Sampling

Sampling and Dispersion

“Novices in the size-measurement field must
understand that most errors in size measurement
arise through poor sampling and dispersion and not
through instrument inadequacies.”
Terry Allen, Advances in Ceramics, Vol 21: Ceramic Powder Science, page 721, The American
Ceramic Society Inc. (1987)
 Relative Errors in Particle Size Distribution Analysis

Dispersion

Sampling and dispersion as with many

other techniques can be significant
Sources of error in laser diffraction
Sampling Analysis.
Error

Instrumentation

0.01 0.1 1 10 100 1000

Particle Size ( mm)

Dr. Henk Merkus (Delft University of Technology) “Quality Assurance in Particle Size Measurement” from Improving Standards
in Particle Size Distribution Measurement, February 17-19, 1997, at the Engineering Research Center for Particle Science and
Technology
 Sampling of Powders
• Results of any test only as good as the
sample
• Segregation or stratification by particle
size is the greatest problem
• Sampling very important in handling
 Some notes on sampling
• To collect a manageable mass (sample)
representative of the whole
• All particles in the total must have
same probability of being included in
sample
• All parts of the total have to be equally
accessible
 Golden rules of sampling
1. Sampling should be done preferably
from a moving stream (both powders
and suspensions)
2. A sample of the whole stream should
be taken for many periods rather than
a part for the whole time
Moving stream
Moving stream
Moving stream
 Sampling on static solids

Powder Sampler (“Thief

Sampler”)
Into powder bed
Sample splitting
Sampling - Use sample splitter, cone and quarter, rotary riffler or
Other technique to ensure sample is representative. Entire
Subsample should be used for each measurement.

Dispersion is aided by knowing your material/system

• Choosing the right solvent
• Agitation/sonication
• Dispersing aids

agglomerated well dispersed

 Sampling
Ref: T. Allen Particle Size Measurement Chapman and Hall 4th Edition 1993 Page 39

Reliability of selected sampling methods (60:40 coarse,fine sand mixture)

Method Estimated max Efficiency

error (%) (%)

Cone & Quartering 22.7 0.013

Scoop Sampling 17.1 0.022
Table Sampling 7.0 0.13
Chute Riffling 3.4 0.56
Spinning Riffling 0.42 36.3
Random Variation 0.25
Dispersion Issues

• Wet or Dry?
• Charge Stabilization and Zeta Potential
• Dispersion Aids
• Cohesion in dry powders
• Agitation, Sonication and Shear
• How do you know when you are well dispersed?
• Pitfalls
• Bubbles, Optical constants and Agglomeration
 Attrition

An example of attrition due to high energy sonication

How much is Enough?
 Particle Size Distribution

Mixed Result Type (%) 12 Number distribution

10
8
6 Volume distribution
4
2
0
0.01 0.1 1 10 100 1000 3000
Particle Size (µm)
PHBA Disp in IPA w/10min bath 1 min probe, 15 Oct 1999 10:39:29
PHBA Disp in IPA w/10min bath 1 min probe, 15 Oct 1999 10:40:20

A common error is reporting the inappropriate size distribution

Which would you use?
D10 D50 D90
Volume 0.194 0.551 1.336
Number 0.040 0.076 0.196
Particle Shape
 Particle shape
most real materials are NOT SPHERICAL

all automated non-imaging particle sizers use a

“SPHERE”
as the shape for which SIZE is calculated
from some measurement.

Hence: all shapes are described by a DIAMETER

Each technique defines its own

EFFECTIVE SPHERICAL DIAMETER (ESD)
Different techniques give different ESD for the same sample
 What is particle shape?
• Very difficult to get a handle on:
• Affect flowability, packing, interactions with
fluid, etc.
• Define type of shape – acicular, granular,
angular etc.
• Define shape factor (coefficient)
V = αvx3 A = αAx2
S = αSx2 S/V = αsv/x
Where α – function of particle shape
 What is particle shape?
One special shape factor “sphericity”

 = surface area of sphere with the same volume

surface area of particle
Commercially available techniques for particle sizing
BIC systems for particle sizing, size range: 2 nm to ~100 µm
 Size range

• “Zero-to-Infinity” System?
– Hybrid systems ? Combine more than one technique
• only one system is required
• combine more than one technique
• operator learning is reduced to ONE
• less bench space is required
• covers the range from 2 nm to 100 µm +
• lower investment
 The Hybrids

Examples
Frauenhofer / High Angle Scattering
Sedimentation / Centrifugation

Problems
Sampling
Changes in resolution
Changes in weighting
Boundary matching
 Limitations
• Diffraction

• Sedimentation

• Density of sample

• Splicing of different techniques

 Diffraction

• Limited to sizes much larger than

wavelength of light source

• Size classes logarithmically spaced

 Sedimentation

• at high end limited by turbulence

• at low end limited by diffusion
• acceleration broadens real size
distribution
• density of sample measured
Classification of Techniques

Size range

Particle Counters

Fractionators

Ensemble Averaging
Particle Counters

Microscopes

Image Analyzers

Zone Counters (EZ, PZ)

