AN OPTICAL PARTICLE AND FLUX MONITOR
FOR STACK EMISSIONS
FINAL REPORT
SDL No. 86-2466-05D
SPECTRON
DEVELOPMENT
LABORATORIES, INC.
-------
AN OPTICAL PARTICLE AND FLUX MONITOR
FOR STACK EMISSIONS
FINAL REPORT
SDL No. 86-2466-05D
3 April 1986
Prepared for:
Environment Protection Agency
Research Triangle Park, NC 27711
SPECTRON
DEVELOPMENT
LABORATORIES, INC.
33O3 HARBOR BLVD.. SUITE G-3. COSTA MESA. CA 92626 • C71 4)
-------
TABLE OF CONTENTS
NO. PAGE
1.0 INTRODUCTION 1
2.0 SUMMARY OF WORK CONDUCTED UNDER PHASE 1 2
2.1 Summary of Tasks 2
3.0 DESCRIPTION OF OPTICAL TECHNIQUES 4
3.1 Laser Sheet Nephelometer 4
3.2 Optical System Design Consideration 6
4.0 DESCRIPTION OF EXPERIMENTS 11
4.1 The Optical Breadboard 11
4.2 Electronic Hardware 13
5.0 EXPERIMENTAL RESULTS 18
6.0 FUTURE WORK 24
7.0 CONCLUSIONS 27
-i- 86246605tc/44
-------
LIST OF FIGURES
NO. PAGE
1 Conceptual Diagram of Laser Sheet Nephelometer 5
2 Perpendicularly Polarized Scattered Light
Intensity as a Fuction of the Size Parameter a 8
3 Collection Mask 9
4 Integrated Intensity Function for Small Angle
Collection for Uniform Latex Particles a s Function
of Diameter 10
5 Optical Breadboard 12
6 Photograph of Optical System 14
7 Electronic Hardware 15
8 Block Diagram of Electonic Interface Box 16
9 Oscilloscope Traces Showing Intensity Peaks for
2.26 ym and 3.30 ym Diameter Uniform Latex
Particles .19
10 Histogram Showing Experimental Data Obtained with
1.7 ym Polystryrene Sphere in 2.8 Seconds 20
11 Histogram Showing Experimental Data Obtained with
2.26 um and 3.30 ym Uniform Latex Particles 21
12 Histogram Showing Experimetnal Data Obtain with
1.74 urn, 2.26 ym, and 3.30 ym Uniform Latex
Particles 22
13 Comparison of Laser Diode Outputs With and
Without the Fiber-Optic Coupler 25
-ii- 86246605tc/44
-------
1.0 INTRODUCTION
We have studied and demonstrated a laser sheet nephelometer tech-
nique which shows great potential for measuring the particles in
stacks. The emphasis in this first phase was to demonstrate the feasi-
bility of the optical technique. To do so, we used conventional helium
neon lasers and optical components which are large and will not fit in
an actual stack.. The prototype developed during the second phase will
make use of small and rugged solid state devices (laser diode, photo-
diode array) resulting in a very compact unit which can be installed in
a stack.
An optical breadboard was constructed and interfaced to a tran-
sient digitizer and a computer. Signals recorded from the transient
digitizer were carefully analyzed by the computer. This way the feasi-
bility of various scattering techniques can be tested, thus arriving to
an optimum method. Polystyrene particles of known size (1.7 ytn to 3.3
pm) were seeded in a flow and measured. These polystyrene particles are
excellent for testing the accuracy and resolution of a size measuring
system since their size and concentration are well known. Notice that a
system must be tested with a known size distribution in order to assess
its performance. The results compared very well with the theoretical
predictions.
-1- 86-2466-05/44
-------
2.0 SUMMARY OF WORK CONDUCTED UNDER PHASE I
In this section a brief account of the effort conducted under
Phase I is reported. This will provide readers an overall view of the
program while each section will discuss in detail the various accom-
plished tasks.
2.1 Summary of Tasks
a) Establish the particulate levels, expected size distribution, and
particle composition of interest to the EPA.
b) Analytical computations of the scattering functions with various
collection apertures.
c) Evaluation of two scattering techniques and the selection of the
most appropriate for the job.
d) Design and construction of an optical breadboard and interfacing
it to dedicated electronics. Size histograms of particles were
obtained with this system.
