Small-Angle X-Ray (SAXS) and Neutron

Scattering to Quantify Nanostructure

Gregory Beaucage
Professor of Chemical and Materials Engineering
University of Cincinnati
Cincinnati OH 45221-0012

20,085 undergraduate students

5,054 full-time graduate
6,739 part-time undergraduate students
3,366 part-time graduate
83.9 percent residents of Ohio

$332.0 million External Grants (2005)

3,000 full time faculty

& 6 SAXS Cameras!!! 6 hrs APS, IPNS, SNS

Roe, Schaefer, Beaucage, Jim Mark etc.
Outline:

General Background, Instruments,

Facilities
1) Specific Scattering Laws.
2) General Scattering Laws.
Guinier’s Law; Porod’s Law;
Unified Scattering
Fractal Scattering
Quantification of Branching in
Aggregates
3) Polydispersity
4) ASAXS for Catalysts
5) Summary
Kammler HK, Beaucage G, Kohls DJ, Agashe N. Ilavsky J
J Appl. Phys. 97(2005) (Article 054309).
Nanostructure from Small Angle X-ray Scattering

θ
 Small- and Wide-Angle X-ray
Scattering Measurements

X-ray
X-ray Source
Source SAXS : pinhole camera : 2-d detector at 1m from the sample
WAXS : pinhole geometry camera : image plate detector at
5cm from the sample
Sample Detector
Chamber
(ND)
(TD)

(MD)

 2D measurements are useful in

determining both size and relative
orientation of various structural
components
Small-Angle X-ray Scattering, (SAXS)
We Get Intensity as
A Function of Angle
(or radial position)

-Collimated Beam -Longer Distance for Lower Angle

-Monochromatic Beam -Large Dynamic Range Detector
-Coherent Beam -Evacuated Flight Path
(-Focusing Optics Perhaps) -Extend Angle Range with Multiple SDD’s

Crystalline Reflections Can Also Be Used

Small-Angle X-ray Scattering at the APS Chicago
We Get Intensity as
A Function of Angle
(or radial position)

Pinhole Cameras at: 12 ID BESSRC

5 ID DND } Variants on Build/Tear
Down Motif
18 ID BIOCAT (Inside Traders)
15 ID CARS
8 ID XOR
9 ID CMC-CAT

Semi-Permanent
33 ID UNICAT (Easily Used)
Small-Angle X-ray Scattering at Other Synchrotrons

ESRF we use ID2 with T. Naryanan

(Probably the best instrument)

Much easier to get time on smaller synchrotrons

We use SSRL (Stanford); CHESS (Cornell)
Nanostructure from Small Angle X-ray Scattering

Time Resolution 10 ms (Synchrotron Facility)

For Flow Through Experiment
(Flame/Liquid/Gas Flow) can be 10 µs

Size Resolution 1 Å to 1 µm (Synchrotron Facility)

Nanostructure from Small Angle X-ray Scattering

3-Techinques are similar

SALS/LS, SANS, SAXS

λ = 0.5 µm λ = 0.1 - 0.5 nm

For light For x-ray/neutron
Contrast, index of refraction, electron density,
neutron cross section
3-Closely related Techniques:
USAXS- Ultra Small Angle Scattering SAXS at
1/1000 º. APS in US and ESRF in Europe.
ASAXS- Anomalous x-ray scattering, vary
wavelength leads to change in contrast due to the
complex absorption spectra.
GISAXS- Promise of high resolution spectra
for surface structures but there are technical issues
with data interpretation.

http://staff.chess.cornell.edu/~smilgies/gisaxs/GISAXS.php
For Small Angle X-ray Scattering (SAXS)
1) Specific Structure
Calculate Pairwise Correlation Function
Calculate Fourier Transform to predict
scattering or direct transform of
measured data to correlation function
and analysis of correlation function.
Svergun (Hamburg) has applied this to protein in
native state
Simple application to monodisperse structures
Sphere Function/Rod Function/Cylinder

G = Nne2
 The Debye Scattering Function for a Polymer Coil

2
I(Q) = 2 (Q "1+ exp("Q))
Q
Q = q 2 Rg2

!
Binary Interference Yields Scattering Pattern.

I(q) ~ N ne2

ne Reflects the density of a

Point generating waves
N is total number of points
 The Scattering Event

I(θ) is related to amount Nn2

θ is related to size/distances

4!
q=
#
sin "
2
( )
2!
d=
q
2) Rather than consider specific structures, we can consider
general scattering laws by which all scatters are governed
under the premises that 1) “Particles” have a size and
2) “Particles” have a surface.
Binary Interference Yields Scattering Pattern.

