on the photograph should have similar shapes, observed intensities in the process of calculating the struc-
without tails or streaks (see Fig. 5). A more sensitive check ture amplitudes, |Fhkl |. The Lorentz factor corrects for
of crystal quality is provided by omega scans of a num- the relative speeds with which different reflections pass
ber of reflections measured with different orientations of through the reflecting position. Since X-rays are polar-
the crystal. If the diffraction pattern falls off rapidly with ized on reflection, and the degree of polarization depends
angle, low-temperature data collection is advisable; how- on experimental conditions, a correction must be applied
ever, some crystals crack when cooled. If suitable crystals to account for how the polarization affects the observed
are scarce, it may be advisable to collect a dataset at room intensities. Frequently, Lorentz and polarization correc-
temperature and then cool the sample to improve the res- tions are combined. Absorption depends on the average
olution of the experiment path-length of both the incident and the reflected beam
and hence may be very different for symmetry-related
reflections.
E. Establishing the Orientation Matrix
Extinction is an interference process. An extinction cor-
The orientation matrix and the cell dimensions are estab- rection is generally made during least-squares refinement,
lished by determining the setting angles of about eight making it a correction applied to the model rather than to
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
134 Crystallography
the observations. A detailed discussion of the absorption, Advances in the phasing of macromolecular data have
extinction, and other factors affecting diffraction intensity been made by the use of a phenomenon called anomalous
can be found in Appendix I. dispersion or anomalous scattering. Single isomorphous
replacement with anomalous scattering (SIRAS) takes ad-
vantage of both the phasing power and anomalous scat-
H. Solving the Phase Problem
tering properties of certain heavy atoms. An advantage of
The central problem in crystallography arises because the the SIRAS technique is that data, in some cases, can be
experimental data yield only the modulus of the structure collected using a conventional Cu-K radiation source.
factor, |F hkl | and not the phase. The phase is required in The SIRAS approach to phasing obviates the need for
order to evaluate the electron density in the unit cell, but crystallization and data collection from multiple samples.
it cannot be measured directly. The technique does require careful data collection but pro-
Several methods have been developed to determine duces all of the information needed to determine phases
the phases of the complex structure factors, |F hkl |, with from a single dataset collected from only one crystal.
no prior knowledge of atomic positions. These meth- A crystal containing an anomalous scattering atom may
ods include multiple isomorphous replacement, single- be used to collect data at multiple wavelengths and allow
isomorphous replacement with anomalous dispersion, the phases to be determined using multiple-wavelength
multiple-wavelength and single-wavelength anomalous anomalous dispersion (MAD). A single crystal can be
dispersion, heavy-atom, and direct methods. With the ex- analyzed at multiple wavelengths, generating a varia-
ception of direct methods, all of these methods take advan- tion in scattering factors that allows direct determination
tage of the scattering properties of heavy atoms (transi- of crystal structures. With careful measurements, even
tion, actinides, and lanthanides). For very large molecules, weak signals from a single crystal can provide the nec-
such as proteins, neither direct nor heavy-atom methods essary phasing information. Data must be collected using
are generally used. Molecular replacement (MR) also pro- tunable X-ray radiation available only at synchrotron
vides a powerful phasing method for the structure analysis. facilities.
A model structure similar to the structure being analyzed A paper by J. Karle (see Bibliography) offers the
is required for MR. possibility of nearly direct phasing for protein crystals.
If the positions of the atoms are known, both the mag- The potential for analysis of the data using this single-
nitude and phase of the structure factor can be calculated. wavelength anomalous dispersion (SAD) technique is yet
The heavy-atom method of structure solution depends on to be explored. Given the collection of sufficiently accurate
this. If at least one atom in the structure is heavy enough anomalous dispersion data, one dataset at one wavelength
to be located in a Patterson function, that position can may provide all the information required to determine the
be used to calculate phase angles. A Fourier summation phases.
using observed structure amplitudes and these calculated Direct methods of phase determination do not depend
phases would reveal the heavy atom and some others. The on any a priori structural information. Phases are deter-
additional atoms are included in the structure factor cal- mined from statistical relationships among the intensities.
culation, providing a better estimate of the phase angles. Powerful computer programs have facilitated the use of
This process can be facilitated by use of the tangent for- these methods for centrosymmetric and noncentrosym-
mula to extend and refine the new phases and is repeated metric structures of moderate size. Considerable effort has
until all of the atoms are located. been made to extend these methods to protein structures,
Multiple isomorphous replacement (MIR) is a powerful and structures containing over 1000 non-hydrogen atoms
method for determination of phases. MIR depends upon have been solved by direct methods.
the phasing power of heavy-metal atoms bound to a com- In general, the conformationally rigid portions of the
pound in such a fashion that the positions of other atoms molecule are most easily located. Once the major features
in the crystal are minimally perturbed (i.e., isomorphous of the molecule have been recognized, the difference map
derivatives). The MIR method requires that the crystallog- provides a powerful means both for checking the accu-
rapher prepare more than one derivative of the parent crys- racy of the partial model and for completing it. A Fourier
tal. These often turn out to be unstable, not isomorphous, summation using F = |F obs F calc | as coefficients and
or to have the metal bound with too low an occupancy phases calculated from the known portion of the struc-
to be useful. Even when successful, this method requires ture yields a map of the discrepancies between the crystal
the collection and reduction of numerous datasets from and the model. An atom present in the crystal but not
multiple crystals. Anomalous dispersion methods can be included in the model will appear as a peak. Difference
combined with isomorphous replacement to circumvent maps are also useful for locating light atoms in the pres-
some of the disadvantages. ence of heavy ones, such as the location of hydrogen atoms
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 135
in compounds of first-row elements or carbon atoms in a trosymmetric crystal is treated as a twin by inversion and
tungsten compound. the contribution of the two components evaluated during
refinement as the Flack parameter. In the case where the
arrangement of atoms in the model and crystal are in agree-
I. Refinement
ment, the contribution of the Flack parameter is zero. If
Once approximate locations have been determined for all, the model and crystal are inverted with respect to each
or almost all, of the atoms in the structure, it must be re- other, the Flack parameter is one and the model needs to
fined or made more precise. The standard method for re- be inverted.
finement of structures, known at atomic resolution (not for
macromolecules), is full-matrix least squares. For struc-
tures that are not known at atomic resolution, or for very K. Derived Parameters
large structures, sparse-matrix or simulated annealing is The parameters produced directly by the least-squares re-
used for refinement. Since the structure factors are not finement are the positions of the atoms and their thermal
linear functions of the parameters, the process is an iter- parameters. Bond lengths, bond angles, and torsional an-
ative one. Hydrogen atoms are not generally included in gles are derived from these positions. An examination of
the early stages of refinement but are included in the final short inter- and intramolecular contacts may provide in-
cycles of refinement. formation about hydrogen bonding, van der Waals forces,
The function minimized is 2 , where is the packing forces, etc.
weight assigned to a particular observation and is the Room-temperature studies of organic compounds gen-
difference between the observed and calculated values of erally show appreciable thermal motion with Ui j values
F hkl (for a refinement based on |F|) or F 2hkl (for a refine- of the order of 0.04 A 2 . This corresponds to a root-mean-
ment based on |F|2 ). A convenient parameter, referred to square vibration amplitude of 0.2 A and is a reminder that
as the R-factor, for monitoring the progress of the refine- some caution is required in comparing bond lengths from
ment is diffraction experiments with those determined by spec-
R= (||Fo | |Fc ||) |Fc | troscopic and theoretical work. Bond lengths can some-
times be corrected for thermal motion, but it is generally
Correct structures generally have R-values under 0.10, and preferable to reduce thermal motion by cooling the crystal.
those that are well behaved are frequently under 0.05. Many investigators routinely collect data at temperatures
in the range of 100 to 200 K.
