. V. mwflmvfim
.‘t‘vgn J .. V
)1! lv. . V .. ‘F.
V «3.3 . . arm}. gamma
. nianfifi

3! a

.

mummy» , V . . V . . . , V V V .. V .“unnflurmmfi
. . ‘ . . V . . . V _ , V . , , 1....
. x V V . V . . .. 33%

...:....V .1,

n. iv...

 

. V . . V V . u. . Jr‘i‘; Desi}... V. t: . . . st: .31!
1.52:; . . . V . . ,fififi “$34”? a V V. V *3; ..

. x. 1; 1.951....

 

 

[7513856

900')"

This is to certify that the

dissertation entitled

COHERENT CONTROL AND ITS APPLICATIONS IN
MOLECULAR RECOGNITION

 

 

presented by

JOHANNA MARIE GALLARDO DELA CRUZ

 

LIBRARY
Michigan State
University

has been accepted towards fulfillment
of the requirements for the

 

 

 

Ph.D. degree in Chemistry

[til i i
W
ajor Professor’s Signature

 

#0 ’(Y [Z], ZOOC

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

 

COHERENT CONTROL AND ITS APPLICATIONS
IN MOLECULAR RECOGNITION

By

Johanna Marie Gallardo Dela Cruz

A DISSERTATION
Submitted to
Michigan State University
in partial fiJlfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

2006

ABSTRACT

COHERENT CONTROL AND ITS APPLICATIONS
TN MOLECULAR RECOGNITION

By
Johanna Marie Gallardo Dela Cruz

The fundamental design of this dissertation involves the use of the laser as a tool
to recognize molecules. The work presented here has relevance to a variety of fields,
including nonlinear optics and spectroscopy, coherent control of laser-matter interactions,
biomedical imaging and analytical chemistry. My work involves a series of experiments
utilizing a femtosecond laser system and an instrument known as a pulse shaper. It entails
new applications of coherent laser control, where tailoring the shape of laser pulses by
phase modulation causes changes in the impact of the excitation field on different
processes such as multiphoton excitation and photoionization.

The application of coherent control methods to biomedical imaging hinges on the
possibility that phase-modulated femtosecond laser pulses can propagate through
biological tissue. The approach used in this dissertation corrects spectral phase distortions
introduced by imaging optics, maintains laser pulse optimization throughout the
experiment and initiates phase manipulation in order to gain selectivity that is analogous
to that which can be achieved by tuning a longer laser pulse. This capability is used to
demonstrate functional imaging by selective two-photon excitation. Furthermore, phase
modulation is exploited as a means to curtail the damaging effects of multiphoton
excitation in biomedical applications, thus increasing the prospect for cellular and

molecular identification, including cancer detection.

The use of phase modulation of femtosecond laser pulses is also shown here as a
means to recognize isomeric molecules based on their ionization and fragmentation
patterns. The selectivity that is attainable with shaped pulse technology is coupled with
the inherent sensitivity of mass spectrometry to not only distinguish between isomers, but
to quantify them in isomeric mixtures. Although chemical identification is carried out
quite efficiently with standard mass spectrometry, the dilemma involving the
identification of isomeric compounds is addressed here by laser control. The
environmental, pharmaCeutical and analytical chemistry fields could stand to benefit from
this “customized” mass spectrometry approach. Additionally, the concept of utilizing

chiral femtosecond pulses as a tool for enantiomeric resolution is introduced.

ACKNOWLEDGMENTS

I am indebted to my advisor, Prof. Marcos Dantus, who has given me so much guidance,
insight and encouragement throughout this work. I am grateful to the members of my
committee, Profs. Gary Blanchard, James McCusker, and Greg Swain, who offered me
their time, advice and support. My colleagues in our research group have also been of
great assistance during my research. I thank the following past and present members of
the Dantus group: Dr. Vadim Lozovoy, Dr. Igor Pastirk, Dr. Matthew Comstock, Dr.
Tissa Gunaratne, Bingwei Xu, Michael Kangas, and Jess Gunn. Many thanks to Trish
Miller who has taken time to read my drafts and make comments. Thanks to the College
of Natural Sciences for awarding me a Dissertation Completion Fellowship, providing me
with the financial means to complete this project. My heartfelt appreciation to the
Department of Chemistry for supporting me as a teaching/research assistant so that I
could complete my graduate studies. Finally, I would like to thank my husband and my
children who have endured this long process with me, always offering support, patience

and confidence.

iv

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. viii

LIST OF FIGURES ................................................................................. ix

KEY TO ACRONYMS/ABBREVIATIONS ................................................... xix

INTRODUCTION ..................................................................................... 1

CHAPTER 1

PRINCIPLES .......................................................................................... 6
1.1 Multiphoton Intrapulse Interference (MII) ................................................ 7
1.2 Pulse Characterization and Compensation ............................................... 11
1.3 Phase Modulation ........................................................................... 17

CHAPTER 2

EXPERIMENTAL TOOLS ........................................................................ 21
2.1 F emtosecond Laser .......................................................................... 22

2.1.1 Ti:S Laser Oscillator System
2.1.2 Regeneratively Amplified Ti:S Laser System

2.2 Pulse Shaper ................................................................................. 27
2.2.1 SLM
2.2.2 Pulse Shaping Apparatus
2.2.2.1 The Linear Setup
2.2.2.2 The Folded Setup
2.2.3 Phase Modulation with the LQ-SLM
2.2.4 Limitations of the SLM

CHAPTER 3
SELECTIVE TWO-PHOTON EXCITATION .................................................. 36
3.1 Probing a Molecule’s Microenvironment ................................................ 39

3.1.1 Introduction

3.1.2 Experimental Section
3.1.3 Results

3.1.4 Discussion

3.1.5 Conclusion

3.2 Selective Two-Photon Microscopy ........................................................ 53
3.2.1 Introduction
3.2.2 Experimental Section

 

3.2.3 Results
3.2.4 Discussion
3.2.5 Conclusion

CHAPTER 4
COHERENT LASER CONTROL IN IMAGING .............................................. 62
4.1 Phase Modulation through Scattering Tissue ............................................ 64

4.1.1 Introduction

4.1.2 Experimental Section
4.1.3 Results

4.1.4 Discussion

4.1.5 Conclusion

4.2 Phase Compensation of Imaging Objectives ........................................... 71
4.2.1 Introduction
4.2.2 Experimental Section
4.2.3 Results and Discussion

4.3 Phase Compensation through Scattering Tissue ........................................ 75
4.3.1 Introduction
4.3.2 Experimental Section
4.3.3 Results and Discussion

4.4 Functional Imaging with Shaped Laser Pulses .......................................... 78
4.4.1 Introduction
4.4.2 Experimental Section
4.4.3 Results and Discussion
4.4.3.1 Functional Imaging
4.4.3.2 Loss of Coherence as a Function of Scattering
4.4.5 Conclusion

4.5 Minimizing Multiphoton Excitation Damage with Shaped Pulses .................... 92
4.5.1 Introduction
4.5.2 Experimental Section
4.5.3 Results and Discussion
4.5.4 Conclusion

CHAPTER 5
ISOMERIC IDENTIFICATION .................................................................... 97
5.1 Isomerism .................................................................................... 99

5.1.1 Structural isomerism
5.1.2 Stereo-isomerism

vi

5.2 Isomeric Identification by Laser Control and MS ..................................... 101
5.2.1 Introduction
5.2.2 Experimental Section
5.2.3 Results and Discussion
5.2.4 Conclusion

CHAPTER 6
CHIRAL PULSES AND CHIRAL RECOGNITION ........................................ 120
6.1 Polarization Control ....................................................................... 125

6.1.1 QWP Technique

6.1.1.1 Experimental Validation of Polarization Condition
6.1.2 LQ-Shaper Technique

6.1.2.1 Experimental Validation of Polarization Condition

6.2 Concept: Chiral Pulses and Binary Polarization ....................................... 133

6.3 Experimental Implementation ............................................................ 135
6.3.1 Polarization shaping with QWP technique: Initial experiments
6.3.2 Problems encountered with QWP technique
6.3.3 Polarization shaping with LQ-shaper technique
6.3.3.1 Simple Linear and Circular Polarization
6.3.3.2 lO-bit Binary Linear and Circular Polarization
6.3.3.3 Chirp Combined with CPL

6.4 Inferences and Future Possibilities for Chiral Pulses .............................. 150
CONCLUSIONS .................................................................................. 151
APPENDIX A
Group Velocity Dispersion .......................................................................... 154
APPENDIX B
Description of components of vacuum system — TOF assembly used for mass
spectrometry experiments ......................................................................... 157
BIBLIOGRAPHY ................................................................................. 166

vii

LIST OF TABLES

Table 6.1. Phase-Amplitude Modulation for LPL and CPL ................................. 131

Table 6.2. Two-bin Binary Polarization Masks ................................................ 143

viii

LIST OF FIGURES

Figure 1.1. M11 concept for two-photon transitions ............................................. 8

Figure 1.2. Calculated time and frequency domain representation of shaped pulses using
sinusoidal (A), quadratic (B) and cubic (C) phase functions. The introduction of spectral
phase modulation causes the amplitude in the frequency-doubled spectrum of the pulse
(with frequencies near 2600) to change. The first column under each phase shows the
spectrum of the pulse and the spectral phase modulation that is introduced by the pulse
shaper. The second column shows the time dependent electric field. The third column
depicts the effects of phase modulation on the frequency-doubled spectrum of the
laser ..................................................................................................... 9

Figure 1.3. Characterization (left) and compensation (right) of spectral phase modulation
in femtosecond pulses using the MIIPS method. MIIPS-generated traces (wavelength, It,
as a function of 6) show changes in the SHG spectrum of the laser. As a general rule,
quadratic phase distortions are determined from differences in the distances between the
four SHG features in the MIIPS scans. Cubic phase distortions are determined by
differences between the slopes of these SHG features. Computer analysis of the trace
retrieves the spectral phase of the input pulse. After a few iterations of characterization
and compensation, the output pulses are TL as evidenced by the parallel, equidistant
features in the MIIPS data (right). Notice that after compensation, differences in the
distances and slopes of uncompensated pulses are greatly diminished ....................... 16

Figure 1.4. Cartoon representation of the goal of phase modulation. The broad bandwidth
frequency-doubled spectrum from TL pulses is represented by a Gaussian (thin line).
The objective is to introduce phase modulation to cause the two-photon spectrum to be
intense only inside the window defined by frequency 2coc, and to minimize all other
regions of the spectrum outside the window ..................................................... 17

Figure. 1.5. Effect of spectral amplitude restriction on SHG. Left: Experimental spectrum
of the laser before (dotted lines) and after filtering with windows of width 20 nm
(window A) and 40 nm (window B). Right: Experimental SHG of TL (dotted lines) and
spectrally filtered pulses, A and B, as indicated in the right panel ........................... 18

Figure 2.1. Four-level scheme of a typical solid-state laser. The photons are first excited
to a state higher in energy than the upper laser state. They then quickly decay down by
relaxation or radiationless decay into the upper laser state. It is important for the pumped
state to have a short lifetime for spontaneous emission compared to the upper laser state.
The upper laser state should have as long a lifetime (for spontaneous emission) as
possible, so that the photons live long enough to be stimulated and thus contribute to the
gain. The fourth step involves depopulation of the lower laser level by a fast decay
(relaxation) process. This greatly decreases the loss of laser photons by stimulated

ix

absorption processes since the photons in the lower laser level have a short lifetime for

spontaneous emission ................................................................................ 23
Figure 2.2. The spatial profile of the laser beam creates the “Kerr lens” .................. 24
Figure 2.3. Ti:S laser oscillator system ........................................................... 25
Figure 2.4. Chirped pulse amplification .......................................................... 27

Figure 2.5. The LQ-SLM. A. Front view with pixel dimensions. B. Side view showing
components of the SLM. C. Direction of polarization as laser pulse goes through each LQ
mask ................................................................................................... 29

Figure 2.6. Schematic design of a linear 4f pulse Shaper using diffraction gratings and an
SLM ................................................................................................... 32

Figure 2.7. Schematic design of a folded pulse Shaper using prisms in lieu of
gratings ................................................................................................ 33

Figure 3.1. Schematic of the laser setup for MII experiments. The output of the Ti:S
oscillator is compressed by the prism pair compressor before arriving at the pulse shaper.
Phase and amplitude shaping take place at the SLM which incorporates a polarization
optic at its input and output windows. Once shaped, the pulses are directed towards the
experimental cell. Characterization of the pulses by MIIPS requires replacing the sample
cell by a thin SHG crystal; alternatively, the pulses are directed to an SHG-FROG ...... 42

Figure 3.2. Experimental MII results and simulations for HPTS in basic and acidic
solutions. Intensity of the two-photon induced fluorescence of HPTS solution at pH 10
(A) and pH 6 (B) are plotted as a function of the phase parameter 8. Simulations
(continuous lines) based on Equation 3.1 are in good agreement with the experimental
data (dots). The arrows at 8 =1/41t (dotted) and 8 = 3/4rr (dashed) represent phase
functions for which the intensity of two-photon fluorescence from HPTS at pH 10 is
similar but shows large differences for pH 6. The optimum two-photon excitation of
HPTS is indicated by vertical lines for the acid and basic solutions .......................... 45

Figure 3.3. Photochemistry of HTPS in basic (pH > 7.5) and in acid solutions (pH < 7.5).
Upon two-photon excitation, the equilibrium species in basic solution (3sPyO') reaches
an excited state that is solvated within 0.4 ps. Fluorescence from 3sPyO" has a lifetime
of 5.3 ns and is centered at 515 nm. The equilibrium species in acidic solution (3sPyOH),
upon two-photon excitation, reaches an excited state that is solvated within 0.3 ps.
Fluorescence from 3sPyOH’ is near 450 nm, however, deprotonation to yield (3sPyOJ)
takes place within 87 ps and is followed by fluorescence centered at 515 nm with a 4.8 ns
lifetime ................................................................................................ 47

Figure 3.4. Laser pulse characteristics in the time and spectral domain. The inset shows
the measured spectrum of the pulse 10.) and the phase ¢(3.) programmed by the SLM.

The pulses have a Gaussian shape with center wavelength M = 842.6 nm and 48 nm
FWHM. The temporal profile for two pulses (dotted and dashed lines) is calculated from
the experimental parameters for 6= U47: and 6 =3/47r, respectively. Notice that the two
pulses have the same temporal profile ............................................................ 49

Figure 3.5. Absorption spectra of HPTS in acidic and basic solution (bold lines). The thin
line is the calculated spectrum of the square of the electric field 053(1)?) of the laser
pulse without phase modulation (TL). The square of the electric field of the laser pulse is
also calculated for shaped pulses with 6 = 1/47r (dotted) and 3/47: (dashed), respectively.
The two spectra are very different because of MII. Notice that in the frequency domain, it
is clear that shaped pulses with 5 = 1/47r favor two-photon excitation of HPTS in pH 10,
while pulses with 5= 3/47rfavor two-photon excitation of HPTS in pH 6...... ....5.....()

Figure 3.6. Schematic experimental setup for selective two-photon microscopy.
Femtosecond laser pulses are compressed and sent to the pulse shaper. The modulated
beam is then focused on the microscope slide with the specimen. The two-photon induced
fluorescence is collected by a microscope objective and imaged on a CCD. . . . . . . . . . .......54

Figure 3.7. Experimental demonstration of pH-sensitive selective two-photon
microscopy. The sample being imaged has an acidic (left side, pH 6) and a basic (right
side, pH 10) region, both labeled with HPTS. A. Image of the sample obtained with TL
pulses. The spectrtun of the 21 fs laser pulses centered at 842 nm and the phase of the
pulse across the spectrum are shown in the middle diagram. The absorption spectra of
HPTS under acidic (dark gray) and basic (black) conditions are shown together with the
laser spectrum corresponding to EU). B. Image of the same sample and location obtained
with pulses that have been optimized for selective excitation of HPTS in an acidic micro-
environment. Notice that only the left region shows significant two-photon excitation.
The spectral phase function used for this experiment is given by ¢(Q)=acos(}Q—é),
where Q is the frequency detuning measured from the center wavelength of the laser
pulse. For this image a'= 1.57:, 7= 20 fs, and 6 = 0.757;. C. Image of the same sample and
location obtained with pulses that have been optimized for selective excitation of HPTS
in a basic micro-environment. Notice that only the right region shows significant two-
photon excitation. For this image a = 1.57r, 7 = 20 fs, and 6 = 0.252. Selective two-
photon excitation from the two pH regions in the sample is observed at specific values of
6 ........................................................................................................ 56

Figure 3.8. Selective two-photon microscopy of pieces of PMMA doped with different
fluorescent probes. A. Image showing two-photon induced fluorescence from both
pieces: C540 doped PMMA (top) and R6G doped PMMA (bottom), obtained with 17 fs
TL pulses centered at 790 mm. B. Image obtained with pulses optimized for selective
C540 excitation. For this image the pulses were shaped with a sinusoidal function with
6=0.311t. C. Image obtained with pulses optimized for selective R6G excitation. For this
image the pulses were shaped with a sinusoidal function with 6=0.741r. Selective two-
photon excitation from the two PMMA pieces is observed at specific values of 6 ........ 57

xi

Figure 3.9. Selective two-photon microscopy of 10 pm blue and 15 ,um green fluorescent
polystyrene microspheres. A. Image showing two-photon induced fluorescence from both
microspheres, obtained with 15 fs TL pulses centered at 790 nm. B. Image obtained with
pulses optimized for selective excitation of the blue microsphere. For this image the
pulses were shaped with a sinusoidal phase with 6=0.757r. C. Image obtained with pulses
optimized for selective excitation of the green microsphere. For this image the pulses
were shaped with a sinusoidal phase with 5=1.257r. Selective two-photon excitation from
the two microspheres is observed at specific values of 6 ...................................... 58

Figure 4.1. Experimental setup. The shaped laser pulses impinge on the sample where
fluorescence is detected in the forward direction. The sample, with or without scattering
tissue, is scanned in the focal plane of the laser while the two-photon induced
fluorescence is detected at each point ............................................................ 66

Figure 4.2. Optimized phase functions and the resulting SHG. The panel on the left
shows the fundamental spectrum of the pulse centered near 830 nm and the three
different pulses evaluated, TL with a flat spectral phase, and shaped pulses optimized for
acidic (BP06) and for basic excitation (BPlO). The panel on the right shows the
experimental SHG spectrum obtained from TL pulses and SHG spectra obtained when
the phase functions BP06 and BPlO were introduced by the pulse shaper. Notice that
BP06 suppresses SHG at long wavelengths and BPIO suppresses SHG at short
wavelengths .......................................................................................... 67

Figure 4.3. Fluorescence signal obtained from two capillaries containing HPTS in
buffered solutions. The signal was obtained after excitation with TL and shaped
pulses ................................................................................................... 68

Figure 4.4. Line scan of sample without (left) and with (right) tissue. The sample,
consisting of 5 capillary tubes containing acidic or basic HPTS and water (control), was
scanned in the x direction. Experimental results obtained with TL (black dots) pulses are
plotted in the negative axes. Difference plots obtained from the shaped laser pulses (open
circles) are plotted in the positive axes. Notice that shaped laser pulses are capable of
selective excitation even when the laser transmits through scattering tissue ............... 69

Figure 4.5. SHG spectra for TL and for shaped pulses (BP06) in the presence (open
circles, intensity multiplied by 20) and absence (black squares) of biological tissue. ....70

Figure 4.6. Compensation of spectral phase distortions caused by a 60x/ 1.45 NA
objective. The error bars indicate :t 1 standard deviation. The upper panel shows the full
range of data, while the lower panel shows the same data over the FWHM of the pulse,
on a reduced scale. The inset (top panel) shows the phase extracted before compensation
with MIIPS ........................................................................................... 73

Figure 4.7. Image quality improvement with spectral phase compensation. Left:
fluorescence image of mouse kidney using femtosecond laser pulses with spectral phase
deformations corrected. Right: fluorescence image of the same section of the sample

xii

when pulses are uncompensated (no spectral phase correction). Both images were
normalized to the same maximum intensity level. Image size is 30 ,um x 30 pm. . .7.....4

Figure 4.8. MIIPS measurement through scattering biological tissue. Panels A and B
show MIIPS traces without and with tissue, respectively. While the overall signal to
noise ratio is significantly decreased, similar phase information is obtained (C). . . . . ......77

Figure 4.9. Coherence degradation and pulse transformation as a function of scattering
path length. As a short pulse of light enters a scattering medium, coherent or ballistic
photons (sharp feature) are lost exponentially. The scattered photons (shaded broad
feature), which lag in time, lose their coherence and are randomly delayed. . . . . . . . . . . ......80

Figure 4.10. Molecular structure and absorption spectra of HPTS in acidic and alkaline
pH. Notice that the loss of the hydroxylic proton leads to a large change in the absorption
spectrum ............................................................................................... 81

Figure 4.11. Experimental setup used for selective two-photon excitation through
scattering tissue. The sample consists of three l-mm capillaries containing an acidic
solution of HPTS submerged in a quartz cell filled with an alkaline solution of HPTS.
The sample is raster scanned (without and with scattering tissue) in the focal plane of the
beam during data acquisition to obtain the images ............................................. 82

Figure 4.12. Images A and B are presented in color. A and B are photographs of the
sample obtained under white light illumination in the absence and presence of biological
tissue. In both of these photographs, the three capillaries in the sample are difficult to
distinguish. C and D are two-photon excitation images obtained by raster scanning the
sample without and with scattering tissue, respectively. The walls of the capillary tubes,
which look like vertical lines, are approximately 300 pm thick and are clearly visible in
both images. Notice that comparable image quality is obtained in the presence and
absence of biological tissue. Images C and D were obtained with TL pulses which are not
capable of discriminating between the two different pH solutions ........................... 85

Figure 4.13. Coherent control is used here to obtain pH sensitive functional imaging.
Panels A and B show images obtained by two-photon excitation using the optimized
phases BP06 and BPIO, respectively. Notice that the signal from the 1 mm capillary tubes
is enhanced for BP06 (A) and suppressed for BPIO (B). The top half of images C and D
are presented in color. Panel C shows a functional image highlighting the contrast
possible using coherent control. This image was obtained by taking the ratio of the data
obtained by BP06/BP10. Panel D shows the functional image obtained when a 1 mm
thick slice of biological tissue was placed in front of the sample. Notice that the presence
of the tissue reduces the overall signal to noise ratio, but the discrimination between
acidic and alkaline HPTS is conserved. The contrast in the functional images can be
further enhanced using false color (top half of Figures C and D). Higher values are
shaded red, while lower values are shaded blue ................................................ 86

xiii

Figure 4.14. Loss of coherent photons as a function of scattering. The transmitted ballistic
photons from TL (squares) and shaped (circles) pulses were measured as a function of
scattering. The phase shape for these measurements was chosen to suppress SHG
generation evenly through the entire spectrum. The intensity of the ballistic photons
measured at 800 nm (solid markers) was found to decay exponentially for both TL and
shaped pulses. The intensity of the ballistic photons was also measured after SHG from a
thin SHG crystal (open markers). A factor of two higher scattering was recorded in this
case given the square dependence of SHG, confirming that only ballistic photons
generate the second harmonic light. The spectrum and phase for the TL and shaped
pulses used in this measurement are shown in the upper inset. The spectra of the SHG
output from TL and shaped pulses are shown in the lower inset. Since both TL and
shaped pulses have similar slopes, they experience the same rate of scattering .......... 88

Figure 4.15. SHG spectra generated for TL pulses and shaped pulses optimized for leO
without a scattering medium and after propagating through a 15% skim milk in water
solution (equivalent to 2.1 scattering lengths). All the spectra were normalized to the
intensity they had in the absence of scattering. Scattering does not change the observed
SHG spectrum. The inset image on the left corresponds to three 15 ,um diameter
FluoSpheresTM behind a 1 mm thick slice of biological tissue. This image demonstrates
that high resolution (micron) can be achieved by the shaped pulses after they transmit
through biological tissue ........................................................................... 89

Figure 4.16. Suppression of three-photon excitation by phase modulation. Top:
fluorescence image of HPTS and L-trp deposited on a quartz cover slip and imaged with
TL 800 nm pulses. Middle: fluorescence from the same samples when laser pulses are
shaped with a sinusoidal phase function crossing zero at 0.25 1:. Bottom: Cross-section of
the fluorescence intensities of both dyes highlights the effect of phase modulation in
minimizing three-photon excitation (right) while essentially maintaining intensity of two-
photon fluorescence (left). These intensities (normalized) are compared with those
obtained using TL pulses. Each image size is 30 ,um (width) x 15 ,um

(height) ................................................................................................ 95

Figure 5.1. Classification scheme of isomers (with examples) ............................... 99

Figure 5.2. Schematic experimental setup for laser control MS for the identification of
isomers. The amplified femtosecond laser is sent to a pulse shaper that is programmed to
measure and compensate any phase distortions present on the laser pulse. Binary phase
functions are introduced at the Fourier plane, where the spectrum of the pulse is fully
dispersed. The phase-shaped laser pulses are focused inside a low pressure gas chamber
where they interact with the sample molecules, causing fragmentation and ionization. A
TOF mass spectrometer is used to record the arrival of the ions according to their mass-
to—charge ratio. Some binary phase functions cause large enough differences in the
fragmentation and ionization patterns that can be used to identify isomers and even
quantify their concentration in a mixture ....................................................... 105

xiv

Figure 5.3. Difference mass spectra of the various sample isomeric pairs tested: 3-
heptene, 4-methyl-2pentene (MP), cresol, and xylene. The maximum differences
obtained from the standard mass spectra (N IST) of isomeric compounds, particularly
ortho and para-xylene, make differentiation by conventional MS difficult. Note the
difference observed in xylene (2.6%) compared to the rest of the compounds tested. The
structures of the compounds are also shown ................................................... 108

Figure 5.4. Laser control mass spectra obtained for ortho- and para-xylene. The top panel
shows results obtained with BPO corresponding to TL laser pulses; the flat phase across
the spectrum of the laser pulse is shown in the inset. The bottom panel shows data
obtained with laser pulses with binary phase BP858, which is shown together with the
spectrum of the pulse in the bottom inset. The mass spectra were normalized so that the
molecular ion (M+) intensity equals unity. In the top and bottom panels, the region for
m/z near 91, corresponding to the tropylium ion (T), was amplified to highlight the
observed differences. For the BP858 shaped pulses, the relative yield of T+ for p-xylene
is 2.1 times greater than that observed for o-xylene .......................................... 109

Figure 5.5. Histograms showing the statistical distribution of the experimentally
measured ratio (M+/T+) obtained for 100 independent measurements each for o- and for
p-xylene. The top graphic shows the histograms obtained using laser pulses with phase
BPO; the bottom graphic shows the histograms obtained using laser pulses with phase
BP858. The insets show the temporal profiles of the two different laser pulses. The
histograms obtained for BP858 show a greater (~2x) molecular-ion-to-tropylium-ion
ratio. Note that in most cases the distribution of measured values is compact, indicating
good reproducibility of the results. The observed differences can be used for fast and
reliable identification between these two isomers .............................................. 111

Figure 5.6. Quantitative concentration measurements of isomeric mixtures using laser
control MS. Each data point represents an average of 100 measurements (0.1 8 total)
obtained for each of six different mixtures. The horizontal axis corresponds to the known
concentration of each of the mixtures. The vertical axis corresponds to the normalized
MTV ratio, which increases linearly with the relative concentration of o-xylene. This
linear relationship provides a quantitative method to determine the concentration of each
of the isomers (o- and p-xylene) in a mixture. Two sets of data obtained in two different
days are shown to illustrate the robustness of this method ................................... 112

Figure 5.7. Ion peak ratios (A+/B+) obtained for several isomeric pairs using binary
shaped pulses. The legend indicates the m/z values (A and B) of the fragments
chosen ............................................................................................... 113

Figure 5.8. Correlation between SHG or ion intensity and complexity of BP function. A.
General exponential decay of SHG intensity as phases are sorted based on the average
SHG intensities recorded during data acquisition of ortho- (black squares) and para-
xylene (open circles). B. Relation between ion intensity of mass peak at m/z 91 and
sorted BP function. C. Relation between ion intensity of mass peak at m/z 106 and sorted
BP function ......................................................................................... 116

XV

Figure 5.9. Left: Correlation of ion ratio intensity (m/z 91/106) and sorted phase function
for ortho- and para-xylene. Ortho-xylene is clearly distinguishable form para-xylene. The
difference is observable for most of the phase functions used. Right: Grouping the sorted
phases according to complexity shows maximum contrast (m/z 91/106) between the
isomers for more complex phase structures. Phase group 1 consists of 2 phases, group 2
contains 4, group 3 has 8, group 4 has 16, group 5 has 32, group 6 has 64, group 7 has
128, group 8 has 256, group 9 has 256, and group 10 consists of 258 BP functions. The
phases were sorted based on their effect on SHG intensity .................................. 117

Figure 6.1. Linear polarization. If the vector of the electric field (measured at a fixed
point of space) oscillates along a straight line then the waves are called plane polarized or
linearly polarized waves. Here, a wave is polarized in a horizontal (left) or vertical plane

(right) ................................................................................................ 126

Figure 6.2. Circular polarization. The superposition of two waves that have the same
amplitude and wavelength and are polarized in two perpendicular planes with a phase
difference of 90° between them generates circularly polarized light. A phase difference of
90° means that when one wave is at its peak, the other one is just crossing the zero line.
At any fixed point in space that is in the line of propagation of this wave, the electric field
vector rotates in a circle while its length remains constant. In this cartoon, the electric
field vector rotates counter-clockwise when looked at from the direction of propagation.
The light generated in this case is left circularly polarized. Clockwise rotation of the field
vector generates right circularly polarized light ............................................... 126

Figure 6.3. Polarization conversion by a quarter waveplate ................................. 128

Figure 6.4. Polar plots representing measurements for LPL and CPL generated from a
QWP. Left. Horizontal (BP-O) and vertical (BP-rr) polarization plots show definite but
slightly warped production of LPL. Right. Even when retardance of the QWP is
modified to conform to specifications for modifying polarization, the light generated is
not circular .......................................................................................... 130

Figure 6.5. Polar plots representing measurements for LPL and CPL generated from the
LQ-pulse shaper technique. Left. Horizontal (BP-O) and vertical (BP-n) polarization plots
show definite production of LPL. Right. When the program is modified to produce
circular polarization, the light generated is elliptical .......................................... 132

Figure 6.6. Phase mask with 0 or 113 retardation binned to 10 super pixels, each 10 pixels
wide, covering the full spectrum of the fundamental laser pulse. The SLM used has 128
pixels ................................................................................................ 135

Figure 6.7. Electric fields for binary polarization pulse BP167 for right and left circular
polarization ......................................................................................... 136

xvi

Figure 6.8. Mass spectra of or-methyl-benzenemethanol (BME). Top: electron ionization
spectra of R- (Open circles) and -S-BME (black circles). Bottom: special fragmentation
mass spectra of R- and S-BME using shaped laser pulses with circular binary
polarization. Peaks A and B refer to fragment ions at m/z 51 and 107, respectively. The
ratio of intensities of these peaks was used to differentiate the enantiomers. Inset:
structures of the R and S enantiomers of BME ................................................ 138

Figure 6.9. Reactions involved in the fragmentation of BME ............................... 139

Figure 6.10. Statistical evaluation of phase and polarization shaping for enantiomeric
selectivity. Top: histograms representing 100 measurements for each enantiomer using
circularly polarized (BP167) laser pulses. Bottom: histograms representinglOO
measurements for each enantiomer using linearly polarized (BP167) femtosecond laser
pulses ................................................................................................ 140

Figure 6.11. Collected mass spectra of R- and S-BME with right circular and left circular
polarization combination. The spectra were normalized to the highest fragment ion peak
at m/z 79. Note the presence of the peak at m/z 91 (encased) for S-BME .................. 144

Figure 6.12. Mass spectra of R- and S-BME using lO-bit polarization shaping. Left:
observed mass spectra of the enantiomers as a consequence of TL LCPL. Right: observed
mass spectra of the same compounds when a more complicated circularly polarized
binary function is applied. The structures of the binary functions used are also shown
(phase) ............................................................................................... 145

Figure 6.13. Changes in intensity of fragment ion peaks at m/z 107 (A) and m/z 51 (B)
as 128 circularly polarized binary functions are scanned across the SLM. The intensities
were normalized to that of m/z 107. C. Plot of the ratios between m/z 51 and 107 for the
128 polarization shaped pulses. Error bars indicate standard deviation of data
(n=10) ................................................................................................ 146

Figure 6.14. Raw TOF data for R- and S-BME showing peaks with intensity differences.
The peaks shown represent fragments at m/z 105, 106 and 107. The inset zooms in on
peaks 105 and 106, highlighting the difference between the two enantiomers ............ 148

Figure 6.15. Mass spectra of R— and S-BME using chirped lO-bit polarization shaping.
Left: observed mass spectra of the enantiomers for positively and negatively chirped left
circularly polarized pulses. Right: observed mass spectra of the same compounds when
positive and negative chirp are applied in addition to a complicated circularly polarized
binary pulse. The structures of the binary functions used are also shown (phase) ........ 149

Figure A-l. Prism arrangement to compensate for positive GVD ........................... 155
Figure A- 2. Major components of TOF mass spectrometry system include the mechanical

roughing pump (not shown), diffusion pump, turbomolecular pump, ion gauge, flight
tube and the MCP detector ........................................................................ 159

xvii

Figure A-3. Diffusion pump mode of operation. A heating coil (1) evaporates the special
diffirsion pump oil in the boiler (2). The oil vapor then ascends through the center of the
pump and shoots at high speed out of a multi-stage system of concentric ring-shaped
nozzles (9-12). The oil vapor shoots out in the shape of a cone (7) and then condensates
against the wall of the pump (3), which is cooled by water flowing through the cooling
coils (4). The oil then returns into the boiler as a thin film along the wall. The air or other
gas in the high vacuum part of the pump (6) is forced to move along with the high speed
oil vapor in the different cascade levels and leaves the pump on the low vacuum side of
the pump (8) ........................................................................................ 160

Figure A-4. A typical ion gauge. It has a heated filament biased to give thermionic
electrons (70c) energetic enough to ionize any residual gas molecules during collisions.
The positive ions formed drift to an ion collector held at about 150V. The current
measures gas number density, a direct measure of pressure ................................. 161

Figure A-S. Lens stack and flight tube assembly. An assembly of grids, steering plates
and lens elements mounted on a flange ......................................................... 163

Figure A-6. MCP principle of operation ........................................................ 164

xviii

BME
BP
BPS
CD
CPA
CPL
CW

DI

EI
FROG
FWHM
GA
GC
GVD
HPTS
ID
IMS/MS
IR
KLM
LC/MS
LCPL
LIF
LPL
LQ '
m/z
MCP
MII
MIIPS
MP

MS

NA
NszLF
NdIYVO4
NIST
ORD
PDT
PMMA
QWP
RCPL
SHG
SLM

KEY TO ACRONYMS/ABBREVIATIONS

a-methyl-benzenemethanol

Binary phase

Binary phase shaping

Circular dichroism

Chirped pulse amplification

Circularly polarized light

Continuous wave

Dissociation followed by ionization
Electron ionization

Frequency resolved optical gating

Full width at half maximum

Genetic algorithm

Gas chromatography

Group velocity dispersion
8-hydroxypyrene-1 ,3,6-trisulfonic acid trisodium salt
Ionization followed by dissociation

Ion mobility spectrometry/mass spectrometry
Infrared

Kerr lens mode locking

Liquid chromatography/mass spectrometry
left circularly polarized light

Laser induced fluorescence

Linearly polarized light

Liquid crystal

Mass to charge ratio

Multichannel or microchannel plate
Multiphoton intrapulse interference
Multiphoton intrapulse interference phase scan
4-methyl-2-pentene

Magnetic resonance imaging

Mass spectrometry

Numerical aperture

Neodymium: yttrium lithium fluoride
Neodymium: yttritun vanadate

National Institute of Standards and Technology
Optical rotatory dispersion

Photodynamic therapy
Polymethylmethacrylate

Quarter waveplate

Right circularly polarized light

Second harmonic generation

Spatial light modulator

xix

SPIDER Spectral interferometry for direct electric field reconstruction
TOF Time of flight

Ti:S Titaniumzsapphire

TL Transform—limited
TMP Turbomolecular pump
UV Ultraviolet

XX

INTRODUCTION

When this research project was conceived, the idea of teaching lasers to produce
photons that are smart enough to recognize molecules seemed implausible. The full
potential of lasers had barely been realized and questions as to whether the utility of
lasers could be expanded from sheer spectroscopic tools to photonic tools for probing
molecules had yet to be answered. Since the first caveman learned to control fire,
researchers have shaped and used laser light in a constantly expanding array of
technologies. Yet “smart lasers” could do much more. This dissertation focuses on our
quest to transform the cutting-edge science of ultrafast laser technology into a versatile
tool that can be used by the biomedical research community on the one hand, and, on the
other hand, open up possibilities for applications in real-time chemical sensing and
detection that will benefit a wide range of disciplines. It is our hope that smart lasers will
find a niche not only in the research laboratory but in the real world. The key to these
aspirations is molecular recognition by coherent laser control.

While in the past, lasers had mainly been used to investigate given physical
systems, currently scientists are exploiting the coherent properties of lasers to manipulate
or control matter. The field of coherent control explores the ability of a coherent laser
source to manipulate the dynamics at various stages of a process. Its rapid development
has been stimulated by the objectives of selective bond breaking and formation. The first
concepts for controlling chemical reactions in the time domain with ultrashort laser
pulses were developed theoretically by Tannor, Rice and Kosloff" 2 and realized
experimentally by Gerber3 and Zewail.4 However, the applications of coherent control are

by no means limited to controlling chemical reactions. This research explores coherent

control as a tool for the selective recognition of molecular components, which could
provide benefits in the biomedical and analytical chemistry fields.

Ultrafast laser technology involves the use of femtosecond (1 fs=10'15 seconds)
lasers to study the properties of matter. The extremely short pulse duration allows one to
create, detect and study short-lived chemical reaction intermediates and transition states.
For the analytical chemist, femtosecond lasers can be used to monitor transient chemical
species in solution or in the gas phase, to image living cells with sub-micrometer
resolution, for laser ablation, and micromachining applications.

Ultrafast pulses consist of a broad range of phase-locked frequency components.
They can be represented as a coherent superposition of several monochromatic light
waves within a range of frequencies that is inversely proportional to the duration of the
pulse. The ultrashort time duration of ultrafast pulses results in a very broad spectrum;
this allows one to access a wide range of different frequencies. Possibilities for
manipulating ultrafast coherent broadband sources in addition to technological
developments in ultrafast lasers have just begun to unleash the potential of the exciting
field of ultrafast pulse shaping.

Pulse shaping involves either the simultaneous or exclusive control of the
amplitude and phase of different frequency components in the spectrum of a femtosecond
laser pulse. This has led to a surge of activity in coherent control. The time—bandwidth
profile, in conjunction with the phase characteristics of the pulse, is commonly referred to
as pulse shape. Many of the frequencies constituting the ultrafast pulse can
simultaneously excite many coherent transitions to the excited states, and the capability

to manipulate them with shaped pulses leads to interesting results. In 1984, Warren

1“

 

 

suggested a theoretical application of shaped pulses in the time domain for nonlinear
optical applications such as population transfer.5 The role of the spectrum of the shaped
pulse was found to be indispensable in the choice of the optimal temporal shape that
would influence whether or not population transfer occurs. Warren tested his proposal
experimentally in 19876 using phase and amplitude shaped nanosecond (100-200 ns)
laser pulses to control I2 fluorescence . A year after this, Weiner, Nelson and co-workers
investigated the effect of pulse shaping on impulsive stimulated Raman scattering with
the objective of mode-selective vibrational amplification in an alpha-perylene molecular
crystal.7’ 8 Following this breakthrough, Broers et. a1. introduced binary amplitude
modulation to demonstrate spectral focusing of second harmonic generation (SHG) and
two-photon absorption of Rydberg states in Rubidium.9’ '0 In 1992, Judson and Rabitz
suggested the use of learning algorithms to optimize the phase and amplitude of
femtosecond pulses to control optical processes.ll Computer feedback algorithms based
on genetic or evolutionary strategy techniques have made it possible for an experiment
itself to select an optimal optical waveform to enhance a particular process. This was
experimentally implemented in 1997,12 and since then, the use of feedback optimized
genetic algorithms (GA) has gained a tremendous following from several research
groups,”3 I while paving the way towards a number of valuable applications.”58 In 1998,
Meshulach and Silberberg demonstrated that femtosecond laser pulses can be tailored to
control two-photon absorption in cesium.59 They soon after showed that the effect of
shaped pulses in exciting sharp transitions in small molecules was absent when large
molecules in solution are involved.60 In 1999, variations in chirp conditions of the laser

pulses were applied to control the fluorescence yield of molecules in different micro-

environments.61 Silberberg and co-workerséz’ 63 later demonstrated the constructive or
destructive nature of interferences between resonant and nonresonant contributions in
two-photon excitation. They revealed that amplitude control and phase modulation with
step-functions could amplify two-photon fluorescence from Rb atoms several times.
Leone and co-workers used feedback and evolutionary algorithms with a phase shaper to
optimize the interference between resonant and nonresonant contributions to multiphoton
excitation of Liz.64 Motivated by the work of Silberberg, the Dantus group began its
pursuit of the laser pulse requirements for controlling multiphoton excitation of large
organic molecules in solution. This effort subsequently resulted in the development of a

165’ 66 which had a direct influence on the experiments

new approach to coherent contro
covered in this dissertation.

By appropriately controlling the arrival time of different frequency components of
an ultrashort excitation pulse, one can readily control ground state and excited state
dynamics, or cause chemically selective molecular excitation or ionization, without
having to resort to wavelength tuning by tweaking the laser, using filters or moving
optics. The creation of tailored pulses and their use in attempts to control atomic and
molecular events is an innovative research area. Recent successes in molecular control
have resurged hopes towards bond-selective chemistry. As pulse-shaping technology
matures further, it will open up exciting new applications in areas as diverse as
biomedical research (imaging, diagnostics and therapeutics) and quantum computation.

Furthermore, computer technology has advanced at a pace that allows us to take full

advantage of progress in optical technology.

In the following chapters, the essential principles (Chapter 1) that governed this
research and the major experimental tools (Chapter 2) that were crucial to the success of
the experiments will be presented. Chapters 3 and 4 discuss the experimental
demonstration of selective two-photon excitation as a tool for the identification of
molecules based on their chemical environment. The last two chapters discuss the
analytical implementation of coherent control in qualitative and quantitative molecular
identification of geometric and structural isomers (Chapter 5) and of optical isomers
(Chapter 6), including the setbacks encountered when exploring enantiomeric resolution
with chiral laser pulses. These impediments, however, do not necessarily suggest the
impossibility of enantiomeric discrimination by chiral pulses. This is just the beginning

of converting this seemingly unattainable dream into reality.

The only way of discovering the limits of the possible is to venture a little way past them
into the impossible."

- Arthur C. Clarke's Second Law

 

 

CHAPTER 1

PRINCIPLES

When considering laser control of large molecules, the large numbers of degrees
of freedom where energy can flow imply that the control field must be complex enough
to track the evolution of the wave packet in a multidimensional potential. The complexity
arises because different frequencies are required to interact with a sample at specifically
coordinated times. Based on this representation, coherent laser control of large molecules
seems extremely difficult. However, the Gerber group showed the capability of certain
shaped pulses obtained from learning algorithms for inducing selective photochemistry in
isolated molecules.”’ 53 Gerber and Motzkus have published adaptive control experiments
on large molecules in condensed phase.” 5 I They have both argued that the molecules are
under coherent control during the pulse, which in both cases extends for a couple of
picoseconds. These results contradict the measurement of electronic coherence time by
the group of Shank, where it was determined that the electronic coherence time in large
molecules in solution decays within the first 70 fs.67 The experiments of Shank were
carried out using the photon echo method, which cancelled the contributions from
inhomogeneous broadening. If inhomogeneous dephasing is taken into account, as it
should for a single pulse experiment, then dephasing occurs faster than 10 fs. This would
imply that the mechanism involved discounts coherent control of the intramolecular
dynamics of the sample molecule.

In the condensed phase, both homogeneous and inhomogeneous broadening,

which affect both the clarity and location of absorption lines, make it virtually impossible

to consider excitation of a specific vibronic transition. Such processes preclude the
creation of pulses that can address a specific vibronic transition in all the molecules in the
bulk sample. Consequently, the perception of coherent control of molecules in the

condensed phase calls for a new paradigm.

1.1 Multiphoton Intrapulse Interference (MII)

The MII approach of coherent control is based on the delivery of small packets of
energy to molecules by means of an ultrafast laser pulse. Its main advantage lies in its
ability to control the order of a transition (two-photon, three-photon or higher) while
maintaining a constant and intense photon flux. Controlling excitation consequently
allows control over the details of how molecules interact with light pulses.

The underlying concept of the M11 method is shown in Figure 1.1. Multiphoton
transitions are optimized when the central bandwidth of the laser pulse, (do, is some
proportion (1/2 for two-photon, 1/3 for three-photon, etc.) of the total energy of the
transition. The large bandwidth of ultrashort pulses permits interference between the
different frequency components.9 The phase of each of the frequencies within the
bandwidth determines how photons combine to achieve multiphoton excitation. Figure
1.1 shows the spectrum of an ultrashort pulse, with amplitude plotted as a function of
detuning from the central frequency. A phase mask, ¢(Q), can be imprinted on the pulse
such that the phase of each frequency component, Q, acquires a specific value. MII
requires phase functions that are comparable to the homogeneously broadened absorption

5

spectrum and the spectral width of the pulse.6 The phase function depicted here is

defined by the following equation,

¢(Q)=acos(;fl—6) 1.1

 

 

to consider excitation of a specific vibronic transition. Such processes preclude the
creation of pulses that can address a specific vibronic transition in all the molecules in the
bulk sample. Consequently, the perception of coherent control of molecules in the

condensed phase calls for a new paradigm.

1.1 Multiphoton Intrapulse Interference (M11)

The MII approach of coherent control is based on the delivery of small packets of
energy to molecules by means of an ultrafast laser pulse. Its main advantage lies in its
ability to control the order of a transition (two-photon, three-photon or higher) while
maintaining a constant and intense photon flux. Controlling excitation consequently
allows control over the details of how molecules interact with light pulses.

The underlying concept of the M11 method is shown in Figure 1.1. Multiphoton
transitions are optimized when the central bandwidth of the laser pulse, (00, is some
proportion (1/2 for two-photon, 1/3 for three-photon, etc.) of the total energy of the
transition. The large bandwidth of ultrashort pulses permits interference between the
different frequency components.9 The phase of each of the frequencies within the
bandwidth determines how photons combine to achieve multiphoton excitation. Figure
1.1 shows the spectrum of an ultrashort pulse, with amplitude plotted as a function of
detuning from the central frequency. A phase mask, ¢(Q), can be imprinted on the pulse
such that the phase of each frequency component, 9, acquires a specific value. MII
requires phase functions that are comparable to the homogeneously broadened absorption

5

spectrum and the spectral width of the pulse.6 The phase function depicted here is

defined by the following equation,

¢(Q)=acos(7Q-—5) 1.1

where (1 refers to the phase amplitude, y is the modulation in the frequency domain

(27W /(Q max—(2min) , with period N) and 6 is the position of the phase mask with

respect to the center of the spectrum of the laser pulse.

 

Spectral Phase
:99) 9.63)

 

 

 

 

 

 

   

 

 

 

Figure 1.1. M11 concept for two-photon transitions.

MII was inspired by the work of Silberberg59 who demonstrated control of two-
photon excitation of a sharp atomic transition using initially a step function, and then a
sinusoidal phase function. Figure 1.2 shows how the fundamental spectrum of a pulse in
the frequency domain is unaffected when a given spectral phase function is used. On the
contrary, the phase changes the pulse in the time domain and may significantly affect the
frequency-doubled spectrum. The addition of a linear phase function across the
bandwidth of the pulse in the frequency domain results only in the delay in time of the
pulse and has no effect on its nonlinear interaction with the molecules. A quadratic phase
modulation (Figure 1.2B) advances or delays the pulse in time, but leads to broadening.
The frequency-doubled spectrum is basically unaffected, except for a substantial decrease

in intensity. Consequently, a correspondingly reduced cross-section to nonlinear

processes is expected. Large changes in the nonlinear excitation processes are expected to
occur when odd-order phase functions such as cubic (Figure 1.2C) and higher
components are used.66 A simple sinusoidal phase function (Figure 1.2A) focuses the
energy in a narrower spectral region. The wavelength where the frequency-doubled

spectrum is maximized can be controlled by shifting the sine function.

. Ill . . _ ,
i 770 ‘ 800 ‘ 830 -5b0 o 500 3190 460 410
l t . . .

 

A. sinusoidal phase

* l

 

 

 

B. quadratic phase , d)" = 6000fs2

1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 7t , nm t, f5 7» . nm
1 C. cubic phase 7 41'": 105fs3
fl . /
l.
770 ‘ 860 ‘ 830 300 o 560 390 ‘ 460 ‘ 410
77018001830600 ('1 560 390 ‘460‘410 tnnm hfs an
k , nm t, fs 1. , nm

 

 

 

Figure 1.2. Calculated time and frequency domain representation of shaped pulses using
sinusoidal (A), quadratic (B) and cubic (C) phase functions. The introduction of spectral
phase modulation causes the amplitude in the frequency-doubled spectrum of the pulse
(with frequencies near 266)) to change. The first column under each phase shows the
spectrum of the pulse and the spectral phase modulation that is introduced by the pulse
shaper. The second column shows the time dependent electric field. The third column
depicts the effects of phase modulation on the frequency-doubled spectrum of the laser.

To illustrate how MII takes place, consider a two-photon process such as SHG.
Assuming that the fundamental pulse is centered at frequency coo, maximum SHG then
takes place at 2m). Silberberg59 first realized that the intensity of SHG at 2m) not only
depended on the intensity of the ultrashort laser pulse at at), but more importantly on the

entire bandwidth of the pulse. He further observed that the relative phase between the

blue-shifted frequencies ((00 — Q) and the red-shifted frequencies (can + Q) determined the
intensity at 2m) according to

co
15).,G (2010) cc [Ema + ammo — Q)exp[i{¢((oo + Q) + ((00 — 9mm 1.2

—oo

In this expression, the interference terms in the brackets imply that there can be
constructive interference when the sum of the phases is zero, or destructive interference
when the sum of the phases is 7:. Even functions lead to a decrease in signal while odd
frmctions increase signal intensity at 2a». This has led to the formulation of an equation
that could be used to calculate the nth order transition probability to a frequency that is

detuned from the central wavelength of the laser pulse by an amount A. The expression:

°° 2
S,, at jg(")(A)lA(")(A)| dA, 1.3

—oo
correlates the nth-order spectral amplitude of the laser pulse A‘") with the nth-order
absorption spectrum of the system, g(")(A). With these equations as basis, the effect of
phase modulation from all multiphoton transitions, not only at net), but through the entire
bandwidth of the pulse n( an+A), can be calculated. This in turn allowed the application
of MII to multiphoton transitions of large organic molecules, including proteins, in the
condensed phase.“’ 66 The first order spectrum of the field (one-photon excitation) is not
affected by the phase function because of the absence of an interference term at this
order. The introduction of phase functions has been shown to achieve much more than
simple attenuation of the laser peak intensity. The nth order spectrum (n-photon
transitions) can be tuned to specific resonances, and phase can be used to attenuate third-

order processes.66

10

Through MII, modest phase modulation can be used to manipulate nonlinear
transitions. This capability can initiate the development of techniques that would enable
shaped femtosecond lasers to selectively excite molecular reactions and alter the course
of biological processes. Accordingly, coherent control through MII can be employed as

an initial step towards molecular identification.

1.2 Pulse Characterization and Compensation

Ultrashort pulse laser technology has advanced remarkably over the past two
decades, expanding its niche in the research community from home-built sub-picosecond
dye systems to contemporary commercial tum-key sub-50 fs amplified systems to the
generation of exceedingly short pulses. Most researches have concentrated on pulses in
the visible or near infrared (IR) part of the light spectrum. At these wavelengths, an
optical wave takes about 3 fs to complete one oscillation, so that state-of-the-art 5-fs laser
pulses correspond to less than two optical cycles. Optical pulses much shorter than this
require substantial effort to produce, and since light propagates as an oscillating
electromagnetic wave, they are fundamentally limited to the single-cycle duration (time
taken for the propagating optical wave to complete one full oscillation) of about 3 fs.
Huge advances in the generation and manipulation of ultrashort laser pulses have also led

“'70 where pulses are in the

to the generation of pulses with attosecond time resolution,
UV/XUV spectral range, and are orders of magnitude weaker and spectrally much
broader than femtosecond pulses. These types of pulses, however, are beyond the scope

of this discussion.

11

Femtosecond pulsed lasers have provided unprecedented insights into molecular
processes such as reaction dynamics. In the last several years, femtosecond lasers (~50-
100 fs) have been indispensable tools in scientific and development laboratories,
including several industrial, communication, and medical applications. Consequently, a
whole new set of more stringent requirements and constraints for the everyday use of
ultrashort laser sources needs to be addressed.

The characteristics and quality of ultrashort pulses play a crucial role in the fields
or processes they are used for or applied to. As the pulses get shorter, however, they
become more prone to spectral phase distortions that take place in the process of their
generation and their subsequent delivery to the target. Every interaction of an ultrashort
pulse with an optical surface such as a dielectric mirror, lens, optical fiber, or microscope
objective leads to spectral phase distortions. These distortions will consequently lead to
increased pulse duration, loss of peak intensity and, in some cases, loss of information.
Shorter pulses suffer more significant distortions as a consequence of group velocity
dispersion (GVD, see Appendix A), whereby different frequency components of the short
pulse travel at different speeds, causing the pulse to eventually fall apart. For this reason,
the characterization and correction of these spectral phase distortions, especially at the
location where the laser meets the sample, are extremely important. These distortions,
which are either intentional or caused by experimental imperfections, are generally
addressed on a case-to-case basis. Careful characterization of the laser pulse’s temporal
profile, bandwidth, intensity, and spectral and temporal phase are essential for successful
application. Due to the uncertainty principle, the time-bandwidth product is a natural

limit and has persistently been an ultimate goal in many applications. The pulses that are

12

at or very close this limit are referred to as transforrn-limited (TL), which means that they
are shortest for a given bandwidth. The atomic and molecular processes initiated and
driven by TL pulses can be relatively easily simulated and understood.

Traditional ways of basic pulse characterization involve measurement of a pulse’s
bandwidth and time profile by means of intensity autocorrelation, where a pulse is split
and measured against itself by nonlinear interaction. The corresponding spectral phase of
the pulse, unfortunately, stays undetermined using these methods. Current established
spectral phase measurement methods include frequency resolved optical gating
(FROG)71’ 72 and spectral phase interferometry for direct electric filed reconstruction
(SPIDER)73’ 74. These methods, including their variants, depend upon the measurement
of autocorrelations or cross-correlations between two pulses to characterize the spectral
phase of a pulse. In FROG, the frequency and time-resolved signals are used to retrieve
spectral phase. SPIDER requires that the laser pulse be split into two beams that are then
upconverted by a heavily chirped pulse. The interference of the upconverted pulses in the
frequency domain is used to reconstruct the spectral phase in the original pulse. These
methods have been implemented for more than a decade, but they suffer from either
complex or demanding set-ups (SPIDER) or cumbersome and. approximate phase-
retrieving algorithms (FROG) that prevent them from being used without the aid of a
laser expert. The first attempts to achieve TL pulses based on feedback dependent on
spectral phase used a GA-controlled shaper for adaptive compression.” 16’ 19’ 23’ 25’ 26' 28' 30’
75 Others implemented time-domain interferometry with an acousto-optic programmable

filter.76 These indirect methods are limited by the lack of sensitivity of the total nonlinear

l3

optical signals to phase deformations outside the full width at half maximum (F WHM) of
the pulse.

The use of fully characterized ultrashort pulses in the experimental setting should
not be disregarded. In this research, the multiphoton intrapulse interference phase scan
(MIIPS) method is used for pulse characterization and compensation of spectral phase
distortions. This technique combines spectral phase characterization and pulse shaping in
one simple set-up. MIIPS takes advantage of the influence that phase modulation has on
the probability of nonlinear optical processes at specific frequencies.” 66 It is a single
beam method that does not require beam splitters, interferometry, autocorrelation or
global maximization. The spectral phase across the pulse can be obtained analytically
with this method. Within a few minutes, the pulses are characterized and compensated to
yield TL or user-specified shaped pulses at the sample. This capability ,is extremely
practical and can be incorporated in any laser setup. Consequently, this method has
proven to be extremely powerfirl for the accurate and reproducible demonstration of the
experiments discussed in this work - selective microenvironment probing, multiphoton
microscopy, functional imaging and discrimination between isomeric compounds -
using ultrashort shaped pulses.

The principle of the measurement involves comparing an unknown phase against
a calibrated phase. The active component of the MIIPS system is a computer-controlled
spatial light modulator (SLM) located at the Fourier plane of a pulse compressor. The
computer introduces a reference phase function and records changes in the frequency-
doubled spectrum of the pulses. To make a precise and accurate measurement of the

spectral phase using MIIPS, the known phase requires an accurately calibrated pulse

14

shaper. The pulse shaper essentially behaves as two back-to-back spectrometers. The
pulse is dispersed with a prism or grating and collimated with a cylindrical mirror. At the
Fourier plane, where all the frequencies are isolated, their phases are manipulated by a
computer-controlled liquid crystal (LQ)—SLM. The SLM applies the reference phase
function to the input pulse, and the resulting pulse is then reconstituted to the time
domain by a second cylindrical mirror and prism/grating. The SLM has a 250-ms
response time. The output beam is analyzed by placing a 0.01-mm-thick beta barium
borate SHG crystal in its path, usually at the place where optimum pulses are required.
For each reference phase function that is introduced by the computer-controlled SLM, the
output spectrum from the SHG is dispersed in a spectrometer and recorded. The reference
phase function is programmed into the SLM and scanned for different values of 6 ranging
from 0 to 4n. For each value of 6, the spectrum of the frequency-doubled pulse changes,
achieving a maximum in the spectral region over which the SLM compensates for the
phase distortions.

The MIIPS data consist of a two dimensional collection of SHG spectra as a
function of 6 (as shown in Figure 1.3). For TL pulses, the data show equally spaced
parallel linear features. Deviations from these parallel lines indicate distortions in the
spectral phase. The spacing between the lines indicates the degree of quadratic phase
distortions. The slope that the maximum SHG features makes in the 2—6 plane

determines the cubic phase modulation.

15

 

 

TE 27: 3n 4r:
6 (rad)

 

Figure 1.3. Characterization (left) and compensation (right) of spectral phase modulation
in femtosecond pulses using the MIIPS method. MIIPS-generated traces (wavelength, 9»,
as a function of 6) show changes in the SHG spectrum of the laser. As a general rule,
quadratic phase distortions are determined from differences in the distances between the
four SHG features in the MIIPS scans. Cubic phase distortions are determined by
differences between the slopes of these SHG features. Computer analysis of the trace
retrieves the spectral phase of the input pulse. After a few iterations of characterization
and compensation, the output pulses are TL as evidenced by the parallel, equidistant
features in the MIIPS data (right). Notice that after compensation, large differences in the
distances and slopes of uncompensated pulses are greatly diminished.

Once the MIIPS system has characterized the pulse and retrieved the phase
distortions inherent to the pulses, it can use this information to drive the SLM such that it
compensates for the distortions. The first step in compensation is to take the phase
determined from the first scan and program it into the SLM with a negative sign so that it
subtracts the distortions. The system then carries out a new phase scan to determine the
remaining spectral phase modulation (usually about 10% of the original). A few such

iterations will typically eventually yield TL pulses.

1.3 Phase Modulation

Complete control over multiphoton transitions entails reaching the maximum
limited by TL pulses at some interesting frequencies and suppressing effects at different
unwanted frequencies, as illustrated in Figure 1.4. The realization of this task lies in the

accurate manipulation of either the phase or amplitude of a laser pulse.

 

 
 
  

       
    

 

 

 

1‘ \

\
P 4

(I)
. .\

c - II
9 II
5 :1
‘1’ :22:
0) RI
\
b
0. r

Frequency

Figure 1.4. Cartoon representation of the goal of phase modulation. The broad bandwidth
frequency-doubled spectrum from TL pulses is represented by a Gaussian (thin line).
The objective is to introduce phase modulation to cause the two-photon spectrum to be
intense only inside the window defined by frequency Zinc, and to minimize all other
regions of the spectrum outside the window.

Amplitude modulation only affects multiphoton transitions by decreasing pulse
energy and consequently decreasing multiphoton effects (see Figure 1.5). In effect,
control over which frequencies should be maximized or minimized is basically
nonexistent. For a spectral width that is about 10% of the available bandwidth, amplitude
restriction produces 100 times less second harmonic intensity than TL pulses.77 Although
the suppression of a large range of frequencies is possible, the low intensity obtained at

the desired wavelength makes this control route experimentally unpractical. Conversely,

phase modulation does not change the overall energy of the pulse. Interference between

17

 

 

fields with different phases, which can be constructive or destructive, allows targeted
regions of the second harmonic spectrum to either be maximized or minimized, as was
shown in Figure 1.2. Finding the specific phase modulation of a fundamental pulse to

control the amplitude of the multiphoton spectrum is crucial for MII control experiments.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0- . windowA - i ‘ —SHGA
— ,’ ‘ window a i- i x —— SHG 13
0 8 l" ,l [m g. *- ,’ 1‘
_ I m _ o
a I c r ‘
S __ r \ .E _ r ‘
“ 0.4 - .' ‘. 0 - l \
E r s I I ‘
T I \ (D '- I ‘\
0.2 "" ' J ‘ ‘ " " \
00 vf’L 1 1) 1AA r 1 ‘1‘. 1 L’l A 1 I~_|_
' 720 760 800 840 880 360 380 400 420 440
Wavelength (nm) Wavelength (nm)

Figure. 1.5. Effect of spectral amplitude restriction on SHG. Left: Experimental spectrum
of the laser before (dotted lines) and after filtering with windows of width 20 nm
(window A) and 40 nm (window B). Right: Experimental SHG of TL (dotted lines) and
spectrally filtered pulses, A and B, as indicated in the right panel.

The question on which pulse shape among numerous alternatives is optimal to
produce a desired outcome is also a matter of concern. The number of different pulse
shapes that can be generated with a 128-pixel phase-only pulse shaper scales as the
number of phase values P that a pixel can introduce raised to the number N of pixels (PN).
If coarse phase steps are taken and the number of phases is reduced to 100, the search

0256. This is an infinitely large search space, compared to the

space would be of size 1
number of atoms in the universe (~1081 atoms), which, in all respects, is quite a huge
number. One of the challenges of pulse shaping technology has been to figure out how to

effectively explore such a vast search space. Although the efficiency of computer

18

learning algorithms, such as GAs, in exploring large spaces for the most favorable pulse
shape is highly regarded, the practicability of this approach in the laboratory setting needs
to be considered. Since the number of experiments that can be carried out with a
femtosecond laser is usually limited to less than 1000 per second, only a small fraction
(~3.6 million pulses) of the huge search space can be tested in an hour. Only in very
special cases or with great luck could one find the optimum pulse shape among the
hundreds of orders of magnitude of alternatives.

A recent study by Rabitz affirms that the continuous control of both the phase and
amplitude of all frequencies within the bandwidth of a pulse by a GA leads to several
excellent, but no intermediate solutions.78 This entails a markedly large search space that
cannot be sampled experimentally. Consequently, all experimental endeavors require
some strategy to reduce the search space by limiting the phases to a specific function (for
example, sinusoidal or polynomial). These approaches have been quite successful but
setting drastic limitations on the search space may not always prove to be valuable. Other
groups have reduced the number of pixels by binning contiguous pixels32’ 3‘7 although the
possibility that much better solutions may have been missed by such a reduction in the
search space is an issue that one needs to be aware of. Efficient search space reduction is
presently one of the most important goals in shaped pulse coherent control.

Initial MII experiments from the Dantus group involved the use of a simple
sinusoidal function because it offered the ability to create desired regions of constructive
and destructive intrapulse interference with a minimum number of control parameters‘55
(see Figure 1.2). The realization that the most important parameter involves controlling

the relative phase between the frequencies that combine to achieve the excitation

19

consequently led to the generation of binary phase (BP) functions with 0 or 1: phase
retardation77 (see Equation 1.2). The implication of this realization in terms of search
space reduction allows one to use many more active pixels. In addition, the redundancy
of phases that are allowed to take on arbitrary values is eliminated. Phase modulation
with BP fimctions is exploited in this research to initiate selective responses from
different molecules as a means for identification. Chapters 3 to 5 involve experimental
applications of phase modulation. Chapter 6 involves a variation of phase modulation,
whereby binary functions are used to control the degree and direction of polarization of a
laser pulse. Polarization shaping with femtosecond laser pulses, which is explored as a
prospective scheme for chiral recognition, will be discussed in greater detail in the final

chapter of this work.

20

 

CHAPTER 2

EXPERIMENTAL TOOLS

The experiments covered by this dissertation were carried out on two different
laser systems. The first system was a Titanium:sapphire (Ti:S) oscillator system
(Kapteyn-Mumane Labs) pumped by a neodymium: yttrium lithium fluoride or Nd:YLF
laser (Spectra Physics: Millenia), capable of producing 10 fs pulses tunable from 780 nm
to 840 nm at 87 MHz repetition rate. The pulse energy from this oscillator is ideally
suited for liquid phase experiments and for biological imaging. This system was the
primary laser source for the experiments in Chapters 3 and 4.

The second was a regeneratively amplified Ti:S laser system. The output from a
continuous wave (cw) neodymium: yttrium vanadate or Nd:YVO4 laser (Spectra Physics:
Millenia) pumps a femtosecond Ti:S oscillator (Kapteyn-Murnane Labs), capable of
producing ~15 fs pulses at 80 MHz. This output is amplified by a regenerative amplifier
(Spectra Physics: Spitfire) pumped by a 1 KHz Nd:YLF laser (Spectra Physics: Evolution
X). The amplified pulses centered at 800 nm have a maximum energy of 0.8 mJ/pulse
and time duration ranging from 35-50 fs. The high photon flux in this system was
essential for the ionization experiments carried out in Chapters 5 and 6.

Each of these laser systems was interfaced with a 128-pixel double mask pulse
shaper. In each of the following chapters, the laser system and pulse-shaping apparatus

used will be re-introduced in the experimental section.

21

2.1 Femtosecond Laser

2.1.1 Ti:S Laser Oscillator System

The generation of ultrashort pulses implies the amplification of a broad frequency
spectrum. A new era in the solid state laser field was born with the characterization of a
laser material based on the titanium doped sapphire crystal, TizAIZO3.79 Shortly after the
characterization of the optical parameters, the successfirl cw lasing action in this material
was demonstrated.80 The broad gain bandwidth, large energy storage density, and
excellent thermal properties of the Ti:S crystal make it an excellent material for
femtosecond laser and amplification systems. Ti:S has an energy level diagram (see
Figure 2.1), which is typical for solid laser materials (four-level lasers). The bandwidth
over which this laser can be obtained is wide enough to allow for 3-fs long pulses. The
absorption band has its peak at around 500 nm and the emission spectrum is peaked
around 800 nm. Since the absorption and emission bands are well separated, losses due to
reabsorption of the laser radiation are minimized.

Two conditions determine the frequency components available to build up a wide
frequency spectrum. The first is the approximately resonance condition for a standing
wave: nM2=L, where L is the length of the cavity. The second is the spectral bandwidth
of the laser, which is mainly limited by the gain profile of the amplifying medium.

In a cavity consisting only of a gain medium, end mirror and output coupling
mirror, all modes are oscillating in random phase. By introducing additional components
in the cavity, it is possible to phase-lock these modes to each other. The generation of

ultrashort laser pulses implies that all the frequency modes should have a well-defined

22

phase relation with respect to each other. In this sense, the concept and implementation of

mode-locking techniques in laser sources has been crucial.8|

 

Relaxation :

     
   

Laser Transition
800 nm

Pump
500 nm

Relaxation

 

 

 

Figure 2.1. F our-level scheme of a typical solid-state laser. The photons are first excited
to a state higher in energy than the upper laser state. They then quickly decay down by
relaxation or radiationless decay into the upper laser state. It is important for the pumped
state to have a short lifetime for spontaneous emission compared to the upper laser state.
The upper laser state should have as long a lifetime (for spontaneous emission) as
possible, so that the photons live long enough to be stimulated and thus contribute to the
gain. The fourth step involves depopulation of the lower laser level by a fast decay
(relaxation) process. This greatly decreases the loss of laser photons by stimulated
absorption processes since the photons in the lower laser level have a short lifetime for
spontaneous emission.

When the phase difference between the modes is constant, the modes are
oscillating in phase and interfere constructively. If we add the electric fields from the
standing modes, the sum as a function of time will be a short pulse propagating back and
forth in the cavity.

At low intensities, the refractive index for a given wavelength is constant. At high
intensities (I), the refractive index varies by n = n1+n21. Due to the fact that the laser has a
spatial intensity variation, the refractive index will be larger in the center of the beam,

where the intensity is the highest. This makes the crystal behave like a lens (referred to as

a Kerr lens). A major breakthrough in the femtosecond laser field was established by

23

Sibbett et al. in 1991, with the discovery of the action of self-mode-locking.82 Also
referred to as Kerr-lens mode-locking (KLM), the method relies on the non-linear
refractive index properties (n;) of the Ti:S crystal. The self-focusing action in the crystal
in combination with an intracavity aperture results in intensity dependent losses and
causes the mode-locking of the laser. This is shown in Figure 2.2. It is common that Kerr

lensing in the amplifying medium is used to obtain both gain and mode-locking.

 

 

 

 

 

 

 

 

 

n = n1+n2|
A V
x l—_
laser
beam .
( Ti:S aperture
"' crystal

 

 

 

Figure 2.2. The spatial profile of the laser beam creates the “Kerr lens”.

Implementation of the KLM scheme and accurate control over GVD allow the
generation of short femtosecond pulses. These features are characteristic of the Ti:S laser
oscillator system used in the first half of this thesis, shown in Figure 2.3.

Pulse durations in prism-controlled Ti:S oscillators are limited to 10 fs, as a result
of the inherent presence of uncompensated higher order dispersion terms used. With the
introduction of KLM lasers, the design of an oscillator that generates femtosecond optical
pulses has reached a stage of remarkable simplicity and robustness. The ability of the

83, 84

Ti:S solid-state oscillator to generate 10 fs pulses as well as its excellent reliability,

24

makes it an attractive ultrafast light source in the laboratory. Not surprisingly, the class of

lasers based on Ti:S has rapidly gained ground in the field of ultrafast spectroscopy.

 

   
   

pump input

curved ‘
mirror 1 )4
a”
’ (ll lens @
l v
' Ti:S crystal

output coupler

 

   
 

  

 
  

reflector
prism 1

 

 

 

Figure 2.3. Ti:S laser oscillator system.

2.1.2 Regeneratively Amplified Ti:S Laser System

High energy laser systems are generally classified by their method of
amplification as regenerative or multi-pass. Regenerative amplification, which describes
the system in our laboratory, provides a more efficient use of the gain and better stability,
but results in greater phase distortions in the output pulse. Multi-pass amplification
avoids some of these phase distortions by operating with fewer optical elements through
which the laser beam transmits several times. Although shorter pulses can usually be
achieved by multi-pass amplification,85 difficulties with alignment and lower pulse to
pulse stability make regenerative amplification systems more practical.

The KHz laser system in our laboratory comprises four parts: an oscillator that

produces femtosecond pulses in the near IR, a stretcher that lengthens the pulses to a few

25

hundred ps, an amplifier stage that increases the pulse energy and a compressor that
reduces the pulses to about 50 fs.