-2- 86-2466-05/44
-------
e) Design and implementation of experiments using polystyrene parti-
cles of known size (in the 1.7 urn to 3.3 ym range) entrained in a
flow of hot air.
f) Experimental evaluation of the scattering function yielding the
size distributions of interest. This evaluation included an
algorithm to discriminate against noise.
g) Data analysis and comparison with the theoretical results.
h) Evaluation of replacing the gas laser with a diode laser and the
photomultiplier tube with a solid state detector. This task
essentially defines the requirements and proposed approach of
Phase II.
-3- 86-2466-05/44
-------
3.0 DESCRIPTION OF OPTICAL TECHNIQUES
For a mass loading of 3 x 10"^ gm/ft^ the concentration of parti-
cles is about 80/cm , if we assume a typical size of 5 ym. This concen-
tration calls for single particle counters in contrast with an ensemble
measurement technique. The main reason is the need for measuring parti-
cle concentration. Techniques such as turbidimetry will not have the
required sensitivity since the total attenuation could be very small.
The single particle counters can measure the concentration from the rate
of individual pulses and the size of the probe volume.
Two single particle counters offer the potential of working in
stacks: a) laser diffraction; b) laser sheet nephelometer. Signal-to-
noise considerations and the simpler algorithms to extract the data lead
us to choose the second technique.
3.1 Laser Sheet Nephelometer
The concept is shown in Figure 1. The center of the laser beam
is identified by the intersection of the images of two collection aper-
tures cut on a special mask at different angles from the center of the
receiving lens.
Two probe volumes are thus formed: a large probe volume viewed
from the shallow angle, and a smaller intersecting probe volume viewed
from the large angle. The configuration is such that particles in the
small probe volume are totally immersed in the large probe volume, thus
avoiding signal masking and ambiguities. Only particles crossing this
intersection are accepted as valid. The light scattered by the
-4- 86-2466-05/44
-------
Receiving Lenses
Mask
Image of
Aperture
i
Ln
Aperture
Figure 1. Conceptual Diagram of the Laser Sheet Nephelometer.
-------
particles through the large angle is used as a trigger, and the light
intensity collected (simultaneously) through the small angle is inverted
to obtain the size of the particle.
The scattered light intensity from particles crossing a Gaussian
probe volume is given by:
2 2
I8(A) = IoK(ct,n,9) exp [-2 (*j + I-) ]
b b
x y
where IQ is the central intensity of the incident beam, a is the size
parameter, n is the index of refraction of the particle and 9 is the
scattering angle, and b and bw are the waist radii of the laser
x y
sheet. Since the laser beam is expanded in the x direction, b can be
X
quite large. The x-dimension of the probe volume intersected by the L
aperture of the receiver is only ~ 1/6 of bx. The exp (-2x /^x ) term
can then be regarded as constant. Therefore,
2
P(y) = I K(a,n,9) exp [ 1-]
b
y
is the pedestal of the signal.
the size (a) is obtained by measuring P(y) and solving for
K(a,n,9). The velocity can also be obtained by measuring the time t or
the particle to cross 2b .
3.2 Optical System Design Consideration
There are several factors to be considered in designing an
optical system for stacks. The dimension of the probe volume is deter-
mined from the particle density. Too large a probe volume could result
-6- 86-2466-05/44
-------
in multiple particle scattering. On the other hand, if the probe volume
is too small an extensive time may be required to collect a significant
amount of data.
Care must also be exercised in selecting the optical configura-
tion since oscillations in the scattering function can be observed for
particles around 1 urn. This problem was circumvented by choosing a near
forward scattering scheme. In this near forward scattering, not only
the Mie oscillations are reduced to a minimum, but the scattering inten-
sity is also insensitive to the variations of the refractive index of
the scatterer. Figure 2 shows the scattered intensity as a function of
the particle size parameter ot for different refractive indices at
different scattering angles. The results indicate that near forward
scattering angles of collection (6 < 5°) offer the desired optical char-
acteristics. Notice that the triggering signal from the large angle can
be subject to oscillations without compromising the size distribution.
The receiving aperture limiting the large and shallow angles is shown on
Figure 3. The scattered light was theoretically evaluated and
integrated over the apertures. Figure 4 shows the Mie calculation of
the scattered light intensity from polystyrene particles (n = 1.59)
collected by the small aperture as a function of particle diameter.
Notice that there are still some oscillations which can produce size
errors of up to 20%. For the most part the curve is smooth and will
yield accurate results.
