|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
REAL-TIME PARTICLE DETECTION USING SUB-THRESHOLD
LASER INDUCED BREAKDOWN DETECTION
WILLIAM PAUL MASON
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
William Paul Mason
For Constantine Panageotes:
Fasooli, fasooli, ghiomeeze tor sakooli.
(A bean, a bean, it fills the bag.)
For Jane Woodward Mason and family:
Somewhere ages and ages hence,
Two roads diverged in a wood and I,
I chose the one less traveled by,
And that has made all the difference.
I wish to thank my colleagues in the lab and particularly Ben Smith, Nicolo Omenetto, and Ron
Whiddon for many stimulating conversations, for the patience of oak, and for sharing insight into
the language of nature.
There are more things in heaven and earth, Horatio, than are dreamt of in your philosophies.
Shakespeare, Hamlet, Act I. Scene V
Special thanks go to Ryan Mohney and Dan Shelby for long hours in the lab and valuable
discussions. Thanks to Sue, Sal, and Lily for strong prayers.
And God saw the light, that it was good: and God divided the light from the darkness.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .............. ...............4.....
LIST OF TABLES ............. ...... ...............6...
LI ST OF FIGURE S .............. ...............7.....
LI ST OF AB BREVIAT IONS ............. ...... .__ ...............9...
AB S TRAC T ............._. .......... ..............._ 1 1..
1 INTRODUCTION ................. ...............12.......... ......
2 PARTICLE DETECTION............... ...............2
3 EXPERIMENTAL SETUP AND DEVELOPMENT .............. ...............38....
4 RE SULT S .............. ...............52....
5 CONCLU SION................ ..............5
APPENDIX LASER CHARACTERISTICS............... ............5
LIST OF REFERENCES ................. ...............60........... ....
BIOGRAPHICAL SKETCH .............. ...............64....
LIST OF TABLES
1-1 Comparison of particle counting techniques. ................ ............... ......... ...._..23
2-1 Particle mode versus diameter. ............. ...............30.....
2-2 Particle composition versus diameter. ............. ...............30.....
LIST OF FIGURES
1-1 Laser energy diagram............... ...............23
1-2 Passive Q-switching............... ..............2
1-3 Active mode locking ................. ...............24................
1-4 Cavity dumping............... ...............24
1-5 Chirped pulse amplification............... .............2
1-6 Rayleigh and Raman scatter ................. ...............25........... ...
1-7 Multiphoton excitation............... ...............2
1-8 Cascade electron ionization .............. ...............26....
2-1 Thre should irradi ance versus pulse wi dth ................. ...............35...........
2-2 Size range of aerosol physics. .............. ...............36....
2-3 Scanning mobility particle sizer ................. ...............37........... ...
2-4 Tapered element oscillating microbalance............... ..............3
2-5 Bernoulli effect on particle focusing............... ...............37
3-1 Laser and detection setup............... ...............46.
3-2 Effect of the HEPA filter. ................ ...............46....... ....
3-3 Note baseline shift. ................. ...............47....... ....
3-4 Drying tower and overall experimental layout. ............. ...............48.....
3-5 Electronics cart. ............ ............ ...............48...
3-6 Target chamber and PMT. ............. ...............49.....
3-7 Comparison of laser pulse and emission lifetime. ............. ...............49.....
3-8 Oscilloscope trace of laser pulse. ................. ...............50.___ ...
3-9 Illustration of near single-mode operation. .....__.....___ .........._ ................50
3-10 Diagram to accompany scatter calculation, using a nominal 500 nm particle close to
the PM T. ............. ...............51.....
4-1 03 November outdoor air background .............. ...............53...._._._....
4-2 03 November outdoor air with HEPA filter ................. ...............53..............
4-3 50 nm Gelman filter. .............. ...............54....
4-4 20 nm Gelman filter. .............. ...............54....
4-5 Short tube background. ............. ...... ...............55..
4-6 Long tube background. .............. ...............55....
4-7 24-hour time series. ............_...... ...............56..
4-8 24-hour time series. ............_...... ...............56..
4-9 24-hour time series. ............_...... ...............57..
4-10 40-hour time series. ............_...... ...............57..
LIST OF ABBREVIATIONS
a Particle radius
A, Transition probability
c From Latin circa, meaning nearby or approximately
c Speed of light, 3 x 10s meters / second
C Constant incorporating partition function Q(T)
cw Continuous Wave
d Beam diameter
D Unfocused laser diameter or particle diffusion coefficient
Do Original droplet diameter
DIAL DIfferential Absorption LADAR
Ei Energy of upper level
f Lens focal length
gi Statistical weight of upper level
h Planck' s constant, 6.6 x 10-34 kg m2 / S
HEPA High Efficiency Particulate-Air
I, Intensity of LIBS spectral line
Io Incident intensity.
K Evaporation constant
L Internal diameter of tubing (4 mm) or resonator cavity length
h Laser wavelength or mean free path (m)
hij Transition wavelength
LADAR LAser Detection And Ranging
LPM Liters Per Minute
LTE Local Thermodynamic Equilibrium
me Electron rest mass
M Molecular species
CL Dynamic fluid viscosity or Dipole moment
MASER Microwave Amplification by Stimulated Emission of Radiation
n Number of modes
No Critical electron density for LTE
n, Number of particles at large distance
ns Nanosecond (10-9 Second)
p Fluid density (1.168 kg/m3 for air at STP)
pP Particle density
pm Vapor density at r,.
pw Power (watts)
pdf Probability distribution function
ps Picosecond (10-12 Second)
r Focal spot radius
o Absorption cross section (m2) Or particle diameter (m)
To Critical temperature
AT Pulse width
trt Round trip travel time in laser resonator
us Mean fluid velocity = 13.26 m/s in these experiments
v Kinematic fluid viscosity
Avesr Free spectral range
Z Degree of ionization
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
REAL-TIME PARTICLE DETECTION USING SUB-THRESHOLD
LASER INDUCED BREAKDOWN SPECTROSCOPY
William Paul Mason
Chair: Nicolo Omenetto
Major Department: Chemistry
Ambient aerosols play an important role in a variety of processes ranging from
semiconductor fabrication and industrial emission monitoring to chemical warfare and the global
climate. Though interest in particle detection is not new, advancing technology provides
heretofore unavailable methods of particle detection. This work lays a historical foundation for
optical development, compares related particle detection techniques, describes related work, and
exercises the potential of the PowerChip laser in real time aerosol monitoring. Following the
work of Kwabena Amponsah-Manager and David Hahn, this effort relates the versatility and
robustness of the PowerChip laser with test chamber and electronics for real-time aerosol
monitoring. For particles with finite absorption at 1.06 Cpm, the PowerChip laser shows itself a
stable and reliable laser source. With engineering development such as weatherization and
battery power, this instrument could become field portable and remotely operable.
