International
Tables for Crystallography Volume B Reciprocal space Edited by U. Shmueli © International Union of Crystallography 2006 |
International Tables for Crystallography (2006). Vol. B. ch. 1.3, pp. 59-64
|
Let ρ be the distribution of electrons in a crystal. Then, by definition of a crystal, ρ is Λ-periodic for some period lattice Λ (Section 1.3.2.6.5) so that there exists a motif distribution with compact support such that where . The lattice Λ is usually taken to be the finest for which the above representation holds.
Let Λ have a basis over the integers, these basis vectors being expressed in terms of a standard orthonormal basis as Then the matrix is the period matrix of Λ (Section 1.3.2.6.5) with respect to the unit lattice with basis , and the volume V of the unit cell is given by .
By Fourier transformation where is the lattice distribution associated to the reciprocal lattice . The basis vectors have coordinates in given by the columns of , whose expression in terms of the cofactors of A (see Section 1.3.2.6.5) gives the familiar formulae involving the cross product of vectors for . The H-distribution F of scattered amplitudes may be written and is thus a weighted reciprocal-lattice distribution, the weight attached to each node being the value at H of the transform of the motif . Taken in conjunction with the assumption that the scattering is elastic, i.e. that H only changes the direction but not the magnitude of the incident wavevector , this result yields the usual forms (Laue or Bragg) of the diffraction conditions: , and simultaneously H lies on the Ewald sphere.
By the reciprocity theorem, can be recovered if F is known for all as follows [Section 1.3.2.6.5, e.g. (iv)]:
These relations may be rewritten in terms of standard, or `fractional crystallographic', coordinates by putting so that a unit cell of the crystal corresponds to , and that . Defining and by so that we have These formulae are valid for an arbitrary motif distribution , provided the convergence of the Fourier series for is considered from the viewpoint of distribution theory (Section 1.3.2.6.10.3).
The experienced crystallographer may notice the absence of the familiar factor from the expression for just given. This is because we use the (mathematically) natural unit for , the electron per unit cell, which matches the dimensionless nature of the crystallographic coordinates x and of the associated volume element . The traditional factor was the result of the somewhat inconsistent use of x as an argument but of as a volume element to obtain ρ in electrons per unit volume (e.g. Å3). A fortunate consequence of the present convention is that nuisance factors of V or , which used to abound in convolution or scalar product formulae, are now absent.
It should be noted at this point that the crystallographic terminology regarding and differs from the standard mathematical terminology introduced in Section 1.3.2.4.1 and applied to periodic distributions in Section 1.3.2.6.4: F is the inverse Fourier transform of ρ rather than its Fourier transform, and the calculation of ρ is called a Fourier synthesis in crystallography even though it is mathematically a Fourier analysis. The origin of this discrepancy may be traced to the fact that the mathematical theory of the Fourier transformation originated with the study of temporal periodicity, while crystallography deals with spatial periodicity; since the expression for the phase factor of a plane wave is , the difference in sign between the contributions from time versus spatial displacements makes this conflict unavoidable.
In many cases, is a sum of translates of atomic electron-density distributions. Assume there are n distinct chemical types of atoms, with identical isotropic atoms of type j described by an electron distribution about their centre of mass. According to quantum mechanics each is a smooth rapidly decreasing function of x, i.e. , hence and (ignoring the effect of thermal agitation) which may be written (Section 1.3.2.5.8) By Fourier transformation: Defining the form factor of atom j as a function of h to be we have If and are the real- and reciprocal-space coordinates in Å and Å−1, and if is the spherically symmetric electron-density function for atom type j, then
More complex expansions are used for electron-density studies (see Chapter 1.2 in this volume). Anisotropic Gaussian atoms may be dealt with through the formulae given in Section 1.3.2.4.4.2.
The convergence of the Fourier series for is usually examined from the classical point of view (Section 1.3.2.6.10). The summation of multiple Fourier series meets with considerable difficulties, because there is no natural order in to play the role of the natural order in (Ash, 1976). In crystallography, however, the structure factors are often obtained within spheres for increasing resolution (decreasing Δ). Therefore, successive estimates of are most naturally calculated as the corresponding partial sums (Section 1.3.2.6.10.1): This may be written where is the `spherical Dirichlet kernel' exhibits numerous negative ripples around its central peak. Thus the `series termination errors' incurred by using instead of consist of negative ripples around each atom, and may lead to a Gibbs-like phenomenon (Section 1.3.2.6.10.1) near a molecular boundary.
As in one dimension, Cesàro sums (arithmetic means of partial sums) have better convergence properties, as they lead to a convolution by a `spherical Fejér kernel' which is everywhere positive. Thus Cesàro summation will always produce positive approximations to a positive electron density. Other positive summation kernels were investigated by Pepinsky (1952) and by Waser & Schomaker (1953).
If the wavelength λ of the incident X-rays is far from any absorption edge of the atoms in the crystal, there is a constant phase shift in the scattering, and the electron density may be considered to be real-valued. Then Thus if then This is Friedel's law (Friedel, 1913). The set of Fourier coefficients is said to have Hermitian symmetry.
If λ is close to some absorption edge(s), the proximity to resonance induces an extra phase shift, whose effect may be represented by letting take on complex values. Let and correspondingly, by termwise Fourier transformation
Since and are both real, and are both Hermitian symmetric, hence while Thus , so that Friedel's law is violated. The components and , which do obey Friedel's law, may be expressed as:
By Section 1.3.2.4.3.3 and Section 1.3.2.6.10.2, Usually is real and positive, hence , but the identity remains valid even when is made complex-valued by the presence of anomalous scatterers.
If is the collection of structure factors belonging to another electron density with the same period lattice as ρ, then Thus, norms and inner products may be evaluated either from structure factors or from `maps'.
Let and be two electron densities referred to crystallographic coordinates, with structure factors and , so that
The distribution is well defined, since the generalized support condition (Section 1.3.2.3.9.7) is satisfied. The forward version of the convolution theorem implies that if then
If either or is infinitely differentiable, then the distribution exists, and if we analyse it as then the backward version of the convolution theorem reads:
The cross correlation between and is the -periodic distribution defined by: If and are locally integrable, Let The combined use of the shift property and of the forward convolution theorem then gives immediately: hence the Fourier series representation of : Clearly, , as shown by the fact that permuting F and G changes into its complex conjugate.
The auto-correlation of is defined as and is called the Patterson function of . If consists of point atoms, i.e. then contains information about interatomic vectors. It has the Fourier series representation and is therefore calculable from the diffraction intensities alone. It was first proposed by Patterson (1934, 1935a,b) as an extension to crystals of the radially averaged correlation function used by Warren & Gingrich (1934) in the study of powders.
Shannon's sampling and interpolation theorem (Section 1.3.2.7.1) takes two different forms, according to whether the property of finite bandwidth is assumed in real space or in reciprocal space.
It was shown at the end of Section 1.3.2.5.8 that the convolution theorem establishes, under appropriate assumptions, a duality between sectioning a smooth function (viewed as a multiplication by a δ-function in the sectioning coordinate) and projecting its transform (viewed as a convolution with the function 1 everywhere equal to 1 as a function of the projection coordinate). This duality follows from the fact that and map to and to (Section 1.3.2.5.6), and from the tensor product property (Section 1.3.2.5.5).
In the case of periodic distributions, projection and section must be performed with respect to directions or subspaces which are integral with respect to the period lattice if the result is to be periodic; furthermore, projections must be performed only on the contents of one repeating unit along the direction of projection, or else the result would diverge. The same relations then hold between principal central sections and projections of the electron density and the dual principal central projections and sections of the weighted reciprocal lattice, e.g. etc.
When the sections are principal but not central, it suffices to use the shift property of Section 1.3.2.5.5. When the sections or projections are not principal, they can be made principal by changing to new primitive bases B and for Λ and , respectively, the transition matrices P and to these new bases being related by in order to preserve duality. This change of basis must be such that one of these matrices (say, P) should have a given integer vector u as its first column, u being related to the line or plane defining the section or projection of interest.
The problem of constructing a matrix P given u received an erroneous solution in Volume II of International Tables (Patterson, 1959), which was subsequently corrected in 1962. Unfortunately, the solution proposed there is complicated and does not suggest a general approach to the problem. It therefore seems worthwhile to record here an effective procedure which solves this problem in any dimension n (Watson, 1970).
Let be a primitive integral vector, i.e. g.c.d. . Then an integral matrix P with det having u as its first column can be constructed by induction as follows. For the result is trivial. For it can be solved by means of the Euclidean algorithm, which yields such that , so that we may take . Note that, if is a solution, then is another solution for any . For , write with so that both and are primitive. By the inductive hypothesis there is an integral matrix V with as its first column, and an integral matrix Z with z as its first column, with and .
Now put i.e. The first column of P is and its determinant is 1, QED.
The incremental step from dimension to dimension n is the construction of matrix V, for which there exist infinitely many solutions labelled by an integer . Therefore, the collection of matrices P which solve the problem is labelled by arbitrary integers . This freedom can be used to adjust the shape of the basis B.
Once P has been chosen, the calculation of general sections and projections is transformed into that of principal sections and projections by the changes of coordinates: and an appeal to the tensor product property.
Booth (1945a) made use of the convolution theorem to form the Fourier coefficients of `bounded projections', which provided a compromise between 2D and 3D Fourier syntheses. If it is desired to compute the projection on the (x, y) plane of the electron density lying between the planes and , which may be written as The transform is then giving for coefficient :
Another particular instance of the convolution theorem is the duality between differentiation and multiplication by a monomial (Sections 1.3.2.4.2.8, 1.3.2.5.8).
In the present context, this result may be written in Cartesian coordinates, and in crystallographic coordinates.
A particular case of the first formula is where is the Laplacian of ρ.
The second formula has been used with or 2 to compute `differential syntheses' and refine the location of maxima (or other stationary points) in electron-density maps. Indeed, the values at x of the gradient vector and Hessian matrix are readily obtained as and a step of Newton iteration towards the nearest stationary point of will proceed by
The modern use of Fourier transforms to speed up the computation of derivatives for model refinement will be described in Section 1.3.4.4.7.
The converse property is also useful: it relates the derivatives of the continuous transform to the moments of : For and , this identity gives the well known relation between the Hessian matrix of the transform at the origin of reciprocal space and the inertia tensor of the motif . This is a particular case of the moment-generating properties of , which will be further developed in Section 1.3.4.5.2.
The classical results presented in Section 1.3.2.6.9 can be readily generalized to the case of triple Fourier series; no new concept is needed, only an obvious extension of the notation.
Let be real-valued, so that Friedel's law holds and . Let be a finite set of indices comprising the origin: . Then the Hermitian form in complex variables is called the Toeplitz form of order associated to . By the convolution theorem and Parseval's identity, If is almost everywhere non-negative, then for all the forms are positive semi-definite and therefore all Toeplitz determinants are non-negative, where
The Toeplitz–Carathéodory–Herglotz theorem given in Section 1.3.2.6.9.2 states that the converse is true: if for all , then is almost everywhere non-negative. This result is known in the crystallographic literature through the papers of Karle & Hauptman (1950), MacGillavry (1950), and Goedkoop (1950), following previous work by Harker & Kasper (1948) and Gillis (1948a,b).
Szegö's study of the asymptotic distribution of the eigenvalues of Toeplitz forms as their order tends to infinity remains valid. Some precautions are needed, however, to define the notion of a sequence of finite subsets of indices tending to infinity: it suffices that the should consist essentially of the reciprocal-lattice points h contained within a domain of the form (k-fold dilation of Ω) where Ω is a convex domain in containing the origin (Widom, 1960). Under these circumstances, the eigenvalues of the Toeplitz forms become equidistributed with the sample values of on a grid satisfying the Shannon sampling criterion for the data in (cf. Section 1.3.2.6.9.3).
A particular consequence of this equidistribution is that the geometric means of the and of the are equal, and hence as in Section 1.3.2.6.9.4 where denotes the number of reflections in . Complementary terms giving a better comparison of the two sides were obtained by Widom (1960, 1975) and Linnik (1975).
This formula played an important role in the solution of the 2D Ising model by Onsager (1944) (see Montroll et al., 1963). It is also encountered in phasing methods involving the `Burg entropy' (Britten & Collins, 1982; Narayan & Nityananda, 1982; Bricogne, 1982, 1984, 1988).
