METHODOLOGY OF X-RAY STRUCTURE DETERMINATION

Crystal structure analysis is based on diffraction phenomena caused by the

interaction of matter with x-rays, electrons or neutrons. X-rays scattered from each atom
in the unit cell contribute to the overall scattering pattern. Since each atom acts as a
source of scattered X-rays, the waves will add constructively or destructively in varying
degrees depending upon the direction of the diffracted beam and the atomic positions.
The theoretical background of x-ray structure analysis is briefly discussed in this chapter.

The Structure factor

The structure factor is the combined scattering of all the wavelets scattered
from all the atoms in the unit cell in a specified direction of the hkl plane. Each of these
waves has amplitude proportional to fj, the scattering factor of the atom and a phase j with
respect to the origin of the unit cell. The scattering vector ‘s’ corresponds to a triplet of
integers (hkl). If fj and rj are the scattering factor and the position vector of the jth atom,
the structure factor Fhkl can be represented as
N
Fhkl =  f j exp(2 ir j .s) (1.1)
j =1

where ‘N’ is the number of atoms in the unit cell. The exponential term expresses the
relative phase of the radiation scattered by each atom j as a function of its position rj in
the unit cell.
The phase angle фhkl associated with Fhkl is given by
tan  hkl =Bhkl/Ahkl (1.2)
The intensity of the diffracted beam I hkl is proportional to the square of the structure

factor amplitude.
i.e. I hkl α Fhkl 2 2
= Ahkl + Bhkl
2
(1.3)
N N
= { f j cos (2 r j .s) }2+ {  f j sin(2 r j .s) }2
j =1 j =1

1
 N N
=  fi f j [cos(2 ri .s) cos(2 r j .s) + sin(2 ri .s)sin(2 r j .s]
i =1 j =1

N N
Or I hkl =  fi f j cos[2 (ri − r j ).s] (1.4)
i =1 j =1

Thus the intensities are found to depend on the inter atomic vectors (ri –rj) and not on the
actual atomic coordinates, which depend on an arbitrary choice of the origin.

Atomic Scattering Factor

X-rays set the electrons of atoms into vibration causing them to act as sources
of secondary radiation. The scattering power of a given atom for a given reflection,
known as the scattering factor (fo) depends on its electronic structure and the angle of
scattering and can be expressed in terms of the scattering power of an equivalent number
of electrons located at the position of the atomic nucleus. The amplitude of coherent
scattering from the total electron density will be obtained by adding the amplitudes for
the electrons taken individually. The change in scattering power due to the thermal

motion is given by exp( − BSin 2 / 2 ) where B = 8 2 U 2 and U 2 is the mean square

amplitude of the atom perpendicular to the reflecting planes. Hence the proper scattering
factor for a real atom is given by
f = f 0 exp(− BSin 2 /  2 )

Atomic scattering factor is defined as the ratio of the amplitude of the coherent
scattered radiation from an atom to that from a single electron situated at the atomic
centre.

The electron density equation

X-ray diffraction pattern gives the image of crystal structure. Since it is the
electrons which scatter the X-rays, they can be seen in the image, giving the value of
electron density at every point in a single unit cell of the crystal. The electron density
(number of electrons per cubic angstrom unit) can be expressed in terms of the structure
factor as:
+
 ( xyz ) =
1
V

h , k ,l =−
F ( hkl ) exp  −2 i ( hx + ky + lz )  (1.5)

2
Where the summation is over all the structure factors F(hkl) and V is the volume of the
unit cell.

Bragg’s Law
A qualitatively simple method for obtaining the conditions for diffraction was
described by Bragg’s Law, considering the diffraction as the consequence of
contemporaneous reflections of the X-ray beam by various lattice planes belonging to the
same family. This gives the geometrical condition under which a diffracted beam can be
observed.

Figure (1.1)shows rays diffracted from lattice planes. To get constructive interference, the
path difference must be an integral multiple of wavelength λ. Hence, Bragg’s Law states
that if the angle of incidence of the nth order diffraction from a set of crystal lattice planes
separated by dhkl is θB, then
n = 2d sin (1.6)

where λ is the wavelength of X-rays.