The inflationary growth of lasers and optical spectroscopy comes only after millennia of
development. The first optical element officially recognized is the Assyrian Layard/Nimrud
lens, dated to 700 BC [BBC, 1999]. However, Robert Temple suggests that lenses existed as
early as 3300 BC in Egypt, evidenced by microscopic carving on a knife handle, and in 2500 BC
by crystal lenses found in the facial wrappings of a mummy [Temple, 2000]. Euclid (325-265
BC) produced the first known writings on optics [Euclid], and both the Greeks and Romans used
crystal and water-fi11ed glass spheres as lenses [Pliny]. The triumph of Archimedes against the
armies of Marcellus is oft-quoted as the first use of optics [Djiksterhuis, 1987]. The veracity of
the claim has been debated for centuries, but Aristophanes (c. 448-c. 385 BC) does mention the
use of a burning glass to start fires in his play "The Clouds" [Aristophanes, 423 BC]. Ptolemy
(AD c. 90-c. 168) measured the changes in the path of light from air to water, air to glass, and
water to glass [Hecht, 2001]. Ibn al-Haitlam, known in the West as Alhazen (965-1040),
developed the first comprehensive alternative to Greek theory. He knew that light traveled in
rays composed of colors and that bodies do not emit visible light, the eyes only receive reflected
light, unless from a source such as a lamp or the sun. Further, he realized that light had a large
but finite velocity and that refraction occurs because light has different velocities in different
media [Lindberg, 1976]. The crystal lenses of 11Ith century Sweden were comparable in quality
to aspheric lenses of the 1950s [Schmidt, et al., 1999]. Rene Descartes developed a theory of
light as a wave traveling through the plenum, a forerunner to the luminiferous aether proposed by
Robert Hooke in 1678. Isaac Newton championed the particle nature of light, but allowed that
particles of light could create waves in the aether to explain diffraction (having been recently
demonstrated by Francesco Grimaldi and later by Thomas Young). In 1845, Michael Faraday
found that the polarization of a beam of light would change in a magnetic field as it passed
through a polarizing medium. In 1847, he proposed that light was an electromagnetic vibration
requiring no medium of propagation. Influenced by Faraday, James Clerk Maxwell ultimately
derived the equations describing electromagnetic waves [Maxwell, 1865]. However, the failure
of the Michelson-Morley experiment to detect the aether, and the inability to explain blackbody
radiation, Einstein's photoelectric effect, and the constant speed of light regardless of reference
frame remained troublesome issues for the wave theory of light. Max Planck in 1900 resolved
the blackbody problem by proposing the quantization of energy, giving credence to the notion of
light as particles, called photons. The packets of discrete energy were called quanta. Thus was
born quantum mechanics. Einstein explained the photoelectric effect by suggesting that the
energy of ej ected electrons was a function of the wavelength of incoming light, while the number
of electrons ej ected was a function of the incoming photon flux (photons per unit time and area).
The invariance of the speed of light was also solved by Einstein with his Special Theory of
Relativity. From the time of Galileo, velocities were considered relative to the velocity of the
observer. Einstein' s theory suggested that this was not truly the case [Thornton and Rex, 2002].
Einstein also derived the so-called "Einstein coefficients" for stimulated absorption and
emission, the first indication that coherent emission may be possible. In short, if a population
inversion can develop with sufficient pumping, coherent radiation will result (Figure 1).
Denoting by subscripts one and two the ground state and first electronic excited state,
respectively, the Einstein coefficients describe both stimulated (B) and spontaneous (A) emission
and absorption. The overall energy balance for a lasing system in a steady state is given by the
change in number of excited atoms through time:
= A,,N, + B,,p(v)N, B,, p(v)N, AN (1.1)
(8mb~ v) 1
p(V) = (1.2)
However, Al2N1 can be neglected by the Boltzmann distribution, assuming the laser is
operating at room temperature. Once the laser achieves steady state (less than one second), the
number of atoms brought to the excited state by pumping is balanced by the de-excitation of
atoms through both spontaneous and stimulated emission. Einstein's derivation of stimulated
emission, the B21 term, is what initially gave rise to the idea of a laser.
Further development by Nikolay Bosov and Alexander Prokhov laid the theoretical
framework to build a device for coherent production of microwaves, now known as the MASER.
Charles Townes, J. P. Gordon, and H. J. Zeiger built the first maser at Columbia University in
1953. Townes and Arthur Shawlow went on to describe the theoretical basis for the LASER
(using visible light instead of microwaves). Theodore Maiman built the first laser in 1960
[Maiman, 1960]. The advantages of lasers over other sources are manifold, including
monochromaticity, coherence, short pulse length, and extremely high intensities. A host of laser
types exist, from CO2 gaS lasers occupying entire buildings to solid state laser pointers, each with
concomitant advantages and disadvantages. Lasers may be broadly defined as continuous wave
(cw) or pulsed. While cw lasers are useful for applications such as welding and have been
developed to the megawatt level, they are large and costly to operate. Pulsed lasers offer the
advantage of short pulsewidths (femtoseconds) and high pulse energies gigawattss). The
femtosecond regime offers a new class of physics under active investigation by many groups.
Unfortunately, femtosecond lasers remain prohibitively expensive. The JDS Uniphase (now
Teem Photonics) PowerChip laser, a 500 ps solid state system, offers a compact, rugged, and
cost effective compromise.
The generation of short pulses is done primarily by three methods: Mode locking, Q-
switching, and chirped pulse amplification. Mode locking was the first technique developed and
takes advantage of the relation:
27r 1 1
C Av B bnancidth
Active and passive mode-locking methods are possible, such as electro-optic modulation
(active) or saturable absorbers (passive). Consider the modulator as a weak shutter. By timing it
to coincide with the round trip time of the cavity,
r = 2L(1.4)
A standing wave develops in the cavity. One packet of photons will bounce back and forth
in the resonator, emitting regular pulses and recharged by the pump. The round trip time of
flight in the laser cavity determines the inter-pulse separation.
The pulse duration At obeys the following relation:
At = Av- (1.5)
for Av the gain bandwidth. For a laser with n output modes,
L = (1.6)
Av = (1.7)
One obtains short pulses by decreasing cavity length or increasing the number of output
modes oscillating in phase. Thus the PowerChip has very short pulses because the cavity length
is short the round trip time of flight in the laser cavity determines the inter-pulse separation.
Since the number of modes that can oscillate depends on the Doppler width of the transition and
the cavity length, optical modulators inside the resonator can cause active mode locking.
Saturable absorbers achieve passive mode locking, as with the PowerChip. To produce more
energetic pulses, Q-switching is used to contain the intensity of emitted photons. Here the
energy in the resonator builds up to a threshold determined by optical switches or nonlinear
crystals. However, active Q-switching requires significant electrical power and is difficult in
practice. In passive Q-switching, the pulse repetition frequency can be changed simply by
changing the pump power, which changes the amount of time needed to reach a threshold in the
passive switching medium.
Passive mode-locking is possible using a Kerr lens that absorbs low intensity light while
passing high intensity transient pulses, leading to mode-locking.
Passive Q-switching depends on the saturable absorber becoming transmissive at a certain
threshold of photon intensity and dumping the photons in one large pulse. The setup is the same
as for active Q-switching, sans drive electronics (Figure 2).
Electro-optic devices can also be used to "dump" the cavity all at once, a form of Q-
switching. For example, Pockels cells will change the path of light in response to a change in
applied voltage across certain faces of the nonlinear crystal (e.g. potassium dihydrogen phthalate,
or KDP). Thus, the population inversion can be built up and then dumped at will (Figure 3).
Active Q-switching is brought about by controlling a saturable absorber, which controls
the quality of the resonator and therefore the transmissivity of the resonator. This change in
transmittivity is tantamount to a change in the quality of the resonator, hence the name Q
(Quality) -switching [McClung and Hellwarth, 1962]. The pump creates a population inversion
with only a small number of photons circulating. When the signal is given, the saturable
absorber becomes transparent and the excited atoms relax, falling into phase with the small beam
of lasing photons to create a giant pulse that exits the resonator all at once (Figure 4).
Chirped pulse amplification is common with femtosecond lasers and essentially stretches
the pulse in time so that it can be amplified without damaging the optical system. The amplified
pulse is then recompressed to exit the resonator (Figure 5). Once the pulse leaves the laser,
several processes can take place. Overall, they fall into three categories: absorption, scatter, and
transmission. The normalized energy distribution equals unity: absorption + scatter +
transmission = 1. Working backwards, transmission is the trivial case. The laser must have a
backstop of some sort to remove excess photons in the event that an absorbing or scattering body
is not present.