References
Ash, J. M. (1976). Multiple trigonometric series. In Studies in harmonic analysis, edited by J. M. Ash, pp. 76–96. MAA studies in mathematics, Vol. 13. The Mathematical Association of America.Google ScholarBooth, A. D. (1945a). Two new modifications of the Fourier method of X-ray structure analysis. Trans. Faraday Soc. 41, 434–438.Google Scholar
Bricogne, G. (1974). Geometric sources of redundancy in intensity data and their use for phase determination. Acta Cryst. A30, 395–405.Google Scholar
Bricogne, G. (1982). Generalised density modification methods. In Computational crystallography, edited by D. Sayre, pp. 258–264. New York: Oxford University Press.Google Scholar
Bricogne, G. (1984). Maximum entropy and the foundations of direct methods. Acta Cryst. A40, 410–445.Google Scholar
Bricogne, G. (1988). A Bayesian statistical theory of the phase problem. I. A multichannel maximum entropy formalism for constructing generalised joint probability distributions of structure factors. Acta Cryst. A44, 517–545.Google Scholar
Britten, P. L. & Collins, D. M. (1982). Information theory as a basis for the maximum determinant. Acta Cryst. A38, 129–132.Google Scholar
Friedel, G. (1913). Sur les symétries cristallines que peut révéler la diffraction des rayons Röntgen. C. R. Acad. Sci. Paris, 157, 1533–1536.Google Scholar
Gillis, J. (1948a). Structure factor relations and phase determination. Acta Cryst. 1, 76–80.Google Scholar
Gillis, J. (1948b). The application of the Harker–Kasper method of phase determination. Acta Cryst. 1, 174–179.Google Scholar
Goedkoop, J. A. (1950). Remarks on the theory of phase-limiting inequalities and equalities. Acta Cryst. 3, 374–378.Google Scholar
Harker, D. & Kasper, J. S. (1948). Phases of Fourier coefficients directly from crystal diffraction data. Acta Cryst. 1, 70–75.Google Scholar
Karle, J. & Hauptman, H. (1950). The phases and magnitudes of the structure factors. Acta Cryst. 3, 181–187.Google Scholar
Linnik, I. Ju. (1975). A multidimensional analogue of a limit theorem of G. Szegö. Math. USSR Izv. 9, 1323–1332.Google Scholar
MacGillavry, C. H. (1950). On the derivation of Harker–Kasper inequalities. Acta Cryst. 3, 214–217.Google Scholar
Montroll, E. W., Potts, R. B. & Ward, J. C. (1963). Correlations and spontaneous magnetization of the two-dimensional Ising model. J. Math. Phys. 4, 308–322.Google Scholar
Narayan, R. & Nityananda, R. (1982). The maximum determinant method and the maximum entropy method. Acta Cryst. A38, 122–128.Google Scholar
Onsager, L. (1944). Crystal statistics. I. Two-dimensional model with an order–disorder transition. Phys. Rev. 65, 117–149.Google Scholar
Patterson, A. L. (1934). A Fourier series method for the determination of the components of interatomic distances in crystals. Phys. Rev. 46, 372–376.Google Scholar
Patterson, A. L. (1935a). A direct method for the determination of the components of interatomic distances in crystals. Z. Kristallogr. 90, 517–542.Google Scholar
Patterson, A. L. (1935b). Tabulated data for the seventeen plane groups. Z. Kristallogr. 90, 543–554.Google Scholar
Patterson, A. L. (1959). In International tables for X-ray crystallography, Vol. II, pp. 9–10. Erratum, January 1962. Birmingham: Kynoch Press. (Present distributor Kluwer Academic Publishers, Dordrecht.)Google Scholar
Pepinsky, R. (1952). The use of positive kernels in Fourier syntheses of crystal structures. In Computing methods and the phase problem in X-ray crystal analysis, edited by R. Pepinsky, pp. 319–338. State College: Penn. State College.Google Scholar
Sayre, D. (1952b). Some implications of a theorem due to Shannon. Acta Cryst. 5, 843.Google Scholar
Sayre, D. (1952c). The Fourier transform in X-ray crystal analysis. In Computing methods and the phase problem in X-ray crystal analysis, edited by R. Pepinsky, pp. 361–390. State College: Penn. State University.Google Scholar
Warren, B. E. & Gingrich, N. S. (1934). Fourier integral analysis of X-ray powder patterns. Phys. Rev. 46, 368–372.Google Scholar
Waser, J. & Schomaker, V. (1953). The Fourier inversion of diffraction data. Rev. Mod. Phys. 25, 671–690.Google Scholar
Watson, G. L. (1970). Integral quadratic forms. Cambridge University Press.Google Scholar
Widom, H. (1960). A theorem on translation kernels in n dimensions. Trans. Am. Math. Soc. 94, 170–180.Google Scholar
Widom, H. (1975). Asymptotic inversion of convolution operators. Publ. Math. IHES, 44, 191–240.Google Scholar