J. Determination of Absolute Configuration A single crystal structure determination provides
valuable information on chemical connectivity, relative
If a particular torsional angle has a positive sign in the conformation, and, under the proper experimental condi-
right-handed enantiomer, it will have a negative sign in tions, absolute configuration. However, an understanding
the left-handed molecule. Thus, determination of the ab- of structurefunction relationships requires correlating
solute configuration in a chiral molecule can be regarded features from a number of different structures. The exis-
as the determination of the correct signs for the torsional tence of computer-searchable databases of structural data
angles. A torsional angle ABCD is positive if a clockwise greatly enhances the possibilities for such comparisons
rotation will cause the bond AB to eclipse the bond CD. (see Section VI).
Determination of absolute configuration by X-ray crys-
tallography requires a structural study in the presence of
dispersive scatterers. Thus, if at least one atom in the struc-
V. COMPARISON OF X-RAY AND
ture is an anomalous (i.e., dispersive) scatterer, Friedels
NEUTRON DIFFRACTION
law breaks down and reflections from two sides of the
same plane are no longer equal. The differences in in-
For a structure with N atoms, each with atomic scattering
tensity are generally small, so careful measurement is re-
amplitude f i and position ri in the unit cell, the structure
quired. Coster et al. demonstrated that a noncentrosym-
factor for the Bragg reflection of index h is
metric crystal structure could be distinguished from its
inverted image using these differences. Later, Bijvoet re-
N
alized this principle was more general and used it to de- F(h) = f i (h)Ti (h) exp(2ih ri )
i=1
termine the absolute configuration of the sodium rubid-
ium salt of (+)-tartaric acid. In 1983, Flack developed a where Ti is the temperature factor of the ith atom. Differ-
method for distinguishing a noncentrosymmetric structure ences between neutron diffraction and X-ray diffraction lie
from its inverse. Using the method of Flack, any noncen- in the scattering amplitudes f i (h) and in the temperature
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
136 Crystallography
parameters Ti (h). In X-ray diffraction, X-rays are scattered periment, positions of hydrogen atoms will be more poorly
by the electrons of an atom, and the scattering factors f i (h) determined than positions of carbon atoms. In the neutron
are strongly dependent on scattering angle. At a scattering experiment on the same compound, hydrogen would con-
angle of 0 , f i is proportional to Z , the atomic number of tribute 24% and carbon 76%. (Because of the negative
the scattering atom. As the scattering angle increases, the sign on the scattering length of hydrogen, a Fourier sum-
scattering factor f i (h) decreases. Thermal motion causes mation will show holes rather than peaks at hydrogen
the scattering to fall off even more strongly. These factors positions.) For deuterobenzene, the contributions from C
limit the resolution available in an experiment and make and D (in the neutron-scattering experiment) are virtually
it difficult to determine the positions of light atoms ac- identical, with each contributing 50% of the scattering.
curately in the presence of heavy ones or to distinguish Thus, neutron diffraction can locate hydrogen and deu-
among heavy atoms that have similar atomic numbers. terium atoms with the same precision as carbon, nitrogen,
In neutron diffraction, the scattering is primarily from and oxygen, while X-rays cannot. For this reason, stud-
the nucleus. Since the diameter of the nucleus is small ies of hydrogen bonding in biologically significant com-
relative to the wavelength of thermal neutrons, the scat- pounds such as amino acids and sugars were among the
tering factor is a constant, characteristic of the particular early experiments using neutron diffraction.
nucleus and independent of scattering angle. There is no As the precision of the experiments and the sophistica-
simple relationship between the scattering amplitudes and tion of the refinements improved, it became obvious that
the nuclear mass or charge. Nuclei with similar atomic there were systematic differences in the positional and
number can have significantly different scattering ampli- thermal parameters from the two experiments that were
tudes. Hence, neutron diffraction can distinguish among much larger than expected from the estimated standard
near neighbors in the periodic table and so is useful in the deviations. The differences are very pronounced for hy-
study of alloys. drogen atoms. Even in the most precise, low-temperature
Table II shows the relative scattering lengths of a num- studies, electron density maxima for hydrogen atoms were
ber of atoms for X-rays and neutrons. If the average con- displaced by as much as 0.2 A from the positions of pro-
tribution to the intensity of a structure factor is tons determined from neutron diffraction. Since the CH
and OH bond lengths in the neutron experiment agree
n
P(i) = f i2
f i2 with spectroscopic measurements, it was recognized that
i=1 the apparent shortening of the bonds to hydrogen observed
in X-ray experiments is a bonding effect. The position of
Then, for a compound such as benzene with equal numbers
maximum electron density does not coincide with the po-
of hydrogen and carbon atoms, carbon will contribute 97%
sition of the nucleus because the formation of the cova-
and hydrogen will contribute 3%. Thus, in the X-ray ex-
lent bond perturbs the electron-density distribution in the
atom.
TABLE II Scattering Amplitudes for Selected Similar, but smaller, effects are observed for first-row
Elements (1012 cm)a atoms, C, N, O, etc. Typically, the discrepancy is of the
order of 0.01 A, but it depends critically on the range
Scattering amplitude
Atomic of Bragg angles included in the X-ray refinement. If the
Element number X-ray Neutron refinement is based on very-high-order data [(sin )/ >
1.00 A 1 ], the discrepancies will be much reduced; how-
1H 1 0.28 0.37
2D 1 0.28 0.67 ever, few organic compounds scatter to such high angles
12 C 6 1.69 0.66 even at liquid-nitrogen temperature.
14 N 7 1.97 0.94 Systematic differences in temperature factors also re-
16 O 8 2.26 0.58 flect the very real differences in the scattering processes
32 S 16 4.51 0.28 in the two experiments. In aromatic molecules, thermal
Clb 17 4.79 0.96 vibration parameters are greater for X-rays than for neu-
Brb 35 9.87 0.68 trons in the plane of the ring and smaller perpendicular to
127 I 53 14.95 0.52 the ring, implying that electron density is smeared in the
238 U 92 25.92 0.84 plane of the ring by covalent bonding and contracted in
the perpendicular direction.
a For X-rays, the scattering amplitude at 0 is given by
Atoms with lone pairs of electrons show significant
(e2 /mc2 ) f 0 or ((0.282 1012 cm) atomic number).
differences in both positional and thermal parameters in
Values for neutrons are taken from International Tables
for X-Ray Crystallography, Vol. IV, 1974, pp. 270271. the two experiments because the lone-pair density is not
b These values are for the elements in their natural centered on the nucleus. These very real differences sug-
isotopic abundances. gest strongly that neutron diffraction is the preferred tool
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 137
for determining atomic parameters, whereas the X-rays drugs, the phenomenon of one-dimensional conduction, or
measure the electron density in the solid. The combina- any other structureproperty relationship) requires the de-
tion of the two techniques provides a means of studying tailed analysis of a large number of structures and correla-
bonding effects (this topic was covered in greater detail in tion with results from other disciplines. Databases provide
Section II.B.3). the kind of information required for structurefunction
studies and are very valuable to the scientific community.
The topics presented here are intended to be illustrative of
VI. RESULTS the many ways that the databases have been used. There
are many other uses not documented here.
The determination of a single crystal structure may answer
A. Crystallographic Databases
a question about the connectivity of a molecule or some
detail of its conformation. The direct results of structure At present there are eight major databases for crystal-
analysis are the positional, thermal, and occupancy pa- lographic results continuously maintained and updated
rameters of atoms in the asymmetric unit. Bond lengths in different laboratories in Europe and North America.
and angles, torsion angles, and intermolecular associations In each case the data are available in machine-readable
(such as hydrogen bonding) are all derived from these ba- form, and considerable effort has been expended to de-
sic structural parameters. While the final R-value that a velop efficient computational algorithms for searching the
structure refines to is often a good indicator of the quality files and correlating the data. Perhaps the most impor-
of a structural determination, a plot showing the thermal tant feature of the databases is that the data have been
ellipsoids (Fig. 6) can also give an indication of the quality checked and errors are corrected when possible or flagged
of a structural determination. Errors in the structure deter- if uncorrected. Some of these databases have been in-
mination and thermal motion in the molecule can distort corporated into commercial packages. Some are avail-
these ellipsoids. able in both printed and computer-readable form. In the
Understanding a complex process (such as the mech- brief summary that follows, they are listed in alphabetical
anism of a reaction, the biological activity of a class of order.