Scattering represents a more complicated picture. Perhaps the most well known form of
scattering is Rayleigh, in which photons are elastically scattered in preferential directions as a
function of frequency to the fourth power, hence blue light is scattered moreso than
red I~aylerg o~ (-) (1.8)
The primary direction of scatter is normal to the incident path, thus the sky overhead
appears blue while sunrise and sunset appear red. In Rayleigh scattering, the atom is excited to a
virtual state lower in energy than the first excited electronic state and rapidly de-excites, emitting
an identical photon in a preferentially radial direction. Rayleigh scattering is dominant in
particles of diameter less than or equal to the incident wavelength (Figure 6).
Raman scattering is the inelastic counterpart to Rayleigh scattering and also involves
excitation to virtual states. Since the transition probabilities are much smaller than Rayleigh
scatter, Raman has very low amplitude. Despite this low intensity, Raman scattering has many
important applications such as non-destructive artifact testing.
Now Mie scattering applies primarily to particles of diameter equal to or larger than the
incident wavelength and results in preferential scatter along the direction of transmission. Mie
scattering is not highly wavelength dependent, so scattered light appears white, as in clouds and
So far the interactions described have dealt with uncharged particles. Consider now a
photon interacting with a charged particle, say, an electron. This interaction is defined by
Compton scatter, which does not typically occur with visible wavelength photons because their
energy is too low to overcome the atomic binding energy. However, X-ray photons have plenty
of energy and can lose energy to electrons:
p = (1.9)
A1 h (1.10)
A = h (1- csB), (1.11)
for 6 = radiation scattering angle. By considering wave theory, AL can arise due to the
Doppler effect [Ditchburn, 1991].
Thomson scatter involves the interaction between a photon and a free charged particle,
though only in the plane of polarization of the incident photon. The magnitude of the oscillation
varies as (cos a), where a is the angle between the incident light and the observer. Such
scattering can give rise to a polarization effect.
Brillouin scatter occurs when light changes its vector due to density changes in its path.
Such density changes can arise from acoustic modes (phonons), temperature gradients, or
pressure gradients. Brillouin scattering occurs in a Pockels cell when using acoustic shutter
frequencies to produce Q-switching.
Photoacoustic scatter is the process wherein light strikes a surface and creates sound
waves. Though first noted by Alexander Graham Bell , it was not developed until the
1970s [Rosenewaig and Gersho, 1973]. In essence, the photon source is modulated at an
acoustic frequency, say, 1 k
transfer their energy to the phonons, resulting in acoustic frequency signals.
LADAR is a comparatively new but burgeoning field. Aerosols at large distances (km)
can be interrogated by scatter and absorption measurements. Information about particle
concentration and composition becomes available upon comparison of retro-reflected scatter at
one wavelength versus a different wavelength, a technique known as DIfferential Absorption
LADAR or DIAL. While standoff analysis of aerosols is important and continues to grow as a
field, interesting physical processes arise when the laser energy is increased to the point where
Plasma composes 99% of the observable universe. Methods for analyzing plasma
emission are well characterized, though plasma spectroscopy is still subj ect to significant
background in many situations. Given this, consider now the application of lasers to particle
detection by plasma formation, which amounts to looking at the final term, absorption. The basis
for LIBS and the present particle detection scheme rests on absorption of laser photons into
particles to cause their ionization and emission.
Two years after the first laser was built, F. Brench and L. Cross proposed the theory of
LBES [Brench and Cross, 1962]. In 1967, Moonke and Moenke-Blankenburg built the first LIBS
instrument [Cremers and Radziemski, 1989]. Though the complete process of LIBS is still not
fully understood, great advances in its application as an atomization/ionization source have been
Consider the absorption of photons into a particle. Classically, one may imagine photons
of light striking the surface of the particle like bullets hitting a target. The larger the target, the
more likely the bullets are to strike. The size of the target is quantified by the absorption cross
section o. In practice, this is related to but not necessarily the same as the geometrical cross
section of the particle. The absorption cross section in essence describes the likelihood of
interaction with a photon of given wavelength. A more meaningful analogy is available in terms
of resonance. The photon interacts with the outer shell electrons of the atom. The electrons can
be seen as point masses vibrating on springs with spring constant k a measure of the strength of
the electron's binding energy. Loosely bound electrons will have a small k and will interact with
relatively low energy photons. Depending on the energy and number of photons, the electron
will excite to a higher energy level and then relax through fluorescence, phosphorescence, or
collisional de-excitation. In multiphoton excitation, complete ionization is possible given
sufficient photon flux (Figure 7). Since the density of free electrons in most materials at STP is
negligibly small, initiation of cascade ionization requires some sort of catalyst such as
multiphoton excitation in a laser.
Multiphoton ionization: nhy + M M' + e- (1.12)
Once free, the electron is accelerated by the electric Hield of the laser, leading to cascade
ionization wherein it collides with an atom or molecule and knocks loose another electron, both
of which then accelerate in the electric Hield and repeat the process, forming a geometrically
growing electron cascade [Radziemski and Cremers, 1989] (Figure 8).
Again, for LBES, photons interact with the electrons of the material. The electrons are
excited to high temperatures in femtoseconds but transfer the energy to the phonon lattice,
resulting in a shockwave and explosive removal of material in a plasma state. In the ns-ps
regime, the laser pulse continues to excite the ej ected material as it leaves the surface, heating
and ionizing the ej ecta to a plasma. This forms an optically opaque plasma with a temperature in
excess of 10,000 K. The plasma temperature can be calculated by solving the following equation
for temperature [R. Harmon, et al., 2005]:
I ekB': (1.13)
For ultrashort pulse lasers (< 1 ps) interacting with a solid surface, all the energy is
deposited at once into the electron lattice, which transfers the energy to the phonon lattice,
resulting in explosive removal of material with virtually no melting, though some researchers
have found extensive ionization [Martin, et al., 2002] (Figure 9).
After about 1 Cps the shock wave decouples from the plasma, leaving the plasma in local
thermodynamic equilibrium [Zeng, et al. 2006]. Thermodynamic equilibrium is defined as a
zero gradient for all intensive properties of the system (temperature, chemical activity, pressure,
etc.). For a gas, this is tantamount to having a specific Maxwell-Boltzmann distribution. Such a
condition is virtually impossible to achieve in LIBS plasma it is a non-equilibrium
phenomenon. However, the approximation suffices for many instances. LTE implies that,
though the system parameters vary across space and time, they vary slowly enough to permit the
assumption of thermodynamic equilibrium about any given point instantaneously.
Thus, the 58 pIJ of the PowerChip may or may not allow LTE. One may use the Griem
criterion to determine if LTE exists [Yueh et al., 2000]:
No~lcm ) >> 30,545x10 I[To- /(K 7n) Z (1.14)
Given a plasma in LTE, the problem of spatial measurement arises. Though the present
effort does not involve LIB S per se, an understanding of the processes at work is instructive.
When an intense ultrashort laser field propagates inside a dielectric medium, it induces a strong
polarization field and high density of electrons and holes. This space-time dependent problem is
intricate because it involves nonlinear effects such as multiphoton excitation, free carrier
absorption, photoemission, electron-phonon interaction, exciton generation, and carrier-carrier
interaction, all in the presence of a high intensity field [Audebert, et al., 1994]. Additionally for
natural substances, the situation is further complicated by inhomogeneous sample composition
and surface irregularities [Harmon, et al., 2005]. Variation in composition manifests as variation
in laser-target coupling convolved with surface roughness variation.
