Title: Dynamic Light Scattering
1Dynamic Light Scattering
- ZetaPALS w/ 90Plus particle size analyzer
Also equipped w/ BI-FOQELS Otsuka DLS-700 (Rm
CCR230)
2- Dynamic Light Scattering (DLS)
- Photon Correlation Spectroscopy (PCS)
- Quasi-Elastic Light Scattering (QELS)
- Measure Brownian motion by
- Collect scattered light from suspended particles
to - Obtain diffusion rate to
- Calculate particle size
3Brownian motion
- Velocity of the Brownian motion is defined by the
Translational Diffusion Coefficient (D) - Brownian motion is indirectly proportional to
size - Larger particles diffuse slower than smaller
particles - Temperature and viscosity must be known
- Temperature stability is necessary
- Convection currents induce particle movement that
interferes with size determination - Temperature is proportional to diffusion rate
- Increasing temperature increases Brownian motion
4Brownian motion
- Random movement of particles due to bombardment
of solvent molecules
5Stokes-Einstein Equation
- dH hydrodynamic diameter (m)
- k Boltzmann constant (J/Kkgm2/s2K)
- T temperature (K)
- ? solvent viscosity (kg/ms)
- D diffusion coefficient (m2/s)
6Hydrodynamic diameter
Particle diameter
- The diameter measured by DLS correlates to the
effective particle movement within a liquid - Particle diameter electrical double layer
- Affected by surface bound species which slows
diffusion
Hydrodynamic diameter
7Nonspherical particles
Rapid
Equivalent sphere
Slow
Hydrodynamic diameter is calculated based on the
equivalent sphere with the same diffusion
coefficient
8Experimental DLS
- Measure the Brownian motion of particles and
calculate size - DLS measures the intensity fluctuations of
scattered light arising from Brownian motion - How do these fluctuations in scattered light
intensity arise?
9What causes light scattering from (small)
particles?
- Explained by JW Strutt (Lord Rayleigh)
- Electromagnetic wave (light) induces oscillations
of electrons in a particle - This interaction causes a deviation in the light
path through an angle calculated using vector
analysis - Scattering coefficient varies inversely with the
fourth power of the wavelength
10Interaction of light with matterRayleigh
approximation
- For small particles (d ?/10), scattering is
isotropic - Rayleigh approximation tells us that
- I a d6
- I a 1/?4
- where I intensity of scattered light
- d particle diameter
- ? laser wavelength
11Mie scattering from large particles
- Used for particles where d ?0
- Complete analytical solution of Maxwells
equations for scattering of electromagnetic
radiation from spherical particles - Assumes homogeneous, isotropic and optically
linear material
Stratton, A. Electromagnetic theory, McGraw-Hill,
New York (1941) www.lightscattering.de/MieCalc
12Brownian motion and scattering
Constructive interference
Destructive interference
13Brownian motion and scattering Multiple particles
14Instrument layout
15Intensity fluctuations
- Apply the autocorrelation function to determine
diffusion coefficient - Large particles smooth curve
- Small particles noisy curve
16Determining particle size
- Determine autocorrelation function
- Fit measured function to G(t) to calculate G
- Calculate D, given n, ?, and G
- Calculate dH, given T and ?
User defined values.
17How a correlator works
- Random motion of small particles in a liquid
gives rise to fluctuations in the time intensity
of the scattered light - Fluctuating signal is processed by forming the
autocorrelation function - Calculates diffusion
18How a correlator works
- Large particles the signal will be changing
slowly and the correlation will persist for a
long time - Small, rapidly moving particles the correlation
will disappear quickly
19The correlation function
- For monodisperse particles the correlation
function is - Where
- A baseline of the correlation function
- Bintercept of the correlation function
- G Dq2
- Dtranslational diffusion coefficient
- q(4pn/?0)sin(?/2)
- nrefractive index of solution
- ?0wavelength of laser
- ?scattering angle
20The correlation function
- For polydisperse particles the correlation
function becomes - where g1(t) is the sum of all exponential decays
contained in the correlation function
21Broad particle size distribution
- Correlation function becomes nontrivial
- Measurement noise, baseline drifts, and dust make
the function difficult to solve accurately - Cumulants analysis
- Convert exponential to Taylor series
- First two cumulants are used to describe data
- G Dq2
- µ2 (D2-D2)q4
- Polydispersity µ2/ G2
22Cumulants analysis
- The decay in the correlation function is
exponential - Simplest way to obtain size is to use cumulants
analysis1 - A 3rd order fit to a semi-log plot of the
correlation function - If the distribution is polydisperse, the semi-log
plot will be curved - Fit error of less than 0.005 is good.