1.5 Resolution
The resolution is given by the minimum value of d which appears in Bragg’s
law. This corresponds to the maximum value of θ for the set of diffraction data. With Mo
Kα radiation and all data collected to a maximum value of θ of 25º, Bragg’s law gives
2sin(25º)/0.71=1/dmin, producing a resolution (dmin) of 0.84 Å. The maximum possible
resolution is λ/2, which occurs when sin θmax=1. For Cu Kα radiation, this will be 0.77 Å,
similar to the resolution obtained with Mo Kα for θmax=25º.

3
The phase problem
Only the structure factor amplitudes are obtained from the measured X-ray
intensities and not their phases, which are essential for the calculation of electron density.
This problem is called the phase problem. To overcome the phase problem, the methods
used are
(i) Patterson search and interpretation techniques;
(ii) direct methods;
(iii) use of anomalous dispersion;
(iv) isomorphous replacement;
(v) molecular replacement etc.
Patterson methods are used when some heavy atoms are present. Direct methods are the
most important in chemical crystallography. The other methods feature in
macromolecular crystallography.

Data reduction
In the relationship between integrated intensity and the square of
structure factor amplitude there are several factors that vary from reflection to reflection
The process of data reduction can be done by applying a number of corrections. The
main geometrical and physical factors that affect the intensities are due to Lorentz,
Polarization, Absorption, Extinction and thermal motion. The procedure followed to
extract the relative structure factor amplitudes from the raw integrated intensities is called
data reduction.

Lorentz correction
Diffraction occurs whenever reciprocal lattice nodes cross the sphere of
reflection. Depending on the method used to record the reflection intensity and on the
position of the reciprocal lattice node, the time required for different nodes to cross the
Ewald sphere are different. The correction applied for this is called Lorentz correction.

4
The Lorentz correction is
L= 1/sin 2θ (1.7)

Polarization correction
The polarization correction depends on the state of polarization of the
incident X-ray beam and on the scattering angle of the diffracted beam. It takes into
account the partial polarization of the unpolarized incident X-ray beam which occurs in
the diffraction process. It does not depend on the crystallographic apparatus but only on
the diffraction angle 2θ and can be expressed analytically by
1 + cos 2 2
P= (1.8)
2
where θ is the Bragg angle of the reflection produced by the specimen and θM ,the angle
of the reflection of the monochromator crystal which was used to select the wavelength.
Lorentz and Polarization corrections put together can be applied as

Lp =
Absorption correction
When an X-ray beam passes through a crystal, its intensity is reduced by
scattering and by absorption. The extent of absorption depends on the path length of the
beam through the crystal, the nature of the atoms in the crystal and the wavelength of the
incident X-ray beam.
The transmission of X-ray beam after passing through a thickness t of an
absorbing crystal is given by the equation [4]
I
T= = exp[−  (t i + t d )] (1.11)
I0
where I0 is the intensity of the incident beam, ti and td the incident and diffracted beam
path lengths and μ is the total linear absorption coefficient for the primary beam which is
a function of wavelength and atomic number. If the shape of the crystal is exactly known,
then it is possible to make correction for absorption by calculating
T= V −1  exp[−  (t i + t d )]dV (1.12)
v

5
where dV is an infinitesimal volume of the crystal

Temperature factor
Thermal agitation at temperature above absolute zero will cause the atoms
and molecules in crystals to fluctuate about their equilibrium position on the space lattice.
An exponential expression by which the scattering of an atom is reduced as a
consequence of vibration is called the temperature factor. For best results, thermal
vibrations are assumed to be isotropic and the values of atomic temperature factor (B) are
the same for all the atoms. In that case, for isotropic vibrations, the observed intensities
are reduced by a factor exp (-2B sin2 θ)/λ2 known as Debye–Waller factor. Here B is given

by B=8π2 U 2 where U 2 is the mean square displacement of the atom from its equilibrium
position. In general, an atom will not be free to vibrate equally in all directions. If it is
assumed that the probability of finding the centre of the atom has a three dimensional
Gaussian distribution, the surfaces of equal probability will be thermal or vibrational
ellipsoids centered on the mean position occupied by the atom.

Extinction
Extinction is the modification of the incident beam as it passes through a
single block of a crystal. It depends on the mosaic structure of the crystal and has two
components; primary extinction and secondary extinction. Primary extinction takes into
account the loss of intensity due to dynamic effects inside every single block. At Bragg
angle, every incident wave can suffer multiple reflections from different lattice planes;
after an odd number of reflections the direction will be same as the diffracted beam: after
an even number of reflections the direction will be the same as the primary beam. Each
scattering causes a phase lag of λ/4. Then the unscattered radiation is joined by doubly
scattered radiation having with a phase lag of π causing destructive interference.
Secondary extinction takes into account the fact that the lattice planes first encountered
by the primary beam will reflect a significant fraction of the primary intensity so that
deeper planes receive less primary radiation. That causes a weakening of the diffracted
intensity, mainly observable for high-intensity reflections at low sinθ/λ values in
sufficiently perfect crystals. Secondary extinction is equivalent to an increase of the linear

6
absorption coefficient and hence is negligible in case of sufficiently small crystals. If the
reflections are affected by the secondary extinction, for some high-intensity reflections,
|Fobs| seems to be less than |Fcal|.
In the crystal structure refinement package SHELXS-97, an extinction
parameter x is refined by least squares, where Fc is multiplied by
k[1 + 0.001* x * FC 2 *  3 / sin2 ]−1/ 4 (1.13)

where k is the overall scale factor.

Methods of Structure Solution.

The phase problem can be solved by several methods. They include Molecular
Replacement method, Anomalous scattering method, Patterson method and Direct
methods
(1) Molecular Replacement Method
The technique of positioning a molecule or a fragment of known structure in a
crystal cell is called molecular replacement method. In principle, inorder to minimize the
difference between Fobs and Fcal, six variables are to be considered. As it is impossible in
practice, first the orientation of similar sub units in a crystal cell is to found out without
any knowledge about the translation t making use of Patterson function. After the correct
orientation can be found, a search for the translation vector can be carried out.
(2) Isomorphous Replacement Method
This method is applicable in the case of two or more crystalline derivatives
that have the same crystal structure but have different heavy atoms. Classic small-
molecule cases have been isomorphous salts of alkaloid and porphyrins with different
metal atoms in the chelated positions.

(3) Anomalous Scattering Method

This method is based on the anomalous scattering properties of certain atoms
under proper conditions. This effect occurs when the wavelength of the incident X-rays is
reasonably near that of an absorption edge of the irradiated element. Sulfur is the lightest
element for which dispersion can be observed with common radiations. So in organic
compounds, the use of this effect is normally combined with the heavy atom method. For

7
inorganic structures, there may be cases in which the anomalous scatterer is not
significantly heavier than the other atoms in the structure and phasing must be based on
the anomalous effects.
(4)Patterson method
The lack of knowledge of the phases of structure factors prevents us from
directly computing an electron density map showing the positions of atoms in the unit
cell. This method can be applied to structures in which one or few atoms are markedly
heavier than the rest. It is possible to find the locations of heavy atoms by use of
Patterson function [14] which do not require a prior knowledge of the phases. Patterson
suggested as an aid to structure determination the use of the function
K = −H (1.15)
where ‘ r ’is the position vector and ‘ H ’ is the scattering vector.
Vectors between heavy atoms appear with weights corresponding to the
square of their atomic numbers and thus stand out strongly against the background of
heavy-light and light-light atom vectors in a Patterson map. This helps to recognize the
heavy atom vector in the Patterson map.

(5) Direct method

‘Direct method’ approach is applicable to molecular structures having equal
atoms or atoms with nearly equal atomic number. In direct methods, structure factor
phases are derived directly from the observed amplitudes through purely mathematical
techniques without assuming any kind of molecular structure. In 1953, Hauptman and
Karle established the basic concepts and probabilistic foundations of direct methods.
The probability relations between amplitudes and phases can be strengthened
and simplified through the introduction of two modified structure factors, either the
Unitary, Uhkl; or the normalized, Ehkl structure factors. Unitary structure factor, Uhkl
satisfies the relationship

N
Fhkl
U hkl = N
=  n j exp(2 iH.r j ) (1.16)
f
j =1
j
j =1

8
 fj
where n j = N
and U hkl  1
f
j =1
j

The normalized structure factor E hkl is given by

2
U hkl
2
E hkl = N
(1.17)
n 2
j
j =1

where  is an integer whose value is generally 1.

2
From the above equation, E hkl = 1 can be derived. (1.18)

*************************************
Thermal motion also affects X-ray intensities. The effect of thermal motion is
to spread the electron cloud over a large volume and thus to cause the scattering power of
real atom to fall off more rapidly than that of the stationary model. Considering a set of
observed intensities falling within a restricted range of sin /  , it was showed from the
2
expansion of Fhkl in terms of f j , that

 N   −2 Bsin2 
 =   f 0 j 2  exp 
2
FH 
 j =1   2 
which can be abbreviated to
 −2 BSin 2 
IH  = K  s exp 
2

 2 
N
where K 2  s =  f 0 j 2 or
j =1

 IH    2 BSin 2 
ln   = 2 ln K −  
  s   
2

(1.19)
If the left hand side of eqn(1.19) is evaluated for each of the shells of constant f, and the
values are plotted against sin 2 /  2 [Wilson plot], it results in a straight line whose
intercept at sin 2 /  2 = 0 is 2 ln K ' and the slope is − 2 B . Thus the values of K and
B are obtained from the Wilson plot.

9
 ln(Irel/Σfj2)

Sin2θ/λ2
Wilson plot for determining scale and thermal parameters
********************************
Structure Invariants and Seminvariants
One of the difficulties in solving the phase problem is defining the origin of
the unit cell properly with respect to the symmetry elements and all the contents of the
cell. Even though shifting the origin does not affect the structure amplitudes it may
change the phase drastically. Hence the selection of phases is associated with choice of
origin. Certain combinations of phases do not change regardless of the arbitrary
assignment of cell origin. They are called structure invariants.
From the observed amplitudes, information on single phases or linear
combination of phases independent of the choice of the origin can be obtained. Since
their value depends only on the structure, they are called structure invariants. The most
general structure invariant is represented by the product
FH1 FH 2 .......FHn = F H1 FH 2 ......FHn exp[i(H1 + H 2 + ....... + Hn )] (1.20)

when H1+H2+……….+Hn=0 (1.21)

When the origin is shifted by a general vector ro, the product of the structure
factors with respect to the shifted origin will be
N
FH' =  f j exp{2 ihH.(r1 - r0 )}
j =1

= FH exp{2 iH.r0 }

= FH exp[i (H − 2 H.r0] (1.22)

10
Thus the modulus remains unchanged while the phase changes by
 = 2 (H1 + H 2 + .......H n ).r0 (1.23)

(H1 + H 2 + .....Hn ) = 0
When (1.24)
H1 '+ Η 2 '+ .......Hn ' = H1 + H 2 + ........Hn

Such combinations of phases are independent of the choice of the origin and are called
structure invariants.

*****************************
Structure semi-invariants are single phases or linear combination of phases
which are invariant with respect to a shift of origin, provided the position of origin is
restricted to those points in the cell which posses the same point symmetry. A structure
semi-invariant can be transformed into a structure invariant by adding one or more pairs
of symmetry equivalent phases. In a given space group possessing the symmetry operator
C = (R,T), the phase φH is a structure semi-invariant if it is possible to find a reflection h
such that
Ψ = φH-φh+ φhR (1.25)
is a structure invariant. ie., H-h+hR = 0
Also, ψ = φH-2πh·T (1.26)
Ψ is independent of any choice of origin, while φH and T will depend on it. The
dependence is such that eqn. (1.26) holds and hence the value of φH will not change if the
origin is moved within those points which maintain the vectors T unchanged.

****************************

Phase relationships
The relation connecting the phases 3  -H + K +  H-K = 0 is possible only when

E
K
-H EK EH-K = real and large

only if |E-H|| EK| and |EH-K| are large.

Then 3  -H + K + H-K = 0

11
The  sign means “is probably equal to”
This is called Triple Phase relation.

For a Centro symmetric structure,

S H SK S H-K  +1 (1.30)

where S is the sign of the phase

Considering the special case when K = -H, from eqn. (1.25)
-H = −H (1.31)

For a structure invariant involving four phases

4 = ( H ) + ( K ) + ( L ) + ( M ) (1.32)

where H+K+L+M=0
Taking into account the neighbourhood concept, in addition to large
EH , E( K ) , E( L ) and E( M ) values (first neighbourhood) there are three additional

magnitudes (second neighbourhood) which strongly influence the value of φ4 [26,27].

They are E( H+K ) , E( H+L ) and E( H+M ) usually referred to as ‘cross terms’.

If the magnitudes E ’s are assumed as known in any of the neighbourhood of

 , the conditional probability distribution P(  ) of  produces an estimate of  .

Reliability of  is high if the variance is small. The conditional probability relation for
Triplet:
1
P1/3 ≈ exp( A Cos  ) (1.33)
2I 0 ( A)
where I0 is the modified Bessel function and A is defined by
3
A= E E E
 23/ 2 H K -H-K
N
and  n =  f0 j n
j =1

where foj is the zero angle atomic scattering factor

12
This distribution has a unique maximum at  =0 in the interval –π to + π so that the most
probable value of φ is zero.
*****************************************************************

Tangent Formula
According to eqn.1.25, the expectation value of φH can be given as
< φH > = φK + φH-K (1.37)
If there exists more than one pair of reflections whose known phases give an indication
for one unknown phase, φH, then
< φH> = the phase of E
K
K EH-K (1.38)

This equation can be expressed in an alternative form by considering separately the real
and imaginary parts of the right-hand side. Thus

E
K
K EH-K =  EK EH-K cos(K + H-K ) + i  EK EH-K sin(K + H-K )
K K

E EK H-K sin(K + H-K )

which gives H = tan −1 K
(1.39)
E E
K
K H-K cos (K + H-K )

Once the phases of a few reflections are obtained using single and triple phase relations,
tangent formula can be used to refine phases and determine new phases by an iterative
procedure.
**** All E’s whose phases are known are included in the summation over K and the
result gives estimates for φ(H). This is repeated for each ‘H’ and new phase values thus
obtained are used in the next iteration. This process is continued till all the phases have
converged.*****

Structure Solution
Once the phases are known we can arrive at the structure by calculating electron density
using the Fourier transform.

13
 +
 ( xyz ) =
1
V

h , k ,l =−
F ( hkl ) exp  −2 i ( hx + ky + lz ) 

****************
The program SHELXS is used to solve the structures by direct methods. This program
employs a multiple permutation single solution procedure which uses phases generated
by random numbers, refines them and examines the phase sets for a single solution. At
first the program averages equivalents and calculates values of Rint and Rsigma using the
equations

F 2
( )
− F2 mean
Rint =
F 2

(1.40)
Rsigma =
  (F 2 )
F 2

The program uses triple phase relations for generating phases and negative
quartets for assigning their figure of merit NQEST. The initial phase refinement is
normally performed using a reflection subset. The best 10% of the subset phase
permutations judged by R and NQEST (called Filter)
Filter = Rα + [0 or (0.25 + NQEST)]2 (1.41)
are then refined using the full set of reflections employed in phase refinement. In contrast
to the subset tangent refinement the full phase refinement is performed with unit weights
and a modified tangent formula.
For centrosymmetric structures, phases are calculated using
 = phaseof [( E −*1 E −*2 ) − wq *t *  ( E −*3 E −*4 E −*5 )] (1.42)

where the first summation is over negative quartets, t is a constant and wq a constant in
the range 1-4, depending on the number of negative quartets used. For ascentric phases,
the standard tangent formula is used until  is greater than its estimated value.This is an
indication that the phase relations are overconsistent. The phase angle obtained from the
the tangent formula is increased or decreased by an angle cos −1 ( /  est ), where the sign
of the correction is chosen to give the best agreement with the negative quartets. This

14
procedure simultaneously minimizes the R and NQUAL , the two figures of merit
calculated at the end of the refinement.

R =
 w[ −  est ]2
(1.43)
 w( )est
2

where the weight w is 1 /[ est + 5]

 (E * E ) *  (E * E * E )
−1 −2 −3 −4 −5

  ( E * E ) *  ( E * E * E ) 
NQUAL = (1.44)
−1 −2 −3 −4 −5

where the outer summations are performed over all the refined reflections and the inner
over triplets and negative quartet relations involving a given reflection. NQUAL
approaches -1 for the correct solution. A combined figure of merit, CFOM is then
calculated.
CFOM = Rα+[0 or (NQUAL-wn], whichever is large. (1.45)
where ‘wn’ is a structure dependent constant. CFOM should be minimum for the best
solution. Only this solution is retained for calculating an E map. The peaks of the electron
density in the electron density map are usually interpreted as the centroid of the atoms.
*******************************
An appropriate structure becomes more accurate by successive Fourier synthesis. The
Fourier cycles not only allow the location of new atoms but also improve the positions of
the model atoms. A difference Fourier series
1
Δρ(R) =  ( ( FO )H − ( FC )H ) exp(−2 iH.R + i(C )H )
V H
(1.46)

shows the difference between the observed and the calculated electron densities. Location
of missing non-hydrogen atoms, if any, is possible by this method.
************************
Since there are so many measured reflections, the ‘trial structure’ parameters,
representing atomic positions and extents of vibration, may be refined to obtain the best
possible fit of observed and calculated structure factors. In any crystal structure analysis
there are many more observations than parameters to be determined. The best parameters
corresponding to some assumed model of the structure are found by minimizing the sum

15
of squares of the discrepancies between the observed values of |F| (or |F|2) and those
calculated for an appropriate trial structure .
The following criteria should be applied for the correctness of a structure.
1. The agreement of individual structure factor amplitudes with those calculated for the
refined model should be consistent with the estimated precision of the experimental
measurements of the observations.
2. A difference map, phased with final parameters for the refined structure, should reveal
no fluctuations in electron density greater than those expected on the basis of the
estimated precision of the electron density.
3. Any anomalies in molecular geometry or packing should be scrutinized with great care
and regarded with some skepticism. }
****************
Structure Refinement
The structural models obtained by Direct methods are almost incomplete as
they represent only a first approximation of the real structure. When all the atoms in the
structure are located at chemically meaningful positions, the correct structural parameters
are obtained by the method of least squares. The refinement is continued till convergence
is attained. The calculated intensities are then compared with the measured intensities and
the best model is that which minimizes M:

M=  w(Fo2 − Fc2 )
2

Residual factors
The quality of the refined model can be judged with the help of residual
factors or ‘R factor’ given below.

R=
F −Fo c

F o

This value is acceptable if found less than 5%.

At the end of the structure analysis atomic coordinates and their thermal parameters will
be obtained. From this molecular geometry, intermolecular interactions etc. can be
calculated.
*************************

16
 Here the scale factors, positional and thermal parameters of the atoms in the unit cell are
adjusted to obtain best agreement between observed and calculated structure factors.

The crystallographic Residual factor is used as an indicator of this. During the initial
stages of refinement, the vibrational thermal parameters are assumed isotropic and the
anisotropic nature of thermal parameters is introduced for non-hydrogen atoms later. The
number of cycles of least square refinement can be increased for best adjusted
parameters. The refinement is continued until the shifts in the parameters are some small
fraction of the estimated standard deviation of corresponding parameters.
The structures reported in this thesis are refined using the program
SHELXL[36].This program automatically rejects systematically absent reflections. The
sorting and merging of the reflection data is controlled by the MERG instruction. The
merging residual values Rint and Rsigma are defined as in eqn. (1.40).The program is meant
for small molecule structures and least square refinement is to be done. First, a complete
set of structure factors is calculated from the atomic model by means of Fourier
transformations. The calculated intensities are then compared with the measured
intensities and the best model is that which minimizes M:

M=  w(Fo2 − Fc2 )
2
(1.47)

Where F is the structure factor and the subscripts o and c stand for observed and
calculated.
In order to minimize the function (1.48) the measured intensities have to be
transformed into structure factors. This involves the extraction of a root which leads to
mathematical problem with very weak reflections or reflections with negative measured
intensities. To overcome this problem, negative measurements must be set to zero or to
an arbitrary small positive value in a refinement against F. When such an approach is
made, very weak reflections also contain information and ignoring them affects structure
determination. Another problem in the use of F values arise from the difficulty of
estimating the σ(F) values from σ(F)2, the latter of which are determined during data
reduction. Least square method is very sensitive to weights applied each reflection.

17
Hence problems with the σ estimation lead to inaccuracies in the refinement. Refinement
against F2 (1.47) does not cause any of these problems. It has some advantages. It makes
the refinement of twinned structure mathematically simpler and refinement against
squared structure factors less likely to be minimum. Hence refinement against F2 is
superior to refinement against F. }
***************************
Residual factors
The quality of the refined model can be judged with the help of residual
factors or ‘R factor’ given below.

R=
F −Fo c

F o

These factors should converge to a minimum during the refinement. The three most
commonly used residual factors are:
The weighted R- factor based on F2: wR(or wR2 in SHELXL), which is most
closely related to the refinement against squared structure factors.
1
  w( Fo2 − Fc2 ) 2  2

wR =   (1.49)
  wFo2 
The weighting factor w is individually derived from the standard uncertainties of the
measured reflections.
The most popular one in the unweighted residual factor based on F: R (or R1
in SHELXL)

R=
F −Fo c
(1.50)
F o

The accepted value is close to 0.01 – 0.05

Finally, there is a Goodness of Fit: GooF, GoF, or simply S

1
  w( Fo2 − Fc2 ) 2  2

S=  (1.51)
 ( N R − N P ) 

18
Where NR is the number of independent reflections and NP, the number of refined
parameters. Theoretically, for a properly adjusted weighting scheme, the value of S
should be close to 1. }

Weighting scheme
The weighting scheme suggested by the program produces a flat analysis of
variance in terms of Fc and suggests updating parameters in the output. It is better to use
unit weights in the early stages of refinement. The updated parameters are incorporated
after all atoms have been found and at least the heavier atoms refined anisotropically. The
weighting scheme is defined as follows.
w = 1 /[ 2 ( Fo2 ) + (a * P) 2 + b * P] (1.52)

where P = ( FO2 + 2 FC2 ) / 3 and a & b are adjustable parameters chosen in such a way that

the  wF 2  values remains constant when the reflections are grouped in different ways.
At the end of the structure analysis atomic coordinates and their thermal parameters will
be obtained. From this molecular geometry, intermolecular interactions etc. can be
calculated.

REFERENCES
1. Ciegg, W; Blake, A.J; Gould R.O and Main,P, (2001) Crystal structure analysis
principles and practice,1st ed.
2. Woolfson, M.M, (1997), An introduction to X-ray crystallography, 2nd ed.
3. Giacovazzo C; Monaco H L; Vitterbo D; Scordavari F; Gilli G; Zanotti G; Catti M,
(1992 ) ‘Fundamentals of Crystallography’,.
5. Ladd F.M.C; Palmer R A, (1993), Structure determination by x-ray crystallography,
3rd ed.

19