Short History of Powder Diffractionw3.esfm.ipn.mx/~fcruz/ADR/Cristalografia/Roisnel... ·...
Transcript of Short History of Powder Diffractionw3.esfm.ipn.mx/~fcruz/ADR/Cristalografia/Roisnel... ·...
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Short History of Powder Diffraction
1895 Discovery of X-rays by Roentgen 1912 von Laue demonstrates that X-rays can be diffracted by crystals1935 Le Galley constructs the first X-ray powder diffractometer1947 Phillips introduces the first commercial powder diffractometer1950’s Powder diffraction used primarily to study structural 1960’s imperfections, phase identification, … largely by metallurgists
and mineralogists1969 Hugo Rietveld develops a method for whole pattern analysis of
neutron powder diffraction data1977 Cox, Young, Thomas and others first apply Rietveld method
to synchrotron and conventional X-ray data
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Commercial Powder Diffractometer
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Uses of Powder DiffractionQualitative Analysis Phase IdentificationQuantitative Analysis Quantitative
Lattice Parameter(indexing&refinement)
Phase Fraction AnalysisStructure Determination Structure
Reciprocal Space Methods Real Space Methods
Structure Refinement StructureRietveld Method
Peak Shape Analysis Peak Shape Crystallite Size DistributionMicrostrain AnalysisAnti-phase domains, stacking faults,…
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Limitations of Powder Diffraction for solving crystal structures
• Single crystal diffraction allows solving crystal structures by using appropriate mathematical algorithms and accurate peak intensities.•The 3D set of reflections obtained from a single crystal experiment is condensed into 1D in powder diffraction pattern. This leads to both accidental and exact peak overlap, and complicates the determination of individual peak intensities.• Indexing may be a bottleneck for starting to solve a crystal structure. Multiphase mixtures complicates the task.• Crystal symmetry cannot be obtained directly from powder diffraction patterns.•Preferred orientation leads to biased peak intensities.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Applications of powder diffractionPhysics, Chemistry, Geology, Industry,...
• Phase identification (mineralogy)
• Quantitative analysis (chemical and phase analysis)
• Texture determination of polycrystalline materials
• Residual stress analysis
• Structure determination
• Microstructure and defects
• Structural behaviour under complex environments
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Neutron/Synchrotron Powder Diffraction for what?
High precision crystallography with powders in complex environments: high pressure, low and high temperatures. Structural studies of phase transition and transformations
Real crystals: defects, diffuse scattering, total scattering analysis.
Magnetic structures and phase diagrams (NEUTRONS)
Time-dependent phenomena: in situ powder diffraction
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Solid State Physics and Chemistry: recent hot topics where powder diffraction has played a
fundamental role
• Transition Metal and Rare Earth Mixed Oxides
• High Tc superconductors and related materials
• Metal-Insulator transitions
• Giant Magneto-Resistance materials
• Low-dimensional systems
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Neutron and synchrotron Application of powder diffraction
• Hydrogen in metals• Magnetic intermetallics• Molecular solids and fullerenes• Composite and technological materials:
cements, ceramics, zeolites, catalysts, opto-electronic• Chemical reactions: solid-solid, liquid-solid, gas-solid• Pharmaceuticals compounds and bio-compatible materials• Supramolecular magnetic solids• Ionic conductors, solid electrolytes. Chemical processes in
solid state batteries.• Aperiodic materials: quasicrystals, modulated structures.• Diluted magnetic semiconductors.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Scheme of a two axis diffractometer
Collimators α1, α2, α3
Mosaicity of monochromator βM
« take-off » angle : 2θM
Parameters determining resolution and intensity
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
A powder diffraction pattern can be recorded in numerical form for a discrete set of scattering angles, times of flight orenergies. We will refer to this scattering variable as : T. The experimental powder diffraction pattern is usually given as three arrays :
The profile can be modelled using the calculated counts: yciat the ith step by summing the contribution from neighbouring Bragg reflections plus the background.
1,2,..., ,i i i i n
T y σ=
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
yi
Position “i”: Ti
Bragg position Th
yi-ycizero
Powder diffraction profile:scattering variable T: 2θ, TOF, Energy
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
∆(Q)=2π2 j / (Q2 Vo) ≈ p DQ
Average separation (in Å-1) between adjacent reflections
j General multiplicity of the Laue classVo Volume of the primitive cell in Å3
Resolution Function in reciprocal spaceDQ= FWHM in Å-1
of Bragg peaks
Well separated reflections means their
maxima are distant more than pDQ
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Constant wavelength neutron powder diffraction
• Resolution function determined by optical diffractometer arrangements and collimators. Coupled to scattering angle and wavelength. • Limited Q-range: depend on the selected wavelength. In a hot source (low lambda) the gain in Q-range is degraded by lower resolution.• Difficult to obtain high resolution in the whole Q-range in a single instrument.• Simple data treatment. Minimal corrections of the raw data before processing.• Simple model of peak shape and faster calculations in data analysis.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Time of Flight neutron powder diffraction
• Resolution function determined by the pulse width and the flight path.• Large Q-range with excellent resolution for long flight path instruments.• Simple to obtain high resolution in the whole Q-range in a single instrument, using several banks. • Spectro-diffractometers allows to study the dynamics together with the structural aspects.• Complex data treatment. Important corrections of the raw data before processing.• Complex models of peak shape and lengthy calculations in data analysis.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Neutrons versus Synchrotron: advantages and drawbacks for Powder Diffraction
Constant scattering length. Contrast.Low absorption: easy sample environmentMagnetic structuresHigh precision in structure refinementModerate resolution
Extremely high resolutionSubtle distortionsIndexing and Structure determinationAnomalous scatteringTexture effects
Neutrons Synchrotron X-rays
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
How and where finding information about powder
diffraction?
