PARTICLE CHARACTERISATION

1
PARTICLE CHARACTERISATION

• Heinrich Hofmann, Paul Bowen
• Powder Technology Laboratory (LTP), Materials
  Department
• Swiss Federal Institute of Technology Lausanne
  (EPFL), Switzerland.

2
Introduction

• Powder characteristics important application
  size always important
• Below 100 nm not always easy
• Methods
• Light scattering (PCS), centrifugation (X-ray,
  light), image analysis (TEM), (laser diffraction
  (agglomerates))
• Introduction sizes, distributions, methods
• Examples
• Milled of gamma alumina 1000 to 50 nm
• Boehmite 20nm
• Spherical silica 20-80 nm
• Gold 15-50 nm
• Iron oxides 10nm
• Conclusions

3
Important Powder Characteristics

• Size and size distribution
• Morphology and shape factor - microscopy and
  image analysis
• Specific Surface Area - gas adsorption (N2 ,
  gt0.1 m2/g)
• Porosity (internal structure) -
  adsorption-desorption of gas, mercury
  porosimetry
• Crystalline Phase - X-ray diffraction
• Chemical Composition (purity, additives)
• Homogeneity
• Density (absolute, apparent) - poured, tapped,
  He pycnometry
• Internal friction - angle of repose, shear
  tester Jenike cell
• Flowability and compressability - Hall flow
  meter (Hausner ratio), pressing (414 MPa)

L. Svarovsky, Powder Testing Guide, Elsevier
Applied Science, London, 1987.
4
Introduction - Why Measure a PSD?

• Why measure particle size distributions -?
• does the powder flow or how quickly does it
  dissolve?
• Du Pont survey by Davies and Broughton on 3000
  products
• 80 involved a powder at some stage of the
  manufacture
• Powder Technology important in the modern
  industrial world.
• Heavily dependent on our capability to measure
  the powder's
• Particle size and its distribution. PSD
• PSD influences its properties, handling and
  domain of application.
• PSD measurement - not enough on its own the
  results must be coupled
• other techniques to correctly interpret and use
  the measured PSD
• microscopy,
• X-ray powder diffraction,
• surface area measurement
• chemical composition analysis

5
Introduction- More whys

• Often PSD measured with aim of relating it to a
  particular property or behaviour
• When choosing a method the application should
  always be born in mind.
• Different methods for PSD measurement often have
  limitations
• If ignored - correlations with the property of
  interest and
• conclusions drawn can be erroneous.
• Verification of batch to batch variation
• Sampling of large batches
• Storage - effects of time on powder properties -
  eg moisture/agglomerates
• The aim of this talk is to
• familiarise the participants with some of the
  methods currently available
• their limitations
• how good are these for absolute or comparative
  PSD measurement

6
Diameters and distributions
If the particle is not spherical life is
already difficult without the particle having to
be small
d
d
Fmax
d
A
d

equivalent circular
A
d

diameter

A projected area
d
7
Particle Diameters
8
Diameters and Distributions

• Frequency or cumulative - Less than - Greater
  than

• Number or volume - distribution base

9
Distributions and Average Diameters

• Averages - central tendency -
• Mean - Mode - Median - for a normal distribution
  all equivalent

• Standard deviation- dispersion or width
• dv the mean volume diameter
• fi is the frequency of particles (as a volume)
  of that diameter and V total volume for all of
  the diameter intervals
• Span d90-d10 / d50

10
Diamètres moyennes de particules.
 
 
Nombre - longueur
Nombre - surface
Longueur - surface
Nombre - volume
Volume - moment (Poids - moment)  
Surface - volume
Écart-type pour chaque diamètre
Surface spécifique

• Mean diameter means NOTHING
• Essential to quote base volume, number,. and
  instrument used

11
Specific Methods - Sedimentation -
Photocentrifuge (1)

• Horiba CAPA-700 - principle
• Sedimentation in liquid phase
• Stokes law to calculate diameter
• Concentration from light absorption
• Homogeneous Suspension
• Gravitation or centrifugation
• Variation in particle concentration followed by
  change in intensity
• For small particles - light scattering
  correction needed
• Light extinction efficiency K(Di) is a function
  of
• Particle size
• R.I. of suspending medium and particle

12
Photon Correlation Spectroscopy - PCS

• Dynamic light scattering (DLS) method or
• Photon correlation spectroscopy (PCS) or
• Quasielastic light scattering (QELS) (as it was
  first termed)

• Dependence of the scattered intensity
• - Proportional R6
• - Particle only twice the size
• will give 64 times the intensity.

