close

Вход

Забыли?

вход по аккаунту

?

Evidence for the microwave effect during hybrid sintering and annealing of ceramics

код для вставкиСкачать
C A V IT Y R IN G D O W N S P E C T R O S C O P Y OF E X C IT E D A R G O N IN A
M IC R O W A V E D IS C H A R G E
A Thesis
Presented to
The Faculty of G raduate Studies
of
The University of Guelph
by
CLAYTON WINSLADE
In partial fulfilment of requirements
for the degree of
Master of Science
September, 2005
© C layton W inslade, 2005
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1*1
Library and
Archives Canada
Bibliotheque et
Archives Canada
Published Heritage
Branch
Direction du
Patrimoine de I'edition
395 W ellington Street
Ottawa ON K1A 0N4
Canada
395, rue W ellington
Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN: 0-494-14608-7
Our file Notre reference
ISBN: 0-494-14608-7
NOTICE:
The author has granted a non­
exclusive license allowing Library
and Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distribute and sell theses
worldwide, for commercial or non­
commercial purposes, in microform,
paper, electronic and/or any other
formats.
AVIS:
L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans
le monde, a des fins commerciales ou autres,
sur support microforme, papier, electronique
et/ou autres formats.
The author retains copyright
ownership and moral rights in
this thesis. Neither the thesis
nor substantial extracts from it
may be printed or otherwise
reproduced without the author's
permission.
L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these.
Ni la these ni des extraits substantiels de
celle-ci ne doivent etre imprimes ou autrement
reproduits sans son autorisation.
In compliance with the Canadian
Privacy Act some supporting
forms may have been removed
from this thesis.
Conformement a la loi canadienne
sur la protection de la vie privee,
quelques formulaires secondaires
ont ete enleves de cette these.
While these forms may be included
in the document page count,
their removal does not represent
any loss of content from the
thesis.
Bien que ces formulaires
aient inclus dans la pagination,
il n'y aura aucun contenu manquant.
i*i
Canada
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
CAVITY R IN G D O W N SPECTRO SCO PY OF EXCITED AR G O N IN A
MICROWAVE DISCHARGE
Clayton Winslade
University of Guelph, 2005
Advisor:
Professor R. L. Brooks
The cavity ringdown absorption spectrum of excited argon in a microwave dis­
charge has been measured between 605nm and 677nm. 73 lines in neutral argon have
been seen, 13 of them not appearing on the NIST spectral database. The 13 lines were
identified by directly referencing the table of atomic energies for argon. As well as
the atomic argon lines and some lines due to atmospheric contaminants, seven strong
broad unidentified features were observed between 634nm and 664nm. A tentative
assignment of these lines to rotationally unresolved, molecularly narrow, vibrational
transitions between unknown E and II states in the argon excimer has been made.
The vibrational spectroscopic constants of the upper state have been calculated (u'f,
= 295 ± 1 1 cm-1 , and u>'ex'e = 24 ± 2 cm-1 ), and are comparable to those of well
known dimer states. It has not been ruled out that the features are transitions from
the 3p54s4p 4 S 3 / 2 state in the negative argon ion to higher lying, extremely short
lived argon negative states.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A cknowledgem ents
I would like to acknowledge the contributions of my parents, Gary and Lorraine
Winslade, w ithout their love and support this work would have been impossible.
I would like to thank Bob Brooks for the opportunity to try my hand at graduate
level experimental atomic physics. Throughout this degree Bob has filled the roles of
a lecturer and a supervisor. In these roles Bob has been a consummate physicist and
considerate of both the experiences of graduate students and his role in supervising
them.
Most of the experimental work in this thesis was done in Nick W estwood’s labo­
ratory, using some of his equipment, but in particular his Nd.YAG pumped dye laser.
For this and his input into this research I would like to thank him.
Rick deLaat was both a teacher and colleague to me, he taught me how to use
the experimental equipment and operated it alongside of me for the m ajority of my
degree. He also provided a few years of good conversation, laughter, and a good
source to vent with. Thank you Rick.
Bryan and Alex were fellow grad students of Bob Brooks who have helped me with
numerous problems both technical and theoretical. Bryan lent his hand whenever I
had problems with the sometimes confusing process of writing in Latex. For all of
their help, conversation, joking and ranting I thank them.
i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Contents
A ck n ow led gem en ts
i
1
1
In tro d u ctio n
1.1
Cavity Ringdown Spectroscopy
.................................................................
2
1.2
History of C R D S ............................................................................................
4
1.2.1
Background to C R D S ..........................................................................
4
1.2.2
CRDS at G u e l p h ................................................................................
5
Microwave Induced P l a s m a s .........................................................................
7
1.3
2
E xp erim en tal D eta ils
2.1
2.2
2.3
Set-up and A pparatus
9
...................................................................................
9
2.1.1
System O v e rv ie w ................................................................................
9
2.1.2
Light S o u rce..........................................................................................
11
2.1.3
S y s te m ....................................................................................................
13
2.1.4
D ata A c q u is itio n ................................................................................
19
P r o c e d u r e ........................................................................................................
21
2.2.1
Laser O p e ra tio n ...................................................................................
21
2.2.2
Cavity A lig n m e n t................................................................................
24
M irro rs...............................................................................................................
27
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
4
5
2.3.1
Mirror C le an in g ...................................................................................
28
2.3.2
Damage Issues
...................................................................................
32
2.3.3
Mirror C a ta lo g u e ................................................................................
33
E xp erim en tal R esu lts
35
3.1
In tro d u c tio n ......................................................................................................
35
3.2
Survey of the ArgonS p e c tru m .......................................................................
35
3.3
Unidentified F e a tu re s ......................................................................................
43
D iscu ssion
48
4.1
In tro d u c tio n ......................................................................................................
48
4.2
Possible Causes of the Broad Features
.....................................................
49
4.2.1
Contamination
...................................................................................
49
4.2.2
Neutral A rg o n ......................................................................................
49
4.2.3
Doubly Excited A r g o n ......................................................................
50
4.2.4
The Negative ArgonI o n ....................................................................
51
4.2.5
The Argon E x cim er............................................................................
52
C onclusions
60
5.1
60
C o n c lu sio n s ....................................
B ibliograph y
62
A
67
A rgon Line List
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Tables
2.1 The current mirror catalogue. All mirrors have a radius of curvature of
lm . The label ’’Set” corresponds to a 1 inch diameter, while "Small”
corresponds to a 1/2 inch diam eter...............................................................
34
3.1 Observed contam inants between 605nm and 677nm..................................
.37
3.2 Centroids, positions, arid linewidths of the seven broad unidentified
features. A typical linewidth in neutral argon is 0.15cm_1.....................
43
4.1 Atomic transitions in close proximity to the broad features....................
49
4.2 The upper state lifetimes of the broad features..........................................
51
4.3 Vibrational constants with experimental and observed spacing between
adjacent broad features of the same vibrational band..............................
57
A .l The generated line list of excited argon between 600nm and 700nm. .
68
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures
1.1
Pulsed light ’’ringing down” in an optical cavity.......................................
2
1.2 Three mirror configuration of a cavity lossmeter. Redrawn from [1] .
5
2.1 Schematic of the experimental s y s te m .........................................................
10
2.2
The attenuating optics of the s y s t e m ........................................................
12
2.3
Experim ental ringdown c a v i t y .....................................................................
15
2.4
Detailed schematic of the m irror m o u n t s .................................................
16
2.5 Gas in p u t/o u tp u t system currently in use, sample gas enters the system
through valve F, and exits through the main vacuum line. Note th at
2.6
separate streams of gas may be put across the face of the mirrors. . .
18
Positioning of the aperture and target on the steel fra m e s ....................
25
2.7 The calculated curve representing the reflectivity versus the ringdown
time of an empty 1.1 m etre cavity................................................................
2.8 Movement of the cleaning tissue across the mirror face
........................
27
30
3.1 The first section of the spectral survey, from 605nm to 621nrn. 01 lines
noted are the 2p33p 5Pi, 2 , 3 to 2p34d 5D° 3 4 transitions............................
38
3.2 The second section of the spectral survey, from 621nm to 637nm. N il
line noted is the 2p23p 4S3/2 to 2p26s 4P i / 2 transition.............................
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
3.3 The third section of the spectral survey, from 637nm to 653nm. 01
lines noted are the 2p33p 5Pi, 2,3 to 2p35s 5S2 22 transitions....................
40
3.4 The fourth section of the spectral survey, from 653nm to 669nm. . . .
41
3.5 The fifth section of the spectral survey, from 669nm to 677nm...............
42
3.6 Unidentified spectral features 1-3 shown with a neutral transition (N)
3.7
(line 74 in Appendix A )...................................................................................
44
Unidentified spectral features 4-7............................................. •....................
45
3.8 Integrated absorption of broad feature 5 versus power at 130mTorr,
with two neutral reference lines.....................................................................
46
3.9 Integrated absorption of broad feature 5 versus pressure at 200W, with
two neutral reference lines...............................................................................
47
4.1 The broad features plotted over the spectral region, in the absence of
all other lines. All data was taken at an argon pressure of 145mTorr,
microwave power of 175W, and using a step scan of size 0.003nm.
. .
53
4.2 Possible vibrational transition scheme between two unknown electronic
energy levels, v" is assumed to be equal to 0 here....................................
55
4.3 Possible vibrational transition scheme between three unknown elec­
tronic energy levels............................................................................................
56
4.4 Polynomial fit of the vibrational series, coe = 295 ± 11 cm -1 , coex e =
24 ± 2 cm ^1 ......................................................................................................
V I
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Chapter 1
Introduction
Cavity ringdown spectroscopy (CRDS) is a highly sensitive spectroscopic technique
using the decay time of light within an optical cavity to measure absolute absorp­
tion. Direct absorption measurements make the technique very useful in analytical
and physical chemistry, and more appealing than other sensitive techniques, such
as laser induced fluorescence th at don’t measure absolute absorption[2]. The theory
behind pulsed CRDS will be explained in section 1.1, and a brief discussion of the
development of the technique in section 1.2.1.
CRDS has lent itself to many environments including: pulsed jets, atmospheric
flame, hollow cathode discharges, laser photolysis reactors et cetera [3]. The choice to
look at microwave discharges wras natural as the plasma discharge may be sustained
by a discharge cavity outside of the optical/sam ple cavity, thereby maintaining clean­
liness. CRDS of microwave discharges for elemental analyses has been developed con­
currently at the University of Guelph and at Mississippi State University by Wang,
W instead and Duan [4] [5] [6]. CRDS of noble gas discharges also provides the spectroscopist with a sensitive way to probe transitions in species such as the argon dimer,
argon hydride, and the negative argon ion. R should be noted th at Wang et al use
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a microwave plasma torch centred in the ringdown cavity, while the experiments de­
scribed in this thesis have the microwave source coupled to the discharge tube using
a discharge cavity.
1.1
C avity R ingdow n S p ectroscop y
If a pulse of light were introduced between two parallel, highly reflective (HR) mirrors
(>99.93%), the pulse would reflect back and forth tens of thousands of times, slowly
leaking out with an exponentially decaying intensity (/), this is shown in figure 1.1.
The cavity loss rate ( ^ ) may be expressed as
Initial P ulse
•>
->
/
\
<-
O
o>
a
n
>
< -
- >
< -
T im e
-
- >
i,
Figure 1.1: Pulsed light ’’ringing down” in an optical cavity.
<n
= IT,c a v i t y
dt,
2L
( 1.1)
where Tcnvny = 2T is the cavity transmission (the total round trip transm isivity), and
T = 1 —R is the transmission of each mirror, assuming uniform reflectivities, and L
is the cavity length [7] [8]. The solution to this equation is (substituting reflectivities)
T
T
I = he
_ t Ll-R)c
L
( 1 .2 )
or expressed using the e-folding decay time (r0), otherwise known as the ringdown
time
/ = he'
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(1.3)
The ringdown time may then be expressed as
L
°
(1.4)
c(l - R)
or, with an absorber (of length ls and absorption a) present
L
(1.5)
c[(l —R ) + cv/s]
By combining equations 1.4 and 1.5 the absorption may be related to the ringdown
time of the cavity with and without a sample present
L ,1 1 ,
a = y ~ { ---------).
lsC T
.
,
(1.6)
T„
To obtain a spectrum the ringdown times of the cavity with and w ithout a sam­
ple present are measured at varying wavelengths, then used to find the absorption
(equation 1.6) for each point.
The sensitivity of this technique is due to its extremely large effective path length
which can be on the order of tens of kilometres [3]. Other advantages are th a t this
technique allows the measurement of absolute absorption intensities at any visible or
near visible wavelength, from ultraviolet to the infrared, and experimental systems are
relatively inexpensive: commercial systems are available for a few thousand dollars.
The main disadvantage to this technique are th a t the mirrors need to be highly
reflective, are costly and are easily damaged. Since the mirrors form the ends of the
sample cavity, contam ination and damage from the sample is of great concern. The
experiments described in this thesis were carried out in a microwave discharge, which
contributed to mirror contam ination and damage. Highly reflective mirrors have a
limited spectral range; studies needing great wavelength range (larger than ~60nm)
require more than one set of mirrors.
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.2
H istory o f C R D S
1.2.1
B ack grou n d to C R D S
Long initially suggested the concept of Cavity Ringdown Spectroscopy in 1967 in the
preface to his book[9], The modern application of CRDS stretches back only as far as
1980 when Herbelin et al published an experimental technique for the measurement
of photon lifetimes in an optical cavity as a method for measuring the reflectivities
of highly reflective mirrors; a system called a cavity lossmeter [1]. This method used
a cw laser and measured the phase shift of a modulated light source as a way of
measuring the cavity decay time. In this initial paper the possibility of using the
technique to measure the absorption of a substance is mentioned. The technique
measures the cavity reflectivity of a two reflector cavity, then the cavity reflectivity
with a third reflector in the system, then extrapolating the reflectivity of the third
mirror. The experimental design is seen in figure 1.2. The m ethod was successful,
allowing an increase in the reflectivities of highly reflective mirrors by a factor of 100
[10]. Deacon Research manufactured and sold cavity lossmeters; some of these were
later altered to be used in a spectroscopic capacity[7].
In 1984 Anderson Frisch and Masser [11] started directly measuring the decay
time of the light intensity in the optical cavity using an optical switch to truncate a
cw laser beam. Working at Deacon Research, Anthony O ’Keefe and David Deacon
developed the spectroscopic technique of pulsed CRDS, discussed previously (sec­
tion 1.1). Deacon and O ’Keefe now operate Los Gatos Research Inc, one of the world
leaders in providing highly reflective (HR) mirrors for CRDS
[8], and commercial
CRDS systems.
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M o d e M a tch in g
L en s S y ste m
^r’s
R1
R3
Optical
Source
T h eta
O p tical M od u lator
R2
O p tica l F ilters
D etector
Figure 1.2: Three mirror configuration of a cavity lossmeter. Redrawn from [1]
1.2.2
C R D S at G u elp h
Cavity ring-down spectroscopy (CRDS) in the Atomic, Molecular, and Optical physics
group at the University of Guelph began in the fall 1997 as an undergraduate research
project by Mike Coughlin. The project was studying the feasibility of CRDS using
existing laboratory equipment, and so a pair of mirrors wrere purchased (the first
entry in table 2.1)[12]. Lauren M acarthur carried on the work in the winter and fall
of 1998, performing an undergraduate research project with R.L. Brooks. Lauren
was involved in software development based upon programs w ritten by D. Tokaryk
[13]. During her project sophisticated mirror mounts were designed by H. Tiedje and
then used by her. The light source was a N2 dye laser. Her experimental research
was the first effort made, and resulted in measurements of the b-X (1-0) and (2-0)
transition of 0 2 [12] [14]. Galen Dunning started his undergraduate research project
in the winter of 1999; he observed water vapour in oxygen gas, and made the initial
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measurements of the molecular oxygen dimer [15].
Steve DeMille performed both undergraduate and graduate work on the system.
As an undergraduate he worked on software modification, and with testing new ex­
perimental mounts designed by H. Tiedje (figure 2.4). The 0 2 dimer was investigated
further resulting in a publication [16]. The apparatus was moved into Nick West­
woods lab, using his pulsed Nd:YAG laser. Rick deLaat and Ruth Tanner, both Nick
Westwoods students, provided assistance in the lab, especially with laser operation.
Steve’s graduate work entailed comparing CRDS to intra-cavity photoacoustic spec­
troscopy and to phase shift CRDS. The overtone spectra ( A i/ch — 6) of propane,
n-butane and neopentane were taken, and the f-values and integrated cross sections
were measured. Neopentane was used in the comparison of the various spectroscopic
techniques [12] [17].
In 2003 an investigation into helium discharges was started; attem pts were made
to find helium hydride, but no evidence of it was seen in microwave discharges of
H2/He mixtures. The 2p [D - 3d lD intercombination transition in He was seen and
attem pts were made to measure its absolute oscillator strength. The dipole 2p lD 3d
transition was used as a reference in this experiment. W ith a well known f-
value the integrated absorption could, in principle, be used to find the number density
of the initial 2p
state, leading to the determ ination of the oscillator strength of
the intercombination line. Saturation effects made the measurement of the dipole
transition untrustw orthy and this experiment was abandoned[18]; these experiments
are discussed in detail in R.H. deL aat’s PhD thesis [19].
Experimental studies of microwave discharges using Cavity Ringdown Spectroscopy
were begun in 2000 by S. DeMille and R. Tanner. The microwave discharge spectrum
of argon was briefly examined before both students left the project. This research was
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a precursor to the study of ArH bands in the region 600-700nm. The argon spectrum
was surveyed again in early winter 2004, 73 transitions in argon were observed, 13 of
them not appearing on the NIST database. The purpose for surveying argon was to
provide a spectrum for reference in experiments into ArH, which was never observed.
In the process of surveying the argon spectrum seven broad features were observed,
which have tentatively been identified as two vibrational series in the argon dimer,
although alternate interpretations are discussed (chapter 4).
1.3
M icrow ave Induced Plasm as
Microwave discharges or microwave induced plasmas have been used extensively since
the early sixties. Early microwave sources were commonly constructed from govern­
ment surplus radar equipment. One of the reasons that microwave discharge cavities
have been used in our experiment, and one of the motivations behind the initial de­
velopment, is th a t there is no need for internal electrodes, allowing simple usage with
our system while keeping contam ination to a minimum [20].
Microwave power is generated at the power supply (section 2.1.3) and is trans­
ferred to the gas via the discharge cavity, an Evenson cavity in this experiment (fig­
ure 2.3). The microwaves fill the cavity and cause free electron oscillations in the
gas. At low pressures a spark is needed to create some initial electrons to initiate the
discharge.
The oscillating electrons excite the argon atoms by means of collisional excitation:
e~ + Ar —>■Ar* + e ~.
(1.7)
The excited argon atoms may engage in three body collisions generating argon excited
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dimers [21] [22] [23]
Ar* + 2 /lr —>Ar*2 + Ar.
(1.8)
It is expected th a t Ar+ is present in the discharge as the plasma will be charge neutral,
however no evidence of it is seen in the spectra (chapter 4 and 3). Apparently the
density of the argon positive ion is too low to observe; according to Palmero et al
[24]this indicates th a t the plasma is in the low density limit, where the 4s and 4p
levels of argon are both excited directly from the ground state.
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 2
Experim ental D etails
2.1
2.1.1
S et-u p and A pparatus
S y ste m O verview
A schematic for the experimental system is shown in Figure 2.1. The light source for
the experiment is a 10 Hz repetition rate, pulsed Nd:YAG pumped dve laser operating
from 600 to 680nm. As shown the light is attenuated after
leaving the
laser and is
directed toward the ringdown cavity. The pulse enters the ringdown cavity where
the intensity ’rings down’, either for an empty cavity or with a microwave discharge
of a sample gas. The dynamic intensity is detected by the silicon photodiode. The
signal is sent to a CAMAC crate where it is digitized and multiple ringdown curves
are averaged. The digitized ring-down curves are then analyzed on a computer; the
ring-down time is determined and saved to file.
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction
Laser Control
Nd: YAG Pumped Dye Laser
Com puter
Attenuator
Assem bly
With DCM Dye
(F ig . 2.2)
(600-680nm )
Closed Cavity (1.1m)
Crq
C
a>
to
zo
o
a>
Detector
L ens
M icrowave
Assem bly
G as F low
End
prohibited without perm ission.
cr
0>
X
a>
W ind ow
E nd
Highly Reflective End Mirrors
W in d ow
CD
Reflectivity: (600-680nm )>99.97% , Radius o f Curvature: lm
CD
p
CO
CO
Microwave
ADC and
Memory Averager
(CAMAC Crate)
Power Supply
Computer with
Curve Fitting Program
2.1.2
L ight Source
The light used in this experiment is produced by a Continuum YG661 Nd:YAG
pumped TDL 60 dye laser. The YAG second harmonic (532 nm) pumps a DCM
dye in methanol, giving tunable visible output between 600 and 680nm. This dye is
long lived, having only been changed once in the past two years of this experiment.
As is common with dyes in the red, it has a broad spectral range allowing a significant
portion of the visible spectrum to be used without having to frequently change dyes.
The laser pulses are 7ns long, and have a maximum power of 60 m J/pulse.
The light from the laser is attenuated and directed down the optical axis of the
ringdown cavity. The attenuating optics are shown in Figure 2.2. The first component
is a prism set in a carbon graphite block. The power transm itted into the prism is
dumped into the carbon graphite block in which the prism is set.
The 4% back
reflection off the prism is directed by a mirror into a set of filters, the first of which is
an infrared filter used to protect the cavity mirrors from infrared damage. The second
set of filters are neutral density filters (OD=0.1, 0.2, 0.3, 0.4, 0.5, 1.0, 2.0) which are
used in combination to m aintain the light, intensity at the detector within an optim al
range for the experiment (Section
2.1.4). The intensity at the detector, which is
im portant to control (Section 2.1.4), will change as both the dye laser power and the
reflectance and transm ittance of the highly reflective mirrors change with wavelength.
Typically one combination of OD filters is used over the lnm to 4nm section of one
experimental run. A variable aperture removes stray light, and reduces the beam
size which in practice results in a cleaner ringdown signal. Typical attenuations are
between 3% and 0.04%.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dump
Graphite Beam
/
1s
tS! c
io
oo
Figure 2.2: The attenuating optics of the system
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.1.3
S y ste m
The experimental ringdown cavity is shown in Figure 2.3. Its purpose is twofold: to
act as a lm high finesse optical cavity and to contain the gaseous sample th a t is being
studied. This cavity consists of two sections of stainless steel tubing (30 cm in length,
50 mm outer diameter) on either end of which the concave mirrors are mounted,
between which a glass (Pyrex) tube (13 mm outer diameter, 60 cm in length) for
the microwave discharge is mounted, and through which the gas and vacuum ports
and pressure gauges are mounted. The tubes are supported by four flat aluminium
frames (2 per tube) which are bolted to the optical table. The frames have some small
amount of lateral movement available for the purpose of assigning the cavity axis in
the initial alignment of the system (Section 2.2.2) and to ensure the absence of shear
forces on the glass tube. The laser pulse is focused through a lens (figure 2.1) such
th at the pulse, after passing through the diverging lens formed by the front mirror,
is focused in the middle of the cavity. This lens can be moved in three dimensions,
as well as rotated in its mount.
The mirror mounts are bolted to the anchor plates on the ends of the steel tubes
(Figure 2.3), and are shown in Figure 2.4 in assembled and exploded view. The mirror
mounts consist of a steel tilt plate used to affix the mount to the steel cylinders, which
provides tilt, and a brass cylinder in which the mirrors are held. Channels allow gas
movement between the front of the mirror, and the area between the m irror and the
window; the window provides the vacuum seal. This ensures the pressure on either
side of the mirror is the same, preventing any deformation of the mirror. The mirror
is held at the top of the brass cylinder, an O-ring and brass ring hold the mirror
firmly in place. The mount is designed for 1 inch diameter mirrors; when 1/2 inch
mirrors are used a 1 inch brass converting ring is used. The 1/2 inch mirrors are
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
secured in the converting ring using Teflon screws. It should be noted th a t using
the window as a vacuum seal may be over-cautious as Los Gatos Research produces
CRDS mounts which uses the mirror as the main seal. In these mounts the mirror
movement is accomplished by applying pressure on three points on the back of the
mirror. From all indications this would work as well as the current system. The seal
between the m irror mount and the steel cylinder is a large O-ring th a t fits between
the brass cylinder and the steel tube and is shown in Figure 2.4 . It is on this O-ring
that the mounts pivot the mirror, while maintaining a seal. The mirror mount is
affixed to the steel framing tubes by bolting the tilt plate on the mirror mount to the
anchor plate. These bolts are 120 degrees apart, with tension supplied by springs,
and afford the range of angular motion needed for mirror alignment. Between the
quartz window and the brass cylinder is an O-ring which serves as one of the main
vacuum seals of the system. Between the window and the tilt plate rests a nylon
spacer preventing any damage to the window.
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tilt P late
QQ.
52 vtxo
cd
-<
O^
3
00
CO
Figure 2.3: Experim ental ringdown cavity
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
□
ooooo
ooooo
i=l .
[1]
CO
■c
T3
*cI
o
a Ja
oI
V
<D
cc
£
o0s)
Exploded
m i/m iiii
CP
i///j//////////t
ooooo
ooooo
DOOOO
u
t!
s
Assembled
ooooc
o
'V
ooooc
ooooo
Figure 2.4: Detailed schematic of the m irror mounts
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The ringdown signal from the optical cavity is measured by a detector.
The
detector assembly consists of a collimating lens in front of a silicon photodiode, a
preamplifier, and a linear offset. The photodiode is used instead of a photom ultiplier
tube as the photodiode has linearity and higher quantum efficiency [16].
The sample gas is introduced through the gas manifold, an integral part of the
two steel tubes and is shown in Figure 2.5. The gas pressure is regulated with a
Granville-Phillips leak valve allowing for fine control of the gas pressures. Typically
gas pressures are between 100 mTorr and 3 Torr. During runs a microwave discharge
is struck in the glass tube centred in the system. The microwaves are generated by a
Microtron 200 power source operating at 2.45 GHz, fed via a 3m coaxial cable to an
Evenson cavity, operating between 15 and 200 W. The cavity has two tuning knobs,
both of which are used to zero the reflected power as much as possible. The discharge
was originally initiated with a Tesla coil with the unfortunate side effect of affecting
the electronics. Now a hand-held Zero-Stat7A/, which generates a small spark by the
piezo effect, is used. Once the discharge has had a 20 minute warm up, multiple
spectral scans are taken, the discharge is turned off and the system is allowed to cool,
then the background scans are made.
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
c3 “O
o o
go
Lu
i-H
<
u
c
£3
3
ao
c
>
’ c3
s
T3
,o
Figure 2.5: Gas in p u t/o u tp u t system currently in use, sample gas enters the system
through valve F, and exits through the main vacuum line. Note th a t separate streams
of gas may be put across the face of the mirrors.
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.1 .4
D a ta A cq u isitio n
The ringdown signal is observed with the silicon photodiode. The signal is then sent
to the CAMAC crate, where a DSP model 2030 transient digitizer converts the analog
signal into an
8
bit digital signal. The signal is only digitized between -256mV and
256mV. Any signal below -256mV is cut off, and any signal above 256mV is wrapped
around to -256 mV making it imperative th a t the signal stay in this range. The
signal should take up more than half of this range to make full use of the
8
bits. The
varying laser powers and reflectances at different wavelengths force the change of the
OD filters about once every 2-4 nm to maintain the signal size. There is a linear
offset on the detector th a t is used to ensure the baseline is close to -256 m V ; when
a discharge is struck this must be adjusted as the baseline will be raised. A memory
averager averages some set amount of curves together, typically 10, 40 or 100 curves.
The computer analyzes the ringdown curves then saves the inverse ringdown time
and wavelengths into columns that are saved in a data file.
number of programs written to run this experiment.
There have been a
CRDSTME2, most recently
modified by Bob Brooks, is the program used for continuous scanning runs, where
the laser is continuously scanned between two wavelengths while the ringdown curves
are analyzed at the points in between. The program CRDSSTP is used for aligning
the mirrors in the cavity, and the program CRDSALL contains both of the functions
of CRDSTME2 and CRDSSTP, but also allows the laser to be operated in a step scan
mode. CRDSALL is limited to 512 d ata points in the final file so all runs must be kept
below this number of points. In the following paragraph the function of CRDSALL
will be described, with the understanding th a t the other two programs may be used
for some of the functions. There exists another difference between CRDSALL and
CRDSTME2, which is th a t CRDSTME2 chooses the part of the ringdown curve
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
between 90% of the maximum and 10% of maximum for analysis. This is done as this
region has the best exponential characteristics. CRDSALL for diagnostic reasons,
chooses and analyses this range as well as the three regions 90%-25%, 40%-10%,
20%-2.5%. The variation among these ranges provides insight into the exponential
character of the ringdown curve, large differences indicating a non-single exponential
curve.
The easiest way to describe the program is by going through the pertinent menu
items step by step. CRDSALL contains five headings, each of which lead to a sub
menu. The first Heading is module p a ra m ete rs where the param eters of the d ata
acquisition may be changed for a run. The sampling interval may be changed from
33 ns to 3.3 ps. This is the length between data points on the d ata curve,
66
ns is
typically used in the experiment, but for strong spectral lines or when the reflectivities
of the mirrors decreases, the sampling interval is set to 33 ns. The record length may
be set between 8192 bins and 64 bins: this param eter is left at 512 bins for almost all
experiments. The individual ringdown curves may be saved along with the inverse
ringdown time normal y saved. This function is used to manually check the exponential
nature of the ringdown curves.
The second main option is scan w hile c o lle c tin g d a ta .
When chosen, the
starting w avelength/the end wavelength, the name of the file, and the scan speed are
entered. Continuous scanning mode is the main method by which coarse spectra are
taken. The laser is set to the determined speed (between 0.0023A/s and 0 .6 A /s), and
scans over the range continuously while the computer takes and analyzes the data.
This method is much quicker than the stepping mode but has some drawbacks. The
curves are averaged over a range of wavelengths. Averaging more curves increases
this range reducing the effective resolution of this method, while increasing the signal
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to noise. The laser scan and the data collection are synchronized by simply pressing
enter at the same time on both the collection PC and the laser control computer.
For high resolution scans the current method is to step the laser, which is the third
option, and unique to the CRDSALL2 program. The d ata PC must be connected
to the laser control PC via an RS232 cable. The program will ask for the beginning
wavelength, the end wavelength, the file name and the step size. The data com puter
tells the laser computer to begin, then the laser is slewed to the beginning wavelength
where the number of ringdown curve acquisitions at each step is made, the d ata being
collected and analyzed. The laser then moves to the next step and collects another set
of data. While being slower this method averages many ringdown curves at a given
wavelength, and has a higher resolution than the continuous scan. The smallest step
that can be taken is O.OOlnm, which is the finest reading the laser offers.
For tuning the ringdown cavity the fourth option is chosen (Tuning). The curves
are analyzed at 33ns,
66
ns, and 133ns bin sizes, and the ringdown time is printed to
screen with each ringdown time. This allows the result of alignment adjustm ents to
be immediately seen, and the difference in the ringdown of the different bin sizes to
be seen.
2.2
P rocedure
2.2.1
Laser O p eration
As mentioned in the preceding section the light source used in this experiment is a
pulsed Nd:YAG pumped dye laser. R is im portant to note th a t proper safety glasses
are worn at all times. The initial step of the experimental procedure is to turn on and
warm up the laser. The pumps th at circulate the laser dye through the oscillating
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and amplifying chambers are turned on; safety features will not allow the laser to
operate otherwise. The Nd:YAG coolant water must be flowing. The main power is
turned on and the operator’s key is then used to turn on the laser. The Nd:YAG
lamps are manually flashed 20 times, then set to automatic repetition at lOhz for 30
minutes as a warm up. Once the flash lamps are warmed up, the shutter between the
pump laser and the dye laser cells must be opened manually. On the laser control
box the shutter is opened (the actual feature is not in operation), the repetition rate
is set to f (lOhz), and the Q-switch is turned on. The dye laser will now be lasing.
W ith the main exit shutter closed the dve laser is allowed to warm up for 30 min.
This 1 hour warm-up time is very im portant as it leaves the laser in a stable state;
the beam characteristics will remain unchanged until the end of the day.
After the one hour warm up procedure, the power of the frequency doubled
Nd:YAG pump beam and of the dye laser is measured. This allows the performance
of the laser to be tracked so th a t maintenance may be performed when needed. The
shutter to the YAG is closed, the cover to the dye laser is lifted, and a Gentec power
meter is placed in the chamber so that, the light from the pump laser will be incident
on the meter instead of being split to the dye oscillator and amplifiers. The YAG
shutter is opened allowing the -532nm doubled light to be incident on the detector.
The power is recorded; typically it takes five minutes for the power reading to equili­
brate. There are minor alignment controls for the YAG in the form of a single toggle
switch. This switch may be used to shift the 532nm frequency doubling crystal to
increase the power of the doubled Nd:YAG, this is done every day. The dye laser
cover is replaced and the cover to the dye doubling chamber is removed. The de­
tector is placed in front of the dye laser output beam path. The dye laser power is
measured at a constant wavelength (generally 635nm) to get a long term feel for the
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
condition of the dye, although if substantial work is being done at higher wavelengths
the power is measured at 660nm. It should be noted th at the dye laser is capable of
being frequency doubled, but is not for the CRDS experiment.
The wavelength calibration is checked periodically by taking a known spectrum
and correcting the laser’s wavelength by comparing the measured positions of the
spectral lines with the actual positions. Near the end of the experiments described in
this thesis, the calibration shifted overnight, so the calibration was checked everyday.
The cause of this shift is unknown at present. The helium line at 667.817nm was
used for this calibration due to its immense strength, and the simplicity of the he­
lium spectrum in this region. The laser was parked overnight at 667.817nm, so th at
it becomes immediately obvious the next day if the calibration was lost overnight.
When working in argon the calibration is checked using argon only for minor drifts.
Due to the richness of the argon spectra any significant calibration errors are almost
impossible to characterize, as one may lose oneself in the forest of argon lines.
Periodically throughout this work there have been power outages. This is a special
case but is im portant as any lengthy experiment may encounter an outage. After an
outage the dye stepper motor must be manually turned off, then the rest of the laser
is turned on as usual. The laser is sent to the wavelength th a t it was parked at the
night before. The stepper motor is then turned on, in this wav the calibration may
be kept to within one nanometre through a power outage. Leaving the stepper motor
on will result in a calibration drift of many nanometres.
When moving across large distances in the continuous scan mode, slew rates of
10A /s or lA /s are common, as are scan rates of 0.6A/s to O.OObA/s. It has been found
th at the top slew rate of 10.4/s causes calibration changes, and so slewing should only
be done at, or below lA /s.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
When the day’s experiments are finished the dye laser is parked at 667.817nm and
the dye laser is shut off in the reverse order from which it was turned on, except there
are no cool-down times. Parking the laser gives a small defence against calibration
problems, a steady wavelength to measure the daily power with, and a wavelength to
set the laser to in case of power outage.
All internal alignments and maintenance of the laser were done by R.H. De Laat,
PhD candidate, Departm ent of Chemistry.
2.2.2
C a v ity A lig n m en t
The most critical experimental procedure is the alignment of the ringdown cavity.
The initial alignment, including the alignment of the steel tubes will be described
with additional comments on smaller alignments.
After the laser pulse exits the
attenuating apparatus described in Section 2.1.2, it is directed down the ringdown
cavity axis. First the support frames for the steel cylinders are bolted to the table,
roughly centred to the laser beam (Figure 2.2.2). Two plastic targets are used in
this initial alignment. The targets have the same diameter as the outside diam eter
of the steel tubes. One of the targets has a small aperture (3 mm) in the middle, the
other a series of concentric rings forming a target. The aperture is placed in the first
frame, the target being placed in the second. The position of the pulsed laser beam
and of the mounts are adjusted such th at the beam goes through the centre of the
aperture, and is incident on the centre of the target. The target is then placed on the
third frame, the aperture staying on the first frame. The position of the third frame
is changed to align it with the pulse path. The fourth frame is aligned as was the
second and third. The target and aperture are then placed in different combinations
on the frames to ensure th a t the beam path runs down the centre of the frames. This
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Frame 4
Frame 3
Frame 2
Frame 1
F rom L igh t S ou rce
Figure 2.6: Positioning of the aperture and target on the steel frames
alignment is im portant not only for the mirrors, for they are adjustable and may cater
to a slightly off axis beam path, but also so th a t there are no shearing forces on the
glass tube in the centre of the optical cavity.
The rear stainless steel tube is affixed to legs 3 and 4, such th a t the final distance
between the mirrors will be 1.1 metres. The mirror mount is bolted to the rear anchor
plate, ensuring th a t the distance between the anchor and tilt plates is uniform. The
main laser shutter is opened such th a t a laser pulse is incident upon the rear mirror
reflecting back towards the laser. The m irror position is adjusted using the three tilt
screws so th a t the reflection is centred on the iris in front of the neutral density filters
(Figure 2.2). The front steel tube is bolted into place, the front m irror mount is bolted
to the front plate, and the sample tube is fixed into its position (Figure 2.3). The
back reflection from the first mirror is also centred on the aperture. The focusing lens
is now placed in front of the system, its back reflection is centred on the aperture,
the refracted reflections are used to align the x-y position as well as the angle of
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the lens; typically the aperture is reduced to its smallest size for this, then restored.
The detector assembly is now mounted behind the rear mirror, a ring-down signal
should now be visible on the oscilloscope with a lifetime 60%-80% of the maximum.
The positions of the lens, the front mirror, and the rear mirror positions are further
adjusted to maximize the lifetime of the signal, while maintaining a single exponential
character to the ringdown curve. W ith relative ease distortions can be introduced to
the curve, even with an increase in the ringdown time. The curve must be analyzed
visually to ensure th a t it has a single exponential character, with as little noise as
possible.
In those cases when the mirrors have been removed for cleaning, the alignment
procedure is simpler as the cavity axis is not affected. The rear mirror mount is bolted
into place, and the reflection is centred on the aperture as usual. Hand pressure on
various points on the rear mount allow exploration of the man} mirror positions
quickly. The front mirror and lens are attached and aligned as usual.
On a day to day basis the ringdown cavity’s alignment is tweaked to maximize the
ringdown time. The need for this is due to slight shifts in the alignment overnight,
and in moving to different ranges the number and kind of neutral density filter are
changed, which slightly changes the alignment of the laser beam. In this case only
the front m irror is adjusted, unless the ringdown time is very low (2 /i/s) then the
lens is adjusted. If the ringdown time is not restored with these efforts then the back
mirror may be adjusted.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reflectivity vs. Ringdown
19
17
15
13
11
9
7
5
3
1 *—*0.998
0.9982
0.9984
0.9986
0.9988
0.999
0.9992
0.9994
0.9996
0.9998
Reflectivity
Figure 2.7: The calculated curve representing the reflectivity versus th e ringdown
time of an empty 1 . 1 metre cavity.
2.3
M irrors
Highly reflective mirrors are an integral component of Cavity Ringdown Spectroscopy.
This experiment uses multi-layer dielectric coated mirrors th a t are widely available
on the commercial market. Figure 2.3 shows the relation between mirror reflectivities
and ringdown times (using equation 1.4), reflectivities m ust be above 99.93%, being
nominally above 99.97% reflective to have a usefull ringdown tim e (1.5-11^/s). Unlike
m etal coated mirrors the dielectric coatings have a negligible absorption, a necessary
feature as the light used in the experiment m ust first pass through the front mirror
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
coating to enter the cavity, then be transm itted through the second m irror for mea­
surement. The drawbacks to multilayer dielectrics are the spectral range is narrow
compared to metal coated mirrors, and the coatings damage readily (Section 2.3.2).
Great lengths are taken to ensure gentle handling and safe transport of the mirrors
especially with regard to the coating.
2.3.1
M irror C lean in g
The cleaning of highly reflective dielectric coated mirrors constitutes a time consum­
ing part of CRDS. During the past
6
years there have been five different m ethods
employed to clean the mirrors. These methods have been adopted and abandoned
as they were either too gentle and time consuming, or too aggressive. Inherent to
the issue of finding a repeatable cleaning process is the compromise between being
so passive as not to adequately clean the surface, and being so aggressive th at the
coatings are damaged. This conflict is compounded by the fact th a t the mirrors are
damaged in use (section 2.3.2), so th at methods apparently gentle to new mirrors will
enhance the damage on used mirrors. The cleaning is made all the more difficult, as
outside of deep damage or significant contamination, mirrors with usable reflectances
appear identical to the naked eye as unusable mirrors. It should be noted th at powder
free latex gloves are worn throughout every cleaning method discussed.
The first step in any cleaning of the mirrors is to remove the mirrors safely from
their holders. The mirror mount (see Figure 2.4) is turned so th a t the tilt plate is
accessible, and held in a recessed block as a base. The tilt plate is removed, being
careful of the window; the window is removed with forceps and placed on lens tissue.
The mount is then turned mirror side up, and the brass holding ring is removed. The
mirror and its O-ring may be stuck to the holding ring, in this case the mirror is simply
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
taken off the ring. If the mirror is not stuck to the holding ring a Teflon stopper is
inserted from below, pushing the mirror out of the brass cylinder. If 1/2 inch mirrors
are in use, the adapting ring is removed before cleaning. Once the mirrors have been
cleaned, the mount is reassembled.
The first method used was one of sequential four solution agitation. The mirror
was placed into a holder, affixed to a long handle. The mirror was then immersed into
a solution and oscillated, gently, back and forth for a period of
20
minutes to
1
hour.
The mirror was then rinsed with clean solution from a squirt bottle, and immersed in
the next solution. The first solution used was 1% detergent in distilled water, the main
cleanser in the process, removing both water soluble and organic deposits. The mirror
is rinsed with clean soap solution, and transferred to a second solution of distilled
water. The mirror is agitated, but for a slightly shorter period of time. The purpose
of the distilled water is simply to remove any trace of detergent. The mirror is rinsed
with clean water, then immersed in the third solution, isopropanol, and agitated for
twenty minutes. The isopropanol removes the water, and may also dissolve some
organic soluble contaminants, and the mirror is rinsed with clean isopropanol. The
fourth immersion is into acetone, again agitation occurs, and the mirror is rinsed with
acetone. The purpose of the acetone is that it is a very good organic solvent, and it
removes the isopropanol. The acetone does not evaporate off the mirror, this would
cause streaking and the deposits of minute amount of solutes, it is blown off before it
has a chance to evaporate using a gentle gust of dry nitrogen gas, having been passed
through a dust filter.
The method of four solution agitation proved to have many problems. It is very
time consuming requiring a full day to clean a single set of mirrors once, including
time to mount the mirrors and check the ringdown time. Considering th a t it takes a
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2.8: Movement of the cleaning tissue across the mirror face
number of distinct cleaning sessions to restore the reflectivity of a mirror, this could
turn into days, sometimes as much as a week for a single cleaning. As mirrors may
be dirtied in as little as a week, this method is too time consuming. While being
potentialy useful for non-discharge work it is found that the mirrors need cleaning
3-4 times as often under discharge than under non-discharge conditions.
To make the m ethod more abrasive, the mirror was scrubbed when immersed.
A sheet of lens tissue was rolled onto a cotton tipped applicating stick and the tip
immersed in the solution. The surface tension of the solution on the mirror surface
held the tip of the lens tissue to the m irror surface. The tissue is passed back and forth
over the surface while moving it laterally across (Figure 2.3.1). After one pass the
tissue is discarded and another prepared. This method is significantly more aggressive
than the previous, but more aggressive methods were employed later to deal with the
contam inants from the discharge. The tissue is pressed a little harder than is safe to
the surface, as the weakening of the tissue bv the solvent destroys the ability to gauge
the pressure. This can result in scarring of the mirror.
When it was decided to become more aggressive with the mirrors, the decision
was made to move toward the widespread practice of dropping and dragging solvents
across the surface of the mirror. A flat bed of lens tissue is made, onto which the
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mirror is placed face up. A strip of lens tissue is placed on top of the mirror, and
a drop of solvent is placed on the mirror surface, through the tissue. The tissue,
being bound to the mirror by surface tension, is dragged across the mirror so th a t
the solvent is slowly dragged off the mirror. The solvent does not evaporate on the
mirror, and the tissue provides a slight amount of abrasion to aid the cleaning. This
is repeated many times, sometimes alternating the solvents used.
Several different solvent combinations have been used in the drop and drag method.
The first combination was isopropanol, for its organic and water solubility, and ace­
tone for its usefulness as a cleaning solvent. This combination was quickly discontin­
ued as it was discovered th at isopropanol while safe with new mirrors amplifies the
existing damage in the mirrors used with discharges, quickly producing visible dam ­
age. The second solvent combination used was acetone and methanol. Both solvents
are relatively gentle on the mirrors, and do not seem to cause damage beyond th a t by
the light surface contact of the tissue. It was found that it takes 2-5 distinct clean­
ing sessions to restore the mirrors to a usable level after long exposure to a helium
discharge.
To reduce the am ount of sessions needed to clean the mirrors it was decided, again
in consultation with groups from UNB, to immerse the mirrors in a 50% solution of
acetone and methanol, and to sonicate the mirrors in this bath for one hour. After
sonication the mirrors are placed onto a lens tissue bed, and dragged on the nonreflective side; this leaves the back surface clean and does not produce any damage.
The fronts of the mirrors are then subjected to a usual drop and drag cleaning,
alternating between acetone and methanol. This method has proved gentle enough
to cause little damage, and aggressive enough to restore mirrors to a usable level
in one cleaning. This is the current cleaning m ethod employed for highly reflective
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mirrors.
The fifth method to be discussed has not been in use for 4 years and was de­
veloped by S.W. DeMille during his M.Sc research. This m ethod is an evaporative
technique using trichloroethylene to do a hard clean of the mirrors. In a fume hood
trichloroethylene (trike) is brought near to its boiling point so th a t it may evaporate
w ithout boiling, which would cause trike to hit the mirror. The m irror is suspended
above the solution at a forty five degree angle. The trike will evaporate, and condense
on the m irror surface, and drip back into the solution, taking contam inants with it.
2.3.2
D a m a g e Issues
T hat the highly reflective mirrors will be contam inated upon use is a m atter of course.
Since a change in reflectivity of 0.01% is noticeable, minor contam ination requires
cleaning. In the presence of a discharge, the contamination becomes complicated by
the effect of the discharge on hydrocarbons. Any oil, dust, and other contam inates can
be cleaned off with relative ease, neither the cleaning procedure nor the contam inants
causing significant damage. The mirrors need cleaning every 1-3 weeks of usage.
When discharging helium, ultraviolet radiation burns the coating on the mirrors.
While it takes a significant time (approximately 1 month of constant usage) under
helium discharge to burn the mirrors, they will be burnt beyond repair. The argon
discharge damages the coatings to a lesser extent than the helium discharge. Mirrors
last 3 to 4 times longer when used in an argon discharge.
This slowly accruing damage compromises the cleaning process.
Cleaning the
mirrors, considering the more abrasive methods required to affect the discharge con­
tam inants, appears to amplify the ultraviolet damage. Cleaning the mirrors is traded
off against slight damage. The methanol and acetone drop and drag seems gentle
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
enough not to damage the mirrors until after long exposure to the discharge. W hen
the damage reaches the point where it is visible, the mirrors are considered to be
unusable.
The damage may be noticed in a few different ways before it becomes visible to
the naked eye. The change in ringdown time is the most obvious way, although it
is hard to tell the difference between radiation damage and general contam ination
in this way. By allowing some acetone to evaporate off the surface of the m irror
one sees preferential evaporation from the undamaged parts of the mirror. W ater
proves more useful than acetone as it evaporates more slowly. A similar method is
one of condensation not evaporation. By holding the mirror (at the edge) between
two ungloved fingers, the hand’s moisture condenses on the mirror surface.
The
condensate on damaged sections scatter light differently than undamaged areas. In
this way mild scratches from cleaning, UY damage, and gross contam inants can be
seen. Looking at the reflection of light off the mirror at grazing incidence may also
reveal some damage.
2.3.3
M irror C a ta lo g u e
There are currently nine pairs of mirrors in the possession of the research group.
The mirrors are bought from two manufacturers, Layertec (a German firm), and
Newport.
The Newport mirrors are $1500 for a pair of 1 inch diam eter mirrors.
These mirrors have been characterized to be more robust to damage and to have
a broader spectral range than the Layertec mirrors, but have been discontinued by
Newport. The Layertec mirrors cost $1000 for a 1 inch pair of mirrors, and $500 for a
1/2 inch pair of mirrors. While these seem more susceptible to damage and to have a
slimmer spectral range than the Newport mirrors, the cost difference make them the
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mirrors of choice, especially as
1 /2
inch mirrors work just as well as
1
inch mirrors. A
third company has mirrors th a t by description and reputation are much more robust,
and have very broad spectral ranges, the company being Los Gatos Research, owned
by CRDS pioneer Anthony O ’Keefe. These mirrors have never been bought as they
cost $2000 a pair. A pair will be bought in the future to see if the rise in quality and
robustness is proportional to the rise in price.
The current inventory follows, listed by our internal labelling system, and noted
with the pertinent information of the manufacturer, listed range, reflectivity and
current state.
M irror
Set
Set 01
Set 02
Set 03
Set 04
Set 05
Small 01
Small 02
Small 03
Small 04
Small 05
M anufacturer
R ange
C on d ition
Newport
Newport
Newport
Layertec
Layertec
Layertec
Layertec
Layertec
Layertec
Layertec
583-663nm
583-663nm
583-663nm
620-680nm
590-660nm
590-660nm
620-680nm
620-680nm
590-660nm
590-660nm
No visible damage
Damaged by glass shards
Centre of mirror visibly damaged.
Burnt and scratched
Visibly burnt
Damaged
Undamaged
Undamaged
t = 2.1 fxs at 650mn
Unused
Table 2 . 1 : The current mirror catalogue. All mirrors have a radius of curvature of
lm . The label ’’Set” corresponds to a 1 inch diameter, while ’’Small” corresponds to
a 1 / 2 inch diameter.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 3
Experim ental R esults
3.1
In trod u ction
This chapter presents the experimental results of cavity ring down spectroscopy of an
argon microwave discharge. As much as possible discussion will be kept to chapter
4. The experimental methods used are those presented in chapter 2. Specific condi­
tions such as gas pressures, microwave powers, scanning method, and averaged curve
numbers will be mentioned where appropriate. The spectra were taken in the range
605nm to 677nm due to the limitations of the reflectivities of the mirrors and the
dye laser range. These limitations are seen experimentally as noise at the extreme
edges of the spectrum, and can be noted at the beginning of figure 3.1 and the end
of figure 3.5.
3.2
Survey o f th e A rgon S p ectru m
The initial spectra obtained were a survey of the region 605nm to 677nm, taken
both to provide a reference in further experiments with argon, and to ensure th a t no
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
unidentified features were present. Experiments searching for argon hydride yielded
no results. Across the survey spectrum 73 argon transitions were observed, originating
from the 3s 2 3p5 4p and 4s configurations. The lines were identified using the National
Institute of Standards and Technology (NIST) spectral database for argon [25]. This
database could not identify all of the observed argon lines as it is based upon emission
rather than absorption spectroscopy. For the purpose of fully identifying the argon
absorption spectrum a spectral line list was generated from a database of atomic
argon energy levels [25], yielding a list of all possible neutral argon transitions. This
list was generated by applying parity and A J = ± 1 ,0 as selection rules, limiting
transitions to the region 600 to 700 nm. These selection rules were chosen as parity
is rigorous, and A J is nearly so [26], but usual coupling schemes are poor in argon so
no other rules were applied (Appendix A). This procedure generated 124 lines, 46 of
which are not in the NIST spectral line list. Of these 46, 13 were seen in absorption
between 605 and 677nm, as was the entire original NIST line list. The generated
list is presented in Appendix A, with indications as to which lines were observed and
which are additions to the NIST database.
In addition to argon, transitions due to contam inants were observed in the spectral
survey. The contam inants observed were atomic oxygen, atomic nitrogen, the positive
nitrogen ion, and the Balmer alpha transition of hydrogen (table 3.1). The presence
of systematic errors in the table are because the experimental positions of these lines
have not been calibrated. Any other transitions from these species are lost in the
background noise of the spectra. No other known species were observed.
There are 7 relatively intense broad features th at have not been identified in any
databases or in the literature. These features will be examined further in section 3.3
(figures 3.6- 3.7).
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Species
Oxygen (Of)
Nitrogen (Nil)
Hydrogen (HI)
E x p erim en ta l
P o sitio n
615.587nm
615.666nm
615.807nm
645.380nm
645.460nm
645.610nm
632.116nm
656.284nm
L iterature
P o sitio n
615.597nm
615.678nm
615.819nm
645.360nm
645.444nm
645.598nm
632.151nm
7 unresolved lines
656.271nm to
656.291nm
Table 3.1: Observed contam inants between 605nm and 677nm.
The full spectral survey is shown over the following pages in figures 3.1 to 3.5.
Argon gas pressures range from 50 - 120 mTorr, at a microwave power of 200 W.
The range of pressures is due to day-to-day variations in equipment operation. As
these spectra were taken over two months, there is significant difference in the noise
and quality of the spectra. The survey was taken using the continuous scan method
at
.
0 01 2
A /s, and each data point analyzed is the average of
10
ring down curves.
Spectra were recorded in 4mn sections at a scan speed of 0 .0 1 2 A /s, using ten laser
shots per point.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
617
618
619
620
621
613
614
615
616
617
609
610
611
612
613
c
D
<
'/
605
606
607
yf-Ywv^vA^-fMVAV<v''-^ - ^ ,w'
-tt/'~''J } ^ /Kv/V-z/v-'-'
608
609
W avelen gth (nm )
Figure 3.1: The first section of the spectral survey, from 605nm to 621nm. 01 lines
noted are the 2p 3 3p 5Pi 2 3 to 2p 3 4d 5 D)°°1,3,4
3 4 transitions.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Feature 1
v«V-v~vyv>AvA'v’
633
634
635
636
637
N il
D
j£
629
630
631
632
633
625
626
627
628
629
621
622
623
624
625
oc
Cl
W avelength (nm )
Figure 3.2: The second section of the spectral survey, from 621nm to 637nm. N il line
noted is the 2p23p 4S ^ 2 to 2p26s 4P i / 2 transition.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
649
650
651
652
653
Feature 4
(sjjufl t-VV) uondiosqv
Ar-'-'Jf t '
■'V-VV-'Wv-//*-\^v,'>v\-A'<v,^ rv‘—
--v^.Vy-
645
646
647
648
649
641
642
643
644
645
Feature 2
Feature 3
I
637
638
659
640
Wavelength (nrn)
Figure 3.3: The third section of the spectral survey, from 637nm to 653nm. 01 lines
noted are the 2p33p 5Pi, 2, 3 to 2p35s 5S2 j 2,2 transitions.
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
666
665
667
669
668
Absorption (Arb. U nits)
Feature 7
661
662
663
664
665
657
658
659
660
66!
655
656
657
Feature 6
653
654
W avelength (nm )
Figure 3.4: The fourth section of the spectral survey, from 653nm to 669nm.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Absorption (Arb. Un
673
674
669
670
675
676
677
671
672
673
W avelength (nm)
Figure 3.5: The fifth section of the spectral survey, from 669nm to 677nm.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Feature
1
2
3
4
5
6
7
Wavelength
nm (air)
634.673
639.258
639.882
645.248
650.636
653.303
664.512
Frequency
cm_1(vac)
15,751.80
15,638.82
15,623.57
15,493.64
15,365.34
15,302.61
15,044.48
±
±
±
±
±
±
±
0.05
0.05
0.03
0.05
0.05
0.05
0.05
linevvidth
cm -1
1.0
0.61
0.41
0.44
0.38
0.38
0.35
Table 3.2: Centroids, positions, and linewidths of the seven broad unidentified fea­
tures. A typical linevvidth in neutral argon is 0.15cmW
3.3
U niden tified Features
The seven distinct unidentified broad features seen in the survey spectra are repro­
duced with greater resolution in figures 3.6 to 3.7. The features are preceded with
a plot of a typical transition in neutral argon. The pressures used for these broad
transitions were between 200 and 300mTorr, with a microwave power of 200W. The
spectral data was taken using the laser step method of d ata acquisition with a step
size of 0.003nm (giving roughly 30 points per line), 40 ring down curves per point.
Each of these features have been fitted with a Lorenztian profile using Microcal7 M
Origin v6.0, with R 2 values of 0.95 or better. The centroids of the transitions were cal­
ibrated off two known Arl transitions each, by taking the neutral spectra in the same
experimental run. Uncertainties of 0.05 cm -1 or less were achieved in this manner.
The calibrated centroids and line widths of the features are shown in table 3.2.
Two experimental studies were performed on these features. In the first, spectra
were taken with much greater concentrations of atmospheric contam inants present by
opening the system to the atmosphere shortly before obtaining the data. No change
in the intensity or character of the lines was observed.
In the second study the power of the discharge was varied while holding the
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4
2.0
1.6
o
0.8
15748
15388
o
15750
15752
15754
15622
15624
15626
2.8
2.4
2.0
-
0.8
Ur.
0.4
15634
15636
15640
15644
15628
Frequency (cm*1)
Figure 3.6: Unidentified spectral features 1-3 shown with a neutral transition (N)
(line 74 in Appendix A).
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Absorption (*10' )
6
6
5
5
4
4
3
3
2
2
1
1
0
0
15490
15492
15494
15496
15498
15366
6
15368
"1
15370
15372
15374
------- '--------r
5
4
3
2
V y -v' H w *
0 --15302
_1_ _ _ _ _ ._ _ _ _ _ 1_ _ _ _ _ i_ _ _ _ _ I_ _ _ _ _ i_ _ _ _ _ I_ _ _ _ _ i_ _ _ _ _ L_
15304
15306
15308
15310
15312
Frequency (cm )
Figure 3.7: Unidentified spectral features 4-7.
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
G
O
Broad Line 5
O
5- h
O
00
£>
<
<D
-t-j
G
W)
<D
Arl line 1
A rl line 2
•
▲
*
J
0
i
20
1
40
i
I
i
I
i
I
i
I
i
I
i
I
i
L
60 80 100 120 140 160 180 200
Microwave Power (W)
Figure 3.8: Integrated absorption of broad feature 5 versus power at 130mTorr, with
two neutral reference lines.
pressure constant and vice versa. The integrated absorption of each of the broad
lines was plotted against power and pressure (respectively) alongside the integrated
absorption of two neutral lines. To ensure th a t the observed phenomenon occurs
in all the lines, the study was carried out on multiple lines w ith identical results.
The plots of integrated absorption versus pressure and power, respectively, are shown
in figures 3.8 and 3.9. Note th a t due to experimental lim itations fewer points of
absorption versus pressure were measured. It is obvious th a t for the broad features
absorption increases with increasing power and increasing pressure at a rate greater
than the two neutral lines. It is assumed th a t the plateau in the neutral lines is due
to the creation of excited states being balanced by collisional relaxation and argon
dimer formation.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
1
1
i
-------- 1------
i
•
•
G
O
CU
!—
i
oC/J
•
Broad Line 5
•
▲
▲
X)
■
<
T3
<D
03
OJQ
D
4—>
■
A
Arl line 1
Arl line 2
▲
■
-
i
60
-
■
i
120
180
Argon Pressure (mTorr)
•
i
240
Figure 3.9: Integrated absorption of broad feature 5 versus pressure at 200W, with
two neutral reference lines.
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 4
D iscussion
4.1
In trod u ction
The survey spectrum shown in the last chapter, and the line list in appendix A,
show the existence and identification of many neutral argon lines. In addition to
these there are lines due to contam ination and seven unidentified broad features.
C ontam ination either from the atm osphere or the gas source, transitions in neutral
argon, or transitions in doubly excited argon were the first causes considered as sources
of the features. The final two possibilities considered were th a t the lines are due to the
argon negative ion or the argon excited dimer (excimer). Each case will be discussed
in the following sections.
It should be noted th at 13 lines in neutral argon were observed th a t are not
present in the NIST database, nor in the literature. This is due to the fact th a t our
measurements are in absorption rather than emission. The fact th at these lines were
identified by simply applying selection rules to the well know energy levels of argon
means th at there is little surprise in seeing additional lines.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2
4.2.1
P ossib le C auses o f th e Broad Features
C o n ta m in a tion
It is unlikely th at a contaminate is the cause of the broad lines as
110
single im purity
could cause all of the broad features. The introduction of air into the system prior to
measurement does not affect the intensity of the features indicating th a t they are not
due to atmospheric contamination. A search for the features through the NIST atomic
database yields no single species th at could be responsible for all of the features, this
is shown in table 4.1. Not only are the possible species different, some of the lines
would be due to species in the lanthanide series (EuII and G dll in table 4.1); such
rarities are not expected to be present.
U nidentified
Feature
1
P o sitio n
(nm, air)
634.673
2
3
4
5
639.258
639.882
645.248
650.636
6
653.303
7
664.512
P o ssib le
Sp ecies
G dll
All
Mgll
Mgll
Fel
Osl
VI
NI
Zrl
Nel
Osl
Xel
EuII
P o sitio n
(nm, air)
634.665
634.668
634.674
634.675
639.2-54
639.886
645.234
650.630
650.636
653.288
653.314
653.316
664.511
Table 4.1: Atomic transitions in close proximity to the broad features.
4.2 .2
N eu tr a l A rgon
In section 3.3, the integrated absorption of the lines was presented versus power
and pressure, respectively (figures 3.8 and 3.9). The behaviour of the unidentified
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
features show marked difference to the behaviour of neutral argon. The broad features
integrated absorption increases at a rate greater than the neutral species integrated
absorption in both cases. The features also have much broader linewidths (shorter
lifetimes) than those expected, and observed, for an atomic transition. Due to these
differences, and the absence of the broad lines in the generated line list, it can be
concluded th a t the broad features are not due to neutral argon.
4 .2 .3
D o u b ly E x cited A rgon
Doubly excited transitions in argon was one of the candidates discussed as a cause for
the broad features. In general doubly excited states will autoionize unless they are
forbidden to do so. The selection rules for autoionization are parity, A J, A L, A S = 0,
understanding th a t the free electron is treated as a 21 where I is any integer [26], The
final state will be .3p5 2P + 2 /, while the initial state will form 3p4[1D ,1 S,'3 P]nln'l'
states. It is clear that the initial state will form singlets, triplets and quintets, while
the final will only form singlets and triplets: only the quintet states will be spin
forbidden to autoionize.
A m etastable quintet state could form the lower level of an absorption transition
to higher lying quintets. R.L.Brooks has performed multiconfiguration Hartree-Fock
calculations and cannot determine if the 3pl4s4p *P3 or the 3pA3d4s 5D 4 lies lowest.
It is expected th a t the doubly excited states have more energy than the ionization
threshold; therefore the formation energy of the doubly excited state should be similar
to th at of the positive ion [26]. No positive ion lines have been observed even though
there are transitions in this region [25], leading to the conclusion th a t there is not
enough energy to form the positive ion collisionally, or by extension, the doubly
excited state.
There is also no reason why doubly excited transitions should be
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
broader th a n the atom ic transitions.
4.2 .4
T h e N e g a tiv e A rgon Ion
The possibility th at the features are due to transitions in the negative argon ion is
feasible. The negative ion has a well known lower level th at is m etastable and could
form the initial state in an absorption transition. The lower level is the 3p54s4p ASz /2
level with a long lifetime of 260ns [27]; its existence makes a transition in negative
argon a possibility. The upper levels should be short lived which is suggested by the
breadth of the features. Table 4.2 shows the line positions together with the upper
state lifetimes (r). These have been calculated from the widths (A v Hz)of the lines
using [28]:
These depend only on the upper state lifetimes as the lower 3p°4s4p 4S3/2 level is long
lived, and will not contribute to the width compared to the upper state. The upper
Line
1
2
3
4
5
6
7
Wavenumber
(cm-1) (vac)
15,751.80
15,638.82
15,623.57
15,493.64
15,365.34
15,302.61
15,044.48
Lifetime
(ps)
5.2
8.7
13
12
14
14
15
Table 4.2: The upper state lifetimes of the broad features.
levels are believed to have 4p2 and 4s3d configurations. Some attem pts at a multiconfiguration Hartree-Fock calculation have been performed by Froese-Fischer and
Brage indicating th a t the calculation is quite involved and will require a significant
effort to confirm or deny these lines as negative ion transitions [29].
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2 .5
T h e A rgon E xcim er
When plotted over the entire survey spectral region the broad features do appear,
primae facae, to be part of a molecular band (figure 4.1). The ground state of the
argon dimer is a very weakly bound
state, fferzberg’s constants for the argon
dimer notes only transitions observed from excited ungerade states to the ground state
[30]. Since the late 1970’s significant work has been done observing and characterizing
transitions in excited states [31] [32] [33] [34] [35]. The first electronically excited state
of Ar2 is 4s 3E J['1„ at 92,393.3cm-1 [30] [36]. Gary Eden has done absorption work on
transitions between 4s 3E+ state to higher mp 3n 9 states (5 < m < 15) [31]. These
transitions lie some lOOnrn above the broad features but provide inform ation about
what could be happening, specifically th at a yet unobserved II9 state could exist, and
form the final state of a transition from 4s 3£ „ l u, and be responsible for the broad
features. As well there is a 4s W+ state lying some 1349 cm-1 higher than 3E+1U,
which could also form the initial level, transm itting to higher lying 1,3ITS states [37],
Duplaa and Spiegelmann calculated the spectroscopic constants of the Rydberg states
of the dimer, agreeing well with experimental results [37]. No two calculated levels
form a transition that agrees with our experimental work to b etter than 500 cm -1 ,
although in general the calculated levels do come within a few hundred wavenumbers
of known transitions.
It is clear in figure 4.1 th a t there are two distinct bands, one including lines 1,
2, 4, 6, 7, and the other including the lines 3, and 5. It is assumed th a t the series
are complete due to the large amount of space bracketing them. It is not clear at
which vibrational level the series begins, but due to the strength of line 7, it will
be assumed to involve the lowest vibrational states. Figure 4.2 shows the possible
vibrational transitions responsible for the broad features. The spacing between lines
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
2 5 8 .1 3
0 1 .0 3
1 4 5 .1 8
2.98
8
o
6
o
eo
X3
<
4
?
0
15000
15250
15750
15500
16000
16250
16500
Frequency (cm ')
Figure 4.1: The broad features plotted over the spectral region, in the absence of all
other lines. All d ata was taken at an argon pressure of 145mTorr, microwave power
of 175W, and using a step scan of size 0.003nm.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 and 6, and 5 and 3 is almost identical, this is evidence th a t while these transitions
may originate on different initial levels, the final levels of transitions (7 and 5), and
(6 and 3) are the same. Since the energies of lines 5 and 3 are larger than 7 and 6,
the former must originate lower. This energy scheme is shown in figure 4.2. It is odd
th a t the first, excited vibrational state seems to be more highly populated than the
ground vibrational state in the lower electronic state, since it would be expected th a t
the ground vibrational state be more highly populated.
The vibrational constants of the 4s 'JE+ and the 4s JE+ states are u e = 296 cm ” 1,
and u ex e — 2.7 cm” 1 [37] [35]. The measured separation between the first and second
vibrational state in the lower level of figure 4.2 (320.9 cm” 1) falls outside the range
of separations given by the above listed u e, and
and 4s
(290.4 cm !), making 4s aE+
unlikely lower states. In fact no known state has vibrational constants
consistent with the separation 320.9 cm” 1 [37] [35]. This and the fact th a t the first
excited state (Figure
4.2) is more heavily populated than the ground state leads
to the conclusion th a t there are two separate ground electronic states whose ground
vibrational states are separated by 320.9 cm ” 1. There are no such states observed in
previous work. This is shown in Figure 4.-3.
The vibrational (u)e) and anharmonicity (ioex e) constants of the upper state can
be found. The equation describing the energy of a vibrational transition is [38]
G(v', v") = T ” - T'e
+io'e(v' + |) -
+ y)2
- w > " + 4 )+ c < ;" r" K + s )2
(4-2)
where T'e and T" are the upper and lower electronic states. It is clear th a t a vibrational
series originating from one initial vibrational state the equation may be simplified to
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
v 1o+4
7
6
4
2
1 Transition Number
5
3
12.98
v 'o+3
145.18
Upper
Electronic
State
v ' O+2
191.03
v ' +1
258.23
15365.34 cm'
258.13
v " +2
L ow er
V
E lectronic
v " +1
State
320.9
Figure 4.2: Possible vibrational transition scheme between two unknown electronic
energy levels, v" is assumed to be equal to 0 here.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Transition Number
V
' +4
O'
112.98
v ’ +3
o
145.18
,
V
Upper
’ +2
o
E lectronic
191.03
State
v 1+1
258.23
258. 13
First L ow er
v"+ l
E lectronic
State
" +1
Second L ow er
Electronic
State
Figure 4.3: Possible vibrational transition scheme between three unknown electronic
energy levels.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(for the purpose of fitting):
G (v',v") = Constants(v")
+co'e(z/ + | ) - w 'x '(z/ + ^)2
(4.3)
Taking the vibrational series of lines 7, 6, 4. 2, 1, and fitting them to equation 4.3
should yield the vibrational constants for the excited state. This is shown in figure 4.4.
The fitted constants (tue, ujex e) are shown in table 4.3 along with observed and
calculated line positions.
The large discrepancy between observed and calculated
C onstants
toe = 295 ± 1 1 cm-1
LUgXg = 24 ± 2 cm-1
E x p erim en ta l
sep aration
258.13 cm "1
191.03cm"1
145.18 cm "1
112.98 cm "1
C alcu lated
sep a ra tio n
247.11 c m "1
199.10 c m "1
151.08 cm "1
103.05 c m " 1
Table 4.3: Vibrational constants with experimental and observed spacing between
adjacent broad features of the same vibrational band.
separations should be noted. The series was fit dropping lines 1 and 7, respectively,
resulting in a worse fit, with average differences from observed numbers being four
times larger.
Reported ujgS of 291 to 310 can"1 for the 1,3n Ui9 states have been
observed [35] [37]. The fitted value for uoe falls within this range.
The we.xe’s of
the upper 1,3n UiS states are between 2.3 and 5.3 cm-1 [35] [37]; these values are an
order of magnitude smaller than the experimentally derived value.
A few of the
E states have uiex e's between 27.5 and 49.2, although the corresponding u e’s are
between 207 and 188 c n r 1. The experimentally derived constants are reasonable in
comparison to those previously determined for states of the argon dimer, however no
direct assignment can yet be made from the known literature [37].
The absence of rotational structure, and the symmetric, Lorentzian character of
the seven broad features is unexpected. While some transitions in argon have been
rotationally resolved [39] [40], many in the literature remain rotationallv unresolved
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15800
15700
Line 1 -
15600
Line 2
15500
Line 4
15400
43
15300
Line 6
15200
Line 7
15100
15000
0
1
3
2
4
5
V '+ I
Figure 4.4: Polynomial fit of the vibrational series. tue = 295 ± 1 1 cm \ ujex e = 24
± 2 cm-1
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
or appear to be made of a blend of rotational lines [34], although these are generally
asymmetric.
States of the argon dimer have rotational constants (B) in the range of 0.1336
to 0.1428cm-1 , with A B ’s of 0.01 c m '1 or less. From the equation for the energy
separation (F) of the Q-branch lines [41], an approximate value for the width of a
Q-branch can be made.
F = A 5 J(J+ 1 )
(4.4)
where ,J is the rotational quantum number. The equation defining the maximum J
value is
”™
I kT
1
um ~ V 2Bh~c ~ 2 “
o „ [T
1
V b “ 2‘
(
.
^
The maximum rotational number for known states is 27 at room tem perature which
is surely much less than the tem perature in the discharge. Assuming a ./ spread of
around 40, if we limit the Q branch to a width of 0.5 c m '1, which is an approximate
width for the broad features, A /I would have to be 0.00001cm-1 . Some of the A B ’s
between states reported by Duplaa are in the 10-5 range; an unresolved Q branch
with muted P and R branches in a E to II transition could explain the absence of
rotational structure and the symmetry of the broad features.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 5
Conclusions
5.1
C onclusions
An absorption spectral survey of an argon microwave discharge has been measured
using pulsed Cavity Ringdown Spectroscopy between 605nm and 677nm. 73 lines
in neutral argon have been seen. 13 of them not appearing on the NIST spectral
database [25], This is due to the fact that most spectral d ata in the database is from
emission experiments; the 13 lines were identified by generating a spectral list from
the energy levels of argon. Measurements were made on a discharge with a mixture
of argon and hydrogen as a search for argon hydride; no evidence of ArH was seen
in the wavelength region 605nm to 677nm. As well as the neutral lines, atmospheric
contam ination of oxygen, helium, and nitrogen were seen and identified. Seven strong
broad unidentified features were observed between 634nm and 664nm. Causes consid­
ered were contamination, unidentified neutral lines, doubly excited argon, the argon
negative ion and the argon excimer. It was concluded th at the features are most likely
vibrational transitions in the argon dimer, although attributing them to transitions in
the negative argon ion cannot be ruled out. Any rotational structure of the features
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
has not been resolved leading to the assumption th a t the transition must be E to
n , with an extremely narrow Q branch and muted P and R branches. Vibrational
level separations in the lower state (320.9 cm -1 ) are close to, but do not agree with
those reported in the literature for other A r2 states (290.4 cm-1 ). This and the fact
th at if there were a single lower state the first excited vibrational state would be more
populous than the ground state lead to the conclusion th at there may be two separate
lower electronic states whose ground vibrational levels are separated by 320.9 cm -1 .
The vibrational spectroscopic constants for the upper state have been calculated, and
are comparable to those of other Ar2 states (<ne = 295 ± 1 1 cm -1 , and ojex e = 24 ±
2 cm-1 ). Rotational structure was not observed in the features, which is comparable
with some published work.
While the dimer is the more likely cause of the broad features, the negative ion
cannot be rigorously ruled out. Absorption transitions from the inetastable 3/i’4.s4/;
453/2 state, to extremely short lived (lifetimes between 5.2 and 15 ps) upper levels
could be the cause of the broad features. A multi-configuration Hartree-Fock calcula­
tion has been attem pted; the calculation is very involved and will require a significant
effort to confirm or deny these lines as negative ion transitions.
W ith the sensitivity afforded by CRDS it is possible th a t many more unobserved
transitions in the argon dimer could be observed. Continuing the study of argon over
the entire visible spectrum could yield additional vibrational series thereby helping
to confirm (or refute) the possible observation of the argon dimer.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bibliography
[1] J.M. Herbelin, J.A. McKay, M.A. Kwok, R.H. Uenten, D.S. Urevig, D.J. Spencer,
and D.J. Benard. Sensitive measurement of photon lifetime and true reflectances
in an optical cavity by a phase-shift method. Applied Optics, 19(1):144—147,
1980.
[2] Richard Engeln, Gert, von Helden. Giel Berden, and Gerard Meijer. Phase shift
cavity ring down absorption spectroscopy. Chemical Physics Letters, 262:105109, 1996.
[3] Giel Berden, Rudy Peelers, and Gerard Meijer. Cavity ring-down spectroscopy:
Experim ental schemes and applications.
International Reviews in Physical
Chemistry, 19(4):565-607, 2000.
[4] Chuji Wang, Susan T. Scherrer, Yixiang Duan, and Christopher B. W instead.
Cavity ringdovvn measurements of mercury and it’s hyperfine structures at 254nm
in an atmospheric microwave plasma: spectral interference and analytical per­
formance. Journal of Analytical Atomic Spectrometry, 20:638-644, 2005.
[5] G.P. Miller and C.B. Winstead. Inductively coupled plasma cavity ringdown
spectrometry. Journal of Analytical Atomic Spectrometry, 12:907-912, 1997.
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[6] Yixiang Duan, Chuji Wang, and Christopher B. Winstead. Exploration of mi­
crowave plasma source cavity ring-down spectroscopy for elemental measure­
ments. Analytical Chemistry, 75(9):2105, 2003.
[7] A. O'Keefe, J.J.Scherer, J.B. Paul, and R.J. Saykally. Cavity ringdown laser
spectroscopy (CRDS) history, development, and applications. In K.W. Busch and
M.A. Busch, editors, A C S Symposium Series 720. American Chemical Society,
1999.
[8] Anthony O ’Keefe and David A.G. Deacon. Cavity ring-down optical spectrome­
ter for absorption measurements using pulsed laser sources. Review of Scientific
Instruments, 59:2544, 1988.
[9] R. K. Long. Preface. In Atmospheric absorption and laser radiation. Engineering
Experiment Station, Ohio State University; Columbus, 1967.
[10] John M. Herbelin and James A. McKay. Development of laser mirrors of very
high reflectivity using the cavity-attenuated phase-shift method. Applied Optics,
20(19):3341-3344, 1981.
[11] Dana Z. Anderson, Josef C. Frisch, and Carl S. Masser. Mirror reflectometer
based on optical cavity decay time. Applied Optics, 23(8): 1238—1245, 1984.
[12] Steven DeMille. Overtone spectroscopy comparing cavity ring-down (CRDS) to
other sensitive laser spectroscopies. M. Sc., University of Guelph, 2004.
[13] Dennis Wayne Tokaryk. Reaction Dynamics of Metastable Helium Molecules and
Atoms at f . 2 K. PhD thesis, University of Guelph, 1992.
[14] Lauren M acArthur.
Cavity ringdown spectroscopy (CRDS).
Undergraduate
Thesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[15] Galen Dunning. Cavity ringdown spectroscopy. Undergraduate Thesis.
[16] H. F. Tiedje, S. DeMille, L. M acArthur, and R.L. Brooks. Cavity ring-down
spectroscopy of transient 0 2-0-2 dimers. Canadian Journal of Physics, 79:773781, 2001.
[17] S. DeMille, R.H. deLaat, R.M. Tanner, R.L. Brooks, and N.P.C. Westwood.
Comparison of crds to icl-pas and phase-shift crds spectroscopies for the absolute
intensities of ch (6vCh = 6) overtone absorptions.
Chemical Physics Letters,
366:383-389, 2002.
[18] P. Macko, D. Romanini, and N. Sadeghi.
Saturation phenomena in cw cav­
ity ringdown spectroscopy, total density depletion by collisional transfers, 2001.
Presented at the Fontiers in Low Tem perature Plasma Diagnostics IV workshop.
[19] Rick H. deLaat. In Progress. Phd, University of Guelph, 2006.
[20] F.C. Fehsenfeld. K.M. Evenson. and H.P. Broida. Microwave discharge cavities
operating at 2450 mhz.
The Review of Scientific Instruments, 36(3):294-298,
1965.
[21] K.H. Becker, N.M. Masoud, K.E. M artus, and K.H. Schoenbach. Electron-driven
processes in high-pressure plasmas. The European Physical Journal D, 2005.
[22] Kunihiko Kasama, Takefumi Oka, Shigeyoshi Arai, Hiroshi Kurusu, and Yoshimasa Hama. Near-infrared absorption bands and kinetic behaviour of rare gas
excited diatomic molecules. Journal of Physical Chemistry, 86:2035-2042, 1982.
[23] Ehud Zamir, David L. Huests, Howard H. Nakano, Robert M. Hill, and Donald C.
Lorents. Visible absorption by electron-beam pumped rare gases. IE E E Journal
of Quantum Electronics, QE-15(5):281, 1979.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[24] A. Palmero, J. Cotrino, C. Lao, and A.R. Gonzalez-Elipe. Gas heating in lowpressure microwave argon discharges. Physical Review E, 66:066401, 2002.
[25] Nist
atomic
spectra
database.
http://physics.nist.gov/PhysR efD ata/A SD /index.htrnl.
[26] R. L. Brooks. 710 Atomic Physics. Guelph-Waterloo Physics Institute, 2002.
Course notes: Physics 710.
[27] T. Anderson. Atomic negative ions: structure, dynamics and collisions. Physics
Reports, 394:157-313, 2004.
[28] Anne P. Thorne. Spectrophysics. Chapm an and Hall, second edition, 1988.
[29] R. Froesse-Fischer and T Brage. Private Communication.
[30] K.P. Huber and G. Herzberg. Molecular Spectra and Molecular Structure IV.
Constants of Diatomic Molecules. Van Nostrand Reinhold Company, 1979.
[31] K.P. Killeen and J.G. Eden. Observation of ns 3E+(1U,0“ )—> nip 3n s Rydberg
transitions of Ar2 (n=4;5< m <15) and Kr2(n=5;6< m <16) in absorption.
Journal of Chemical Physics, 83(12):6209, 1985.
[32] K.P. Killeen and J.G. Eden. Gerade Rydberg states and ns 3E+(1U,0“ ) pho­
toionization spectra of the rare gas dimers (n=2-6). Journal of Chemical Physics,
84(11):6048, 1986.
[33] D.C. Shannon and J.G. Eden. Intracavity, laser absorption spectroscopy of Ar2
A nu5p and A au5p Rydberg states. Journal of Chemical Physics, 89(11):6644,
1988.
65
with permission of the copyright owner. Further reproduction prohibited without permission.
[34] D.J. Kane, S.B. Kim, D.C. Shannon, C.M. Herring, and J.G. Eden. Rydberg
states of the Ar2 molecule. Journal of Chemical Physics, 96(9):6407, 1992.
[35] Marshall L. Ginter and J. Gary Eden. Rydberg states of the rare gas dimers.
Canadian Journal of Physics, 82:762, 2004.
[36] P.R. Herman, P.E. LaRocque, and B.P. Stoicheff. Vacuum ultraviolet laser spec­
troscopy. v. rovibronic spectra of Ar2 and constants of the ground excited states.
Journal of Chemical Physics, 89(8) :4535 4589, 1988.
[37] P. Duplaa and F. Spiegelmann. A psuedopotential hole-particle treatm ent of
neutral rare gas excimer systems, ii. the rydberg states of the ar*2 dimer. Journal
of Chemical Physics, 105:1500, 1996.
[38] R. L. Brooks. Atomic and Molecular Physics. University of Guelph, A utum n
2002. Course notes: Physics 4120.
[39] C.M. Herring. J.G. Eden, and M.L. Ginter. The 5/7T 3n s t— a 3E+ and 5f a
3E^ <— a 3E ,| systems of Ar2. Journal of Chemical Physics. 108(13):5426, 1998.
[40] C.M. Herring, S.B. Kim, and J.G. Eden. Rotational analysis of the 7pa 3E+ <—
a 3E J syst em of the Ar2 molecule. Journal of Chemical Physics, 101 (6) :4561,
1994.
[41] Gerhard Herzberg. Molecular Spectra and Molecular Structure L Spectra of Di­
atomic Molecules. Van Nostrand Reinhold Company, second edition, 1950.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A ppendix A
Argon Line List
The following pages show the generated line list in its totality from GOOnm to 700nm.
This list was generated by taking the first X energy levels from the NIST database
and subtracting their energies from all of the higher lying levels th a t would generate
lines in the region Xcm” 1 to Ycm-1 . The lines generated from that process were
then tabulated with the lower and upper levels designated. Any line not meeting
the selection rules of parity and A J = ± 1 .0 was deleted. These selection rules were
chosen as parity is rigorous, and A J is nearly so [26], but usual coupling schemes are
poor in argon so no other rules were applied. A similar list was made for the argon
positive ion, with only one new addition. As no ionic argon lines were seen, this list
will not be included.
All of the new lines not shown on the original NIST spectral list are denoted as
* * * while all the new lines th at are not observed are denoted as *.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table A .l: T he generated line list of excited argon between 600nm and 700nm.
CO
<N
ll
II
ll
CN
in
CN
cn
CN
CN
CN
CN
CN
cn
in
II
II
Final Level
O
in
II
CN
cn
LJ
cn
C/3
C/3
00
00
*d
<N
o
CN
<N
CO
in
C/3
r-
xT
in
x)
VO
00
vo
CN
L— J
CN
in
<N
CN
II
o
o
o
o
Oh
Oh
m
in
ir>
CX
CX
Cl
cn
CN
co
cn
cn
CN
00
cn
cn
CN
CO
Oh
Pl
uo
CX
cn
m
Oh
m
P-,
>n
Ph
CO
cn
CO
cn
CO
cn
II
IIi
II
CN
CN
cn
CN
in
II
II
in
m
CN
CO
CN
CO
CN
CO
cn
II
II
II
II
CN
<N
CN
CN
CN
cn
cn
cn
in
H-s
1-^
CN
in
I—
■
> l—
J
o
Ph
in
CX CX CX CX cx
cn CNcn <Ncn CNcn cn
CN
CO
cn
cn
o
o
Oh
CX Cx
cn cn
cn
C N C N O C O -H C N O C N C O O
CN
II
o
<N
rT
CN
CN
CN
in
cn
cn
in
cn
cn
in
CN
CN
d
VO
CN
CN
T3
VO
din
VO
*d "d
vo
d
CN
Cl
CN
CN
CN
d
CO
©
c7 c7
cn
II
Oh
o
m
in
Cl
CX
CL
cn
Cl
0-
cx
rt*
Cx
Oh
CX cx
cn cn
cx
cn
m
■d
VO
C/3
CO
-H
©
Cl
©
cx cx cx
cx
cn
cx
cn
<N
c/3
Cl
Cl
d- d- Tf*
cx
cn
CN
C/3
©
CX
CX
cn
CN
C/1
cx
Cl
d-
©
CX
Cx
cn
CO
CO
CO
II
CN
CN
cn
CN
cn
CN
cn
CN
in
CN
cn
CL
d©
Cl
d©
CX
cx
cx
cn
cx
cn
cx
cn
cn
Cx
CN
©
CX
cx
cn
©
CX
cx
cn
TN
C/3
cn
cn
cn
cn
cn
On
On
ON
00
On
ON
in
in
CN
rin
o
o
VO
cn
cn
rr~
VO
CO
On
oo
cn
oo
CN
in
q
in
oo
o
Ov
t-H
J—
H
^H
o
VO
O
VO
o
•n
in
CN
o
VO
rCN
H
CN
cd
o
VO
Ov
it
VO
in
cn
o
vo
o
eq
cn
o
cn
CN
CN
cn
d
O
vo
cn
00
CN
cn
d
o
VO
din
cn
in
d
o
VO
00
Tf
ON
d
o
vo
o
VO
VO
O
o
o
NO
oo
00
cn
00
i-H
00
vq
CN
O
VO
ov
r~q
cn
O
vo
o
CN
cn
q
cn
o
vo
On
00
VO
m
ON
(N
O
OO
CN
CN
d
d
i>
o
o
VO
r —1i
o
VO
i t
<N
m
t-~
o
Ov
OV
h-H
VO
VO
VO
VO
*
*
*
o
o
o
cn
o
CN
it
r“H
On
cn
o
VO
d
o
VO
CN
cx cx cx
Tt
rf T}-
cn
i t
CO
II
cn
m
o
vq
_ i
CO
CN
cn
cn
dcn
in
in ^ m
CN
II II
cn
00
O
o
q
in
CN
II
CX
vo
in
<N
cn
O
VO
in
©
(N
cn
o
VO
CN
cn
cn
o
VO
CN
Oh
cn
CX
i t
CN
cn
cn
cn
o
cn
Line
nm (air)
c/3
©
Cl
O
Oh
cn
Position
nm (vac)
CN
C/3
cx
CL
Oh
Oh
CX CX CL CX CX CX cx
cn <Ncn CNcn CNcn CNcn CNcn CNcn
II
CN
o
Oh
CN
H-5
CN
o
Oh
CN
CN
C/3
Number
<N
CN
CX
Initial Level
CN
CN
o
in
II
II
o
o
i/)
CN
CO
CN
i t
II
o
(N
"d
CN
cn
d -
©
CN
OX
Tf
CN
©
CX
©
cx
cn
C/3
00
cn
cn
00
CN
o
CN
in
o
VO
cn
CN
rq
in
o
VO
o
r~
oo
in
o
VO
CN
cn
ov
•n
o
vo
VO
*
*
*
*
i t
in
o
VO
oo
oo
cn
O
vd
o
vo
Ox
CX cx
cn cn
N
cn
Ov
OV
cn
Cx
©
Oh
cn
o
CN
CX
©
C/3
o
cn
cn
cx
cn
i t
cx
cn
CN
CO
CN
cx
cn
VO
o
cn
Oh
CN
CX
CN
C/3
Cl
cn
CN
CO
i
Oh
vo
in
CO
cn
CX
o
t"
q
•n
o
vo
i-H
fH
cn
Oh
CN
C/3
cn
o
in
r-~
i t
vd
o
VO
o
•n
o
OV
CN
it
VO
vd
o
VO
vd
o
VO
< N c n T t - t / n v o i > o o o v O
-^ rd c n d -in v o r^ o o o N
t-H
t
*H
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
t
-H
CN
cn
cn
li
►
■X
CN
il
CN
II
II
<N
o
o
o
O
O
(N CN FT CN CN
cn
cn
cn
cn
o
i—
04
j
un
04
04
W)
-O
o
"O
r-
00
c/3
r-
CX
Ah
Dh
CX
a ,
cn
cx
cn
CX
cn
cn
cn
cn
y^.
y^.
CN
(N
CN
CN
CN
cn
cn
in
CN
II
li
cx
cn
cn
y^
<N
~o
03
r-
VO
f— 1
CN
II
o
o
FT
"O
Ph
cx
cn
cn
Oh
Ph
cx
cn
cn
Dh
cx
cx
cn
03
cn
cn
y^V
,
y^i
y^.
II
II
O
II
CN
(N
CN
CN
d
y^>,
il
11
CN
cn
03
03
03
03
cn
cn
cn
cn
cn
y^^
^
CN
II
CN
03
03
03
cn
cn
cn
y^!
y^i
y^^
CN
CN
cn
cn
cn
cn
II
CN
CN
CN
cn
cn
cn
CN
in
<N
CN
CN
Cn
cx
CX
Ah 04CX
04
in in
CX
cn
04
CX
cn
04
Cx
cx
©
CX
o
CX
in
cx
cn
CX
04
03
cn
cn
vo
rin
in
vd
o
VO
VO
VO
CN
Ov
vd
O
VO
VO
Ov
O
vo
vO
m
m
(N
rvd
o
vo
VO
Tt"
ov
VO
r"
o
ri
Ov
CN
O
o
o
vo
VO
*
*
*
—<
CN
CN
CN
cn
CN
o
cx
in
Cx
cn
03
o
cx
T}- Tf Tf Tf Tj-
cn
o
cx
cx
CX
04
CX
cn
04
CN
CN
cn
cn
cn
©
©
©
o
in
in
in
in
CX
Tj-
©
©
©
m
in
in
CX CX CX cx CX cx
cx
04
cx
CO
04
CX
©
—
CX
©
cx cx cx
03
03
03
03
03
03
03
03
03
03
03
03
cn
cn
cn
cn
cn
cn
cn
cn
cn
cn
cn
cn
y-H
cn
in
cn
o
Ov
in
in
cn
vo
in
o
vo
VO VO Ov
VO <n r VO oo
cn OV >—I
i-H i—l <N
^“H i-H
VO vo VO
r-~ VO
i-H
ro
vq
CN
<n
t —H i-H
VO VO
CN
r00
r-*
o
CN
Nin
Tf
i-H
VO
00
cn
CN
in
in
r-H
vo
00
in
in
rH
^■H
cn
CN
r-
t
-H
(N
Tf-
CN
Tt
Ov
t"
00
m
00
r~
rVO
Ov
VO
■'t
cn
O
oo
oo
Ov
iq
o
O
cn
cn
oo
Ov
o
vo
Ov
O
VO
o O
vo vo
^H
r-
cn
in
VO
CN
Ov
O
VO
>n
OV
*
*
*
*
*
*
vo rCN
©
CX
cx
nf
cn
cn
m
CN
o
CX
cx
03
cn
CN
Ov
O
VO
o
CX
cx
cn
03
vo
in
C~;
oo 00
O
o
VO vo
©
CX
cx
cn
cn
VO
Tj- Tt* Tf
cx
cx
cn
03
o
cx
CN
cx
cn
in
cx
cn
cx
cx
cn
04
CX
cn
Tf
(N
CO
cx
cx
cn
Oh
04
CX
03
o
o
cx
Oh
CN
cx
cn
<N
i“*H
00
Cx
r*-
CN
04
CX
in
o
II
CN
04
-a
Cx
cn
cn
11
04
cx
cn
cn
II
04
cx
cn
CX
cn
T T
cn
Oh
cx
cn
y^.
in
Oh
CX
cn
03
cn
cn
O h
cn
cx
FT
cn
Oh
Oh
cn
o
CN
vo
Oh
cx
o
cT
~o
Oh
O h
o
M
d
VO
<N
il
CN
04
in
vo
cn
T3
in
r-
CX
o
CN
~d
O
cn
r — '<
cn
in
Oh
Oh
CX
cn
y^.
CN
cn
Oh
(N
o
04
04
(A
r~ i
CN
in
cn
04
*d
o
CN
O
CX
cn
<N
»n
04
c/T
II
o
FT CN
cn
o
Oh
cn
o
©
CN
II
>—>
CN
(N
il
\\
04
04
<yf T3 "O
in in in oo in SO
o£
CN
o
il
FT
o
CN
o
II
>
—5
CN
o
FT
CN
II
CN
Ov
cn
cn
o
o
vo
00
CN
04
04
04
in
cx
cn
in
cx
cn
in
cx
cn
04
VO
i-H
<
o
i—
H
VO
04
04
o
d
i—H *—i
VO vo
CX
cn
H—1 o
Ov in
00
o CN
cn
it
O
VO
O
00
CN
vo
in
CN
cn
o
o
d
o
i-H
VO
i—
H
VO
o
i—
H
VO
i-H
vo
^H
VO
r-
CX
cn
04
vo
CX
cn
04
o
in
cn
CX
cn
04
r^
in
cn
CN
O
cn
04
CN
CN
i—
H i—
4
VO vo VO
CX
cn
04
CX
cn
04
Ov
CN
vo
CN
^H
vo
cn
CN
cn
cn
cn
cn
in
cn
vo rcn
cn
CX
cn
04
^-H
vo
o
CN
CX
cn
04
cx
cn
04
Tf rq cn
cn ■H"
i-H OV
VO
vo vo
*
*
*
*
OV
CN
cx
cn
VO
NO
in
*
*
*
00
cn
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ov
cn
o
CN
(N (N m
o
o
CN
cn
CN
ON
rT
cn
o
rT
in
o
CN
c/T
xj
in
r-
CL
CL,
CL
a.
cn
a,
cn
GO
CO
CN
cn
CN
cn
CN
(N
II
cn
CN
CN
II
(N
II
II
II
o
o
o
o
o
o
o
c
o
o
o
r™i o
CN CN CN (N FT r T rT rT rT rT rT CN rT
n
cn cn
cn
cn
cn i n
L J
t l
CN
CN
CN
CN
cn
CN) CN
CN
CN
orT
X3
x f x f X3 x f t T x f XJ X3 x f
xT
in in
in in in
VO in in
rVO in
Ov m
o
o
FT
in
CN
CN
CN
-a
in
cn
II
o
o
CN
wn
CN
o
o
CN
cn
d
in
X5
VO
xf
in
-a
m
CL
Cl
Cl
C L C l
FT
o
cn
CN
cn
cl
cn
CN
CO
cx,
cn
CN
CO
a.
cn
CN
CO
cl
cn
CN
CO
CL CL CL CL Cl
cn cn cn cn cn
CN
CN
CO
cn
cn
cn
cn
cn
cn
(N
11
CN
i-H
(N
(N
II
CN
II
CN
cn
CN
cn
CN
cn
CN
cn
II
II
CN
cn
II
CN
cn
C L C L C L C L C L C L C l
CO
cn
<N
CN
CN
CO
CO
cn
CN
CL CL
^
^
CN
CN
CN
O — O — O Cl
CL
CL
CN
CN
Cl
o
Cl
Tf
—
o
Cl
o
o
o
CO
CN
II
II
II
II
<N
CN
<N
i-H
CN
cn
CN
cn
cl
cl
CN
CN
CN
Cl
CN
CO
CL
N
CN
CO
CO
i—H cn
CN
i-H
in
m
in
vd
i-H i—
H
vo
VO
xf
cn
cn cn
rXf
xf
rH
r-;
vo
vo
OV
CM
oo
VO
vd
*“H
VO
CL
i-H
o
i-H
CL
CL
cn
o
CN
cl
Cl
^
CN
CN
^
o
—■
o
^
o
o
CO
CL
CN
CO
CN
GO
cn
cn
cn
vo
C\
i—H
Xf
cn
in
On
O
cn
l>
Ov
O
CN
VO
VO
i-H
vo
vo
r»H
in
o
Ov
CM
CN
CN
00
00
1—H
*“ H
vo
00
Xf
i—
H
vo
CN
CO
xl-
CO
CN
CO
CN
GO
CO
CN
xj-
in
xf
vo
xf
CL
CL
N"
Cl
CL
Tt*
CO
CN
CO
c
o
CL
n
CL
cn
N
CO
OV
00
IT)
r-
cn
Ov
vo
O
o
Ov
cn
cn
o
CM
CM
vo
cn
CM
't
OV
OV
CO
OV
in
in
't
VO
■vt
H
CM
Ov
CM
VO
VO
00
i-H
xf
<N
cn
cn
xf
xf
CN
CN
VO
oo
CN
cn
(N
cn
On
xf
CN
CN
cn
cn
00
r-
II
cn
in
CN
NO
r-
CO
o
r-~ ' t
't oo
■t ■t
VO
VO
CM
CM
vo
xf
in
vo
vq
vo
(N
cn
ri- H
r-H
in
CN
CN
NO
xf
CN
NO
in
xf
CN
VO
CM
CM
vo
in
cn
vq
o
in
cn
i-H
00
CN
VO
VO
CN
VO
*
*
*
Tf
II
cn
VO
CM
VO
II
cn
CM
CM
CM
Ov
xf
o
m
CO
cn
CN
cn
cn
On
in
CO
cn
CO
cn
II
CN
CN
cn
o
oo
CM
vo
CM
00
cc
©
Oh
in
Cl
cn
CN
CO
«n
Cl
cn
CN
CO
©
CL
CN
CO
cn
CN
II
CN
cn
in
CL
cn
CN
in
CL
<N
m
cn
in
xf
in
wn
in
CO
cn
cn
CO
r—
o
vo
r-
r~- m
vo
r-
oo
vo
ov
CM
VO
•N"
VO
o
oo
00
CM
VO
Ov
CM
VO
in
CO
cn
y
—
H
VO
cn
CN
©
CO
On
©
cn
vo vo
o
'It
m
o
Ov
O
—J
cn
vo vo
Ov
C3
cn
xf
xf
t-H
co
CL
CN
in
o
cn
VO
CL
CL
CN
CN
in
m
cn
CL
cn
cn
cn
cn
r-
Xf
OV
Cl
cn
CN
CO
Ov
CL
t
-H 00
OV
Xf
cn
vo NO
CM
CM
VO
CL
xf
o
©
Oh
CN
cn
vd
CM
VO
CL
CL
CL
XCL
f- X
l"
CM
in
o
in
OV
cn
vo
in
rin
xf
On
vd
cn
vo
VO
vd
cn
cn
vo
VO
vo
cn
cn
i-H
vd
cn
vo
cn
vo
cn
xf
in
*
*
*
*
*
*
in
CN
CN
CO
cn
II
cn
i-H CN
i—H x f
^H
vo
CN
CO
cn
CL
CL CL CL CL CL
cn cn CNcn cn CNcn
CN
(N
cn
*
*
*
cn
rN
CN
cn
CO
C L O h
CN
Oh CNCL CNOh CNOL CNOh CNO-i CNOh CNOh
CN
CN
CN
CN
CN
W)
lO
ID io> «/-» in
i/->
»n
«n
in in
CL CL CL CL CL cl Cl cl Cl Cl cl cl cl cl
cn cn cn cn cn cn cn cn cn cn cn cn cn cn
Cl
CN
CO
cn
cn
cn
C L C L C l
CL Cl
cn cn
CN
CO
CO
cn
CL
Cl
CN
CN
CO
cn
>—5
(N
CL CL CL CL
cn cn cn cn
<N
cn
CN
CL
^
CO
Oh
r~
in
oo
in
ov
in
o
VO
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VO
CN
VO
vo vo
cn
o
rT
cn
<N
<N
II
*■"->
o
CN
cn
in
<N
II
II
©
©
rT
VO vo
o
©
rT
bh
c/3
vo
rT
cn
04
CO
m
Ph
ir>
Ph
04
m
m
CN
©
rT
Ph
rT
CN
cn
in
cn
cn
cn
m
in
04
CN
in
04
W) -a
in
©
©
Ph Ph Oh Ph Ph Ph
in
in
m
*n
cn
il
H—j
<n
II
©
i!
©
ii
©
CN
II
II
©
©
rT
cn
r_,n
CN
CN
rT
CN
cn
>n
cn
cn
cn
in
cn
04
04
04
04
xf x f
vo in
C/^
r-
©
©
Ph
in
Ph
04
m
in
*n
rT
in
II
>
—>
~
CN
in
04
c/T
c/3
Ph
Ph
r-
©
Oh
CN
©
04
in
©
Ph
II
o
rT
04
XJ
xf
in
©
V)
CN
©
cn
in
CL,
ii
©
©
04
04
CO
II
rT
bh
©
II
CN
04
xf
in
©
Ph
cn
II
H-j
©
04
04
in
cn
cn
CN
il
h
■n”^
in
in
II
>—>
©
CN
in
04
xf
T3
N- in
©
P
CN
©
Ph
04
Ph
in
CL CL CL CL CL CL CL CL CL CL CL CL CL CL Cl CL CL CL CL CL Cl CL
cn cn cn cn cn cn cn cn . cn cn cn cn cn cn cn cn cn cn cn cn cn cn
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
04
00
C/3
C/3
C/3
c/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
C/3
cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn cn
04
^
CN
CN
^
^
CN
O
II
CN
II
O
II
CN
CN
CN
CN
CN
CN
CN
CN
CN
cn
In
cn
in
cn
^
II
II
CN
CN
CN
cn
cn
^
cn
CN
CN
cn
in
^
CN
CN
CN
CN
CN
cn
CN
II
O
11
cn
CN
CN
CN
CN
in
cn
cn
04
CL
CL
N"
N*
CL
CL
CL
N"
©
CL
N"
CL
CL
CL
N-
N-
Ph
Cl
Cl
CL
CL
N-
N-
Ph
Cl
CL
Cl
04
Cl
CL
N"
©
©
CL
CL
N-
—
CL
N©
CL
cn
Ph
CL
N-
—
©
CL
CL
N"
©
©
©
CL
CL
N*
CL
in
CL
N©
CL
04
00
cn
r-
i-H
00
cn
r-*
o
Tf
cn
vq
00
cn
VO
N-
<N
00
N-
vq
00
cn
vo
C/3
04
C/3
©
CL
04
m
IT)
in
C/3
C/3
cn
cn
cn
cn
cn
cn
cn
cn
cn
cn
CN
in
oo
00
i-H
Ov
in
N*
OO
Ov
cn
cn
Ov
VO
Ov
cn
in
in
NOV
in
^H
m
On
o
00 00 ov O n
cn
N
in
NN* HVO vo vo vo NO vo vo VO
CN 00
i—
H ov
i-H o
O
O
in
cn
t }vo
NO
H"
00
cn
i—
H
in
in
rm
*—
H
CN
in
in
vq
vd
cn
m
N’
m
VO
N-
Ov
VO
T—
H
in
VO
cn
in
VO
oo
i-H
OV
Ov
cn
in
VO
VO
o
Ov
vq
N"
in
VO
N" m
vo in
N"
O
00
O
00
^■H
VO
i-H
CN
cn
cn
O
v cn
cn
N- NVO vo
VO
On
cn
cn
oo
vd
00
rOV
o
00
Nvo VO
vo
cn
Ov
CN
00
N*
VO
Ov
00
i-H
in
00
N-
vo
o
cn
vo O Nr- On vo
in
O in
Ov O i-H
N- in i n
VO
NO
N" r r- cn
r-; CN
in
VO
VO
VO
00
VO
OV
VO
O
r-
CN
O
cn
r-
N-
in
r-
VO
r-
r-
r-
cn
C/3
cn
C/3
O CN NO
in
O ON
in
Hvo
vd
00
Ov
m
i-H
N"
VO
cn
00
vd
in
m
VO
•n
NO
r-
_
T-H
cn
cn
^-H
CN
cn
in
vo NO
OO
cn
oo
in
in
vo NO
NO
ON
in
NO
CL
H"
it
C/3
00
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
*
*
cn
t j-
oo
00
C/3
cn
cn
cn
NO
NO
NO
00
ON
in
NO
NO
On
m
NO
in
cn
On
00
m
NO
NO
NO
00
it
r-~ o
ON o
*
*
*
*
CN
00
CL
N*
in
*
*
00
C/3
cn
VO
f“H
*
o
CL
m
<n
NO
NO
H"
00 On
m
in
vo NO
*
*
Ov
cn
CL
N-
CL CL CL CL CL
cn 0 4cn 0 4cn 0 4cn 0 4cn
cn
*
*
oo
C/3
CL
>n
cn
in
vo vo VO
*
*
*
*
*
*
VO
C/3
CL
N©
CL
04
04
cn
NO
N"
04
cn
VO
VO
00
04
Cl
Tt*
04
CL CL CL CL CL CL CL CL CL CL CL CL CL CL .CL CL CL
cn 0 4cn cn cn cn cn cn cn cn cn cn cn cn cn 04cn 04cn 0 4cn
04
CL
in
00
vo
00
M
n
- < ^ n c n
—
ii
ii
ii
ii
o*'®
©
II
ii
© ©^ ©
o P ' I c n c n ^ h
II
ii
ii
® ©^ ©
ii
o
O
—
C N c N ’-1
It S?
^ ^
ii
II
*- O
—
ii
o
II
II
O
o ** o
O o
O O
rT <N FT rT rT
T
—
H
cn
cn
r 7 < N < N ( N l c 7 ( N ( N c N ' ^ C i < N c T ( N < N c 7 ^ r T
^ c j c n i n c n d i n c n d i n m m j d i
— c n c n c ^ d
<N
c/o
<N
“O
<N
“O
1 <N
“O
</3
<N
"O
c T ” ' <N
"O
(N
"O
go
(N
“O
r d "1 ^
*0
<N
co CL
c V ” * <N
*0
co^
in - *
"O
CN
co
m
—
o
—
—
—
m m
IN
cn
^
CN
CO
C/D
cn
^
CN
00
cn
cn
c N c n c N d
II
II
►
—■>
II
*— J
N
C
CL
>— 5
N
C
O-i
C
—
m
—
m
CN
<y}
CN
CO
cn
cn
cn
—
in
—
—
c n c n c n t -h
H -j
t— >
t-n
*“ >
i— i
cn
II
II
•— >
CO
II
N
C
—
N
C
N
CL
N
C
N
O- i
O
J
C
m m
C
N
Cl
N
—
m m
—
N
C
II
N
C
IN
CO
cn
^
II
N
CL
C
m
cn
CO
cn
II
CN
CO
II
)— }
CN
CO
cn
—
N
C
—
C
N
CL
N
C
—
C
N
O^
C
C
N
CX,
^
f
N
m m
C
N
ii
f
CL
C
N
—
s
j
CL
C
in
i
C
CN
CN
d
in
T3
in
CS <S CN CS
m
II
c
m
m
m
m
N
C
N
n
i n
C
N
C
i n
N
m
C
N
i n
C
N
i n
C
N
m
C
N
i n
C
*
N
n
C
N
i n
C
N
i n
C
—
i n
N
C
—
d o i o c N
II
N
C
N
n
i n
C
II
cn
^
II
t— >
C
O-i
l
m
N
CN
CO
CO
cn
^
CO
CN
CO
CN
CO
CN
CO
—
—
cn
►—j
h -^
*— )
cn
^
II
t— j
H —j
cn
^
II
II
h—j
cn
cn
^
II
II
^ ^ n c N c T c N r T ^
- H c n m ^ H c n ^ c n c n
C
N
CO
cn
^1
f
N
—
N
i n
C
N
C
N
CL CL
C
—
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
C U C L h CLh CLh Ph C U C L Q h Qh Ph Qh Qh C L O h Ph
C
—
cn
^1
I— j
co
N
m
CO
cn
^ , - ,
II
t— >
CN
CO
cn
T f N c T c T r j f N
— c n u n c n u n ^ ^ m
CX,
C
IN
CO
CO
II
^ ^ ^ ^ ^ N ^ ^ c
c n i ^ u n —^ c n c n
C
—
CN
d
un
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
Q-i P ~ ( CL i Cu Cu CLi C U C U^ D - i ^ C U^ P h ^ Oh & - ( C U
O h <NCu <NPh <NP-<i CvJC U QIN - i CND-i
^ (N
in
in
n
in
in
in
in
irP^ in
in
in
m
in
in
in
in
in ^ in
in
in
in
in
m
n
in
in
CL CL CL CL CL CL CL CL CL CL CL CL CL CL
CL CL CL CL CL CL, c l
cn cn cn cn cn <Ncn CNm CNcn cn cn cn cn cn INcn CNcn CNcn cn CNcn cn cn cn
<N
GO
o
CN
tT
t ' - ^ T f i n N o i n i n N O T t - i n ^ r ^ ^ i n C ' ^ T t - N O
o
Q-i
in
CL
cn
CN
7i
11
>—3 *—
>
CN
II
J= 3
n
h
N
i n
C
N
C
N
CL
N
C
N
C
m m
o
Ph
i n
»
C
C
N
N
C
i n
N
C
N
C
CL
C
. m
o
o
O- ( Oh
n
N
CL
N
C
—
C
i n
o
Qh
N
C
*
n
N
C
N
CL
CL
Cl
N
N
^
C
m
C
C
—
o
o
CLOh
N
1
N
i n
C
N
—
PL
i n
cl
cl
cl
cl
cl
cl
cl
cl
cl
cl cl
cl
cl
cl
Cl cl
cl
cl
cl
cl
cl
o -,
c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n
N
C
C
O
N
C
c
O
N
C
C
O
N
C
(
O
N
C
<
O
N
C
C
O
N
C
<
O
N
C
(
O
N
G
<
O
N
C
C
O
C
N
C N
O C O
t
N
C
(
O
N
C
(
O
N
C
(
O
N
C
C
O
N
C
r
O
a
C
<
O
N
C N
( N
C N
C O C O C O C O
c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n c n
00
rNO
H
o
0 0
N-
cn
in
00
in
00
rcn
ON
in
NO
o
NO
NO
d
NO
NO
cn
in
l>
o
o
m
o
d
NO
NO
r —(
cn r00 NO
in
NO
o
o
NO
NO
NO
NO
NO
NO
VD
in
<N
VO
VO
Tt
OO
o
CN
cn
vo
VO
VO
r —i
ON
cn
cn
NO
NO
VO
VO
00
CN
OV
vq
IT)
VO
VO
r~
rVO
o
vd
VO
VO
ov
VO
VO
cn
—H
t ;
Tf r~ fH
in
00 00 CN
mi
vd
N-’
OV
VO
vo
VO
VO
vo
vo
in
o
ti vo
vo
VO
cn
o
CN
t I-
i- H
— o
00
■
'3
- <N
-it
00
m
Ov
ov
oo
oo
00
00
CN
CN
vd
VO
vo
VO
VO
VO
VO
NO
NO
■'t
00 o
»n o
<n
ov
cn
vo
CN
m -'dcn
vo
Ov vq
0 0
OV
vd
vo
vo
vo VO
vo vo
CN
0 0
VO
vo
*
vo
vo
VO
VO
VO
vo
vo
vo
^H
r~
0 0
oo
oo
ov
CN
ON
OV
VO
VO
ov
vo
vo
NO
m o
o CN
oo
cn
— O
Ov O
VO r-
vo
vo
Tt
p-
cn
CN
r-
o
o
t-~
VO
i— H
r^
cn
r-
o
’"I
CN
t>
NO
CN
rcn
o
ON
OO
CN
CN
rNO
cn
oo
CN
in
rNO
in
rNO
NO
x>
ON
cn
r> NO
r-
CN
cn
NO
vq
in
r-
vo
00
CN
o
NO
oo
CN
in
in
NO
NO
NO
in
r~
VO
t>
r-
r-
*
*
*
CN
*
*
*
*
*
*
OT-H(NcnTj-wnvor-oo
t ^ o o o N O ^ - < c N c n T j - i n N O t ^ o o o \ 0 0 0 0 0 0 0 0 0
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-H
CN
O
il
II
CN
II
CN
II
cn
II
(N
<N
II
o
CN
li
(— 5
~
CN
CN
CN
CN
cn
-a
o
P
L.
CL
cn
C/3
cn
CN
o
CN
cn
d
d-
in
CN
CN
d
vo
o
d
in
o
d
in
o
d
Tf
o
O
CN
CN
CN
in
in
in
d
d-
d
d -
O
o
CN
o
o
CL
cn
C/3
cn
CL
cn
CL
cn
CL
cn
CN
C/3
CN
C/3
CN
C/3
cn
cn
cn
CN
CL
cn
CL
cn
C/3
CN
C/3
cn
cn
CN
CL
cn
C/3
cn
CN
II
CL
cn
CL
cn
O
CL
CN
CN
C/3
CN
C/3
CL
cn
CL
cn
cn
cn
cn
II
II
CN
CN
CN
II
O
CN
CN
CN
in
cn
i—i
d
n-
d- in
d
m
CL CL CL CL <NCL CNCL CL CL CL CNCL
CN
CN
cl
O
O O O
CN
CN
Oh
Cl
cn
Ph
Ph
CN
CN
d
Oh
CN
CL
cn
CL
cn
CL
cn
cn
cn
cn
CN
CN
C/3
CN
r—
n °
CN
CN
CN
cn
CN
CN
cn
cn
CN
CN
CN
CN
CN
cn
cn
cn
CN
cn
CN ?N
P. in
CN
CL
CL
CL
CL
CL
d* d- Tf d“ dO
—
CL
CN
CL
cn
O
CO
Oh
CL
cn
o
—
CL
Cl
cn
o
CL
Cl
cn
CN
C/ 3
CN
C/3
CN
C/3
CN
C/3
cn
cn
cn
cn
cn
d-
p
H CN
—1 V
D
Ov
00
VO
CN
Ov
Ov
CL
d"
CL
CL
o
CL
CL
CL
CL
cn
c/3
cn
CL
cn
CL
cn
OV
CN
CL
(N
CL
cn
so
rso
Os
0
cn
cn
O
Os
SO
C
N C
N
00
00
so
so
VC
vo
N"
00
f~
s
—
H
0
1-H
(N
cn
(H
y— i
p^H
p—H
i-H
p—H
?—H
CL
cn
(N
CL
cn
CL
cn
C/3
cn
C/3
cn
cn
cn
cn
OV
cn
00
m
Ov
00
0
d* Ov
0
ir-H 0
d-
00
00
00
CN
P“H
OO
O
r-
0
os
r~—1
00
CL
CN
i/ “3
rrcn
10
00
vo
00
vo
00
00
SO
lO
s o
1—“ 4
^H
p—H
^H
i-H d"
cn
CN
C/3
00
00
00
CN
C/3
CL
CL
CN
00
10
00
00
CN
OV
CL
CN
CL
cn
C/3
cn
cN
CL
CN
Cl
cn
C/3
cn
d- rvo
vc
cn
Os
in
OS
vo V
C vo V
C so
00
00
Os
oo
00
00
SO
r-H
^■H
SO
SO
0
CN
Os
SO
CN
OS
SO
00
OV
i- H
1— i
t—H
p—H
0
0
Os
00
CL
CN
c/3
CL
d- n- d- n- n- n-
C/3
10
00
os
CL
CL
cn
00
VO
t"-
CL
CN
C/3
00
V
D
t>
CN
lO
vo v d
0
0
00
00
00
rr- rV
C V
C vo vo
C vo V
C V
O
O'
CN
cn
CL
CN
00
m
Os
cn
Os
SO
m
Os
cn
cn
in
Os
SO
CL
CN
Cl
cn
C/3
cn
CL
CN
Cl
cn
C/3
cn
O O
in
CN
cn
T jIT )
so
Os
SO
OS
O
O
t''
H
CN
SO
OS
SO
SO
SO
T—H
in
cn
r-~
sd
os
SO
CL
CN
cl
cn
C/3
cn
(N
f-H
CN
(N
OS
Os
SO
r-"H
d- it
as
os
so
p—H
C
N cn
O
C
N C
N C
N C
N C
N
^H
^H
r—
H
p -H
p h
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 891 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа