University of Southern California Dissertations and Theses

Development of a 20-femtosecond tunable ultraviolet laser source towards study of the photochemistry of liquid water

(USC Thesis Other)

Development of a 20-femtosecond tunable ultraviolet laser source towards study of the photochemistry of liquid water

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content DEVELOPMENT OF A 20-FEMTOSECOND TUNABLE
ULTRAVIOLET LASER SOURCE TOWARDS STUDY
OF THE PHOTOCHEMISTRY OF LIQUID WATER
by
Askat Jailaubekov
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(CHEMISTRY)
December 2003
Copyright 2003 Askat Jailaubekov
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY O F SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This thesis, written by
/j$kai J~a i I & u $e ko v
under the direction o f h thesis committee, and
approved by all its members, has been presented to and
accepted by the Director o f Graduate and Professional
Programs, in partial fulfillment of the requirements fo r the
degree of
Master of Science
Director
Date, D ecem b er 1 7 . 2 0 0 3
Thesis Committee
j ' 4 Chair
v i i O W v
V'iAali
/ f
< ]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Contents
List of Table and Figures......................................................................................... iii
A b stract ................. vii
Chapter 1. Generation of femtosecond UV pulses................................................ 1
1. Introduction ......................................................................................1
2. Phase matching in hollow waveguides ...... 3
3. Third harmonic generation (TH G ) ....... ........10
4. R esults ....... ...12
5. Discussion ...... .17
6 . References for Chapter 1 ..........................................................................21
C hapter 2. Femtosecond UV laser so u rc e .................................................... ..22
1. Introduction ........................ 22
2. Installation.................................................................................. .....23
2.1 W avelength durability............................................................. 23
2.2 Experim ental ............29
2.3 Results ........35
2.3.1 Power stability, shot-to-shot noise, m ode .............. 35
2.3.2 S eed power dependence ..................................... 36
2.3.3 Performance and setup for 240 nm ...........38
3. Ultrashort pulse characterization. ...... 40
3.1 Spectral broadening............................................................ 40
3.2 Method of pulsewidth determination .................. 44
3.3 Autocorrelation setup ...... ...47
3.4 Results ...... 50
3.5 Discussion .............54
4. Summary ....... ....56
5. References for Chapter 2 ................................................................... ...... ..57
C hapter 3. Liquid w ater photochem istry ...... 59
1. Background ............... 59
2. Preliminary UV pump-UV probe results................ 62
3. References for Chapter 3 .......67
Bibliography......................... . . . . 6 8
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Table and Figures
Table 1.
Figure 1.1
Figure 1.2
Figure 1.3
Figure 1.4
Figure 1.5
Figure 1.6
Figure 1.7
Figure 1,8
Inventory of parts to build UV hollow core fiber source.
a) Difference-ffequency four-wave mixing process, b) For example,
four-wave mixing of second-harmonic light 2 © with fundamental to
from an amplified Ti:S laser. Generated UV light appears at third
harmonic 3© of the laser (near 267 rnn).
Calculated transmissions of the EHi i fundamental mode and that of the
subsequent, linearly polarized, high-order modes EH n (wn-5.5201) and
EHi3 (ui3=8.6537) for a capillary waveguide with a diameter 150 pm, at
800 nm, as a function of the capillary length (in cm).
Calculated transmissions of the EHn fundamental mode for a capillary
waveguides with the diameters 75, 100 and 150 pm, at 800 nm, as a
function of the capillary length (in cm).
Calculated EHn mode propagation in fiber with diameter 75 micron at
three different wavelengths 267, 400 and 800 nm, as a function of the
capillary length (in cm).
Schematic layout for THG. Fundamental beam (©) and second
harmonic (2©) are mixed to generate third harmonic (3©). In reality,
separate focusing elements are used for © and 2© beams. Here, BBO is
a nonlinear crystal, and HWP is a half-waveplate.
Pressure tuning curve for 150-micron fiber, 20 cm long. Solid line
represents experimental curve of the 267 nm light intensity. It is fitted
with calculated phase matched signals for the 6 lowest-order modes,
assuming that all driving beams are propagated in EHn •
Calculated phase-matched optimum pressures for three hollow core
diameters. Maximums are located at 60 torr, 140 torr and 250 torr for
150, 100 and 75 micron fiber, respectively. Note, the intensities are not
to scale.
Pressure tuning curve for 75-micron fiber, 30 cm long. Solid line
represents experimental curve of the 267 nm light intensity. It is
compared with the calculated phase matched signal (dash line),
assuming that all beams are propagated in EHn.
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2,5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 2.9
Cascaded mechanisms. Generation of 4® and 5© harmonics via
DFFWM. Here, co=8 GOnm, 2©=400, ©=267, 4©=200. First mixing 3®=
2(0+2©-®. Cascaded mixing, 4©=3©+2©-®, or 4©=3©+3c o -2c q . 5© can
be generated by a similar two step cascaded event.
Signal/Seed wavelength dependence. 1 /^uv_S ignai-2 /7 . 4 o o n m - 1/7-ir.
Wavelength units are presented in reciprocal scales. Right-hand side
scale shows the energy of the UV photon in eV. For example, if the
wavelength of the seed is 1.2pm the wavelength of the output is 240 nm
or 5.17eV.
Linear dependence of the optimum pressure on seed wavelength (square
dots). Phase-matching conditions were calculated assuming that all
beams are in EHn mode, propagating in the hollow core fiber with 75-
micron core diameter. The upper curve linked to RHS scale shows
UV/Seed wavelength dependence. For example, if Lseed ~ 1000 nm the
needed optimum pressure is 295 torr, the generated UV output appear
-250 nm.
Signal/Seed dependence in terms of different options of the OP A
output.
Optical layout.
Argon-gas chamber and pressure controlling system.
Dependence o f the UV output power (267 nm) on the seed (800nm)
intensity. The lower curve (triangle dots) represents “bad” alignment,
the upper curve - “good” alignment (square dots). Experimental data
fitted with first order exponential decay.
Spectral broadening of the 240 nm signal, FWHM is equal to 4 nm. The
fiber (75 micron, 35 cm long) was pumped by 45 mW at 400-nm and
seeded with IR pulses at 1200 nm, 10-20 mW power. The time
bandwidth in brackets shows transform-limited pulse duration for a
particular spectral bandwidth, assuming Gaussian spectrum.
Spectral bandwidth of 800-nm (a) and 400-nm (b) beams. For both
graphs the solid line represents the spectra of pulses emerging from
fiber, dash line is a spectrum of input beams The output bandwidth
shown are collected with the time delay set to generate UV signal in
DFFWM, i.e. there is a depletion due to wave-mixing. All data
collected for 35 cm length fiber and 75 micron core diameter. The time
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
bandwidth in brackets shows transform-limited pulse duration for a
particular spectral bandwidth, assuming Gaussian spectrum.
Output spectra of third harmonic generated with hollow core fiber (solid
line) and non-linear BBO crystal (dash line). The fiber parameters: 35
cm, 75 micron core. Crystal thickness 80 pm. The time bandwidth in
brackets shows transform-limited pulse duration for a particular spectral
bandwidth.
Output spectra of UV signal at three different relative delays of blue
(400 nm) and red (800nm) pulses. Experimental data (square dots)
represent time-zero delay, indicating 7.5 nm FWHM bandwidth. Blue
(triangle dots) and red (round dots) curves correspond to cases when
blue pulse precedes red pulse or vice versa.
Output spectrum of the UV signal emerging from 150-micron-core
fiber, 20 cm long. The intensities of 400-nm beam and 800 nm were
76mW and 42mW, respectively, assuming that these powers were
actually enter the fiber. FWHM about 10 nm can be extracted from this
graph, which correspond to lOfs transform limited relation.
Optical layout for autocorrelation setup.
Absorption intensity autocorrelation trace in a fused silica sample.
Assuming Gaussian pulse shape, the pulsewidths of 500 fs can
estimated.
Experimental autocorrelation trace (solid line) in fused silica after
compression, Prism separation is 28.5 cm . The experimental curve was
fitted with Gaussian function adjusting the height and the width (dotted
line). Thickness of silica sample is ~3 mm long. FWHM of the
autocorrelation trace is 41 fs, pulsewidth if fitting as Gaussian is 29 fs.
Experimental autocorrelation trace (solid line) in liquid water after
compression, Prism separation is 28.5 cm. The experimental curve was
fitted with Gaussian function adjusting the height and the width (dotted
line). Thickness of liquid water jet 30-50 pm.
Primary events in the first picosecond. By analogy with the gas phase
data it is assumed that within - 1 0 fs the water molecule dissociates into
H and OH. First two fragments of photo-ionization channel are initially
a positively charged hole H2 0 + and e". reacts quickly ( ~ 1 0 0 fs)
with another water molecule to form H3 0 + and OH fragments. After a
period o f ~ 1 ps the system is fully relaxed and a certain ejection length
separates the charged hole (now H3O4 ) and trapped e'aq. This separation
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2.10
Figure 2.11
Figure 2.12
Figure 2.13
Figure 2.14
Figure 2.15
Figure 2.16
Figure 3.1
Figure 3,2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
depends on the photo-ionization energy. The question sign between two
main channels represents other possible mechanisms o f formation of the
solvated electron and accompanying products. For example, it has been
suggested there exists a dissiociative ionization o f excited water
molecule (H(hot) + H2O -» HsO+ + e"aq ).
The static absorption spectra of e'aq, OHaq, Haq, and OH~aq from 200 to
1200 nm (Ref. 2). Notice that the absorption spectrum of the solvated
electron has been reduced by a factor of 10. For example, the extinction
coefficients for the solvated electron and OH radical at 267 nm (4.65
eV) are -1000 M^cm"1 and -400 M ^cm '1 , respectively.
Pump-probe technique to study femtosecond transient absorption
spectroscopy of pure liquid water.
Experimental data for the geminate recombination o f solvated electron
after two-photon excitation of pure water. The pump and probe
wavelengths are 300 nm and 700 nm, respectively.
The first picosecond of the transient absorption observed at 267 nm.
Here (1+1) near t = 0 corresponds to one photon pump + one photon
probe process, (2 + 1 ) is the two photon excitation followed by one
photon probe. In the experiment the pump and the probe powers were
-200 nJ and -20 nJ, respectively. The autocorrelation trace reveals the
FWHM of 8 8 fs, which corresponds to 62 fs pulsewidth. The measured
value for the maximum is 230 pV.
The pump-probe absorption signal for the time delays from -200 fs to -
130 fs and from 200 fs to 3 ps. The former signal, where the probe
precedes the pump pulse, serves as a background level. The
autocorrelation trace is deliberately omitted to contrast the two signal
levels. The signal is enlarged in order to demonstrate the small transient
absorption effect.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
The dissociation or ionization of a water molecule in liquid water is amongst
the most fundamental events in aqueous solution chemistry. Because of the ultrafast
time scales (10-100 fs) involved, the dynamics have not been directly probed by
experiment. There is much theoretical discussion on this issue, however, the
mechanistic details remain rather controversial. Femtosecond spectroscopies provide
a route to explore liquid water dynamics, however, typically to initiate such small
molecule system requires -10 fs pulses in the “photochemical” ultraviolet (UY)
region (<300 nm), a difficult spectral range for femtosecond sources. The
development of ultrashort-pulse light sources in the short-wavelength region of the