Given such difficulties, why bother with LIBS at all? Though LBES is often touted for
little or no sample prep, caution must be exercised in some cases. For example, inhomogeneous
matrices can present spectral complexity and create difficulty in interpretion. Further, heavy
surface contamination such as grease or dirt can cause wide variation in signal intensities and
interference. Surface films and surface roughness can also create skewed results. Bearing these
caveats in mind, particle counting and the natural extension to LBES offer many advantages such
as light weight, solid state electronics, no vacuum requirement, and real-time analysis.
Comparison to other techniques shows the usefulness of LIBS as summarized in Table 1:
Armed now with a conceptual understanding of plasma formation and LIBS, the properties
of particles may be addressed. For small particles, some transmission and scatter may occur,
depending on the shape and composition of the particle. For silicon dioxide (SiO2), We aSSume
that transmission and reflection are negligible. Because the pulse duration is so long (500 ps)
relative to the time it takes to ej ect material from the bulk (femtoseconds), the initial part of the
pulse excites electrons and breaks up the particle into clouds of molecules which continue to be
irradiated by the later part of the pulse to form a plasma. The emission from this plasma,
Bremsstrahlung, fluorescence, and recombination, can be used for simple particle detection by
the photomultiplier tube. The photomultiplier tube is composed of a material that emits electrons
when struck by photons of sufficient energy. In the case of the Hamamatsu R 647, the material
responds to photons with wavelengths between 300 and 650 nm. The response is controlled in
part by the applied voltage, a sensitivity selector of sorts. The price of high sensitivity is an
increase in false hits from, for example, cosmic particles or noise fluctuations that are amplified
by the high voltage required for sensitive measurements.
Particle detection is desired in a host of applications, from semiconductor fabrication to
chemical/biological warfare agent detection and global climate modeling. The following chapter
details particle characterization and detection.
Table 1-1: Comparison of Particle Counting Techniques.
Fast Intersystem Crossing (Singlet to Triplet)
Fast Intersystem Crossing
Singlet Excited State
Singlet Ground State
Figure 1-1: Laser Energy D
Gain Medium Saturable Absorber
Figurel-2: Passive Q-Switching
Figure 1-3: Active Mode locking
Figure 1-4: Cavity Dumping
Figure 1-5: Chirped Pulse Amplification
Image Credit: http ://www.nsu.ru/psj/lector/lotov/terawatt/cp~i
Rayleigh and Raman Scatter.
""""""' Virtual State
Figure 1-7: Multiphoton Excitation
Figure 1-8: Cascade electron ionization
Pulse strikes surface
Brehmsstrahlung emission LIBS (> 1 Cps)
Figure 1-9: Laser-material interaction.
The term aerosol derives from the term for hydrosol, meaning a solid colloidd) suspended
in liquid (solution). Aerosols are important to many aspects of life, including health, global
climate, and military weapons. Epidemiological studies found associations between particulate
air pollution and human health [Schwartz, et al., 1996]. People for some time argued that air
pollution only sped up the inevitable, killing only those who would soon die anyway. However,
though mortality increases as air pollution increases, it is not followed by a deficit when
pollution decreases, implying that pollution not only harvests from the vulnerable pool, but
recruits new people into the pool. [Zanobetti, et al., 2002].
Aerosols also have an impact on the environment. Their primary effect is to alter the
scatter and absorption of solar radiation, leading to either warming or cooling depending on the
fraction scattered versus absorbed. The secondary effect manifests by altering the scattering
properties and longevity of clouds [Penner, et al., 2001]. Without aerosols in the atmosphere,
very few clouds would form [Zalabsky, 1974]. As aerosol number increases in a cloud, water in
the cloud is spread over many more droplets, each of which is proportionally smaller. Clouds
with smaller droplets reflect more light and last longer, since it takes more time for droplets to
coalesce and fall.
According to Raes, et al. , primary particles can be emitted directly into the
atmosphere as particles (primary process) or formed in the atmosphere from gas-to-particle
conversion (secondary process). Atmospheric aerosols range in size from a few nm to Cpm in
diameter. Once airborne, particles evolve in size and composition through condensation,
evaporation, coagulation, chemical reaction, or activation within supersaturated water vapor to
form cloud and fog droplets. Particles smaller than one Cpm range from 10 10,000 cm-3, while
particles with diameters greater than 1 Cpm are typically < 10 cm-3
A primary aerosol is emitted into the atmosphere as a particle, whereas secondary aerosols
are formed in the atmosphere by gas+ particle conversion [Raes, et al., 2000].
Particle formation can be categorized by diameter (cp):
cp > 1 Cpm = primary formation
cp < 1 Cpm = secondary formation
Strong overlap exists for particles of diameter 0. 1 1.0 pm.
Combustion soot is typically 5 20 nm, but coagulates rapidly to form fractal aggregates
which collapse to more stable structures of tens of nm due to the capillary forces of condensing
vapors. Shah et al.  found that lubricating oil was a primary fraction in diesel emission,
and increased while the engine was accelerating, versus cruising.
The Kelvin effect plays an important role in particle formation. The equilibrium vapor
pressure over a spherical particle increases with decreasing radius of curvature; hence
equilibrium vapor pressure above molecular clusters formed by random collisions is much larger
than that above a film or flat surface. Consequently, molecular clusters tend to evaporate. Small
particles (<1 Cpm) diffuse to the Earth's surface, a process that becomes less efficient with
increasing cp. For 0. 1 < cp < 1 Cpm, dry removal is very slow, so these particles tend to accumulate
in the atmosphere. They are removed mostly by cloud activation and precipitation [Willeke and
Particle behavior and composition are weak functions of the particle diameter (Tables 2
Table 2-1: Particle mode versus diameter.
Diameter (um) Mode
cp>1 um Coarse
Table 2-2: Particle composition versus diameter.
Diameter (um) Composition
0.1<(p<2.0 Sulfate, nitrate, heavy metal
(p>2 Geologic material, pollen
To develop a new particle detector, consideration of previous devices is important.
Important work on the detection of aerosols began with Aitken in the 19th century [Aitken,
1923], who determined that most atmospheric aerosols were less than 100 nm in diameter and
ranged from hundreds to tens of millions per mL depending on the cleanliness of the air.
Interestingly, the Wilson cloud chamber was developed as a result of Wilson being moved by the
sighting of a Brocken specter while working at the meteorological observatory atop Ben Nevis in
Scotland. So struck was he that he studied cloud formation and condensation in the laboratory.
The result was the cloud chamber, which is also one of the most sensitive particle counters for
aerosol measurements. Particle detection methods encompass a range of responses:
Aerodynamic particle sizer measures the velocity of particles in accelerating air
flow using two laser beams and scatter detectors at various angles.
Photoacoustic spectroscopy chopped laser illuminates ambient air. Particles
absorb energy from the beam and transfer it as heat to surrounding air. The
expansion of heated gas produces a sound wave at the same frequency as the
chopper. This acoustic signal is detected by microphone and is proportional to the
amount of light absorbed.
Electrical aerosol analyzer collects particles according to size dependent mobility
in electric Hield, then detected by deposition of charge on an electrometer.
Differential mobility particle sizer classifies particles by their mobility in an
electrical Hield and counts them with a condensation nuclei counter in a range of
Scanning mobility particle sizer a complex version of the Differential Mobility
Analyzer (Figure 11), including a radioactive ionization source and Condensation
Particle Counter. Because the Condensation Nuclei Counter (CNC) cannot classify
particles by size, it is combined with the DMA to give both particle size and
number. One drawback of the SMPS is that it can take up to 300 s to obtain size
distributions, since the particles need time to form a hydration shell in order to be
detected in the chamber. Further, CNCs are overwhelmed by particle
concentrations greater than class 1000 environments [Particle Measurement
Systems website, 2006].