1ISO 133211996 Particle size analysis -- Photon
correlation spectroscopy
23Cumulants analysis
- Third order fit to correlation function
- b z-average diffusion coefficient
- 2c/b2 polydispersity index
- This method only calculates a mean and width
- Intensity mean size
- Only good for narrow, monomodal samples
- Use NNLS for multimodal samples
24Cumulants analysis
25Polydispersity index
- 0 to 0.05 only normally encountered w/ latex
standards or particles made to be monodisperse - 0.05 to 0.08 nearly monodisperse sample
- 0.08 to 0.7 This is a mid-range polydispersity
- gt0.7 Very polydisperse. Care should be taken in
interpreting results as the sample may not be
suitable for the technique (e.g., a sedimenting
high size tail may be present)
26Non-Negatively constrained Least Squares (NNLS)
algorithm
- Used for Multimodal size distribution (MSD)
- Only positive contributions to the
intensity-weighted distribution are allowed - Ratio between any two successive diameters is
constant - Least squares criterion for judging each
criterion is used - Iteration terminates on its own
27Correlation functionCorrelograms
- Correlograms show the correlation data providing
information about the sample - The shape of the curve provides clues related to
sample quality - Decay is a function of the particle diffusion
coefficient (D) - Stokes-Einstein relates D to dH
- z-average diameter is obtained from an
exponential fit - Distributions are obtained from
multi-exponential fitting algorithms - Noisy data can result from
- Low count rate
- Sample instability
- Vibration or interference from external source
28Correlation functionCorrelograms
29Data interpretationCorrelograms
- Very small particles
- Medium range polydispersity
- No large particles/aggregrates present (flat
baseline)
30Data interpretationCorrelograms
- Large particles
- Medium range polydispersity
- Presence of large particles/agglomerates (noisy
baseline)
31Data interpretationCorrelograms
- Very large particles
- High polydispersity
- Presence of large particles/agglomerates (noisy
baseline)
32Upper size limit of DLS
- DLS will have an upper limit wrt size and density
- When particle motion is not random (sedimentation
or creaming), DLS is not the correct technique to
use - Upper limit is set by the onset of sedimentation
- Upper size limit is therefore sample dependent
- No advantage in suspending particles in a more
viscous medium to prevent sedimentation because
Brownian motion will be slowed down to the same
extent making measurement time longer
33Upper size limit of DLS
- Need to consider the number of particles in the
detection volume - Amount of scattered light from large particles is
sufficient to make successful measurements, but - Number of particles in scattering volume may be
too low - Number fluctuations severe fluctuations of the
number of particles in a time step can lead to
problems defining the baseline of the correlation
function - Increase particle concentration, but not too high
or multiple scattering events might arise
34Detection volume
Detector
Laser
35Lower particle size limit of DLS
- Lower size limit depends on
- Sample concentration
- Refractive index of sample compared to diluent
- Laser power and wavelength
- Detector sensitivity
- Optical configuration of instrument
- Lower limit is typically 2 nm
36Sample preparation
- Measurements can be made on any sample in which
the particles are mobile - Each sample material has an optimal concentration
for DLS analysis - Low concentration ? not enough scattering
- High concentration ? multiple scattering events
affect particle size
37Sample preparation
- Upper limit governed by onset of
particle/particle interactions - Affects diffusion speed
- Affects apparent size
- Multiple scattering events and particle/particle
interactions must be considered - Determining the correct particle concentration
may require several measurements at different
concentrations
38Sample preparation
- An important factor determining the maximum
concentration for accurate measurements is the
particle size
39Sample concentrationSmall particles
- For particle sizes lt10 nm, one must determine the
minimum concentration to generate enough
scattered light - Particles should generate 10 kcps (count rate)
in excess of the scattering from the solvent - Maximum concentration determined by the physical
properties of the particles - Avoid particle/particle interactions
- Should be at least 1000 particles in the
scattering volume
40Sample preparation
- When possible, perform DLS on as prepared sample
- Dilute aqueous or organic suspensions
- Alcohol and aggressive solvents require a
glass/quartz cell - 0.0001 to 1(v/v)
- Dilution media (1) should be the same (or as
close as possible) as the synthesis media, (2)
HPLC grade and (3) filtered before use - Chemical equilibrium will be established if
diluent is taken from the original sample - Suspension should be sonicated prior to analysis
41Checking instrument operation
- DLS instruments are not calibrated
- Measurement based on first principles
- Verification of accuracy can be checked using
standards - Duke Scientific (based on TEM)
- Polysciences
42Count rate and z-average diameter Repeatability
- Perform at least 3 repeat measurements on the
same sample - Count rates should fall within a few percent of
one another - z-average diameter should also be with 1-2 of
one another
43Count rateRepeatability problems
- Count rate DECREASES with successive measurements
- Particle sedimentation
- Particle creaming
- Particle dissolution or breaking up
- Resolution
- Prepare a better, stabilized dispersion
- Get rid of large particles
- Coulter
44Count rateRepeatability problems
- Count rate is RANDOM with successive measurements
- Dispersion instability
- Sample contains large particles
- Bubbles
- Resolution
- Prepare a better, stabilized dispersion
- Remove large particles
- De-gas sample
45Z-average diameterRepeatability problems
- Size DECREASES with successive measurements
- Temperature not stable
- Sample unstable
- Resolution
- Allow plenty of time for temperature
equilibration - Prepare a better, stabilized dispersion
46Repeatability of size distributions
- The sized distributions are derived from a NNLS
analysis and should be checked for repeatability
as well - If distributions are not repeatable, repeat
measurements with longer measurement duration
47References
- http//www.bic.com/90Plus.html
- http//www.brainshark.com/brainshark/vu/view.asp?t
extM913802pi62212 - http//www.malvern.co.uk/malvern/ondemand.nsf/frmo
ndemandview - http//www.brainshark.com/brainshark/vu/view.asp?t
extM913802pi96389 - http//www.brainshark.com/brainshark/vu/view.asp?t
extM913802pi73504 - http//physics.ucsd.edu/neurophysics/courses/physi
cs_173_273/dynamic_light_scattering_03.pdf - http//www.brookhaven.co.uk/dynamic-light-scatteri
ng.html - Dynamic Light Scattering With Applications to
Chemistry, Biology, and Physics, Bruce J. Berne
and Robert Pecora, DOVER PUBLICATIONS, INC.
Mineola. New York - Scattering of Light Other Electromagnetic
Radiation, Milton Kerker, Academic Press (1969)
48(No Transcript)
49Evaluating the correlation function
- If the intensity distribution is a fairly smooth
peak, there is little point in conversion to a
volume distribution using Mie theory - However, if the intensity plot shows a
substantial tail or more than one peak, then a
volume distribution will give a more realistic
view of the importance of the tail or second peak - Number distributions are of little use because
small error in data acquisition can lead to huge
error in the distribution by number and are not
displayed
50Correlogram from a sample of small particles
51Correlogram from a sample of large particles
52Extra
- Time-dependent fluctuations in the scattered
intensity due to Brownian motion - Constructive and destructive interference of
light - Decay times of fluctuations related to the
diffusion constants --- particle size - Fluctuations determined in the time domain by a
correlator - Correlation average of products of two
quantities - Delay times chosen to be much smaller than the
time required for a particle to relax back to
average scattering