spectrum is therefore a matter of great importance. In this work we demonstrate the
implementation of a unique femtosecond UV laser source for spectroscopy studies in
the condensed phase currently with 20 fs time resolution.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 1. Generation of femtosecond UV pulses
1. Introduction
Tremendous progress has been made in the development of high peak-power
ultrafast lasers over the last decade. Lasers are now capable of generating terawatt
peak powers with 25 fs pulsewidths in the near IR [1-3]. However for many
applications femtosecond pulses with energies of more than a microjoule (gigawatt
peak-power) in the UV are required. This is a spectral region o f great interest for
studies o f ultrafast chemical dynamics in small molecules. Taking advantage of
powerful IR lasers the generation of ultrashort-pulse light in the UV region is almost
exclusively performed currently in nonlinear crystals using frequency conversion
techniques such as second harmonic generation (SHG) and sum-frequency mixing
(SFM). For achieving high conversion efficiencies of one wavelength to another
these processes should be phase-matched, i.e., both fundamental and generated lights
must travel with the same phase velocity in nonlinear medium. Otherwise, there will
be destructive cancellation of generated light. If the phase-matching condition is
satisfied the generated light is reinforced and in principle, its intensity scales with
length of crystal. This allows to obtain up to several percent conversion efficiency in
the deep UV [4,5]. However, for spectrally ultrabroad femtosecond pulses (<50 fs)
overall velocity of the pulse envelope, or group velocity, should also be taken into
account. For instance, even if the phase velocities are matched due to phase
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
matching, the group velocity o f the different pulses differ greatly especially in UV
region, which leads to temporal walk-off of the fundamental and the generated
pulses. Thus, group velocity mismatch (GVM) mostly determine a pulse broadening.
It is also worth mentioning that because of the spectral dependence phase-matching
condition has bandwidth limitations. As crystal becomes longer the spectral
acceptance becomes smaller, and as a result, the spectral bandwidth that can be
doubled gets limited. All these disadvantages are related with length of the nonlinear
crystal. It implies that in order to avoid the problem of the generated pulse
broadening very thin crystal should be used. As it was previously mentioned, that
results in low efficiency. To summarize, nonlinear crystals are practical in the visible
and infrared region, whereas UV frequency generation occurs with undesirable
tradeoff between pulse broadening and intensity. Finally, for inorganic crystals
optical transparency typically cuts off at about 2 0 0 nm; certainly, by this wavelength
it is unlikely that one can achieve 1 0 fs pulses using nonlinear crystals.
Frequency conversion using gases as a nonlinear optical media can resolve
the problem of the pulsewidth limitations in the UV [6 ], Moreover, many gases are
transparent well into the vacuum ultraviolet (VUV), and even shorter wavelengths
can propagate with moderate absorption in a low-pressure gas. However, the
conversion efficiency in gases is relatively low, because of the short interaction
length and poor phase-matching. To obtain high-efficiency conversion, Mumane and
co-workers have demonstrated phase-matched generation in the UV by use of guided
waves in a long argon-filled hollow core fiber (capillary waveguide) [7]. Using a
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
four-wave mixing process, 3© = 2co + 2© - to (Figure 1.1), they were able to
generate sub-10 fs pulses with energies greater than a microjoule [8]. Since our lab is
equipped with a commercial laser system (Hurricane, Spectra Physics) generating
100 fs pulse, which has 5 times smaller peak-power intensity than Mumane’s home-
built multi-pass amplifier generating 20 fs pulses, the primary question was whether
it is possible to implement such a technique in our lab and obtain comparable
intensities and bandwidths in the UV. In this chapter we demonstrate that this
method can be efficient and suitable even with standard laser systems. In order to
establish this method as a new UV source, to carry out non-linear spectroscopy,
significant improvement for the robustness and intensity of the apparatus required,
these steps are detailed in this Chapter.
2. Phase matching in hollow waveguides
Hollow core waveguides differ significantly from conventional optical fibers:
hollow dielectric waveguides (capillaries) guide laser beams through Fresnel
(grazing incidence) reflections at the inner wall of the capillary instead of by total
internal reflection. Whereas the mode structure is similar to that of a step-index
optical fiber, all the modes are inherently lossy because of partial transmission at the
walls of the capillary [9]. The throughput of the hollow core fiber, the ratio of the
output to the Input intensities, is given by:
( 1)
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a)
ft
pump
©signal ©pump"'" ©pump " ©seed
b)
'' c o (seed)
f
2co(pump) 1
A
3co(signal)
2 © (pump) J ,
3®=2®+2© -G )
Figure 1.1 a) Difference-frequency four-wave mixing process, b) For example,
four-wave mixing of second-harmonic light 2co with fundamental ©
from an amplified Ti:S laser. Generated UV light appears at third
harmonic 3® of the laser (near 267 nm).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The field loss rate a n m (or attenuation coefficient) for the mode with azimuthal and
radial modal indices m and n in a capillary of radius a is
a„m =
nm
O 2 ' ? v 2 +i
\ 2 tz j
n
a
V =
v - I
glass
n air
(2)
where un m is the modal constant, X is the wavelength, v is the ratio between refractive
indices of the external (fused silica) and internal (gas) media. The lowest-order mode
in a capillary waveguide is the linearly polarized EHn mode and its modal constant
wn=2.405. The mode EHn is closest to the free-space TEMqo Gaussian mode. As it
is seen from Figure 1.2, for sufficiently long waveguides the mode discrimination is
very high so that only the fundamental mode EHn can propagate. The strong
dependence of the transmission from the capillary radius and the laser wavelength
(Figure 1.3, 1.4) should also be considered when capillary length is selected.
By countering diffraction of a wave, a waveguide adds a geometrical
component to the wavevector [10]. The propagation constant for a capillary filled
with a medium of index n is given by
2 nn( X )
X
1
2
X
\ 2
2 n a
(3)
2 n 2 n
~~X~
P 8 (X ) - -
2
nm
X
An a "
In the approximate form for propagation constant above, the index of refraction for a
gas is written in the form n(X) = 1 + P8(X), where P is the pressure of the gas, S(X)
contains the gas-dispersion function. Thus, in this approximation the k-vector now
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
% (8 0 0 , 1.45332 , 1 5 0 , 1, 8.6537 )
150 jam, 800 nm
0.9
0.8
E H ,
0.7
0.6
0.5
0.4
0.3
0.2
t i i
o .i
0
0 40 60 80 1 0 0 20
1 0 0
Figure 1.2 Calculated transmissions of the EHn fundamental mode and that of the
subsequent, linearly polarized, high-order modes EH 12 (wi2=5.5201) and
EH b (wi3=8.6537) for a capillary waveguide with a diameter 150 pm, at
800 nm, as a function of the capillary length (in cm).
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
EH
150 /im
0.7
.4048 ) ° ’6
100 , 1 , 2.4048 ) Q5
100 /dm , 75 , 1, 2.4048 )
0.3
0.2
75 /dm
o
100
o 100
Figure 1.3 Calculated transmissions of the EHn fundamental mode for a capillary
waveguides with the diameters 75, 100 and 150 pm, at 800 nm, as a
function of the capillary length (in cm).
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75 f i m , E H
0.7
, 75 , 1, 2 .4 0 4 8 ) °-6
267 nm
^ 0.5
400 nm
0,4
0.3
0.2
800 nm
o
1 0 0
100 0
Figure 1.4 Calculated EHn mode propagation in fiber with diameter 75 micron at
three different wavelengths 267, 400 and 800 nm, as a function of the
capillary length (in cm).
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
consists of three sources of dispersion: vacuum, gas and waveguide. For a particular
nonlinear mixing process, the total phase mismatch (Ak) is given by the vector sum
of all waves involved, where the k-vector for each wave is given by Eq. 2. By
adjusting the gas pressure, waveguide radius and spatial mode in which the beams
propagates, the phase mismatch can be tuned to achieve the phase-matching
condition of Ak = 0. For example, for our particular case o f difference-frequency
four-wave mixing process, coS ig n ai = 2©p u m p - coS eed 5 the phase mismatch is Ak = 2kp u m p
- kseed - ksig„ai, or, where all beams in the same spatial mode ( un m - u )
A k 2 nP
f o 5 ^
pump @ seed U signal
\ A - / l 7 / l • 7
\ pump seed signal J
A n a 2 ^ ‘^'p u m p ^ s e e d ^ signal )
Akgas Akm o,J e (4)
The phase mismatch results from a gas dispersion term (Akgasac P) minus a
■ y
modal dispersion term (Akm o de°c 1/a ). Since Akm o de>0, and in gases with normal
dispersion Akgas>0, there will exist an optimum pressure Popt, for which Ak = 0.
Thus, the advantage of the phase-matching technique using hollow waveguides is
that the mode of the generated light is determined by phase matching, making it
possible to balance the modal and the material phase mismatches.
The signal field in non-depleted-pump approximation [10,11]
y ? ikiX0)E 2 p u m p K e e d [exp(z'Ak3z) - 1] ^
signal \ Z ) ~ i ,
A k3
where z is interaction length. Or
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hence, the intensity of the signal depends quadratically on interaction length. At the
same time, the phase mismatch Ak is dependent on the pressure P, the wavelength A,
the modal constant u and the core radius a, therefore, the intensity is also dependent
indirectly on the pressure:
3. Third harm onic generation (THG)
The schematic layout of experimental setup that allows us to perform phase-
matched four-wave mixing in a hollow core fiber-giving rise to third harmonic
generation is shown in Figure 1.5. Complete information on experimental setup and
instrumental performance is given in Chapter 2 (section 2.2). An amplified laser
system capable of producing 1 0 0 -fs pulses at 800 nm was used for the experiments.
About 20-30% of the fundamental beam (800 nm) was converted into second
harmonic (400 nm) using BBO doubling crystal. After separation of the two colors
with a dichroic beam splitter, the 400-nm beam was sent to a retroreflector. It was
placed on a translation stage which controls the relative time delay of the two pulses.
The polarization of the residual 800-nm beam was rotated by means of a half-wave
plate to maintain the same polarization of the harmonic. The two colors were then
recombined at another dichroic mirror and focused into the fused silica hollow core
EIgnat {z',P)= z 2 sine (7)
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
800 nm
100 fs, 300 jiJ
BBO
HWP
Cell filled with pressure-tunable Argon
'r f -
3co
Hollow core fiber:
75-150 pm ID x 30 cm
Detector
Figure 1.5 Schematic layout for THG. Fundamental beam (co) and second
harmonic (2co) are mixed to generate third harmonic (3co). In reality,
separate focusing elements are used for c o and 2© beams. Here, BBO is
a nonlinear crystal, and HWP is a half-waveplate.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fiber (75, 100 and 150 jam core diameter, 20-40 cm long). The hollow core fiber was
mounted in a cell, which was, in turn, evacuated and filled with argon gas. In
experiments with the range of fibers studied, we adjusted the argon pressure in the
cell to maximize the phase-matched 267-nm (3to) signal. The UV output of the
hollow core fiber was separated from the 400-nm and 800-nm beams using a
dichroic mirror and directed into a calibrated thermal power meter or photodiode.
4. Results
For the first set of experiments, the two color input beams were focused into
a hollow waveguide of 150-pm core diameter, 20 cm long. For input (output)
energies of 180pJ (74 pj) at 2© and 55pJ (37pJ) at ©, the throughputs were
measured to be 41% and 6 8 % for the 400- and 800-nm beams respectively. For
dimensions of this capillary, much higher transmission of the lowest spatial mode is