Electrical low pressure impactor offers real time size distribution and
concentration measurement from 30 nm to 10 pm. The electric current carried by
charged particles into each impactor stage is measured by a sensitive electrometer.
One anticipates the difficulty of measuring charged particles at low pressure
(requiring some measure of pumping), and interference effects from electronic
noise. However, it has the advantage of being able to measure rapid changes in
both particle size and concentration.
Tapered element oscillating microbalance operates by changing the frequency
of oscillation as mass accumulates on the cantilever (Figure 11). As particles
accumulate the frequency co changes as co, = kI- The balance lasts for about 3
weeks, with a V/2 hour equilibration time before data can be taken.
The classic methods of atomic absorption/emission spectroscopy are well-characterized,
but suffer heavy background emission and comparatively low sensitivity. Scattering, absorption,
and emission techniques such as Rayleigh, Raman, Fourier-Transform Infrared Absorption,
LBES, LAser Detection And Ranging (LADAR), fi1ter impaction, and mass spectrometry offer a
host of techniques for particle detection. While each technique has some advantage and some
disadvantage, LBES offers a few important advantages over other techniques. Mass spectrometry
provides high resolution and good limits of detection, but requires the use of vacuum pumps,
adding to the cost and bulk of the system, and experiences peak broadening due to excess kinetic
energy [Tolocka, 2004]. Light scattering techniques are effective only to the point at which the
particle diameter equals the wavelength. For particles smaller than the wavelength used (below
~300 nm), such techniques are less reliable [Maynard, 2000]. Techniques like LADAR offer
remote detection, but have poor sensitivity and minimum detectable particle sizes (~300 nm).
CNCs are sensitive, but require periodic refilling with alcohol and are easily saturated in dirty
environments. LIBS provides a good detectable size range and good portability, for the price of
sensitivity. However, particle focusing could ameliorate this problem [Wu, 2006].
Research has shown a bimodal particle distribution in the atmosphere, while other research
has shown that particles in the accumulation region (around 0.1 Cpm in diameter and smaller) are
most harmful to humans. These are also the most difficult to count continuously and the most
difficult to filter. Filtration and detection methods are many and wide ranging, from Raman and
fluorescence to condensation nuclei counters and impaction.
Having considered particle counting, particle transport bears remark. Particle transport
both through the air and through tubing is complex. Physical phenomena such as
thermophoresis, turbulence, and adhesion are factors. Velocity focusing, diffusion, and
Brownian motion also play into the scheme (Figure 12). Brownian motion of particles results
from collision with other particles whose velocity is proportional to the square root of
temperature. Equating kinetic energy and the thermal energy gives a relation between velocity
-my k,T (2.1)
v =~i (2.2)
Because their mass is so small, Brownian particles do not settle from a given volume; they
are kept "afloat" by the thermal motion of the particles around them. In other words, they are
perpetually diffusing at a rate given by:
Thermophoresis is brought about by temperature gradients in a given volume. Higher
temperature will increase the volume between particles, causing cooler particles to move to
cooler regions of the gradient. Note that energy must be added to the system to maintain the
gradient. Similarly, eddy and turbulence focusing arise from pressure gradients, similar to
thermophoresis. In addition to various transport phenomena, particles may adsorb to surfaces by
the following two processes:
Electrostriction may arise between particles with a charge or a strong permanent dipole
moment. In liquids, such charges will be solvated, but in the gas phase electrostriction can be a
significant factor. Electrostriction between two or more particles, usually of opposite sign, is a
phenomenon known as accretion.
Simple friction can play a role, such as in HEPA filters, where particles are mechanically
trapped by small gaps in a medium. Diffusion and adhesion coexist in a dynamic equilibrium,
suggesting that a sudden shift in some relevant parameter (e.g. temperature or pressure) could
shift the particle equilibrium and result in a "pulse" of free particles or a sudden shift in average
particle diameter. Also, particle concentration decreases with increasing tube length. The rate of
decrease depends on several factors, including type of tubing [Willeke and Baron, 1993].
Carranza, et al.  found that particle transport efficiency was greater than 95% for particles
from 0.1 to ~1.5 Cpm in diameter for their system.
Deliquescence, a sharp rise in liquid water content at ~ 55% relative humidity, is a result
of the hygroscopicity of particles and is relevant for many salts such as calcium chloride and
magnesium chloride. Because the particles are strongly hygroscopic, under high humidity they
can absorb enough moisture to dissolve themselves. This could be a factor in situations where
particle solutions are created and then dried to form aggregates. The ability of the tower to dry
such salts may be questioned.
Clearly a host of physico-chemical processes are at work. Desirous to further develop
particle detection capabilities with the PowerChip laser, several experiments were performed.
Note here that M.D. Cheng  failed to produce meaningful results from the reference
standards while testing the sub-threshold setup similar to the present study. The present work
may add further strengthen the overall effort of particle detection with lasers. Though this effort
recapitulates much of Cheng, Hahn, and Amponsah-Manager' s work, the use of the PowerChip
laser and the discussion of particle distribution (Poisson versus diffusion mediated) may lend
insight into the investigation.
LBES offers some advantages because it can both count particles and characterize them
by constituents. Several particle counting systems also provide spectral analysis (e.g. Hahn
 and Cheng ). Cheng proposed the use of sub-threshold breakdown for aerosol
detection (akin to sub-threshold breakdown in liquids). That is, the laser power is set just below
the breakdown threshold of air. When a particle enters the laser focus, the breakdown threshold
of air is decreased, causing plasma formation. The tacit assumption is made that the absorption
cross section of the particle is greater than that of the background carrier gas. The advantage of
using sub-threshold pulses is a reduction in background noise, since a plasma forms only in the
presence of particles of nontrivial absorption cross-section. Following this line of work, we
sought to further characterize and expand the capabilities of such a system. The next chapter
describes those efforts, including initial development, enhanced chamber setup, aerosol nebulizer
and flow gas, and long term measurements.
Fr-3rTauwr. GP~rf Wy iles slaser-lnd..rt~PMII sbeadown lm:.r19"F
10-9 10-8 10-7 10-6 10-5 10-4
PULSE TIME TIME TO BREAKDOWN (SEC)
Figure 2-1: Threshold irradiance versus pulse width.
001 0.1 10 le LOD
Suctng SkEswi~tinkpearecU.,/ Sloke ITransinc
Charyqp Dilfumn / Combined / Field
on,u. Ave hrolr ~ TB~ Hr*eAd Nont~iad
depostulon Datuden imp ;Youdonadatin
Sunphq D altrnkesu /ArnokMIC onesu
FNtunnon~l Dllffusn D.I LPEU~IrnahnnO necpo
0.01 0.1 LO IQ 100
Fnlure 2-2. Slze ra nsc ohercrool oroucntio urnilticrt imm i nli.d l982)
Figure 2-2: Size range of aerosol physics.
Figure 2-3: Scanning Mobility Particle Sizer.
Figure 2-5: Bernoulli Effect on particle focusing.
Figure 2-4: Tapered Element Oscillating Microbalance.
Gas particles bouncing against a surface at low velocity
Gas particles bouncing against a surface at high velocity
EXPERIMENTAL SETUP AND DEVELOPMENT
Manager  worked on particle detection with the JDS MicroChip laser and nebulized
solutions of simple salts. Particle focusing occurred by nebulizing and desolvating particles,
then passing the air flow through a narrow pipette tip. The stream was interrogated by a 5 k
MicroChip laser at 50 CLJ / pulse. Results were promising but inconclusive. Inspired by the work
of Smith, Hahn, Omenetto, Amponsah-Manager, and Cheng, the present study investigated the
feasibility of the PowerChip laser for ambient aerosol monitoring at 1 k
The PowerChip laser offers the advantage of relatively high repetition rate, short pulse
width, solid state passive Q-switching, air cooling, and short cavity length, yielding essentially
single mode output. Again, all of these properties play well into a field portable design as well
as complimenting the LIBS detection. The PowerChip laser is amenable to particle detection
because its short cavity length allows longitudinal mode spacing greater than the gain bandwidth;
it enj oys virtually single mode operation with no frequency beating (Figure 16).