13
Brookhaven X-ray Disc Centrifuge

• Stokes law to calculate diameter X-ray
  absorption to detect particle concentration

• Advantages
• X -ray absorption proportional to mass - volume
• Good correspondance with standard powders
• Reasonably quick - 10mins (for 0.3 -3.0 µm)
• Large size range 0.01 -300µm
• Limitations
• Need 2-3 g of powder
• Suspension concentrated - particle-particle
  interactions
• Often difficult to match liquid viscosity-
  particle density for single run measurement

T. Allen, in  Particle Size Analysis , Eds.
Stanley-Wood, N.G. and Lines, R.W., p.498, Royal
Society of Chem, Cambridge, 1992
14
Porosity - Agglomeration Factor, Fag - Number, FN

• Fine powders have the tendency to form
  agglomerates (during forming) or aggregates
  (synthesis)
• Define an agglomeration factor Fag or
  agglomeration number, FN ,

Primary particle
Particle measured

• Fag, very good indication of the degree of
  agglomeration allows comparison between powders
  and treatments

- agglomerate or aggregate
dv50 - median diameter (volume, µm), dBET is an
average diameter (mm) calculated from specific
surface area, SSA (m2/g) measured by nitrogen
adsorption (model BET), r powder density
(g/cm3), VBET volume of sphere from dBET , Vs
volume of powder in agglomerate of given size,
exluding pore volume (estimated from nitrogen
desorption pore volume)
German International Journal of Powder
Metallurgy Vol. 32 4 365-373 (1996)
15
Gamma Aluminas Sintering Studies

• Two Gamma aluminas studied
• CR 125 Baikowski France (gt 99.99 Al2O3) pure
  doping studies
• Degussa C , Germany (gt 99.6 Al2O3) - fine

16
Gamma Alumina - Powder Characteristics

• Degussa C , - (Germany gt 99.6 Al2O3),

25 nm
17
Horiba Photocentrifuge - Boehmite
Boehmite 1 - Condea - (AlOOH) 213 m2/g
dBET 9.4 nm

• Correction for light scattering essential
• Very good quality data needed
• Takes 2 days with Horiba CAPA700

18
Comparison - Boehmite - Horiba - XDC - PCS
Boehmite 1 - Condea - (AlOOH) 213 m2/g
dBET 9.4 nm

• Distribution narrow - (sv50/dv50 2.2)
• Median diameters very similar
• Resolution with XDC best
• XDC 2hrs
• PCS 10 -20mins
• (single angle, CONTIN)
• Horiba 2 days!

v50
19
Gamma Alumina
Gamma Alumina 1- Degussa C - 92 m2/g dBET
19.2nm

• Dv50 higher values
• Horiba finer overcorrects?
• PCS Intensity R6
• Dv50 bigger sensitive to size range and small
  agglomerate population?
• Hydrodynamic density estimated from porosity
  measured by nitrogen adsorption desorption (NAD)

20
Attrition Milling and PSD Gamma Alumina CR125
(XDC)

• Attrition milling significant
  de-agglomeration still agglomerates dV50 50
  nm
• Hydrodynamic density from NAD (nitrogen
  adsorption desorption)
• How reliable is dv99 ? (Agglomerates very
  important in ceramic processing)

21
Effect of Milling Parameters on PSD

• Looked at zirconia beads and the effect of their
  size on PSD
• 0.5mm and 1.25 mm
• 1 hr and 3hrs milling
• Polyacrylic acid (PAA) pH6
• Preliminary results with 1.25 mm beads

• Preliminary results showed needed 2 repetitions
  for 95 confidence level
• 8 experiments in total

22
Effect of Milling parameters on dv50 et dv99
Dv50 (nm)

• Dv50 decreases from 60 to 50 nm between 1 and 3
  hrs but
• No statistically significant effect when smaller
  beads used

• Dv99 significantly effected by all parameters
  and
• Best result when using 0.5mm beads for 3hrs 99
  of particles below 140nm
• Expected trends but interesting to see for
  commercial nanosized powder

23
Spherical Silica - Real size ??

• AlOOH very good agreement between XDC, PCS,
  Horiba
• Use spherical particles to illustrate accuracy
• Counted 1000 particles image analysis program

• Klebosol Clariant
• Sold initially as 50 nm particles - dBET 50nm
• Porosity?

24
Spherical Silica - Real size ??

• Golden Rule No. 3 even if different methods give
  same result it is not necessarily "absolute " 
• Only parameter density

• Take into account porosity - NAD 1.75 g/cm3

• Sold as acid dispersion -measured in - 0.005M HCl
• double layer 5 nm
• 1.52 g/cm3

25
Spherical Silica - Real size ??

• Using density 2.2 g/cm3
• Added salt KNO3 0.1m
• double layer lt1 nm
• No significant difference
•  anomalous  stabilty of silica sols even at
  iep?
• Often attributed to hydration layer thickness?

26
Spherical Silica - Real size ??

• Interparticle energy calculations suggest
  thickness of around 0.5 nm enough
• Dense silica particles normally synthesised at
  basic pH
• End point for Klebosol acidic pH near 2 where
  porous polymeric gels form
• Perhaps a  fuzzy  hydrated layer porous gel
  giving  steric stability ?

• Have studyied smaller sizes ( 35/50 nm and
  12/18nm) and batch to batch variation
• Density/porosity reasonably consistent but have
  to characterise each batch
• Trying to modify  fuzzy layer  by ageing in
  acid or base no yet fully resolved

27
Silca Spheres 80, 50, 25 nm

• Compared - XDC (1.77 g/cm3), TEM, PCS, HRSEM
• Very coherent results over whole distribution

50 nm particles
28
Silca Spheres 80, 50, 25 nm

• dn mean no. diameter sensitive to width
• 80, 50, 25 nm silica spheres
• Results overall very good
• ? standard devaition of distribution
• All narrow and good agreement for our model
  spheres
• Except arbitrary hydrodynamic density
• For 25 nm PCS protocol took some time to develope
  before agreement with other methods

29
Gold particles size distribution
BBI GC30
BBI GC50
BBI GC15
gold (BBI,UK)
HRSEM Micrographs C. Soare
30
Iron Oxides approaching 10 nm

• Iron oxides synthesised by preciptation for
  biomedical applications superparamagnetic
• Properties strongly linked to size narrow
  distribution key
• TEM counted 100 particles

31

Particle characteristics
Nominal particle sizes
silica (Clariant, F) - 80, 50, 25 nm
(10-30w 5-16v) gold (BBI,UK) - 50, 30, 15
nm (4.5x10-4v) iron oxide - 10 nm
(0.25v) zinc sulfide - 5 nm

• Used 4 methods for silica just PCS EM for
  others
• XDC X-ray Disc centrifuge (Brookhaven)
• PCS Photon Correlation Spectroscopy
• TEM transmission electron microscopy
• HRSEM High Resolution Scanning electron
  microscopy

32
Iron Oxides 10 nm comparison

• XRD line broadening, TEM, PCS
• PCS after optimisation
• all liquids filtered at 20nm
• Data collection optimised (baselines)
• Data anlaysis CONTIN
• XRD, TEM volume and number 9-10 nm
• PCS 13-14nm

• Hydrodynamic surface layer seen by other
  researchers (Jolivet)
• Here 2nm  layer  thickness
• Parker (Bath, UK) molecular modelling in conc.
  salt (0.5 M)
• 1 nm structured water but 2.5 nm structured ions

33
Goethite Surface in Contact with Electrolyte
Solution
Professor Steve Parker S. Kerisit PhD
student Department of Chemistry University of
Bath Bath BA2 7AY United Kingdom Tel 044
(0)1225 386505 Fax [phone] http//www
.bath.ac.uk/chsscp
NaCl 0.5 M
Kerisit S, Cooke DJ, Marmier A, Parker SC ,
CHEMICAL COMMUNICATIONS (24) 3027-3029 (2005)
34
Goethite Surface in Contact with Electrolyte
Solution
NaCl 0.5 M
35
PCS - Iron Oxide- PVA adsorption

• For applications needed steric stabilisation
  PVA worked but how much and what thickness?
• Once  optimised  can use PCS eg PVA adsorbed
  layer thickness
• Suggests sautration at ratio 1.4 PVA/FeOx
• Supported with zeta potentail measurements
  saturates near 1.4 also

Mobility m2s-1V-1
36
Lower limits ? example of ZnS

• Synthesis of ZnS doped with Mn for bio-assay
  fluorescence applications

5 nm

• XDC no change in baseline!
• all particles lt 14 nm, rh 4.1,
• lt 24 nm, rh 2.0

• XRD line broadening 3.6 nm - Analytical
  Ultracentrifuge (AUC, rh 3.2 g/cm3 ),
• Image analysis (IA) 500 particles in TEM 3-17nm
  Absorption UV-visible 4 nm
• PCS 6-7 nm avec  optimisation de methode  -
  Diffcile pour une taille absolute mais coherent

37
Conclusions

• For narrow distributions methods coherent in
  15-100nm range
• More agglomerated and broader distributions
  powder - less agreement
• Horiba (photocentrifuge) undersizes
  sensitivity to correction?
• PCS oversizes if small population of
  agglomerates?
• X-ray disc centrifuge - XDC - best resolution and
• Probably best precision BUT
• Hydrodynamic density must be accurately known for
  accurate results
• For lt 10-15 nm but life gets very difficult
• TEM, PCS, AUC best approach
• but detection difficult
• particles or dust must work cleanly
• hydrodynamic density difficult to assess
• Essential to use more than one method
  compliment PSM with images, SSA, XRD