predicted by Eq.l (97% and 88% at 400 and 800 nm, respectively). Figure 1.6 shows
the 3© output power versus argon gas pressure. To define the conversion efficiency,
we will use the power of 2© and © exiting the fiber and the theoretical throughput to
estimate the entering power. This assumes all problems are with in proper light
coupling into the fiber. For 400-nm, 76 pJ is then calculated to enter the waveguide.
Energies up to 1.5 pJ were generated in UV, which corresponds to 2% conversion
efficiency, expressed as a fraction of pump 400-nm beam. It is also typical [7] to
express the conversion efficiency in ratio to the fraction of 400-nm light leaving the
waveguide. In this case, the correction to the calculated efficiency is negligible.
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Intensity, mW
exp
EH,
1.5 -
EH
EH
EH,
EH,
EH
0.5 -
0.0
100 0 200 300 400 500 600 700
Pressure, Torr
Figure 1.6 Pressure tuning curve for 150-micron fiber, 20 cm long. Solid line
represents experimental curve of the 267 nm light intensity. It is fitted
with calculated phase matched signals for the 6 lowest-order modes,
assuming that all driving beams are propagated in EHn.
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The experimental data is fitted with the values calculated using Eq. 7. For the
given calculations we assumed that both 800-nm and 400-nm beams are in the lowest
spatial mode EHi i. The positions of the peaks are in good agreement with theoretical
calculations. Two overlapped peaks with maxi mums at 80 torr and 148 torr in the
experimental curve are assigned to phase-matched signals at EHn and EH21 spatial
modes, respectively. The broad peaks at high pressures are due to phase matching to
high-order modes in the fiber [7], The interesting observation is that the UV intensity
is greater for high-order spatial modes than that of for the lowest order EHn mode.
The expected predominance in propagation of EHn mode over high-order modes is
not seen. In fact, at short wavelengths the transmission of high-order modes and the
lowest mode are now comparable. For example, for the given fiber (diameter 150
pm, 20 cm long) calculated values of transmission for different modes at 267 nm are:
EHU (99%), EH21 (96%), EH31 (94%), EH1 2 (93%), EH41 (90%), and EH2 2 (8 8 %).
Being guided by the result of the Eq. (6) that the interaction length increases
the signal power quadratically, we tried to repeat the same conditions as used in
Mumane’s reported experiment (70 cm fiber length) [8], However, no further
enhancement in conversion efficiency in UV, especially in EHn output, was
observed. This mostly can be explained by poor throughput for long fibers in our
setup, due to imperfect straightness of the fiber. In principle, this could be improved
by having a thick-wail capillary waveguide instead of flexible fiber. Thus, use of
increasing interaction length, as an enhancement factor, was not successful.
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Nevertheless, pressure-dependence of the intensity that has been observed
above suggests another method. Being interested in increasing the power of the
lowest spatial mode EHn, which is the only mode suitable for practical applications,
one can aim to shift optimum pressure of the EHi i mode to high-pressure values.
This can be achieved by varying another dimensional parameter of the fiber, the core
diameter (see Eq. 4). Figure 1.7 shows how the phase-matched maximum shifts with
the diameter change. Pursuing this goal, several fused silica fibers of smaller than
150 pm diameter were tested and our best results, to be reported later in this work,
were obtained with a 30 cm long fiber with 75 pm bore diameter.
Under these conditions, the input (output) powers of 79 pJ (9 pJ) at 2to and
70 pJ (1.5 pJ) at co, the measured (theoretical) throughputs o f 400-nm and 800-nm
beams were 11% (68%) and 2% (20%), respectively. While the mode coupling into
the fiber was poorer than desired, the experimental results for the phase-matching
curve of the generated 3oo showed an excellent agreement with predicted value for
the optimum pressure (253 torr) of the phase-matched EHn mode. The good 267-nm
output mode quality, which is very nearly Gaussian, and high intensity, 5 times
exceeding previous results, allowed us to use a standard thermal power meter rather
than a photodiode. Assuming that only 13 pJ (i.e. the ratio of 9 pJ to 68%) of 400-
nm light actually enters the waveguide, the conversion efficiency is equal to 22%, or
31% expressed as a fraction of 400-nm light leaving the fiber. The measured
pressure-tuning curve (Figure 1.8) has a clear maximum at 248 torr, corresponding to
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Intensity, a.u.
150 (.tm 100 (im 75 nm
0 100 200 300 400 500
Pressure, Torr
Figure 1.7 Calculated phase-matched optimum pressures for three hollow core
diameters. Maximums are located at 60 torr, 140 torr and 250 torr for
150, 100 and 75 micron fiber, respectively. Note, the intensities are not
to scale.
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.8 mW output intensity. The slight difference of 5 torr between calculated and
experimental values is within the accuracy of pressure controller calibration. The
difference in width between predicted and experimental curves might reflect the
situation when propagation losses are not negligible. The theoretical dependence was
derived for non-depletion approximation and therefore exhibits a broader width in
the phase-matching pressure curve [10]. In contrast to the first series of experiments
with 150 pm fiber, where phase-matched EHu and EH21 modes had 68 torr
separation, now theory predicts the maximum for EH21 at 640 torr. However, due to
the sensitivity of the power meter (~0.1mW) no other secondary peaks o f the high-
order modes were observed up to 700 torr.
5. Discussion
The real challenge working with hollow core fibers is getting high laser light
coupling into the fiber and good transmission through the fiber, especially in two-
beam input experiment. Almost all experimental throughputs are 2-10 smaller than
theoretical values. Such a big deviation can be understood considering two main
processes. The first process, coupling, stems from the fact that before being
propagated, the laser beam has to be focused and launched into a fiber. For example,
for the fibers with 75 pm core diameter the launching system and focusing spot were
not fully optimized, therefore considerable losses may be assigned to back reflection
of the light from the front wall of the fiber. Moreover, only proper mode matching of
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
— exp
calc EH
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
200 400 0 100 300 500
P ressure, torr
Figure 1.8 Pressure tuning curve for 75-micron fiber, 30 cm long. Solid line
represents experimental curve of the 267 nm light intensity. It is
compared with the calculated phase matched signal (dash line),
assuming that all beams are propagated in EHn.
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the input beam to the EHn mode of the hollow fiber gives high coupling efficiencies.
In most cases, experimental 800-nm beam throughputs are closer to their theoretical
values than 400-nm throughputs. The mode quality of 400-nm beam might be a little
distorted from Gaussian shape. Thus, imperfect mode matching of the second
harmonic beam is responsible for the lower than expected throughput. Another factor
omitted from consideration of wave propagation in hollow waveguides in section 2
are bending losses. The relationship for the field loss rate a, which governs the
bending losses in hollow waveguides, is aoc 1/R where R is a bending radius. Most
of commercially available fused silica hollow core fibers have veiy small outer
diameter of less than 1 mm and they are very flexible. Therefore it is very difficult to
maintain straightness of the fiber when its length exceeds 70-80 cm.
Earlier in this chapter (Eq. 7) we showed that the power of generated UV
signal depend quadratically upon the interaction length L. Using Equation 5 the
power dependence can be extended for other parameters. In the limiting case of low
losses and zero phase mismatch, and since wave vector kj is proportional to 1/a 2 the
equation can be reduced to [11]
(8)
a
where Pw and P 2m are the powers of fundamental and second harmonic beams,
respectively. We will use this to analyze the obtained UV powers. It is now obvious,
and as it has been proven in the last experiment, that the radius of the hollow core is
the one of the most important parameters affecting power of four-wave-mixing
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
signal. Ideally, if switching of fibers form 150- to 75-micron didn’t change the power
throughputs the UV power would be multiplied 16 times. According to Eq. 8 the
same multiplication requires 150-micron fiber to be at least 4 times longer. Thus, this
change may partially resolve the problem of bending losses by avoiding sufficiently
long fibers.
Another interpretation of Eq. 8 might be given in term of peak intensity, P im ~
L2 I2. We will use it to make a comparison with the data from Mumane’s group [8],
which used 3 times shorter laser pulses (35 fs) and a 140-micron diameter fiber, 70
cm length. In our experiments by switching the hollow core diameter to 75 micron
we increase the peak intensity by factor of 4. Assuming the same average powers for
the driving beams, our results for signal intensity should only be separated by L
factor. Comparing two sets of data for length 30 cm versus 70 cm and 2.8 pJ versus
10 p i we conclude that these data are in a good agreement with Eq. 8. Thus our setup
achieves the same order of conversion efficiencies.
In conclusion, we have demonstrated a very efficient and robust technique to
obtain femtosecond UV pulses. As will be seen in Chapter 2, these pulses have
unique bandwidth properties, while the beam exhibits excellent technical
characteristics, such as the nearly Gaussian mode and the shot-to-shot noise.
Currently, pulse energies up to 4 jjJ can be generated, approaching 40% conversion
efficiency. With several straightforward improvements in coupling efficiency and
dispersion compensation In the pump, it should be possible to increase conversion
efficiency of output energy. Coupling, for example, could be improved by use of
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
tapered input to reduce reflection losses [12]. Finally, we need to optimize the
dimensions of the fiber, the length and diameter, once coupling losses have been
mostly eliminated.
6. References for Chapter 1
[1] S. Backus, C. Durfee, M. M. Mumane, and H. C. Kapteyn, Rev, Sci. Instrum.
69, 1207 (1998).
[2] A. Rundquist, C. Durfee, Z. Chang, G. Taft, E. Zeek, S. Backus, M.
Mumane, H. Kapteyn, I. Christov, and V. Stoev, Appl, Phys. B 65, 161
(1997).
[3] S. Backus, C. Durfee, G. Mourou, H. C. Kapteyn, and M. M. Mumane, Opt.
Lett. 22, 1256 (1997).
[4] J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, S. E. Bradforth, Chem.
Phys. Lett, 298, 120 (1998).
[5] J. A. Kloepfer: PhD Dissertation, USC (2002).
[6] S. Backus, J. Peatross, E. Zeek, A. Rundquist, G. Taft, M. M. Mumane, and
H. C. Kapteyn, Opt. Lett. 21, 665 (1996).
[7] C. G. Durfee, S. Backus, M. M. Mumane, and H. C. Kapteyn, Opt. Lett. 22,
1565 (1997).
[8] C. G. Durfee, S. Backus, H. C. Kapteyn, and M. M. Mumane, Opt. Lett. 24,
697 (1999).
[9] E. A. J. Marcatili and R. A. Schmeltzer, Bell Syst. Tech. J. 43, 1783 (1964).
[10] C. G. Durfee, L. Misoguti, S. Backus, H. C. Kapteyn, and M. M. Mumane, J.
Opt. Soc. Am. B 19, 822 (2002).
[11] A. M. Zheltikov, Physics-Uspekhi 45, 687 (2002).
[12] I. K. Ilev , R. W. Waynant, Rev. Sci. Instrum. 70, 3840 (1999).
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C hapter 2. Femtosecond UV laser source
1. Introduction
Recent interest in fused silica hollow core fibers started when it was proved
that they had utility in the compression femtosecond optical pulses o f high energy
[1]. Compression becomes possible due to spectral broadening obtained by pulse
propagation along a fiber waveguide filled with noble gases, followed by group
velocity compensation. Optimizing all the parameters of the experiment (pressure,