-Consider that an in-phase

wave scattered at angle θ was
in phase with the incident
wave at the source of
scattering.
-This can occur for points
separated by r such that
|r| = 2π/|q|
- q = 4 " sin $
# 2
Binary Interference Yields Scattering Pattern.

-For high θ, r is small

Binary Interference Yields Scattering Pattern.

-For small θ, r is large

For an isotropic sample we consider scattering as
arising from the probability of the random placement
of a vector r in the scattering phase.
For an isotropic sample we consider scattering as
arising from the probability of the random placement
of a vector r in the scattering phase.
 For an isotropic sample we consider scattering as
arising from the probability of the random placement
of a vector r in the scattering phase.

Rather than random placement of the vector we can hold

The vector fixed and rotate the particle
 For an isotropic sample we consider scattering as
arising from the probability of the random placement
of a vector r in the scattering phase.

Rather than random placement of the vector we can hold

The vector fixed and rotate the particle
 For an isotropic sample we consider scattering as
arising from the probability of the random placement
of a vector r in the scattering phase.

Rather than random placement of the vector we can hold

The vector fixed and rotate the particle
For an isotropic sample we consider scattering as
arising from the probability of the random placement
of a vector r in the scattering phase.

Rather than random placement of the vector we can hold

The vector fixed and rotate the particle
The particle becomes a probability density function
from the center of mass.

That follows a Gaussian Distribution.

# "3r 2 &
p( r) = exp%% 2 (
(
$ 4Rg '
 The particle becomes a probability density function
from the center of mass.

Whose Fourier Transform is Guinier’s Law.

# "3r 2 & # q 2 Rg2 &
p( r) = exp%% 2 (
( ) I(q) = G exp% " (
$ 4Rg ' $ 3 '
G = Nn 2e
 Guinier’s Law Pertains to a Particle with no Surface.
# "3r 2 & # q 2 Rg2 &
p( r) = exp%% 2 (
( ) I(q) = G exp% " (
$ 4Rg ' $ 3 '
G = Nn 2e

Any “Particle” can be Approximated as a Gaussian

probability distribution in this context.
!
 # "3r 2 & # q 2 Rg2 &
p( r) = exp%% 2 (
( ) I(q) = G exp% " (
$ 4Rg ' $ 3 '
G = Nn 2e

Guinier’s Law can be thought of as the

First Premise of Scattering:
! All “Particles” have a size reflected by the radius of gyration.
 The Debye Scattering Function for a Polymer Coil

2
I(Q) = 2 (Q "1+ exp("Q))
Q
Q = q 2 Rg2

For qRg << 1

2 3 4
! Q Q Q
exp("Q) = 1" Q + " + " ...
2! 3! 4!

Q $ q 2 Rg2 '
I (q) = 1" + ... # exp& " )
3 % 3 (

Guinier’s Law!
 I(q) ~ N ne2
ne Reflects the density of a
Point generating waves
N is total number of points

At the other extreme we consider a surface.

 I(q) ~ N ne2
ne Reflects the density of a
Point generating waves
N is total number of points

At the other extreme we consider a surface.

The only location for contrast between phases is
At the interface r ne =
4" 3
r
3

!
 I(q) ~ N ne2
ne Reflects the density of a
Point generating waves
N is total number of points

At the other extreme we consider a surface.

We can fill the interface with spheres of size r

r N = S/(πr2) ne =
4" 3
r
3

!
 Porod’s Law can be thought of as the
Second Premise of Scattering:
All “Particles” have a surface reflected by S/V.
(dp = (S/V)-1)
#S&
2
# 3 &2 4 2"n % (
e
# S & 4 "r 16S"r $V '
I (q) ~ Nn e2 ~ % 2 (% ( ~ )
$ "r '$ 3 ' 9 Vq 4

N = S/(πr2) r ne =
4" 3
r
! 3

!
For a Rough Surface: 2 ≤ ds < 3
(This Function decays to Porod’s Law at small sizes)

2
2 " S %" 4 (r % 3
6)d s S
I (q) ~ Nn ~ $ d s '$
e ' ~ Sr *
# r &# 3 & q 6)d s

S 4" 3
N ~ ds ne = r
r r 3

! !
Sphere Function

For qR >> 1

<sinqR> => 0
<cos2qR> => 1/3

G
I (q) " 4 4 Porod’s Law for a Sphere!
q R
 First and Second Premise of Scattering

# "3r 2 & # q2R 2 &

p( r) = exp%% 2 (( ) I(q) = Gexp%" g (
$ 4Rg ' $ 3 '
G = Nn 2e

!
Particle Size, dp

Structure of flame made silica nanoparticles by ultra-snall-

angle x-ray scattering. Kammmler HK, Beaucage G,
Mueller R, Pratsinis SE Langmuir 20 1915-1921 (2004).
Many Things can Happen between the “Particle” Size
And the “Particle” Surface. Consider a “Linear” Aggregate.
Many Things can Happen between the “Particle” Size
And the “Particle” Surface. Consider a “Linear” Aggregate.

Overall Rg
Many Things can Happen between the “Particle” Size
And the “Particle” Surface. Consider a “Linear” Aggregate.

S/V

Overall Surface Area (Sum of Primaries)

Many Things can Happen between the “Particle” Size
And the “Particle” Surface. Consider a “Linear” Aggregate.

At intermediate sizes the chain is “self-similar”

df
Mass ~ Size
df
" R2 %
z ~$ '
# R1 &
 Many Things can Happen between the “Particle” Size
And the “Particle” Surface. Consider a “Linear” Aggregate.

At intermediate sizes the chain is “self-similar”

I(q) ~ N ne2 I(q) ~ N ne2

df
N = Number of " R2 %
N ~$ '
Intermediate # rint &
Spheres in the " rint %
df

Aggregate ne ~ $ '
# R1 &
ne = Mass of inter. !
df df
sphere 2
" rint % " R2 % " R2 % d f )d f
Nn e ~ $ ' $ ' ( I(q) ~ $ 2 ' q
# R1 & # R1 & ! # R1 &
 Linear Aggregates

Beaucage G, Small-angle Scattering from Polymeric

Mass Fractals of Arbitrary Mass-Fractal Dimension,
J. Appl. Cryst. 29 134-146 (1996).
 Branched Structures

PDMS/Silica Hybrid Material

Guo L, Hyeon-Lee J, Beaucage G J. Non-Cry. Solids 243 61-69 (1999)
 Branching in different systems

• Long Chain and Short Chain

• Model Branched Polymers (Stars, Hyperbranched, Dendrimers)

• Branching governed by kinetics (nano-scale aggregates)

 1) Mass Fractal dimension, df.

df
& 2R # z is mass/DOA
z = '$ ! dp is bead size
$d ! R is coil size
% p"

Random aggregation (right) df ~ 1.8;

Randomly Branched Gaussian df ~ 2.5;
Self-Avoiding Walk df = 5/3 Nano-titania from Spray Flame

2R/dp = 10, α ~ 1, z ~ 220

Problem:
df = ln(220)/ln(10) = 2.3
Disk df = 2
Gaussian Walk df=2

A Measure of Branching is not Given.

 2) Fractal dimensions (df, dmin, c)
and degree of aggregation (z)

df
-F F & R#
dp z~$ !
$d !
% p"
R
d min
& R#
-F F p = z min ~$ !
$d !
% p"

c d f d min
z~ p = p
dmin should effect perturbations & dynamics, transport
electrical conductivity & a variety of important features.
Beaucage G, Determination of branch fraction and minimum dimension of frac. agg. Phys. Rev. E 70 031401 (2004).
Kulkarni, AS, Beaucage G, Quant. of Branching in Disor. Mats. J. Polym. Sci. Polym. Phys. 44 1395-1405 (2006).
 Fractal aggregates are springs

Ogawa K, Vogt T., Ullmann M, Johnson S, Friedlander SK, Elastic properties of nanoparticulate chain
aggregates of TiO2, Al2O3 and Fe2O3 generated by laser ablation, J. Appl. Phys. 87, 63-73 (2000).
A Scaling Model for Branched Structures
Including Polyolefins

42
!Br = = 0.66
64

z! p
"Br = = 1 ! z1 c !1
z
Mole Fraction of Branches

Beaucage G, Determination of branch fraction and minimum

dimension of fractal aggregates Phys. Rev. E 70 031401 (2004).
 3) Neutron & X-ray Scattering

We can “Build” a Scattering Pattern from Structural

Components using Some Simple Scattering Laws

I(θ) is related to amount Nn2

θ is related to size/distances
θ
4!
q=
#
sin "
2
( )
2!
d=
q
Guinier’s Law

(!q Rg )2 -2
I (q) = G e 3 Power Law
!d f
I (q) = B f q

Thin Disk Gaussian Chain

Persistence is distinct from chain scaling
Branching has a quantifiable signature.
 R2
G2

Bf
df

G1
R1
 df
Bf R g ,2
R2 d min =
G2 !(d f 2 )

G2

Bf
df

G1

R1
Branching dimensions are obtained by combining local scattering laws
 Small Angle Scattering

-df -df

-4 -1

Aggregate Primary Chain Persistence

Silica Aggregates Deuterated-PHB

Guinier’s Law Power Law Scattering Vector, q

& (' q Rg )2 # !d f
4)
I (q ) = G exp$ ! I (q) = B f q q=
&' #
sin $ !
$ 3 !
% " ( %2"
 Branch Content from Scattering
df
c & R2 # p dmin
p ~ z ~ $$ !! R2

% R1 "
pp=<<z z
R1
cc==1d f
ddmin = df
min = 1
c dmin

G2 d min ! d f
z= d f = c ! d min z! p ' R2 $
G1 (br = = 1 ! %% ""
z & R1 #
G Beaucage, Physical Review E (2004)
A S Kulkarni and G Beaucage, Journal of Polymer Science, Part B: Polymer Physics (2006)
 df = -2
c x dmin

dmin 2 1 2
c 1 2 2

df
Bf R g ,2
d min =
G2 !(d f 2 )

Topological information can be extracted from this feature

arising from combining Local Scattering Laws

G Beaucage, Physical Review E, 70, 031401 (2004)

AS Kulkarni, G Beaucage, J Polym Sci, Part B: Polym Phy, 44, 1395 (2006)
 4) Branched Polymers at Thermal
Equilibrium: Model Systems for LCB

“2D Slice”
PDI ~ 1.05
F5

F2

For Polymers dmin is the

Thermodynamically Relevant
Dimension (5/3 = 1.67 or 2)

df = dmin c
~ thermo x branching
Mole fraction of Branches
 Multi-Arm Star Polymers

Horton et al., Macromolecules, 22, 681 (1989)

US Jeng, TL Lin, LY et al. App Phys A, 74, S487 (2002).

# Sample Rg (Å) df dmin c φbr Theoretical φbr

1 3 Arm Star 6.87 2.00 1.33 1.49 0.39 0.33

2 6 Arm Star 78.27 2.00 1.25 1.59 0.71 0.67
3 18 Arm Star 4.37 2.27 1.61 1.40 0.81 0.88
 Hyperbranched Polymers
SANS on Hyperbranched Polymers: Beaucage model correctly describes transition
from good-solvent to Θ - solvent collapsed state for the minimum path dimension.

a) b)
E. De Luca, R. W. Richards, I. Grillo, and S. M. King, J. Polym. Sci.: Part B: Polym. Phys. 41, 1352 (2003).
Geladé, E. T. F.; Goderis, B.; et al. Macromolecules, 34, (2001).
AS Kulkarni, G Beaucage, Macromolecular Rapid Communications, accepted, (2007).
 Cyclic Polymers
Casassa Equation

qRg
2 2
2 '(q R $ 2
P(q) =
q Rg
exp%
% 4
g
"
" ! exp x 2
( )
dx
& # 0

E F Casassa, Journal of Polymer Science: Part A, 3, 605, (1965)

 V Arrighi, S Gagliardi et al., Macromolecules, 37, (2004)
S Gagliardi, V Arrighi et al., Applied Physics A, (2002)
S Gagliardi, V Arrighi et al., Journal of Chemical, 122 (2005)
 This Approach Can Quantify LCB in Polyolefins.
(Talk by Kulkarni/Beaucage Tomorrow T34 0668 3:30)
Mole Fraction NMR Model Based on
Mn Index of Branches Mole Fraction of
nbr/104
(g/mol) LCBI Branches
φbr C

PE 0 46,500 0.04 0 0

PE 2 11,500 0.91 0.39±0.005 0.36

PE 3 37,900 2.27 0.63±0.004 0.91

4) Model Polydisperse Simulations

Mw/Mn ~ 4
Large aggregates
Growth kinetics show
dmin => 1
df => 1.8 for RLCA
Predicted previously
by Meakin
 Branched Aggregates

APS UNICAT
Silica Premixed Flames
J. Appl. Phys 97 054309
Feb 2005

5mm LAT 16mm HAB

Typical Branched Aggregate
dp = 5.7 nm
z = 350
c = 1.5, dmin = 1.4, df = 2.1
φbr = 0.8

Beaucage G, Determination of branch fraction and minimum

dimension of fractal aggregates Phys. Rev. E 70 031401 (2004).
 - Behavior is Similar to Simulation
df drops due to branching

- Aggregate Collapse

Height above burner - Entrainment High in the Flame

Mechanics Depend on Structure TEOS:H2O:HCl
Structure on Growth Chemistry
Vapor
R
Reacting
Aerosol

a L HMDS

Mean Primary Particle Size, "a"

Mean Aggregate Size, "R" Particle/Cluster, Denser
Specific Surface Area related to 1/ a
"R" is related to " a ", "N", and Structure Cluster/Cluster, Looser
"Structure" is Related to Growth Mechanism
"a" is Related to Early Stage If Particles
"N" and "R" to Later Stage Don't Stick Well
Packing is Dense
"Reaction Limited"
N = α (R/a)d f How Dense?
1≤ df ≤ 3 If Particles
( /a) = (R/a)dmin Minimum Path Dim. Stick Immediately
1≤ dmin ≤ df Packing is Loose
"Diffusion Limited"
C = d f/dmin How Branched?
Linear C = 1; Reg. C = df
 Nano-Aggregates Can Act as Springs

a
σ E dmin is Dimension
3+dmin
EAggregate = E Oxide(a/R) R of Stressed
Path
ε
From: T. A. Witten, M. Rubinstein, R. H. Colby
J. Phys. II France 3, 367 (1993).

Static Stressed Released

100nm F 100nm 100nm

From: S. K. Friedlander, H. D. Jang, K. H. Ryu

Appl. Phys. Lett. 72 173 (1998).
 Summary of Witten/Rubinstein/Colby Theory
for Mechanics of Springy Aggregates in Elastomers

a
dmin is Dimension
R of Stressed EAggregate = E Oxide(a/R)3+dmin
Path

Aggregates are only Effective

below a Critical Size, Rcritical
Rcritical = a(EOxide/ERubber)1/(3+dmin )
A Critical Concentration is Predicted Beyond which
There is No Higher Reinforcing Effect, φcritical
φcritical = (R/a)df-3

The Modulus for a Critical Concentration Composite, Ecomposite

is given by:

Ecomposite = EOxide φcritical(3+dmin)/(3-df)

Test these Propositions using Tuned Nano-Composites

 Complex Scattering Pattern (Unified Calculation)

2! N = Number Density at Size “d”

q= I (q ) = N (d )ne2 (d ) ne = Number of Electrons in “d” Particles
d
Guinier’s Law

Particle with No Interface

I (q ) = N (d )ne2 (d )
& ' q 2 Rg2,1 #
I (q ) = G1 exp$ !
$ 3 !
% "
G = N! e2 V 2 ~ R 6

R8
Rg2 ~
R6
Guinier and Porod Scattering Spherical Particle
With Interface (Porod)

I (q ) = BP q !4
BP = 2"N! e2 S

S ~ R2

Q = " q 2 I (q )dq = N! e2 R 3

Q R3
dp = = 2
2!BP R
Structure of Flame Made Silica Nanoparticles
By Ultra-Small-Angle X-ray Scattering
Kammler/Beaucage Langmuir 2004 20 1915-1921
 Polydispersity Index, PDI
Polydisperse Particles

BP Rg4
PDI =
1.62G
12
& ln (PDI )#
' = ln (' g )= $
% 12 !"

2 12
& 5R # g
m = $ 14' 2 !
$% 3e !"
Particle size distributions from small-angle scattering
using global scattering functions, Beaucage, Kammler,
Pratsinis J. Appl. Cryst. 37 523-535 (2004).
 Linear Aggregates

& ' q 2 Rg2, 2 #

I (q ) = G2 exp$ !
$ 3 !
% "
df
G2 & R2 #
z= = $$ !!
G1 % R1 "
!d f
I (q) = B f q

G2 d f
Bf = df
!(d f 2 )
R g ,2

Beaucage G, Small-angle Scattering from Polymeric

Mass Fractals of Arbitrary Mass-Fractal Dimension,
J. Appl. Cryst. 29 134-146 (1996).
 1) Fractal dimensions (df, dmin, c)
and degree of aggregation (z)

df
-F F & R#
dp z~$ !
$d !
% p"
R
d min
& R#
-F F p = z min ~$ !
$d !
% p"

c d f d min
z~ p = p

dmin should effect perturbations & dynamics.

Beaucage G, Determination of branch fraction and minimum
dimension of fractal aggregates Phys. Rev. E 70 031401 (2004).
 Linear Aggregates

& ' q 2 Rg2, 2 #

I (q ) = G2 exp$ !
$ 3 !
% "
df
G2 & R2 #
z= = $$ !!
G1 % R1 "
!d f
I (q) = B f q

G2 d f
Bf = df
!(d f 2 )
R g ,2

Beaucage G, Small-angle Scattering from Polymeric

Mass Fractals of Arbitrary Mass-Fractal Dimension,
J. Appl. Cryst. 29 134-146 (1996).
 Branched Aggregates

d min
& R2 # 1
p = $$ !! =z c

% R1 "
df
c=
d min
d min ! d f
' R2 $
(Br = 1 ! %% ""
& R1 #

G2 d min
Bf = df
!(d f 2 )
Beaucage G, Determination of branch fraction and
minimum dimension of fractal aggregates Phys. Rev.
Rg , 2
E 70 031401 (2004).
Large Scale (low-q) Agglomerates

I (q ) = BP q !4
Small-scale Crystallographic Structure
 Branched Aggregates

APS UNICAT
Silica Premixed Flames
J. Appl. Phys 97 054309
Feb 2005

5mm LAT 16mm HAB

Typical Branched Aggregate
dp = 5.7 nm
z = 350
c = 1.5, dmin = 1.4, df = 2.1
φbr = 0.8

Beaucage G, Determination of branch fraction and minimum

dimension of fractal aggregates Phys. Rev. E 70 031401 (2004).
Supported Catalysts
 Nobel Metals (Gold)
Become
Reactive Catalysts
When of 1-6 nm Size

Onset of catalytic activity of gold

clusters on titania with the appearance
of non-metallic properties.
Valden M, Lai X, Goodman DW
Science 281, 1647-1650 (1998).
 Solution Method
For Au/Support Oxide
(Using HAuCl4)

20 nm

Gold Catalysts Prepared by coprecipitation

for low-temperature oxidation of hydrogen
and of Carbon Monoxide.
Haruta M, Yamada N, Kobayashi T, Iijima S,
J. Of Catalysis 115, 301-309 (1989).

480 Citations

Size- and support-dependency in the

catalysis of gold. Haruta M, Catalysis
Today 36, 153-166 (1997).
592 Citations
Consider Support Particle with Deposited Domains
We can obtain: Mean Size,
Polydispersity,
State of Aggregation
For Both Particle Types.
This can be done in situ in almost
any environment that can be brought to the synchrotron.

Measurements with ETHZ (Eveline Bus, Jereon Van Bokhoven)

ESRF (T. Narayanan)
Option 1: Brute Force/Lab Source

dp, nm σg

3.97 1.35

14.9 1.08
Consider Support Particle with Deposited Domains
Option 1: Brute Force
How do the particles vary with concentration gold?
Measurements with ETHZ (Eveline Bus, Jereon Van Bokhoven)
ESRF (T. Narayanan)
Consider Support Particle with Deposited Domains
Option 1: Brute Force
In situ versus ex situ measurements.
Measurements with ETHZ (Eveline Bus, Jereon Van Bokhoven)
ESRF (T. Narayanan)

Desirable: In Situ Study/Contrast variation for Gold

 Consider Support Particle with Deposited Domains
Option 2: Anomalous Scattering/Synchrotron
In situ anomalous small-angle x-ray scattering from metal particles in supported-
metal catalysts. I Theory and II Results. Brumberger H, Hagrman D, Goodisman
J, Finkelstein KD, J. Appl. Cryst. 38 147-151 and 324-332 (2005).

Haubold et al. 1999

(Ge)

Goerigk et al. 2003

G.Goerigk and D.L.Williamson

http://www.desy.de/~jusifa/solarzellentechnik.htm
Scattered Intensity Depends on Contrast, G (For Each Phase)

G = N! e2 V 2

& 1.62 PDI #

BP = G$ 4
!
$ R !
% g "
I (q ) = BP q !4
" e2 (E )=
& ' q 2 Rg2,1 #
I (q ) = G exp$
$
%
3
!
!
"
[n1 f1 (E )! n2 f 2 (E )]2

Particle size distributions from small-angle scattering

using global scattering functions, Beaucage, Kammler,
+ {
+ }
Pratsinis J. Appl. Cryst. 37 523-535 (2004).
“f” Depends on Wavelength
Sintering of Ni/Al2O3 catalysts studied by
anomalous small angle x-ray scattering.
Rasmussen RB, Sehested J, Teunissen HT,
Molenbroek AM, Clausen BS
Applied Catalysis A. 267, 165-173 (2004).
- =
Particle Size Distributions
From SAXS
Particle Size Distribution Curves From SAXS

Assumption Method
i) Assume a distribution function.
ii) Assume a scattering function (sphere)
iii) Minimize calculation
Particle Size Distribution Curves From SAXS
Assumption Method.
i) Assume a distribution function.
ii) Assume a scattering function (sphere)
iii) Minimize calculation

Sintering of Ni/Al2O3 catalysts studied by

anomalous small angle x-ray scattering.
Rasmussen RB, Sehested J, Teunissen HT,
Not unique & Molenbroek AM, Clausen BS
Applied Catalysis A. 267, 165-173 (2004).

Based on assumptions
But widely used & easy to understand
Particle Size Distribution Curves From SAXS
BP Rg4
PDI =
Unified Method 1.62G
i) Global fit for BP and G. & ln (PDI )#
12

ii) Calculate PDI (no assumptions & ' = ln (' g )= $ !

% 12 "
unique “solution”) 12
2
iii) Assume log-normal distribution & 5R # g
m = $ 14' 2 !
for σg and distribution curve $% 3e !"
(or other models)
iv) Data to unique solution
Solution to distribution

Advantages
Generic PDI (asymmetry also)
Global fit (mass fractal etc.)
Direct link (data => dispersion)
Use only available terms
Simple to implement
Particle Size Distribution Curves
from SAXS

PDI/Maximum Entropy/TEM Counting

Particle size distributions from

small-angle scattering using global
scattering functions, Beaucage,
Kammler, Pratsinis J. Appl. Cryst.
37 523-535 (2004).
Particle Size Distribution Curves From SAXS
Maximum Entropy Method
i) Assume sphere or other
scattering function
ii) Assume most random solution
iii) Use algorithm to
guess/compare/calculate
iv) Iterate till maximum “entropy”
Advantages
No assumption concerning
distribution function
No assumption for number of modes
Matches detail PSD’s well
Related Alternatives
Regularization

Particle size distributions from small-angle scattering

using global scattering functions, Beaucage, Kammler,
Pratsinis J. Appl. Cryst. 37 523-535 (2004).
Software for My Collaborators/Students
(And Me)
Particle Size Distribution Curves From SAXS
All Methods are available in Jan Ilavsky’s Igor Code
http://www.uni.aps.anl.gov/usaxs/
Anomalous Scattering
Particle Size Distribution Curves From SAXS
All Methods are available in Jan Ilavsky’s Igor Code
http://www.uni.aps.anl.gov/usaxs/

Unified Fit (Not all implemented)

Particle Size Distribution Curves From SAXS
All Methods are available in Jan Ilavsky’s Igor Code
http://www.uni.aps.anl.gov/usaxs/

Sphere (or any thing you could imagine) Distributions

Particle Size Distribution Curves From SAXS
All Methods are available in Jan Ilavsky’s Igor Code
http://www.uni.aps.anl.gov/usaxs/

Maximum Entropy/Regularization Code (Jemian)

Zeolites/Spherical Colloids
Many Other Parts to Scattering
Are Not Covered
For Instance:
Zeolites:
-Pore Structure
-Nano-Structure
-Micron Structure
-4
-4

Ethyl acrylate Polyethylacrylate

Benzyl Peroxide in 2 nm

Zeolite 13X Zeolite Pores

1 nm

Pu Z, Mark JE, Beaucage G, Maaref S, Frisch HL, SAXS Investigation of

PEA Composites J. Polym. Sci., Polym. Phys. 34 2657 (1996).
Keep in Mind:

-SAXS Measurement is Generally Easy

-SAXS Analysis is Generally Difficult

-A Reasonable Model is Mostly Needed

-You will Generally Have to Understand

What is going on.

(-This is not a good Technique for Those

Interested only in Verifying)