The apertures shown on Figure 3 insure that particles detected by
the large angle optics are unobstructedly measured by the shallow angle
optics. Thus, the signals from which the size is obtained are totally
unmasked.
-7- 86-2466-05/44
-------
FL-1
icr
10 -
10-
H
M
CO
§
H
!3
H
S
es
W
H
51
u
CO
10
10
10
1.56 - ±2.9 -x. 10~7
1.66 - 12.9 x 10~7
1.96 - 12.9 x 10~7
1.66
1.66
10
15
20
25
30
SIZE PARAMETER a
Figure 2. Perpendicularly polarized scattered light intensity as a
function of the size parameter a
-8-
85-2409-01/13
-------
Figure 3. Collection Mask
-9-
86-2456-03
-------
86-2466-03
53
O
W
CO
53
O
PL,
CO
w
pi
10.0.
8.0"-
6.0.
4.0
2.0-
1.0-
0.8-
0.6'
0.4
0.2 —
0.1
3.30±0 .12pm
r I
. 2 3
DIAMETER (y m)
1
4
Figure 4.
Integrated Intensity Function for Small Angle
Collection for Uniform Latex Particles as a
Function of Diameter.
-10-
-------
4.0 DESCRIPTION OF EXPERIMENTS
4.1 The Optical Breadboard
The optical breadboard is shown in Figure 5. The 25 mW HeNe
laser emits a beam of light (X = 633 nm) 1.25 mm in diameter, with a
full angle divergence of 0.7 m rad. This beam is then steered by three
turning mirrors to define the system's optical axis. The first cylin-
drical lens L, is located about 430 mm from the exit of the laser, and
expands the x-dimension of the laser beam. The second cylindrical lens
L£ [210 mm past Li] contracts the y-dimension of the laser beam. The
third lens L-J is a spherical lens and is located 113 mm from lens L,2«
The above configuration results in a collimated laser sheet with dimen-
sions 2.3 mm in the x-direction and 0.36 mm in y-direction. These
mirrors and lenses together with the laser define the transmitter.
The second part of the optical breadboard is the receiver. The
first component of the receiver is the mask which defines the collection
apertures. This mask is shown in Figure 3. . This mask, and the light
collected through it, is imaged by the spherical lenses L^ and Lr,
through a pinhole. The pinhole is 200 mm from Lc and is 800 \im in dia-
meter. Lenses Lg and Lj are relay lenses used to help separate the
images of the two sections of the mask; so that they can be steered into
their respective photomultiplier tubes; this is done with mirrors M. and
Me. There is one lens in each photomultiplier tube housing, Lg and LQ,
which images the light onto each photocathode.
An electric current is generated when the scattered light falls
onto the photocathode. The signal is then analyzed in the electronic
-11- 86-2466-05/44
-------
25 mW Helium-Neon Laser
Probe
Volume
5
L7 L8
and
N
Focal Length 250
(mm)
Diameter 50
(mm)
AO 73 100 200 220 70 13 13
15 28 50 75 52 45 5 5
\
.M-
Cylindrical
Cylindrical
x^Axis
00
i
N>
ON
O
U>
Figure 5. Optical Breadboard
-------
processor. A photograph of the complete optical system is shown on
Figure 6.
To verify the dimensions of the probe volume, a string of mono-
disperse water droplets was generated (from a Berglund-Liu tonodisperse
Droplet Generator TSI Model //3050) and traversed in the
z-direction. The test showed that the signals from the small angle were
virtually unmasked (90%) for particles just outside the large angle
probe volume. To test the system dispersions of monodisperse poly-
styrene particles were generated and measured. These particles come in
a suspension and were diluted in water. The diluted suspension was
nebulized, thus producing a mist which carried the particles. This two-
phase mist was introduced by a compressor into a heating chamber where
the water was evaporated leaving the polystyrene particles. These par-
ticles were dispersed into the probe volume where the size and velocity
are measured.
4.2 Electronic Hardware
Figure 7 shows the system used to record the signals from the
detectors. A ten-times amplifier and a low-pass filter are used to con-
dition the signal before it enters an electronic interface box. This
box was built specifically for this test (see Figure 8 for block
diagram). The interface box contains scaling amplifiers to adjust the
signal level to the optimum recording level. A threshold trigger
circuit detects the presence of a pulse on the large angle channel which
initiates a gated clock signal consisting of 64 clock pulses with a 1.6
microsecond separation. This clock is used by the LeCroy event recorder
-13- 86-2466-05/44
-------
Figure 6. Photograph of Optical System
-14-
86-2456-03
-------
Large Angle
Photomultiplier Tube
— x!0^-«>
Low
Pass
Ln
I
Electronic
Interface
Box
Large Angle
Trigger Signal
Digitizing Clock
Small Angle
Intensity Signal
Channel A
Lecroy
;Digitizer
GPIB
Interface
IBM
PC
Channel B
Small Angle
Photomultiplier Tube
——»• xlO^>—°
Low
Pass
00
ON
i
ro
O
U>
Figure 7. Electronic Hardware
-------
00
O^
I
fo
-O
O
Channel A
I nput.
(Large Angle)
500
Nanosecond
Delay
Line
Gain
Inverting Amplifier
Voltage
Comparator
Count Gate
1.25 MHz
Crystal
Oscillator
Channel B
Input
(Small Angle)
Gain
Inverting Amplifier
Channel A
- Output
(To LeCroy)
Channel A
Monitor
500
Nanosecond
Delay
Line
Gated
Clocks
(To LeCroys
64 Clock
Pulses
Per Event
Channel B
o Output
(To LeCroy)
Channel B
Monitor
FigureS. Block Diagram of Electronic Interface Box
-------
to digitize and store an 8-bit value for the incoming signal for each
clock pulse. Thus, Channel A and Channel B signals are stored digitally
in the memory of the LeCroy. Each pulse consisted of 64 digitized
values covering a time span of 102.4 microseconds. A software program
from LeCroy Research called "Waveform Catalyst" is used on an IBM PC to
control the collection of data and to transfer data fields via a GPIB
interface from the digitizer to the computer.
-17- 86-2466-05/44
-------
5.0 EXPERIMENTAL RESULTS
A sample of the results obtained during Phase 1 are presented
here. Polystrene particles of 1.7 ym, 2.26 ]m and 3.3 ym in diameter
were tested in experiments. Figure 9 shows a typical record of a mix-
ture of 2.26 urn and 3.3 ]sm particles. The upper trace obtained from the
small angle signal can be classified into two intensity peak levels.
The larger peaks are corresponding to the scattered signal from 3.3 ym
while the smaller ones are due to 2.26 pm particles. This classifica-
tion is not present in the lower trace which is corresponding to the
large angle signal. We also applied some criteria in validating data.
For example, a sudden increase of signal followed by a negative slope is
deemed to be invalid. In this case, the signal is acquired after par-
ticles pass through the peak intensity of the Gaussian profile of the
laser beam. This will result in erroneous data.
Figure 10 shows a histogram of monodisperse polystyrene particles
of 1.7 ym. Different combinations of polystyrene particles can be
diluted into water, thus producing monodisperse, bimodal and trimodal
distibution. Figure 11 shows the results corresponding to a bimodal
distribution of 2.26 ym and 3.3 ym. Figure 12 shows a histogram of a
trimodal distribution of 1.7 ym, 2.26 ym and 3.3 ym. Throughout this
experiment, the gain of the small angle PMT was kept constant. Thus,
the histogram of 1.7 ym polystyrene in Figure 10 can be used as a cali-
bration source. Signal amplitudes can be converted into particle size
diameters using the integrated Mie calculations; this is shown in the
abscissas of Figures 11 and 12. The arrows indicate the corresponding
size from the theoretical calculations.
-18- 86-2466-05/44
-------
small peaks: 2.26ym
large peaks: 3.3pm
Small Angle
100 mV/div
200 p s/div
Large Angle
100 mV/div
200 p s/div
Figure 9. Oscilloscope Traces Showing Intensity Peaks
for 2.26 pm and 3.30 pm Diameter Uniform
Latex Particles.
00
o\
i
10
*»
ON
o\
i
o
U)
-------
100 -
80 -
60-
§
u
40-
20-
100
Peak Voltage (mV)
Figure 10. Histogram Showing Experimental Data Obtained with
1.7 ym Polystyrene Spheres in 2.8 Seconds.
-20-
86-2456-03
-------
50 —
40-
H
Z
§
u
30 —
20 —
10 _
1.6 1.8 2.0
3.4 3.0
3.55
DIAMETER
Figure 11. Histogram Showing Experimental Data Obtained
with 2.26 jJm and 3.30 pm Uniform Latex Particles.
oo
I
N>
-------
N)
to
50 —
3.30 pm
3.55
DIAMETER (
Figure 12. Histogram Showing Experimental Data Obtained
with 1.74 pm, 2.26ym, and 3.30ym Uniform
Latex Particles.
oo
i
I
o
U)
-------
Q ^
The probe volume of the small angle aperture is 0.6 x 10 urn .