FIGURE 6 Thermal ellipsoid plot from a crystallographic structure determination. Hydrogen atoms are plotted as
small balls with an arbitrary radius; non-hydrogen atoms are plotted as ellipsoids, with the axes corresponding to the
thermal parameters for the atom. Notice that the ellipsoids gradually increase in size and become more asymmetric
moving out the chain from C8 to C16. The ellipsoids can become large and more asymmetric for noncovalently bounds
groups such as the coordinated nitro group (N2 , O1 , O2 , and O3 ). In both of these cases, the axes of the ellipsoids
are aligned with expected vibrational modes and do not indicate any major problems with the structure. If the axes
were poorly aligned and/or the ellipsoids more asymmetric, this could be an indication of problems with the structural
determination.
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
138 Crystallography
1. Cambridge Structural Database The data stored include chemical name, chemical for-
mula, density, lattice parameters, space groups, atomic
(http://www.ccdc.cam.ac.uk/prods/csd/csd.html; Cam-
coordinates, oxidation state, temperature factors, remarks
bridge Crystallographic Data Center (CCDC), University
regarding conditions of measurement, R-values, and
Chemical Laboratory, Lensfield Road, Cambridge, CB2
bibliographical references. Online access to the file is
1EW, U.K.) The Cambridge Structure Database (CSD)
available. It is also possible to lease the entire database.
is the largest database of experimentally determined or-
The database now contains over 53,000 entries and is
ganic and metallo-organic crystal structures in the world;
updated twice a year.
inorganic carbon compounds (such as carbonates and
cyanides) are excluded. The CCDC also provides a suite
of graphical search, retrieval, data manipulation, and vi- 5. Metals Data File
sualization software for use with the database. The CSD
(Manager, CAN/SND, Canada Institute for Scientific and
contains bibliographic data, tables of connectivity, atomic
Technical Information, National Research Council (NRC)
positions, cell dimensions, and quality indicators for vir-
of Canada, Ottawa, Canada K1A OS2.) The Metals file
tually all three-dimensional structures of organic com-
contains structural data for metal and alloy structures de-
pounds published since 1935. The CSD currently contains
termined since 1913 based on either powder or single-
over 210,000 entries and is growing at a rate of approx-
crystal diffraction. Under an exclusive license from the
imately 15,000 entries per year. The CSD is available to
NRC, Toth Information Systems has maintained and up-
scientists throughout the world.
dated the database. If available, the following information
is included: formula, cell dimensions, structure type, Pear-
2. Crystal Data son symbol, atomic coordinate, temperature parameters,
(http://www.nist.gov/srd/nist3.htm; Crystallographic Sec- occupancy factors, R-values, method of refinement, in-
tion, National Bureau of Standards, Washington, D.C. strument used, radiation, and bibliographical information.
20234.) This database contains lattice parameters for all New software for manipulating the file has been developed
crystals whose dimensions have been reported by X-ray, by Toth Information Systems.
neutron, or electron diffraction on single crystals or fully
indexed powders. Data include name, formula, cell di- 6. Nucleic Acid Database
mensions, space groups, number of molecules in the cell,
density (measured and calculated), bibliographical data, (http://ndb-mirror-2.rutgers.edu/NDB/ndb.html; Dr. H. M.
crystal habit, melting point, etc. Crystal Data accepts data Berman, Department of Chemistry, Rutgers University,
from the other databases. 610 Taylor Road, Piscataway, NJ 08854-8087.) The goal
of the Nucleic Acid Database (NDB) project is to assem-
ble and distribute structural information about nucleic
3. Electron Density Data Base acids. Online access to the NDB is freely available. A
(Prof. H. Burzlaff, Lehrstuhl fur Kristallographie, Institut variety of tools have been developed in conjunction with
fur Angewandte Physik der Universitat, Bismarckstrasse the NDB to provide a robust user interface.
10, D-91054 Erlangen, Germany.) This database contains
accurate structure factors for crystal structures whose elec- 7. Powder Diffraction File
tron densities have been carefully determined. This is the
type of data required for studying bonding effects, co- (http://www.icdd.com/ productsold/pdf2.htm; Interna-
valency in organometallics, and other details of electron tional Center for Diffraction Data (ICDD), 1601 Park
distribution. Lane, Swarthmore, PA 19081.) The Powder Diffraction
File (PDF) is a compilation of powder diffraction pat-
terns produced and published by the Joint Committee on
4. Inorganic Crystal Structure Database
Powder Diffraction Standards of the ICDD. The PDF is
(http://crystal.fiz-karlsruhe.de/portal/cryst/ab icsd.html.) the worlds largest and most complete collection of X-ray
The Inorganic Crystal Structure Database (ICSD) was powder diffraction patterns. The 1999 release of the PDF
initiated in 1978 at the Institute for Inorganic Chemistry contains about 115,000 patterns, of which 20,000 are or-
at the University of Bonn. Today the database is produced ganic and 95,000 inorganic. Included in the database are
by FIZ Karlsruhe (P.O. Box 2465, D-76012 Karlsruhe, calculated powder patterns for almost 40,000 compounds.
Germany) in cooperation with NIST. Inorganic is The PDF is used for identification of crystalline materi-
defined to exclude metals, alloys, and compounds with als by matching d-spacings and diffraction intensity mea-
CH and CC bonds (with the exception of graphites). surements. Each pattern includes a table of interplanar
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 139
140 Crystallography
acetals, the geometry of nearly 2400 ethers and esters was Examples of this include mapping of the methionine bind-
investigated. The results show clearly that there is sub- ing site in the enzyme S-adenosyl ATP transferase by
stantial, systematic variation in CO bond lengths. The Marshall and colleagues, and the postulating of a com-
fragment of interest is defined as R1 OR2 . Data are di- mon site of action for gamma-butyrolactone analogs and
vided into four categories, depending on whether R1 is picrotoxinin. On the basis of this model, it has been pro-
methyl, primary, secondary, or tertiary, and each of those posed that the convulsant activity of these compounds is
categories is then divided according to the effective elec- related to their ability to block the passage of chloride ions
tronegativity of the group R2 . The four categories of R2 through channels. The model would appear to be applica-
are alkyl, aryl, enol/ether, and acyl/carboxylic ester, giving ble to a number of convulsant and anticonvulsant drugs.
16 categories in all. The shortest CO bonds (1.418 A) are
found in compounds where R1 is methyl and R2 is alkyl.
E. Crystallography and Molecular Mechanics
Within each of the four categories of R1 , the length of the
R1 O bond increases with increasing electronegativity of Molecular mechanics is an empirical method for calcula-
the R2 group. The effective electronegativity of a group tion of properties of molecules such as molecular geome-
can be estimated from the pK a of its conjugate acid. In the try, heat of formation, strain energy, dipole moment, and
subset of 2-(aryloxy)-tetrahydropyrans, there is a linear re- vibrational frequencies. Different programs use different
lationship between the length of the exocyclic CO bond parameter sets and reproduce these physical properties
and the pK a of the leaving group. This implies a linear with different degrees of fidelity. Parameters are assumed
relationship between the bond length and the free energy to be transferable from one type of molecule to another.
of activation for the hydrolysis reaction or any other reac- The geometrical parameters used in molecular-
tion in which the CO bond is cleaved. The consequences mechanics programs are frequently derived from crys-
of these generalizations to the chemistry and reactivities tallographic data. However, the predictive value of the
of acetals and glycosides are fully explored in these pa- method is limited by the datasets used to derive the pa-
pers. In the third paper in the series, the authors plot the rameters. For instance, a parameter set derived from un-
reaction coordinate for six aryl tetrahydropyranyl acetals. crowded hydrocarbons is not likely to predict structures of
Both relative free energy and pK a are plotted as a func- crowded hydrocarbons with satisfactory accuracy. Tran-
tion of the bond length. The Morse function is also plotted, sition states, small rings, unusual states of hybridization,
as is a reaction coordinateenergy contour diagram. This and other electronic effects may require special treatment.