Bv, 1 31
While Amponsah-Manager  worked mostly with solids, some particle sampling was
done. This work extends his initial investigation. The test chamber (Figure 17) was built and
used in all subsequent experiments. Several modifications were made over time to improve its
performance. Light tightness is critical since the PMT is sensitive and low background is
desirable. Tightness was improved through the addition of a laser beam tube, black electrical
tape around the housing seam, and an aluminum guard ring over the laser entry port. A mirror
placed opposite the PMT (in the bottom of the chamber) reflected light towards the PMT to
increase signal. In the future, an elliptical mirror with one focus at the laser focus and the other
focus on the PMT might prove a better choice. Lastly, the chamber was wrapped in black felt to
further reduce stray room light. Note that a beam expander would improve the focus
characteristics (sharper focus), though this might push the system into breakdown regardless of
whether a particle was present and would decrease the effective plasma volume. Depending on
the desired experimental setup, a variable diameter beam expander could serve as a simple power
control. In the limit of perfect system function, particles below a given diameter (all other
parameters, e.g. absorptivity, being equal) could be made "invisible" to the system. If the
absorption cross section is too small to create a plasma, the particle will pass undetected. Above
some minimum diameter, the cross-section will be large enough to form plasma.
The signal passed through an SR 570 amplifier, boxcar, SR 245 signal processor, and out
to a computer for data storage and display (Figure 16). Initial tests demonstrated the
effectiveness of HEPA filters at reducing the number of events detected in ambient air (Figure
The system measured the emission of light from plasma formed when a particle was struck
by the laser pulse and converted to plasma. The number of photons expected for nominal 500
nm diameter, 2% (weight) silicon dioxide particles is given by:
62g 1000cm3 I
100x10b = 2mg silica in 100 CLL solution.
100g 1L cm3
particle 9 particles
0.002g -= 2.9x10
2.9x109 particles 0.1mL I min = .x0 particles
mL min 60 sec sec
Ambient air traveled through the target chamber by laboratory exhaust flow at roughly 0.2
LPM as measured by a rotometer. If the PMT detected sufficient photons, the signal continued
through preamplifier, boxcar, and postamplifier before readout on a stripchart or storage in
computer memory. Data processing included counting the number of events greater than a
certain threshold. Initially this threshold was set at 3 standard deviations (30) from the mean, but
minor (< 0.05 V) step fluctuations in the background forced a change to the less sensitive but
more trustworthy measure of 0.2 V as threshold (Figure 15). This was well above any baseline
fluctuations. A power supply regulator (SOLA MCR 1000, Mini/Micro Computer Regulator,
catalogue # 63-13-210-05) was installed to eliminate background shifts, but was unsuccessful at
doing so. The threshold remained at 0.2 V.
Computer sampling rate:
1,000shots 3,600 sec
Expect: -24hours = 8.64x107 shots
Found: 3,317 fls-1,0 iles = 4.98x107shots
Ratio = 0.576 ~ 58% duty cycle.
This is very repeatable for all runs, leading to the conclusion that it represents the computer
writing data to memory, during which time it cannot simultaneously take data.
The amount of light scattered by the particle into the PMT we expect to be negligible
compared to the 0.2 V threshold. The following calculation bears this out. Because the particle
diameter (500 nm) is less than the laser wavelength (1,064 nm), Rayleigh scatter dominates.
Assume a 900 cone of emission circling the equator of the particle normal to the laser beam
containing all of the scattered light. Further, assume that approximately one quarter (- sr ) of
the emission is detected by the PMT:
For a 58 1064 nm laser,
v=C = 2.83x10'4 s1
E,, = hv = 6.6x10 3 g-n 2.83x10 's = 1.87x10-19.
6J ph ,,photons
pulse 1.87x10-19 pulSe
Ph Pulses =
~ref lectivity2til attenuation, EfficiencyP~f Gain Angle, re
pulse pt sec
pulse 4 se ),,,~ sec
ExpectedCurrent = 4.9x10 = 2.94x10 A
Expectedyoltage = 2.94x10 'A- 5002 = 14.7pVy
This is well below the 0.2 V threshold for particle hits. Using conservative values for the
reflectivity, attenuation, efficiency, and hit rate, we are confident that we do not detect
significant scattered light.
Now consider the Reynolds and Knudsen numbers regarding flow characteristics:
Reynolds Number = Re U, (3.2)
kg na 0.004n?
R, = 1.168 13.26- = 3480.4
e a kg
m S 1.78x105
The results suggest that the flow is unstable. This finding prompted a change to a longer
drying tube, permitting a lower drying gas flow rate because the particles would have more time
Knudsen Number Kn (3.3)
L (rZ~o PL)
Kn= K -50x10-9n t101,300Pa =1.8x105
1/~(100x10 -9 2)
Therefore continuum mechanics are valid (for Kn < 1), as expected.
Consider now the time to to completely evaporate water:
to [Hahn et al., 2001] (3.4)
Then to 7.14x10 's = 71pUS .
The lower flow rates should also help reduce effects such as velocity and eddy focusing,
plus increase the dwell time within the PMT viewing angle.
The particle dwell time can be calculated as follows:
Dwell time = (Flow rate) (Interaction volume)
-mi -e / 0.2cn??)- (0.2cnt)= 1.5x10- 'SOC
This should provide plenty of time for the detectors, which has a response time of a few
When the composition of particles is considered, the difficulty of matrix effects comes to
bear. Defined as variations in laser-target coupling secondary to variable sample composition
and surface characteristics [Harmon, et al., 2005], matrix effects can yield transients in plasma
emission that could affect the observed hit rate. Indeed, Harmon found that surface roughness is
a primary factor in experiment repeatability. The surfaces of aerosols are known to vary greatly
from spherical to needle-like [Hinds, 1999].
A final consideration is the distribution of particles, taken to follow a Poisson distribution
P(x) = (3.5)
The Poisson distribution assumes no correlations in either space or time, an assumption
which may be called into question by point source emission of particles, e.g. from a vehicle
passing close to the detector. The diffusion equation may be a more appropriate model. The
properties of the diffusion equation are complex, but a brief treatment follows. For a probability
distribution function of a single particle P, use the heat equation:
PT = DAP (3.6)
If the diffusion coefficient D is not constant but depends on P, then one gets the nonlinear
diffusion equation. The random traj ectory of a particle subj ect to the particle diffusion equation
is Brownian motion. To treat it, place a particle in R = 0 at t = 0 and find the pdf associated with
R to be:
P(R, T)= G(R, T) = (4i ~t)e e (4D) (37
For R2 = Rx2 Ry2 Rz2 (3.8)
At t = 0, P(R, T) is singular with a pdf for the particle at R = 0 given by the Dirac delta
function. The solution of the diffusion equation subject to this initial condition is the
Green Function G(R, T) given above. This treatment can be extended to a large number
of particles by a decomposition of Green functions giving the time evolution of the
particles. Such a decomposition can be generalized to any diffusive process like heat
transfer or momentum diffusion, which is the phenomenon at the origin of viscosity in
liquids. If the Poisson distribution does not hold, the diffusion equation could serve as a
new model [Willeke and Baron, 1993].
Having laid a theoretical foundation, a review of the experiments conducted is indicated.