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Downloading of SoftwareThe above software and resources can be downloaded via the CCP14
site:http://www.ccp14.ac.uk
Graphical tutorial run-through of most of this software is located via (“look before you try”):
http://www.ccp14.ac.uk/tutorial/
Freely available softwaretools for crystallography
Lachlan M.D. Cranswick ([email protected])CCP14 Project for Single Crystal and Powder Diffraction
http://www.ccp14.ac.uk [email protected]
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Armel Le Bail page: http://sdpd.univ-lemans.fr
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
In scattering experiments, the incident particle (neutron, electron, photon...) experiences a change in its momentum and energy.
(h/2π)Q =(h/2π)(kF-kI)= h s hν=EF-EI
In the following we shall be concerned with elastic scattering (hν=0) for which kF=kI= 2π/λ and Q= Q = (4π/λ) sinθ , θ being half the scattering angle.
“crystallographic scattering vector”: s = Q/2π.
Some equations and formalism (1)
kI=2π/λ uI
kF=2π/λ uFQ= kF - kI
2θ
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
In the kinematics theory (first Born approximation), the amplitude of the wave scattered by an object is the Fourier transform (FT) of its scattering density (SD) ρ(r) measured in cm-2. Any object can be considered as constituted by atoms of SD ρaj(r) centred at positions Rj; the SD and the corresponding scattered amplitude and intensitycan be written as:Scattering density: ρ(r) =∑j ρaj(r-Rj)Scattered amplitude: A(s)=FT[ρ(r) ]
A(s)= ∑jexp2πisRj∫ρaj(u) exp2πisud3u
A(s) = ∑ j fj(s) exp2πi s Rj
I(s) = A(s)A(s)* = ∑ i ∑ j fi(s) fj(s)* exp2πi s (Ri-Rj)
Some equations and formalism (2)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Magnetic scattering of neutrons
Dipolar interaction (µn , m): vector scattering amplitude
( ) ( ) ( )2
12
Q m Qa Q Q mM er f
Qγ
= −
kI=2π/λ uI
kF=2π/λ uF
Q= kF - kI
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Magnetic scattering of neutrons
( ) ( ) ( ) ( )2
12
Q m Qa Q Q m Q mM er f p f
Qγ ⊥
= − =
( ) ( ) 3exp( )Q r Q r rmf i dρ= ∫m
m⊥
Q=Q eOnly the perpendicularcomponent of m to Q=2πh
contributes to scattering
p=0.2696 10-12 cm
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
For powders, we have to average the intensity for all possible orientations of an object with respect to the incident beam. Theintensity depends on the length, s, of s and the whole set of inter-atomic distances Rij=Ri-Rj; and is given by the Debye formula in terms of Q=2πs:
I(Q) = ∑i ∑ j fi fj sinQRij /(QRij)
Some equations and formalism (3)
Debye Formula
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Total scattering density of the infinite object can be decomposed in the following way:
ρ∞(r) =∑n ρc(r-Rn) = ρc(r) * ∑n δ(r-Rn) = ρc(r) * z(r)
For a finite crystal z(r) must be replaced by z(r)g(r) , where g(r) is the shape factor of the crystal defined as g(r)=1 for r inside the crystal, and g(r)=0 for r outside. We define G(s)=FTg(r). The scattered amplitude for a finite crystal is:
A(s)=FTρf(r) Z(s) = FTz(r) = 1/Vc ∑H δ(s-H)
A(s)= FTρc(r) * z(r)g(r) = F(s)Z(s) * G(s) = F(s)/Vc∑HG(s-H)
Laue conditions:s=H ⇒Bragg Law (s=s=2sinθ/λ, H=1/dhkl)
Some equations and formalism (4)Scattering Amplitude of a Finite Crystal
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
I(s)=F2(s)/V2c ∑ HG(s-H) ∑ H'G*(s-H') ≈ NF2(s)/(VVc)∑HG2(s-H)
G2(s) is Fourier transform of the auto-correlation function:
V.η(r)=∫g(u)g(r+u)d3u
rThe interpretation of η(r) is straightforward: it represents the fraction of the total volume shared in common between the object and its "ghost" displaced by the vector r. Obviously, η(0)=1 and decreases as r increases.