work by Kirby et al. is one of the most comprehensive These caveats notwithstanding, the method has enjoyed
attempts at deriving structurereactivity relationships so considerable success.
far available. It is not uncommon for molecular-mechanics calcula-
tions to be used to provide a starting point for the re-
finement of electron diffraction data. A recent example
D. Drug Design
is provided by the study of bicyclo[3.2.0]heptane. Many
The coupling of the resources of databases with molec- other examples are available in the literature.
ular graphics devices and QSAR (quantitative structure Molecular mechanics can also be useful in interpre-
activity relationships) techniques raises a tantalizing pos- tation of crystal structures, particularly in differentiating
sibility that therapeutically useful new structures might be electronic and steric effects or in estimating the effects
predictable. Detailed comparison of structures of known of packing forces. In the structure of lepidoptrene, a CC
agonists and antagonists is the major strategy currently bond was observed to have a length of 1.64 A in the X-ray
in use. Similarities in the three-dimensional structure of experiment, while molecular mechanics predicted a length
portions of the molecules allow identification of the fea- of 1.57 A. Mislow and colleagues demonstrated that the
tures required for binding. Differences in other portions additional lengthening of the bond beyond the amount ex-
of the structure may account for the agonist/antagonist re- pected from steric strain was caused by through-bond
sponse after binding. Once the pharmacophore has been coupling of adjacent systems.
identified, binding studies can be combined with structural Although molecular-mechanics calculations can be use-
comparisons to map the receptor site. ful in assessing which aspects of a structure control con-
A sum or superposition of active molecules can be used formation, it should be pointed out that conformational
to define the available volume within a receptor. Inac- parameters from molecular mechanics tend to be less
tive molecules then define excluded volumesvolumes reliable then bond lengths or bond angles. Also, confor-
occupied either by the receptor itself or by a cofactor. mational parameters in crystal structures are most influ-
Any new drug must be designed to present the correct enced by packing forces. Molecular-mechanics calcula-
pharmacophore and to occupy only the available volume. tions have been used to suggest alternative conformations
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 141
(having energy similar to that observed in the crystal) integrated path-length for incident and diffracted beams
that might exist in the gas phase or in solution. Addi- for all reflections.
tionally, molecular-mechanics has been used to optimize For crystals with no reentrant angles, the most com-
the geometry of a pharmacophore in model studies of mon methods are the analytical method and the method
drugreceptor binding, to evaluate the interaction ener- of Gaussian quadrature. In both cases, the objective is to
gies between dinucleoside monophosphates and cationic evaluate the integral,
intercalators such as ethidium bromide, and to interpret 1
conformational polymorphism. Although some caution is A= exp(T ) d
V crystal
obviously necessary, molecular mechanics and crystallog-
This can be done analytically if the crystal is divided into
raphy can provide complementary information in a variety
a number of polyhedra in each of which the path-length
of cases.
is a linear function of the coordinates. De Meulenaer and
Tompa first programmed this method in 1965. Calcula-
tion time is independent of the severity of absorption. The
APPENDIX I: FACTORS
factor-limiting accuracy in cases of severe absorption is
AFFECTING INTENSITIES
the precision of measuring the crystal, particularly in its
shortest direction.
A number of geometrical factors influence the intensities
Gaussian quadrature is a numerical integration method
observed in a diffraction experiment. The most impor-
that evaluates an integral by summing an appropriate poly-
tant of these are Lorentz, polarization, absorption, and
nomial. It uses a nonisometric grid in which the inter-
extinction corrections. The first three corrections are nor-
val is subdivided symmetrically about the midpoint, with
mally applied to the observed intensities in the process
large spacings near the middle and smaller ones toward
of calculating the structure amplitudes |Fhkl |. Many re-
the edges. This tends to put the maximum number of grid
searchers feel that the absence of this correction is the sin-
points near the surface of the crystal where the change of
gle largest source of systematic error in crystal structures
absorption with path-length is largest. The number of grid
in the current literature. Extinction is a correction made
points determines the precision of the calculation. Thus,
during least-squares refinement and so is a correction ap-
if a 4 4 8 grid gives a precision of 2% in calculated
plied to the model rather than to the observations. How-
transmission for = 2.5 cm1 , a grid of 8 8 16 is
ever, the correct formulation for extinction depends on
required to produce the same precision if = 5.0 cm1 .
both the polarization and the path-length (calculated dur-
By choosing sufficient points, the Gaussian method can
ing absorption correction) and so is discussed with these
reproduce the analytical result to any desired precision;
other correction terms.
however, for strongly absorbing crystals, the analytical
method is the method of choice.
A. Absorption When a crystal has reentrant angles, neither of these
methods can be used, and one possible resort is numeri-
The absorption of X-rays by crystals obeys the relation
cal integration with an isometric grid; however, the grid
I = I0 exp(t), where is the linear absorption coef-
must be very fine in order to achieve reasonable precision.
ficient, in units of (length)1 , and t is the thickness. The
For crystals mounted in capillaries, crystals with irregu-
general effect is to reduce the intensity of reflections at low
lar shapes, and other pathological cases, absorption may
sin . If the crystal is centrosymmetric in cross-section,
be dealt with by measurement of a transmission surface
neglect of this correction will have little effect on posi-
as proposed by Huber and Kopfmann in 1969. However,
tional parameters, although scale and thermal parameters
these problems can be dealt with by the program DIFABS,
may be very strongly affected. If the crystal cross-section
developed by Walker and Stuart in 1983, which models the
is not centrosymmetric, all structural parameters will be
absorption surface by a Fourier series in polar coordinates.
systematically wrong. Neglect of absorption is probably
The coefficients are obtained by minimizing the sum of
the largest single source of systematic error in published
squares of residuals between observed and absorption-
structures.
modified values of the structure factors. The chief virtue
Absorption corrections can be done by collecting
of the method is that it can be used even when the crystal
-scans and calculating an empirical correction or by
is no longer available
indexing and measuring crystal faces and calculating a
face-indexed absorption correction. For a face-indexed
B. Lorentz Factor
correction, the size and shape of the crystal must be pre-
cisely determined. Once the crystal shape is established, The Lorentz factor corrects for the relative speeds with
a number of techniques are available for calculating the which different reflections pass through the reflecting
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
142 Crystallography
position. The intensity of a reflection produced by a mov- parallel or perpendicular to the equatorial plane of the
ing crystal depends on the time taken for the corresponding diffractometer. If the monochromator is in the incident
reciprocal lattice point to pass through the sphere of re- beam and mounted so that its axis is perpendicular to
flection. Using 1/(S cos ) as the angular velocity, the the equatorial plane, the component that was attenuated
definition S = 1/d, and Braggs law, = 2d sin , the cor- by reflection from the monochromator is attenuated again
rection takes the form: by the sample, so that the polarization correction for the
twice-reflected beam becomes:
d/( cos ) = /(2 sin cos ) = /( sin 2)
p = 1 + cos2 2m cos2 2 1 + cos2 2m
The factor 1/(sin 2) is the Lorentz factor.