Particle Scanners
Measure one particle at a time
VERY HIGH RESOLUTION

 slow, time consuming

 operator “fatigue”

 sampling problem (coincidence)

 statistics
Classification of Techniques

Size range

Particle Counters

Fractionators

Ensemble Averaging
Fractionators

Sieves

Hydrodynamic Chromatography

Field Flow

Sedimentation

Centrifugation
 Specification
- Quantitative
s - Qualitative

 Size range *)  Sample preparation time

 Throughput  Analysis time

 Accuracy  Data reduction /

interpretation time
 Precision
 Clean-up time
 Reproducibility

 Resolution
Definitions
ACCURACY is a measure of how close a
measured value is to the true value

PRECISION is a measure of the variation

in the repeated measurements

RESOLUTIONis a measure of the minimum

detectable differences between features
in a size distribution

REPRODUCIBILITY
is a measure of the variation between
different machines or sample preparations
or operators, etc.
Sampling - Use sample splitter, cone and quarter, rotary riffler or
Other technique to ensure sample is representative. Entire
Subsample should be used for each measurement.

Dispersion is aided by knowing your material/system

• Choosing the right solvent
• Agitation/sonication
• Dispersing aids

agglomerated well dispersed

 Sampling
Ref: T. Allen Particle Size Measurement Chapman and Hall 4th Edition 1993 Page 39

Reliability of selected sampling methods (60:40 coarse,fine sand mixture)

Method Estimated max Efficiency

error (%) (%)

Cone & Quartering 22.7 0.013

Scoop Sampling 17.1 0.022
Table Sampling 7.0 0.13
Chute Riffling 3.4 0.56
Spinning Riffling 0.42 36.3
Random Variation 0.25
Measurement and Data Representation
When I make the Measurement – how do I communicate it
to other scientists?

We need common agreed upon methodologies!

#1 do I sample the material in the same way?

#2 do I make the measurement in the same way?
#3 do I represent data in the same way?
#4 do I interpret the data in the same way?
 Relative Errors in Particle Size Distribution Analysis

D
i
sp
er
si
on

Sampling and dispersion as with many

other techniques can be significant
Sources of error in laser diffraction
S
a
mp
li
ng Analysis.
Ero

I
ns
tr
ume
n
ta
ti
on

0
.
010
.
11 1
010
010
0
0
P
a
rt
i
cle
Si
zem
(m)

Dr. Henk Merkus (Delft University of Technology) “Quality Assurance in Particle Size Measurement” from Improving Standards
in Particle Size Distribution Measurement, February 17-19, 1997, at the Engineering Research Center for Particle Science and
Technology
 Common Mistakes in Particle Characterization

The most common mistake in particle characterization is

not doing your homework!

• Know your particles

• Know the system in which they are to be used
• Know the physical principles of your analysis technique
- and its limitations!
• Determine what information you require
• Think critically
 Other Common Mistakes

• Sampling –
• using a non-representative sample
• using too small a sample
• using a contaminated sample
• Not looking at your particles under the microscope
• Using the wrong technique
• Making measurements under conditions different
than those in system under study
• Lack of understanding of agglomeration and
cohesion
 Common mistakes - continued

• Not understanding the chemistry and physics of your material

• Not calibrating the instrumentation
• The instrument must be right! – Black Box Addiction.
• Poor sample preparation – impatience
• Not reading the directions (Instrument manual)
• No man is an island – doing it all yourself
• When in doubt – call the technical representative!
• Contamination
• Poor data interpretation - not considering alternative explanations
• Stopping when you see the answer you like or expect
• Giving up too soon

Not Cleaning Up After Yourself !!

Particle properties - Density
 Densities

Particle density

Absolute density

Bed density

Bulk density
 solid material

pores

Particle density
 solid material

pores

hydrodynamic volume

mass of particle
Particle density 
hydrodynam ic volume of particle
 solid material

pores

Absolute density
 solid material

pores

material volume

mass of particle
Absolute density 
volume of solids material making up the particle
Bed density =

Bulk density =
 Bed cross-sectional area A

Bed height, H
Bed volume = HA

Bed density =

Bulk density =
 Bed cross-sectional area A

Bed height, H
Bed volume = HA

mass of particles in a bed

Bed density =
volume occupied by particles and voids between them

mass of particles
Bulk density = volume occupied by particles and voids between them
Relevant particle size:

equivalent sphere diameter

hydrodynamic
derived from a measurement
diameter
technique involving
hydrodynamic interaction
between the particle and fluid.
 Relevant particle size:

equivalent sphere diameter

hydrodynamic
derived from a measurement
diameter
technique involving
hydrodynamic interaction
between the particle and fluid.

Or, from sieve diameter or similar:

1
xp 
mi
 xi
 Relevant particle size:

equivalent sphere diameter

hydrodynamic
derived from a measurement
diameter
technique involving
hydrodynamic interaction
between the particle and fluid.

Or, from sieve diameter or similar:

1 Harmonic mean
xp  of the mass
mi
 xi distribution
 Relevant particle size:

equivalent sphere diameter

hydrodynamic
derived from a measurement
diameter
technique involving
hydrodynamic interaction
between the particle and fluid.

Or, from sieve diameter or similar:

1 Harmonic mean
xp  of the mass
Surface-volume
mi
 xi distribution
mean diameter

D(3, 2)