The oscillator of the KHz laser system delivers laser light pulses of 15 fs with
energy of 2 nJ. Since this power is not sufficient for the applications described in
Chapters 5 and 6, the laser pulses are amplified using the Chirped Pulse Amplification
(CPA) technique.” 87 In CPA (Figure 2.4), an ultrashort laser pulse is stretched out in
time prior to introducing it to the gain medium using a pair of gratings that are arranged
so that the low-frequency component of the laser pulse travels a shorter path than the
high-frequency component does. After going through the grating pair, the laser pulse
becomes positively chirped, that is, the high-frequency component lags behind the low-
frequency component, and has longer pulse duration than the original by a factor of 103 to
105. The resulting stretched pulse is then safely introduced to the gain medium and
amplified by a factor of approximately 106. Finally, the amplified laser pulse is
recompressed back to the original pulse width through the reversal process of stretching
(compression), achieving orders of magnitude higher peak power than laser systems
could generate before the invention of CPA.

Most of the spectral phase distortions accumulated by the laser pulse inside the
regenerative amplification cavity can be compensated by a pulse shaper. For this reason,
a pulse shaper was introduced between the oscillator and regenerative amplifier to
compensate these distortions and offset spectral narrowing in the amplifier. The result is a

very stable source with a pulse duration of ~30 fs (33 nm bandwidth).

26

 

 

compressor

 

 

 

 

 

 

 

 

\ - regenerative ;
\\ amplifier

 

 

 

 

 

 

 

 

Figure 2.4. Chirped pulse amplification

2.2 Pulse Shaper

Many approaches in femtosecond pulse shaping are dedicated to the synthesis of
laser waveforms by spectral filtering of a spatially dispersed frequency spectrum. The
earliest available pulse modulation scheme still in use is frequency chirping, where
frequency components in a laser pulse are arranged with a certain phase ordering. This
can be easily achieved by dispersing the ultrafast pulses through a pair of gratings or
through propagation in an optical fiber. As a consequence, an otherwise bandwidth-
limited ultrafast pulse is stretched in time. Most practical applications of shaped pulses
require a pulse shaping device that can be used to construct desired pulses with
appropriate amplitude and phase modulations. The idea of pulse shaping entails the
modulation of the incident electric field of the laser by a mask in the frequency domain.

This results in an outgoing shaped spectral electric field.

27

 

3i)

Sp

I (II.
C0111
mm

51mm

Femtosecond laser pulses can not be directly shaped in the time domain because
of their short duration. There are currently no electronic devices which work with such
speeds. For this reason, the indirect technique for shaping ultrafast laser pulses involves
splitting the pulse into its frequency components, modulating these in the required
manner and recombining them to obtain the shaped pulse. This requires the use of
modulators like LQ modulators. The method we used in this work — pulse shaping with
an SLM at the Fourier plane - is described below. A pulse shaper is implemented to

manipulate the electric field that interacts with the molecules.

2.2.1 SLM

The femtosecond pulse shaper using a LQ-SLM was pioneered by Heritage and
Weiner.7’ 88 The LQ-SLM used in this research is a commercially available device
(Cambridge Research Institute, Inc.) — a double mask SLM unit”’ 90 which consists of
two linear LQ arrays of 128 pixels each, separated by 2.2 mm, with independent control
of each pixel. Figure 2.5 (A and B) illustrates the front face and side view of the SLM.
The two LQ arrays are perpendicular to each other and are oriented 45° to the incoming
light. They are placed between two silica substrates coated with indium tin oxide, an
optically transparent material which is electrically conductive. One substrate is the
ground plate and the other is patterned with a linear array of 128 electrodes (width of
100nm each) that define the pixels. The 3 gm spacing between the electrodes cannot be
controlled. Each pixel in the first array is superimposed directly in front of the
corresponding pixel in the second array. These arrays allow the independent or

simultaneous control of phase and amplitude. The LQ components in each array are

28

birefringent. Depending on the polarization of the incoming light, a voltage can initiate
pure phase retardation (delay) or a combination of phase retardation and polarization

rotation.

A B - r n I '1
I 1- AR-coated cover
.. _ — 2- polarizer
° 3 3 3 ° 3- substrate buffed at :450
4- quuid crystal
4 4 3

 

g
Q

 

5- indium tin oxide patterned substrate
6- index-matching epoxy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

out

I’ “ C L01 L02

 

 

 

l:(>

 

   

 

’ 12.8 mm ‘
.f calibration pixel
. Milwlllllllllllll . Irtied to 127)
0 , , 127
/' K
I I] polarizer polarizer
2 mm

 

 

 

311m
l‘-'| ? g
100 um

Figure 2.5. The LQ-SLM. A. Front view with pixel dimensions. B. Side view showing
components of the SLM. C. Direction of polarization as laser pulse goes through each LQ
mask.

 

 

 

 

Figure 2.5 C depicts how this SLM unit operates. Light incident on the SLM first
goes through a polarizer and transmits to the first LQ array where its phase undergoes
retardation and polarization is rotated in one direction. As the light hits the second LQ
array, it undergoes more retardation in phase and is polarized in the Opposite direction.

On average, LQ-SLM systems provide retardance equal to 41:. The second polarizer is

29

utilized for amplitude modulation. Polarization rotated away from the incident, horizontal
direction causes the attenuation of light by this second polarizer. Rotation by 90° results
in zero transmission. This dependence on polarization is employed in the calibration of
the SLM. I

Control over the retardance introduced by each pixel is established from the
voltages introduced in the SLM. The basic factor is determining the overall differences
between pixels and the changes in retardance. Calibration of the SLM is accomplished by
obtaining a frequency-dependent calibration curve of the laser spectrum. Increasing the
voltage, V, of one of the LQ masks while keeping the other mask at a constant voltage
generates a transmission function, T, which can be used to calibrate the dependence of

retardance on voltage.

T = c032[”(RA(VA)/I‘ RB(VB))] 2.1

 

RM refers to the retardance as a function of voltage for each LQ array, A and B. The
total phase retardation, ¢, obtained from changes in T, can then be calculated by

¢ = ”IRAQ/10+ RB(V3))/1 2.2
Needless to say, the accuracy by which phase can be controlled by the pulse shaper is

contingent on a well-calibrated SLM.

2.2.2 Pulse Shaping Apparatus

The operation principle of a pulse shaper is based on optical Fourier
transformations from the time domain into the frequency domain, and vice versa. This
section focuses on the two pulse shaping setups used in this work. All the optical

elements in the pulse shaping apparatus are symmetrically separated by the focal length

30

 

In",
‘.-
.I' 17]

of the lenses, f, so that the pulse leaving the setup has the same characteristics as the
pulse entering the setup (4]). The collimating lenses and grating pair are set up in a 4f
configuration”, with the SLM being placed at the center. The whole pulse shaping setup,
without the SLM, introduces no dispersion if the 4f condition is met, and is called a zero-
dispersion compressor. This dispersion-free condition takes into account that the lenses
are thin and free of chromatic aberrations. A grating spreads the pulse, so that different

spectral components map onto different spatial positions.

2.2.2.1 The Linear Setup

The linear 4f-setup (Figure 2.6) uses two gratings and two lenses. The first grating
spectrally disperses the femtosecond pulse. The frequencies are collimated by the first
lens, which Fourier transforms the pulse from the time domain into the spectral domain.
In the Fourier plane of the lens, the spectral components are spatially separated along one
dimension. They are focused by the second lens onto the second grating. An inverse
Fourier transform into the time domain takes place after passing the second lens.
Cylindrical lenses are required in order to avoid damaging the polarizers in the SLM. If
the lenses are not separated by 2f, the beam is divergent and a temporal chirp leaves the
setup. If the gratings are not situated exactly in the focal plane of the lenses, an
additional spectral chirp, where the frequencies are spatially separated within the laser
beam profile, is generated. This distorts the temporal phase of the pulse and cannot be
compensated.

This setup was used for the experiments on selective ionization and

photofragmentation described in Chapters 5 and 6, with the amplified femtosecond laser

31

system described in Section 2.1.2, which provides laser pulses with a spectral bandwidth
of up to A)» = 27 nm at FWHM. The intensity of the laser beam after passing through this

configuration decreases by about 40%.

 

Fourier Plane

 

 

 

 

Figure 2.6. Schematic design of a linear 4f pulse shaper using diffraction gratings and an
SLM

2.2.2.2 The Folded Setup

For pulses with larger bandwidths, provided by the femtosecond laser oscillator
system described in Section 2.1.1, the folded pulse shaping system was used. The shorter
pulses make gratings too dispersive, so a setup that is less dispersive is more
advantageous. In this setup, prisms, which are about an order of magnitude less
dispersive than gratings, were used to disperse the frequency components of the pulse.
Instead of cylindrical lenses, we used cylindrical plane concave mirrors where their inner
concave sides were covered with a dielectric layer, leading to a reflectivity of 99%. The
overall intensity losses of the folded setup are about 25%. The off-axis aberration caused

by the diffraction of the spectrum by the prism onto the cylindrical mirrors is minimized

32

by two folding mirrors. Figure 2.7 shows the folded pulse shaping setup, which was used

for the experiments on selective two-photon excitation described in Chapters 3 and 4.

 

input output

i\i Ill ligfj

 

 

 

W““““\V
cylindrical SLM cylindrical
mirror ’ folding folding mirror
mirror mirror

 

 

 

Figure 2.7. Schematic design of a folded pulse shaper using prisms in lieu of gratings.

2.2.3 Phase Modulation with the LQ—SLM

The programmable LQ-SLM used in this work is capable of both phase and
amplitude modulation. The incident light that enters the pulse shaper is horizontally
polarized. This ensures parallel polarization with the extraordinary axis of the LO (x-
axis) at the input face of the SLM. At the output face, a polarizer selects only the
horizontally polarized light. As the input light field passes through the first LQ array, its

polarization is rotated in one direction according to the phase retardance 136'. As it passes

through the second array, a voltage-dependent phase retardance A¢B occurs and
polarization is rotated in the other direction. The polarization can also be rotated in such

a way that the attenuation is compensated while the phase shifts. The average phase

33

retardance is given by exp[i(Aqf'+ A¢BV2I The amplitude is altered according to the

cosine of the differential retardance Agf' — Ad)”.

The output electric field 133”!" “I of the nth pixel after leaving the SLM is given
by,
Egu’P'“ = 5;,"Pu’ -cos((A¢A — 13¢” )/2) - exp(i(A¢A + A¢B )/2) = E},"P“’R,,, 2.3
where cos((A¢‘4- A¢B)/2) is the amplitude modulation term, and exp(i(A¢A+ A¢B)/2) is

the phase modulation term. Rn=exp(iw(a))); JTn—(ai) is the filter function applied to the

nth pixel; Tn(a)) is the transmission and 111,,(a2) is the phase filter of each element. For
w(a))=0 the spectral filter function allows amplitude-only modulation, while for Tn(w)=1
the filter function can modulate only the phase mask.

Pure phase modulation is achieved when the cosine in the expression is equal to l,
a condition that is achieved by keeping AW = A¢B. Amplitude modulation is also
accompanied by phase modulation. As the SLM possesses two LQ arrays, the second
one compensates for the unwanted phase modulation so that pure intensity attenuation of
the optical waveform is obtained. Neither of the two masks alone is able to modulate
phase or amplitude, but their combination allows independent phase and amplitude

modulation.

2.2.4 Limitations of the SLM
The number of pixels in LQ modulators limit. the resolution of the modulation.
The finite width covers a range of wavelengths, reducing the fidelity of shaping. The

optical setup, particularly the focal spot size that is projected across the lens towards the

34

SLM, if not given careful attention, can further reduce spectral resolution. The dead
spaces or gaps between electrodes also add artifacts to the pulse train. Although typical
LQ-SLM systems provide 4n retardance, in theory, only 2n retardance is required since
the phase can be folded (distributed into separate 0-2 1: segments) to produce any type of
total retardance. Folding the phase can, however, introduce additional unwanted
modulation into the pulse because of the abrupt phase changes required for folding.91 In
addition, the limited resolution of the shaper prevents clean phase jumps from one pixel
to the next. Nonetheless, the excellent agreement between experimental and theoretical
predictions obtained for most nonlinear optical experiments implies that these

shortcomings are quite trivial.

35

CHAPTER 3

SELECTIVE TWO-PHOTON EXCITATION

Multiphoton excitation draws on an old idea in quantum theory and, in the course
of its infusion in microscopy, is currently being applied to a number of modern scientific
problems. This chapter dwells on the possibility of utilizing MII to attain selective
excitation of fluorophores without the need for wavelength tuning, optical setup
modifications, and selective visualization with filters. The basic idea is to exploit the
high peak intensity of the femtosecond laser to attain multiphoton excitation, and to make
use of its wide spectral bandwidth to access a wide range of frequencies that could be
manipulated through intrapulse interference to bring about selective two-photon
excitation. Fluorescence selectivity through multiphoton excitation with shaped
femtosecond laser pulses could have large implications in terms of the ability to
recognize and monitor the environment and dynamics around a fluorophore in a complex
biological system.

The most commonly used type of multiphoton excitation is two-photon excitation
(described by Gbppert-Mayer in 1931), where simultaneous absorption of two low-
energy photons leads to a single quantum event. The powerful advantages of two-photon
excitation microscopy92 arise from the basic physical principle that absorption depends
on the square of the excitation intensity. In practice, two-photon excitation is generated
by focusing a single pulsed laser through a microscope. As the laser beam is focused, the
photons become more crowded, but the only place at which they are crowded enough to

generate an appreciable amount of two-photon excitation is at the focus. Photons are not

36

only crowded spatially (by focusing) but also temporally (by using mode-locked laser
pulses). This gives the needed excitation intensities for multiphoton absorption, but limits
the average power at the focus to less than 10 mW. The limited excitation region reduces
phototoxicity because photodamage is largely confined to the focal volume.

In two-photon fluorescence microscopy, the excitation wavelength is in the near
IR, a region of the spectrum in which there is virtually no absorption in cells or most
chemical systems. The most widely used light source for multiphoton microscopy is the
Ti:S laser which normally has a pulse width of ~1 00 fs and a repetition rate of ~ 80 MHz,
allowing the high photon density and flux required for two photon absorption and is
tunable across a wide range of wavelengths. The use of longer wavelength excitation
light slows the rate of photobleaching of the sample, allowing the use of higher
magnification objectives. In addition, longer wavelength light can penetrate deeper into a
specimen. In fluorescence microscopy, excitation light incident on the specimen can be
scattered (Rayleigh scattering) to various degrees before it reaches the focal plane, and
the resulting fluorescence can also undergo scattering as it returns through the specimen
toward the detector. These scattering effects combine to reduce the collected fluorescence
signal. The simplest approximation of Rayleigh scattering shows that the amount of
scattered light is inversely proportional to the fourth power of the light's wavelength.
Thus 400 nm light (one-photon) light would be expected to undergo several times more
scattering than 800-nanometer (two-photon) light.

Successes associated with two-photon microscopy have also created interest in
exploring applications based on three-photon excitations. For a three-photon process, a

longer excitation wavelength such as those common in optical communications can be

37

used. The cubic dependence of the three-photon process on the incident light intensity
provides a stronger spatial confinement, so that a higher contrast in imaging can be
obtained.

The notion that phase modulation can be used to attain selectivity in two-photon
excitation is motivated by the need to keep the laser characteristics constant throughout
the excitation process. Tweaking the laser or moving optical elements to attain the
desired excitation wavelength for a particular fluorophore can be averted by controlling
the phase of a laser pulse. The following Sections show how the use of a pulse shaper for
two- or three-photon microscopy provides the flexibility of selective probe excitation or
maximum signal enhancement by controlled modulation of the spectral phase of

femtosecond pulses.

38

3.1 Probing a Molecule’s Microenvironment'

3.1.1 Introduction

This work focuses on the design of tailored femtosecond pulses to achieve control
of nonlinear optical excitation in large molecules based on the concept of MII.“' 66 The
goal is to elucidate well-defined and reproducible pulse shapes that can be used to
enhance or suppress particular nonlinear optical transitions in large molecules such as
laser dyes and proteins in solution. The use of MII to probe the local and microscopic
environment of molecules by selective two-photon laser induced fluorescence (LIF) is
discussed in this section.

There has been much interest in the use of tailored ultrafast laser pulses to control
the excitation of large molecules in condensed phase.“ 5 l’ 93’ 94 Coherent control of large
molecules in condensed phase has been considered very challenging because of the large
numbers of intra and intermolecular degrees of freedom involved.95 Further
complications arise from the extremely fast homogeneous (< 100 fs) and inhomogeneous
(< 10 fs) electronic dephasing in these systems.96 Many of these studies have been based
on the concept of pulse shaping, whereby a learning algorithm optimizes the spectral
phase of the laser field.ll These methods have the advantage of convergence to an
optimal pulse shape. However, because of the sensitivity of high-intensity multiphoton
excitation to a large number of experimental parameters (for example pulse intensity
fluctuations, saturation, transverse-mode quality, pulse bandwidth, and spectral phase
which is affected by transmission through lenses and windows and reflection from

dielectric mirrors), it is difficult to ascertain principles from the results of these

 

' Reproduced in part with permission fi‘om J.M. Dela Cruz, I. Pastirk, V.V. Lozovoy, K.A. Walowicz, M.
Dantus, Multiphoton Intrapulse Interference 3: Probing Microscopic Chemical Environments J. Phys.
Chem. A 2004, 108, 53-58. Copyright 2004 American Chemical Society.

39

I.

experiments. Furthermore, it has never been shown that results from these types of
experiments can be reproduced.

The influence of phase on two-photon excitation was explored by Broers et al., who
proposed that certain phase firnctions could focus the second harmonic spectrum in the
frequency domain like a Fresnel lens focuses light in the spatial domain. They
demonstrated this principle by controlling the two-photon excitation of Rb atoms.lo
Meshulach et al. used the same principle to demonstrate coherent control on the two-
photon excitation of Cs atoms.59 Their work focused on the narrow two-photon
absorption resonance; they showed that the coherent two-photon enhancement was absent
in the broadband excitation of large molecules in solution.59 Work from the Dantus group
demonstrated that the changes introduced by pulse shaping on two- and three-photon
transitions arise from destructive interference, not on focusing in the frequency domain.66
The theoretical framework that was conceived as a result of this observation led to the
development of MII for controlling multiphoton excitation in large molecules in
solution.“’ 66 The fimdamental advantage of this method was the early realization that
very short pulses, with concomitant large bandwidths, would be needed to control the
nonlinear excitation of complex systems. Unlike in the gas phase, the existence of
inhomogeneous broadening in the condensed phase prevents the coherent manipulation of
vibronic levels. MII could control the delivery of energy to the molecule with little or no
dependence on intramolecular dynamics.

The first application of phase tailored pulses for sensing pH involved intense linearly
chirped laser pulses that saturated the one-photon resonant transition of a pH-sensitive

dye.61 The ratio between the LIF signals obtained for positively and negatively chirped

4O

pulses was found to vary from 1.9 to 1.2 for different pH values. However, this ratio did
not vary monotonically, being the same for pH 7 and pH 9.

The method presented here is completely different from the early experiments of
Buist et al.‘51 This method involves very weak pulses inducing two-photon transitions.
The samples are thus transparent at the carrier wavelength of the pulse, the transitions are
not saturated, and the results are independent of probe molecule concentration. The
mechanism we used for pH probing is understood and theoretical simulations are in close
agreement with the data.

The probability for two-photon excitation, 3(2), using phase-modulated pulses at
frequency a) =A+ (on, where can is the carrier frequency, can be calculated from the

expression66

2
5(2) cc [gm (2a))l [5(a) + Q)E((o — main dd), 3.1

where E(a)) = |E(a))|exp(i¢((o) is the electric field of the laser. In this expression, we
assume a broad nonlinear two-photon absorption spectrum, g(2)(2a)) and ignore
intramolecular dynamics to simulate our experimental results. Consequences of this
assumption are addressed later.

Here we demonstrate the M11 method, which couples the sensitivity of
multiphoton excitation on the spectral phase of the laser pulses to probe microscopic
chemical environment-induced changes in the multiphoton excitation spectrum of
sensitive reporter molecules. We carry out the optimization of the required phase

firnctions in solution and provide theoretical simulations.

41

3.1.2 Experimental Section

Experiments were performed using a Ti:S laser oscillator (K&M Labs) capable of
producing pulses with bandwidths up to 80 nm FWHM. The output of the oscillator was
evaluated to make sure there were no changes in the spectrum across the beam and was
compressed with a double prism pair arrangement shown in Figure 3.1. The pulse shaper
uses prisms as the dispersive elements, two cylindrical concave mirrors, and an SLM
(CR1 Inc. SLM-256), composed of two 128-pixel LQ masks in series. The SLM was

placed at the Fourier plane.90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SHG-FROG ..
:Compressor : g
A i 1 113,3
Spec2000 i ; a
el ;. : : l1 E
O.
1 ' O.
l ' CU
---------_---:::::::::,:::::-- (D.
, I:
x/D \' E
‘ ‘ \ 12 £-

: PulseShaper :

 

Figure 3.1. Schematic of the laser setup for MII experiments. The output of the Ti:S
oscillator is compressed by the prism pair compressor before arriving at the pulse shaper.
Phase and amplitude shaping take place at the SLM which incorporates a polarization
optic at its input and output windows. Once shaped, the pulses are directed towards the
experimental cell. Characterization of the pulses by MIIPS requires replacing the sample
cell by a thin SHG crystal; alternatively, the pulses are directed to an SHG-FROG.

Because the retardance introduced by the LQ is frequency dependent, pulse
shaping of ultrashort pulses (with > 100 nm total bandwidth) requires calibration of each

pixel element. At the end of the calibration, our data acquisition program maps the

42

spectrum of our laser on the SLM in a frequency rather than wavelength scale. An
algorithm has been programmed to independently control amplitude and phase at each
frequency with the SLM. This allows us to define a Gaussian shape or spectrum in the
frequency domain using an amplitude mask, and then to scan the phase function without
affecting the spectrum of the pulse. After compression and pulse shaping, ~ 200 pJ pulses
were used to interrogate the samples.

The reproducibility of pulse shaping experiments depends on accurate
characterization of the phase of the laser pulses. For this task we used the MIIPS
method97 to measure both the sign and magnitude of phase distortions in the pulse. This
method was described in Chapter 1.2. For the experiments in this study, we started with
very broad bandwidth pulses (~80 nm FWHM). We then defined Gaussian spectra of the
desired width and center wavelength by amplitude shaping. Second, we measured the
quadratic and cubic phase distortions in the pulse by seaming a phase function defined
by Equation 1.1, where the amplitude at was set to 1.57rrad, and 7 was set to the
approximate pulse duration (20 fs).

After MIIPS, the final pulse shaping step involved the controlled phase
modulation of TL pulses. This step involved introducing phase functions determined by
Equation 1.1. The laser was focused by a 50 mm focal length lens onto a 10 mm path
length quartz cell containing the dye solution. Fluorescence perpendicular to the
excitation beam was collected near the front face of the cell to minimize dispersion in the
sample. We took advantage of the narrow acceptance angle of the optical fiber for spatial
filtering and imaged the signal onto an Ocean Optics spectrometer (82000). The

experiment involved scanning the parameter 6 and detection of the fluorescence of the

43

sample (spectrally integrated). Signal integration time was typically 500 msec per

spectrum for each value of 6, and the data scan, with 256 different phase positions was

averaged over 3 cycles of 6 from 0 to 471'rad.

The samples used were 10-5 M aqueous solutions of 8-hydroxypyrene-1,3,6-
trisulfonic acid trisodium salt (HPTS or 3sPyOH) (Sigma-Aldrich) in different pH buffer
solutions. The pH 6 buffer was prepared from 0.1 M KH2P04 (250 mL) in 0.2 M NaOH
(14 mL); pH 8 buffer was prepared from the same solutions, but with more (115 mL) of
the NaOH; while the pH 10 buffer solution was prepared from 0.025 M NazB407
(250 mL) in 0.2 M NaOH (27 mL). The buffer solutions were diluted to 500 mL with
distilled water. UV-visible absorption spectra of the dye solutions at different pH

environments were obtained with a Hitachi U-4001 UV-vis spectrophotometer.

3.1.3 Results

Figure 3.2 illustrates the dependence of two-photon LIF on phase modulation.
These measurements were obtained using 23 fs TL pulses that were shaped to have a
Gaussian spectrum centered at 842.6 nm with a bandwidth of 48 nm FWHM and spectral
phase defined by Equation 1.1. Figure 3.2A shows the experimental two-photon LIF
intensity from HPTS in a pH 10 solution (dots) as a function of 6. The signal shows a
sinusoidal modulation. Figure 3.28 shows the experimental two-photon LIF intensity
from HPTS in a pH 6 solution (dots) as a function of 6. The signal is modulated,
however, there is a clear phase shift compared to pH 10. The shift comes about from pH-
induced changes in the two-photon absorption spectrum. Experiments obtained for pH 8

were similar to those obtained at pH 10 but they contained a small shift that made them

44

 

 

 

distinguishable. It is crucial to underline that the laser spectrum and intensity remain

unchanged throughout the experiments and that the emission spectrum from the dye is

independent of pH.

 

 

 

 

 

 

 

 

 

«1
a)
O 5
C E "
(D E
8 * BE ‘
9 - o 1 \ e. '. pH-G d
O 5 3 I .' . .0
2 08 '- ' I: .0 e '-
U— - V. l'. : '. '. -
0.6 - I '1 ‘. 1. -
- : 1, / -
. : -
0'4 ’- Y . Y 4 4 r _
0 1 1: 2.011: 3.0x

Figure 3.2. Experimental MII results and simulations for HPTS in basic and acidic
solutions. Intensity of the two-photon induced fluorescence of HPTS solution at pH 10
(A) and pH 6 (B) are plotted as a function of the phase parameter 6. Simulations
(continuous lines) based on Equation 3.1 are in good agreement with the experimental
data (dots). The arrows at 6 = 1/41r (dotted) and 6= 3/41r (dashed) represent phase
functions for which the intensity of two-photon fluorescence from HPTS at pH 10 is
similar but shows large differences for pH 6. The optimum two-photon excitation of
HPTS is indicated by vertical lines for the acid and basic solutions.

Simulations of the experimental signal were performed using Equation 3.1
without adjustable parameters. The spectrum of the pulse and the absorption spectrum of

HPTS were measured experimentally. An examination of the molecular structure of

HPTS shows that it has no center of symmetry. Since parity restrictions may be relaxed

in this case, the similarity between one-photon and two-photon absorption spectra is

45

expected. The spectral phase ¢J(.Q) imprinted by the phase mask was the same used for
the simulations. Both experimental and theoretical data were normalized such that the
signal intensity is unity and the background observed is zero. Generally, the experimental
data (dots) agree quite well with the calculated response (continuous line) of the dyes in
all pH environments (see Figure 3.2). Differences found for pH 6 are discussed below.
We repeated our experiments with 50 fs pulses centered at 760 nm. No
differences in phase dependence were observed for the three different pH solutions, as
expected and as corroborated by simulation. The absorption spectrum of HPTS in the
three different pH environments (pH 6, 8 and 10) is relatively flat near 380 nm. These
control measurements indicate that the phase dependence arises primarily from

differences in the two-photon excitation spectrum and not from intramolecular dynamics.

3.1.4 Discussion

The MII method is used in this experiment to control two-photon excitation. More
specifically, the interference process, which is controlled by the phase introduced in the
pulse, suppresses two-photon transitions at specific frequencies within the two-photon
bandwidth.66 In other words, the interference can be used to tune multiphoton transition
probability at specific frequencies. This effect can be envisioned from scanning SHG
intensity using MIIPS. As 6 is scanned the maximum intensity in the second harmonic
signal scans as well from longer to shorter wavelengths.“’ 66 By considering this effect
together with the two-photon absorption spectrum, the expected signal can be calculated

with very good agreement.

46

I

pH > 7.5

O' '03 S

O 2th, 0.4 ps
——>
0‘ 515 nm,5.3 ns

:80.

“033 303- ‘03s
pH<75
'03 OH '03s *
0 2th, 0.3 ps OCH
0‘ 00
O ~450 nm 030
‘—
‘033 $03: 'ogs
+ WI - H”18:C:3 ps

'O3S i O' 'O3S
515 nm 4. 8 ns

“033 $03: -033

30:90

Figure 3. 3. Photochemistry of HTPS in basic (pH > 7. 5) and in acid solutions (pH < 7. 5).
Upon two—photon excitation, the equilibrium species in basic solution (3sPyO') reaches
an excited state that rs solvated within 0. 4 ps. Fluorescence from 3sPyO has a lifetime
of 5. 3 ns and rs centered at 515 nm. The equilibrium species in acidic solution (3sPyOH),
upon two-photon excitation, reaches an excited state that is solvated within 0. 3 ps.
Fluorescence from 3sPyOH is near 450 nm, however, deprotonation to yield (3sPyO )
takes place within 87 ps and 15 followed by fluorescence centered at 515 nm with a 4. 8 ns
lifetime.

HPTS is a highly water-soluble, pH-sensitive dye with a pKal of ~7.5 in aqueous
solution.98 The photochemistry of this system is shown in Figure 3.3. When in alkaline

medium, pH > 7.5, acid-base equilibrium is totally displaced toward the anion form

47

(3sPyO') of the dye. The electronic character of 3sPyO' remains unchanged after photo-
excitation, and corresponds to a singlet excited state.99 Fluorescence fi'om this state
undergoes a fast 0.4 ps Stokes shift and has a maximum at 515 nm and a lifetime of 5.3 i
0.1 ns.100

In acidic medium, pH < 7.5, the conjugate acid form (3sPyOH) dominates in the
ground state. Two-photon excitation yields the excited state (3sPyOH*). Fluorescence
from this state undergoes a fast 0.3 ps Stokes shift and has a maximum near 450 nm.
HPTS has a very high fluorescence quantum yield (almost 100% in both alkaline and
acidic solution, when excited with light of wavelengths shorter than 400 nm).101 Proton

99’ '00 to yield the conjugate base

transfer in the excited state takes place in 87.5 ps
(3sPyO'I) (see Figure 3.3), with a fluorescence maximum at 515 nm and a lifetime of 4.8
i 0.5 ns. Therefore, in both acidic and basic environments, the emission spectrum is
unchanged. The nature of the excited state dynamics involves a three-step mechanism of
the proton transfer reaction, whereby an intermediate state (unionized acid form) with
significant charge transfer character converts to the fluorescent excited anionic state.'00
There has been speculation about the manipulation of intramolecular dynamics of
large molecules in condensed phase. In Figure 3.4, we show the spectrum, phase and time
profiles for the shaped pulses with 6 = 1t/4 (dotted line) and Bit/4 (dashed line). Notice
that both of the pulses show identical time structure. The periodicity of the structure has
an average period of ~ 50 fs and lasts for about 200 fs. Given the similarities between the
two pulses, the control mechanism does not depend on temporal differences. The typical

electronic dephasing time of laser dyes in solution is shorter than ~100 fs, preventing

coherent manipulation by these shaped pulses. Based on these observations and the fact

48

that the simulations, which neglect intramolecular dynamics, are in excellent agreement
with the experimental data (especially pH 10), we conclude that intramolecular dynamics

plays a minimal role in the control that is observed.

 

 

 

 

 

 

 

 

 

7‘ l’>'\‘:"'-.. /‘ 2 "z ‘ 0'3
’..° / ' V It
‘7‘: \\ )T ..... ‘ '9. ,3 E
1‘ A ,1 i
<5, I l « 0.2
o — ’ I
' I I :3
If; i l
('1'; ‘l/u‘ ‘1' ‘1
I . o: ‘l ‘\
a 2": . . . . ‘~ .00
-200 -100 0 100 200
t, fs

Figure 3.4. Laser pulse characteristics in the time and spectral domain. The inset shows
the measured spectrum of the pulse 10‘.) and the phase ¢(7L) programmed by the SLM.
The pulses have a Gaussian shape with center wavelength 3.0 = 842.6 nm and 48 nm
FWHM. The temporal profile for two pulses (dotted and dashed lines) is calculated from
the experimental parameters for 6 = U47: and 6=3/47r, respectively. Notice that the two

pulses have the same temporal profile.

Some discrepancy between experiment and simulation is seen for the data
obtained under acidic conditions. The simulations were obtained using the measured
linear absorption spectrum of HPTS at pH 6 (see Figure 3.5). It is possible that the two-
photon spectrum is significantly different from the one-photon absorption spectrum.
Alternatively, fast dynamics may be causing the difference. However, the time scale for

proton transfer is much slower than the shaped pulses.

49

 

 

 

 

 

pH=6 H=10

1 - p -1
:- =2
mi “1
r5 2.
s . e
E- 8
8 5.“.
.O
<

o . . . . 3’5" . . - , . . o

350 400 x , nm 450 500

Figure 3.5. Absorption spectra of HPTS in acidic and basic solution (bold lines). The thin
line is the calculated spectrum of the square of the electric field (|E(2)(}c)|2) of the laser
pulse without phase modulation (TL). The square of the electric field of the laser pulse is
also calculated for shaped pulses with 6= 1/47r (dotted) and 3/47r (dashed), respectively.
The two spectra are very different because of MII. Notice that in the frequency domain, it
is clear that shaped pulses with 6 = 1/47r favor two-photon excitation of HPTS in pH 10,
while pulses with 6 = 3/47rfavor two-photon excitation of HPTS in pH 6.

In Figure 3.5, we show the absorption spectra of HPTS at pH 6 and pH 10. The
spectrum of the electric field E20) is proportional to IEmOQIZ, and is shown for a TL
pulse (thin line) and for phase modulated pulses with 6 = 71/4 (dotted line) and 370’4
(dashed line). Notice that two-photon transitions at 410 nm take place preferentially when
6 = 7114, favoring probe molecules in a pH 10 environment. Two-photon transitions at 430
nm take place preferentially when 6 = 371/4, favoring probe molecules in a pH 6
environment. This explanation in the frequency domain is consistent with the observed

results and simulations. Note that phase shaping does not tune the wavelength of the laser

50

 

 

 

SI

Ol

5P6

in il
tam.

Come

pulse. The changes in the two-photon probability for different frequencies arise from the
interference, which is controlled by the spectral phase imprinted in the pulses.“’ 66

Pulse shaping and, in particular, the use of MII provide a number of advantages
for multiphoton microscopy. (a) A very short pulse laser is set up once and optimized
once (no subsequent tweaking). (b) The short pulse duration has the advantages of higher
multiphoton excitation efficiency and broader bandwidth. (c) As the phase is optimized
for selective excitation, maximum efficiency is maintained; there is no need for tuning,
realigning, tweaking, compensating for chirp, and renorrnalizing. (d) The system is
computer-controlled and can take advantage of phase compensation at the sample using
MIIPS, as described here. (e) MII can be used to control two- or three-photon excitation.
(1) During the selective excitation experiments the intensity and spectrum of the laser
remain constant; therefore, one-photon processes like absorption, reflection, and

scattering remain constant as well. (g) Selective photochemistry can be realized through

shaped pulse excitation and this can be used for selective multiphoton microlithography.

3.1.5 Conclusion

Selective excitation of large fluorescent probes is achieved by optimizing the
overlap between the power spectrum of the pulse and the two-photon absorption
spectrum of a molecule. Phase modulation of ultrashort pulses through MII can be used
to probe chemical changes that take place within cells as they respond to external changes
in their environment. The effect can be further enhanced using even shorter (larger
bandwidth) femtosecond pulses. The quantification of changes in intracellular ion

concentration can be accomplished with the use of fluorescent indicator dyes102 based on

51

shifts in their two-photon excitation spectrum. MII has demonstrated applicability in the
study of biochemical phenomena, and this is validated by its significant role in selective
microscopy (discussed in Section 3.2). Practically all of the cellular processes can
potentially be affected by changes in intracellular pH. The absorbance and/or emission
properties of fluorescent dyes are affected by their environment, including solvent, pH,
and conjugation to other macromolecules (e.g., nucleic acids). Changes in intracellular
pH are often one of the responses of cells to externally applied agents, including grth

3 The principles of MII can be similarly applied to other

factors and pathogens.10
biologically relevant local environment parameters, like sodium or calcium ion

concentration.