For a monochromator mounted with its axis parallel to
the equatorial plane, reflection from the monochromator
C. Polarization
attenuates the beam normal to the equatorial plane, while
X-rays are polarized on reflection, and the degree of polar- reflection from the sample attenuates the parallel compo-
ization depends on experimental conditions. Neutrons are nent, so that
not polarized on reflection from ordinary crystals. X-rays
p = cos2 2m cos2 2 1 + cos2 2m
produced from an X-ray tube or a rotating-anode gener-
ator are unpolarized; that is, all directions of the electric These formulas assume that the monochromator crystal
vector normal to the direction of propagation are equally is an ideally mosaic crystal. For a perfect or non-mosaic
represented. Thus, the beam can be regarded as composed crystal, the factor cos2 2m should be replaced by |cos m |.
of two components, one polarized parallel to the reflection In practice, the polarization ratio seldom corresponds
plane and one perpendicular. The relative intensities of the to either of these ideal values and may not even lie be-
two components are tween them. This has led some investigators to recast the
equations in the form:
I|| = I = 12 I0
p = (1 + K cos2 2 )/(1 + K )
For an ideally imperfect crystal, the intensity of radia-
tion scattered in a particular direction is proportional to for the monochromator axis perpendicular to the equato-
sin2 , where is the angle between the electric vector rial plane, and
and the direction of observation. It follows that the par-
p = (K + cos2 2 )/(1 + K )
allel component will be attenuated by reflection but the
perpendicular component will not. Thus, the relative in- for the parallel orientation, where K is the actual measured
tensities of the two components after reflection at an an- value of the polarization ratio for the monochromator in
gle will be I = 12 I0 sin2 (90 2 ) = 12 I0 cos2 2 and question. The value of K is different for different wave-
I = 12 I0 sin2 90 = 12 I0 . Thus, the beam is partially po- lengths. For routine structural work with MoK radiation,
larized at all angles and completely polarized at 2 = 90 . the error in assuming that the monochromator is an ideal
For data monochromated by means of a filter, the mosaic will generally be small. However, maximum error
polarization correction is occurs at = 45 and so is important in the case of very
precise studies that rely on high-angle data. Vincent and
p = 12 (1 + cos2 2 )
Flack have developed a method for determining K , the
Frequently, Lorentz and polarization corrections are com- polarization ratio, without special equipment.
bined to give:
L p = (1 + cos2 2 )/2 sin 2 D. Extinction
These corrections are applied to the observed intensity to As early as 1922, Darwin realized that absorption is not
derive structure amplitudes: the only effect that attenuates the X-ray beam as it passes
through the crystal. He described two phenomena, which
|Fhkl | = Ihkl /L p
he designated as primary and secondary extinction, and
In many modern diffractometers, monochromatization of showed how they could be treated mathematically.
the primary beam is achieved by Bragg reflection from Primary extinction is an interference process. If a set
a suitable crystal. Three of the most commonly used of planes is in a position to reflect, the reflected rays may
monochromator crystals are quartz (d = 3.35 A), LiF also be reflected a second time. Since there is a phase
(d = 2.01 A), and highly oriented graphite (d = 3.35 A).
change of /2 on reflection, a beam that has been reflected
The monochromator may be installed in the incident beam n times will be exactly out of phase with one that has
or the diffracted beam and may be mounted with its axis reflected (n2) times. This causes the reflected intensity
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 143
to be proportional to |F| rather than |F|2 . A crystal for It contains both a breadth parameter and a misalignment
which this is true is called an ideally perfect crystal. Such parameter. If g, then the broadening of the diffrac-
crystals are rare. It is much more common to encounter tion peak is dominated by mosaic spread, and we have a
crystals where I |F|n where I < n < 2 but nearer to 2. type I extinction. Type II extinction, when g, is less
Darwin modeled the phenomena of extinction by as- commonly encountered and corresponds to the situation
suming that crystals were made up of mosaic blocks, where the misalignment is small and the breadth of the
slightly misaligned with respect to one another. In perfect diffraction peaks is controlled by small domain size. Sep-
crystals, the blocks are assumed to be large and the mis- aration of the terms is possible if two determinations are
alignments small. Secondary extinction occurs because made on the same crystal with different wavelengths. This
the planes first encountered by the incident beam reflect is rarely done.
so strongly that deeper planes receive less radiation and Coppens and Hamilton first extended the treatment of
so reflect with less power than they otherwise would have extinction to an anisotropic model in 1970. The crystal
done. This effect is pronounced for strong reflections with is modeled as if it were composed of ellipsoidal parti-
|F/V | of the order of 0.1 1024 cm3 . cles whose misorientations follow a Gaussian probability
According to the mosaic model, the effect is expected in distribution. Since there is no need for the distribution
crystals where the mosaic blocks are small with respect to of mosaic blocks to obey any symmetry in the crystal,
the size of the crystal. A mosaic crystal in which the blocks symmetry-equivalent data are not averaged in this
are sufficiently misaligned that secondary extinction is treatment.
negligible is called an ideally imperfect crystal. Since such
crystals are seldom encountered in experiments, it is con- E. Anomalous Dispersion
venient to correct for extinction in least-squares refine-
ment. In most current programs, the correction is based When the wavelength of the incident X-ray beam is close
on Zachariasens 1967 formalism for isotropic extinction, to the absorption edge of a scattering atom, the atomic
in which the mosaic blocks are assumed spherical: scattering factor for that atom becomes complex:
f = f 0 + f + i f
2 p1 1/4
Fc = k|Fc | 1 + 2r Q 0 Tp
where k is the scale factor, Fc is the calculated value of the where f 0 is the normal scattering factor for wavelengths far
structure factor, r = [1 + (/g)2 ]1/2 where = 2t/3, t from the absorption edge, and f and i f are correction
is the mean path length in a single domain, and g is re- terms. The quantity f is usually negative, and i f
lated to the mosaic spread distribution and is frequently is always /2 radians ahead of the real part in phase.
assumed to be Gaussian. For X-rays, For structural work, the corrections are assumed to be
2 3 independent of scattering angle. In addition, f = 0 for
e FK wavelengths longer than the absorption edge.
Q0 =
mc2 V sin 2 Four aspects of anomalous dispersion important in nor-
mal structural work in crystallography are
where e2 /mc2 is the classical radius of an electron, and K
is the polarization ratio. For neutrons, 1. Determination of absolute configuration
2 2 2. Solution of the phase problem (structure solution)
F
Q0 = 3. Distinguishing among atoms of similar scattering
V sin 2
power
The term T is the mean path-length in the crystal and 4. Avoiding systematic errors in structures with polar
represents an integration of incident and diffracted beams space groups
over all diffraction paths in the crystal. This is normally
evaluated during the calculation of absorption correction. Knowledge of the absolute configuration is extremely im-
If absorption is small in the crystal, T may be arbitrar- portant in physiologically active materials, since biolog-
ily set to some value such as 0.03 cm. Finally, the term ical systems discriminate strongly between enantiomor-
pn is a polarization term generally assumed to have the phous forms of their substrates. Anomalous scattering
form 1 + cos2n 2 appropriate to filtered radiation. If a with phase change causes a breakdown in Friedels law,
monochromator is used, the appropriate form of the po- and Ihkl = Ihkl . This effect was first exploited for deter-
larization factor should be incorporated into the extinction mination of the absolute configuration of the sodium ru-
calculation. For very precise work, the actual polarization bidium salt of (+)-tartaric acid using Zr radiation ( =
ratio of the monochromator should be determined experi- whose wavelength is slightly shorter than that of
6.07 A),
mentally. For extinction correction, the term refined is r . the absorption edge of Rb ( = 6.86 A). The experiment
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
144 Crystallography
coupled with the known relationship between the stere- F. Scaling of Data
ochemistry of (+)-tartaric acid and (+)-glyceraldehyde
Virtually all methods of solving structures require a rea-
showed that Fischers arbitrary choice had been correct.
sonable estimate of the relative scale factor between the
With modern data collection techniques, the determination
observed and calculated structure factors. Most computer
of absolute configuration is relatively uncomplicated. In
programs that calculate scale and temperature factors are
favorable cases, the method can be applied to a compound
based on Wilsons 1942 equation:
with no atom heavier than oxygen when the incident ra-
diation is CuK ( = 1.54 A). Rabinovich and Hope have Ihkl
= f i2
determined the absolute sign of the torsional angles in i
the achiral compound 4, 4 -dimethylchalcone, C17 H16 O. which says that the local average value of the intensity is
These authors believe that the determination of absolute equal to the sum of the squares of the scattering factors. It is
configuration may be possible with hydrocarbons. assumed that the average is taken over a sufficiently narrow
In structure solution, anomalous scattering without range of (sin )/ so that the f values can be treated as
phase change is formally equivalent to isomorphous re- constants:
placement. An anomalous-difference Patterson function
Ihkl = k f i2 exp[2B(sin2 /2 )]
is analogous to an isomorphous-difference Patterson and
i
so contains peaks only for vectors between anomalously
scattering atoms and vectors between anomalous scatter- Ihkl f i2 = k exp[2B(sin2 /2 )]
ers and normal scatterers. Vectors between normal scat- i
terers do not appear.