Some equations and formalism (5)Scattered Intensity of a Finite Crystal
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Defects: average lattice + structure factor depending on unit cell
I(s) =∑n(∑ m FmF*m+n) exp2πi s Rn
Fm is the structure factor of the cell m. Taking into account the long range homogeneity of the object, the average value pn=⟨FmF*
m+n⟩ is independent of m. The number of terms in the inner sum is given by Vη(Rn)/Vc and the equation can be transformed to:
I(s) = V/Vc ∑ nη(Rn) ⟨FmF*m+n⟩ exp2πi s Rn=
= N ∑nη(Rn) pn exp2πi s Rn
Some equations and formalism (6)Scattered Intensity of a Real Crystal
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Average structure factor: F= ⟨F⟩ =1/N ∑mFm
local fluctuation of the structure factor: φn=F-Fn
pn=F2+⟨φmφ*m+n⟩ = F2+Φn,
The intensity formula can be transformed to:
I(s) = IBragg + IDiffuse =
N F2 ∑ nη(Rn) exp 2πi s Rn + N ∑ nη(Rn) Φn exp2πi s Rn =
N F2/(VVc) ∑ HG2(s-H) + N ∑ n Φn exp2πi s Rn
In the last expression, we have made the approximation η(Rn)=1 because
Φn decreases with n faster than η(Rn).
Some equations and formalism (7)Scattered Intensity of a Real Crystal
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Fm=F exp2πi s um, with the constraint: ∑ mexp2πi s un=0
Defining: ζ(Rn, s)= ⟨exp2πis(um-um+n)⟩.
The intensity formula can be transformed to:
I(s) = I(s) = N F2 ∑ nη(Rn) ζ(Rn, s) exp2πi s Rn
If s=H+∆s the scattered intensity around a Bragg peak is given by:
IH(∆s) =N F2H ∑nη(Rn) ζH(Rn) exp2πi ∆s Rn ≈ F2
H Ωx(∆s)
size strain
Some equations and formalism (8)Strained Crystals
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Reciprocal space associated to an Ideal Powder: set of spherical
shells whose intersections with the Ewald sphere produces scattering
distributed in cones having as a common axis the incident beam.
Single crystal intensity around a single Bragg peak:
IH(∆s) ≈ F2H Ωx(∆s) 3D
For a powder (after spherical averaging), set of equivalent nodes :
IH(∆s) ≈ ∑H F2H ΩH (∆s) 1D
Some equations and formalism (9)Powder Average
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The Rietveld Method for refinement of crystal and
magnetic structures
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
profile
The model to calculate a powder diffraction pattern is:
( )h hh
ci i iy I T T b= Ω − +∑( ) 1x dx
+∞
−∞Ω =∫
Profile function characterized by its full width at half maximum (FWHM=H)and shape parameters (η, m, ...)
( ) ( ) ( )x g x f x instrumental intrinsic profileΩ = ⊗ = ⊗
The profile of powder diffraction patterns
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The profile of powder diffraction patterns
( )h hh
c i iiy I T T b= Ω − +∑Contains structural information: atom positions, magnetic moments, etc( )h h II I= β
( , )h PixΩ = Ω β Contains micro-structural information: instr. resolution, defects, crystallite size, ..
( )Bi ib b= β Background: noise, diffuse scattering, ...
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
, ,h
( )h hci i iy s I T T bφ φ φφ
= Ω − +∑ ∑
Several phases (φ = 1,nφ) contributing to the diffraction pattern
, ,h
( )h h
p p p p pci i iy s I T T b
φφ φφ
= Ω − +∑ ∑
Several phases (φ = 1,nφ) contributing to several (p=1,np) diffraction patterns
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The Rietveld Method consist of refining a crystal (and/or magnetic) structure by minimising the weighted squared difference between the observed and the calculated pattern against the parameter vector: β
22
1( )
n
i i cii
w y yχ β=
= −∑
21i
iwσ
=2iσ : is the variance of the "observation" yi
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Least squares: Gauss-Newton (1)Minimum necessary condition:
A Taylor expansion of around allows the application of an iterative process. The shifts to be applied to the parameters at each cycle for improving χ2 are obtained by solving a linear system of equations (normal equations)
2
0∂=
∂χβ
( )icy β 0β
0
0 0
0
( ) ( )
( )( )
A b
ic ickl i
i k l
ick i i ic
i k
y yA w
yb w y y
β β
β
=
∂ ∂=
∂ ∂∂
= −∂
∑
∑
βδ
β β
β
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Least squares: Gauss-Newton (2)
The new parameters are considered as the starting ones in the next cycle and the process is repeated until a convergence criterion is satisfied. The variance of the adjusted parameters are calculated by the expression:
The shifts of the parameters obtained by solving the normal equations are added to the starting parameters giving rise to a new set
01 0= + ββ β δ
1( ) ( )Ak kk
N - P+C
2 2ν
22
ν
σ β χ
χχ
−=
=
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Least squares: a local optimisation method
• The least squares procedure provides (when it converges) the value of the parameters constituting the local minimum closest to the starting point
• A set of good starting values for all parameters is needed
• If the initial model is bad for some reasons the LSQ procedure will not converge, it may diverge.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Profile R-factors used in Rietveld Refinements
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
R-factors and Rietveld Refinements (1)
, ,
,
100obs i calc i
ip
obs ii