Many experimental series were run to determine the effectiveness of particle detection with the
PowerChip laser. The first was with a HEPA filter on ambient outdoor particles. The hit rate
showed a marked decrease with installation of the filter, suggesting that the detection system was
functioning nominally (Figure 14). A drying tower was assembled and operated using
compressed air (Figure 16).
The time responses of the laser beam and the detector were measured to verify that plasma
and not scattered light was detected. Because the plasma lifetime is long compared to the laser
pulse microsecondss versus picoseconds, respectively), we conclude that plasma, not scatter, is
observed. Further, a calculation of expected scatter shows that it should produce a negligible
signal at best (Figure 18).
2ilf 2 1064x10-9 0.02nt)
Calculation of Beam waist = coo = r .x06 a=27p
58x1-6 xx0m 27x0m2 7w
Photon flux: 'P= _, 500x10 s -1040 photon
zi3h (2.7x10 -6 n 6.6x10-34 k m 3x108
The expected particle hit rate may be estimated as follows:
For 2 % by weight silicon dioxide particles of 500 nm diameter and density = 1.05
2g 1000nzL g
100x10-6L -1.05 =
100g L mL
For the mass of one particle, find the voli
V=4 ar3 4 a 250x10- cm .51
Then find the mass of a single particle:
0.0021 g of silica in 100 CLL of solution.
6.55x10-14 3 -1.05 g
To give the total number of particles in 10 mL:
0.002 g -= 2.9x10' particles
Calculate the flow rate using the nebulization rate:
2.9x10"' particles 0.100nzL 30, particles x0 particles
10nzL min min sec
Given a counting efficiency of 1:106, this amounts to roughly 5 hits per second.
Experiments were performed using undiluted samples and approximately this rate was
observed. A more sensitive detector or higher particle concentrations must be use to see
significant hit rates.
Table 3-1: Flow regimes.
Reynolds number, Re Flow
Figure 3-1: Laser and detection setup.
10 HEPA Fi
0 50 100 150
Figure 3-2: Effect of the HEPA filter.
200 250 300
[Reprinted with permission from Xihong Wu]
Nubulizer 2 LPM II
26 April 2006
0 10000 20000 30000 40000 50000 60000
Figure 3-3: Note Baseline Shift.
1 ii 1 / I I
Figure 3-4: Drying tower and overall experimental layout.
; ; ;iiii----;;--
............ .....:: .:-::::::--;;;;;;;;;;I*-"~
Figure 3-5: Electronics cart.
Figure 3-6: Target chamber and PMT.
typical particle hit
Figure 3-7: Comparison of laser pulse and emission lifetime.
0 02 -
0 00 -
a, -0 02 i
-0 06 -
-0 08 ~
-0 00000010 0000000 00000001 00000002 00000003 00000004 00000005
X Axis Title
Figure 3-8: Oscilloscope trace of laser pulse.
Free spectral range
Figure 3-9: Illustration of near single-mode operation.
Figure 3-10: Diagram to accompany scatter calculation, using a nominal 500 nm particle close
to the PMT.
Experimental results over the past 18 months served to characterize and improve system
performance, find environmental trends, and determine limits of detection for particle size. Note
that the limit of detection for particle concentration is poorly posed because detection with
decreasing concentration simply implies a longer time interval between hits, assuming
cpntinuous flow. The question is valid and crucial for batch applications.
Styrofoam was tested, but no evidence of interaction was found, so silica powder was
used. Long time-series data collection campaigns sought long term environmental effects, as in
Tolacka et al.  who found a bimodal particle distribution over 24 hours due to increased
human activity in the morning and evening. No such trends emerged in this study, excepting a
clear correlation between rainfall and decreased particle counts. Interesting peaks were found,
such as transient spikes one order of magnitude higher than surrounding peaks. No satisfactory
explanation of their origin was found.
The trials did establish the reliability of the PowerChip and associated electronics for
potential field applications to particle monitoring. A test for plasma current (MFP = 40 Cpm) was
unsuccessful. Amponsah-Manager  reported a small current, but did not describe the
setup used. The effect of adhesion in tubing was measurable (Figures 19 and 20), but was
neglected in the present study.
Tests were conducted using Gelmann filters of varying pore size, though no clear
correlations were found. This could be due to dirty filters, fluctuating background signal, or
Number of h ts per 1 mmn as fundticn et time
5 10 15
30 35 40) 45 50]
03 November Outdoor Air Background
Number of Ri1s per 1 min as function of time
10 15 20 25 30 35 40 45 50
.Figure 4-2: 03 November Outdoor Air with HEPA Filter.
rr.Jruber 3I ral, Fer I rrln ii FI.ncart) ofl lne
Figure 4-3: 50 nm Gelman Filter.
20 SeYpember 2008 20 nm Filer
Number of flits per 1 min as fuvnclion o~ftime
Figure 4-4: 20 nm Gelman Filter.
20 September~ 20116 Shod~ Tube Backpound
Nrsaonle o iMpr 1 minas fundlon ofume
Figure 4-5: Short Tube Background.
20~ September 200 Long Tube Backgrundl
Nrsnaerof hiM pr minas fundlon ofume
Figure 4-6: Long Tube Background.
One Week, Day 2
Number of hits per 1 mmn as function of hime
Figure 4-7: 24 hour time series.
One Wieek,. Day 3
Number of hits per 1 min as function of lime
Figure 4-8: 24 hour time series.
I J I 1. 1 I
1000 1200 la
0 200 400 6000 800 1000) 1200 1400 19001
Figure 4-10: 40 hour time series.
One Week, Day 4l
Number of hits per 1 mmn as function of time
Figure 4-9: 24 hour time series.
One Week,. Day 5
Number of hits per 1 mmn as function of time
00 000 800
The PowerChip laser was used to form plasma on ambient aerosol particles which were
detected by a photomultiplier tube and stored to a computer. A variety of subtleties arose in the
development of the system, ranging from making the detector housing light-tight to changing the
drying tower to make the flow properties more laminar. A 5-day sampling campaign produced
no clear environmental cycles and only one correlation; that between rainfall and decreased
particle counts. However, PowerChip features such as high repetition rate, short pulsewidth,
stable output, and nearly single-mode operation render it useful for many applications including
real-time particle monitoring
By taking long time-series data of ambient air and studying the size and concentration
dependence of particle counting, the PowerChip laser shows itself a steady and reliable source
for plasma excitation. The PowerChip would be a strong candidate for continuous aerosol
Future work may include more precise determination of minimum particle size and
improvement in efficiency through, for example, particle focusing [Wu, 2006; Erdmann, et al,
2005] and combination of the laser with an iCCD or other spectrometer to gain spectral
identification along with particle counting. Interesting experiments would include using the
spectrometer to calculate the plasma temperature to correlate the particle size versus breakdown
energy from Weyl . Lastly, particle beam focusing could dramatically improve particle
transport efficiency and, concomitantly, detection efficiency.
JDS Uniphase PowerChip Laser
-500 picosecond pulsewidth
1 kHz repetition rate
Hamamatsu R647 PhotoMultiplier Tube (PMT)
-Diameter = 13 mm
-Wavelength Range = 300 650 nm
-Gain 1.4 x 106
BG 3 8 filter glued to end of PMT to block laser light
~300-650 nm FWHM transmittance
LIST OF REFERENCES
Aitken, J., Collected scientific papers of John Aitken, L. L. D., F. R. S., Ed. C. Knott,
Cambridge University Press, London, 1923.
Alfarra, M., Insights into atmospheric organic aerosols using an
Aerosol~a~ss Spectrometer, PhD. Dissertation, University of Manchester, Manchester,
Amponsah-Manager, K., N. Omenetto, B. Smith, I. Gornushkin, and J. Winefordner, J.
Anal. At. Spectrons., 20 (2005) 544-551.
Amponsah-Manager, K., PhD. Dissertation: M~icroChip Lasers as sources for la~ser-
induced breakdown spectroscopy: Plasma characteristics and analytical
performance, University of Florida, Gainesville, Florida, 2005.
Aristophanes, Clouds, 423 BC. Translation by Ian Johnston, Malaspina University-
College, Nanaimo, BC, Canada, 2006.
Audebert, P., Ph. Daguzan, A. Dos Santos, J. Gauthier, J. Geindre, S. Guizard, G.
Hamoniaux, K. Krastev, P. Martin, G. Petite, A. Antonetti, Phys. Rev. Lett., 73 (14), 03
BBC, Thursday, 01 July 1999, David Whitehouse
Bell, A. G., Phil. Mag., 11 (510), (1881).
Brench, F. and L. Cross, Appl. Spectrosc., 16 (5), (1962).
Carranza, J., B. Fisher, G. Yoder, D. Hahn, Spectrochina. Act. PartPPPPP~~~~~~~PPPPPP B, 56 (2001) 851-864.
Cheng, Meng-Dawn, Talanta, 61 (2003) 127-137.
Cheng, Meng-Dawn. FuelProcessing Technology, 65-66 (2000) 219-229.
Ditchburn, R. W. Light, Dover Publications, New York. 1991.
Djiksterhuis, E. J, Archimedes, Princeton University Press, Princeton, 1987.
Erdmann, N., A. Dell'Acqua, P. Cavalli, C. Gruning, N. Omenetto, JP Putaud, F. Raes,
R. Van Dingenen, Aerosol Science and Technology, 39 (5), May 2005, 377-393.
Einstein, A. Relativity: The Special and General Theory, translated by R. Lawson,
Pi Press, New York, 2005.
Euclid of Alexandria, Optcs D. Joyce, translator, Clark University, Dept. of Math and
Computer Sciences, Copyright 1998, accessed October 2006.
http:.//al eph0.clarku. edu/~dj oyce/j ava/el ements/el em ents.html
Gornushkin, I., K. Amponsah-Manager, B. Smith, N. Omenetto, J. Winefordner, App.
Spectros., 58 (7), July 2004, 762-769.
Hahn, D., J. Carranza, G. Arsenault, H. Johansen, K. Hencken, Rev. Sci. Inst., 72 (9),
Hahn, David W., App. Phys. Lett., 72 (23), 08 June 1998, 2960-2962.
Harmon, R., F. De Lucia, A. Miziolek, K. McNesby, R. Walters, and P. French,
Geochemistry: Exploration, Environment, Analysis; 5 (1), February 2005, 21-28.
Hecht, E. Optics (4th ed.), Pearson Education, 2001.
Hinds, W., Aerosol Technology: Properties, Behavior, and Measurement of Airborne
Particles, 2nd Edition, Wiley-Interscience, New York, January 1999.
Lindberg, D. C., Theories of Vision from al-Kindi to Kepler, University. of Chicago
Press, Chicago, 1976.
Maiman, T., Journal of the Optical Society ofAmerica. 50, 11 (1960) 1134-1134.
Martin, F., R. Mawassi, F. Vidal, I. Gallimberti, D. Comtois, H. Pepin, J. Kieffer, and H.
Mercure, Appl. Spect. 56 (11), 2002.
Maxwell, J. C., Philosophical Transactions of the Royal Society of London 155 (1865) 459-512.
Maynard, A., Phil. Trans. R. Soc. Lond., 358 (2000) 2593-2610.
McClung, F.J. and Hellwarth, R.W. Journal ofAppliedPhysics, 33 (3), (1962).
Particle Measurement Systems,
http://www. pmeasuring. com/support/papers/particlemonitoring/airgldCD8Kw
gCFSVNVAody~r2Lw, 15 October 2006.
Penner, J., D. Hegg, R. Leaitch, Eny. Sci. Tech., 35 (15), August 1, 2001.
Petrucci, G.; Farnsworth, P.; Cavalli, P.; Omenetto, N., Aerosol Science and' Technology,
33, (1-2), 105-121, (2000).
Raes, F., R. Van Dingen, E. Vignati, J. Wilson, J.P. Putaud, J. Seinfeld, P. Adams,
Atmospheric Environment, 34 (2000).
Radziemski, L. and D. Cremers, Laser Induced Plasmas: Physical, Chemical, and
Biological Applications, New York, M. Dekker, 1989.
Rosenewaig, A. and A. Gersho, J. Appl. Phys. 47, 64-69, (1976).
Schmidt, O., K. Wilms, B. Lingebach, Optonsetry and vision science, 76(9), September
Schwartz, J., D. Dockery, L. Neas, Journal of the Air and Waste Management
Association, 46 (10), October 1996, 927-939.
Secundus, Gaius Plinius (23-79), Naturalis Historia, John Bostok and H. T. Riley, Eds,
Book XXXVII, p. 6396 Perseus Digital Library Project. Ed. Gregory R. Crane.
Tufts University. http ://www.perseus.tufts.edu. 30 October 2006.
Shah, S., D. Cocker, Aerosol Science and Technology, 39, (2005) 519 526.
Shawlow, Arthur and Charles Townes, Physical Review, 112 (6), 15 December 1958,
Temple, Robert. The Crystal Sun, Century Books, London, 2000.
Thornton, S. and A. Rex, Modern Physics for Scientists and Engineers, 2nd Ed.,
Thompson Learning, Inc., United States, 2002.
Tolacka, M., D. Lake, M. Johnston, A. Wexler, Atmospheric Environment, 38 (2004) 3263 -
Twomey, S., Atmospheric Aerosols, Elsevier Scientific Publishing Company, New York,
Weyl, G. "Physics of laser-induced breakdown: an update" in L. Radziemski and D.
Cremers, Eds., Laser induced plasmas and applications, M. Dekker, New York, 1989.
Willeke, K. and P. Baron, Aerosol Measurement: Principles, Techniqlues, and
Applications, Van Nostrand Reinhold, New York, 1993.
Wu, Xihong, Theoretical design, construction, and experimental characterization of a versatile
apparatus for detecting individual aerosol particles, PhD. Dissertation, University of
Florida, Gainesville, Florida, 2006.
Yueh, Feng-Yu, J. Singh, H. Zhang, in Encyclopedia of Analytical Chemistry, R. Myers,
ed., Wiley, New York, 2000.
Zalabsky, R. and S. Twomey, Bulletin of the American M~eteorological Society, 55 (6), (1 974)
Zanobetti, A., J. Schwartz, E. Samoli, A. Gryparis, G. Touloumi, R. Atkinson, A. Le
Tertre, J. Bobros, M. Celko, A. Goren, B. Forsberg, P. Michelozzi, D. Rabczenko,
R. Aranguez, and K. Katsouyanni, Epidemiology, 13 (1), January 2002.
Zeng, X., X. Mao, S. Mao, S. Wen, R. Greif, and R. Russo, Applied Physics Letters 88,
William Paul Mason was born in Gardner, Massachusetts on 15 May 1973. The youngest
of five children, William graduated from Oakmont Regional High School in 1991 and obtained a
B.A. in outdoor leadership from Prescott College in 1995. After working as a mountain guide
and EMT, he returned to Northern Arizona University to obtain a B.S. in environmental
chemistry. Upon graduation he commissioned in the U.S. Air Force and served his first
assignment at Kirtland AFB, New Mexico as a scientist. From there, he was selected to obtain a
master' s degree and attended the University of Florida.
When not studying, William enj oys outdoor activities ranging from skydiving to cave
diving, and is a budding musician with bass guitar and violin.