Anomalous dispersion with phase change ( f = 0), loge Ihkl f i2 = loge k 2B(sin2 /2 )
can be used to determine the phase angles from noncen- i
trosymmetric crystals. In the case where the position of
Thus, a plot of loge [Ihkl / i f i2 ] versus (sin2 )/2 will
the anomalous scatterer is known, the procedure requires give a straight line of slope 2B and intercept k.
that differences in intensity between be measurable for a A number of conditions must hold so that the values of
significant number of intensities. B and k are reasonable. The sampling interval should be
Anomalous dispersion provides an elegant means to small, so 40 to 50 intervals of (sin )/ are required for
distinguish among near neighbors in the periodic table three-dimensional datasets. Weak reflections must be in-
that would otherwise have similar scattering power. Al- cluded. Elimination of all reflections with I < 3 (I ) will
loys such as -brass (CuZn) and Cu2 MnAl were early cause the average intensity to be too high in the high ranges
examples of structures determined by this technique. The of (sin )/, and that will cause an underestimate of the
tunability of the synchrotron source allows choice of a temperature factor and a corresponding overestimate of
wavelength close to the absorption edge of an element the scale factor. Including weak reflections at half of the
in the sample to maximize the anomalous component of local minimum observed intensity is better than leaving
scattering. them out, but a Baysian fill is probably a better strat-
Neglecting to correct for anomalous dispersion will ob- egy. Large excursions on both sides of the best-fit line
viously introduce a small error in the magnitudes and are quite common and reflect the facts that Wilsons for-
in the phases of the calculated structure factors. It was mula was derived for a random distribution of atoms and
long assumed that such errors would have little effect on real structures contain many repetitions of certain bonded
atomic positions. However, Cruickshank and McDonald and nonbonded distances. The inflection points (the re-
have pointed out that neglect of the correction will always gions in the graph where the experimental points cross the
cause errors in thermal parameters, and in the case of po- straight line) are relatively constant for different types of
lar space groups very serious errors in coordinates can structures. Hall and Subramanian recommend an inflec-
arise. The size of the error varies directly with f and tion point least squares in which the least-squares line
inversely with the resolution of the data. For a moderately is fitted to 15 points: the five lowest angle points, five
heavy atom such as Co (Z = 27), the error in coordinates points nearest to (sin2 )/2 = 0.15, and five points near-
can be as large as 0.06 A in the experiment done with
est (sin2 )/2 = 0.26.
Cu radiation. Neglect of anomalous dispersion will cause
for structures with atoms no
errors of the order of 0.005 A
G. Thermal Diffuse Scattering
heavier than oxygen with Cu radiation or sulfur with Mo
radiation. The error caused by including f and choos- Thermal diffuse scattering (TDS) arises mainly from
ing the wrong enantiomer is twice as large. low-frequency acoustic modes in the crystal. First- and
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 145
second-order TDS cause the scattering density to peak un- postulate trial structures and see whether the pattern of in-
der Bragg peaks, with the degree of peaking related to the tensities matched. The method cannot generally be applied
velocity of sound in the crystal. The effect is not removed to molecular structures in low-symmetry space groups.
in normal data-reduction techniques and is different in
different directions in the crystal. In normal structure de-
termination, TDS is ignored. The result is a systematic B. Transform Methods
decrease in apparent thermal parameters. Since TDS in- A crystal can be regarded as convolution of the lattice
creases with increased (sin )/, it enhances the apparent with the unit-cell contents. By the convolution theorem,
intensity of high-order diffraction data. the transform of a convolution is the product of the trans-
In very precise work, such as the determination of forms. The diffraction pattern is the transform of the crys-
charge-density distribution, it is extremely important that tal structure and so must be the product of the delta func-
the effect be eliminated or accounted for. It should be noted tion representing the reciprocal lattice and the transform
that the amount of TDS included in a diffraction profile of the unit cell contents.
will generally be different for X-ray and neutron experi- From the properties of a delta function we know that
ments, since it depends on such experimental conditions the product f (x) (x x0 ) has values only at x0 since
as primary beam divergence, wavelength spread crystal (x x0 ) = 0 if x = x0 . Thus, the diffraction pattern of
dimensions, and counter aperture. Extensive calculations a crystal can be regarded as the transform of the unit-cell
are required to correct for TDS, and most formulations contents sampled at the points of the reciprocal lattice.
demand that the elastic constants of the crystal be known. This implies that a direct plot of the weighted recipro-
However, cooling the crystal can reduce the effect. If cal lattice can give some information about the structure.
is defined as I (TDS)/I (Bragg), cooling from room tem- The method is used frequently to solve the structures of
perature to liquid-nitrogen temperature will reduce by polynuclear aromatic hydrocarbons. In these compounds
a factor of 5. Cooling to liquid-helium temperature will the dominant features of the diffraction pattern are the ben-
reduce by another factor of five. Facilities for X-ray zene transform and the fringe function showing the sepa-
diffraction experiments down to liquid-nitrogen tempera- ration of the molecules. The molecules generally crystal-
ture are fairly common. Helium cryostats are rare. lize in centrosymmetric space groups with one short axis
roughly perpendicular to the plane of the molecule.
Figure 8a shows the calculated transform of a regular
APPENDIX II: METHODS OF hexagon, the transform of a benzene ring. It is character-
STRUCTURE SOLUTION ized by positive and negative regions with strong positive
peaks at a distance 0.8 A 1 from the origin. If the hexagon
A. Trial-and-Error Methods
is tilted along an axis, the distance perpendicular to the
The earliest structures that were determined by X-ray axis of tilt will be foreshortened in crystal space. The cor-
diffraction were mineral structures with relatively high responding distances in the transform will be elongated.
symmetry. Intensities were measured as strong, medium, Simple geometry will allow calculation of the angle of tilt.
and weak, and most of the atoms sat on special positions Determination of the separation is relatively simple.
in the cell. From knowledge of the density of the mate- Consider two centrosymmetric molecules related by a cen-
rial, its chemical formula, and the space group, one could ter of symmetry with a separation of molecular centers
FIGURE 8 Transform of a hexagon. The circle has a radius of 0.8 A 1 . (a) Regular hexagon; (b) tilted hexagon;
(c) geometric construction to determine tilt of hexagon in crystal space.
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
146 Crystallography
FIGURE 9 (a) Transform of naphthalene with correctly oriented reciprocal lattice; inset shows naphthalene orien-
tation used to calculate the transform. (b) Weighted reciprocal lattice of naphthalene for comparison with transform.
Reflections such as 202 and 801, which lie on regions of the transform where density is changing, are much more
sensitive to orientation than are those such as 203 that lie in the middle of strong areas.
of 5 A. Since the molecules are identical in shape and The location of the heavy atom can be determined from
orientation, the combined transform will be that of a sin- the Patterson function:
gle molecule crossed by straight fringes. Regions in the 2
combined transform are strong only if the corresponding P(uw) = (Fhkl )2
V h k l
region in the single transform is strong, but weak regions
arise either from weak regions in the single transform or [cos 2 (hu + k + lw)]
from the zeros of the fringe system. A line perpendicular
to the fringes is the direction of the line joining the cen- = p(x yz)
ters of molecules, and the separation of the molecules is volume of cell
reciprocal to the spacing of the fringes. p(x + u, y + , z + w) d
If the transform and the reciprocal lattice are drawn on
the same scale, the correct relative orientation of one to the The Patterson function can be calculated directly from the
other can be established by matching strong areas of the intensities with no previous knowledge of the phases. It
diffraction pattern with strong areas in the transform. For is the self-convolution of the electron density. This means
fine adjustment of the orientation, attention must be paid that a peak at uw represents a vector between two atoms
to those reflections most sensitive to orientation effects, whose separation is equal to the vector distance from the
those lying on rising or falling regions of the transform. origin to the point uw. The weight of that peak is propor-
See, for example, reflections 202 and 801 in Fig. 9. By tional to the product of the atomic numbers of the atoms
contrast, the 203 reflection lies well within a strong area at each end of the vector. Vectors between heavy atoms
of the transform and its value will not be affected by even tend to dominate these maps and hence allow the position
fairly large changes in orientation. of the heavy atom to be determined.
Figure 10 shows a hypothetical molecule with one
C. Heavy-Atom Methods
heavy atom and its Patterson function. Note that the Pat-
In X-ray diffraction, the scattering power of an atom is terson function is always centrosymmetric. If the heavy
proportional to the square of the atomic number, Z 2 . If a atom were iodine and the light atoms were carbons, II
molecule contains a heavy atom (high Z ) and that atom vectors would have weight 2809; IC vectors, 318; and
can be located, then a set of phase angles can be calculated CC vectors, 36.
for the dataset that are approximations to the true phases. If two molecules are related by a center of symmetry as
A Fourier synthesis calculated with observed structure in Fig. 11, then in addition to the intramolecular vectors
amplitudes and phases appropriate to the heavy atom will near the origin of the vector maps, there are intermolecu-
give a map that contains the heavy atom, some light atoms, lar vectors. The IC and CC vectors are double-weight
and some noise. Phases based on the known atom posi- vectors because both molecules give the same pattern of
tions are better estimates of the true phase than the heavy vectors. However, the II vectors, which represent vectors
atom alone. The iterative procedure is repeated until all between iodine atoms across the center of symmetry, are
atoms are located. single-weight peaks.
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 147
FA = FL + FAR
FB = FL + FBR
148 Crystallography
FIGURE 12 Isomorphous replacement. (a) The difference FAB between the structure factors for the two isomorphs
A and B provides the center for |FB |. (b) The ambiguity in phase can be resolved by a third derivative.
Crystallography 149
FIGURE 13 Anomalous dispersion. The first circle is FN . The vectors FH and FH are drawn and their vector sum
becomes the center for FN H . This gives a twofold ambiguity, which can be resolved in the (a) centrosymmetric case
by noting that the phase must lie on the real axis. In the (b) noncentrosymmetric case, the ambiguity can be resolved
by using the negative anomalous component as shown in (c) or an isomorphous derivative (not shown). FNH and
FN+H refer to Fhkl and F hkl , respectively.
a
where Q is the ratio f 2 / f 2,h
n
, |F1,h
n
| is the magnitude of the tural information are known as direct methods. It is not
structure factor for the normally scattering atoms, |F2,h n
| is necessary to determine all of the phases. In general, about
the normal part of the structure factor for the anomalously 10 the number of non-hydrogen atoms in the cell is suf-
scattering atoms, and 1,hn
and 2,hn
are the associated phase ficient.
angles. There is a second equation for the Friedel mate, Virtually all of the direct-method programs currently
and it is the same as the previous one except for a minus available make use of normalized structure factors, E(h),
sign before the last term. Thus, there will be two indepen- which correct the structure factor F(h) for fall-off with
dent equations for each wavelength at which data are col- angle caused by the temperature factor and the scattering
lected, plus a third equation resulting for the trigonometric factors of the atoms,
identity sin2 x + cos2 x = 1. This set of equations can be 1/2
1/2
solved if data are collected at two or more wavelengths. E(h) = F(h) h f i2
Generally, one dataset is collected at energies below the
absorption edge so that f = 0; a second set is collected If the shapes of the scattering curves are similar, the values
above the absorption edge, with f = 0 and f hav- of E(h) can be calculated from the relation:
ing the same value as for the first experiment. The actual
1/2
N
values for the real and imaginary parts of the scattering E(h) = h Z j (2 ) 1/2
exp(2ih r j )
factor are determined experimentally before the choice of j=1
wavelengths is made. The obvious advantage is that the
where N is the number of atoms in the cell, and h is the av-
isomorphism is exact. In principle, the data can be col-
erage intensity multiple of the hth reflection. The average
lected on one crystal; in practice, radiation damage will
value of |E|2 (h) = 1. The most frequently used relation-
require the use of several different crystals. A generaliza-
shipin direct methods for centrosymmetric structures is
tion of the theory can take into account any number of
the 2 or triplet relationship:
types of anomalous scatterers and any number of anoma-
lous scattereres within each type. S(E h ) S E k E hk
k
F. Direct Methods
where S(E h ) is the sign of the reflection hkl, and means
Methods of structure solution that attempt to evaluate the probably equal to. The probability associated with this
phases of the structure factors without recourse to struc- relationship is
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
150 Crystallography
The symbol kr implies that k ranges only over those vec-
3/2
P+ (E h ) = + tanh 3 (2 )
1
2
1
2 |E h | E k E hk tors associated with large |E| values.
k The process of choosing initial origin-determining
N phases is similar to the centrosymmetric case, but an ad-
where n = n
i=1 Z i and Z is the atomic number. For
example, ditional enantiomorph-determining reflection must also
be specified. The symbols are assigned in the same way
h = 633 and k = 790 as before and result in assignments such as h = 2a b.
These are converted to numerical values, and each set is
hk = 163
then expanded and refined using the tangent formula. The
In this example from a centrosymmetric structure, if all first computer programs for structure solution used this
of these reflections are strong, it is probable that the sign method. The symbolic addition procedure generates only
of F(633) is positive. The probability increases as the a few alternatives for the values of the phases which must
magnitude of the normalized structure factors increase. be considered because the number of resulting unknown
The steps in solving a centrosymmetric structure (where symbols is usually no more than three for four.
the phases can have values of 0 or ) are In multisolution (i.e., multitrial) methods, a small num-
ber of phases are assigned arbitrarily to fix the origin
1. Evaluate E terms. This includes calculation of the and, in the case of noncentrosymmetric space groups,
normalized structure factors and sorting of the reflections the enantiomorph. Additional reflections are each as-
among eight subgroups defined by the parity of the h, k, signed many different starting values in the hope that
and l indices. one (or more) of these sets of starting conditions will
2. Form lead to a solution. Some programs use random-number
2 relationships with the strongest
reflectionsthose that have E values greater than some generators to set starting values for some 20 to 200
arbitrary value such as 1.2 or 1.5. phases, which are then extended and refined by the tangent
3. Determine phase. Historically, phases were deter- formula:
mined by the symbolic addition method. In this method, wk |E k E hk | sin(k + hk )
origin-determining reflections are given signs and a few tan h = k
k k |E k E hk | cos(k + hk )
w
others,
chosen from the strongest |E| values with the most
The weighting function in the tangent formula is useful
2 interactions, are given symbols (a, b, c, etc.). Signs
can be used because, in the case of a centrosymmetric in some approaches. A new phase h is assigned a weight
structure, the phases can only be 0 or . Thus, Sh means the that is the minimum of 0.2h and 1.0:
sign of the structure factor for the h reflection. The value w0 = min(0.2h , 1.0)
for Sh could arise in one of several ways. One way is its
assignment as described above. Alternatively, Sh could ac- Although this allows rapid development of a phase set for
quire a value through the triplet relationship as follows: If each trial, it tends to lead to an incorrect centrosymmetric
Sk is know to be positive, and Shk is known to be a, then: solution in the case of polar space groups. For HallIrwin
weights, w = w0 f (/est ), where f = 1.0 for < est and
Sh = Sk Shk = a decreases for > est . This weighting tends to conserve
the enantiomorph and has been incorporated into more
During the course of the analysis, relationships appear
recent versions of the multisolution programs. The major
among the symbols, such as ac = e. Manipulation of these
difference in treatments for centric and acentric datasets
relationships usually allows the number of unknowns to
lies in the values assigned to the extra reflections in the
be reduced at the end of phase determination.
starting set. In the centric case, the possible values are 0
4. Calculate the E mapa Fourier summation using
and ; in the acentric case, general reflections have four
E values as coefficients and phases determined in step 3.
possible values, /2 and 3/2.