1 ^
This is equivalent to a limiting particle density of 1.6 x 10 I/cm .
The actual trigger probe volume defined by the large angle aperture is
much smaller. Experimental data were obtained by adjusting polystyrene
particle density to 6 1/cm^ with a velocity of 8 m/s. Thus, a 50 1/s
data rate was achieved.
-23- 86-2466-05/44
-------
6.0 FUTURE WORK
As stated in the introduction, the emphasis in this first phase
was to demonstrate the feasibility of the optical technique. This sec-
tion discussed the design modifications needed to implement a working
prototype to be used in a stack.
In the transmitter, the 25 mW helium-neon laser will be replace
by a 50 mW CW GaAIAs laser diode in a fiber coupled package. A power
supply will need to be purchased, and optional features could include
monitoring photodiodes to indicate the output power on a continuous
basis and a thermo-electric cooler to keep the diode at room tempera-
ture. This laser diode was chosen with the fiber-optic output coupler
because of the high output power needed for particle sizing, and the
uniformity of the output beam from the fiber optic. Figure 1 compares
the output of the laser diode with and without the fiber-optic output
coupler.
The output from the fiber-optic coupler diverges quickly. There-
fore, a special laser diode collimating lens must be used, such as
Melles Griot Optical Components (Product Number 06-GLC-001). This will
yield a collimated laser beam 4 mm in diameter. A pair of cylindrical
lenses of focal lengths, 100 mm and 12.7 mm, will then compress the y-
dimension of the laser beam into a laser sheet 0.5 mm thick. The total
overall length of the transmitter is estimated to be 130 mm.
Modifications to the receiver are minimal, the mask, aperture and
receiving lenses will remain the same. The relay lenses will not be
needed because the photomultiplier tubes will be replaced with solid
-24- 86-2466-05/44
-------
FAR FIELD OF SPECTRA DiODE LASERS
Fiber Coupled and
Fiber Stub Lasers
Open Heat Sink and
Window Package Lasers
A A
25 15 505 15 25
FAR FIELD ANGLE 9 (deg)
L
25 15 505 15 25
FAR FIELD ANGLE 0,i (deg)
FIGURE 13. Comparison of Laser Diode Outputs With
and Without the Fiber-Optic Coupler.
-25-
86-2466-04
-------
state silicon avalanche photodiodes with integral fiber-optic cables
(for example, RCA Product Number C30902EQC). The input of each fiber-
optic cable will be positioned directly behind the aperture and angled
to received the focused light from the respective collection angle as
defined by the mask in from of the receiving lens.
A pre-amplifier module will be designed to interface the received
signals with an IBM Digitizer Board that has been designed and con-
structed at Spectron. This digitizer board will replace the LeCroy
Digitizer used during the previously reported experiments, and will be
physically located inside the IBM personal computer. This will greatly
reduce the cost and size of the electronic hardware used to collect,
store, and analyze the data.
-26- 86-2466-05/44
-------
7.0 CONCLUSION
It has now been shown that the laser sheet nephelometer used in
the previously reported experiments can be designed into a small rugged
system using solid state components. The electronic hardware can also
be reduced in size and the cost can be decreased, especially in the case
of large quantities. These proposed advantages can put the proven
nephelometer technique into the environments required by the Environ-
mental Protection Agency.
-27- 86-2466-05/44
-------
SPECTRQN
DEVELOPMENT
LABORATORIES, INC.
33O3 HARBOR BLVD.. SUITE G-3. COSTA MESA. CA 92636-1 579
86-7730-03/44
86-2446-05
3 April 1986
Mr. Walter H. Preston
EPA
RD-675
Washington, DC 20460
Dear Mr. Preston:
Enclosed please find two (2) copies of the final report entitled, "An
Optical Particle and Flux Monitor for Stack. Emissions".
If you have any questions or comments, please contact me.
Sincerely,
SPECTRON DEVELOPMENT LABORATORIES, INC.
Leader,
Particle Characterization Group
/mk
Enclosure>-—Two copies of SDL No. 86-24lT6-<15 (Final Report)
*ec: Dana G. Lloyd, Contracting Officer (2) copies
EPA Library (1) copy
------- |