y yR
y
−=
∑∑
1/ 22, ,
2,
100i obs i calc i
iwp
i obs ii
w y yR
w y
−
=
∑
∑
R-pattern
R-weighted pattern
1/ 2
2,
( )100expi obs i
i
N P CRw y
− + =
∑
Expected R-weighted pattern
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
R-factors and Rietveld Refinements (2)
Reduced Chi-square
Goodness of Fit indicator
2
2 wp
exp
RRνχ
=
wp
exp
RS
R=
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
R-factors and Rietveld Refinements (3)
Two important things:
• The sums over “i” may be extended only to the regions where Bragg reflections contribute
• The denominators in RP and RWP may or not contain the background contribution
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Crystallographic R-factors used in Rietveld Refinements
, ,
,
' '100
' '
obs k calc kk
Bobs k
k
I IR
I
−=
∑∑
Bragg R-factor
, ,
,
' '100
' '
obs k calc kk
Fobs k
k
F FR
F
−=
∑∑
Crystallographic RF-factor.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Crystallographic R-factors used in Rietveld Refinements
Provides ‘observed’integrates intensities for calculating Bragg R-factor
In some programs the crystallographic RF-factor is calculated using just the square root of ‘Iobs,k’
,, ,
,
( )( )' '
( )i k obs i i
obs k calc ki calc i i
T T y BI I
y B Ω − − = −
∑
,,
' '' ' obs k
obs k
IF
jLp=
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The structural information contained in the integrated
intensities
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
2h h
I L pO ACF=
Integrated intensities are proportional to the square of the structure factor F. The factors are: Lorentz-polarization (Lp), preferred orientation (O), absorption (A), other “corrections” (C) ...
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The Structure Factor contains the structural parameters
( ) ( ) 1
2h h h t rn
j j j jsj s
F O f T exp i Sπ=
= ∑ ∑
( , , ) ( 1, 2, ... )r j j j jx y z j n= =
( ) ( ) ( ), , , , 1, 2, ...h Ts s Gs
h k l h k l S s N= = =
sinexp( )j jT B2
2
θλ
= −
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Structural Parameters(simplest case)
( , , )r j j j jx y z= Atom positions (up to 3nparameters)
jO Occupation factors (up to n-1 parameters)
jB Isotropic displacement (temperature) factors (up to n parameters)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Structural Parameters(complex cases)
As in the simplest case plus additional (or alternative) parameters:
• Anisotropic temperature (displacement) factors
• Anharmonic temperature factors
• Special form-factors (Symmetry adapted spherical harmonics ), TLS for rigid molecules, etc.
• Magnetic moments, coefficients of Fourier components of magnetic moments , basis functions, etc.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Magnetic scattering:
Intensity (non-polarised neutrons)
Magnetic interaction vector:
*hhhhh MM ⊥⊥ ⋅+= *NNI
( )( ) ( ) ( )( )hMeehMehMeM h ⋅−=××=⊥
kHh += ⇐ Scattering vector
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
( ) ( ) ( ) 1
2k kM h h S H k t rn
j j j js j j jssj s
p O f T M exp i Sπ ψ=
= + − ∑ ∑
The magnetic structure factor:
Magnetic structures
2kk
m S kRlj j lexp iπ= −∑
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Electronic crystallization in a Li battery material: columnar ordering of electron and
holes in the spinel LiMn2O4
J. Rodríguez-Carvajal, G. Rousse, Ch. Masquelier and M. HervieuPhysical Review Letters, 81, 4660 (1998)
LiTd [Mn2]Oct.O4 : Mn3.5+
High temperature: mixed valence state
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
LiMn2O4 : Electron Diffraction
320 K 230 K
J. Rodríguez-Carvajal et al, PRL, 81, 4660 (1998)
Cubic Fd3ma = 8,248 (1) Å
Orthorhombic Fddda = 24.7435(5) Åb = 24.8402(5) Åc = 8.1989(1) Å
x 9
0 12 0
12 0 0
040
400
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
30 40 50 60 70 80
230K
350K
2 theta (°)
LiMn2O4 : Neutron Diffraction, 3T2 (LLB)
Charge ordered state
Orthorhombic Distortion Superstructure reflections
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
<Mn-O> = 1,915(4) ÅMn(5) = 3.90+
∆ = 6.1
a
bc
<Mn-O> = 1,903(4) ÅMn(4) = 4.02+
∆ = 4.6
64 « Mn4+ »80 « Mn3+ -like»
8 delocalised holes
<Mn-O> = 1,996(4) ÅMn(2) = 3.27+
∆ = 19.4
<Mn-O> = 2,020(5) ÅMn(3) = 3.12+
∆ = 36.6
<Mn-O> = 2,003(2) ÅMn(1) = 3.20+
∆ = 20.6
LiMn2O4 : Partial Charge Ordering
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
b
222
33
33
33
33
22
22
22
22
1 133
33
33
33
22
22
1
22
22
33
33
33
33
22
22
122
22
33
33
33
33
1
22
22
1
a
2
1 1 1 1 1
1 1
1 1
1 1
1 1 1 1 1
111
x
x x
x xx
x
x
x xx x
xxx x x x x
x x
x
x
xxxx
x x x
xx
x
xx x
x x
x
x xx
xxxxxx
x x x x x xx x x x x x
x
x x
xx
xx
x
x
x
x
x
Mn3+
Mn4+
X/ Li+
c
Mn(3) Mn(2)
LiMn2O4 : Partial Charge Ordering
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Nature of the phase transition in LiMn2O4
Nature of the phase transition in Nature of the phase transition in LiMnLiMn22OO44
The structural transition is driven by a charge ordering process. The Coulomb repulsion is however of secondary importance compared to electron-lattice coupling via theJahn-Teller effect
The structural distortions and average distances around the Mn ions support the atomic scale charge ordered state below the transition temperature.
The distorted pyrochlore lattice of Mn-ions is then half doped with a partial charge ordering at low temperature. The magnetic ground state is incommensurate and very complex.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Behaviour of Rietveld R-factors, and other indicators, versus counting statistics for
perfect and biased models
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
0
5
10
15
20
25
30
35
40
2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5
R-factors (RP, R
WP, R
Bragg and R
F)
cRp(B)cRwp(B)RbRF
R(%
)
Log(Counting Time)
NIntdp=16, Nprof=8, Nref=393Neff=(173, 244, 301)
Correct model
RM (systematic errors): Behaviour of R-factors versus counting time (correct model)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
RM (systematic errors): Behaviour of RWP factors versus counting time (biased peak shape)
0
10
20
30
40
50
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
R-factors (RWP
) RwpcRwpRwp(B)cRwp(B)R
WP(%
)
Log(Counting Time)
NIntdp=16, Nprof=8, Nref=393Neff=(173, 244, 301)Biased peak shape
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
0
2
4
6
8
10
12
14
2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5
R-factors (RBragg
and RF)
RbRF
R(%
)
Log(Counting Time)
NIntdp=16, Nprof=8, Nref=393Neff=(173, 244, 301)Biased peak shape
RM (systematic errors): Behaviour of RBragg and RFfactors versus counting time (biased peak shape)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
RM (systematic errors): Behaviour of reduced Chi-square versus counting time (correct model)
0,600
0,800
1,000
1,200
1,400
2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5
R-factors(χ 2)
Chi2Chi2(B)
Red
uced
χ2
Log(Counting Time)
NIntdp=16, Nprof=8, Nref=393Neff=(173, 244, 301)
Correct model
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
0
10
20
30
40
50
60
2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5
R-factors(χ2)
Chi2Chi2(B)
Red
uced
χ2
Log(Counting Time)
NIntdp=16, Nprof=8, Nref=393Neff=(173, 244, 301)Biased peak shape
RM (systematic errors): Behaviour of reduced Chi-square versus counting time (biased peak shape)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The peak shape function of powder diffraction patterns
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The peak shape function of powder diffraction patterns contains the Profile Parameters
(constant wavelength case)
( , ) ( , )h P h Pi ix T Tβ βΩ = Ω −
( ) 1x dx+∞
−∞Ω =∫
The cell parameters are included, through Th, within the profile function. They determine the peak positions in the whole diffraction pattern.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
2( ) exp( )G GG x a b x= −
2
2 ln2 4ln2G Ga b
H Hπ= =
Gaussian function
Integral breadth:1
2 ln 2G
Ha
πβ = =
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
2( )1
L
L
aL xb x
=+
2
2 4L La b
H Hπ= =
Lorentzian function
Integral breadth:1
2L
Ha
πβ = =
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Fwhm
BG
x0
x
I
Comparison of Gaussian and Lorentzian peak shapes of the same peak height “I” and same width “Fwhm”
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Convolution properties of Gaussianand Lorentzian functions
1 2 1 2
2 21 2 1 2
( , ) ( , ) ( , )
( , ) ( , ) ( , )
L x H L x H L x H H
G x H G x H G x H H
⊗ = +
⊗ = +
( , ) ( , ) ( , , )L G L GL x H G x H V x H H⊗ =
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
( ) ( ) (1 ) ( )pV x L x G xη η′ ′= + −The pseudo-Voigt function
The Voigt function
( ) ( ) ( ) ( ) ( )V x L x G x L x u G u du+∞
−∞= ⊗ = −∫( ) ( , , ) ( , , )L G L GV x V x H H V x β β= =
( ) ( , , )pV x pV x Hη=
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Mapping: Pseudo-Voigt ⇔ VoigtThomson-Cox-Hasting formulation
( , ) ( , )G LH F H Hη =
1( , ) ( , )G LH H F H η−=
5 4 3 2 2 3 4 5( 2.69269 2.42843 4.47163 0.07842 )G G L G L G L G L LH H H H H H H H H H H= + + + + +2 3
1.36603 0.47719 0.11116L L LH H HH H H
η = − +
2 30.72928 0.19289 0.07783LHH
η η η= + +
2 3 1/ 2(1 0.74417 0.24781 0.00810 )GHH
η η η= − − −
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
2 22tan tan
cosG
GIH U V Wθ θ
θ= + + +
tancosL
YH X θθ
= +
Profile Parameters(simple cases)
Parameters controlling the Full-Width at half maximumU, V, W, IG, X, Y
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Microstructural effectsAnisotropic peak
broadening
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
2 2 2 22( (1 ) ) tan tan
cosG
G STIH U D V Wξ θ θ
θ= + − + + +
[ ( )]( ) tancos
ZL ST
Y F SH X Dξ θθ
+= + +
Modeling the Gaussian and Lorentziancomponents of the profile function in terms of microstructural parameters
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
xy
LhΩ
h
The intrinsic profile of a particular reflection due to size effect has an integral breadth , the Scherrer formula:
gives the volume-averaged apparent size of the crystallites in the direction normal to the scattering planes. This apparent size has a perfectly defined physical interpretation:
in terms of the normalized column-length distributionpV(L):
*
1cosV
S S
D λβ θ β
= =
( ) 3
1,...
1 1 ,i
Vi N i C
D L x y dN V=
= ∑ ∫∫∫ h r
0
( )VVD L p L dL
∞
= ∫
Sβ
Anisotropic broadening due to size effects
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Special expressions for anisotropic broadening due to size effects:
Microstructural effects are simulated in the Rietveld Method using the Voigt approximation(Langford & Louër).
• Infinite needles • Infinite platelets• Finite cylinders
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Spherical harmonics to simulate the average form of crystallites
( )cos1 cos ; /sinlmp lm
lmp
ma P p
mDΦ
= Θ = + − Φ ∑ h
hhh
( ),Θ Φh h : Polar angles of reciprocal vector h w.r.t. crystal frame
( )cos
cossincos lmp lm
lmp
mkFWHM a Pm
λθ
Φ = Θ Φ
∑ hh
h
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The intrinsic profile of a particular reflection due to a straineffect has an integral breadth , the apparent strain is defined as
We shall use the so called maximum strain, that is derived from the apparent strain as:
cotDη β θ=Dβ
14 2 D
de dd
η β∆= = =
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Anisotropic strain broadening
Phenomenological model: strains considered as fluctuations and correlation between metric parameters J. Rodríguez-Carvajal et al (J. Phys. Cond. Matt. 3, 3215 (1991)
( )2
1 ;hkl ihkl
M M hkld
α= =
The metric parameters αi (direct, reciprocal or any combination) are considered as stochastic variables with a Gaussian distributioncharacterized by :• the mean ⟨ αi⟩ and • the variance-covariance matrix Cij
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The mean and the variance of the function Mhklare given by (JRC et al , J. Phys. Cond. Matt. 3, 3215 (1991) ):
( );hkl iM M hklα=
( )2
,hkl ij
i j i j
M MM Cσα α
∂ ∂=
∂ ∂∑If the metric parameters are taken as the coefficients of the quadratic form: 2 2 2
2
1
hkl
Ah Bk Cl Dkl Ehl Fhkd
= + + + + +
( )
2
4
H K Lhkl HKL
HKLH K L
M S h k lσ+ + =
= ∑P. W. Stephens,J. Appl. Cryst. 32, 281 (1999)
Cij contains 21 parameters, 15 independent
αi=A,B,C,D,E,F
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
S_400 S_040 S_004 S_22022.04(78) 17.74(57) 0.016(2) -38.8(1.2)Lorentzian Parameter: 0.093(2)
Nd2NiO4, LT
A-strain h k l43.4585 0 1 248.1172 1 0 2 7.1018 1 1 0 5.9724 1 1 1 4.1383 1 1 2 9.7952 0 0 4 4.0162 1 1 379.5271 0 2 087.5578 2 0 0
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Portion of the neutron diffraction pattern of Pd3MnD0.8 at room temperature obtained on 3T2 (LLB, λ = 1.22 Å). On top, the comparison with the calculated profile using the resolution function of the instrument. Below the fit using IsizeModel = -14. Notice that only the reflections with indices of different parity are strongly broadened. An isotropic strain, due to the disorder of deuterium atoms, is also included for all kind of reflections.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Classically, crystal structure determination is considered as a process to determine the “phases” of the structure factors
( )
( ) ( ) ∑
∑Φ+−=
−=
hhh
hh
hrr
hrr
iexpF
iexpF
πρ
πρ
2
2
For a centrosymmetric structure is 0 or 1/2
The knowledge of all phases for the measured structure factors provides a density map from which the structure is derived (chemically recognised).
hΦ
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
The resolution of the “phase problem” is the goal of the crystal structure determination methods
Direct Methods tackle the problem looking for phase relations (tangent formula) between
structure factors of different reflections
Direct methods need a high number of reflections and good resolution (powders)
Direct methods are generally very efficient
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
But …sometimes direct methods fail in
solving particular structuresor
cannot be applied because poor data quality (low resolution)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Structure factor calculation
If the chemical composition and the space group are known we have to determine just the 3n variables
( ) ( ) [ ] ∑∑=
=s
js
n
jjjj SiexpTfOF rthhh π2
1
( ) ( )n,...,jz,y,x jjjj 21==r
( ) [ ] rs
js
n
jjjjrobs SiexpTfOF ∑∑
=
≈ rthh π21
( ) ( )N,...,rl,k,h rr 21==h
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Direct space methods:
•Look directly for atom positions explaining the experimental data
•Minimize a reliability factor with respect to the
“configuration vector” or “chromosome”
nnn z,y,x,...z,y,x,z,y,x 222111=ϖ
( ) ( ) ( )∑=
−=N
rrcalcrobs,FFcR
1
22 ϖϖ hh
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Solving Crystal structures by Powder diffraction and direct space methods(1) Indexing the
powder diffraction pattern: DICVOL, TREOR, ITO,…(WinPLOTR)
(2) Extracting integrated intensities (FullProf)
Space Group determination
(4) If ExPo fails thenUse Simulated
Annealing (FullProf) or Genetic Algorithms
(3) Use ExPo to solve the structure
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Simulated Annealing (SA):
The SA method is a general purpose optimization techniquefor large combinatorial problems introduced by:
Kirpatrick, Gelatt and Vecchi, Science 220, 671-680 (1983).
The function, E(ω) to be optimized with respect to the configuration described by the vector state ω is called the “cost” function.
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Simulated Annealing (SA):
The SA method applied to structural problems:
•J. Pannetier, J. Bassas-Alsina, J. Rodríguez-Carvajaland V. Caignaert, Nature 346, 343-345 (1990)
•J.M. Newsam, M.W. Deem and C.M. Freeman, Accuracy in Powder Diffraction II. NIST Special Publ. No. 846, 80-91 (1992)
•J. Rodríguez-Carvajal, Physica B 192, 55-69 (1993)
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
beginInitialise (set to zero useful quantities, do preliminary calculations )τ = 1do
doPerturb the system:ωold → ωnew, ∆=E(ωnew)-E(ωold)
if ∆ ≤ 0 then accept, elseif exp(-∆/Ττ) > random[0,1] then acceptif accept then Update (replace ωold by ωnew)
until equilibrium is approached closely enough (Ncyc)Tτ+1 = f(Tτ) (decrease temperature, usually Tτ+1 = q Tτ, q≈0.9)τ = τ + 1
until stop criterion is true (maximum τ, convergence, low % accepted...)end
The Simulated Annealing Algorithm
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
COMM Ab initio structure solution of PbSO4 (Simulated Annealing, data D1A-ILL)! Files => DAT-file: pb_san, PCR-file: pb_san!Job Npr Nph Nba Nex Nsc Nor Dum Iwg Ilo Ias Res Ste Nre Cry Uni Cor Opt Aut
1 0 1 0 0 0 0 0 0 0 0 0 0 12 3 0 0 0 0!Ipr Ppl Ioc Mat Pcr Ls1 Ls2 Ls3 Syo Prf Ins Rpa Sym Hkl Fou Sho Ana
0 0 1 0 1 0 0 0 0 1 0 0 0 0 0 0 0!!NCY Eps R_at R_an R_pr R_gl Thmin Step Thmax PSD Sent0
1 0.10 1.00 1.00 1.00 1.00 15.0000 0.0200 120.0400 0.000 0.000!
12 !Number of refined parameters!-------------------------------------------------------------------------------! Data for PHASE number: 1 ==> Current R_Bragg for Pattern# 1: 7.86!-------------------------------------------------------------------------------PbSO4
!!Nat Dis Mom Pr1 Pr2 Pr3 Jbt Irf Isy Str Furth ATZ Nvk Npr More
5 0 0 0.0 0.0 1.0 0 4 0 0 0 0.00 0 7 0!P n m a <--Space group symbol!Atom Typ X Y Z Biso Occ In Fin N_t Poi /CodesPb PB 0.81174 0.23348 0.83479 1.42124 0.50000 0 0 0 0
11.00 21.00 31.00 0.00 0.00S S 0.93358 0.23348 0.32454 0.41603 0.50000 0 0 0 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
! Sc1 Sc2 Sc3 Sc4 Sc5 Sc61.531 0.000 0.000 0.000 0.000 0.000
0.00 0.00 0.00 0.00 0.00 0.00! a b c alpha beta gamma
8.485130 5.402066 6.964059 90.000000 90.000000 90.0000000.00000 0.00000 0.00000 0.00000 0.00000 0.00000
! Limits for selected parameters (+ steps & BoundCond for SA):1 0.0000 1.0000 0.0500 1 x_Pb2 0.0000 1.0000 0.0500 1 y_Pb_SO13 0.0000 1.0000 0.0500 1 z_Pb4 0.0000 1.0000 0.0500 1 x_S5 0.0000 1.0000 0.0500 1 z_S6 0.0000 1.0000 0.0500 1 x_O17 0.0000 1.0000 0.0500 1 z_O18 0.0000 1.0000 0.0500 1 x_O29 0.0000 1.0000 0.0500 1 z_O210 0.0000 1.0000 0.0500 1 x_O311 0.0000 1.0000 0.0500 1 y_O312 0.0000 1.0000 0.0500 1 z_O3
! T_ini Anneal Accept NumTemps NumThCyc InitConf8.000 0.900 0.008 60 0 0
! NCyclM Nsolu Num_Ref Nscalef Algor150 1 71 1 2
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Extracting Magnetic Intensities
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
0.00
0.20
0.40
0.60
0.80
1.00
0 5 10 15 20 25
LiMn2O
4
Ph_Mn2a1Ph_Mn2a2Ph_Mn2a3Ph_Mn2a4Ph
ases
(mod
2π)
1/T
Behavior of parameters in Simulated Annealing runs
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Average step ...
0.00
0.20
0.40
0.60
0.80
1.00
1.20
0 10 20 30 40 50 60
LiMn2O
4
<Step>
<Ste
p>
t
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Crystal structure of DFe2(P2O7)2 (D:Sr,Pb)
Pb
O22
O12
O22
O14
O23
Pb
O13
O14
O22
O12
O21O23
O13
O13
O23
O14
O12
O22
O12
O21O23
O13
O14
O22
O12
O13
a
b
c
Pb +2
Fe+3
P+5
O-2
_Triclinic P 1
a=4.7982(2) Å
b=7.1125(2) Å
c=7.8360(3) Å
α=89.816(2) Å
β=87.590(2) Å
γ=73.132(2) Å
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Magnetic susceptibility of PbFe2(P2O7)2
0
10
20
30
40
50
60
-100 -50 0 50 100 150 200 250 300
1/χ
(g/e
mu)
Temperature
PbFe2(P
2O
7)2
0
0.01
0.02
0.03
0 10 20 30 40 50
χ
Temperature
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Observed versus calculated neutron powder diffraction patterns of SrFe2(P2O7)2 on 3T2
Room temperature, 3T2λ= 1.225 Å
1228 reflections, 55 free parameters
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Observed versus calculated neutron powder diffraction patterns of PbFe2(P2O7)2 on G42
Room temperature, G4.2λ= 2.343 Å
285 reflections, 46 free parameters
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Observed neutron powder diffraction patterns of SrFe2(P2O7)2 on G41 as a function of temperature
(001) (010) (011) & (0 1 -1)
(100)T= 1.4 KT= 6.0 KT=11.0 KT=16.0 KT=19.0 KT=22.0 KT=25.0 KT=28.0 KT=34.0 KT=36.0 KT=45.0 K
Strongest magnetic reflections
FePO4impurity
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Observed neutron powder diffraction patterns of PbFe2(P2O7)2 on G41 as a function of temperature
(001)(010) (011) & (0 1 -1)
(100)T= 1.4 KT= 5.0 KT=10.0 KT=14.5 KT=17.0 KT=20.0 KT=22.0 KT=24.5 KT=25.5 KT=26.5 KT=27.5 KT=39.0 K
Strongest magnetic reflections
FePO4impurity
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Observed versus calculated neutron powder diffraction patterns of SrFe2(P2O7)2 on G4.1
T=1.4 K, G4.1λ= 2.43 Å
Nuclear reflections of FePO4Magnetic reflections of FePO4
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Observed versus calculated neutron powder diffraction patterns of PbFe2(P2O7)2 on G4.1
T=1.5 K, G4.1λ= 2.43 Å
Nuclear reflections of FePO4Magnetic reflections of FePO4
Mérida November 2003 Taller intensivo: El método de Rietveld, FullProf
Magnetic structure of SrFe2(P2O7)2 and PbFe2(P2O7)2
SrFe2(P2O7)2 at 1.4K PbFe2(P2O7)2 at 1.5K
Atom Atomic position Magnetic moment components in triclinic cell (µB)
x y z Mx My Mz Mx My Mz
Fe(1) 1 e ½ ½ 0 –4.75(4) 1.5(1) 0 –4.49(5) 1.5(1) 0
Fe(2) 1 b 0 0 ½ 4.75(4) –1.5(1) 0 4.49(5) –1.5(1) 0