 |
 |
Introduction
SPARCLE is an advanced Optical Particle Counter, OPC. In a
conventional OPC, an individual particle is illuminated by a light
source the scattering intensity is then measured. This measurement is
then related to the size of the particle from a calibration
curve. Unfortunately the light intensity is also a function of the
particles refractive index. For a synthetic aerosol made from one
material this is of little consequence but for a mixed aerosols such
as those found in the atmosphere errors can be as large as a factor of
three (Willeke 1993). SPARCLE turns
this dependence on it's head, inferring both the refractive index and
particle size, offering a unique concept for aerosol measurement.
No aerosol instrument currently exists with this capability and this
makes SPARCLE an ideal instrument to study the radiative impact of
aerosols a recognised area of uncertainty in the assessment of climate
change (IPCC, 2007). Integration of
individual particle quantities can give properties of the aerosol,
such as, particle size distribution, the total aerosol phase function
and the single scatter albedo.
Instrument history
Phase I instrument
A proof-of-concept Stratospheric Aerosol Composition and Loading
Experiment (SPARCLE) instrument was designed and developed by Dr Gareth
Thomas to measure the complex refractive index and size of individual
aerosol particles. Basic principles were observed and reported during
this work at University of Canterbury, New Zealand under the supervision of
Dr Grainger. Development the mathematical tools and algorithms for the
instrument were also undertake. Much was learned about the basic
principles of the instrument. Since this work Dr Thomas and Dr Grainger
(his supervisor) moved to Oxford. Detailed information on the phase I instrument can be found in Thomas 2003.
Phase II instrument
In 2007 HEFCE funding was secured to develop the instrument
further, the main aim of this project was to move design from a
stratospheric instrument into one suitable for use in troposphere
studies. Dr Dan Peters then applied the lessons learned on the
instrument design by Thomas and building a breadboard implementation of
the new design. In addition new design tools were developed to predict
the instrument performance, by both Dr Peters and Mr Andrew Smith (as
part of his PhD). Detail information on the Phase II instrument can be found in Peters 2009.
Basic principles
The instrument measures both the intensity of the scattered laser
light from one particle at a time using a photomultiplier tube, PMT,
and the angular light scattering pattern using a linear detector array
LDA. The light scattering pattern allows individual particle
refractive index and particle size to be derived.
Due to the large amount of data from the LDA signal, the signal is
pre-processed to reduce the number of measured parameters (termed the
measurement vector) required in the analysis algorithm (to find the
particle size and refractive index).
 | |
Fig. 1. Example LDA
signal.
 | |
Fig. 2. Smoothed LDA signal (in
black), crosses show the points used in the particle analysis. Red
line is the curve fitted by Mie theory to these data
points.
 | |
Fig. 3. The Fourier transform of the
LDA signal is shown in black. The red line shows the Fourier transform
of the fitted Mie function from the analysis.
An example LDA signal for the Phase I instrument is shown
in Fig.1. To reduce the number of measured parameters passed to the
analysis the signal is transformed in the following way to be
represented by around 21 parameters rather than all the LDA
pixels. First the signal is smoothed to remove the high frequency
spatial variations, this is plotted as the black line in Fig.2. From
this smoothed curve a 20 points are sampled and these are then
used in the analysis, these points are plotted in Fig.2. as crosses.
To represent the high frequency information a Fourier transform is
taken of the LDA signal minus the smoothed LDA signal (Fig3. black
line). The peak of the Fourier transform is then found, and this
spatial frequency is used one the measured parameters. Thus the
analysis algorithm input parameters are reduced from the number of LDA
pixels to just 21 parameters.
These are then fitted to Mie theory. The fits for the example LDA
signal when compared to Mie theory as shown as the red line in the
above plots. For more detail on the analysis method see Thomas 2003.
Instrument performance
Measurement signal-to-noise
 | |
Fig. 4. Predicted Phase II detector performance.
Using modelling software it is possible to obtain size resolved
instrument performance plots-these are shown in Fig.4. The PMT
detector allows the possibility of detecting the smallest particles,
whilst the LDA affords the opportunity to resolve the scattering
pattern of the particles.
Fig. 4 shows the LDA is sensitivities to particles 100nm in radius and
above. The LDA covers the coarse and fine fraction with the PMT
providing sensitivity from 80nm radius. The limiting factor for the
LDA performance is the dark noise, for the PMT it is Rayleigh
scattering from the surrounding air molecules.
Both detectors are not photon limited. The photon limit for the
detectors is 30nm for the LDA and 15nm of the PMT (the radius where
the blue and black lines cross in Fig. 9). This indicates that some
performance enhancements may be possible with further development of
the instrument optics. For further discussion of the instruments
performance see Peters 2009.
Refractive index and particle sizing range
 | |
Fig. 5. Number of degrees of freedom for the Phase II instrument.
Having sufficient signal to construct a measurement vector is one
thing, but what will this mean in terms of the instruments ability to
determine refractive index and particle size? To help answerer this
question we can compute the number of degrees of freedom of the
analysis system. This analysis does not include the signal-to-noise
ratio but rather calculates the number of independent parameters that
can be retrieved given sufficient signal-to-noise. Thus when
interpreting this plot we must also referrer to the signal-to-noise
plots to get the complete picture of the instruments performance.
The result of this analysis are shown in Fig. 5 for a number of
different refractive index values and particle sizes found in the
atmosphere. Were the plot shows three degrees of freedom, the complex
refractive index (real and imaginary parts) and the particle size can
be determined, these areas are shown in red. The green areas
designate were only two independent parameters can be determined,
i.e. particle size and the real part of the refractive index. The blue
areas show were size only can be obtained.
Thus the instrument can provide the full refractive index and particle
size measurements for nearly all refractive index values for particles
200nm radius and above, and there is good signal-to-noise from
both the LDA and PMT in this range to make this measurement possible
(see Fig. 4.)
Future work
The current LDA signal is not limited by the noise shown in Fig. 4 but
rather by the digitization noise of the electronics. Thus additional
gain is required to ensure the instrument is limited by shot noise.
In addition to allow extend operation, or operation by inexperience
personnel the instruments flow control need to be modified for closed
loop control. This will allow high quality number density information
to be obtain which is not currently possible as the flows must be
monitored and pump flows manually adjusted.
Links and references
IPCC: Climate Change 2007 - The Physical Science Basis, Contribution of Working Group I to the Fourth Assessment Report of the IPCC, Cambridge University Press
Peters D., Grainger R.G., Smith A, Final Report: Characterising near surface aircraft PM, Omega, 2009
Thomas, GE,A new instrument for stratospheric aerosol measurement, PhD Thesis, University of Canterbury, 2003.
Willeke, K., Baron P. A., Aerosol
measurement: principles, techniques, and applications, Van Nostrand
Reinhold, 1993.