fiber length, inner diameter), Svelto and co-workers have been able to achieve
compressed pulses of 10 fs duration using output of a Ti:Sapphire laser with 140 fs
pulses. This technique was extended to single color compression at near 800 nm [2,
3] and 400 nm [4]. Even though the spectral broadening mechanisms and conditions
were different for single color and difference-frequency four-wave mixing
(DFFWM) experiments, researchers from Mumane’s group have shown that phase-
matched UV pulses at 270 can be compressible down to 8 fs [5], However, in order
to achieve such durations their experimental setup was pumped by 800-nm pulses
with duration of 35 fs. One of the purposes of this Chapter is to show that even with
100 fs pulses the DFFWM process can generate an ultrabroad bandwidth pulses, and
it is possible to compress them with relatively simple technique.
Another useful opportunity, broad wavelength tunability of UV laser light,
opens up with gas-filled hollow-core fiber technique. If DFFWM using strong pump
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
pulses (400 nm) is seeded with a tunable IR light (e.g. 600-2200 nm, instead of the
fixed wavelength at 800 nm), the signal can be continuously tuned throughout the
deep UV (220-300nm). Moreover, the generated UV signal can be further mixed
inside the fiber with 400 nm (pump) and IR seed to generate VUV signal. For
example, 267 + 400 - 800 = 200 nm. This is called a cascaded mechanism [15] (see
Figure 2.1).
All these outstanding features, in combination with high conversion
efficiencies demonstrated in Chapter 1, will lead to the development of a versatile
and unique UV laser source. However, before being used as a spectroscopy tool
other important parameters are required, such as power stability and mode quality.
Only these characteristics make it possible to carry out precise non-linear optical
spectroscopy, e.g., a measurements o f pulse duration, which remain a nontrivial task
in the UV region. In this Chapter I will describe detailed installation o f the source,
concentrating on realization of tunability and measurements of the bandwidth and
pulsewidths of generated light.
2. Installation
2.1 Wavelength tunability
There are two methods of obtaining different wavelengths in UV using
difference-frequency four-wave mixing process (DFFWM). The first method is a
conceptually straightforward extension of the work presented in Chapter 1. Using a
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
400
400
= 267mn
A , sjo n a t = 200niTl
400nm
200nm
Figure 2.1 Cascaded mechanisms. Generation of 4co and 5© harmonics via
DFFWM. Here, ©=800nm, 2©=400, ©=267, 4©=200. First mixing 3©=
2©+2©-©. Cascaded mixing, 4®=3©+2©-©, or 4©=3©+3©-2©. 5© can
be generated by a similar two step cascaded event.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
two-color frequency-mixing scheme, where 400 nm serves as a pump and tunable
light at 500-2200 nm as a seed (couvsignai = 2x©p u tn p - coseed ), the output UV light is
generated from 330 nm to 220 nm. Figure 2.2 shows the dependence of the signal
wavelength on the seed wavelength assuming a 400-nm pump. Thus, the third
harmonic generation (THG) studied in Chapter 1 is basically a special case of this
general method. However, we have to remember that for each specific wavelength
mixing there is an optimum pressure, which makes this process phase-matched. The
graphs in Figure 2.3 give information on what pressure is needed to obtain high
power signal.
The second method is more complicated and it relies on generation of new
UV fields (and VUV) from cascaded four-wave mixing events. The generated UV
light propagating further in a hollow core fiber starts to mix with pump and seed by
another DFFWM process (see example in Figure 2.1). This is second mixing is the
first member of a cascade. However, generated VUV pulse energies are small
relative to the UV signal from the first mixing because the pressure is optimized only
for the first event.
We will concentrate primarily on the first method, as this is what we have
implemented so far. The tunable seed light (500 - 2200 nm) can be taken from
Optical Parametric Amplifier (OPA) (Figure 2.4). In spite o f such a huge coverage of
the spectrum, OPA power capabilities are limited. Pulse energies in the far-IR, for
example, do not exceed 1 pJ. Theoretically, even at these energies seed photons
should induce a field for stimulated emission of two photons, seed and signal. In
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
U V signal wavelength, nm
220
240
260
• i r
280
320
0.6 0.8 1 1.2 1.4 1.61.8 22.2
■ 5.6
•5.4
- 4 .8
::~4.4
C
<
w
co'
3
S.
CD
3
C D
c3
> <
CD
<
■4.0
S eed w av elen g th ,
Figure 2.2 Signal/Seed wavelength dependence. 1 /iW v_signai=2//Uoonm - 1/^ ir-
Wavelength units are presented in reciprocal scales. Right-hand side
scale shows the energy of the UV photon in eV. For example, if the
wavelength of the seed is 1.2pm the wavelength of the output is 240 nm
or 5.17eV.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Pressure, Torr
— Pressure
— nmSigna!
220
500-
-2 4 0
450-
-260
400-
-2 8 0
350-
300
-300
250--®
-320
200
600 800 1000 1200 1400 1600 1800 2000 2200
S e e d W av elen g th , nm
Figure 2.3 Linear dependence of the optimum pressure on seed wavelength (square
dots). Phase-matching conditions were calculated assuming that all
beams are in EHn mode, propagating in the hollow core fiber with 75-
micron core diameter. The upper curve linked to RHS scale shows
UV/Seed wavelength dependence. For example, if kse e d = 1000 nm the
needed optimum pressure is 295 torr, the generated UV output appear
-250 nm.
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
U V output, nm
S eed W avelength, |im
Figure 2.4 Signal/Seed dependence in terms of different options of the OPA
output.
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
practice the threshold may be higher causing UV signal intensities to be smaller.
Hence, seed power studies can give the answer over what range whether it is possible
to implement tunable source.
2.2 Experim ental
The UV source optical layout is shown in Figure 2.5. The output (0.7 mJ,
lOOfs, 800nm, 1kHz, horizontal polarization) of a ThSapphire regenerative amplifier
system (Spectra-Physics Hurricane) is first sent into Optical Parametric Amplifier
(Spectra-Physics, OPA-8QOCF). 50% of the input 800 nm is split (1) to pump the
OPA and the remainder is delivered unused to the front port of the OPA.
O f this 0.35 m l 800 nm, the beam is further split into two legs. In the first
leg, approximately 3% of energy is transmitted in beamsplitter 2 (see specifications
in Table 1), to produce a probe wavelength (e.g. white light continuum) for
femtosecond pump-probe experiments. The remaining 97% (~300ju.J) is reflected and
used for pumping the UV source. High-energy laser mirror 3 directs the beam into
Galilean telescope, which consists of one plano-convex lens 4 (f = 10 cm) and plano
concave lens 5 (f = -5 cm). With the current focal length ratio beam diameter can be
down-collimated from 5 mm to -2.5 mm (1/e2 ), to pump BBO crystal 6 (0.5 mm,
type I). More than 100 p j of 400-nm light is converted from fundamental beam. This
corresponds to conversion efficiency >30 %. The position of focusing lens 4
mounted on a translation stage varies the divergence of the beam emerging from
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
w
§
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2 .5 Optical layout.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
# Name Parameters Vendor Catalog Namber
2 Beamsplitter Reflectance 97%, Trans.3% CVI BS1 -800-97-1012 -45-P-UV
3 800mn mirror bandwidth 64 nm CVI TLM1 -800-45-P-103 7
4 Plano-Convex Lens f = 100 mm Melles Griot 01 LQP 007
5 Plano-Concave Lens f = - 50 mm Melles Griot 01 LQS 003
6 BBO crystal 500 micron, Type I, cut angle 28.7
7,12,31 Dichroic mirror R400/T800 CVI BSR-48-1012
9,10,32 400 nm mirror bandwidth 64 nm CVI TLM 1-400-45-S-103 7
11 Plano-Convex Lens f= 250 mm Melles Griot 01 LQP 025
14,16 Gold mirror 95% Reflectance CVI PG-PM-1037-C
17 IR lens AR coating (1050-1600 nm) CVI PLCX-25.4-128.8-UV -1050-1600
18-21 Silver mirror 95% Reflectance CVI PS-PM-1037-C
22 Fused Silica Window 1mm Thickness, AR 355-532 nm CVI W2-PW-0504-UV-355-532-0
23 Fused silica fiber ID 75 micron, OD 363 micron Polymicro TSP075375
24 Duran Capillary Tube 6mm OD X 0.4mm ID, 59" Long Quark Enterprises Schott Item# 38535
25,26 XY translation stage 3 micron sensitivity Newport 460A-XZ
27 Fused Silica Window 1mm Thickness, AR 225-308 nm CVI W2-PW-0504-UV-225-308-0
28 Long wave pass dichroic R266/T400 CVI LWP-45 -Rs266-T s400-P W -1025-UV
29,30 UV mirror bandwidth 45 nm, s-pol. CVI TLM1-250-45-S-1037
33 Vacuum/Pressure Controller 1-760 torr valve Cole-Parmer Non catalog item (NCI# 79SX)
35 Needle Valve Brass LA valve B-4JN
37 Argon gas prepurified
Table 1. Inventory of parts to build UV hollow core fiber source.
telescope, thereby controlling the efficiency of conversion. In order not to damage
the BBO crystal the position of the focusing lens 4 is translated so that the typical
energies of 400-nm pulses separated from 800-nm pulses by dichroic mirror 7 are
70-80 pJ. Pump beam is then transmitted through the iris 8, It is used to adjust the
focal spot diameter to match that of the lowest-order mode of fused silica waveguide.
Mirrors 9 and 10 are mounted on translation stage, which controls the relative time
delay between pump and seed pulses. Preliminary alignment of pump beam along the
propagation axis is performed using focusing lens 11 (f = 25 cm) and recombining
mirror 12,
After separation with dichroic mirror 7, the 800-nm beam is sent to waveplate
13 to rotate the polarization from horizontal to vertical. Polarizer 15 is optional part
in the setup, it can be used for power-tuning studies. The beam alignment along the
propagation axis is adjusted using mirrors 14 and 16. Focusing lens 17 (f = 25 cm)
with antireflection coating optimized for the IR is mounted on XYZ translation stage,
which enables micro-alignment.
As stated earlier, the 0.35 mJ of 800-nm beam is pumping OPA, to give
wavelength in the range of 1200 -1600 nm (OPA signal) and 1600-2400 nm (OPA
idler). Different types of frequency mixing options extend this range from 350 nm to
1200 nm (OPA user manual). For realization of UV tunability the output of OPA is
delivered using silver mirrors 18-21. After recombination of pump and seed, beams
are launched into argon gas chamber and on into the fiber. Entrance window 22 has
an antireflection coating optimized for 400 nm. After propagation in the fiber the
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Vacuum pump 38 37 Argon
Figure 2.6 Argon-gas chamber and pressure controlling system.
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
output beam passes through the exit window 27, which has antireflection coating
optimized on UV region (225-308 nm). Long-wave-pass dichroic 28 separates UV
signal from 400 nm and 800 nm beams. Broadband UV laser mirrors 29 and 30 form
a multiple-reflection mirror system for cleaning UV light from 400nm. Remaining
two-color output transmitted through the mirror 28 is then separated by dichroic
mirror 31.
Argon-gas chamber and pressure controlling system is built the following
way. (Figure 2.6) Fused silica fiber 23 is threaded inside a standard capillary 24
which is, in turn, mounted by holding clamps on two XY translation stages 25, 26.
These stages perform the main alignment of the fiber. Poly-flo tubing 36 is
connected to stainless steel tube 34 and ultra-torr fittings, which are fitted on both
ends of the capillary 24. Vacuum/pressure controller 33 and needle valve 35
maintain a constant flow inside the tube 34, thereby keeping a constant pressure
along this tube. The pressure gradient is performed by high-pressure argon tank 37
and vacuum pump 38. Pressure controller turned out to be very stable tool. If the
needle valve is correctly adjusted the stability of the pressure can last during the time
of the experiment without any problem, however, one disadvantage of this controller
is inaccuracy in calibration (+/-5 torr).
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3 Results
2.3.1 Power stability, shot-to-shot noise, mode
Technical characteristics of UV output beam were tested by measuring