A recent modification to the multisolution approach
In the noncentrosymmetric case, the solution is more dif-
consists of phase refinement alternated with the impo-
ficult since the phase can take any value between 0 and
sition of constraints by peak picking in real space. This
2. Hence, a different set of relationships was developed
approach, referred to as a dual-space method, differs from
for this case:
other multisolution methods in that the phases are ini-
h = k + hk
kr tially assigned values by computing structure factors for
a randomly positioned set of atoms. The occurrence of
kr |E k E hk |(k + hk ) two Fourier transformations per cycle results in an al-
h =
kr |E k E hk | gorithm that is computationally intensive. However, the
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
Crystallography 151
new approach also extends the limits of these programs or incorrect thermal parameters. Holes indicate electron
to much larger structures. It has proven to be capable of density in the model where none exists in the crystal (see
solving complete structures containing as many as 2000 Fig. 14).
independent non-hydrogen atoms (provided that accurate It is common to use difference maps to find hydrogen
diffraction data have been measured to a resolution of atoms once the positions of the heavier atoms have been
1.2 A or better and some heavy atoms are present in the refined isotropically. In centrosymmetric structures, the
structure). phases are either positive or negative and are usually cor-
There has been progress in the experimental evalua- rect for all but the smallest structure factors by this stage.
tion of triplet-phase invariants. The phenomenon of si- Hydrogen atoms will have electron densities in the range
multaneous diffraction has long been considered a nui- 0.6 to 0.9 e A 3 . Phase errors in the noncentrosymmet-
sance in single-crystal work. If three reflections lie on ric case cause difference maps to be less well defined. At
the sphere of reflection simultaneously, there is a power the end of refinement, a difference map should show no
transfer that tends to enhance such weak reflections at the significant features.
expense of strong ones. In 1977, it was discovered that Since it is much faster to calculate a Fourier synthe-
the shape of the simultaneous diffraction profile is sen- sis than to carry out least squares, refinement by differ-
sitive to the phase of the triplet that gave rise to it. The ence synthesis was popular before the advent of modern
sense of the asymmetry is opposite for positive and nega-
tive triplets. This effect has been observed with ordinary
mosaic crystals of relatively heavy scatterers (ZnWO4 )
using CuK radiation from a fine-focus tube. This tech-
nique could be enhanced by the use of synchrotron radia-
tion, since it is desirable that the beam be monochromatic,
intense, and highly collimated. It is envisioned that exper-
imental phase determination could be used to establish a
starting set of 50 or so triplets and that the tangent for-
mula or some similar technique could then expand these
relationships.
A. Difference Fourier
A Fourier synthesis with coefficients F = |F0 | |Fc | re-
flects differences between the crystal and the model. The FIGURE 14 Difference map. Comparison of observed, calcu-
difference map is routinely used during structure solution lated, and difference electron density for (a) an atom displaced
by 0.1 A from its correct position, and (b) a temperature factor
to check the integrity of the model as it is being developed.
overestimated. (c) Difference density observed when an isotropic
Large peaks in the map (Z /3 to Z /2) correspond to atoms temperature factor is used in the model and the thermal vibra-
not yet included in the model. Smaller peaks may indi- tion of the crystal is actually anisotropic. Dashed contours are
cate a slightly misplaced atom or wrong scattering type negative.
P1: GJC/LOW P2: FYK Final Pages
Encyclopedia of Physical Science and Technology EN004G-160 June 8, 2001 19:47
152 Crystallography
Crystallography 153
atoms are uncorrelated. If systematic effects such as ab- The rigid-body models require 12 parameters for cen-
sorption and extinction have been ignored, if the qual- trosymmetric molecules and 21 for noncentrosymmetric.
ity of the crystals is poor, or if the weighting scheme is In the former, both the translational and librational axes
inappropriate, these parameters may have little physical pass through the center of gravity of the molecule. In the
meaning. noncentrosymmetric case, the libration axes do not nec-
Hirshfeld has proposed a convenient test for reason- essarily coincide, and an additional tenser, S, is required
ableness of a set of vibration parameters. Since bond- to account for the correlation between the librational and
stretching vibrations can be assumed to have much higher translational motions.
energy than other intramolecular modes such as torsion or Two indicators are used to test the goodness of fit be-
angle bending, bonds can be considered rigid to a first tween the refined values of the Ui j parameters and those
approximation. If two atoms A and B are bonded, it is calculated from the rigid-body model. The root-mean-
expected that the component of vibration along the bond square difference between observed and calculated values,
will be the same for both atoms. For carbon atoms, agree- given by (Ui j )2
1/2 , would be considered excellent in
2 is expected for low-temperature 4 2 3 2
210 A but only fair at 210 A . R =
ment within 0.001 A the
region of
structure determinations where bonding effects have been |Ui j |/ Ui j would be considered excellent at 0.04.
accounted for. In the spherical atom approximation, Obviously, the interatomic vectors in a crystal structure
the discrepancy is more likely to be of the order of can be interpreted as bond distances only if thermal vibra-
0.005 A 2 . In molecular crystals, vibrations of neighbor- tions are negligible. To correct for the effects of thermal
ing atoms are correlated to some degree. For relatively vibration, a model for that vibration is required. Often it is
rigid molecules, translational and vibrational oscillations better to reduce the effects of thermal motion by working
are the major contributors to the internal modes of vi- at the lowest available temperature than to attempt an a
bration. If the rigid body model is a good approxima- posteriori correction.
tion, it is possible to correct bond lengths for thermal
vibration.
The reason for the correction can be seen in Fig. 15. SEE ALSO THE FOLLOWING ARTICLES
If a molecule is undergoing librational motion, with a
root-mean-square amplitude of libration about an axis CRYSTAL GROWTH CRYSTALLIZATION PROCESSES
through 0, then an atom at a distance l from the center of FERROMAGNETISM MICROSCOPY PHASE TRANS-
FORMATIONS, CRYSTALLOGRAPHIC ASPECTS X-RAY
libration sweeps out an arc. However, the X-ray analysis
will place the center of the atom at C, the centroid of the SMALL-ANGLE SCATTERING X-RAY, SYNCHROTRON
AND NEUTRON DIFFRACTION
electron distribution. Thus, bond distances are systemati-
cally shortened by an amount PC = l l cos l2 /2.
In three dimensions, the corrections are additive. BIBLIOGRAPHY
Carter, C. W., Jr., and Sweet, R. M., Eds. (1997). Methods in Enzymol-
ogy, Vol. 277, Macromolecular Crystallography. Academic Press,
New York.
Dunitz, J. D. (1979). X-Ray Analysis and Structure of Organic
Molecules, Cornell University Press, Ithaca, NY.
Ewald, P. P., Ed. (1962). Fifty Years of X-Ray Diffraction, N.V.A.
Oosthoeks Uitgeversmaatschappij, Utrecht.
Journal of Physics D, Applied Physics. (1991). Special issue on structural
aspects of crystal growth, 24(2), February 14.
Karle, J. (1994). J. Chem. Inf. Comput. Sci. 34, 381390.
Ladd, M. C. F., and Palmer, R. A., Eds. (1980). Theory and Practice of
Direct Methods in Crystallography, Plenum, New York.
McPherson, A. (1982). Preparation and Analysis of Protein Crystals,
Wiley, New York.
Schmidt, P. W., Ed. (1983). Proc. symp. small-angle scattering. Trans.
Am. Crystallogr. Assoc. 19.
Schoenborn, B. P., Ed. (1985). Proc. symp. structure determination with
synchrotron radiation. Trans. Am. Crysrallogr. Assoc. 21.
FIGURE 15 Rigid-body motion. An atom at a distance l from the Wyckoff, H. W., Hirs, C. H. W., and Timasheff, S. N., Eds. (1985). Meth-
center of libration sweeps out in an arc. Least-squares analysis ods in Enzymology, Vol. 114, Diffraction Methods for Biological
places the atom at C rather than P. Macromolecules. Academic Press, New York.