52

3.2 Selective Two-Photon Microscopyb

3.2.1 Introduction

Two-photon microscopy has provided researchers with unique possibilities for
fluorescence imaging and photochemistry. It offers attractive advantages, including
higher resolution, background-free signal, lower background scattering, better penetration

92’ '04' 105 which arise from the

in thick samples, and reduced photon-induced damage,
basic physical principle that the absorption depends on the square of the excitation
intensity. Two-photon microscopy has been shown to be amenable to multiple probe
staining, whereby two-photon transitions excite different probe molecules that emit at

"’5’ ‘06 and for functional imaging of living cellsm'“0 Here, we

different wavelengths
demonstrate how phase-modulated femtosecond pulses can selectively excite one type of
probe molecule only, leaving the others in their ground state. Selectivity in multiphoton
excitation is achieved by M11 as outlined below.65’ 66’ 1“ Selective excitation can be used

to enhance contrast and to achieve functional imaging of samples stained with fluorescent

probes sensitive to their microscopic chemical environment.

3.2.2 Experimental Section

The experiments are carried out using a Ti:S oscillator (K&M Labs) laser system
capable of generating pulses as short as 11 fs after a double-pass prism compressor. The
amplitude profile and spectral phase of the pulse are tailored using a computer-controlled
pulse shaper of a design similar to those described by Weiner.90 The shaper uses a pair of

SF-10 prisms as the dispersion elements and 0.4 m cylindrical mirrors for collimation. A

 

b Reproduced in part with permission from I. Pastirk, J.M. Dela Cruz, K.A. Walowicz, V.V. Lozovoy, M.
Dantus, Selective two-photon microscopy with shaped femtosecond pulses, Opt. Express 2003, 11, 1695-
1701. Copyright 2003 Optical Society of America.

53

schematic of the experiment is shown in Figure 3.6. Phase modulation is introduced by a
computer-controlled SLM (SLM-256, CRI Inc.), located at the Fourier plane of the
shaper setup. For the experiments shown here, the pulses were centered near 800 nm. The
spectral phase of each pulse was calibrated using the MIIPS method, first to obtain TL
pulses and then to introduce the desired phase.” 111 The spectral phase function used for

this experiment is given by Equation 1.1.

 

 

 

   
 

 

 

 

 

[—W—lLiquisz
I cooled CCD ' : L2

1 =
a Objective I 0
Sample : g
..... .C

Q

O.

(D

(P.

i:

E

9

u.

 

Pulse Shaper

 

 

 

Figure 3.6. Schematic experimental setup for selective two-photon microscopy.
Femtosecond laser pulses are compressed and sent to the pulse shaper. The modulated
beam is then focused on the microscope slide with the specimen. The two-photon induced
fluorescence is collected by a microscope objective and imaged on a CCD.

The shaped laser pulses, with energy of 0.2-1 ml per pulse and 87 MHz repetition
rate were focused mildly, to a spot size of ~ 200 microns in diameter, on the different
samples. Two-photon induced fluorescence was collected by an infinity corrected
apochromatic microscope objective (APO 100X, Mitutoyo) and imaged into a liquid

nitrogen cooled charge-coupled device (CCD) detector using an InfiniTube (Proximity

Series, Infinity Photo-Optical). Experiments were carried out on three different types of

54

 

 

 

 

 

 

ml“

SE11

lc

Ill

13

ml:

samples. The first sample was prepared with acrylonitrile-vinylidene chloride polymer (1
mL) cured in acidic (pH 6) or basic (pH 10) buffered conditions and stained with the pH-
sensitive fluorescent probe HPTS (10'3 M) (Sigma-Aldrich). Both acidic and basic HPTS
mixtures were smeared to an approximate thickness of 0.07:0.03 mm on contiguous
regions of a microscope glass slide, and then allowed to cure for 20 minutes. Some
mixing of the solvent water may have occurred at the interface between the two regions.
Local changes in the fluorescence intensity observed for TL pulses may reflect changes
in the local sample thickness. The second sample consists of pieces of
polymethylmethacrylate (PMMA) doped with cournarin (C540) and rhodamine (R6G)
fluorescent probes”2 (104M). The third sample consists of a mixture of 10 um blue and
15 ,um green fluorescent FluoSpheres (Molecular Probes). The images were captured
without optical filters in the collection optics or post image manipulation. The

background, detector counts when the laser is off, was subtracted.

3.2.3 Results

Figure 3.7A shows an image of acidic and basic HPTS solutions obtained with TL
pulses. The spectrum of the pulse and the spectral phase (flat line) are shown in the
middle panel. The right panel shows the absorption spectra of HPTS under acidic and
basic conditions, including the laser spectrum corresponding to E2(t). Figure 3.7B shows
an image obtained with laser pulses optimally shaped for selective excitation of HPTS
under acidic environment. The spectral phase used to acquire this image is shown in the
middle panel; the spectrum corresponding to this phase is shown in the right. Figure 3.7C

shows an image obtained with laser pulses optimally shaped for selective excitation of

55

 

HPl
the I
tom

inl

Wile
PUI‘Sl
km:
111 a
Photr

pilOlO

{I
lIIIOIEsr

HPTS in a basic environment. The spectral phase used to acquire this image is shown in
the middle panel; the right panel shows the spectrum corresponding to this phase. The
contrast ratio between pH lO/pH 6 in these images varies from 1:1 in Figure 3.7A to 1:10

in B to 64:1 in C.

 

 

 

 

 

 

 

  
  

 

 

 

 

 

 

 

 

 

 

Ina/\Lw
/ g r:
. g "5
9142,... .5 E
‘. a s
\‘ , .11 E (a).
(D .
g :5
1:
+7! '1
" || "
50pm «~ ’7' .- a
— o ‘ ne' .
eoo Mnm) 930 350 Mnm) soc

 

pH=6 pH=10

Figure 3.7. Experimental demonstration of pH-sensitive selective two-photon
microscopy. The sample being imaged has an acidic (left side, pH 6) and a basic (right
side, pH 10) region, both labeled with HPTS. A. Image of the sample obtained with TL
pulses. The spectrum of the 21 fs laser pulses centered at 842 nm and the phase of the
pulse across the spectrum are shown in the middle diagram. The absorption spectra of
HPTS under acidic (dark gray) and basic (black) conditions are shown together with the
laser spectrum corresponding to EU). B. Image of the same sample and location obtained
with pulses that have been optimized for selective excitation of HPTS in an acidic micro-
environment. Notice that only the left region shows significant two-photon excitation.
The spectral phase function used for this experiment is given by ¢(Q)=acos(}Q—6),
where Q is the frequency detuning measured from the center wavelength of the laser
pulse. For this image a = 1.57r, y= 20 fs, and 6 = 0757:. C. Image of the same sample and
location obtained with pulses that have been optimized for selective excitation of HPTS
in a basic micro-environment. Notice that only the right region shows significant two-
photon excitation. For this image a = 1.57:, y = 20 fs, and 6 = 0257:. Selective two-
photon excitation from the two pH regions in the sample is observed at specific values of

The experimental results in Figure 3.8 show selective excitation of different

fluorescent probes. Figure 3.8A shows the data obtained using TL pulses. The two

56

imaged pieces of PMMA exhibit similar amounts of two-photon LIF. When the spectral
phase of the pulse is modified to optimize selective excitation of the C540-doped sample,
we observe that only the top fragment fluoresces, as shown in Figure 3.8B. When the
spectral phase of the pulse is optimized for selective excitation of the R6G-doped sample,
Figure 3.8C, we observe strong fluorescence from the bottom fragment while the other
side, containing the C540 probe, does not fluoresce. Note that the spectrum and intensity
of the laser remain constant during the experiment for all three cases. The only change
occurring is the subtle modulation of the spectral phase of the pulse, which controls
MII.65’ 66 Contrast ratios (IR6G/IC540) (normalized intensities corrected to background)
range from 1:9 for pulses optimized for C540 to 8:1 for pulses optimized for R6G.
Selective multiphoton excitation minimizes the cross talk between different fluorescent

probes and does not require the use of multiple filters and detectors for imaging.

C540

 

R6G

Figure 3.8. Selective two-photon microscopy of pieces of PMMA doped with different
fluorescent probes. A. Image showing two-photon induced fluorescence from both
pieces: C540 doped PMMA (top) and R6G doped PMMA (bottom), obtained with 17 fs
TL pulses centered at 790 nm. B. Image obtained with pulses optimized for selective
C540 excitation. For this image the pulses were shaped with a sinusoidal function with
6=0.317t. C. Image obtained with pulses optimized for selective R6G excitation. For this
image the pulses were shaped with a sinusoidal function with 6=0.741r. Selective two-
photon excitation from the two PMMA pieces is observed at specific values of 6.

The previous experiment showed selective two-photon excitation of different

probe molecules. To further illustrate the potential of coherent control for functional

57

imaging we assessed the possibility of selectively exciting blue- and green-fluorescent
microspheres, which are typically used for targeted staining of biological samples.
Fluorescence observed from these microspheres is shown in Figure 3.9. The top part of
the image shows the green-fluorescent microsphere (15 ,um diameter) which has an
absorption maximum at 450 nm, while the bottom part represents the blue-fluorescent
microsphere (10 um diameter) with an absorption maximum at 365 nm. The image in
Figure 3.9A was acquired with TL pulses. Under these conditions, large two-photon
induced fluorescence signal is observed from both spheres. For the image in Figure 3.9B,
the phase of the pulses was modulated such that the intensity of fluorescence from the
blue microsphere is maximized while minimizing fluorescence from the green
microsphere. For the image in Figure 3.9C, we modulated the phase of the pulses such
that the fluorescence from the green microsphere is maximized and that of the blue one
minimized. The observed contrast ratios (IBLUE/IGREEN) are 1:3 and 4:1 for the different

phases.

 

Figure 3.9. Selective two-photon microscopy of 10 ,um blue and 15 um green fluorescent
polystyrene microspheres. A. Image showing two-photon induced fluorescence from both
microspheres, obtained with 15 fs TL pulses centered at 790 mm. B. Image obtained with
pulses optimized for selective excitation of the blue microsphere. For this image the
pulses were shaped with a sinusoidal phase with 6=0.757r. C. Image obtained with pulses
optimized for selective excitation of the green microsphere. For this image the pulses
were shaped with a sinusoidal phase with 6=1.257r. Selective two-photon excitation from
the two microspheres is observed at specific values of 6.

58

3.2.4 Discussion

Selective excitation with significant contrast ratios has been achieved here by
optimizing the overlap between the power spectrum of EU) and the two-photon
absorption spectrum.“ “ The addition of a computer-controlled pulse shaper to the
multiphoton microscope provides a number of important advantages. First, the pulse
shaper can be used to compensate unwanted phase distortions at the sample. Linear chirp,
for example, has been shown to reduce signal intensity in two-photon microscopy.”3.
With a pulse shaper, linear, quadratic, cubic and higher order chirp can be compensated
to obtain the most efficient excitation.97 Second, the pulse shaper can be used to control
the output spectrum of the laser pulses by amplitude modulation. Here, for example, we
used pulses centered at 790 or at 840 nm. Third, the pulse shaper can be used for
selective probe excitation. Because the spectrum of the laser remains constant, phase
modulation does not affect one photon processes such as absorption, reflection and
scattering. Selective excitation minimizes possible cross talk between different
fluorescent probes in the sample. Finally, the pulse shaper can be used to prevent three-
photon and higher order nonlinear optical processes such as continuum generation.
Higher order processes usually lead to sample degradation, and in the case of living
samples, to DNA damage.1 14’ ”5 Suppression of three-photon transitions of four orders of
magnitude has been achieved using the M11 method“ and this suppression can be coupled
with optimization of two-photon signal from living specimens.

Recently Ogilvie et. al. adopted the sinusoidal phase modulation approach
towards selective two-photon excitation microscopy to image live eGFP (green

fluorescent protein) labeled Drosophila embryos. Pulse shaping, in this case, allowed

59

selectively excitation of eGFP or of endogenous fluorescence.”6 The ability to separate
endogenous two-photon induced fluorescence, which is ubiquitous in biological samples,
is an important achievement that greatly enhanced the contrast of the obtained images.
Although selective excitation can be carried out by wavelength tuning of
narrower-bandwidth lasers, the more practical approach of modulating the phase of a
broad-pulse femtosecond laser has a number of advantages. In wavelength tuning,
spectral phase measurements at every desired excitation wavelength are required and
movement of optics causes problems when trying to overlap images obtained at different
wavelengths (a pixel registration problem). By phase modulation, the whole process of
spectral phase measurement and compensation is carried out in advance and excitation at
different wavelengths within the bandwidth of the ultrashort pulse is achieved, requiring

no physical movement of optical materials.

3.2.5 Conclusion

We have demonstrated selective two-photon microscopy of fluorescent probe
molecules using phase-modulated pulses as short as 15 fs. This method can be used to
selectively excite either different probe molecules or identical probe molecules in
different environments. In addition, this method can be used for selective excitation of
luminescent probes such as quantum dots,117 metallic nanoparticles, and single

molecules.1 18

The same principle can be extended to achieve functional imaging of
semiconductor microchips by two-photon laser induced conductivity.”9’ '20 Having a

pulse shaper for multiphoton microscopy provides the flexibility of selective probe

excitation or maximum signal enhancement by controlled modulation of the spectral

60

phase of the femtosecond pulses. Even for fluorescent labels with very similar absorption
spectra, pulse shaping has the capability of selective excitation.45 This level of selectivity
and enhancement can be adapted to different modes of two-photon and three-photon

microscopyl‘zl'124

61

CHAPTER 4

COHERENT LASER CONTROL IN IMAGING

For many years, lasers in the medical field were seen primarily as surgical tools
for the ablation, cutting, or coagulation of tissue. But in the last decade some of the most
successful medical-laser applications have been noninvasive (e.g. skin rejuvenation, hair
removal, PDT, low-level laser therapy, and, more recently, laser treatments of psoriasis,
acne, cellulite, and dental bacteria). In addition, laser-based diagnostic tools such as
optical coherence tomography125 and multiphoton microscopy92 have gained acceptance
within the medical community for noninvasive imaging, while optical sensors are finding
a place in the detection and monitoring of chronic and other diseases. Recent research
efforts have focused on making medical diagnosis cheaper, faster and more effective by
developing new physics-based imaging techniques and by using powerful computers to
process and analyze biomedical information. New technological developments are being
increasingly driven by the needs of the medical profession, and optics is playing an
increasingly important role.

Optical spectroscopy is a standard tool in laboratory analysis that is used to detect
and monitor the concentration of substances, such as oxygen, by measuring the
absorption, fluorescence or Raman spectra. But performing spectroscopy in biological
tissue is very different from making measurements on a solution in a test tube.
Quantitative optical measurements in biological tissue are extremely difficult because
tissue scatters light strongly, typically within 100 pm. Scattering within tissue results

from macroscopic and microscopic constituents such as cellular membranes, refractive

62

index mismatching between intra- and extracellular fluids, mitochondria, ribosomes and
fat globules. This limitation has prevented the widespread application of optics in clinical
medicine.

If the problems associated with scattering can be overcome, optical imaging could
provide a cheaper technology that would allow patients to be screened for diseases such
as skin and breast cancer, osteoarthritis and diabetes, as well as revealing much more
about the working processes of organs, including the brain.

In this chapter, the possibility of working around the impediment of tissue
scattering to selectively excite molecules based on their chemical microenvironment is
presented (Section 4.1). The sensitivity of nonlinear imaging methods to the spectral
phase of the pulses is highlighted; controlling this phase can lead to significant
improvements and new possibilities in functional biomedical imaging (Section 4.4). The
importance of delivering accurate pulses to the sample for imaging, identification and
possibly therapeutic purposes is also demonstrated (Sections 4.2 and 4.3). In addition, the
advantage of spectral phase modulation in terms of minimizing the damaging effects of

three-photon excitation on biological samples is discussed (Section 4.5).

63

4.1 Phase Modulation through Scattering Tissuec

4.1.1. Introduction

In the last decade a number of reports have demonstrated controlled excitation of
atoms and even the controlled photochemistry of isolated molecules.“ 37’ 59 More
recently these reports have included large organic molecules in solution.12’45’ 47’ 5' Phase
shaped femtosecond pulses90 have the potential for influencing the quantum-mechanical
laser-molecule interactions, thereby controlling population transfer between different
electronic states of large organic molecules.”’ 9" 95’ 126429 The benefits of coherent laser
control of large organic molecules are beginning to permeate to the biological field
through selective two-photon chemical microenvironment probing and microscopy. Here
we question if coherent control methods, in particular phase shaping, can be used in
situations where the laser transmits through scattering biological tissue.

Coherent control depends on the interference among multiple excitation
pathways; this interference is controlled by modulating the phase of the excitation field.
Control schemes therefore depend on the precise phase modulation of the laser pulses,
which, through constructive and/or destructive interference, optimize the desired outcome
and minimize other pathways. The dependence of coherent control on the accurate phase
structure of the pulse, however, suggests that these approaches would not be applicable to
situations involving transmission through scattering media.

This section presents a study that evaluates the possibility of using coherent
control as the laser transmits through scattering tissue. In particular, the ability to achieve

selective two-photon excitation of a pH-sensitive probe molecule using phase-shaped

 

° Reproduced in part with permission from J.M. Dela Cruz, 1. Pastirk, M. Comstock, M. Dantus,
Multiphoton intrapulse interference 8. Coherent control through scattering tissue, Opt. Express 2004, 12,
4144-4149. Copyright 2004 Optical Society of America.

64

femtosecond pulses in the presence and absence of scattering tissue is evaluated. The
motivation for the experiment is that coherent control schemes could enhance methods
such as two-photon imaging, which has provided images with higher resolution, lower

92’ '30 Selective two-photon

background scattering, and better sample penetration.
excitation could be used to help localize two-photon induced PDTm, and restrict its

therapeutic effects to diseased tissue, while leaving surrounding tissue undamaged.

4.1.2 Experimental Section

The setup for the experiment is shown schematically in Figure 4.1. The
experiment was carried out using a Ti:S oscillator (K&M Labs) laser system. The laser
spectrum was centered at about 830 nm, with a bandwidth of 90 nm (FWHM),
corresponding to 11 fs pulse duration. The average energy per pulse entering the
objective was 1 nJ. The laser pulses were calibrated at the sample, after the objective,
using the MIIPS method.97 The spectral phase functions used for this experiment are
based on binary phase shaping (BPS), a method that assigns phase values of zero or 7: to
frequency components of the pulse, and is ideally suited for controlling multiphoton
transitions.77 Optimization of the phase structure was achieved by accounting for the
excitation spectra of probe molecule in acidic and basic environments, the characteristics
of the laser pulse, and M11.“’ “ A learning algorithm was then used to optimize
excitation—maximizing the excitation of one form of the probe molecule while
minimizing excitation of the other form of the probe, HPTS fluorophore in acidic or basic
solution, respectively. Two optimized BP functions were introduced into the pulse

shaper. The sample was scanned in the focal plane of the laser, focused by a 5x long-

65

working distance objective. The capillary tubes (1 mm id) were filled with 10'3 M HPTS
aqueous pH6 and leO buffered solutions. A fifth tube was filled with water to serve as a
control to demonstrate the absence of signal due to scattered light. The tubes were placed
side by side and oriented vertically in a 2 mm quartz cell filled with water. The
fluorescence was focused onto a spectrometer with detection wavelength set at 515 nm
(the emission wavelength of HPTS) where it was detected and averaged point by point. A
1.5-2.0 mm slice of biological tissue (raw breast of chicken) was placed between the front
face of the cell and a microscope slide. The tissue was slightly compressed to a uniform
thickness of about 1 mm. A drop of Nujol was used to reduce aberrations caused by index

of refraction mismatch. Approximately, 1 mm of tissue is equivalent to 10 scattering

lengths.

Sample Capillaries

Scattering Tissue

Fluorescence
Detector

  
 

Shaped pulses

Figure 4.1. Experimental setup. The shaped laser pulses impinge on the sample where
fluorescence is detected in the forward direction. The sample, with or without scattering
tissue, is scanned in the focal plane of the laser while the two-photon induced
fluorescence is detected at each point.

66

The spectral phase for TL pulses and the BP functions that were used to enhance
excitation of HPTS in an acidic or a basic environment are shown in Figure 4.2. The SHG
spectrum obtained from a thin nonlinear optical crystal, also shown in Figure 4.2,
provides a diagnostic test of the effects of phase modulation. Unlike the TL pulses, the
shaped pulses preferentially excite one of the two different pH solutions. The spectrum of
the laser pulse does not change upon phase modulation, only its potential to cause two-
photon excitation or generate second harmonic radiation“ “ The change in the spectrum

is caused by destructive interference induced by the phase mask.”

 

 

..L
o

p
m
4‘ l A

TL

Milli!" “rill

BP1O A

................ LA

720 760 800 840 880 920 '
Wavelength (nm) Wavelength (nm)

9’
9’

 

9’
h

  
 

F’
k)

 

 

Normalized intensity (a.u.)

 

 

 

F’
c:

Figure 4.2. Optimized phase functions and the resulting SHG. The panel on the left
shows the fundamental spectrum of the pulse centered near 830 nm and the three
different pulses evaluated, TL with a flat spectral phase, and shaped pulses optimized for
acidic (BP06) and for basic excitation (BPlO). The panel on the right shows the
experimental SHG spectrum obtained from TL pulses and SHG spectra obtained when
the phase functions BP06 and BPIO were introduced by the pulse shaper. Notice that
BP06 suppresses SHG at long wavelengths and BPIO suppresses SHG at short
wavelengths.

4.1.3 Results
When TL pulses were used, large two-photon induced fluorescence signals at 515

nm, the peak emission wavelength of HPTS in both pH solutions, were observed. The

67

contrast that could be achieved was tested by recording the signal observed when the
laser was focused on each of the two HPTS solutions. This test was carried out using two
capillary tubes and no scattering tissue. As can be seen in Figure 4.3, BP06 preferentially

excites the pH6 solution, and BP10 preferentially excites the leO solution.

 

.3
o
1

 

Normalized intensity (a.u.)
.0
01

_o
q:

 

 

pH6 pH 10

Figure 4.3. Fluorescence signal obtained from two capillaries containing HPTS in
buffered solutions. The signal was obtained after excitation with TL and shaped pulses.

Figure 4.4 shows a plot of the signal obtained from the setup indicated in Figure
4.1 with and without scattering tissue. The signal from TL pulses was multiplied by —1
and plotted as a function of position. The peak height obtained from both acidic and basic
solutions is essentially the same. The small variations are caused by small shifts in the
position of the focal plane. The capillary containing pure water gives no signal as
expected. The difference signal obtained from the two-photon fluorescence signals
(optimized for acid pH minus optimized for alkaline pH) obtained using the two masks is
shown in Figure 4.4 as open circles. For this measurement the signal obtained for pH 10
samples was normalized to one. Notice that the contrast, a measure of efficiency of

coherent control, is quite high. In the absence of tissue the contrast ratio is 10:1. More

importantly the contrast is found to survive the presence of ~ 1 mm of tissue. The

68

contrast obtained (6:1) has been essentially maintained, although the overall signal

decreases exponentially with the tissue thickness.

 

(a.u.)

No tissue .-, With tissue

 

-0— Difference 0' '. —o- Difference .. o_

hoton fluorescence

 

 

 

 

Position (mm) Position (mm)

Figure 4.4. Line scan of sample without (left) and with (right) tissue. The sample,
consisting of 5 capillary tubes containing acidic or basic HPTS and water (control), was
scanned in the x direction. Experimental results obtained with TL (black dots) pulses are
plotted in the negative axes. Difference plots obtained from the shaped laser pulses (open
circles) are plotted in the positive axes. Notice that shaped laser pulses are capable of
selective excitation even when the laser transmits through scattering tissue.

4.1.4 Discussion

The goal of this research was to establish if phase-shaped pulse strategies
designed to control the excitation and chemistry of molecules survive transmission of the
laser through scattering tissue. The results of our experiment indicate that the ballistic
photons, which are the ones responsible for the two-photon fluorescence detected, retain
the shaped pulse characteristics required for selective excitation. This was confirmed by
experimentally measuring the SHG spectrum of the shaped pulses before and after
transmission through biological tissue. Figure 4.5 shows the SHG spectrum of TL pulses
and the SHG spectrum of shaped (BP06) pulses with and without tissue. The result

obtained with tissue gave a much weaker signal that was multiplied by a factor of 20 for

the illustration. Note that the SHG spectrum of the shaped pulses suffered little or no

69

change other than attenuation. SHG and two-photon excitation are achieved only by the
coherent ballistic photons. This is why selective two-photon excitation is possible

through the tissue.

 

.0 .
a:
l .1

.°
N

 

 

 
   

Normalized intensity
0
A

“$33 I 410 420 1‘ 431) 440
Wavelength (nm)
Figure 4.5. SHG spectra for TL and for shaped pulses (BP06) in the presence (open
circles, intensity multiplied by 20) and absence (black squares) of biological tissue.
4.1.5 Conclusion
Coherent control methods can be used to improve biomedical applications of laser
technology that presently use TL femtosecond pulses. In particular, optical coherence
tomography, PDT, and functional imaging stand to gain the most from the results
observed because coherent control should be able to provide better contrast and allow the
use of more intense laser pulses while damage to healthy tissue is minimized by
intrapulse interference. This method is useful for depths up to 3 mm. Delivery of the
beam to internal organs could be achieved using optical fibers. For some imaging
techniques, it isn’t possible to collect fluorescent light in the forward direction, as done

here, and one must collect fluorescent photons in the backward direction

(epifluorescence). The method used here could be implemented in that modality; an

70

inverted microscope could be used for this purpose. Bringing these experiments to actual

biomedical applications is a goal we hope to achieve in the future.

4.2 Phase Compensation of Imaging Objectivesd

4.2.1 Introduction

The nonlinear properties of a pulse are very sensitive to spectral phase distortions
that are either inherent to the laser system or are introduced by lenses, optical fibers,
mirrors, or microscope objectives. Eliminating these distortions is especially critical in
cases where reproducibility is an important criterion for femtosecond laser applicability.
These distortions need to be corrected to ensure the fast progress of femtosecond laser
excitation in biomedical imaging applications.

One of the greatest obstacles to fully developing the promising technique of
multiphoton imaging is the significant amount of phase distortion introduced by high
numerical aperture (NA) microscope objectives. To date, the characterization of such
distortions has been accomplished for sub-10 fs pulses in objectives with NA up to
0.85,132 and for up to 1.30 NA objectives if longer pulses are usedm'136 Compensation
of the quadratic term of the characterized distortions was accomplished in some cases
through precompensation. However, higher order phase distortions, introduced by the
objectives and even by the prisms necessary to precompensate for the quadratic phase
distortions, cannot be corrected by these methods, nor can the phase distortions be

quantified. In this section, the ability of MIIPS to successfully quantify and compensate

 

d Reproduced in part with permission from B. Xu, J .M. Gunn, J.M. Dela Cruz, V.V. Lozovoy, M. Dantus,
Quantitative investigation of the MIIPS method for simultaneous phase measurement and compensation of
femtosecond laser pulses. J.0pt. Soc. Am. B 2006, 23, 750-759. Copyright 2006 Optical Society of

America.

71

for all orders of arbitrary phase distortions introduced into a sub-10 fs pulse by high NA

microscope objectives is presented.

4.2.2 Experimental Section

A periscope was used to bring the laser beam (800 nm, 10 fs, 250 mW average
power) into the rear port of a Nikon TE2000-U inverted microscope. The beam was
reflected upward through the objective and focused onto an SHG crystal. Three
objectives were tested: a Nikon Plan Fluor ELWD 20x/0.45 NA objective, a Nikon Plan
Fluor ELWD 40x/0.6O NA objective, and a Nikon Plan Apo TIRF oil immersion
60x/1.45 NA objective (Nikon immersion oil, Type NF). Phase characterization and
compensation were carried outwith MIIPS. To test the viability of pulse characterization
and compensation on imaging, a mouse kidney section (FluoCells® prepared slide #3
mouse kidney section with Alexa Fluor® 488 WGA, Alexa Fluor® 568 phalloidin,

DAPI) was imaged with either phase-corrected TL pulses or with uncompensated pulses.

4.2.3 Results and Discussion

MIIPS alone was sufficient to compensate for the phase distortions introduced by
20x/0.45 NA and 40x/0.60 NA microscope objectives. To allow for the best
compensation of higher order phase distortion introduced by the 60x/1.45 NA objective, a
pair of SFlO prisms was used to reduce the quadratic phase contributions (~104 fsz).

Figure 4.6 shows the compensation results for the 1.45 NA objective. In both
panels, the solid line shows the average phase residue from five independent

measurements. The errors bars show the standard deviation for every fifth data point.

72

The top panel shows the full set of acquired data; the lower panel shows the same data
over the FWHM of the fundamental laser spectrum with a reduced scale. Note that the
phase is compensated to within 0.1 rad over the entire bandwidth of the pulse (770-860
nm), indicating that the distortions introduced by high NA objectives are well
compensated over this range. The extremely narrow standard deviation observed in the
reduced range provides evidence of a very high degree of reproducibility of the MIIPS

method, even when highly dispersive materials are utilized. The statistical phase error

 

for the 60x/l .45 NA objective was calculated to be 0.026 rad over the full range of data,
where the uncompensated phase distortion was ~40 rad. Considering the significant
degree of distortion introduced by this high NA objective (see inset of Figure 4.6), the

precision by which phase is retrieved by MIIPS is excellent.

40

“/2- ,'\.. €20
, a go

,7 \\..,-20

r ‘40

 

Phase (rad)
O
i'

 

 

 

 

 

 

 

. ........ ......... ........ l
780 800 820 840 860
Wavelength (nm)

 

Figure 4.6. Compensation of spectral phase distortions caused by a 60x/1.45 NA
objective. The error bars indicate a: 1 standard deviation. The upper panel shows the full
range of data, while the lower panel shows the same data over the FWHM of the pulse,
on a reduced scale. The inset (top panel) shows the phase extracted before compensation
with MIIPS.

73

The images shown in Figure 4.7 demonstrate the importance of spectral phase
compensation. The compensation, and any subsequent application of phase, is
accomplished through the introduction of computer instructions. The laser beam is never
tweaked or scanned, reducing the potential for error and the ease of use, as both actions
require a significantly higher level of expertise and may result in changes of the phase
distortions and, in some cases, diminished registration in multiple images. No
fluorescence wavelength filters were used to separate contributions from the fluorescent
labels in the sample. Comparing between the two images, it is clear that phase distortion
leads to a significant loss of signal (Figure 4.7, right) that cannot be compensated simply
by increasing the laser power. This is because phase distortions cause changes in the
wavelength region where multiphoton excitation takes place“ “ The benefits of pulse
characterization and compensation are realized with the improvement in image quality

(Figure 4.7, left).'37

 

Figure 4.7. Image quality improvement with spectral phase compensation. Left:
fluorescence image of mouse kidney using femtosecond laser pulses with spectral phase
deformations corrected. Right: fluorescence image of the same section of the sample
when pulses are uncompensated (no spectral phase correction). Both images were
normalized to the same maximum intensity level. Image size is 30 ,um x 30 ,um.

74

4.3 Phase Compensation through Scattering Tissuee

4.3.1 Introduction

When a laser beam transmits through a scattering medium like biological tissue,
the coherent component of the electric field, although maintaining its directionality,
degrades exponentially with tissue depth. This occurs as a result of the heterogeneous
nature of tissue structure. Although femtosecond lasers have increasingly been used for
medical diagnostics and therapeutics as well as for imaging applications, an underlying
uncertainty exists as to the spectral phase distortions the laser pulse experiences as it
propagates through several scattering lengths of tissue. In this section, the ability of the
MIIPS method to measure spectral phase deformations caused by transmission through

scattering biological tissue is investigated.

4.3.2 Experimental Section

The procedure involved measuring phase distortions inherent in a sub-10 fs laser
pulse, correcting for these distortions to obtain TL pulses, and then comparing the
spectral phase obtained when the laser transmits through a 500-;1m slice of raw chicken
breast tissue or through a similar thickness of glass. Preliminary experiments using a
variable scattering medium (skim milk in water) indicated that the retrieved spectral
phases had very little dependence, other than on the overall intensity, on the degree of

scattering. This is indicative of the ability of the coherent component of the laser or the

 

‘ Reproduced in part with permission from B. Xu, J .M. Gunn, J.M. Dela Cruz, V.V. Lozovoy, M. Dantus,
Quantitative investigation of the MIIPS method for simultaneous phase measurement and compensation of
femtosecond laser pulses. J.0pt. Soc. Am. B 2006, 23, 750-759. Copyright 2006 Optical Society of

America.

75

 

ballistic photons to maintain phase information as the laser penetrates a turbid, scattering
medium.

4.3.3 Results and Discussion

Our results with tissue (Figure 4.8) confirm that despite the presence of
inhomogeneities that significantly affect the coherence of a femtosecond laser pulse,
phase distortion is minimal. As expected, a reduction in signal to noise ratio resulting
from the exponential decay of ballistic photons as the pulse transmits through several
scattering lengths of tissue was observed. Notwithstanding this limitation, phase
information is still comparable to that obtained when light goes through an optical
medium like glass. F igure 4.8A and B show the MIIPS traces obtained for glass and
tissue samples, respectively. The presence of tissue slightly decreases the range of
wavelengths that can be compensated by MIIPS, but phase correction is nonetheless
achieved. Figure 4.8C shows the phase residue, averaged over five trials, of the pulse as
it went through glass and tissue. Note that the phase is compensated to within 0.02 rad
over the entire FWHM of the pulse. With additional iterations, MIIPS can fully
compensate for distortions introduced by a scattering medium. More significantly, the
method characterizes and corrects for distortions at the position of the sample.
Consequently, the phase of a pulse can now be known at the position of the sample, even
under poor signal to noise conditions, without autocorrelation or interferometry.

‘38 show that even for signal to noise ratios as low as one, it is possible to

Simulations
retrieve the spectral phase with high precision by smoothing the two-dimensional MIIPS
data before analysis. The good resistance to noise and the fact that it is independent of

mode quality are additional significant advantages of the MIIPS method.

76

 

 

 

Wavelength (nm)

Wavelength (nm)

 

 

 

 

1 Phase residue C

i — through glass
0.11 -..- through tissue

Phase (rad)
o
g
:5
I

IO
.
A
L441-1-1 1 1 A

 

 

mJ

 

I ' ''''''''

"760T“ 300 820 840
Wavelength (nm)

Figure 4.8. MIIPS measurement through scattering biological tissue. Panels A and B
show MIIPS traces without and with tissue, respectively. While the overall signal to
noise ratio is significantly decreased, similar phase information is obtained (C).

77

 

4.4 Functional Imaging with Shaped Laser Pulsesf

4.4.1 Introduction

Interest in coherent laser control has grown steadily, given its potential for
influencing quantum-mechanical laser—molecule interactions.”’ 9" 95’ ”6"” With the use
of pulse shapers and computer learning algorithms, scientists have controlled the

excitation of isolated atoms and molecules“: 37' '39

and, more recently, large molecules in
solution.“ 45’ 5' Population transfer between different electronic states of large organic
molecules can be maximized or minimized through manipulation of the spectral phase of
ultrashort laser pulses“ “ The benefits of coherent laser control of large organic
molecules can be extended to the biological field through selective two-photon chemical

0 Control of laser-matter interactions

microenvironment probingIll and microscopy.l4
could revolutionize biomedical applications. Coherent control schemes, for example,
could enhance methods such as two-photon imaging” '30 which has provided higher
resolution, lower background scattering, and better sample penetration than traditional
techniques. Selective two-photon excitation would be desirable for distinguishing healthy
from cancerous tissue based on differences in their chemical properties such as pH, or for
activating a PDT agent only when it is absorbed by cancer cells by using two-photon
activated PDT.131 These improvements would be possible only if shaped laser pulses
maintain their unique properties as they transmit through scattering biological tissue.

Coherent control of laser-induced processes is based on the quantum interference

7

among multiple excitation pathways.12 Progress in this field has been fueled by

advances in pulse shaping technology, allowing control of the phase and amplitude across

 

f Reproduced in part with permission from J.M. Dela Cruz, 1. Pastirk, M. Comstock, M. Dantus, Use of
coherent control methods through scattering biological tissue to achieve functional imaging. P. Natl. Acad.
Sci. USA 2004, 101, 16996-17001. Copyright 2004 by The National Academy of Sciences of the USA.

78

the bandwidth of ultrashort laser pulses.“

Control schemes depend on introducing
intricate phase structures on the laser pulses that, by means of constructive and/or
destructive interference, optimize the desired outcome and minimize other pathways.
This dependence on the accurate phase structure of the pulse suggests that these methods
may not be applicable to situations involving transmission through scattering biological
tissue.

When a laser beam passes through a turbid scattering meditun, its coherence (a

property of waves defined as the degree with which waves are in step with each other)

degrades exponentially as I coherent = 10 exp(—pex), where 1/,u. is the scattering length

of the medium and x is the sample thickness. The transformations of the pulse and the
exponential loss of coherence are shown schematically in Figure 4.9. The coherent
component of the electric field, which is sometimes referred to as the ballistic photons,141
propagates through and maintains its directionality. The incoherent component, or
diffusive photons (shaded broad region in Figure 4.9), loses its directionality because of
scattering and can no longer be used for high-resolution multiphoton imaging.142 In this
context, one can also identify an intermediate component referred to as the snake
photons, whose degree of coherence is still debated.’43 Scattering of a laser in biological
tissue results from the spatial variations of the sample caused by the different cellular

structures and substructures with different indices of refraction, causing changes in the

directionality of portions of the beam and introducing various delays.

79

 

 

 

VScauenn

ooooooo
............

    

BaHEMC

 

 

Intensity
Cb

 

 

 

 

Photon Arrival Time 1'“

 

Figure 4.9. Coherence degradation and pulse transformation as a function of scattering
path length. As a short pulse of light enters a scattering medium, coherent or ballistic
photons (sharp feature) are lost exponentially. The scattered photons (shaded broad
feature), which lag in time, lose their coherence and are randomly delayed.

Coherent control of nonlinear optical processes based on phase-only shaping can
be achieved in scattering biological tissue. To verify this hypothesis, selective two-
photon excitation of a pH-sensitive probe molecule, HPTS, was optimized using phase-
shaped femtosecond pulses. Optimization of the laser-pulse phase structure was based on
the spectroscopic changes exhibited by HPTS in acidic and alkaline environments. Figure
4.10 shows the spectroscopy and chemical structure of HPTS. The hydroxylic proton has
a pKa of ~7.5 and is promptly lost in response to an increase in the pH of its local
environment. The absorption maximum of the protonated species changes from 400 to
450 nm upon deprotonation (Figure 4.10). Interestingly, the fluorescence maximum is
515 nm in both acidic and alkaline pH, because the pKa of the excited state molecule is
much smaller, leading to fast deprotonation in all but highly acidic environments” 1“
Laser-pulse optimization depends on the characteristics of the laser pulse (central

wavelength, spectral phase, and pulse duration) and on MII“’ “ which leads to the

suppression of two-photon excitation at certain wavelengths. An evolutionary learning

80

algorithm129 was used to obtain the greatest excitation selectivity between the two HPTS
species (acidic and alkaline). The selectivity achieved by two optimal phases during
functional imaging with and without the presence of tissue was tested. These phases,
BP06 and BP10, maximize two-photon excitation of pH 6 or 10 solutions, respectively.
This section presents results that confirm selective two-photon excitation after a beam
propagates through biological tissue. To substantiate these results, we measured the rate
of coherence loss with tissue depth, characterized the spectral phase of the pulses after
they were transmitted through biological tissue, and characterized the signal intensity as a

function of scattering and the resolution expected for possible biomedical applications of

“035 CH 'oas 0'
H+
.'_.
4.

laser control.

 

   

.033 $03 035 $03

A 020‘ I—pH6 A—pH8 o—pH9
:l a- pH 10
“3' 0.16‘
8 .
t: 0.12
m
e 0.08
if
< 0.04‘ - °\°\°

0.00t ' ' *l

 

 

r -' v ' v I v ' v ' v ' r t v

360 380 400 420 440 460 430 500
Wavelength (nm)

 

 

Figure 4.10. Molecular structure and absorption spectra of HPTS in acidic and alkaline
pH. Notice that the loss of the hydroxylic proton leads to a large change in the absorption
spectrum.

81

4.4.2 Experimental Section

The sample for the experiment is shown schematically in Figure 4.11. It consists
of three capillary tubes (id, 1 mm) filled with an acidic solution of HPTS placed in an
alkaline solution of HPTS. Frozen raw chicken breast was sliced to a thickness of ~1.5—
2.0 mm. The tissue was thawed and placed between the front face of the cell and a glass
plate and slightly compressed to a uniform thickness of ~0.5 mm. It was observed that the
degree of compression did not affect the nonlinear optical signal. A drop of index
matching fluid was used between the tissue and the glass. A transparent mask with the
printed letters “MSU” was placed in front of the capillaries and behind the tissue as a

reference.

 

7

2P Fluorescence
Detection

 

   
  

 

Scattering Biological

Tissue \

Shaped pulses

capillaries pH 6 soln. I

2P Fluorescence
Detection

 

 

 

,\

Figure 4.11. Experimental setup used for selective two-photon excitation through
scattering tissue. The sample consists of three l-mm capillaries containing an acidic
solution of HPTS submerged in a quartz cell filled with an alkaline solution of HPTS.
The sample is raster scanned (without and with scattering tissue) in the focal plane of the
beam during data acquisition to obtain the images.

82

The experiment was carried out with a Ti:S oscillator (K&M Labs) laser system.
The laser spectrum was centered at ~830 nm, with a bandwidth of 95 nm (FWHM),
corresponding to a 10-fs pulse duration. The large bandwidth is desirable to overlap the
absorption of the two HPTS species (see Figure 4.10) by two-photon excitation with the
single pulse. The laser was sent to a prism-pair compressor and a broad-bandwidth pulse
shaper.90 The laser pulses were calibrated and compensated to TL pulses at the position
of the sample after a microscope objective by using the MIIPS method.97 This step is
critical because it removes unwanted phase deformation introduced by the objective and
other optics in the path of the laser beam. Starting from TL pulses, the condition in which
all frequency components are in step with each other with a zero phase difference (a
blank canvas), ensures reproducible results. The optimized spectral phase functions used
for this experiment are based on BPS.77 By restricting all of the possible values to just
two, the search parameter space for the optimum pulses is reduced by ~100 orders of
magnitude”, ensuring a more thorough search by the GA.

Images were obtained by raster scanning the sample at the focal plane of the laser
with and without scattering tissue. The laser was focused on the capillaries by a 20x
objective (0.45 NA, Plan Fluor extended long working distance, Nikon) (see Figure
4.11). The energy per pulse entering the objective was 1 nJ. The resulting fluorescence
from the sample was filtered with a broadband filter (BG40) and a spectrometer with
detection wavelength set at 515 nm (24 nm resolution) and averaged point by point using
a lock-in amplifier to obtain a two-dimensional (6 x 8 mm) image. Sub-micrometer
resolution could be achieved by this method; however, the goal here is mainly to

demonstrate that selective excitation is maintained through scattering tissue.

83

4.4.3 Results and Discussion

4.4.3.1 Functional Imagi_ng

Imaging of the capillary tubes requires laser pulses that selectively excite HPTS in
acidic or alkaline solution. Nonselective excitation, as would be obtained from TL pulses,
cannot distinguish between the two solutions because HPTS in both environments has an
approximately equal integrated cross-section of excitation and fluoresces at 515 nm. The
spectral phase for TL pulses (a flat line) and the BP functions that were used to enhance
excitation of HPTS in an acidic or an alkaline environment were shown earlier in Figure
4.2. The SHG spectrum obtained from a thin nonlinear optical crystal (Figure 4.2, right)
provided a diagnostic test of the effects of phase modulation. SHG has the same square
dependence on the intensity of the laser pulses as two-photon excitation. Unlike the TL
pulses, the shaped pulses preferentially excite one of the two different pH solutions. The
complicated BP introduced in each of the shaped pulses involves several frequencies at
which the phase of the light is retarded by 7r, which is equivalent to 1.3 fs. These
seemingly small variations lead to the observed changes in the SHG spectrum. However,
the spectrum of the laser pulse does not change upon phase modulation. The potential to
cause two-photon excitation or generate second harmonic radiation at specific
wavelengths is, however, changed“ “

Photographs of the sample are shown in Figures 4.12A and B without and with
biological tissue, respectively. The capillaries containing HPTS in acidic buffer are
barely distinguishable from the alkaline HPTS solution in which they are immersed.

Two-photon images obtained with TL pulses without and with biological tissue are

displayed in Figures 4.12C and D, respectively. The walls of the capillary tubes (300 ,um)

84

appear black because fluorescent solution is excluded by the wall. This observation
confirms the confocal effect gained by two-photon excitation.” 13° Note the comparable

quality of imaging with and without tissue.

 

Figure 4.12. Images A and B are presented in color. A and B are photographs of the
sample obtained under white light illumination in the absence and presence of biological
tissue. In both of these photographs, the three capillaries in the sample are difficult to
distinguish. C and D are two—photon excitation images obtained by raster scanning the
sample without and with scattering tissue, respectively. The walls of the capillary tubes,
which look like vertical lines, are approximately 300 pm thick and are clearly visible in
both images. Notice that comparable image quality is obtained in the presence and
absence of biological tissue. Images C and D were obtained with TL pulses which are not
capable of discriminating between the two different pH solutions.

The full effect of coherent control when using BPS is shown in Figure 4.13. The
image obtained with phase BP06 enhances the signal from the capillaries (Figure 4.13A),
whereas the image obtained with phase BPIO enhances the signal from the surrounding
solution (Figure 4.13B). The functional image shown in Figure 4.13C was obtained by
dividing the data from Figure 4.13A by the data from Figure 4.13B. The image in Figure
4.13D corresponds to the functional image obtained in the presence of biological tissue.
The very high contrast between the three capillaries containing acidic solution and the

surrounding alkaline solution is accented by false coloring in the top half of the images

85

(Figures 4.13C and D). The contrast between the two solutions is very clear even when

the laser has transmitted through the tissue.

 

Figure 4.13. Coherent control is used here to obtain pH sensitive functional imaging.
Panels A and B show images obtained by two-photon excitation using the optimized
phases BPO6 and BPIO, respectively. Notice that the signal from the 1 mm capillary tubes
is enhanced for BP06 (A) and suppressed for BPIO (B). The top half of images C and D
are presented in color. Panel C shows a firnctional image highlighting the contrast
possible using coherent control. This image was obtained by taking the ratio of the data
obtained by BP06/BP10. Panel D shows the firnetional image obtained when a 1 mm
thick slice of biological tissue was placed in front of the sample. Notice that the presence
of the tissue reduces the overall signal to noise ratio, but the discrimination between
acidic and alkaline HPTS is conserved. The contrast in the functional images can be
firrther enhanced using false color (top half of Figures C and D). Higher values are
shaded red, while lower values are shaded blue.

4.4.3.2 Loss of Coherence as a Function of Scattering.

The striking functional images shown in Figure 4.13 raise questions regarding the
propagation of shaped and unshaped pulses through scattering media. The experiments
described below are aimed at establishing the general physics responsible for the
observation and at estimating the limits to firture applications.

First, the loss of ballistic photons was measured as a function of scattering. For

this experiment, the microscope objective focused the beam onto a 25-,um diameter

86

pinhole after transmission through a variable scattering medium consisting of a solution
of skim milk in water in a 1-cm path-length cell. A 20% increase in the concentration of
milk decreased the intensity of the transmitted light by 1 order of magnitude. The light
emerging from the pinhole was measured by using a photomultiplier and a lock-in
amplifier. The loss of unscattered photons from TL pulses was compared with that from
shaped pulses. For this experiment, a phase mask that did not favor two-photon excitation
at a particular wavelength was used to demonstrate that there is nothing unique about the
rate of coherence loss for shaped or unshaped pulses. The results from these
measurements are plotted in Figure 4.14. The scattering of photons from both TL (filled
squares) and phase-shaped (filled circles) pulses was found to increase exponentially as
the percentage of skim milk in the solution was increased.

Phase shaping does not affect the rate of decoherence, which is caused by the
spatial inhomogeneities in the scattering medium. The loss of photons arises from their
random collision with particles in the scattering medium, and phase shaping does not
alter the scattering probability. Because the scattering of photons is caused by spatial
inhomogeneities in the sample, it cannot be corrected or prevented by spectral phase
shaping. Replacing the pinhole with an SHG crystal (20 ,um ,B-BaBZO4 type 1 crystal)
confirmed that only the ballistic photons participate in generating second harmonic
light.144 The SHG spectra of the TL and shaped pulses are shown in Figure 4.14 Lower
Inset. Note that the shaped pulses generate a weak, evenly distributed spectrum. The
intensity dependence of the integrated second harmonic light from TL and shaped pulses
was found to decay exponentially with a slope that is a factor of 2 greater than that

obtained for the fundamental wavelength. This result is consistent with the expected

87

dependence for two-photon response of Isno = Inexp(-2pex). These measurements, which
range over more than 5 orders of magnitude in intensity, establish that only the ballistic
photons are coherent and are responsible for SHG, that they are lost exponentially, and
that spectral phase shaping does not affect this firndamental property. The data in Figure
4.14 show that it is possible to transmit ballistic photons with spectral phase information
to depths at which SHG intensity has dropped by 6 orders of magnitude. Additional
efforts in noise suppression would allow us to go even deeper, or to collect back-scattered

signal.

  

 

Intensity, a.u.

 
  

 

 

LAnAu

 

Intensity (a.u.)
a
2"

_L

I'D

L '
Anna-AI A A Anna:

 

  
 

/shaped \
3 ' 4'00 420
1E-63 . nm

I f I ' I ' I ' I ' I ' U

0 10 20 3O 40 50 60
Percent skim milk in water, 1cm path (%)

 

 

 

 

 

Figure 4.14. Loss of coherent photons as a function of scattering. The transmitted ballistic
photons from TL (squares) and shaped (circles) pulses were measured as a function of
scattering. The phase shape for these measurements was chosen to suppress SHG
generation evenly through the entire spectrum. The intensity of the ballistic photons
measured at 800 nm (solid markers) was found to decay exponentially for both TL and
shaped pulses. The intensity of the ballistic photons was also measured after SHG from a
thin SHG crystal (open markers). A factor of two higher scattering was recorded in this
case given the square dependence of SHG, confirming that only ballistic photons
generate the second harmonic' light. The spectrum and phase for the TL and shaped
pulses used in this measurement are shown in the upper inset. The spectra of the SHG
output from TL and shaped pulses are shown in the lower inset. Since both TL and
shaped pulses have similar slopes, they experience the same rate of scattering.

88

Figure 4.15 depicts the SHG spectrum generated with TL pulses and the spectrum
generated by using BP10 shaped pulses without a scattering medium and with a solution
of 15% of skim milk in water in a l-cm path length cell. Note that the frequency-doubled
spectrum of the shaped pulses is unaffected (except in overall intensity) by scattering.
There is a slight change that arises from the wavelength dependence of Rayleigh
scattering after four scattering lengths, which results in a slight attenuation at shorter
(~750 nm) compared with longer (~850 nrn) wavelengths. In addition to the retention of
phase information, the modulation of phase does not lead to a loss of optical resolution.
Figure 4.15 Inset, shows the image of 15 ,um diameter microspheres (FluoSpheresz450-
480, Molecular Probes). The fluorescent probes were deposited on a microscope slide,
covered with a l-mm-thick slice of biological tissue and topped off with a second

microscope slide.

 

TL

 
   

Shaped
0 Before scattering
. After scattering

  

 

 

 

 

380 400 420 440
Wavelength (nm)
Figure 4.15. SHG spectra generated for TL pulses and shaped pulses optimized for leO
without a scattering medium and after propagating through a 15% skim milk in water
solution (equivalent to 2.1 scattering lengths). All the spectra were normalized to the
intensity they had in the absence of scattering. Scattering does not change the observed
SHG spectrum. The inset image on the left corresponds to three 15 pm diameter
FluoSpheresTM behind a 1 mm thick slice of biological tissue. This image demonstrates

that high resolution (micron) can be achieved by the shaped pulses after they transmit
through biological tissue.

89

The sample was raster scanned in front of the laser, which impinged first on the
tissue and then on the spheres. The image, obtained by using BP10, showed no loss of
resolution when compared with that obtained by using TL pulses. The coherent control
method presented for functional imaging is less expensive and simpler to carry out than
current methods of multiphoton imaging. The pulse shaper compensates the pulses
automatically by using MIIPS and introduces the Optimum phases required for selective
excitation. This method should be contrasted to an effort in which two expensive 100-fs
pulsed lasers tuned to 800 and 900 nm and aligned to focus at the exact same spot would
be used to reproduce the observed results. Alternatively, the use of an expensive tunable
lOO-fs laser would require tuning, alignment, and compression after the objective —
procedures that are typically time-consuming and require a certain degree of expertise.
Therefore, the ultrashort laser oscillator with a pulse shaper provides a more powerful

and flexible platform.

4.4.5 Conclusion

The successful use of coherent control to demonstrate functional imaging through
a scattering medium (Figure 4.13) leads us to expect that selective two-photon excitation
and imaging could be achieved at greater depths of tissue (3—4 mm).'“' 145 By using
amplified laser pulses, measuring two-photon fluorescence through greater scattering
lengths (~4 mm) of tissue is possible. For medical applications, endoscopes can be used
to deliver the laser pulses to some of the internal organs, allowing therapeutic treatment
of internal organs. The maximum amount of laser light that can be used on living tissue is

regulated by safety concerns (American National Standards Institute Z163.l) and

90

ultimately limits the maximum depth that could be achieved by the method described
here. Three-photon-induced laser damage, which could result from the use of intense

ultrashort pulses)“ ”7

can be suppressed by orders of magnitude by using BPS, while
preserving the large two-photon excitation yield obtained by these pulses. This
suppression results from lengthening Of the pulse due to the phase structure and fi'om the

destructive interference for three-photon excitation pathways.“

The purpose of our
experiment was to determine whether coherent control methods based on spectral phase
modulation and MII could be used under typical scattering conditions encountered in
biological imaging to provide discrimination for laser-based imaging or therapeutic
prnposes. Given that an affirrnative answer to this question is found, possibilities afforded
by coherent control could be exploited. In the experiment presented here, functional
imaging (as shown here with pH, but also possible with Ca“, Na“, biomarkers, and drug
gradients) was demonstrated by optimizing the spectral phase of the laser pulses by using
a pulse shaper, without laser tuning or fluorescence filters. These advances could enhance
noninvasive two-photon imaging methodsm’ ”8 by providing contrast mechanisms based
on selective-probe excitation while minimizing the risk of damaging healthy tissue. Our

findings should inspire new possibilities for the use of shaped pulse lasers in the

biomedical field.

91

4.5 Minimizing Multiphoton Excitation Damage with Shaped Pulses

4.5.1 Introduction

Reduced photodamage is widely acknowledged as one of the main advantages of
two-photon excitation microscopy on biological specimens.92 However, for femtosecond
pulses, it has also been noted repeatedly that cells are prone to massive damage149
particularly when excitation intensity at the specimen plane is increased above 10 mW.‘46
When laser power is increased above this threshold, nonlinear photodamage sets in.
Nondestructive multiphoton imaging of living specimens is restricted to a certain optical
window contingent on the molecular multiphoton cross section and fluorescence quantum
yield of the fluorophore, the detector sensitivity, and the onset of cell damage. Although
phototherrnal damage appears unlikely in most cells in which water is the major near IR
absorber, destructive nonlinear photochemical effects and severe cell damage by optical
breakdown effects are nonetheless imminent. Since nonlinear excitation scales linearly
(for two-photon) or quadratically (for three-photon) with the inverse of the pulse
duration, the availability of picosecond and femtosecond lasers has accelerated the
development of multiphoton microscopy. Femtosecond lasers, especially, have been
promising excitation sources for nonlinear optical imaging because of their high peak
power and greater penetration ability when compared to lasers required for linear
excitation. However, the expectation that a factor of ten reduction in pulse duration
would lead to one order of magnitude greater signal for two-photon excitation was not
realized, and shorter pulses were actually Observed to cause greater laser induced

15

damage.1 Two-photon excitation of endogenous cellular absorbers can result in the

formation of destructive oxygen radicals and singlet oxygen and can lead to indirect

92

DNA damage.150 Three-photon absorption can result in DNA and protein excitation. A
major absorption band of nucleic acids is at 260 nm, which corresponds to three-photon
excitation from a Ti:S femtosecond laser, the most common excitation source in
multiphoton imaging. Excitation of this band results in direct DNA damage. Although
three-photon-absorption should not occur frequently considering the laser parameters
used in our experiments, the realization of its high damaging potential in biological
imaging, should instigate further developments in multiphoton microscopy in as far as
femtosecond irradiation is concerned.

Since selective excitation by phase modulation is a potentially viable tool for
multiphoton biomedical imaging, its usefulness can be finther expanded to the possibility
of suppressing damage brought about by three-photon excitation. Preserving the integrity
of biological specimens and ensuring that no damage is being caused by femtosecond

illumination is one of the aspects that need to be addressed in phase modulation.

4.5.2 Experimental Section

The efficacy of phase modulation for maximizing two-photon excitation while
minimizing three-photon transitions was demonstrated by imaging a 30x30 ,um section
of a slide containing the chromophores HPTS and L-tryptophan (L-trp) mixed with an
aqueous polyvinyl alcohol (4%) solution on glass cover slips. The laser (12 fs pulses
centered near 800 nm, 97 MHz, 250 mW average power) was brought in the rear of a
Nikon TE2000-U inverted microscope, and focused on the sample with a 60x/1.45 NA
objective. The emission was imaged onto a CCD camera (Andor, iXon DV887). To

generate the final image, the sample was raster scanned in 0.3 pm steps with a digital

93

piezo controller (PI, E-710.3CD). Sinusoidal phase functions were used for the selective

imaging in this experiment.

4.5.3 Results and Discussion

The role of three-photon absorption in photodamage seems to be particularly
relevant when the corresponding one-photon wavelength range for three-photon
absorption with our laser system is considered. Three-photon absorption can result in the
excitation of DNA and amino acids in proteins and this can lead to cell death. With phase

modulation, however, three-photon induced laser damage“: '47

can be suppressed by
several orders of magnitude. This suppression results from lengthening of the pulse due
to the phase structure and the destructive interference of three-photon excitation
pathways.“ In addition, the large two-photon excitation yield obtained by these pulses is
preserved.

Figure 4.16 shows how phase modulation can be used to achieve selective
suppression of three photon excitation while maintaining the efficiency for two-photon
excitation. In this case three-photon excitation from L-trp, an amino acid with absorbance
maximum at 280 nm and fluorescence at 350 nm is suppressed by almost 90%. The two-
photon dye (HPTS), on the other hand, maintains its intensity at the same phase where
three-photon excitation is decreased. The effect of phase modulation on these two
fluorescent molecules is clear when comparing them with the intensities obtained with

TL pulses. Although phase modulation slightly decreases two-photon fluorescence

intensity, its effect on three-photon excitation is significant.

94

Trp (shaped)

 

 

 

99
Am

9
N

  

 

 

 

 

Normalized intensity (a.u.)

 

 

 

 

.0
o

I. J I V l' 1’ r
5 1o 15 20 25 30 5 1o 15 20 25 30m

0

Figure 4.16. Suppression of three-photon excitation by phase modulation. Top:
fluorescence image of HPTS and L-trp deposited on a quartz cover slip and imaged with
TL 800 nm pulses. Middle: fluorescence from the same samples when laser pulses are
shaped with a sinusoidal phase function crossing zero at 0.25 7:. Bottom: Cross-section of
the fluorescence intensities Of both dyes highlights the effect of phase modulation in
minimizing three-photon excitation (right) while essentially maintaining intensity of two-
photon fluorescence (left). These intensities (normalized) are compared with those
Obtained using TL pulses. Each image size is 30 ,um (width) x 15 11m (height).

The results presented here were generated using pulses that were modulated by
sinusoidal phase functions. The greater contrast that is achievable using binary phase
modulation urges us to explore the possibility of yet enhancing the contrast achievable

with two- and three-photon excitation. The search for these Optimal phase functions

requires phases that would theoretically lengthen the excitation pulse and consequently

95

focus two-photon excitation on desired frequency regions while defocusing three-photon

excitation away from regions that can cause biological damage.138

4.5.4 Conclusion

Multiphoton fluorescence microscopy is becoming one of the methods of choice
for dynamic imaging of living cells and tissues. The technique is particularly useful in
biological systems where UV excitation would not otherwise be possible due to the light
transmission characteristics of optical systems. In addition, side effects such as
photobleaching and photodamage are minimized in multiphoton excitation, and occur
only in the immediate region surrounding the focal volume. Although these capabilities
have been advancing at a rapid pace, a significant amount of work, especially in the area
of completely eradicating damage caused by higher order fluorescence excitation, needs
to be completed.

With phase modulation of femtosecond laser excitation, the selective suppression
of three-photon processes can be achieved while maintaining the efficiency of two-
photon excitation. It is worth noting that it is possible to deliver the accurately shaped
pluses that are required for selective two-photon activation through thick scattering
biological tissue. Selective two-photon excitation and suppression of multiphoton
induced damage are two capabilities that will enhance the potential applications of two-

photon imaging for cancer detection and two-photon PDT.

96

CHAPTER 5

ISOMERIC IDENTIFICATION

Multiphoton ionization with a femtosecond laser exhibits a number of specific
features that make it a unique ionization technique for analytical and scientific mass
spectrometry (MS). Among these is the possibility of highly efficient selective ionization
of certain compounds out of complex mixtures while ionization of unwanted substances
is suppressed, a feature that can be used to distinguish between isomers. Other interesting
aspects include the ability to control the degree Of fragmentation even for large
molecules15 " 152. In this chapter, the femtosecond laser is used as the ionization source for
MS experiments involving isomeric compounds.

MS, in all its many variants, is an enormously important analytical tool which
makes possible the identification of the chemical compounds present in a sample, and in
many cases enables assurance of the absence of certain compounds in that sample. It has
developed rapidly from being a research tool used primarily in physical organic
chemistry to the point where biochemical and medical research applications not only
account for a high proportion of its usage but also direct much of the effort concerned
with technique development. Recent major developments have increased and continue to
expand its scope of application in medical research and other areas. These include the
effective use of high mass spectrometric sensitivity and the development of techniques
for the analysis of compounds of high mass and low volatility which were previously

impossible to analyze.

97

In most standard implementations of MS, fragmented mass spectra are typically
obtained from the energetic ionization process involving collision between neutral gas
molecules and free electrons that are typically accelerated to energies between 50 to 150
eV. For most organic compounds, electron ionization (El) has its highest cross section in
the region of 70 eV, where enough excess energy is transferred to the molecule to
generate fragmentation of the ion. Such fragmentation can be very useful in providing
structural information. For this reason, there exist extensive libraries of mass spectral data

at the standard ionization condition of 70-eV EI.153

There are, however, disadvantages to
this approach. First, the analysis can become very difficult in the case of complex
mixtures. Moreover, it is Often the case that different species of the same basic molecular
type yield practically identical fragmentation patterns. Many isomeric compounds yield
virtually indistinguishable mass spectra. In such a case, positive identification of an
isomeric molecule requires a technique by which one can measure the parent molecular
mass and distinguish the molecule from its isomers through its fragmentation pattern.
This is even more important when one wants to analyze a mixture of such compounds.
The premise of this study is that certain phase-shaped laser pulses could control
the ionization and fragmentation of isomeric molecules so that positive identification of
isomers could be accomplished. Shaped femtosecond laser pulses are explored as an
unconventional ionization technique to improve isomer specificity in MS. Qualitative

discrimination of structural and geometric isomeric compounds, including the potential

for quantitative analysis of mixtures of isomers are investigated in this chapter.

98

5.1 Isomerism

The mass spectra of certain isomeric compounds have barely enough information
that helps to ascertain the identity of such molecules. While close examination of certain
minor peaks in the spectrum could possibly help identify one isomer from the other,
small changes in instrumental parameters could mask these effects, and the value of this
practice with real samples is doubtful.

lsomers are molecules with the same molecular formula, but different

arrangements of atoms. There are different types of isomers, as shown in Figure 5.1.

 

 

Chain

WA

 

 

 

 

Structural

 

 

 

 

Isomerism

 

 

 

 

Positional

MB! *8!

 

 

 

 

Funcfional

\/'\g /I\—\-—OH

 

 

 

 

 

Stereo

Geometrical

\_//\/

 

 

 

 

 

Optical

OH OH
NH; NH;

 

 

 

 

Figure 5.1. Classification scheme of isomers (with examples).

5.1.1 Structural Isomerism
Structural isomerism occurs when two or more compounds have the same
molecular formula, but different structures, which tend to give the molecules different

chemical and physical properties. Structural isomerism can either be chain, positional or

functional.

99

i

The possibility of branching in carbon chains leads to chain isomerism. The
molecule butane, C4Hio, for instance, has two isomers - one where all carbon atoms lie in
a straight chain and the other where the chain is branched. In position isomerism, the
basic carbon skeleton remains unchanged, but important groups are moved around on that
skeleton. For example, the structural isomers of C3H7Br include one where the bromine
atom is on the end of the chain, and one where Br is attached to the middle carbon.
Position isomers also occur on benzene rings, as in ortho-, para-, or meta-nitrotoluene. In
firnctional group isomerism, the isomers contain different functional groups - that is, they
belong to different families of compounds. For example, the molecular formula C3H60
could be either propanal (an aldehyde) or propanone (a ketone). There are other
possibilities as well for this same molecular formula, including a carbon-carbon double

bond (an alkene) and an -OH group (an alcohol) in the same molecule (see Figure 5.1).

5.1.2 Stereoisomerism

Stereoisomerism occurs when the atoms in a molecule can have different
arrangements in space. The bond structure of stereoisomers is the same, but the
geometrical positioning of atoms and ftmctional groups in space differs. This class
includes optical isomerism where different isomers are mirror images of each other, and
geometric isomerism where functional groups at the end of a chain can be twisted in
different ways. Geometric isomerism or Cis-trans isomerism describes the orientation of
functional groups at the ends of a bond around which no rotation is possible. Such bonds
are typically double bonds, but they can also be part of a ring structure which prevents

rotation. Cis isomers and trans isomers of a substance have different physical properties.

100

 

 

Trans isomers generally have higher boiling points and higher densities. This is because
the molecules of trans isomers can line up and fit together better than the cis form.
Optical isomers of a given compound are often identical in all physical properties except
the direction in which they rotate light. The molecules of optical isomers are
asymmetrical and are mirror images of each other, as signified by L-alanine and D-
alanine (see Figure 5.1). Molecules like these are said to be chiral and the different forms
are called enantiomers. The concept of optical isomerism, which in itself constitutes a

field of chemistry (chiral chemistry), will be discussed further in the next chapter.

5.2 Isomeric Identification by Laser Control and MSg

5.2.1 Introduction

While it is true that several analytical techniques, including MS, are widely used
for molecular identification, the possibility of utilizing quantum-mechanical laser-matter
interactions to selectively discriminate molecules, in particular, isomers, could enhance
the utility of MS. This chapter introduces a coherent laser control technique that is
coupled with MS to distinguish between pairs Of isomers, which by standard MS alone,
are difficult to discern.

Knowledge of the exact isomeric configuration of a molecule can be important
since different isomers can have widely varying toxicological properties and
environmental impacts. Analysis of extremely toxic environmental pollutants requires
high sensitivity, typically in the parts-per-billion range. MS, being the most sensitive of

analytical methods, has been used for all kinds of chemical analyses, from detection of

 

3 Reproduced in part with permission fi'om J.M. Dela Cruz, V.V. Lozovoy, M. Dantus, “Isomer
identification by mass spectrometry with shaped femtosecond laser excitation,” J. Phys. Chem. A 109,
2005, 8447-8450. Copyright 2005 American Chemical Society.

101

I.

pollutants like dioxin154 to finding high-grade petroleum in sludge.15 5 It is also used to
detect steroids in the urine of athletes156 and contaminants in herbal supplements.
However, conventional MS alone cannot be used for isomer-specific analysis of certain
toxic chemicals. This method of ionization falters in showing reproducible differences
between isomers. The ability of spectroscopic analytical methods such as IR, Raman and
nuclear magnetic resonance spectroscopy to distinguish between pairs or sets of isomers
is indisputable. Nevertheless, the orders of magnitude lower detection limit of MS, with
the possibility for achieving single ion sensitivity, should give MS an added advantage
over these methods. Another significant asset of MS is its capability of handling complex
mixtures in tandem with GC.157’ '58

Although GC/MS is widely used for a variety of environmental, biological, and
other applications, the interpretation of the resulting data can be complex and time
consuming.159 While many laboratories have such instrumentation, its use for isomeric
identification is not trivial. Significant time and effort must be put into configuring the
instrument for analysis; standards must be prepared for each target compound; and the
GC temperature program must be optimized to provide efficient separation of
compounds. The response function must be determined from the analysis of the
standards. The entire process is prone to numerous pitfalls such as contamination,
carryover, and misinterpretation of the data.

Isomeric differentiation has been achieved via selective self-ion/molecule
reactions and tandem MS in ion trap mass spectrometers, but the whole process takes

160, 161

several minutes. Isomeric compounds of pentane have, for instance, been

distinguished by monitoring the products of their reactions with mass-selected ions

102

 

generated from the individual isomers in an ion-trap.162 Resonance enhanced
multiphoton ionization employing molecular cooling in a supersonic beam has also been
used to separate isomers, although, a tunable UV laser source and a special coupling of a
GC column and a supersonic beam valve are necessary.163 Chemical reactivity has been

coupled with MS for identification!“ '65

though the universality of this approach is
uncertain.

Femtosecond lasers have long been used as a source Of ions in MS; in most cases
these studies were aimed at measuring ultrafast photo-dissociation dynamical“ ‘67 The
energy deposited by the laser pulse on the molecules sets off a number of competing
processes, primarily ionization, fragmentation, and redistribution of the energy among the
different degrees of freedom in the molecule. It has recently been shown that the relative
yield of the resulting fragment ions can be controlled to some extent by shaping the
femtosecond laser pulses.32 Specific phase functions that determine the relative phase of
different frequency components in the pulse can be introduced using a pulse shaper.90 At
present, there is no theoretical model that provides a link between pulse shaping and its
effects on molecular ionization and fragmentation. While such a model is still to be
developed, empirical searches guided by computer learning algorithms, as suggested by
Judson and Rabitz,ll had been used to explore the extent of selectivity.” 167’ I“

The Dantus group has developed a method based on a 15-minute (typically)
exhaustive evaluation of an entire subset of phase functions which introduces a varying
degree of M1165 ’ “ occurring among excitation pathways during laser-induced ionization

and fragmentation. To facilitate the search and increase its efficiency, BP functions are

used. Only a small number of frequency regions (pixels) are modified in this case. In fact,

103

—-

as few as 8-bit BP functions have been shown to effect changes (two orders of
magnitude) on the relative yield of particular fragment ions. With pulse shaping, the
sensitivity of laser-molecule interactions to differences in electronic and molecular
structures of compounds can be amplified.169 This method of laser control, which is
coupled with the detection sensitivity of MS, can be used to identify isomers and quantify
their relative concentration in isomeric mixtures. The structural isomers (ortho and para)
of xylene, one of the high production volume chemicals produced in the United States,
were evaluated with this method. A quantitative assessment of the relative concentrations

of these isomers in mixtures also demonstrates the analytical ability of this method.

5.2.2 Experimental Section

Prior research from the Dantus group which involved femtosecond pulse shaping
and MS, established the multidimensional capability of femtosecond laser control in
identifying a number of chemical compounds based on the unique spectral fingerprints
generated from different BP-shaped pulses.169 In this chapter, sub—second and
reproducible isomer identification is achieved with the experimental setup shown in
Figure 5.2.

An amplified 800 ,uJ/pulse femtosecond laser (35 fs) is sent to a pulse shaper
programmed to measure and characterize the laser pulses and to correct unwanted phase
distortions on the laser pulses using the MIIPS method.” ”0 The correction of phase
distortions to generate TL pulses is critical for experimental precision and reproducibility

because the nonlinear interaction between the laser pulses and the molecules is very

104

sensitive to phase variations in the laser pulse. It is this nonlinear interaction that leads to

ionization and fragmentation.

 

 

 

 

  
 

Pulse shaper

 

   

 

 

‘ l \

'8 laser Binary shaped pulse

Figure 5.2. Schematic experimental setup for laser control MS for the identification of
isomers. The amplified femtosecond laser is sent to a pulse shaper that is programmed to
measure and compensate any phase distortions present on the laser pulse. BP firnctions
are introduced at the Fourier plane, where the spectrum of the pulse is fully dispersed.
The phase-shaped laser pulses are focused inside a low pressure gas chamber where they
interact with the sample molecules, causing fragmentation and ionization. A TOF mass
spectrometer is used to record the arrival of the ions according to their mass-to-charge
ratio. Some BP functions cause large enough differences in the fragmentation and
ionization patterns that can be used to identify isomers and even quantify their
concentration in a mixture.

BP functions are introduced at the Fourier plane of the pulse shaper, where the
spectrum of the pulse is fully dispersed. Each phase function contains ten groups of ten
pixels. A phase retardation equal to 0 or 7; is assigned to the ten corresponding regions in
the spectrum of the pulse. Because only the phase of the pulse is modulated, the energy
and the spectrum Of the excitation laser field stay unaltered throughout the experiment.
The phase-shaped laser pulses are then focused inside a low pressure gas chamber where

they interact with gaseous sample molecules causing ionization and fragmentation.

105

The experimental data were Obtained using a vacuum system (see Appendix B)
equipped with a time-of-flight (TOF) mass spectrometer. Several samples, which
included the ortho and para isomers of cresol and the cis/trans isomers of the alkenes 3-
heptene and 4-methyl-2-pentene (MP) (Sigma Aldrich) were tested. Preliminary test runs
were performed on these isomers before the main samples, ortho- and para-xylene, were
tested. These compounds were used without further purification. The molecules were
leaked into a high-vacuum chamber by a precision valve (Pressure ~ 8x10'6 Torr) and
interacted with the femtosecond laser beam in the focal volume. The chamber was
pumped by a 220 US turbomolecular vacuum pump where fast flow prevents the
accumulation of fragments. The voltages on the extractor and repeller plates were kept at
1600 and 2500 V, respectively. The ions were allowed to travel in a drift-free zone of
about 24 inches before being detected by a microchannel plate. Each mass spectrum
obtained represents the signal average resulting from 128 laser shots. The laser was
attenuated to 130 pJ/pulse at the TOF chamber entrance.

A total of 1024 different phases representing all possible lO-bit BP functions were
evaluated. The phases are referred to in the text as BPX, where X corresponds to the
decimal value of the binary number; i.e., the phase function 0000000000 corresponds to
BPO and 1101011010 corresponds to BP858, where 1 is replaced by 11 retardation at the
corresponding position in the spectrum. Some BP functions cause large enough
differences in the fragmentation and ionization patterns that can be used to identify
isomers and even quantify their concentration in a mixture. The mass spectral responses

of the xylene isomeric pair to a number of phase functions were observed. The BP

106

functions which provided either minimum or maximum ion ratio differences between the

isomers were further analyzed.

5.2.3 Results

1 show

For the pairs of isomers tested, the standard 70-eV EI mass spectra”
significant differences in terms of relative abundances of the molecular ion and its
fragments. Based on existing mass spectral libraries (NIST), these isomeric compounds
did show considerable (~10.5% to 30%) differences in certain ion yields. Among the
samples studied, xylene gave the smallest difference (<3% for all peaks), making it
virtually impossible to differentiate between its ortho and para isomers using standard
MS. Figure 5.3 shows the differences in relative abundances of each fragment ion
between the pairs of isomers studied. The plots shown were derived from standard NIST
datam and were obtained by subtracting the 70-eV EI mass spectrum of one compound
from that of its isomer. Before these difference spectra were obtained, all peaks in the
original mass spectra were normalized to the highest mass ion peak for each isomeric
sample. The statistical reliability of the mass spectral data used, specifically in terms of
the peak intensities, were not provided. However, mass accuracies of 5 ppm or less for

small organic molecules and 0.01% for large samples are typical for most MS

measurements.

107

 

 

trans-3-heptene - cis-3-heptene

 

trans-MP - cis-MP

 

 

 

 

  

   

 

02‘ max ~ 10.5% 0-2' max ~ 19.8%
t- c- .
0.1 - 0.1 - t- c-
o 1 I 1
o . .
r:
o l .
53 0-0- 0.0-
E 20 60 8 100 . 20 4O 60 80
0 i
-O.1e -0.1-
-O.2- -0.2-
-0.3- para-cresol - ortho-cresol 0.1 para-xylene - ortho-xylene
q‘ max ~ 26% max ~ 2.6%
-0.2
°‘ C“ 0.054
8 -
g -0.1 p- o-
?» 0.0i
E 0 0 20 6O 80 100
D ' l
m/z
. -0.054
.011.
.
-0.2- -0.1-

 

Figure 5.3. Difference mass spectra of the various sample isomeric pairs tested: 3-
heptene, 4-methyl-2pentene (MP), cresol, and xylene. The maximum differences
obtained from the standard mass spectra (NIST) of isomeric compounds, particularly
ortho and para-xylene, make differentiation by conventional MS difficult. Note the
difference observed in xylene (2.6%) compared to the rest of the compounds tested. The

structures of the compounds are also shown.

Since the error associated with the intensity of a single fragment ion peak can
sometimes be unpredictable, analysis of data obtained with the femtosecond laser was
focused on deciphering mass spectral differences of isomeric pairs of molecules based on

ratios of intensities between two fragment ion peaks. Such peaks were chosen on the

basis of maximum ion ratio differences obtained between the isomers.

108

 

 

a ortho-xylene e para-xylene tropyliumion(T"')

ere O

127 - - - -
100 -----
. 760 8790"" 840
0.4«
. ”4414—91

3432+

 

 

 

 

 

Ion counts
0
‘1’

 

 

 

 

 

 

 

 

 

 

 

 

0.2d
(LOW
20 40 60 80 100
T+
1'2: 202 ----- l' M+
1.0~ BP858A l
3 0.81 I ”It
15 i 100 ----- 1)
760 800 840
c .
2 0.4-
‘ 9
0.2- 1
0.0M
20 40 60 80 100
m/z

 

 

 

Figure 5.4. Laser control mass spectra obtained for ortho- and para-xylene. The top panel
shows results obtained with BPO corresponding to TL laser pulses; the flat phase across
the spectrum of the laser pulse is shown in the inset. The bottom panel shows data
obtained with laser pulses with phase BP85 8, which is shown together with the spectrum
of the pulse in the bottom inset. The mass spectra were normalized so that the molecular
ion (M) intensity equals unity. In the top and bottom panels, the region for m/z near 91,
corresponding to the tropylium ion (T), was amplified to highlight the observed
differences. For the BP858 shaped pulses, the relative yield of T+ for p-xylene is 2 times
greater than that observed for o-xylene.

109

The mass spectra obtained by femtosecond laser phase control for the xylene
isomers were analyzed based on the ratio between the peak value obtained for the
molecular ion (M+, m/z=106) and the tropylium ion172 (T, m/z=91). All peaks were
normalized to the molecular ion peak. Figure 5.4 shows that the application Of a complex
BP-shaped pulse resembling BP858 results in a greater relative yield of the tropylium ion
for para-xylene compared to that of ortho-xylene. The differences observed for the other
major fragment ions were not as significant, and were not further analyzed.

Figure 5.5 shows the statistical distribution of the ratios between the molecular
ion and the tropylium ion Obtained for each isomer, highlighting the contrast achieved
using binary shaped pulses. The ratios are represented by histograms of 100 independent
measurements for each sample and for each phase. For TL pulses, a difference between
the two isomers is clear. However, the difference observed for BP858 is much greater.

The ratio between ionic fragments at m/z 106 and 91 (106/91) for ortho-xylene
(11.21 i 0.40) is about a factor of 2 greater than para-xylene (5.62 i 0.17). When
compared to TL pulses, the shaped pulse enhanced fragmentation leading to a lower
M+/T+ ratio for para-xylene. Interestingly, the Opposite trend was observed for ortho-

xylene. TL pulses also gave relatively greater ion ratios for ortho-xylene (10.53 i 0.17)

than for para-xylene (8.18 i 0.14), but the difference between the isomeric ratios is less.

110

 

 

 

 

 

 

 

 

 

 

 

 

. 1-0 BPO

50 J 0000000000

40 - $0.5 Ho-xylene
E .15
8
0 30‘ o o .
g . '“lime(rs)“°
g 20+
0) .

p-xylene r _
10‘
4 6 T a 1'0 f2
1.
50-
3 1101011010
. ,7, o-xylene

40- 5“ r
...a C '1
c ‘ ._
8 __._AAu“-Ae___
o 30- 0.0‘ ‘ .
2 i 5mlim e(fs) 50°
3
m 20-
(o . \

10‘ § p—xylene

 

 

 

 

 

 

 

 

ii 10 12
Ion ratio (m/z 106/91)

Figure 5.5. Histograms showing the statistical distribution of the experimentally
measured ratio (MV'TI) obtained for 100 independent measurements each for o- and for
p-xylene. The top graphic shows the histograms obtained using laser pulses with phase
BPO; the bottom graphic shows the histograms obtained using laser pulses with phase
BP858. The insets show the temporal profiles of the two different laser pulses. The
histograms obtained for BP858 show a greater (~2x) molecular-ion-to-tropylium-ion
ratio. Note that in most cases the distribution of measured values is compact, indicating

good reproducibility of the results. The Observed differences can be used for fast and
reliable identification between these two isomers.

The quantitative performance of our method was assessed by the analysis of six
different mixtures of the two xylene isomers. The mixtures were probed by the phase-
shaped pulses, and the ratio of the molecular ion to tropylium ion peaks was recorded.

The average of two independent sets of measurements plotted as a function of

concentration are shown in Figure 5.6. The linear behavior of the data agrees with the

111

analytical formula (M*‘/T+),,.,,,.,..i,.d = [(M*/T+),xp,..m,m. — (M*rr*)o] / [(M"/r*)1 —
(M+/T+)o], where (M+/T*)cxp¢,imenm is the measured peak ratio for each sample mixture.
(M+/T+)o is the peak ratio measured for a pure para-xylene sample, and (MIT) is the
measured peak ratio for pure ortho-xylene. The normalized molecular ion to tropylium

ratios directly provide the relative concentration of the ortho isomer in the mixture.

 

A104 Brunt ¥
El; . orun2 ,”
E08] @,
306‘ 1” ‘
.9 04‘ é
<3 ' ,1?
+ ,’
t02 ,
+ ’1
E ,
00 §

 

 

 

0:0 - 01.2 ' 0:4 . 0:6 . 0:8 - 1T0

Relative concentration o-xylene
Figure 5.6. Quantitative concentration measurements of isomeric mixtures using laser
control MS. Each data point represents an average of 100 measurements (0.1 8 total)
obtained for each of six different mixtures. The horizontal axis corresponds to the known
concentration of each of the mixtures. The vertical axis corresponds to the normalized
MJ'/T+ ratio, which increases linearly with the relative concentration of o-xylene. This
linear relationship provides a quantitative method to determine the concentration of each
of the isomers (o- and p-xylene) in a mixture. Two sets of data obtained in two different
days are shown to illustrate the robustness of this method.

The maximum peak ratios obtained for the isomeric pairs of 3-heptene, cresol,
xylene, and MP are shown in Figure 5.7. The reported ratios were derived from selected
major fragment-ion peaks observed for each isomeric pair, thus ensuring statistically

acceptable standard deviations. Although minor peaks provided sizeable ion ratios, their

usefulness for quantitative work is diminished by their greater susceptibility to noise.

112

 

 

 

1.2 - .9 00- O t-3-heptene
o .,.- A=98, 8:27
3:: 1.0-
13 "s A A0- ‘ p-cresol
g 0.8‘ 0.; '_.° A=90, B=39
TD 06- g" op- o o-xylene
: A=91, B=106
< 0.4- o‘ ,."

,. Vt- v c-MP
0.21 A ~.w._.,..-.-..,.,.-’ A=15, 8:84
V“
0.0 -

 

Figure 5.7. Ion peak ratios (A+/B+) obtained for several isomeric pairs using binary
shaped pulses. The legend indicates the m/z values (A and B) of the fragments chosen.
The maximum ratio difference observed for xylene in particular, as seen in the
figure above, ranged from 0.09 for the ortho isomer to about 0.20 for para-xylene. This
corresponds to a difference of roughly 50%, which, when compared to the NIST
difference (~3-4%) (see Figure 5.3), makes an amazingly straightforward distinction
between the two isomers. Although the rest of the isomers already showed disparities in
their NIST spectra, these differences were enhanced several times more with shaped

femtosecond laser irradiation.

5.2.3 Discussion

5.2.3.1 Photoionization and fragmentation of xylene

As a consequence of multiphoton absorption, the fragmentation of molecules can
be effected by two distinct mechanisms- the process of dissociation followed by
ionization (DI, ladder switching) or ionization followed by dissociation (ID, ladder
climbing).'73’ ‘74 Based on the mass spectra obtained and because of the short pulse
duration of the laserm, the xylene molecules fragment by the ID route: the molecules

absorb a number of photons to reach a dissociative state below the ionization level. If the

113

laser pulse length is longer than the lifetime of the state, then the molecular fragments
form neutral moieties. Depending on the laser intensity, these fragments may absorb
additional photons within the initial laser pulse to ionize or further fragment. For xylene
and most organic compounds, ionization energies range from 8 to 9 eV. The absorption
of six photons of energy hv of 1.55 eV each should be sufficient to induce ionization. As

7

shown by measurements of photoelectron energies,l 5 molecules ionize as soon as the

sum of absorbed photon energies exceeds the lowest ionization potential. Subsequent
absorption by the ion results in extensive fragmentation.l76'179

Structure evaluation of xylene using laser control mass spectrometry is
anticipated, owing to the fact that an appreciable population of the molecular ion survived
to give an indication of the molecular weight of the compound. Parent ion dominance
(electron is predominantly emitted from the phenyl ring) and limited dissociation in the
mass spectra of the xylene isomers are attributed to the slow dissociative state of the
compound, highlighting its thermodynamic stability. Even the tropylium ion, which in El
MS is the most abundant ion (even greater than the molecular ion itself), was particularly
low in yield. Fragment ions with m/z <91 also produced rather low yields. The ionization
process is presumably completed within the duration of the laser pulse, leaving very little
time for the molecule to fragment to any significant extent. An observed power
dependence of the ion peaks suggests that molecular fragmentation of xylene follows the
ID mechanism, as suggested by Kosmidis et. al. in their study of the nitrotoluene

isomers. ‘80

114

5.2.3.2 Laser-based isomer recognition

It is interesting to understand the mechanism responsible for the observed laser-
based recognition. The nonlinear optical polarizability of molecules is highly dependent
on molecular structure.181 Strong laser fields have been found to cause charge buildup at
the ends of extended conjugated systems resulting in ionization by tunneling or barrier
suppression.182 A greater nonlinear optical polarizability increases a molecule’s
susceptibility to undergo fragmentation. This is consistent with our results for xylene, and
with the slightly lower ionization energy of para-xylene (8.44 i 0.05 eV) compared to
that of ortho-xylene (8.56 i 0.04 eV). This observation was also evident in the other
isomeric samples, where the structures of trans-3-heptene and trans-MP make them more
polarizable than their isomeric partners. Consequently, the yield of most of their ion
peaks is relatively greater than those of their isomeric counterparts.

The close correlation between SHG intensity and peak ion intensity gives the
impression that the effects observed are intensity dependent. Figure 5.8A shows a plot of
SHG intensity vs. BF function. The phases were sorted according to decreasing SHG
intensity, highlighting the exponential decay of intensity with complexity of phase
function. Figures 5.8 B and C show a corresponding relationship between ion peak
intensities (m/z 91 and 106) and the sorted phase functions. Regardless of the suggested

similarities, differences between ortho- and para-xylene were still apparent.

115

 

 

 

 

   

 

 

 

A 20
g 16-
(I)
g ' - ortho-xylene
g 12' o para-xylene
% -
w 8‘
E _
.2 4-
0.
6 260 400 600 800 1000
Binary phase (sorted)
B 7
0.7-
m/z 91 5.
0.6- ,
05 ‘ 5‘
~ ‘ 3
80.4 9
E03 5 3‘
L502 2 2'
0.1 1-
0 200 400 6&1 800 1000 V 0 200 460 600 800 1000
Binary phase (sorted) Binary phase (sorted)

 

 

 

Figure 5.8. Correlation between SHG or ion intensity and complexity of BP function. A.
General exponential decay of SHG intensity as phases are sorted based on the average
SHG intensities recorded during data acquisition of ortho- (black squares) and para-
xylene (open circles). B. Relation between ion intensity of mass peak at m/z 91 and
sorted BP function. C. Relation between ion intensity of mass peak at m/z 106 and sorted
BP function.

The temporal profiles for the different laser pulses used here are quite diverse, as
represented by TL and BP 858 laser pulses. (see insets of Figure 5.5). The observed
outcome of the reaction could be attributed to the time-frequency structure of the pulse
Although the total ion yield was about 6 times lower for shaped pulses than for TL

pulses, the differences observed between the isomers appear to have increased with pulse

shaping.

116

Figure 5.9 indicates how ortho- and para-xylene can be distinguished based on the
phase structure of the laser pulse. The ion ratios (m/z 91/106) for the isomers are clearly
different when all 1024 phases are considered. The complexity of the structures of the
BP functions that provided maximum contrast between isomeric pairs may have a
temporal effect on the responses of the molecules. When the sorted phase functions are
grouped to provide a statistical indication of peak intensity vs. phase complexity, the ion
ratios obtained clearly distinguish one isomer from the other (Figure 5.9, right). Although
a direct link between the BP modulation of the pulse and the selective molecular response
has yet to be established, the observed differences allowed us to devise a quantitative

measurement of the relative concentration in a mixture of these isomers.

 

 

 

 

 

 

—I— ortho-xylene

0.13- 0 Para-xylene
a . a 0.12 -
g g 011-
9 . v 0.10-
.9 .9
9 9 0.09-
.5 . .5 o.oe<

0.07«

0 260 460 600 800 1000 1 i 1 1'5 0 1'0
binary phase phase group

 

 

Figure 5.9. Left: Correlation of ion ratio intensity (m/z 91/106) and sorted phase function
for ortho- and para-xylene. Ortho-xylene is clearly distinguishable form para-xylene. The
difference is observable for most of the phase functions used. Right: Grouping the sorted
phases according to complexity shows maximum contrast (m/z 91/106) between the
isomers for more complex phase structures. Phase group 1 consists of 2 phases, group 2
contains 4, group 3 has 8, group 4 has 16, group 5 has 32, group 6 has 64, group 7 has
128, group 8 has 256, group 9 has 256, and group 10 consists of 258 BP functions. The
phases were sorted based on their effect on SHG intensity.

117

One of the major objectives of our work is the extension of MS to problems in
biomedical chemistry. A future research effort is directed towards developing laser
control methods for elucidating the structure of peptides and proteins. In metabolomics,
the coupling of the high sensitivity of femtosecond laser control with the high resolution
of MS could provide a method for determining a rough estimate of the number of
metabolites present. MS could offer a valuable first indication of the identities of
molecules, while laser control contends with its inability to separate molecules of the
same molecular mass. Even as emerging MS technologies continue to cope with the
expanding demand for applications in a wide range of areas, the complexity associated
with these techniques is a drawback. Controlled fragmentation with femtosecbnd laser
pulses is a viable alternative to most existing hyphenated (e. g. GC/MS, LC/MS, IMS/MS)
and tandem MS methods, in terms of its ability to rapidly discriminate between structural

isomeric compounds.

5.2.5 Conclusion

The principle of molecular control by pulse shaping with BP fimctions expands
the capabilities of conventional MS to include fast, accurate, and reproducible qualitative
and quantitative isomeric identification. This approach provides a new dimension in MS
that is highly sensitive to molecular structure, as required by current biological and
environmental applications. While there are a number of technological platforms
available for chemical, including isomeric identification, none are perfect, with each
method providing different merits. Although the perceived notion is that coherent control

using femtosecond lasers is too complicated and too expensive to be useful to biologists

118

 

 

and environmental scientists, awareness of its potential capabilities for analytical work
and its eventual applications in the biomedical field substantiate further investigation of
this technology. The accessibility of one-box “no-expert-required” femtosecond lasers
and the imminent development of laser systems with integrated pulse shapers should
further reinforce the usefulness of laser control in MS. The question remains as to
whether shaped femtosecond laser technology could address isomeric identification and
quantification issues that cannot be answered by any other existing analytical method.
Perhaps the next real challenge is whether this technology could pave the way for the

resolution of optical isomers through laser control.

119

 

CHAPTER 6

CHIRAL PULSES AND CHIRAL RECOGNITION

The idea that something seemingly so simple such as optical isomerism can have
such far-reaching ramifications is amazing. To think that the simple rotation around a
carbon bond could determine whether or not something was going to be recognized, the
affinity with which it would be bound, and the efficiency of the attached complex is truly
astounding. With optical isomerism, there is no difference in connectivity. The isomerism
has to do with the arrangement of the atoms in space. It arises from the handedness
property (chirality) of molecules through the presence of a chiral center. Optical isomers
are non-superimposable mirror images of each other; a set of optical isomers are called
enantiomers.

The recognition of chiral compounds is a fundamental phenomenon that is
observed in all biological systems, ranging from single cell organisms to the most
complex of animals. In nature, only the L form (left-handed) of amino acids are used to
make peptides or proteins, and only the D form (right-handed) of sugars exist, but there
are no such constraints in the laboratory. The specificity and efficacy of many
biologically important reactions are based on chiral interactions. The biological activity
of chiral substances often depends upon their stereochemistry, since the living body is a
highly chiral environment. A large percentage of commercial and investigational
pharmaceutical compounds are enantiomers, and many of them show significant
enantioselective differences in their pharmacokinetics and pharmacodynarnics. The

importance of chirality of drugs has been increasingly recognized, and the consequences

120

of using them as racemates (equal amounts of left- and right- handed stereoisomers of a
chiral molecule) or as pure enantiomers have been frequently discussed in the
pharmaceutical literature during recent years. With increasing evidence of problems
related to stereoselectivity in drug action, enantioselective analysis requires intensive
research. Producers of pharmaceuticals are required to evaluate the effects of individual
enantiomers and to verify the enantiomeric purity of chiral drugs that are produced. The
pharmaceutical properties and toxicity must be established independently for both
enantiomers even if the drug is to be marketed as a single enantiomer. As a result, chiral
recognition has been an area of intense interest, especially in the past decade. Significant
progress has been made in understanding chiral recognition from both fimdamental and
application oriented perspectives, yet a complete understanding of chiral recognition has
yet to materialize.

Most of the pharmaceutical and pharmacological studies of stereoselectivity of
chiral drugs before the mid-eighties involved pre-column derivatization of the
enantiomers with chiral reagents, forming diastereomers. The diastereomers were
subsequently separated in the normal or reversed phase mode of chromatography.

Great efforts have been devoted to the deve10pment of better methodology for
enantioselective chromatography during the past decade, and have resulted in new chiral

stationary phases, pioneered by Pirkle183

. Chiral agents were derivatized and immobilized
on the surface of the support (silica gel mostly), and served as the in situ chiral
discriminators during the chromatographic process. The preference of chiral stationary

phases lies in the inherent advantages of any chromatographic separation, such as the

speed of the analysis, the possibility to analyze or purify the enantiomers in complex

121

mixtures, the reproducibility of the analysis and its flexibility. Moreover, analytical
chromatographic systems can be adapted to preparative separations, in which pure
enantiomers can be collected.

In addition to their distinct practical applicability, chiral stationary phases can
uniquely contribute to studies of the nature of molecular recognition. Since the
differential retention of enantiomers in the chromatographic system employing chiral
stationary phases can be attributed only to chiral discrimination by the chiral sites, these
interactions can be isolated and explored. It has been shown that chromatographic
parameters obtained by chiral stationary phases can be sensitive to very subtle differences
between the enantiomers. Chiral stationary phases can be tailor-made to accommodate
specific studies of chiral recognition between molecules. A review by Taylor and
Maher184 describes in detail the principles underlying chiral discrimination by the various
chiral chromatographic systems, utilizing chiral agents either in the mobile or the
stationary phase. Another review by Gubitz185 describes the application of chiral
stationary phases to chiral drugs, emphasizing the main principles of chiral discrimination
of the various categories of stationary phases known so far.

In addition to its use in the analysis and detection of illicit drugs, the versatility
and simplicity of capillary electrophoresis is leading to its increasing application in the
pharmaceutical industry.186’ '87 Of particular relevance is the ability to develop methods
which are capable of resolving optical isomers. In some respects the limiting factor is
more on the detection side of the technique and the advances in capillary electrophoresis
over the next few years will probably be in this area. The range of detection techniques

commonly employed at present, UV absorption, fluorescence, electrochemical detection

122

 

and MS, is being extended and expanded by the application of such techniques as LIF
and luminescence-energy transfer using lanthanide ions.188

It is well-established that in order to obtain a chiral result from any reaction, it has
to contain a chiral component or it has to be subjected to some external chiral influence.
A well-known example of this principle is the use of circularly polarized light (CPL) as a
chiral agent in certain photochemical reactions, where the handedness of the light
determines the handedness of the reaction product. Enantiomers were first distinguished
by their optical activity. This observation can be traced back to the early (circa 1815)
observations of Biot regarding the ability of optically active samples to rotate the plane
of polarization of light. Louis Pasteur deduced in 1848 that the handedness of molecular
structure is responsible for optical activity. There are two optical manifestations of
molecular chirality that can be studied using modern-day polarimetric instruments. The
first is optical rotatory dispersion (ORD), which is the differential retardation of right-
circularly (RCPL) versus left-circularly polarized light (LCPL) as it traverses a medium.
The other, circular dichroism (CD), is the differential attenuation (absorbance) of RCPL
versus LCPL. CD measurements are simpler because they require only the detection of
intensity variations. ORD manifests itself primarily far off resonance from any absorption
bands while CD dominates near an absorption band. To be CD active, a molecule must be
structurally asymmetric and exhibit absorbance. Increased relative absorption of LCPL
results in a positive CD signal, while a negative signal is the result of RCPL being more
highly absorbed. The most commonly studied molecules are proteins where CD signals
are sensitive to protein secondary and tertiary structure. CD is measured with a CD

spectropolarimeter, which is able to measure accurately in the far UV at wavelengths

123

 

 

down to 190-170nm. The difference in the left and right handed absorbance is very small
(usually in the range of 0.0001 absorbance units). Although CD can monitor quite
accurately relative changes due to influence of environment on sample (pH, denaturants,
temperature etc.) and is non-destructive, the amount of sample necessary for analysis
(200pL of 0.5 mg/mL solution in standard cells) is a limitation of the method in terms of
its sensitivity. Another consideration is that oxygen absorbs strongly below about 200
run, so very extensive purging with pure (oxygen-free) nitrogen (>16 L/min) is necessary
for these measurements. The requirement for very dilute, non-absorbing buffers to allow
measurements below 200 nm is an added drawback to this procedure. As a consequence
of these limitations, enantiomeric resolution with the use of chiral light still requires
extensive exploration.

The use of CPL for the preferential production of a chiral compound was
proposed in the 19th century by Le Bel and by van’t Hofflsg’ '90. This idea was confirmed
in the early 1900’s, initiating the field of asymmetric photochemistry, which has been a
subject of extensive reviewsm' '92 Photochemistry with CPU” ‘94 led to the partial
destruction or the optical activation195 of racemic mixtures as well as asymmetric
syntheses.196’ '97 Our vision of enantiomeric resolution can be considered in the context of
asymmetric photodestruction, as reported by Kagan,198 who achieved 20% optical purity
by carrying out photolysis of racemic camphor to 99% completion.

The sections that follow describe the concept of polarization control to produce
chiral pulses and our attempts to generate such pulses with the intention of triggering

differences in the photoionization and fragmentation of enantiomeric molecules.

124

6.1 Polarization Control

The electric field vector of a plane wave propagating in the positive 2 direction
may be resolved into two orthogonal components, Ex and By. Depending on the relative
amplitude and phase (05) of Ex and By, the plane wave may be linearly, circularly or
elliptically polarized. The magnitudes and phases of field components can be described

by the equation,

E = [xEx exp(i¢x) + yEy exp(i¢y )]exp[i(kz — wt)] 6.1

When Ex and By are in phase, their strengths are always equal or related by a
constant ratio, so the direction of the electric vector will always fall on a single line in the
plane. This is linear polarization. The direction of this line will depend on the relative
amplitude of the two components. This direction can be in any angle in the plane, but the
direction never varies. When the two orthogonal components have exactly the same
amplitude and are exactly 90° out of phase, the light generated is circularly polarized. In
this case one component is zero when the other component is at maximum or minimum
amplitude. The direction of rotation is dependent on which of the two phase relationships

exists. Figures 6.1 and 6.2 depict linear and circular polarization.

125

 

 

 

 

\
\-..\.....
.L
\
\.__\.-..)

 

I

,' , E I," ’
W 7\ A /\ 0'
,' \/ \J \/ 1V" ,'

 

 

I
I
Iv

 

Figure 6.1. Linear polarization. If the vector of the electric field (measured at a fixed
point of space) oscillates along a straight line then the waves are called plane polarized or
linearly polarized waves. Here, a wave is polarized in a horizontal (lefi) or vertical plane

(right).

 

 

 

 

 

 

 

Figure 6.2. Circular polarization. The superposition of two waves that have the same
amplitude and wavelength and are polarized in two perpendicular planes with a phase
difference of 90° between them generates circularly polarized light. A phase difference of
90° means that when one wave is at its peak, the other one is just crossing the zero line.
At any fixed point in space that is in the line of propagation of this wave, the electric field
vector rotates in a circle while its length remains constant. In this cartoon, the electric
field vector rotates counter-clockwise when looked at from the direction of propagation.
The light generated in this case is left circularly polarized. Clockwise rotation of the field
vector generates right circularly polarized light.

When the two components are not in phase and either do not have the same
amplitude and/or are not 900 out of phase are elliptically polarized because the sum

electric vector in the plane will trace out an ellipse.

126

The generation of circularly polarized femtosecond pulses at the sample position
called for tests on two different techniques for delivering polarized laser pulses. The first
technique employs the conventional method of using a quarter waveplate (QWP) to
convert linearly polarized light (LPL) to CPL. The second technique takes advantage of
the SLM to introduce phase fimctions that could produce either LPL (horizontal or

vertical) or CPL (LCPL or RCPL).

6.1.1 QWP Technique

When light propagates through a medium such as glass its speed is reduced by the
ratio v/c = l/n, where c is the speed of light in vacuum and n is the index of refraction of
the medium. In uniaxial birefringent materials such as crystal quartz, there are two
indices no and no ("ordinary" and "extraordinary"). no applies to light polarized
perpendicular to the optic axis of the crystal while nc applies to light polarized along the
optic axis. These indices are a function of wavelength and temperature. no and ne differ
by an amount sufficient to produce phase shifts between polarization components, as
utilized in waveplates. Passing plane polarized light ( or LPL) through a birefringent plate
(in the z-direction) splits the light into two plane polarized beams oscillating along
different axes (fast along x, slow along y). When one of these beams is retarded by 90°,
the two beams, which become 90° out of phase are added together, and result in CPL in
one direction. By inverting the two axes such that the alternate beam is retarded, CPL in
the other direction is generated. A QWP (see Figure 6.3) consists of a carefully adjusted
thickness of a birefringent material such that the light associated with the larger index of

refraction (no, slow axis) is retarded by a quarter wavelength with respect to that

127

associated with the smaller index (no, fast axis). The material is cut so that the optic axis
is parallel to the front and back plates of the plate. Adjusting the plane of the incident
light so that it makes a 45° angle with the optic axis gives equal amplitude to waves
propagating in the fast and slow axes. Any LPL which strikes the plate will consequently

produce CPL.

 

  

input polarization plane

 
   

circularly polarized output

 

 

crystalline optic axis direction

 

Figure 6.3. Polarization conversion by a quarter waveplate.

To maintain linear polarization with a QWP in place, the slow axis is oriented at a
zero degree angle with the optic axis. This condition generates horizontal LPL. When the
slow axis is oriented at 900 relative to the optic axis, the light produced is vertical LPL.

Defects in a QWP can alter the intended state of polarization of light. Small
fabrication errors can lead to some deviation from a quarter wave retardance. Polarization
properties of this optical element can vary with the angle of incidence, temperature, or
with the wavelength of light. The narrow free spectral range of QWPs restricts their
operational usage. Misalignment of this polarization element would also lead to defects in

the polarization condition of the emerging light.

128

 

 

 

6.1.1.1 Experimental Validation of Polarization Condition

In order to test the polarization condition of a laser beam going through a QWP,
measurements of the intensity of the fundamental laser wavelength were taken. A QWP
was mounted a few inches from the last component of the pulse shaper. A polarizer was
then placed after the QWP. A spectrometer was used to measure the laser intensity. Two
different phase functions were introduced to the SLM — one where all phase values (128
pixels) corresponded to zero (designated as BP-0), and the other where all phase values
were equal to n (designated as BP-rt). In our femtosecond laser system, the input to the
pulse shaper has a stable, well-characterized linear polarization. Light that enters the
SLM is horizontally polarized. The output polarizer of the SLM was purposely removed
with the intention of synthesizing shaped pulses with frequency-dependent polarization.
The introduction of a flat phase function (BP-O) to the SLM allows only horizontal LPL
to go through and exit the pulse shaper. On the other hand, when BP-‘lt is etched onto the
SLM, only the vertical component of light is allowed to leave the shaper.

Figure 6.4 illustrates polar plots of the measurements made when the polarizer
was rotated in increments of 20 degrees. When the QWP was oriented at 0°, the results
were as expected for LPL (left). BP-0 delivered horizontal LPL and BP-n generated
vertical LPL. The imperfection in the measurements obtained in terms of non-
symmetrical intensities may be attributed to the not nearly perfect alignment of the optics
leading to the detector. The results for CPL were a cause for concern. When the QWP
was rotated to 45°, the light generated was clearly not circularly polarized (Figure 6.4,

right). Although BP-O, which contributes to the horizontal component of the light,

129

seemed to produce elliptical polarization, the contribution made by the vertical
component (BF-1t) drastically affected the polarization condition of the outgoing light.
Moreover, the output spectral polarization was distorted by elements with spectrally
dependent transmission or reflection, such as the diffraction grating component of the
pulse shaper or the dielectric mirrors in the setup. The effect of these elements on
polarization cannot be easily compensated for by the use of standard polarization

compensators.

 

1.01

 

 

 

L . 1 r1

Normalized intensity
0
O

 

 

 

 

 

 

Figure 6.4. Polar plots representing measurements for LPL and CPL generated from a
QWP. Left. Horizontal GBP-O) and vertical (BP-n) polarization plots show definite but
slightly warped production of LPL. Right. Even when retardance of the QWP is
modified to conform to specifications for modifying polarization, the light generated is
not circular.

6.1.2 LQ-shaper Technique

The other polarization modification technique used involved the use of the pulse
shaper to introduce the retardance corresponding to the production of either LPL or CPL.
Computer-controlled femtosecond polarization pulse shaping was first reported by
Brixner and Gerber in 2001 , where a 256-pixel double mask SLM pulse shaper was used

to modulate the degree of ellipticity as well as the orientation of the elliptical principal

130

axes within a single laser pulse.199 The light’s polarization state reaches different linear
and elliptical orientations with varying degrees of ellipticity while keeping intensity
levels in the two polarization components equal. As noted by Brixner and Gerber, full
control of the temporal polarization of ultrashort pulses requires four independent degrees
of freedom for each frequency component: total intensity, total phase, ellipticity, and
orientationm’ 200 Silberberg used three shapers to control the polarization state while

20'. The method used in this section, which, in the

keeping spectral intensity constant.
interim, is. named the LQ-shaper technique, utilizes a digital binary approach to
manipulate the electric field of the laser. The incident x-polarized light (Ex) exits with
field components (Ex ’, Ey') that depend on the phase retardance (05A and ¢3) introduced

by the LO elements whose slow axis is oriented at 45° and -45°. The generation of

linearly polarized output along the x (horizontal) or y (vertical) axes is based on the

equations: Ex =(1/2)Ex(e'¢A +e'¢B) and Ey = (1/2)Ex (e’¢A —e'¢B ). For LCPL or

 

 

RCPL, the field components are E'L =Jl/2(Ex —iEy) and ER =Jl/2(Ex +iEy).

The binary states that were used to generate the desired phase, amplitude (T=|Ex|2) and

polarization are listed in Table 6.1.

Table 6.1. Phase-Amplitude Modulation for LPL and CPL

 

 

 

 

 

 

 

 

Binary State Phase Amplitude LPL CPL

0 0:0 T=1 E'x=1,E'y=0 E'L=1,E'R=0

1 ¢=n T=° E'x=0,E'y=1 E'L=0,E'R=1
0.1“”, 13“” 4» = 0, o = 0 T = 0, 'r = 0 Ex = 0, By: 0 Era/4, E'R= -1r/4
9A“), ¢B(l) (l = 71/2» ‘1’ = 7V2 T = 71/2, T = —1r/2 E'x = 0, E’y = 1r E'L= —1r/4, E'R= n/4

 

131

 

6.1.2.1 Experimental Validation of Polarization Condition

The measurements made using the LQ-shaper technique indicated much better
control over the outgoing polarization condition of the laser. For this exercise, a polarizer
was placed in the beam’s path after the pulse shaper, and the laser intensity was recorded
with a spectrometer. The polarizer was yet again rotated in 20° increments while running

the phases BP-O or BP-rr. The results of the measurements are shown in Figure 6.5.

 

J .L

 

Normalizeg intensity
0

..L
O O
L A .L - 1 A l i l A I

 

 

 

 

 

 

 

Figure 6.5. Polar plots representing measurements for LPL and CPL generated from the
LQ-pulse shaper technique. Left. Horizontal (BP—O) and vertical (BP-rr) polarization plots
show definite production of LPL. Right. When the program is modified to produce
circular polarization, the light generated is elliptical.

Compared to the QWP measurements, the plots show more accurate and symmetrical
production of LPL. Although the results for CPL indicate elliptical polarization, they give

the impression that the potential of the LQ-shaper technique to deliver the required

polarization is much more reliable than the QWP technique.

132

 

6.2 Concept: Chiral Pulses and Binary Polarization

The prospects for a wider range of applications of laser control have increased
further due to innovations in pulse shaping technology. Femtosecond polarization pulse
shaping199 allows the manipulation of electromagnetic waves in a filndamentally new
way. The polarization state of light, as well as the intensity and oscillation frequency, can
be varied as a function of time within a single femtosecond laser pulse. The following
sections (6.2 and 6.3) exploit the possibility of experimentally realizing enantiomeric
resolution by controlling the polarization state of femtosecond laser pulses while
introducing binary functions at the SLM.

A number of theoretical studies have proposed the use of lasers for the resolution
and synthesis of optically active compounds. Shapiro and Brumer proposed chirality
control based on the quantum coherences between rovibronic states to selectively enhance
production of right-handed or left-handed optical isomers.202'204 These theoretical
methods have yet to be accomplished in the laboratory. It was predicted that circularzos’
206 or variable elliptical polarization207 would be required to control the relative
populations of enantiomeric molecular excited states. Fujimura, Gonzalez and Manz
considered the use of a sequence of linearly polarized short pulses to resolve a racemic
mixture.2°8'2” Gonzalez later proposed the use of a linearly polarized IR pulse acting on
pre-aligned molecules followed by a UV photolysis pulse to distinguish between
enantiomeric compounds using MS. This proposal was based on the observed differences
in fragment ions when tailored laser pulses were first used to selectively fragment a

complex molecule.32 On the experimental front, the ability to generate intense short

pulses with varying ellipticity was introduced by Wefers and Nelson.212 The research

133

 

groups of Gerber46' 199‘ 200’ 213, Weiner2 14, and Silberberg215 have explored the preparation
of phase- and polarization-shaped pulses. Such laser pulses are expected to generate
torques on molecules that could lead to more sophisticated control of chemical reactions,
including enantiomeric synthesis. However, these goals have not been demonstrated.
Efforts from the Dantus research group towards the use of tailored laser pulses
for molecular identification at low concentration‘69 have led to the idea of preparing
chiral electric fields to induce differential fragmentation of enantiomeric compounds.
These laser pulses hypothetically have an electric field with a time-varying ellipticity. By
coupling these intense laser pulses with a mass spectrometer, achieving highly sensitive
chiral resolution is sought. Photonic control over optically active samples is a dream that
has been considered theoretically for two decades.2°2'204
Conceptually, a chiral pulse causes asymmetric switching between right- and left-
elliptically polarized light, supposedly leading to differential laser-induced fragmentation
and ionization in enantiomerically different molecules. This perception can be traced
back to the optical activity of these molecules. The inference that can be made is that a
right-handed enantiomer will couple to the chiral electromagnetic field differently than a
left-handed enantiomer. Therefore, an intense pulse with a specific succession of right-
and left-optical torques would thereby play the role of a chiral photonic enzyme,
preferentially cleaving one of the enantiomers. Based on this physical model, the order of

the applied set of torques should determine which of the enantiomers is preferentially

cleaved.

134

6.3 Experimental Implementation

In this section, the initial implementation of chiral resolution with polarization
shaped laser pulses is introduced. Although the data obtained could have led us to the
conclusion of the efficacy of generating chiral pulses for enantiomeric discrimination,

subsequent discouraging experimental results gave us a contradicting message.

6.3.1 Polarization shaping with QWP technique: Initial experiments

Experiments were carried out using a TOF mass spectrometer in which the
electron ionization source was replaced by amplified femtosecond-laser pulses that were
polarization-shaped using a modified phase-amplitude pulse shaper.90 The laser pulse was
dispersed in the frequency domain. This arrangement allows one to control the phase and
ellipticity of the light transmitting through each of the 128 pixels in the dual mask pulse
shaper. The number of pixels was reduced to 10 super-pixels (each 10 pixels wide, see
Figure 6.6) which limited the shaper to produce LPL (horizontal or vertical). A QWP was
then used after the pulse shaper to convert the linear polarization, vertical and horizontal,

into circular polarization, right and left, respectively.

 

phase mask

 

 

 

 

 

 

11011110I0|0I1|1lfi
0 W 127

pixels

 

 

 

Figure 6.6. Phase mask with 0 or it retardation binned to 10 super pixels, each 10 pixels
wide, covering the full spectrum of the fundamental laser pulse. The SLM used has 128
pixels.

135

The binary polarized shaped pluses were then focused on a small concentration of
molecules in the gas phase (approximately 104 molecules interact with each laser pulse).
The resulting ions and fragment ions were detected and identified by their TOF to the
detector. Theoretically, the combination of binary shaping and polarization changes with
the QWP results in fields with rapidly switching CPL, as shown in Figure 6.7. The
electric fields shown are for the binary polarization pulse BP167, which was used to
obtain most of the initial data in this report. When only half of the field is flipped to y
polarization, the resulting fields are simpler to visualize. Two such cases (xy and yx) are

shown. The prevalent rotation of the field is indicated with arrows.

 

 

 

 

 

Figure 6.7. Electric fields for binary polarization pulse BP167 for right and left circular
polarization.

Rotation of the optical-axis angle of the QWP allowed us to determine the
ellipticity of polarization of the pulses used in the experiment. Circular polarization was
obtained when the QWP was oriented at 45°. When oriented at 90°, linear polarization
was attained. Changing the optical axis angle provided a reliable control for our
experiments because it only affected the polarization of the beam, without influencing

other variables such as energy or dispersion.

136

 

In order to find the greatest enantiomeric specificity, an experimental evaluation
of the effect of 512 different circularly polarized shaped pulses on the relative peak ratios
of the different fragment ions of our samples was first performed. These pulses
correspond. to half of all possible 10-bit polarization fimctions that can be obtained by
combining right- and left-circular polarization. From these 512 experiments, the pulses
that gave the greatest difference in ion product ratios (one fragment ion peak vs. another
peak) between the right-handed (R-) and left-handed (S-) enantiomers were obtained.
Mass spectra for both compounds were then generated using these pulses. The binary
polarization function BP167 was among those that resulted in the highest contrast
between the compounds.

”I for the enantiomeric compounds

The standard electron ionization mass spectra
tested, (R)-(+)- and (S)-(—) or-methyl-benzenemethanol (BME) (specific rotation,(R)-(+)-
BME = +42°, (S)-(—)-BME = —42°), are shown in the top panel of Figure 6.8, with both
enantiomers giving almost identical results. The peaks A and B represent the fragment

ions at m/z 51 and m/z 107, respectively. The mass spectra of the enantiomers obtained

using polarization-shaped femtosecond-pulse excitation are shown in the bottom.

137

 

 

 

 

o . 3:107
100 ..
(R)
‘ "I H

O) (D
O O

A
9-.

 

 

 

N
O

 

 

 

9

 

Relative abundance (%)

 

 

 

 

 

 

 

 

 

Figure 6.8. Mass spectra of or-methyl-benzenemethanol (BME). Top: electron ionization
spectra of R- (open circles) and -S-BME (black circles). Bottom: special fragmentation
mass spectra of R- and S-BME using shaped laser pulses with circular binary
polarization. Peaks A and B refer to fragment ions at m/z 51 and 107, respectively. The
ratio of intensities of these peaks was used to differentiate the enantiomers. Inset:
structures of the R and S enantiomers of BME.

The chemistry involved in the photoionization and fragmentation of the above
molecule is shown in Figure 6.9. The molecular ion peak at m/z 122 (C3H90H+) was
significantly reduced with laser pulse excitation (~120 ,uJ/pulse) when compared to

electron ionization, leading us to speculate that ionization of the molecule with TL pulses

was quite extensive.

138

 

 

 

 

+ “0‘0”
-CH3 C ..
H
mlz = 122 mlz = 107
-C=O
+ +
.2H+
———>
mlz = 79 mlz = 77
-C2H3
V <——
m/z = 43 \ mlz = 51
4111*
+
V + ——'>-C _
V _
mh=39 mk‘27

 

 

 

Figure 6.9. Reactions involved in the fragmentation of BME

The dominant fragments were at m/z 77 (CoH5+) and 79 (CoH7+), which are also
substantial fragments in the electron ionization mass spectrum. The abundances of the
smaller fragments with peaks at m/z 27(C2H31), 43 (C3H7+) and 51(C4H3+) were enhanced
by femtosecond-laser laser excitation. The analysis in this section concentrates on two of
the peaks, a small molecular weight fragment at m/z 51 (designated here as A) and a large
fragment ion corresponding to loss of a methyl fragment with m/z 107 (C7OH7+,
designated as B). Figure 6.8 highlights the substantial difference in the intensities of
peaks A and B between R- and S-BME when circularly polarized BP167 pulses are

applied. The A/B ratio obtained for R-BME was about 1.3. A significant increase (~2.2)

139

in the A/B peak ratio was obtained when measurements were made on S-BME. These
values correspond to a 60% difference in NB ratios between the two enantiomers.

The statistical significance of these initial results is apparent in Figure 6.10 where
one hundred mass spectral measurements are made for each enantiomer using BP167.
The histogram representation of these measurements shows that the S enantiomer
undergoes greater fragmentation than its right-handed counterpart. This difference
vanished, as expected, when the optical axis of the QWP was rotated to yield linear

polarization-shaped pulses (BP167) (Figure 6.10, bottom).

1:] R (NB) 3 (NB)
‘C-BP _

   

N
O

RIS=1 .7102

A
-‘f‘-

 
 

V/l/l/l/l/I/I/I/l/l/I/l/I/A

Sample count (%)
O

 

 

 

 

 

 

 

 

..O - 0"
:1
..—fi

     
  

N
.°

V

/

 

A
9‘ -

     

A

 

’.’/////////

o II
# 'IIIIIIIIIIIIIIIIIIIIII

   

'l'Ou'ua

 

 

Sample count (%)
O

 

'I’lll’l’d ’IIIIIIIIII

 

C€IZZW£?_/(””’m

 

 

Figure 6.10. Statistical evaluation of phase and polarization shaping for enantiomeric
selectivity. Top: histograms representing 100 measurements for each enantiomer using
circularly polarized (BP167) laser pulses. Bottom: histograms representing 100
measurements for each enantiomer using linearly polarized (BP167) femtosecond laser
pulses.

140

Analysis of the initial experiments showed that the fragment ion at m/z 107 in
BME, corresponding to the loss of a methyl group, exhibited the greatest sensitivity to the
chiral laser pulses. Pure RCPL or LCPL did not achieve a significant level of asymmetric
photodestruction, this being consistent with previous attempts.216 It was also observed
that the number of polarization flips introduced by the shaper correlated with the degree
of contrast observed. This indicates the possibility of higher levels of discrimination if
polarization phase functions with higher bit-resolution are used. Assessment of a number
of electric fields that were found capable of enantiomeric resolution showed a linear
polarization component that preceded the large rotating component, analogous to a
propeller. The direction and curvature of the propeller component were believed to
influence enantio-selective photochemistry. The high selectivity shown in these
asymmetric photodestruction experiments cannot be explained based on the circular
dichroism of the sample. The use of intense laser fields inducing multiphoton excitation

has been shown to increase the anisotropy factor by two orders of magnitude.2'7

6.3.2 Problems encountered with QWP technique

Although this technique initially indicated the potential to indicate differences
between two enantiomers that are large enough to be able to distinguish between them,
the question regarding the possibility of preparing laser pulses capable of resolving
enantiomeric pairs remains unanswered. The lack of experimental reproducibility has
been the most frustrating obstacle to ascertaining the feasibility of this method for
resolving chiral substances. The success of our earlier observations and the ensuing

attempts to replicate the results is, for the most part, puzzling. The possibility of an

141

unmonitored modification in the characteristics of the laser pulse during experimentation
may be a contributing factor to the results obtained. Although experimental parameters
such as sample pressure, room temperature and laser intensity were closely monitored
before the actual data acquisition, the probability that small changes in these parameters
could have significantly affected molecular fragmentation is uncertain. The use of a QWP
to deliver the desired polarized pulses may have introduced an electric field that affected
the molecules significantly. The optic was tilted slightly (angle unmeasured) from its
orthogonal (to beam) position to introduce the assumed circular polarization. Very slight
deviations from this particular angle may have significantly impacted the results of

subsequent experiments.

6.3.3 Polarization shaping with LQ-shaper technique

Due to the nature of the polarized light generated by the QWP technique and the
fact that precision was at such an uncompromising standpoint, our next task was to test
whether the LQ-shaper technique would be a more feasible and reproducible way to
attain enantiomeric discrimination. The experimental implementation was similar to what
was done earlier using the former technique, except that the QWP was taken out of the
setup, and binary functions were imprinted onto the LO arrays of the SLM to generate

either LPL or CPL. A number of phase functions, including chirp masks were tested.

6.3.3.1 Simple Linear and Circular Polarization

The response of the sample enantiomers to simple pulses tailored to produce

combinations of LPL and CPL was first tested. The polarization masks were generated

142

based on 128-pixel 2-bin resolution binary functions. Sixteen masks were introduced into

the SLM. These masks are shown in Table 6.2.

Table 6.2. Two-bin binary polarization masks

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SLM pixel 1 to 64 SLM pixel 65 to 128
Linear horizontal Linear horizontal
Linear horizontal Linear vertical

Linear vertical Linear horizontal
Linear vertical Linear vertical
Left circular Left circular
Left circular Right circular
Right circular Lefi circular
Right circular Right circular

Linear horizontal Left circular
Left circular Linear horizontal
Linear horizontal Right circular
Right circular Linear horizontal
Linear vertical Left circular
Left circular Linear vertical
Linear vertical Right circular
Right circular Linear vertical

 

 

 

 

Figure 6.11 shows mass spectra collected when the mask right circular-left
circular was applied on R- and S-BME. The differences in relative yields for both
enantiomers appear significant, especially the absence of the peak at m/z 91 (tropylium
ion) in the R- enantiomer. There may be a slight effect of background on the heights of

each fragment ion in the plot below; however, it is very clear that the tropylium ion is

143

present in the S- enantiomer. Although this could be an indication of successful chiral
recognition, the reproducibility factor is still a main constraint. A major dilemma occurs
when the set of 16 masks are run several times more. Based on the data gathered, it
seems that as the S- enantiomer vapor is irradiated longer with the laser, the tropylium
ion peak diminishes in yield. This occurrence may be a major factor in elucidating the
lack of reproducibility in the data. Why this phenomenon occurs is an issue that needs

further contemplation.

 

 

   
   

 

1
j . R-BME
1'0 oS-BME " u
'0
'1; 0.8‘ o o
g 1 ll .
E 0.6-
9
3
‘6 0-2‘ l a

 

mlz

Figure 6.11. Collected mass spectra of R- and S-BME with right circular and lefi circular

polarization combination. The spectra were normalized to the highest fragment ion peak
at m/z 79. Note the presence of the peak at m/z 91 (encased) for S-BME.

6.3.3.2 10-bit Binary Mum Circular Polm

The next task was to test more complicated 10-bit BP functions generated to
produce either LPL or CPL. The phases were prepared from a set of 1024 10-bit binary
functions, where the first 64 and the last 64 phases from this original set were picked to
create a new phase list — a set consisting of 128 phase functions repeated ten times. This

time, no tr0pylium ion was observed in either of the enantiomers. However, with all ion

144

peaks normalized to the m/le7 fragment ion peak intensity, the S-enantiomer seemed to
be generating higher yields of the lower-mass major fragment ion peaks (m/z 43, 51, 77
and 79). Figure 6.12 shows the fragmentation patterns observed when left circularly
polarized TL pulses (left) and circularly polarized binary (right, phase shown in inset)
shaped pulses were used. Once again, although there is a clear distinction between the
two samples tested, the experiment lacked reproducibility. It is interesting to note that the
intensity of the 107 peak decreases significantly from TL to shaped pulses. In addition,
while this decrease is observed, increases in the yields of the lower-mass fragment ions

are also apparent with the polarization shaped pulses.

 

'. I R-BME 7 mm
1 .8" O S'BME “ o
i ‘ o

phase

 

 

 

 

 

 

 

 

 

Figure 6.12. Mass spectra of R- and S-BME using 10-bit polarization shaping. Left:
observed mass spectra of the enantiomers as a consequence of TL LCPL. Right: observed
mass spectra of the same compounds when a more complicated circularly polarized
binary function is applied. The structures of the binary functions used are also shown
(phase).

When the intensities of the fragment ions are plotted against the BP functions
tested, the differences between R-BME and S-BME are more apparent. Figure 6.13

shows three such plots. Based on the observed data, differences in intensities of peak 107

145

 

 

 

(A) occur in the middle section of the set of 128 phase functions. In the case of peak 51
(B), the bigger differences lie primarily at the edges (first and last) of the set of functions.

The ratios between the peaks at m/z 51 to those at m/z 107 are shown in Figure 6.13C.

 

Normalized intensity
.0 9 .o .o
N A O) on

9
o

 

 

 

 

 

A

Q
-1.

 

Intensity (mlz 51:107)

.°.°.°
Ac»

 

 

10-bit binary phase

 

 

 

Figure 6.13. Changes in intensity of fragment ion peaks at m/z 107 (A) and m/z 51 (B)
as 128 circularly polarized binary functions are scanned across the SLM. The intensities
were normalized to that of m/z 107. C. Plot of the ratios between m/z 51 and 107 for the
128 polarization shaped pulses. Error bars indicate standard deviation of data (n=10).
These ratio differences (R/S) range from 1.13 i 0.04 to 1.34 :l: 0.06. It is also
apparent from Figure 6.13 that as the binary function increases in complexity (middle
section of phase set), the ratio differences between the enantiomers also increase;

however, the problems of noise and baseline measurement inaccuracies become more

well-defined and the issue of reproducibility is again a matter of concern.

146

The symmetrical parabolic traces observed in the same figure show that the
intensity of fragmentation is independent of the direction of polarization of the pulses
(LCPL to RCPL). Although the expected reversal in fragment ion peak intensities as the
polarization direction was flipped from left to right circular was not observed, the
capability of the LQ-shaper method to produce the intended polarization pulses was
undeniable.

The experiment was repeated, this time, with a modification in the path length of
the beam going towards the entrance of the mass spectrometer. The optical path length
was shortened by about 10 feet. The results of this experiment showed that R- and S-
BME had no significant differences in fragment ion yields except for the peaks at m/z 105
and 106. Figure 6.14 shows the TOF spectra obtained for the samples. Notice the
difference in intensities of the fragment ions at m/z 105 and 106. It is logical to think that
these differences should be expected, especially when the structures of these compounds
are considered. Once the hydroxyl group and the hydrogen, which are both attached to
the chiral carbon, are removed, the resulting fragments are those with m/z 105 and 106.
Further fragmentation of the molecular ion eliminates its chiral nature, and thus no
differences in the relative yields of the other fragment ions are observed. Unfortunately,
a closer look at the standard EI mass spectra of the sample compounds also shows slight

differences in abundances of these two fragment ion peaks.

147

 

 

 

 

 

 

 

 

 

J/
....
0.014- + o —R-BME
. 0.0061 c/ —— S-BME
0.012- ‘ \H
‘ 0.004« '
0.010a ‘ . rn/z107
E 0.002~
.92 ~
a); 0.008“ 2730 - 2740 - 27230 ' 2760 H '
.2
2.6
g 0.006- -
« / m/z106 "
m/2105
0.004“ i
0.002-
I
'T'I'I‘I'F'Ifr'I'I—‘l
2700 2720 2740 2760 2780 2800
Time of flight

 

 

Figure 6.14. Raw TOF data for R- and S-BME showing peaks with intensity differences.
The peaks shown represent fragments at m/z 105, 106 and 107. The inset zooms in on
peaks 105 and 106, highlighting the difference between the two enantiomers.

6.3.3.3 Chirp Combined with CPL

When the order of the different frequency components of the pulse is reversed, a
reversal of fragmentation yields of the enantiomers is expected. This hypothesis was
tested by introducing circularly polarized chirped laser pulses, including both positive
and negative chirp. Mass spectra were collected using the same 10-bit binary polarization
functions used in Section 6.3.3.2, with the addition of positive or negative chirp masks.

Much to our disappointment, there didn’t seem to be any reversal in the intensities

of the fragment ions. Figure 6.15 shows data collected when CPL pulses were chirped to

148

 

 

either +900 or —900 fsz. The negatively chirped circularly polarized pulses indicated
greater fragmentation, with the R enantiomer fragmenting at a slightly higher yield than
S-BME. In terms of positively chirped pulses, the S enantiomer appears to have a

generally higher yield than R-BME, but this difference is trivial when the noise level of

the data is factored in.
2 0 I R: + chirp
- ‘ o S: + chirp — ‘
. a R: _ chirp ”ha“ , 11ml“ $
V S: — chir P 33° 2
1.6~ p - 9 E
'2 ‘ ‘ V
d.)
-- A
:12 . 1
a: 0 l
I” i t
E l
E
O
z

 

 

 

 

 

 

Figure 6.15. Mass spectra of R- and S-BME using chirped lO-bit polarization shaping.
Left: observed mass spectra of the enantiomers for positively and negatively chirped lefi
circularly polarized pulses. Right: observed mass spectra of the same compounds when
positive and negative chirp are applied in addition to a complicated circularly polarized
binary pulse. The structures of the binary functions used are also shown (phase).

Analysis of all data sets recorded showed no significant statistical differences
between the enantiomers and between positive and negative chirp effects. While one run
of the different polarization masks could highlight differences between R- and S-BME,
these differences are reduced considerably when error plots of the mass spectra are
constructed. Interestingly, the effect of binary polarization shaping on the increase in

yield of the lower-mass fragment ions and the decrease in the m/2107 peak, which were

also observed in earlier experiments where chirp was not an experimental parameter,

149

does insinuate that some control over the fragmentation patterns of the molecules studied
is possible. The issue of reproducibility is, regrettably, the major obstacle to the success
of these experiments. The main source of error in the experiments could be attributed to
laser intensity fluctuations which are amplified by about an order of magnitude when
taking into account the 6- to 8-photon process initiating ionization and fragmentation.
Pulse-to-pulse stability limits the precision with which one can estimate the effect of a
particular control experiment. Intensity variations have critical consequence on the
statistical evaluation of an experiment. For an nth order nonlinear optical experiment, the
average value obtained after N measurements on a signal, where fluctuations are of

magnitude iais determined by in cJN('°‘5).91

6.4 Inferences and Future Possibilities for Chiral Pulses

The reliable identification of a large number of enantiomeric species by coherent
control calls for further studies. The fact that initial experiments were not completely
reproducible requires extensive improvement in the technique and considerable
adjustment of experimental parameters. The positive results gleaned from early tests are
nonetheless encouraging. The task of distinguishing between enantiomers and the overall
intention behind chiral recognition by coherent control are, without any doubt, worth

advocating.

150

 

 

 

CONCLUSIONS

Controlling laser-molecule interactions has become an integral part of developing
the future of devices and applications in spectroscopy, microscopy, optical switching, and
photochemistry. A number of applications are dependent on multiphoton transitions.
Coherent control of multiphoton transitions could bring about a significant improvement
of these methods. In microscopy, multiphoton transitions are used to activate different
contrast agents and suppress background fluorescence; with coherent control, the
potential for selective probe excitation could ascertain its valuability. The potential for
two-photon excitation in PDT has been demonstrated by researchers to result in two key
benefits: deeper penetration depth of light and new PDT treatments of skin melanoma
without the addition of a drug or dye. The successful demonstration of selective two-
photon excitation as a tool for molecular recognition is, in my mind, a major
accomplishment of coherent control in terms of making an impact in fields other than
physical chemistry. With the first few steps towards our vision for biomedical
applications accomplished, future endeavors of coherent control research towards
selective activation of PDT drugs and noninvasive detection and therapy of cancerous
cells should be imminent.

In photochemistry, different dissociative states are accessed through multiphoton
transitions. Coherent control could be used to select the reaction pathway and
consequently yield specific products. Multiphoton excitation of molecules to highly
excited electronic states, each with a different decay pathway, makes laser control of

molecular reactivity possible, provided selectivity, combined with the sensitivity of

151

 

 

analytical methods such as MS, can be achieved. The combination of the physical
principles of coherent control and the analytical merits of MS could prove useful for
future applications in chemical detection.

The prospect of using femtosecond laser control as a means for molecular
recognition in the biomedical and analytical fields has been elucidated in this dissertation.
Although our objective for enantiomeric resolution has not been realized, the potentially
huge impact of successful chiral recognition by coherent control, especially in the

biomedical, pharmaceutical and environmental areas, merits anticipation.

152

APPENDICES

153

APPENDIX A

Group Velocity Dispersion

When short laser pulses propagate through absorptive or dispersive media, the
possibility for dispersion is high. Different frequency components propagate with
different velocities in all materials. GVD takes place because the index of refraction of
light increases with increasing frequency for most materials in the visible region of the
spectrum. This causes the low frequency components in a pulse envelope to propagate
faster than the higher frequency components. The pulse envelope is temporally broadened
and develops a positive chirp. Conversely, if a pulse travels through an anomalously
dispersive medium, high frequency components travel faster than the lower ones, and the
pulse becomes negatively chirped, decreasing in frequency with time. The majority of
pulse broadening in ultrashort pulse lasers is caused by the positive GVD of the gain
medium. The consequence is that a short pulse consisting of many wavelengths is
stretched in time. In order to compensate for this effect, negative GVD, where short
wavelengths propagate more rapidly than longer wavelengths, is needed. A practical
method for doing this is to introduce pairs of prisms into the cavity.218 Although the
prism material will itself contribute positive dispersion, it is possible to configure the
prism pairs so that the overall contribution is negative. The figure below illustrates this

configuration using four prisms.

154

 

 

Figure A-l. Prism arrangement to compensate for positive GVD

The arrangement of the prisms in Figure A-l permits negative GVD to be
achieved using materials with positive GVD. Light is spectrally dispersed by the first
prism. The lower frequency components of the light must travel through more glass in the
second prism and are delayed relative to the high frequency components resulting in
negative chirp. A second pair of prisms is needed to spatially recombine the beam. The
layout of the four prisms allows the GVD to be adjusted from positive to negative by
translating the prisms so that the beam must travel through less glass. The prisms are cut
and orientated so that the rays are incident at minimum deviation and Brewster’s angle to

minimize losses. The total dispersion of the prism sequence is calculated as

2
D = 11161—2}: , where L is the physical length of the light path, P is the optical path length.
C d/l

The derivative dZP/ d/l2 is a function of the angular divergence, the refractive index of
the prism material and the apex separation, l, of the prisms. Provided the prism separation
is sufficiently large, the positive dispersion of the material can be balanced. Therefore, by
changing the prism positions it is possible to vary the total dispersion of the cavity from

positive to negative.

155

Another way of controlling GVD, without the need of introducing intracavity
glass, is by using chirped dielectric mirrors as the source of broadband negative GVD.2'9’
22° Chirped mirrors reflect each wavelength from a different depth through the dielectric
coating, which is made up of multiple stacks of varying thickness. Such mirrors have

facilitated the generation of sub-5 fs laser pulsesm

156

 

 

APPENDIX B

Description of components of vacuum system -— TOF assembly used for mass

spectrometry experiments

A.1. Vacuum System

After a molecule is introduced into the mass spectrometer, the molecule is first
ionized and then fragmented. Following this, the ions can be selected and counted.
However, for proper operation, the interior of the mass spectrometer must be evacuated.
The ion source, mass analyzer (TOF), and detector all are under vacuum. The vacuum
system makes it possible for ions to move from the ion source to the detector without
colliding with other ions and molecules.

Achieving a ' vacuum sufficient for the effective operation of our mass
spectrometer system requires a multi-step pumping scheme with a rotary mechanical
roughing pump followed by a high vacuum diffusion pump and turbomolecular pump
(TMP)..

The roughing pump in our TOF-MS system is a mechanical vacuum pump that
works in collaboration with the high vacuum diffusion pump in the preliminary stage of
the pumping process. The roughing pump is connected to the main chamber of the
diffusion pump through a separate pumping line and valve. The chamber is pumped down
to a pressure of about 10'2 to 10'3 torr , at which point the valve to the roughing pump is
closed. Then a second valve that connects the diffusion pump to the chamber is opened,

at which point the diffusion pump takes over.

157

 

The diffusion pump is necessary to attain pressures below 10'3 to 10'4 Torr. The
vacuum diffusion pump in our setup is a stainless steel chamber containing vertically
stacked cone-shaped jet assemblies. It works by capturing gas molecules by vaporizing
and condensing a pool of low vapor pressure silicone oil located at the base of the
chamber. The oil is heated to boiling by an electric heater located beneath the floor of the
chamber. The vaporized oil is forced up through a central column in the pump and is
vented downward through jets in conical sheets against the inner sides of the pump
chamber. As the vapor jet impacts the outer cooled shell of the diffusion pump the gas
entrained in the jet flow coalesces, carrying the entrained pumped gases into the base of
the pump where the gas pressure is increased and pumped by the rough pump from the
diffusion pump outlet. The coil around the pump is for cooling water. The cooling water
is necessary to cool and condense the oil fast enough to prevent thermal runaway and
permit operation over longer periods of time. With a diffusion pump, vacuum pressures
between 10'5 to 10'8 Torr can be attained.

The TMP is placed in-line with the roughing pump and is used to obtain and
maintain high vaculun. This pump works on the principle that gas molecules can be given
momentum in a desired direction by repeated collision with a moving solid surface. The
TMP effectively functions as a turbine. It rotates very rapidly (IO-20,000 rpm) while the
turbine wings mechanically pump the residual gasses from the high vacuum. Gas
molecules from the inlet of the pump are hit and brought towards the exhaust in order to
create or maintain a vacuum. When switched on, this pump generates a characteristic
high pitch humming. The machine can be a very versatile pump able to generate

intermediate vacuum (~10'4 Torr) up to ultra-high vacuum levels (~10'lo Torr). When

158

switched off, a free flow of air can take place through the device (unless some valve is
shut), but unlike the diffusion pump, normally no backflow of dirty pump oil debris can

take place. The TMP in our system has a pumping speed of 220 US.

MCP
detector

turbomolecular
pump —'.4 '

. diffusion pump

 

Figure A-2. Maj or components of TOF mass spectrometry system include the mechanical
roughing pump (not shown), diffusion pump, turbomolecular pump, ion gauge, flight
tube and the MCP detector.

159

 

 

   

   
   

  
 
  

  
 

than":

llllllmnumlll” ll

 

1 2
2
1
1- heater 6- gas molecules
2- boiler 1- vapor jet
3- pump body 8- backing vacuum
4- cooling coil connection

5- high vacuum flange 9-12- nozzles

 

 

 

Figure A-3. Diffusion pump mode of operation. A heating coil (1) evaporates the special
diffusion pump oil in the boiler (2). The oil vapor then ascends through the centre of the
pump and shoots at high speed out of a multi-stage system of concentric ring-shaped
nozzles (9-12). The oil vapor shoots out in the shape of a cone (7) and then condensates
against the wall of the pump (3), which is cooled by water flowing through the cooling
coils (4). The oil then returns into the boiler as a thin film along the wall. The air or other
gas in the high vacuum part of the pump (6) is forced to move along with the high speed
oil vapor in the different cascade levels and leaves the pump on the low vacuum side of
the pump (8).

An Ion Gauge, which is connected to our mass spectrometer assembly, is used to
measure pressure. It consists of three distinct parts, the filament, the grid, and the
collector. The filament is used for the production of electrons by thermionic emission. A
+ve charge on the grid (anode) attracts the electrons away from the filament; they

circulate around the grid passing through the fine structure many times until eventually

160

they collide with the grid. Gas molecules inside the grid may collide with circulating
electrons. The collision can result in the gas molecule being ionized. The collector
(cathode) inside the grid is -ve charged and attracts these +ve charged ions. Likewise they
are repelled away from the +ve grid at the same time. The number of ions collected by
the collector is directly proportional to the number of molecules inside the vacuum
system. Absolute pressure is determined by the frequency of gas molecules interfering
with cathodic electrons attracted to the grid.. The resulting ionization of the gas (collected
ion current) is measured to determine the pressure. For most of the MS experiments in
this dissertation, the pressure of molecules in the vacuum system ranged from 8x10'6 to
1x10'5 torr. The ion gauge was kept on throughout the experiments to allow us to

monitor pressure changes.

 
 
  

Microammeter

 

lon collector 3+
Electron -

 

 

 

 

 

 

 

 

 

- collector
Milliammeter
Electron
I I emitter 1 50V
ll——lu-—'l'

20V

Figure A-4. A typical ion gauge. It has a heated filament biased to give thermionic
electrons (70e) energetic enough to ionize any residual gas molecules during collisions.
The positive ions formed drift to an ion collector held at about 150V. The current
measures gas number density, a direct measure of pressure.
A.2. TOF Assembly (D-677, R.M. Jordan Co., Inc.)

A TOF mass spectrometer uses the differences in transit time through a drift

region to separate ions of different masses. It operates in a pulsed mode so ions must be

produced or extracted in pulses. An electric field accelerates all ions into a field-free drift

161

region with a kinetic energy of qV, where q is the ion charge and V is the applied voltage.
Since the ion kinetic energy is 0.5mv2, lighter ions have a higher velocity than heavier
ions and reach the detector at the end of the drift region sooner.

An electric field between the repeller and the extractor electrode accelerates the
formed ions past the ground electrode, and through the field-free TOF region (flight
tube), and onwards to the detector. The extractor and ground electrodes are thin metal
plates with a small opening at the middle. The voltages on the extractor and repeller
plates were kept at 1600 and 2500 V, respectively. At the top of the flight tube is a dual
plate detector (multichannel plate or MCP) which consists of a material with a very low
work function, such that a cascade of electrons is released from the top whenever an ion
strikes the lower surface. The subsequent electron pulse from the top of the detector plate
strikes an anode, which passes the signal through some amplification electronics to the

counting electronics where the pulses are recorded.

162

 

fl n-B—> signal out
L ]
7r W A, dual plate

detector

 

 

 

 

 

 

 

 

 

> flight tube
24"
V ground plate
ionization extractor plate
region repeller plate

 

 

 

Figure A-5. Lens stack and flight tube assembly. An assembly of grids, steering plates
and lens elements mounted on a flange.

The detector used in our mass spectrometry experiments is a C-701 18 mm dual
MCP detector (R.M. Jordan Co., Inc.) It has a 50 ohm coaxial output and provides high
gain (106 to 107) with sub-nanosecond rise time. It is fitted with an input grid (90%
transparency) wired to an external SHV (standard high volume) coaxial UHV (ultra high
voltage) feedthrough that can be operated from external power supplies or voltage
dividers. This presents a flat, field free plane to the incoming ions. The maximum voltage
across each plate is 1000 Volts (V), with a 1000 minimum gain per plate.

The MCP consists of a two-dimensional periodic array of very-small diameter
glass capillaries or channels fused together and sliced in a thin plate. A single incident

particle (ion, electron, photon etc.) enters a channel and emits an electron from the

163

channel wall. Secondary electrons are accelerated by an electric field developed by a
voltage applied across both ends of the MCP. They travel along their parabolic
trajectories until they in turn strike the channel surface, thus producing more secondary
electrons. This process is repeated many times along the channel; as a result, this cascade
process yields a cloud of several thousand electrons, which emerge from the rear of the
plate. If two or more MCPs are operated in series, a single input event will generate a

pulse of 108 or more electrons at the output.

 

 

 

 

 

 

 

 

Channels

output

 

     
     

chanr‘rtlwall (““0“
Input electron. EEA’ , ,P/ . o u
electrode

 

 

 

‘ strlp current ’
.1 al
'I ‘I

Figure A-6. MCP principle of operation

 

 

 

 

164

BIBLIOGRAPHY

165

(1)
(2)
(3)
(4)

(5)
(6)

(7)

(8)

(9)

(10)

(11)
(12)

(13)
(14)
(15)
(16)

(17)

BIBLIOGRAPHY

Tannor, D. J.; Rice, S. A. J. Chem. Phys. 1985, 83, 5013-5018.
Tannor, D. J.; Kosloff, R.; Rice, S. A. J. Chem. Phys. 1986, 85, 5805-5820.
Baumert, T.; Gerber, G. Isr. J. Chem 1994, 34, 103—114.

Zewail, A. H.; Casati, G.; Rice, S. A.; Cherqui, M.; Tannor, D. J.; Kobayashi, T.;
Rabitz, H. Adv. Chem. Phys. 1997, 101, 3-45.

Warren, W. S. J. Chem. Phys. 1984, 81, 5437-5448.

Lin, C. P.; Bates, J.; Mayer, J. T.; Warren, W. S. J. Chem. Phys. 1987, 86, 3750-
3751.

Weiner, A. M.; Leaird, D. E.; Wiederrecht, G. R; Nelson, K. A. Science 1990,
247,1317-1319.

Weiner, A. M.; Leaird, D. E.; Wiederrecht, G. R; Nelson, K. A. J. Opt. Soc. Am.
B 1991, 8, 1264-1275.

Broers, B.; Noordam, L. D.; Vandenheuvell, H. B. V. Phys. Rev. A 1992, 46,
2749-2756.

Broers, B.; Vandenheuvell, H. B. V.; Noordam, L. D. Opt. Commun. 1992, 91,
57-61.

Judson, R. S.; Rabitz, H. Phys. Rev. Lett. 1992, 68, 1500-1503.

Bardeen, C. J.; Yakovlev, V. V.; Wilson, K. R.; Carpenter, S. D.; Weber, P. M.;
Warren, W. S. Chem. Phys. Lett. 1997, 280, 151-158.

Meshulach, D.; Yelin, D.; Silberberg, Y. Opt. Commun. 1997, 138, 345-348.
Yelin, D.; Meshulach, D.; Silberberg, Y. Opt. Lett. 1997, 22, 1793-1795.
Meshulach, D.; Yelin, D.; Silberberg, Y. J. Opt. Soc. Am. B 1998, 15, 1615-1619.

Baumert, T.; Brixner, T.; Seyfried, V.; Strehle, M.; Gerber, G. Appl. Phys. B-
Lasers O 1997, 65, 779-782.

Brixner, T.; Strehle, M.; Gerber, G. Appl. Phys. B-Lasers O 1999, 68, 281-284.

166

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)
(29)

(30)

(31)

(32)

(33)

Brixner, T.; Oehrlein, A.; Strehle, M.; Gerber, G. Appl. Phys. B-Lasers O 2000,
70, SI 19-Sl24.

Efimov, A.; Moores, M. D.; Beach, N. M.; Krause, J. L.; Reitze, D. H. Opt. Lett.
1998, 23,1915-1917.

Efimov, A.; Moores, M. D.; Mei, B.; Krause, J. L.; Siders, C. W.; Reitze, D. H.
Appl. Phys. B—Lasers O 2000, 70, $133-$141.

Takasago, K.; Takekawa, M.; Suzuki, M.; Komori, K.; Kannari, F. IEEE J. Sel.
Top. Quant. 1998, 4, 346-352.

Ohno, K.; Tanabe, T.; Kannari, F. J. Opt. Soc. Am. B 2002, 19, 2781-2790.

Zeek, E.; Maginnis, K.; Backus, S.; Russek, U.; Mumane, M. M.; Mourou, G.;
Kapteyn, H. C.; Vdovin, G. Opt. Lett. 1999, 24, 493-495.

Zeek, E.; Bartels, R.; Mumane, M. M.; Kapteyn, H. C.; Backus, S.; Vdovin, G.
Opt. Lett. 2000, 25, 587-589.

Zeidler, D.; Homung, T.; Proch, D.; Motzkus, M. Appl. Phys. B-Lasers O 2000,
70, $125-$131.

Armstrong, M. R.; Plachta, P.; Ponomarev, E. A.; Miller, R. J. D. Opt. Lett. 2001,
26, 1152-1154.

Armstrong, M. R.; Plachta, P.; Ponomarev, E. A.; Ogilvie, J. P.; Nagy, A. M.;
Miller, R. J. D. Appl. Phys. B-Lasers O 2002, 74, $127-$132.

Baltuska, A.; Fuji, T.; Kobayashi, T. Opt. Lett. 2002, 27, 306-308.
Baltuska, A.; Kobayashi, T. Appl. Phys. B-Lasers O 2002, 75, 427-443.

Legare, F. L.; Fraser, J. M.; Villeneuve, D. M.; Corkurn, P. B. Appl. Phys. B-
Lasers O 2002, 74, $279-$282.

Baum, P.; Lochbrunner, S.; Gallmann, L.; Steinmeyer, G.; Keller, U.; Riedle, E.
Appl. Phys. B-Lasers O 2002, 74, $219-$224.

Assion, A.; Baumert, T.; Bergt, M.; Brixner, T.; Kiefer, B.; Seyfried, V.; Strehle,
M.; Gerber, G. Science 1998, 282, 919-922.

Bergt, M.; Brixner, T.; Kiefer, B.; Strehle, M.; Gerber, G. J. Phys. Chem. A 1999,
103, 10381-10387.

167

(34)

(35)

(36)

(37)

(3 8)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

Glass, A.; Rozgonyi, T.; Feurer, T.; Sauerbrey, R.; Szabo, G. Appl. Phys. B-
Lasers O 2000, 71, 267-276.

Daniel, C.; Full, J .; Gonzales, L.; Kaposta, C.; Krenz, M.; Lupulescu, C.; Manz,
J.; Minemoto, S.; Oppel, M.; Rosendo-Francisco, P.; Vajda, S.; Woste, L. Chem.
Phys. 2001, 267, 247-260.

Brixner, T.; Kiefer, B.; Gerber, G. Chem. Phys. 2001, 26 7, 241-246.

Levis, R. J .; Menkir, G. M.; Rabitz, H. Science 2001, 292, 709-713.

Levis, R. J .; Rabitz, H. A. J. Phys. Chem. A 2002, 106, 6427-6444.

Damrauer, N. H.; Dietl, C.; Krampert, G.; Lee, S. H.; Jung, K. H.; Gerber, G. Eur.
Phys. J. D 2002, 20, 71-76.

Bergt, M.; Brixner, T.; Dietl, C.; Kiefer, B.; Gerber, G. J. Organomet. Chem.
2002, 661, 199-209.

Daniel, C.; Full, J.; Gonzalez, L.; Lupulescu, C.; Manz, J.; Merli, A.; Vajda, S.;
Woste, L. Science 2003, 299, 536-539.

Homung, T.; Meier, R.; Zeidler, D.; Kompa, K. L.; Proch, D.; Motzkus, M. Appl.
Phys. B-Lasers O 2000, 71 , 277-284.

Zeidler, D.; Frey, S.; Kompa, K. L.; Motzkus, M. Phys. Rev. A 2001, 64, Art. No.
023420.

Ando, T.; Urakami, T.; Itoh, H.; Tsuchiya, Y. Appl. Phys. Lett. 2002, 80, 4265-
4267.

Brixner, T.; Damrauer, N. H.; Niklaus, P.; Gerber, G. Nature 2001, 414, 57-60.

Brixner, T.; Damrauer, N. H.; Krampert, G.; Niklaus, P.; Gerber, G. J. Opt. Soc.
Am. B 2003, 20, 878-881.

Weinacht, T. C.; White, J. L.; Buckbaum, P. H. J Phys. Chem. A 1999, 103,
10166-10168.

Weinacht, T. C.; Bartels, R.; Backus, S.; Buckbaum, P. H.; Pearson, B. J.;
Geremia, J. M.; Rabitz, H.; Kapteyn, H. C.; Mumane, M. M. Chem. Phys. Lett.
2001, 344, 333-338.

Bartels, R. A.; Weinacht, T. C.; Leone, S. R.; Kapteyn, H. C.; Mumane, M. M.
Phys. Rev. Lett. 2002, 88, Art. No. 033001.

168

 

 

 

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)
(60)

(61)

(62)

(63)

(64)

(65)

Zeidler, D.; Frey, S.; Wohlleben, W.; Motzkus, M.; Busch, F.; Chen, T.; Kiefer,
W.; Matemy, A. J. Chem. Phys. 2002, 116, 5231-5235.

Herek, J. L.; Wohlleben, W.; Cogdell, R. J.; Zeidler, D.; Motzkus, M. Nature
2002, 417, 533-535.

Bartels, R.; Backus, S.; Zeek, E.; Misoquti, L.; Vdovin, G.; Christov, I. P.;
Murnane, M. M.; Kapteyn, H. C. Nature 2000, 406, 164-166.

Vajda, S.; Bartelt, A.; Kaposta, E. C.; Leisner, T.; Lupulescu, C.; Minemoto, S.;
Rosendo-Francisco, P.; Woste, L. Chem. Phys. 2001, 26 7, 231-239.

Schafer-Bung, B.; Mitric, R.; Bonacic-Koutecky, V.; Bartelt, A.; Lupulescu, C.;
Lindinger, A.; Vajda, S.; Weber, S. M.; Woste, L. J. Phys. Chem. A 2004, 108,
4175-4179.

Pearson, B. J .; White, J. L.; Weinacht, T. C.; Buckbaum, P. H. Phys. Rev. A 2001,
63, Art. No. 063412.

Omenetto, F. G.; Taylor, A. J .; Moores, M. D.; Reitze, D. H. Opt. Lett. 2001, 26,
938-940.

Kunde, J.; Baumann, B.; Arlt, S.; Morier-Genoud, F .; Siegner, U.; Keller, U.
Appl. Phys. Lett. 2000, 77, 924-926.

Homung, T.; Meier, R.; de Vivie-Riedle, R.; Motzkus, M. Chem. Phys. 2001, 26 7,
261-276.

Meshulach, D.; Silberberg, Y. Nature 1998, 396, 239-242.
Meshulach, D.; Silberberg, Y. Phys. Rev. A 1999, 60, 1287-1292.

Buist, A. H.; Muller, M.; Ghauharali, R. I.; Brakenhoff, G. J.; Squier, J. A.;
Bardeen, C. J .; Yakovlev, V. V.; Wilson, K. R. Opt. Lett. 1999, 24, 244-246.

Dudovich, N.; Dayan, B.; Feaeder, S. M. G.; Silberberg, Y. Phys. Rev. Lett. 2001,
86, 47-50.

Dudovich, N.; Oron, D.; Silberberg, Y. Phys. Rev. Lett. 2002, 88, Art. No.
123004.

Ballard, J. B.; Stauffer, H. U.; Amitay, Z.; Leone, S. R. J. Chem. Phys. 2002, 116,
1350-1360.

Walowicz, K. A.; Pastirk, I.; Lozovoy, V. V.; Dantus, M. J. Phys. Chem. A 2002,
106, 9369-9373.

169

(66)

(67)
(68)

(69)

(70)

(71)
(72)

(73)
(74)
(75)

(76)

(77)

(78)
(79)

(30)

(81)
(82)

(83)

Lozovoy, V. V.; Pastirk, 1.; Walowicz, K. A.; Dantus, M. J. Chem. Phys. 2003,
118, 3187-3196.

Bardeen, C. J.; Shank, C. V. Chem. Phys. Lett. 1993, 203, 535-539.
Antoine, P.; LHuillier, A.; Lewenstein, M. Phys. Rev. Lett. 1996, 77, 1234-1237.

Antoine, P.; Milosevic, D. B.; LHuillier, A.; Gaarde, M. B.; Salieres, P.;
Lewenstein, M. Phys. Rev. A 1997, 56, 4960-4969.

Agostini, P.; DiMauro, L. F. Rep. Prog. Phys. 2004, 67, 813-855.
Trebino, R.; Kane, D. J. J. Opt. Soc. Am. A 1993, 10, 1101-1111.

Trebino, R.; DeLong, K. W.; Fittinghoff, D. N.; Sweetser, J. N.; Krumbugel, M.
A.; Richman, B. A.; Kane, D. J. Rev. Sci. Instrum. 1997, 68, 3277-3295.

Iaconis, C.; Walmsley, I. A. Opt. Lett. 1998, 23, 792-794.
Iaconis, C.; Walmsley, I. A. IEEE J. Quantum Elect. 1999, 35, 501-509.
Siegner, U.; Haiml, M.; Kunde, J .; Keller, U. Opt. Lett. 2002, 2 7, 315-317.

Monmayrant, A.; Joffre, M.; Oksenhendler, T.; Herzog, R.; Kaplan, D.; Tournois,
P. Opt. Lett. 2003, 28, 278-280.

Comstock, M.; Lozovoy, V. V.; Pastirk, 1.; Dantus, M. Opt. Express 2004, 12,
1061-1066.

Rabitz, H. A.; Hsieh, M. M.; Rosenthal, C. M. Science 2004, 303, 1998-2001.
Moulton, P. F. J. Opt. Soc. Am. B 1986, 3, 125-133.

Sanchez, A.; Fahey, R. E.; Strauss, A. J.; Aggarwal, R. L. Opt. Lett. 1986, 11,
363-364.

Kuizenga, D. J .; Siegman, A. E. IEEE J. Quantum Elect. 1970, 6, 694-698.

Spence, D. E.; Evans, J. M.; Sleat, W. E.; Sibbett, W. Opt. Lett. 1991, 16, 1762-
1764.

Asaki, M. T.; Huang, C. P.; Garvey, D.; Zhou, J. P.; Kapteyn, H. C.; Mumane, M.
M. Opt. Lett. 1993, 18, 977-979.

170

 

(84) Spielmann, C.; Curley, P. F.; Brabec, T.; Krausz, F. IEEE J. Quantum Elect.
1994, 30, 2431-2431.

(85) Backus, S.; Peatross, J .; Huang, C. P.; Mumane, M. M.; Kapteyn, H. C. Opt. Lett.
1995, 20, 2000-2002.

(86) Maine, P.; Strickland, D.; Bado, P.; Pessot, M.; Mourou, G. IEEE J. Quantum
Elect. 1988, 24, 398-403.

(87) Backus, S.; Durfee, C. G.; Mumane, M. M.; Kapteyn, H. C. Rev. Sci. Instrum.
1998, 69, 1207-1223.

(88) Heritage, J. P.; Weiner, A. M.; Thurston, R. N. Opt. Lett. 1985, 10, 609-611.
(89) Wefers, M. M.; Nelson, K. A. Opt. Lett. 1993, 18, 2032-2034.

(90) Weiner, A. M. Rev. Sci. Instrum. 2000, 71, 1929-1960.

(91) Dantus, M.; Lozovoy, V. V. Chem. Rev. 2004, 104, 1813-1860.

(92) Denk, W.; Strickler, J. H.; Webb, W. W. Science 1990, 248, 73-76.

(93) Cerullo, G.; Bardeen, C. J.; Wang, Q.; Shank, C. V. Chem. Phys. Lett. 1996, 262,
362-368.

(94) Bardeen, C. J.; Yakovlev, V. V.; Squier, J. A.; Wilson, K. R. J. Am. Chem. Soc.
1998, 120, 13023-13027.

(95) Rice, S. A. Nature 2001, 409, 422-426.

(96) Bardeen, C. J .; Rosenthal, S. J .; Shank, C. V. J. Phys. Chem. A 1999, 103, 10506-
10516.

(97) Lozovoy, V. V.; Pastirk, I.; Dantus, M. Opt. Lett. 2004, 29, 775-777.

(98) Wolfbeis, O. S.; Furlinger, E.; Kroneis, H.; Marsoner, H. Fresenius Z. Anal.
Chem. 1983, 314, 119-124.

(99) Tran-Thi, T. H.; Gustavsson, T.; Prayer, C.; Pommeret, S.; Hynes, J. T. Chem.
Phys. Lett. 2000, 329, 421-430.

(100) Tran-Th1, T. H.; Prayer, C.; Millie, P.; Uznanski, P.; Hynes, J. T. J. Phys. Chem.
A 2002, 106, 2244-2255.

(101) Schulman, S. G.; Vincent, W. R.; Underberg, W. J. M. J. Phys. Chem. A 1981, 85,
4068-4071.

171

 

(102)

(103)
(104)

(105)

(106)

(107)

(108)

(109)

(110)

(111)

(112)

(113)

(114)

(115)

(116)

Haugland, R. P.; Johnson, I. P. In Fluorescent and Luminescent Probes, 2nd ed.;
Mason, W. T., Ed.; Academic Press, 1999, pp 647.

Putnam, R. W. In Cell Physiology Source Book, 2nd ed.; Sperelakis, N., Ed.;
Academic Press, 1996, pp 738.
Konig, K. J. Microsc. - Oxford 2000, 200, 83-104, Part 102.

So, P. T. C.; Dong, C. Y.; Masters, B. R.; Berland, K. M. Annu. Rev. Biomed.
Eng. 2000, 2, 399-429.

Dunn, K. W.; Sandoval, R. M.; Kelly, K. J.; Dagher, P. C.; Tanner, G. A.;
Atkinson, S. J .; Bacallao, R. L.; Molitoris, B. A. Am. J. Physio]. -Cell Ph. 2002,
283, C905-C916.

Wang, J. W.; Wong, A. M.; Flores, J.; Vosshall, L. B.; Axel, R. Cell 2003, 112,
271-282.

Oertner, T. G. Exp. Physio]. 2002, 87.

Kasischke, K. A.; Vishwasrao, H.; Heikal, A. A.; Webb, W. W. Biophys. J. 2002,
82, 495A-495A 2420 Part 2422.

Rose, C. R.; Kovalchuk, Y.; Eilers, J.; Konnerth, A. Pfiug. Arch. Eur. J. Phy.
1999, 439, 201-207.

Dela Cruz, J. M.; Pastirk, I.; Lozovoy, V. V.; Walowicz, K. A.; Dantus, M. J.
Phys. Chem. A 2004, 108, 53-58.

C540 is 3-(2'-Benzothiazoyl)-7-diethylaminocoumarin. R60 is a mixture of
rhodamine 590 (2-[6-(ethylamino)-3-(ethylimino)-2, 7-dimethyl-3H-xanthen-9-ylj-
ethylester, monohydrochloride), DOCI (3,3 '-Diethy]oxacarbocyanine Iodide) and
DCM (4-Dicyanmethylene-2-methyl-6-(p-dimethylaminosmyl)-4H-pyran).

Wolleschensky, R.; Feurer, T.; Sauerbrey, R.; Simon, 1. Appl. Phys. B-Lasers O
1998, 6 7, 87-94.

Bems, M. W.; Wang, Z.; Dunn, A.; Wallace, V.; Venugopalan, V. Proc. Nat].
Acad. Sci. USA 2000, 97, 9504-9507.

Shafirovich, V.; Dourandin, A.; Luneva, N. P.; Singh, G; Kirigin, F.; Geacintov,
N. E. Photochem. Photobiol. 1999, 69, 265-274.

Ogilvie, J. P.; Debarre, D.; Solinas, X.; Martin, J.; Beaurepaire, E.; Joffre, M. Opt.
Express 2006, 14, 759-766.

172

(117)

(118)

(119)
(120)

(121)

(122)

(123)

(124)

(125)

(126)
(127)
(128)
(129)
(130)

(131)

(132)

(133)

(134)

Larson, D. R.; Zipfel, W. R.; Williams, R. M.; Clark, S. W.; Bruchez, M. P.;
Wise, F. W.; Webb, W. W. Science 2003, 300, 1434-1436.

VandenBout, D. A.; Yip, W. T.; Hu, D. H.; Fu, D. K.; Swager, T. M.; Barbara, P.
F. Science 1997, 277, 1074-1077.

Xu, C.; Denk, W. Appl. Phys. Lett. 1997, 71, 2578-2580.
Osborn, D. L.; Leone, S. R. J. Appl. Phys. 2001, 89, 626-633.

Maiti, S.; Shear, J. B.; Williams, R. M.; Zipfel, W. R.; Webb, W. W. Science
1997, 275, 530-532.

Yelin, D.; Oron, D.; Korkotian, E.; Sega], M.; Silberberg, Y. Appl. Phys. B-Lasers
O 2002, 74, S97-S101.

Watanabe, T.; Akiyama, M.; Totani, K.; Kuebler, S. M.; Stellacci, F.; Wenseleers,
W.; Braun, K.; Marder, S. R.; Perry, J. W. Adv. Funct. Mat. 2002, 12, 611-614.

Dyba, M.; Hell, S. W. Phys. Rev. Lett. 2002, 88, Art. No. 163901.

Huang, D.; Swanson, E. A.; Lin, C. P.; Schuman, J. S.; Stinson, W. G.; Chang,
W.; Hee, M. R.; Flotte, T.; Gregory, K.; Puliafito, C. A.; Fujimoto, J. G. Science
1991, 254, 1178-1181.

Zare, R. N. Science 1998, 279, 1875-1879.

Gordon, R. J .; Rice, S. A. Annu. Rev. Phys. Chem. 1997, 48, 601-641.

Rice, S. A.; Shah, S. P. Phys. Chem. Chem. Phys. 2002, 4, 1683-1700.

Rabitz, H. Science 2003, 299, 525-527.

Denk, W. .1. Biomed. Opt. 1996, I, 296-304.

Fisher, W. G.; Partridge, W. P.; Dees, C.; Wachter, E. A. Photochem. Photobiol.
1997, 66,141-155.

Jasapara, J .; Rudolph, W. Opt. Lett. 1999, 24, 777-779.

Amat-Roldan, 1.; Cormack, I. G; Loza-Alvarez, P. ,Artigas, D. Opt. Lett. 2004,
29, 2282- 2284.

Fittinghoff, D. N.; Millard, A. C.; Squier, J. A.; Muller, M. IEEE J. Quantum
Elect. 1999, 35, 479-486.

173

(135)

(136)

(137)

(138)
(139)

(140)

(141)

(142)

(143)

(144)
(145)
(146)
(147)

(148)

(149)

(150)

(151)

(152)

Muller, M.; Squier, J. A.; Wolleschensky, R.; Simon, U.; Brakenhoff, G. J. J.
Microsc. - Oxford 1998, 191, 141-150.

Fittinghoff, D. N.; Squier, J. A.; Barty, C. P. J.; Sweetser, J. N.; Trebino, R.;
Muller, M. Opt. Lett. 1998, 23, 1046-1048.

Dela Cruz, J. M.; Lozovoy, V. V.; Dantus, M. J. Photochem. Photobiol. A, 2006
180, 307-313.

Lozovoy, V. V.; Dantus, M. ChemPhysChem 2005, 6, 1970-2000.
Weinacht, T. C.; Buckbaum, P. H. J. Opt. B-Quantum S. O. 2002, 4, R35-R52.

Pastirk, I.; Dela Cruz, J. M.; Walowicz, K. A.; Lozovoy, V. V.; Dantus, M. Opt.
Express 2003, 11, 1695-1701.

Yoo, K. M.; Alfano, R. R. Opt. Lett. 1990, 15, 320-322.

Hee, M. T.; Izatt, J. A.; Jacobson, J. M.; Fujimoto, J. G.; Swanson, E. A. Opt.
Lett. 1993, 18, 950-952.

Parys, B.; Allard, J. F.; Desmullier, F.; Houde, D.; Cornet, A. J. Opt. A-Pure
App]. Op. 2004, 6, L23-L27.

Beaurepaire, E.; Oheim, M.; Mertz, J. Opt. Commun. 2001, 188, 25-29.
Theer, P.; Hasan, M. T.; Denk, W. Opt. Lett. 2003, 28, 1022-1024.

Koester, H. J .; Baur, D.; Uhl, R.; Hell, S. W. Biophys. J. 1999, 77, 2226-2236.
Hopt, A.; Neher, E. Biophys. J. 2001, 80, 2029-2036.

Szmacinski, H.; Gryczynski, 1.; Lakowicz, J. R. Biospectroscopy 1998, 4, 303-
310.

Konig, K.; So, P. T. C.; Mantulin, W. W.; Gratton, E. Opt. Lett. 1997, 22, 135-
136.

Cunningham, M. L.; Johnson, J. S.; Giovanazzi, S. M.; Peak, M. J. Photochem.
Photobiol. 1985, 42, 125-128.

Kilic, H. S.; Ledingham, K. W. D.; Kosmidis, C.; McCanny, T.; Singhal, R. P.;
Wang, S. L.; Smith, D. J.; Langley, A. J.; Shaikh, W. J. Phys. Chem. A 1997, 101,
817-823.

Szaflarski, D. M.; ElSayed, M. A. J. Phys. Chem. 1988, 92, 2234-2239.

174

 

(153)

(154)

(155)

(156)

(157)

(158)

(159)
(160)
(161)
(162)

(163)

(164)
(165)

(166)

(167)

(168)

In most 70 eV E1 experiments,approximately 1400 kJ/mole (I5 e10 of energy is
transferred duringthe ionization process. There is, however, a distribution of
energy and as much as 2800 kJ/mole (30 eV) is transferred to some molecules.
Since approximately 960 kJ/mole (10 e10 is required to ionize most organic
compounds and a typical chemical bond energy is 290 1:1/mole (3 e10, extensive
fiagmentation is often observed in 70 eVEI mass spectra.

Harless, R. L.; Oswald, E. O.; Wilkinson, M. K.; Dupuy, A. E.; McDaniel, D. D.;
Tai, H. Anal. Chem. 1980, 52, 1239-1245.

Hawthorne, S. B.; Langenfeld, J. J.; Miller, D. J .; Burford, M. D. Anal. Chem.
1992, 64, 1614-1622.

Axelson, M.; Sahlberg, B. L.; Sjovall, J. .1. Chromatogr. 1981, 224, 355-370.

Burlingame, A. L.; Shackleton, C. H. L.; Howe, 1.; Chizhov, O. S. Anal. Chem.
1978, 50, R346-R384.

Allison, J.; Holland, J. F .; Enke, C. G.; Watson, J. T. Ana]. Instrum. 1987, 16,
207-224.

Bruckner, C. A.; Prazen, B. J .; Synovec, R. E. Anal. Chem. 1998, 70, 2796-2804.
Wu, H. F .; Chuan, Y. J. Rapid Commun. Mass Sp. 2003, 1 7, 1030-1036.

Wu, H. F.; Wu, W. F. Rapid Commun. Mass Sp. 2003, 17, 2399-2406.
Kascheres, C.; Cooks, R. G. Ana]. Chimica Acta 2002, 215, 223-232.

Zimmermann, R.; Lerner, C.; Schramm, K.; Kettrup, A.; Boesl, U. Eur. J. Mass
Spectrom. 1995, I, 341-351.

Brodbelt, J. S.; Liou, C. C.; Donovan, T. Anal. Chem. 1991, 63, 1205—1209.
Bjarnason, A. Anal. Chem. 1996, 68, 3882-3883.

Dantus, M.; Jannsen, M. H. M.; Zewail, A. H. Chem. Phys. Lett. 1991, 181, 281-
287.

Baumert, T.; Buhler, B.; Thalweiser, R.; Gerber, G. Phys. Rev. Lett. 1990, 64,
733-736.

Cardoza, D.; Baertschy, M.; Weinacht, T. C. Chem. Phys. Lett. 2005, 411, 311-
315.

175

(169)

(170)

(171)

(172)

(173)
(174)
(175)
(176)

(177)

(178)

(179)

(180)

(181)

(182)

(183)

(184)
(185)

(186)

Pastirk, 1.; Kangas, M.; Dantus, M. J. Phys. Chem. A 2005, 109, 2413-2416.

Xu, B.; Gunn, J. M.; Dela Cruz, J. M.; Lozovoy, V. V.; Dantus, M. J. Opt. Soc.
Am. B 2006, 23, 750-759.

Stein, S. E.; NIST Mass Spec Data Center In NIST Chemistry Web Book, NIST
Standard Reference Database Number 69; Linstrom, P. J ., Mallard, W. G., Eds;
National Institute of Standards and Technology: Gaithersburg, MD, 2005.

Malow, M.; Penno, M.; Weitzel, K. M. J. Phys. Chem. A 2003, 107, 10625-
10630.

Gedanken, A.; Robin, M. B.; Kuebler, N. A. J. Phys. Chem. 1982, 86, 4096-4107.
Yang, J. J .; Gobeli, D. A.; ElSayed, M. A. .1. Phys. Chem. 1985, 89, 3426-3429.
Meek, J. T.; Jones, R. K.; Reilly, J. P. J. Chem. Phys. 1980, 73, 3503-3505.
Zandee, L.; Bernstein, R. B. J. Chem. Phys. 1979, 70, 2574-2575.

Rockwood, S.; Reilly, J. P.; Hohla, K.; Kompa, K. L. Opt. Commun. 1979, 28,
175-178.

Fisarrick, G. J.; Eichelberger, T. S.; Heath, B. A.; Robin, M. B. J. Chem. Phys.
1980, 72, 5571-5580.

Robin, M. B. Appl. Opt. 1980, 19, 3941-3947.

Kosmidis, C.; Ledingham, K. W. D.; Kilic, H. S.; McCanny, T.; Singhal, R. P.;
Langley, A. J .; Shaikh, W. J. Phys. Chem. A 1997, 101, 2265-2270.

Lezius, M.; Blanchet, V.; Ivanov, M. Y.; Stolow, A. J. Chem. Phys. 2002, 117,
1575-1588.

Smith, S. M.; Markevitch, A. N.; Romanov, D. A.; Li, X. S.; Levis, R. J.;
Schlegel, H. B. J. Phys. Chem. A 2004, 108, 11063-11072.

Pirkle, W. H.; Pochapsky, T. C.; Giddings, J. C., Grushka, E., Brown, P. R., Eds.;
Marcel Dekker, Inc.: New York, 1987; Vol. 27, pp 73-127.

Taylor, D. R.; Maher, K. J. Chromatogr. Sci. 1992, 30, 67-85.
Gubitz, G. Chromatographia 1990, 30, 555-564.

Holland, L. A.; Chetwyn, N. P.; Perkins, M. D.; Lunte, S. M. Pharm. Res. 1997,
14.

176

(187)

(188)
(189)

(190)

(191)
(192)
(193)
(194)
(195)
(196)
(197)

(198)

(199)

(200)

(201)
(202)
(203)
(204)
(205)
(206)

(207)

Altria, K. D.; Elgey, J.; Lockwood, P.; Moore, D. Chromatographia 1996, 42,
332-342.

Rieutord, A.; Prognon, P.; Mahuzier, G. Analusis 1996, 24, 349-360.
Le Bel, J. A. Bull. Soc. Chim. France 1874, 22, 337.

Van't Hoff, J. H. Die Lagerung der Atome im Raume, 2nd ed.; Vieweg:
Braunschweig, 1894.

Rau, H. Chem. Rev. 1983, 83, 535-547.

Inoue, Y. Chem. Rev. 1992, 92, 741-770.

Mitchell, S. J. Chem. Soc. 1930, Part 2, 1829-1834.

Mitchell, 8.; Dawson, 1. M. J. Chem. Soc. 1944, Part 2, 452-454.
Stevenso.K.L.; Verdieck, J. F. J. Am. Chem. Soc. 1968, 90, 2974-2978.
Nicoud, J. F.; Kagan, H. B. Israel J. Chem. 1977, 15, 78-81.

Ward, R. S. T etrahedron-Asymmetr. 1995, 6, 1475-1490.

Balavoin, G.; Moradpou, A.; Kagan, H. B. J. Am. Chem. Soc. 1974, 96, 5152-
5158.

Brixner, T.; Gerber, G. Opt. Lett. 2001, 26, 557-559.

Brixner, T.; Krampert, G.; Niklaus, P.; Gerber, G. Appl. Phys. B-Lasers O 2002,
74, $133-$144.

Polachek, L.; Oron, D.; Silberberg, Y. Opt. Lett. 2006, 31, 631-633.
Shapiro, M.; Brumer, P. J. Chem. Phys. 1991, 95, 8658-8661.

Shapiro, M.; Frishman, E.; Brumer, P. Phys. Rev. Lett. 2000, 84, 1669-1672.
Gerbasi, D.; Shapiro, M.; Brumer, P. J. Chem. Phys. 2001, 115, 5349-5352.
Salam, A.; Meath, W. J. J. Chem. Phys. 1997, 106, 7865-7868.

Salam, A.; Meath, W. J. Chem. Phys. 1998, 228, 115-129.

Asaro, C.; Brumer, P.; Shapiro, M. Phys. Rev. Lett. 1988, 60, 1634-1637.

177

 

 

(208)

(209)

(210)
(211)
(212)

(213)

(214)

(215)

(216)

(217)

(218)
(219)
(220)

(221)

Fujimura, Y.; Gonzales, L.; Hoki, K.; Manz, J.; Ohtsuki, Y. Chem. Phys. Lett.
1999, 310.

Gonzalez, L.; Hoki, K.; Kroner, D.; Leal, A. S.; Manz, J.; Ohtsuki, Y. J. Chem.
Phys. 2000,113, 11134-11142.

Kroner, D.; Shibl, M. F.; Gonzales, L. Chem. Phys. Lett. 2003, 372, 242-248.
Kroner, D.; Gonzales, L. Chem. Phys. Lett. 2004, 298, 55-63.

Wefers, M. M.; Nelson, K. A. J. Opt. Soc. Am. B 1995, 12, 1343-1362.

Brixner, T.; Krampert, G.; Pfeifer, T.; Selle, R.; Gerber, G.; Wollenhaupt, M.;
Graefe, O.; Horn, C.; Liese, D.; Baumert, T. Phys. Rev. Lett. 2004, 92, Art. No.
208301.

Nelson, R. D.; Leaird, D. E.; Weiner, A. M. Opt. Express 2003, 11, 1763-1769.

Oron, D.; Dudovich, N.; Silberberg, Y. Phys. Rev. Lett. 2003, 90, Art. No.
213902.

Krishnamurthy, M.; Mathur, D. Phys. Rev. A 2000, 61, Art. No. 063404.

Nikogosyan, D. N.; Repeyev, Y. A.; Khoroshilova, E. V.; Kryukov, I. V.;
Khoroshilov, E. V.; Sharkov, A. V. Chem. Phys. 1990, 14 7, 437-445.

Fork, R. L.; Martinez, O. E.; Gordon, J. P. Opt. Lett. 1984, 9, 150-152.
Stingl, A.; Spielmann, C.; Krausz, F .; Szipocs, R. Opt. Lett. 1994, 19, 204-206.
Szipocs, R.; Ferencz, K.; Spielmann, C.; Krausz, F. Opt. Lett. 1994, 19, 201-203.

Nisoli, M.; DeSilvestri, S.; Svelto, O.; Szipocs, R.; Ferencz, K.; Spielmann, C.;
Sartania, S.; Krausz, F. Opt. Lett. 1997, 22, 522-524.

178