power stability, shot-to-shot noise and spatial mode quality. First, a power meter was
set to collect the data for two periods of time, 3 hours and 3 min. Using statistical
analysis the estimated stabilities were 5% and 3% for long and short time periods,
respectively. The mean power of 2.8 mW was the same in both measurements. At the
same time, the minimum incremental resolution of the power meter is 0.1 mW, and
within the range of 2.8 mW this corresponds to -3% accuracy. Hence, the power
stability for the observation time of 3 minutes was limited by the power meter. In a
second measurement, the power was registered by a photodiode.
The output signal was displayed on an oscilloscope. Using this method we
were able to detect both the power stability (slow noise) and shot-to-shot noise. The
former turned out to be 2.5%; the latter was much less than 1%, estimated by taking
the ratio of the linewidth of the signal at the peak to its height. This unprecedented
shot-to-shot noise for the multiple stage UV sources favors of generation using
hollow-core fibers over nonlinear crystals. Finally, pinhole transmission experiments
were performed to find out the beam mode quality. The UV laser beam with 6 mm
diameter was focused using a lens (f = 15 cm) into a 50-micron pinhole. The nearly
Gaussian mode of the beam allowed us to transmit 3mW input power with 97%
throughput. With such a tight focusing, assuming a conservative 100 fs for the pulse
duration, the peak intensities as high as 1.6 TW/cm2 can be achieved.
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3.2 Seed power dependence
While considering feasibility of tunability experiments, the seed-power
dependence curve can help to determine the minimum required power for generation
of UV signal. Figure 2.7 shows the power study of the UV signal intensity at 267nm
depending on seed-beam intensity at 8Q0nm. Two experimental data sets, which
correspond to two different alignment situations, are fitted with a single exponential
function. The curves slowly approach their saturation limit of 3.6 and 3.3 mW for the
upper and the lower curves respectively. The depletions of the pump and the seed
power may be involved in the observation o f a slow saturation with seed power.
Interestingly, a finite seed power is required for good output - this may be
due to propagation losses. However, there is no need to deliver greater than 10 mW
of well-aligned seed light under current operating conditions. Perhaps with
improvements in coupling efficiencies this can be reduced to the 2-5 mW range. This
is important because the OPA outputs vary with output wavelength in 1-20 mW
power region. This illustration 800-nm seed-dependence shows that even at theses
energies (1-20 mW) it is possible to obtain moderate powers of UV signal by
DFFWM. However, careful attention should be paid to alignment. Further
improvements in increasing power throughput of the pump in the fiber will also
enhance the effect.
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
267nm power, mW
4.0-j
3.5-
3.0-
2.5-
2.0 -
1.5-
1.0 -
0.5-
0 - 0 — !— |— i— |— i— |— i— |— i— i— i— i— i— i— i— r
0 10 20 30 40 50 60 70
80Qnm power, mW
Figure 2.7 Dependence of the UV output power (267 nm) on the seed (800nm)
intensity. The lower curve (triangle dots) represents “bad” alignment,
the upper curve - “good” alignment (square dots). Experimental data
fitted with first order exponential decay.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3.3 Performance and setup for 240 nm
For future experiments on chemical system, tunability of a UV source is
highly desirable. For fiber alignment purposes, it is easy to start with the maximum
power we can obtain from OPA - this occurs near 1200 nm in the OPA signal output
(Figure 2.4). The measured power was -3 0 mW at the output o f the OPA. After
passing through all the optics (see OPA beam path 18-21,17,12) the power at the
entrance of the fiber was 20 mW. When launched into the fiber, throughput for pump
400-nm pulses was measured to be -20% (45 mW input). IR transmission through
the fiber (75 micron, 35 cm long) is expected to be no better than 5% in the range of
1000-2200 nm. In our first experiment, <0.1 mW of 1200 nm power was detectable
at the exit of the fiber. Using a prism the UV light was separated from 800nm and
400nm beams.
Experimentally, the optimum pressure was at 338-351 torr for EHn- This
result is in a very good agreement with the theoretical value (339 torr). Despite little
IR seed propagates the full length of the fiber, the power o f 240-nm beam was still
measured to be -1 mW. Comparing this result with previous 800 nm seed
dependence, the powers are different only by a factor of two. This is still quite good
result taking into account dramatic change in fiber propagation by switching the seed
wavelength from 800 nm to 1200 nm.
The shot-to-shot noise of the 240-nm beam evaluated using oscilloscope was
similar to the data obtained for 267-nm, about 1% peak-to-peak. Interestingly, the
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
FWHM = 4nm (30fs)
re
in
c
re
60-
5 0 -
4 0 -
30-
2 0 -
10 -
0
T T T T
234 236 238 240 242 244
W avelength, nm
246 248 250
Figure 2.8 Spectral broadening of the 240 nm signal, FWHM is equal to 4 nm. The
fiber (75 micron, 35 cm long) was pumped by 45 mW at 400-nm and
seeded with IR pulses at 1200 nm, 10-20 mW power. The time
bandwidth in brackets shows transform-limited pulse duration for a
particular spectral bandwidth, assuming Gaussian spectrum.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measured noise of the seed 1200 nm beam was higher than that of the UV signal
energies from the fiber. This fact might indicate that the noise of the generated UV
light depends only on the pulse-to-pulse stability of the pump (400 nm) beam. The
spectrum of the 240-nm pulses is shown in Figure 2.8. In this case the FWHM is
estimated to be 4 nm (694 cm4 ), which is rather smaller than obtained in 267 nm
data (1050 cm'1 ). This indicates that XPM has smaller effect due to the lower
intensity of the seed power (see details in section 3.1). However, from the example
above, the relative time delay can be adjusted so that the maximum of the bandwidth
is centered, resulting in broader FWHM.
Finally, this initial experiment confirms that the OPA can be used as a
tunable seed and that the wavelengths on Figure 2.4 from 500 nm to 800 nm
(SFM(pU m p + id ier); SHGS ignai) and 1100 nm to 2000 nm (either Signal; or Idler) at least
can be used to produce 267-310 nm and 220-245 nm UV wavelength range. The
region from 800 nm to 1100 nm is considered to be problematic due to small power
output from the OPA, less than 1-5 mW.
3. U ltrashort pulse characterization
3.1 Spectral broadening
Only detailed knowledge of the spectrum and the phase of the ultrashort
pulse gives complete information about the structure of the pulse envelope. It has
been recently shown [12] that the phase of the generated UV pulse is likely to be
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
well behaved, and the pulse can be compressible to nearly transform-limited pulse
durations. Therefore, for simple estimations of minimum possible pulse duration we
use time-frequency bandwidth relation Av-At = 0.44, assuming Gaussian pulse shape.
The accuracy o f this assumption will be tested with current and future experiments In
this case the information about spectral broadening is very important.
When an intense optical pulse is propagated in an optically nonlinear
medium, self-phase modulation (SPM) occurs. The phase of the pulse is modulated
because of the intensity-dependent change of the refractive index of the medium.
Thus, the spectrum of the pulse is broadened by chirp arising from this temporal
change o f the phase. The typical broadened spectra of the output pulses for the 800-
nm, and 400-nm beams are shown in Figures 2.9 (a) and (b), respectively. The output
IR spectrum shows amplitude modulations, which are considered to be characteristic
of a pure self-phase modulation process [1], Making a comparison between the input
and the output spectra for both beams, one can find that the FWHM of the output
spectrum is approximately 3 times broader than that of the input spectrum.
In order to estimate spectrum broadening of the UV light emerging from the
fiber, when produced by DFFWM compared to typical SFM in crystal, the original
800-nm and 400-nm beams were mixed in the nonlinear crystal using SFM to
generate the third harmonic at 267 nm (3to = 2® + co). Figure 2.10 shows that spectra
of the UV pulses generated in the hollow core fiber are more than 4 times broader
than those produced in a typical THG setup. O f course, the reduced bandwidth in the
crystal conversion process is due to GVM bandwidths in the crystal, however, the
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
output FWHM = 33nm (28fs)
input FWHM = 10.7nm (90fs)
0.6 -
9-
£
^ 0.4 -
0.2 -
0.0
750 760 770 780 790 800 810 820 830 840 850
Wavelength, nm
- — — output FWHM = 11.5nm (20 fs)
input FWHM = 3.2nm (74 fs)
u
C D
Q.
£
<
0.4 -
0.2 -
0.0
380 390 400 410 42Q 430
W avelength, nm
Figure 2.9 Spectral bandwidth of 800-nm (a) and 400-nm (b) beams. For both
graphs the solid line represents the spectra o f pulses emerging from
fiber, dash line is a spectrum of input beams The output bandwidth
shown are collected with the time delay set to generate UV signal in
DFFWM, i.e. there is a depletion due to wave-mixing. All data
collected for 35 cm length fiber and 75 micron core diameter. The time
bandwidth in brackets shows transform-limited pulse duration for a
particular spectral bandwidth, assuming Gaussian spectrum.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fiber: FWHM = 7.5nm (14 fs)
crystal: FWHM = 1.8 nm (57fs)
0 .4 -
Q.
0 .2 -
0.0
256 260 268 264 272 276 280
Wavelength, nm
Figure 2.10 Output spectra of third harmonic generated with hollow core fiber (solid
line) and non-linear BBO crystal (dash line). The fiber parameters: 35
cm, 75 micron core. Crystal thickness 80 pm. The time bandwidth in
brackets shows transform-limited pulse duration for a particular spectral
bandwidth.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fiber output is considerably broader than could be achieved if GVM were eliminated
in crystal frequency mixing. Both conversion of the broadened spectra of the
fundamental and the second harmonic and direct cross-phase modulation driven by
the 800-nm and the 400-nm pulses are considered to be responsible for the
broadened UV signal spectrum. XPM, where change in the refractive index of the
medium induced by another pulse, is an effect of at least a two-beam interaction.
Therefore its character depends on the relative timing between pulses. Figure 2.11
shows this behavior of the UV spectrum with respect to the time delay between
pulses. When the pump pulse precedes the seed, the short wavelength part of the blue
(400 nm) is mixed with the long wavelength part of the red pulse (800 nm), resulting
in UV output that is shifted to short wavelengths. As the delay reduced, the situation
is reversed, and the UV output can be tuned in this manner over 5 nm. However,
increasing the power of one of the driving beams significantly broadens the UV
spectrum due to cross-phase modulation [10, 11]. In our case, for example, in fibers
with 150-micron diameter the FWHM spectrum of up to 10 nm can be achieved
(Figure 2.12).
3.2 Method of pulsewidth determination
Next step on the way of characterizing of the new femtosecond source is to
determine the duration of pulse. The determination of the duration of visible or near-
infrared laser pulses with femtosecond duration is standard procedure in ultrafast
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
—a — blue pulse first
—■—zero delay
— red pulse first
FWHM = 7.5nm
a.
0.4-
0 .2 -
0 . 0 -
256 260 264 268 272 276 280
Wavelength, nm
Figure 2.11 Output spectra of UV signal at three different relative delays of blue
(400 nm) and red (800nm) pulses. Experimental data (square dots)
represent time-zero delay, indicating 7.5 nm FWHM bandwidth. Blue
(triangle dots) and red (round dots) curves correspond to cases when
blue pulse precedes red pulse or vice versa.
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Intensity (counts)
4 0 0 0 -
3500 -J
3000-
2500-
2 0 0 0 -
1500-
1 0 0 0 -
500-
0-
— I—
250
— I—
280
— I—
290 260 270
W avelength (nm)
Figure 2.12 Output spectrum of the UV signal emerging from 150-micron-core
fiber, 20 cm long. The intensities of 400-nm beam and 800 nm were
76mW and 42mW, respectively, assuming that these powers were
actually enter the fiber. FWHM about 10 nm can be extracted from this
graph, which correspond to lOfs transform limited relation.
4 6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
spectroscopy. Due to phase-matching or GYM. issues, methods applied in the IR and
visible spectrum cannot be extended to the UV region. Recently, several methods
were suggested for pulse determination, which allows one to extend them further in
the UV to vacuum-ultraviolet (VUV) [8, 9].
Autocorrelation technique exploiting two-photon absorption (TPA) in a fused
silica was chosen as a method of characterizing ultrashort UV pulses. Since a lot of
information on TPA of fused silica using high-intensity femtosecond pulses at near
267 nm is available [9, 12], the choice of fused silica as a sample material was
considered preferable.
The main idea of the pulse compression is to propagate broadened pulses
through a device with negative dispersion, which can be created using grating or
prism pairs. These optical systems produce different amounts of GVD, TOD and
FOD depending on their characteristics and geometrical arrangement [16]. In our
studies we will use the prism compressor, the most common optical tool for pulse
compression.
3.3 Autocorrelation setup
The optical layout based on interferometric arrangement for pulse
autocorrelation measurement by TPA with noncollinear geometry is shown in Figure
2.13. After passing through the exit window 1 of the argon-filled chamber, the output
beam is collimated using lens 2. Long-wave-pass dichroic mirror 3 separates UV
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2.13 Optical layout for autocorrelation setup.
light from the pump and the seed beams. The mirrors 4 and 5 are used to align and
sent the beam to the beginning of the interferometer, where the beam is divided into
two legs. Approximately 90% of original energy is reflected by beamsplitter 6 to
produce a pump pulses. Remaining 10% is transmitted and used for probing.
Dispersion compensation plate 7 is placed to equalize both legs. The mirrors 9 and
10 are mounted on the translation stage, which controls the relative time delay
between the two pulses. Noncollinear geometry is performed using two mirrors II
and 12 so that both pump and probe beams travel a certain distance parallel to each
other until they encounter a focusing lens 13. Focal length of the lens and the
distance between beams are chosen so that the crossing angle between the beams is
approximately 5-10°. After focusing in the fused silica sample 14 pump beam is
blocked by stopper 16. The probe beam is detected by the photodiode 19 following
the UV filter 18.
Pulse compression modification is performed as follows. Dichroic
beamsplitter send the UV beam to the calcium fluoride prism 20, so that the
incidence angle is exactly at the Brewster angle. After refraction in the second prism
21, the beam is reflected from mirror 22 and using mirror 4 is sent to the beginning
of the interferometer. The prisms are mounted on translation stages, which controls
the beam insertion into the prism. Both the distance between the prisms and the
insertion are optimized for the best compression.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 Results
A first set of experiments was done with fused silica as the two-photon
absorption medium. The energies of the pump and probe pulses were 700 nJ and 50
nJ respectively. At these high energies of the pump the focusing spot started to
initiate red fluorescence. To avoid this effect, the plate of the fused silica was placed
slightly closer to the lens away from the focus of the two beams. The probe beam
was re-aligned to cross the pump beam in the glass. Figure 2.14 shows a typical
shape for the autocorrelation trace without compression. Based on the Gaussian
shape of the pulse, pulsewidth of 500 fs at FWHM can be retrieved from the
experimental curve.
A second set of experiments was performed with prism pair compressor.
Since the polarization of the UV light in this preliminary set-up was vertical (s-
polarization with respect to prism surface), total transmission losses were -70% .
This will soon be corrected so that the correct Brewster configuration in the prism is
achieved and losses minimized. The energies of the pump and the probe turned out to
be 200 nJ and 20 nJ respectively. The best result of the compression with prism
separation of 28.5 cm is shown in Figure 2.15. The trace corresponds to pulsewidth
of 29 fs. The experimental data now can be fitted with Gaussian. The obvious feature
of experimental curve is the deviation from the Gaussian fit and the appearance of so
called pedestal. It is non-symmetric with respect to time-zero time delay. The
pedestal is a characteristic of a modulated spectrum [1], All further trials to obtain
pulsewidths below 30 fs resulted in distorting the shape of the absorption curve.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
FWHM 710 fs (500+/-10 fs)
7 -
6 -
5 -
0 >
4 -
5 3-
> -
0-
-1500 -1000 -500 0 500 1000 1500
Time delay, fs
Figure 2.14 Absorption intensity autocorrelation trace in a fused silica sample.
Assuming Gaussian pulse shape, the pulsewidths of 500 fs can
estimated.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10-,
fu sed silica
FWHM 4 1 fs (29 fs)
c
.2
e-
o
S3
<
4 -
-100 100 -200 0 200
Tim e delay, fs
Figure 2.15 Experimental autocorrelation trace (solid line) in fused silica after
compression, Prism separation is 28.5 cm . The experimental curve was
fitted with Gaussian function adjusting the height and the width (dotted
line). Thickness of silica sample is -3 mm long. FWHM of the
autocorrelation trace is 41 fs, pulsewidth if fitting as Gaussian is 29 fs.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
C 5
C
o
4 3
& 3-
o
C O
5 2-
1 -
O '
-1 -
liquid w ater
FWHM 34fs (24 fs)
-200 -100 100 200
Tim e delay, fs
Figure 2.16 Experimental autocorrelation trace (solid line) in liquid water after
compression, Prism separation is 28.5 cm. The experimental curve was
fitted with Gaussian function adjusting the height and the width (dotted
line). Thickness of liquid water jet 30-50 pm.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
With increase of peak intensities, high non-linear effects might be responsible for
that odd behavior of the autocorrelation trace.
Recently, a new stable liquid jet was employed in our lab. Liquid water films,
produced by gravity drop along a wire guide may have as narrow as 30 pm thickness
[14]. The FWHM of the autocorrelation trace using two-photon absorption in liquid
water (Figure 2.16) is 34 fs, which corresponds to pulsewidth o f 24 fs (note also that,
assuming a sech2 pulse shape, a pulse duration of 22 fs can be estimated). The
pedestal is lower than in the case with the fused silica, and it is also more symmetric.
We believe that due to the small path length this autocorrelation trace is more
representative of the pulse shape.
3.5 Discussion
A good understanding and careful control of dispersion are essential in order
to deliver ultrashort pulse to the sample. Since for all transmissive optics used in our
autocorrelation setup the information about dependence of the refractive index on the
wavelength is known, the group velocity dispersion (GVD) and the third-order
dispersion (TOD) can be calculated. For the sake of simplicity, we limit our
consideration to only these two terms. The two main optical materials, which
generated UV light was transmitted through, are fused silica and air. The calculated
GVD (TOD) at 267 nm are equal to 196 fs2 /mm (59 fs3 /mm) and 0.101 fs2 /mm
" 3
(0.033 fs /mm) for the fused silica and air, respectively. If we now compare these
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
values with the dispersions for the same materials at IR wavelength, for example,
800 nm, GVD (TOD) are found to be 36 fs2 /mm (28 fs3 /mm) and 0.021 fs2 /mm (0.01
fs7mm) for the fused silica and air, respectively. Thus, the GVD for both optical
media is approximately 5 times greater in UV region than that o f in IR. As a result,
the pulse broadening effect increases dramatically for short-wavelength pulses.
In the experiments without compression the total length of propagation in the
fused silica was 9.7 mm, while the distance traveled through air was measured to be
2820 mm. Accepting 50 fs2 /bounce (30fs3 /bounce) for the GVD (TOD) on each
dielectric mirror (according to manufacture specifications), we can estimate the total
accumulated GVD (TOD) for the setup, 2435 fs2 (815 fs3). In fact, this value has a
practical meaning. It can be reversed in the prism compressor. Since the GVD and
TOD of the prism pair system greatly depend on the prism separation distance, one
could use the total accumulated GVD of the setup to find correct separation for the
prism pair, which fully compensates the GVD. However, the problem is that
generated pulses emerging from the fiber have unknown phase shift. For now, we
state the assumption that the pulses emerge with small phase error and the acquired
GVD and TOD is dominant and should be compensated. Though the prism
compensation cancels the GVD it adds more negative TOD. In our case, for a prism
3 3
separation the magnitude of the residual TOD is on the order of - 5x10 fs . Hence,
the appearance o f the pedestal in the autocorrelation traces (see Figures 2.15, 2.16)
can be explained as originating from the residual TOD.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Careful attention should be paid to the autocorrelation setup. Two legs of the
setup must be totally symmetric, so that there is no dispersion disbalance between
them. For example, it has been noticed that the autocorrelation measurements are
very sensitive to how the two-beam input is directed into the focusing lens. In
principle, this can be improved by changing the lens on a curved mirror. Also all
transmissive optics should be thinner. A few other instrumental improvements can be
made to achieve nearly transform-limited pulses. First, a grating pairs are known to
be useful for UV pulse compression. They introduce the negative GVD at very
modest separation leading to compact design, however, suffer from losses of close to
50% in total. Sometimes both grating and prism pair are used to compensate both
GVD and TOD. Second, chirped mirrors become available commercially, which can
provide second- and third-order compensation. They have excellent reflectivity and
can be used in combination with prism/grating pairs, however they available for IR
and visible wavelengths.
4. Sum m ary
We have implemented a very promising femtosecond UV laser source,
demonstrating exceptional results on the pulse-to-pulse stability and the mode
quality. Specifically, these make it possible to carry out precise pulsewidth
measurements, which in combination with compression techniques allows us to
achieve 20-fs pulses. The important experiments with OP A seeding confirm the
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
possibility of continuous tuning of “photochemical” UV light ( 4- 6 eV). Both
tunabiiity and excellent spatial mode of the UV beam will help pulses to initiate two-
photon processes, so that excitation deep into the one photon VUV range (6-11 eV)
is achievable.
5. References for C hapter 2
[1] M. Nisoli, S. Silvestri, and O. Svelto, Appl. Phys. Lett. 68, 2793 (1996).
[2] M, Nisoli, S. Stagira, S. De Silvestri, O. Svelto, S. Sartania, Z. Cheng, M.
Lenzner, Ch. Spielmann, and F. Krausz, Appl. Phys. B 65, 189 (1997).
[3] S. Sartania, Z. Cheng, M. Lenzner, G. Tempea, Ch. Spielmann, F. Krausz,
Opt. Lett. 22, 1562 (1997).
[4] O. Duhr, E. T. Nibbering, and G. Korn, G. Tempea, F. Krausz, Opt. Lett. 24,
34(1999).
[5] C. G. Durfee, S. Backus, H. C. Kapteyn, and M. M. Mumane, Opt. Lett. 24,
697 (1999).
[6] A. Rundquist, C.G. Durfee, Z. H. Chang, C. Heme, S. Backus, M. M.
Mumane, H. C. Kapteyn, Science 280, 1412 (1998).
[7] A. Paul, R. A. Bartels, R. Tobey, H. Green, S. Weiman, I. P. Christov, M. M.
Mumane, H. C. Kapteyn, S. Backus, Nature 421, 51 (2003).
[8] J.-F. Ripoche, B. S. Prade, M. A. Franco, G. Grillon, R. Lange and A.
Mysyrowicz, Applications o f High Field and Short Wavelength Sources,
Plenum Press, New York (1998).
[9] A. M. Streltsov, J. K. Ranka, and A. L. Gaeta, Opt. Lett. 23, 798 (1998).
[10] C. G. Durfee, S. Backus, H. C. Kapteyn, and M. M. Mumane, Opt. Lett. 24,
697 (1999).
[11] C. G. Durfee, S. Backus, M. M. Mumane, and H. C. Kapteyn, Opt. Lett. 22,
1565 (1997).
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[12] L. Misoguti, C. G. Durfee, S. Backus, M. M. Mumane, and H. C. Kapteyn,
Ultrafast Phenomena XIII, ed. Miller, Springer-Verlag (2003).
[13] A. Dragomir, I. G. Mclnerney, and D. N. Nikogosyan, P. G. Kazansky, Appl.
Phys. Lett. 80, 1114(2002).
[14] M. I. Tauber and R. A. Mathies, J. Phys. Chem. A 105, 10952 (2001).
[15] L. Misoguti, C. G. Durfee, S. Backus, R. Bartels, M. M. Mumane, and H. C.
Kapteyn, Phys. Rev. Lett. 87, 013601 (2001).
[16] S. Backus, C. Durfee, M. M. Mumane, and H. C. Kapteyn, Rev. Sci. Instrum.
69, 1207 (1998).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C hapter 3. Liquid w ater photochemistry
1. Background
Understanding the nature of water reactivity on a molecular scale is one of
the most fundamental and challenging problems in chemistry and biology. A very
interesting example of this is the reaction dynamics of primary photo- and radiation
induced chemistry of pure water. From early studies of liquid water photolysis, it is
known that radiation at photon energies higher than 6.5 eV results in the generation
of solvated electrons. This value is very low in comparison with the gas-phase
ionization threshold (12.6 eV). The formation of e‘aq at excitation energies
significantly lower than the Bom-Oppenheimer threshold implies involvement of
nuclear displacement or even secondary reactions of photo excited water molecule
with adjacent molecules. However, it is still unclear how exactly the ejection of the
electron takes place [1-3], Our main interest in this problem, thus, can be stated in
theses two questions: 1) what are the primary photo-events and 2) what is the
mechanism of the solvated electron formation.
A simple model for description of initial processes following photo excitation
of a water molecule in the condensed phase is shown in Figure 3.1. In this
assumption, two main channels, dissociation and ionization, compete on an ultrafast
time scale (<100 fs). In the latter channel an electron is ejected from its parent
molecule to some distance. By ~1 ps all products (OH, H, e'aq, H3 0 + aq) in both
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 0 fs
H + OH
100 fs
[ O H ? H 3 0 +(aq) , e (aq) ]
Figure 3.1 Primary events in the first picosecond. By analogy with the gas phase
data it is assumed that within ~ 10 fs the water molecule dissociates into
H and OH. First two fragments of photo-ionization channel are initially
a positively charged hole H 2 0 + and e'. H 2 0 + reacts quickly (-100 fs)
with another water molecule to form HsO+ and OH fragments. After a
period of - 1 ps the system is fully relaxed and a certain ejection length
separates the charged hole (now H 3 0 + ) and trapped e'aq. This separation
depends on the photo-ionization energy. The question sign between two
main channels represents other possible mechanisms of formation of the
solvated electron and accompanying products. For example, it has been
suggested there exists a dissiociative ionization of excited water
molecule (H(hot) + H2O -* HsO+ + e'a q ).
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cm
2000
1500
v 1000
2
w 500
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
Energy (eV)
Figure 3.2 The static absorption spectra of e‘ aq, OHaq, Haq, and OH”a q from 200 to
1200 nm (Ref. 2). Notice that the absorption spectrum of the solvated
electron has been reduced by a factor of 10. For example, the extinction
coefficients for the solvated electron and OH radical at 267 nm (4.65
eV) are -1000 M ^cm '1 and -400 M ^cm '1 , respectively.
6 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
channels become solvated. According to the absorption spectra (Figure 3.2) only e'a q
and OH can be probed easily by transient absorption spectroscopy to monitor the
population of the species. While the absorption spectrum of H 3 0 + aq has never been
identified, H 2G+ may absorb in the visible [4].
The general idea of the femtosecond transient absorption spectroscopy of
liquid water using pump-probe technique is presented in Figure 3.3. Two-photon UV
excitation followed by an IR probe pulse allows us to observe the trapping and
solvation of the electron at the early times (100 - 1000 fs), and geminate
recombination (recombination of e‘aq with either OH or H 3 0 + formed from its parent)
in the longer time scale (1-100 ps, Figure 3.4). These two processes determine the
important characteristics, the trapping time and the ejection length [5,6]. In addition,
experiments with different pump energies from 8 to 12 eV, where the ejection length
is considered to vary dramatically and direct dissociation channel competes strongly
with ionization, provide additional Insights into nature of ionization mechanism and
primary photo-events.
2. Prelim inary UV pump-UV probe results
It turned out that the autocorrelation setup (Figure 2.13, liquid water jet as a
sample) used for pulsewidth measurements was also suitable for one color pump-
probe experiments, employing, for example, two-photon excitation at 9.3 eV (2x267
nm) followed by probe pulse at 4.65 eV (267 nm). According to the absorption
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Two-photon
pump
One photon
probing
H ,0
▲
2xhvuv=
8 - 12 eV
(OH)'
" S
(e )
h V lR ,V Is|'
h v
UV
transients
time delay
Figure 3.3 Pump-probe technique to study femtosecond transient absorption
spectroscopy of pure liquid water.
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Transient absorption, a.u.
0.18 — |
0.16-
0.14-
0 .1 2 -
0 .1 0 -
0.08-
0.06-
0.04-
0 .0 2 -
100000 20000 40000 60000 80000 0
Time delay, fs
Figure 3.4 Experimental data for the geminate recombination of solvated electron
after two-photon excitation of pure water. The pump and probe
wavelengths are 300 nm and 700 nm, respectively.
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
FWHM 88 fs (62fs)
4 -
3 -
o
m
800 1000 -400 -200 0 200 400 600
Tim e delay, fs
Figure 3.5 The first picosecond of the transient absorption observed at 267 nm.
Here (1+1) near t = 0 corresponds to one photon pump + one photon
probe process, (2+1) is the two photon excitation followed by one
photon probe. In the experiment the pump and the probe powers were
-200 nJ and ~20 nJ, respectively. The autocorrelation trace reveals the
FWHM of 88 fs, which corresponds to 62 fs pulsewidth. The measured
value for the maximum is 230 pV.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Absorption, a.u.
0 .2 -
0.0
100000 200000 300000 -200 -150
Time delay, fs
Figure 3,6 The pump-probe absorption signal for the time delays from -200 fs to -
130 fs and from 200 fs to 3 ps. The former signal, where the probe
precedes the pump pulse, serves as a background level. The
autocorrelation trace is deliberately omitted to contrast the two signal
levels. The signal is enlarged in order to demonstrate the small transient
absorption effect.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
spectra (Figure 3.2), at the probe wavelength there are only two species absorb, e'a q
and. OHaq. Thus, probing in the UV allows us to observe simultaneously the two
photoproducts.
Typical absorption data is shown in Figure 3,5 and the photoinduced transient
absorption of water is shown in Figure 3.6. We believe that both absorption of
solvated electron and OH radical are responsible for the small growth in signal in
positive time delay scale. If we improve our pulse compression setup to minimize the
power losses we, in principle, can deliver 3 pJ pump pulses to the sample that is 15
times greater than the pulse energies (0,2 pJ) used in Figure 3.5. Since the intensity
of the transient absorption scales quadratically with the pump intensity it should be
possible to increase the signal of the absorption 225 times.
3. References for Chapter 3
[1] R. A. Crowell, and D. M. Bartels, J. Phys. Chem. 100, 17940 (1996).
[2] C. L. Thomsen, D. Madsen, S. R. Keiding, and J. Thogersen, and O.
Christiansen, J. Chem. Phys. 110, 3453 (1999).
[3] D. M. Bartels, and R. A. Crowell, J. Phys. Chem. A 104, 3349 (2000).
[4] Y. Gauduel, S. Pommeret, A. Migus, A. Antonetti, J. Phys. Chem. 93, 3880
(1989).
[5] V. H. Vilchiz, A. C. Germaine, V. A. Lenchenkov, J. A. Kloepfer, and S. E.
Bradforth, J. Phys. Chem. A 105, 1711 (2001).
[6] J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine, and S. E.
Bradforth, J. Chem. Phys. 113, 6288 (2000).
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bibliography
-B-
1. S. Backus, C. Durfee, G. Mourou, H. C. Kapteyn, and M. M. Mumane, Opt.
Lett. 22, 1256 (1997).
2. S. Backus, C. Durfee, M. M. Mumane, and H. C. Kapteyn, Rev. Sci. Instrum.
69, 1207(1998).
3. S. Backus, J. Peatross, E. Zeek, A. Rundquist, G. Taft, M. M. Mumane, and
H. C. Kapteyn, Opt. Lett. 21, 665 (1996).
4. D. M. Bartels, and R. A. Crowell, J. Phys. Chem. A 104, 3349 (2000).
-C-
5. R. A. Crowell, and D. M. Bartels, J. Phys. Chem. 100, 17940 (1996).
-D-
6. A. Dragomir, J. G. Mclnemey, and D. N. Nikogosyan, P. G. Kazansky, Appl.
Phys. Lett. 80, 1114 (2002).
7. O. Duhr, E, T. Nibbering, and G. Korn, G. Tempea, F. Krausz, Opt. Lett. 24,
34(1999).
8. C. G. Durfee, S. Backus, H. C. Kapteyn, and M. M. Mumane, Opt. Lett. 24,
697 (1999).
9. C. G. Durfee, S. Backus, M. M. Mumane, and H. C. Kapteyn, Opt. Lett. 22,
1565 (1997).
10. C. G. Durfee, L. Misoguti, S. Backus, H. C. Kapteyn, and M. M. Mumane, J.
Opt. Soc. Am. B 19, 822 (2002).
-G-
11. Y. Gauduel, S. Pommeret, A. Migus, A. Antonetti, J. Phys. Chem. 93, 3880
(1989).
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-I-
12. I. K. Ilev , R. W. Waynant, Rev, Sci. Instrum. 70, 3840 (1999).
-K-
13. I. A. Kloepfer: PhD Dissertation, USC (2002).
14. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, S. E. Bradforth, Chem.
Phys. Lett., 298, 120 (1998).
15. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine, and S. E.
Bradforth, J. Chem. Phys. 113, 6288 (2000).
-M-
16. E. A. J. Marcatili and R. A. Schmeltzer, Bell Syst. Tech. J. 43, 1783 (1964).
17. L. Misoguti, C. G. Durfee, S. Backus, R. Bartels, M. M. Mumane, and H. C.
Kapteyn, Phys. Rev. Lett. 87, 013601 (2001).
18. L. Misoguti, C. G. Durfee, S. Backus, M. M. Mumane, and H. C. Kapteyn,
Ultrafast Phenomena XIII, ed. Miller, Springer-Verlag (2003).
-N-
19. M. Nisoli, S. Silvestri, and O. Svelto, Appl. Phys. Lett. 68, 2793 (1996).
20. M. Nisoli, S. Stagira, S. De Silvestri, O. Svelto, S. Sartania, Z. Cheng, M.
Lenzner, Ch. Spielmann, and F. Krausz, Appl. Phys. B 65, 189 (1997).
-P-
21. A. Paul, R. A. Bartels, R. Tobey, H. Green, S. Weiman, I. P. Christov, M. M.
Mumane, H. C. Kapteyn, S. Backus, Nature 421, 51 (2003).
-R-
22. 3.-F. Ripoche, B. S. Prade, M. A. Franco, G. Grillon, R. Lange and A.
Mysyrowicz, Applications o f High Field and Short Wavelength Sources,
Plenum Press, New York (1998).
23. A. Rundquist, C.G. Durfee, Z. H. Chang, C. Heme, S. Backus, M. M.
Mumane, H. C. Kapteyn, Science 280, 1412 (1998).
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24. A. Rundquist, C. Durfee, Z. Chang, G. Taft, E. Zeek, S. Backus, M.
Mumane, H. Kapteyn, I. Christov, and V. Stoev, Appl. Phys. B 65, 161
(1997).
-S-
25. S. Sartania, Z. Cheng, M. Lenzner, G. Tempea, Ch. Spielmann, F. Krausz,
Opt. Lett. 22, 1562 (1997).
26. A. M. Streltsov, J. K. Ranka, and A. L. Gaeta, Opt. Lett. 23, 798 (1998).
-T-
27. M. J. Tauber and R. A. Mathies, J. Phys. Chem. A 105, 10952 (2001).
28. C. L. Thomsen, D. Madsen, S. R. Keiding, and J. Thogersen, and O.
Christiansen, J. Chem. Phys. 110, 3453 (1999).
-V-
29. V. H. Vilchiz, A. C. Germaine, V. A. Lenchenkov, J. A. Kloepfer, and S. E.
Bradforth, J. Phys. Chem. A 105, 1711 (2001).
-Z-
30. A. M. Zheltikov, Physics-Uspekhi 45, 687 (2002).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 1420368
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 1420368
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Femtosecond dynamics of electron photodetachment from coordinated anions: Hexaferrocyanide and tribromocuprate in solution

PDF

Infrared spectroscopy and ab initio studies of carbon dioxide van der Waals complexes

PDF

Helioseismic inversion procedures to probe the chemical composition and thermodynamics of the sun

PDF

Electron detachment of singly charged anions

PDF

Ultrafast electronic deactivation of DNA bases in aqueous solution

PDF

Design and synthesis of a new phosphine pincer porphyrin

PDF

Femtosecond laser studies of biological systems

PDF

Accurate calculations of bound- and quasibound-state energies of some three -body systems, and cross sections for photodecay

PDF

Holographic RG flows from IIB supergravity

PDF

Derivatization chemistry of mono-carboranes

PDF

Interaction of low -energy electrons with beams of sodium clusters, nanoparticles, and fullerenes

PDF

I. Layered nano fabrication. II. Adhesion layers for hippocampal neurons

PDF

Dye laser characterization of two films of baceriorhodopsin

PDF

Fluorinated carbocations and carboxonium ions

PDF

A measurement of the temperature dependent work functions of alkali metal clusters and implications of the scaling law

PDF

A transgenic mouse model for SCLC: Expression of Hel-N1 in mouse lung

PDF

Energy transfer dynamics in novel macrocyclic polymers: A comparative study of depolarization and exciton annihilation using ultrafast time resolved spectroscopy

PDF

Grain-size and Fourier grain-shape sorting of ooids from the Lee Stocking Island area, Exuma Cays, Bahamas

PDF

Spectroscopy of hydrogen and water clusters in helium droplets

PDF

A measurement of sodium cluster polarizabilities in an electric deflection experiment

Asset Metadata

Creator Jailaubekov, Askat (author)

Core Title Development of a 20-femtosecond tunable ultraviolet laser source towards study of the photochemistry of liquid water

School Graduate School

Degree Master of Science

Degree Program Chemistry

Degree Conferral Date 2003-12

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chemistry, physical,OAI-PMH Harvest,physics, optics

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Bradforth, Stephen (committee chair), [illegible] (committee member), Kresin, Vitaly V. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-310969

Unique identifier UC11328298

Identifier 1420368.pdf (filename),usctheses-c16-310969 (legacy record id)

Legacy Identifier 1420368.pdf

Dmrecord 310969

Document Type Thesis

Rights Jailaubekov, Askat

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA