.4. 2..., at m
p ¢ 01‘0“"

I fit a a
r 3.410.
2

 

“I. my. '
““32: "

u in:
hit 343 t
5.1.»! 5.: t.
«u ‘

P

\-'\JV

Il‘..

3.. ‘ Wagfi... ..
. .6 11h 11‘:

0.9....f .7. ..«l.\ltvl.tfl .‘ . . w .
3. q 32‘. .

stmi , mfifi

 

 

[him

This is to certify that the
dissertation entitled

FEMTOSECOND SPECTROSCOPIC STUDIES OF
VIBRATIONAL COHERENCE FROM
BACTERIOCHLOROPHYLL SOLUTIONS AND LIGHT-
HARVESTING PROTEINS

presented by

Katherine R. Shelly

has been accepted towards fulfillment
of the requirements for the

Ph. D. degree in Chemistry

 

 

WW 476%

 

Major Professor’s Signature

Z 7741.4 2005

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution

Liam?
Michigan State
University

 

 

 

 

o-c-o-I-n-o—-oo.-—o—-o-o-o-o-o-o-n-o-o-o-a-o-o-o-o-o...- . -A A

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:/C|RC/DateDue.indd-p.1

 

FEMTOSECOND SPECTROSCOPIC STUDIES OF VIBRATIONAL COHERENCE
FROM BACTERIOCHLOROPHYLL SOLUTIONS AND LIGHT-HARVESTING
PROTEINS
By

Katherine R. Shelly

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department Of Chemistry

2006

ABSTRACT
By

Katherine R. Shelly

The femtosecond dynamic-absorption technique was used to Obtain pump-probe
transients with impulsive excitation of the Qy absorption band at room temperature in
polar and non-polar solutions of bacteriochlorophyll a. The dynamics of
bacteriochlorophyll a in protein-derived binding sites was characterized in experiments
with the B777 and 8820 subunits of the LP“ light-harvesting complex from
Rhodospirillum rubrum G9. In all of the systems with strong axial ligation, from
pyridines in solution or histidine residues in the protein, two time regions of vibrational
coherence were Observed: a rapidly damped portion, in the < 800-fs time region, and a
slowly damped portion, on the < 8-ps time scale. Only the rapidly damped portion was
observed in acetone or l-propanol solutions. The slowly damped portion is assigned to
narrow line shapes arising from intramolecular vibrations of the BChl macrocycle. The
rapidly damped portion is assigned to inhomogeneously broadened line shapes arising
from intermolecular modes with polar and non-polar components in the surrounding

solvent or protein environment.

To BI

for your unconditional encouragement and patience

iii

ACKNOWLEDGMENTS

The work I have completed at Michigan State University is a result of collective
efforts from my parents, family, and friends. I have received infinite support from the
people who surround me, for which I would have never made it through the last five
grueling years without.

This dissertation is dedicated first and foremost to my husband, B], for giving me the
confidence to make it through this tumultuous endeavor. Your encouragement was and
still is unmatched by any other. This degree is certainly as much yours as it is mine.

I also want to thank my graduate colleagues. You have made this experience one I
will never forget. I will not only walk away from here with my degree, but I will have
your friendship, which I value even more.

Finally, I would like to thank Professor Warren Beck. It is only with your guidance
that I close this chapter of my life. You have provided me with the opportunities and
knowledge that made my graduate experience more fulfilling than I could have ever

imagined.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ........................................................................................................... ix
INTRODUCTION ............................................................................................................... 1
CHAPTER 1
INTRODUCTION .................................................................................................................... 3
1.0 Photosynthetic Reaction Centers ............................................................................. 3
1.0.1 Charge-Transfer Reactions in Photosynthetic Reaction Centers .................... 3
1.0.2 Structure of the Photosynthetic Reaction Center ............................................ 4
1.1 Electron-Transfer Theory ......................................................................................... 7
1.2 Coupled Modes in Photosynthesis ......................................................................... 10
1.3 Assignments for the Coupled Modes ..................................................................... 12
1.4 Resonance Raman Activity in Intermolecular Modes ........................................... 13
1.5 Research Design ..................................................................................................... 15
1.6 References .............................................................................................................. 16
CHAPTER 2
THEORY AND EXPERIMENTAL METHODS FOR THE STUDY OF VIBRATIONAL
COHERENCE .......................................................................................................................... 22
2.0 Introduction ........................................................................................................... 22
2.1.Theory .................................................................................................................... 22
2.1.1 Wave Packets ................................................................................................ 22
2.1.2 Resonance Raman Spectroscopy ................................................................... 23
2.1.3 Dynamic-Absorption Spectroscopy .............................................................. 25
2.1.4 Frequency Domain vs. Time Domain ........................................................... 26

2.2 Experimental: F emtosecond Pump-Probe Spectrometer Design and

Characterization ..................................................................................................... 27
2.2.1 Ti:Sapphire Oscillator ................................................................................... 27
2.2.2 Extracavity Pulse Compression .................................................................... 31
2.2.3 Rapid-Scanning Modified Mach—Zehnder Interferometer ........................... 31
2.2.3.1 Layout of Interferometer ...................................................................... 31

2.3 Data Acquisition .................................................................................................... 37
2.4 References .............................................................................................................. 42

CHAPTER 3
VIBRATIONAL COHERENCE FROM THE DIPYRIDINE COMPLEX OF
BACTERIOCHLOROPHYLL a: INTRAMOLECULAR MODES IN THE 10—220-cm'l REGIME,

INTERMOLECULAR SOLVENT MODES, AND RELEVANCE TO PHOTOSYNTHESIS ................ 45
3.0 Introduction ............................................................................................................ 45
3.1 Experimental Procedures ....................................................................................... 46
3.1.1 Sample Preparation ....................................................................................... 46
3.1.2 Continuous-Wave Spectroscopy ................................................................... 47
3.1.3 Femtosecond Spectroscopy ........................................................................... 47
3.2 Results .................................................................................................................... 47
3.2.1 Continuous-Wave Spectroscopy ................................................................... 47
3.2.2 Dynamic-Absorption Experiments ............................................................... 49
3.2.3 Rapidly-Damped Vibration Coherence from Bacteriochlorophyll a and
Pyridine ......................................................................................................... 51
3.3 Data Analysis ......................................................................................................... 57
3.3.1 Fourier-Magnitude Spectrum Estimation ..................................................... 57
3.3.2 Modeling of the Slowly Damped Components ............................................. 60
3.3.3 Modeling Of the Rapidly Damped Component ............................................. 60
3.4 Discussion .............................................................................................................. 62
3.3.1 Observation of Two Damping Time Scales .................................................. 62
3.5 References .............................................................................................................. 67
CHAPTER 4
INTERMOLECULAR VIBRATIONAL COHERENCE BETWEEN BACTERIOCHLOROPHYLL 0
AND POLAR AND NONPOLAR SOLVENTS ............................................................................ 70
4.0 Introduction ............................................................................................................ 70
4.1 Experimental Procedures ....................................................................................... 70
4.1.1 Sample Preparation ....................................................................................... 70
4.1.1.1 Bacteriochlorophyll a in polar solvents ............................................... 70
4.1.1.2 Bacteriochlorophyll a in nonpolar solvents ......................................... 71
4.1.2 Continuous-Wave Spectroscopy ................................................................... 71
4.1.3 Femtosecond Spectroscopy ........................................................................... 71
4.2 Results .................................................................................................................... 72
4.2.1 Continuous-Wave Spectroscopy ................................................................... 72

vi

4.2.2 Dynamic-Absorption Experiments ............................................................... 78

4.3 Discussion .............................................................................................................. 91
4.3.1 Structural Assignments ................................................................................. 92
4.3.2 Line Shapes for Intermolecular and Intramolecular Modes in Solution and
Proteins ....................................................................................................... .98
4.4 References ............................................................................................................ 100
CHAPTER 5
INTERMOLECULAR VIBRATIONAL COHERENCE FROM BACTERIOCHLOROPHYLL-
CONTAINING PROTEINS, B777 AND 3820 ........................................................................... 103
5.0 Introduction .......................................................................................................... 106
5.1 Experimental Procedures ..................................................................................... 106
5.1.] Sample Preparation ..................................................................................... 106
5.1.1.1 Isolation and Preparation of B820 Subunits ...................................... 109
5.1.1.2 Isolation and Preparation of B777 Subunits ...................................... 112
5.1.2 Continuous-Wave Spectroscopy ................................................................. 112
5.1.3 Femtosecond Spectroscopy ......................................................................... 1 12
5.2 Results .................................................................................................................. 113
5.2.1 Continuous-Wave Spectroscopy ................................................................. 113
5.2.2 Dynamic-Absorption Experiments ............................................................. 116
5.3 Discussion ............................................................................................................ 123
5.3.1 Comparison of Rapidly-Damped Vibrational Coherence in B777 and
8820 ............................................................................................................ 123
5.3.2 Assignments for Components in the Vibrational Coherence in B777 and
B820 ............................................................................................................ 126
5.3.3 BChl—Protein Interactions in B820 and in the Photosynthetic Reaction
Center .......................................................................................................... 127
5.4 Conclusions and Future Work ............................................................................. 130
5.5 References ............................................................................................................ 134

vii

Table
3.1

3.2

4.1

5.1

5.2

5.3

LIST OF TABLES

Frequencies, normalized deconvolved intensities, and damping
constants for the underdamped oscillatory components observed in the
pump-probe signal from bacteriochlorophyll a in pyridine solvent

Log-normal line shape parameters for the solvent-mode distribution

Model parameters for the rapidly damped vibrational coherence
Observed in BChl a solutions

Model parameters for the rapidly damped vibrational coherence
observed in BChl-containing proteins

Potential residue candidates for the interaction between the BChl dimer
and the surrounding protein medium in 3820

Potential residue candidates for the interaction between the BChl dimer
and the surrounding protein medium in Rhodobacter sphaeroides

viii

Page
61

64

90

122

129

132

Figure
1.1

1.2

1.3

2.1

2.2

2.3

2.4

2.5

LIST OF FIGURES

Page
The photosynthetic reaction center from Rhodopseudomonas viridis,
rendered from RCSB Protein Data Bank file laij. This view shows the
topology of the cytochrome (green, top), L (blue), M (pink), and H
(green, bottom) subunits 5

Chromophores in the reaction center from Rhodopseudomonas viridis.
The arrows represent the direction of the electron transfer reactions 6

Diabatic harmonic potential-energy curves for the reactant (E3) and
product (Eb) of an electron-transfer reaction as a function of the
displacement of one of the normal modes of vibration n; ’1" marks the

mode-specific reorganization energy, and —AE° is the driving force

Excited-state and ground-state coherent wave-packet motion in the
dynamic absorption experiment. The excited-state and ground-state
potential-energy surfaces are drawn as parabolas that are displaced with
respect to a generalized multimode coordinate; rg and re mark the
equilibrium ground-state and excited-state geometries, respectively.
Thick arrows represent the resonant pump-laser field; thin arrows show
the direction that the wave packets evolve during the first passage on
the two surfaces. The numbers indicate event times, starting with
ground-state probability density, creation Of the excited-state wave
packet by the pump field, creation of the ground-state wave packet by
the pump field, and evolution during the first vibration 24

The Mira 900-F tizsapphire laser oscillator and pulse compressor

employed in the dynamic-absorption spectroscopy experiments.

Symbols: M1, output coupler; BP1-BP2, Brewster prisms; M2-M3, flat

cavity mirrors; BRF, birefringent filter; L1, focusing lens; M4-M5,

curved mirrors; TiAle3, titanium: sapphire crystal; M6, flat mirror;

M7, flat end mirrors; p1-p2, SFIO prisms 28

Output intensity spectrum of 60-fs (sechz; 12-nm (fwhm) spectral
width) pulses from the tizoscillator 29

Zero-background autocorrelation trace of a 60-fs (sechz) pulse Obtained
from the tizsapphire oscillator 30

Extracavity compressor employed in the dynamic-absorption
spectroscopy experiments. Symbols: Ml-M6, silver mirrors; M7, half
silver mirror; pl-p2, SF 10 Brewster-angled prisms 32

2.6

2.7

2.8

2.9

2.10

3.1

3.2

3.3

3.4

Modified Mach-Zehnder rapid-scanning interferometer employed in the
dynamic-absorption spectroscopy experiments. Symbols: M1-M9, silver
mirrors; BSl-BS2, beam splitters; C1-C2, compensation optics; M2,
half-wave plates; Pl-P2, Glan-laser calcite polarizers; FEM-90,
photoelastic modulator

Sample and detection portion of the Modified Mach-Zehnder rapid-
scanning interferometer employed in the dynamic-absorption
spectroscopy experiments. Symbols: M1-M6, silver mirrors; BSI , beam
splitters; C l , compensation optics; P1, Glan-laser calcite polarizers; x,
B-barium borate crystal; PMT, photomultiplier tube; PD, photodiode

Multiple traces Obtained simultaneously during dynamic-absorption
experiments of bacteriochlorophyll a in pyridine. The probe beam was
detected at 757 nm. Two complete scans of the rapid-scanning delay
stage are shown as a function of real time. Top: Pump-probe delay
determined from the position of the rapid-scanning delay stage. Middle:
Zero-background autocorrelation. Bottom: Single-wavelength dynamic-
absorption transient

Zoomed traces from figure 2.8. obtained from bacteriochlorophyll a in
pyridine. The probe beam was detected at 757 nm. A portion of one
scan of the pump-probe delay is shown as a function of real time. T op:
Probe delay as determined from the position of the rapid-scanning delay
stage. Middle: Zero-background autocorrelation. Bottom: Single-
wavelength dynamic-absorption transient

Single-wavelength dynamic-absorption transient of bacteriochlorophyll
a in pyridine solvent displayed as a function of probe delay. The probe
beam was detected at 757 nm

Continuous-wave absorption and fluorescence spectra from
bacteriochlorophyll a in pyridine plotted as relative dipole

strengths, A(v)/ v and F (v) / v3 , respectively. The spectra are

normalized to unit area. The fluorescence spectrum was excited at 790
nm (12700 cm")

Pump-probe signal Obtained from BChl in pyridine solvent

Triple exponential fit (top) and corresponding residual oscillatory part
of the pump-probe signal from BChl in pyridine (bottom)

Hanning-windowed Fourier-magnitude spectrum of the 200-8000-fs
range

33

34

38

40

41

48

50

52

53

3.5

3.6

3.7

4.1

4.2

4.3

4.4

4.5

4.6

Pump-probe signal obtained under the same optical conditions from
BChl in pyridine solvent and in neat pyridine. The top graph shows the
range corresponding to the population change in the BChl ground state.
The y axes employ a consistent relative-intensity scaling

Oscillatory signal with a frequency of 200 cm'l and a damping time of
100 fs (top panel). Oscillatory signal from top panel multiplied by a
Hanning window function described by equation 3.1(middle panel).
Hanning-windowed Fourier-magnitude spectrum of the ZOO-cm'I
oscillation (bottom panel)

Expanded view of the rapidly damped oscillation from BChl in pyridine
solvent (see figure 3.5.) superimposed with a model (dashed curve)
arising from a distribution of mode frequencies, as shown in the bottom
panel

Continuous-wave absorption spectra at 22 °C from BChl a in
cyclohexane solvent at BChlzpyridine ratios of(a)1:0.6, (b)1:1.2, and
(c)1:100. The vertical line represents 580 nm

Continuous-wave absorption spectra at 22 °C from BChI a in three neat
solvents: (a) pyridine,(b) 1-propanol, and (c) acetone

Structure of bacteriochlorophyll a molecule, the dominant chromophore
in purple non-sulfur photosynthetic light-harvesting proteins

Continuous-wave absorption and fluorescence spectra from BChl in
pyridine solvent plotted as the dipole strength, A(v)/v and F (v)/ v3 ,

respectively, and normalized to unit area. The dashed curve shows the
intensity spectrum emitted by the titanium-sapphire oscillator. The
shaded region shows the range of the transmitted probe light that was
passed by the monochromator to the detector in the dynamic-absorption
experiment

Dynamic-absorption transient from BChI in pyridine. The inset shows
an expanded view of the O—l-ps portion of the signal. The ordinate is
scaled relative to the magnitude of the pump-induced change in
transmission that follows the initial (0-5 ps) exponential decay arising
from solvation dynamics

Expanded view of the rapidly damped oscillation Observed in the
dynamic-absorption transient from BChl a in pyridine (see figure 4.5)
and three Optimized single-modulation-comlponent models employing
different line shapes: (a) a discrete 225-cm" cosinusoid (equation 4.1)
with the damping time 7= 92 fs, (b) a Gaussian distribution of damped
cosinusoids with a center frequency of 225 cm'l and a width (fwhm) of

xi

55

59

63

73

74

76

77

79

4.7

4.8

4.9

4.10

50 cm", and (c) a log-normal distribution of damped cosinusoids with a
center frequency of 285 cm'l and a width (fwhm) of 97 cm". The
models for (b) and (c) were defined by equation 4.2, with the
distributions [(60) set as a Gaussian and as a log-normal, respectively.
The scaling of the ordinate is relative to the magnitude of the pump—
probe ground-state depletion signal, as indicated in figure 4.5

Modeling the rapidly damped vibrational coherence Observed in BChl (1
solutions with a distribution of slowly damped cosinusoids, [(60), as
defined by equation 4.2. A log-normal distribution L(a)), (a), defined by
its center frequency, (00, width, Am, and asymmetry (or skew), p, is
sampled over its full width. Each sample defines the intensity of a
Lorentzian line shape that corresponds in the time domain, (b), to a
cosinusoid at the sampled frequency a) with a homogeneous (or
intrinsic) damping time, 7, here set arbitrarily to 1.5 ps. Summing of the
cosinusoids Obtained by sampling over the full width of the distribution
[,(w), (c), results in a rapidly damped waveform, (d), that resembles the
vibrational coherence observed in BChl solutions

Expanded view of the rapidly damped oscillation Observed in the
dynamic-absorption transient from BChl a in pyridine (see figures 4.5
and 4.6) superimposed with a model defined by the sum Of two
independent log-normal distributions L(a)) of damped cosinusoids, each
defined by equation 4.2. The scaling of the ordinate is relative to the
magnitude of the pump—probe ground-state depletion signal, as in figure
4.5. Bottom: Plots of £(w) for the two components Observed in pyridine

and their sum, 914(60) (thick curve). The model parameters are provided
in table 4.1

Expanded view of the rapidly damped oscillation observed in the
dynamic-absorption transient from BChl a in acetone superimposed
with a model defined by the sum of two independent log-normal

distributions [(0) of damped cosinusoids, each defined by equation 4.2.

The scaling Of the ordinate is relative to the magnitude of the pump—
probe ground-state depletion signal, as in figure 4.5. Bottom: Plots of
L(a)) for the two components Observed in acetone and their sum, 914(60)
(thick curve). The model parameters are provided in table 4.1

Expanded view of the rapidly damped oscillation Observed in the
dynamic-absorption transient from BChl a in l-propanol superimposed
with a model defined by the sum of three independent log-normal

distributions [.(w) Of damped cosinusoids, each defined by equation 4.2.

The scaling of the ordinate is relative to the magnitude of the pump-
probe ground-state depletion signal, as in figure 4.5. Bottom: Plots of
film) for the three cnmnnnents observed in l-nmnannl and their sum.

xii

81

82

85

86

4.11

4.12

4.13

5.1

5.2

5.3

5.4

5.5

L(w) for the three components observed in 1-propanol and their sum,
914(0)) (thick curve). The model parameters are provided in table 4.1

Expanded view of the rapidly damped oscillation observed in the
dynamic-absorption transient from BChl a with a single pyridine axial
ligand in cyclohexane superimposed with a model defined by the sum Of
two independent log-normal distributions L(a)) of damped cosinusoids,
each defined by equation 4.2. The scaling of the ordinate is relative to
the magnitude of the pump—probe ground-state depletion signal, as in
figure 4.5. Bottom: Plots of £(w) for the two components observed in

cyclohexane and their sum, 914(50) (thick curve)

A possible model obtained using a molecular mechanics calculation
with the MMX force field and the PCModel program for the interaction
between two non-ligated pyridine solvent molecules and the rt-electron
density Of the BChl macrocycle in the dipyridine complex

A possible model obtained using a molecular mechanics calculation
with the MMX force field and the PCModel program for the interaction
between acetone solvent dipoles and the n-electron density of the BChl
macrocycle

Model of the LHl light-harvesting complex generated from the
calculations of Schulten and coworkers. Only the Chromophores are
shown here. (Refer to figure 5.3 for the details of how each pair of BChl
are divided by a pair of membrane-separating 0t helices.)

Crystal structure of the LH2 complex from Rhodopseudomonas
acidophila as reported by Caffrey and coworkers. The a and [3 units are
shown in red and blue, respectively, the B800 pigments in purple, the
B850 pigments in green, and the carotenoids in yellow. The structure
can be found under the accession number 2FKW in the RCSB Protein
Data Bank

Model of the B820 subunit from LHl based on the calculations of
Schulten and coworkers

Hypothesized model for the B777 subunit formed from the B820
subunit of LH 1. LHI is the result of dissociating B820 into two
polypeptides, each with a single ligated BChl macrocycle, so there are
two possible B777 structures in a given sample

Absorption spectra illustrating the formation of B777 subunits from

detergent-solublized chromatophores isolated from R. rubrumG9, top
panel. At low concentrations of n-octyl-rac-2,3-dipropyl sulfoxide

xiii

88

89

95

96

104

105

107

108

5.6

5.7

5.8

5.9

5.10

5.11

(ODPS), only the chromatophores are present, as indicated by the
absorption maximum at 873 nm. At intermediate detergent
concentrations, chromatophores, B820 subunits and B777 subunits are
all present within the sample. The B820 population is indicated by the
formation of the absorption peak at 820 nm. B777 subunits are
exclusively present at an increased concentration of ODPS detergent
(~3.0% (w/v)). The bottom panel shows an expanded view of the Qy
absorption band, so that the conversion of the initial 870 nm peak to the
777 nm peak can be more easily observed

Absorption spectra illustrating the formation of B820 subunits from
detergent-solublized chromatophores isolated from R. rubrum G9 over a
period of 17 hours, top panel. The chromatophores were solublized with
0.75% (w/v) octyl [3 -D-glucopyranoside (0G). The bottom panel shows
an expanded view of the Qy absorption band. No chromatophores are
present in the sample 22 hours after the addition of 0G solution

Continuous-wave absorption and fluorescence spectra from B777 in
ODPS plotted as the dipole strength, A(v)/ v and F (v)/v3 , respectively,
and normalized to unit area. The dashed curve shows the intensity
spectrum emitted by the tizsapphire oscillator. The shaded region shows

the range of the transmitted probe light that was passed by the
monochromator to the detector in the dynamic-absorption experiment

Continuous-wave absorption and fluorescence spectra from B820 in OG
plotted as the dipole strength, A(v) / v and F (v)/v3 , respectively, and
normalized to unit area. The dashed curve shows the intensity spectrum
emitted by the tizsapphire oscillator. The shaded region shows the range
of the transmitted probe light that was passed by the monochromator to
the detector in the dynamic-absorption experiment

Pump-probe signal obtained from B777 in ODPS at room temperature
Pump-probe signal obtained from B820 in 0G at room temperature

Expanded view of the rapidly damped oscillation observed in the
dynamic-absorption transient from B777 in ODPS (see figure 5.9)
superimposed with a model defined by the sum of two independent log-
normal distributions £(a)) of damped cosinusoids, each defined by
equation 5.2. The scaling of the ordinate is relative to the magnitude of
the pump—probe ground-state depletion signal, as in figure 5.9. Bottom:
Plots of [.(w) for the two components observed in ODPS and their sum,

914(co) (thick curve). The model parameters are provided in table 5.1

xiv

110

111

114

115

117

118

120

5.12

5.13

5.14

5.15

Expanded view of the rapidly damped oscillation Observed in the
dynamic-absorption transient from B820 in 0G (see figure 5.10)
superimposed with a model defined by the sum of two independent log-
normal distributions [,(w) of damped cosinusoids, each defined by
equation 5.2. The scaling of the ordinate is relative to the magnitude of
the pump—probe ground-state depletion signal, as in figure 5.10.
Bottom: Plots of L( w) for the two components observed in 0G and their

sum, 914((0) (thick curve). The model parameters are provided in table
5.1

Space-filling (CPK) representation of the BChl dimer in the B820
subunit from LHl based on the calculations of Schulten and coworkers

Model of the BChl dimer in the B820 subunit from LHl based on the
calculations Of Schulten and coworkers. The polar residues in the
vicinity Of the dimer are shown as line structures; tryptophan 41,
tryptophan 43, tryptophan 48, and tyrosine 41. These residues are
possible candidates for the interaction between the n-electron density of
the dimer and the protein medium. In order to simplify the figure, the
axially coordinated histidines, H32 and H39, are not shown

The primary electron donor, P, from the reaction center from
Rhodobacter sphaeroides. The polar residues in the vicinity Of the
special pair are shown as line structures; serine 190, serine 205, serine
224, tryptophan 156, asparagine 187, threonine 186, and histidine 168.
These residues are possible candidates for the interaction between the 1:-
electron density of the special pair and the protein medium. In order to
simplify the figure, the axially coordinated histidines, H173 and H202,
are not shown

XV

121

125

128

131

INTRODUCTION

Resonance Raman spectroscopy has been used to examine ground-state and excited-
state vibrational dynamics of a variety of complex systems, such as the photosynthetic
reaction center. While this technique offers valuable information about the equilibrium
geometry and dynamics as well as the electronic structure of the chromophore, it has
some limitations. A significant limitation in resonance Raman spectroscopy is that broad
low-frequency line shapes arising from inhomogeneously broadened features are not
easily discerned from the baseline.

Dynamic absorption spectroscopy, a femtosecond pump-probe, time-domain form of
resonance Raman spectroscopy, affords us a spectroscopic tool that is capable of
detecting the broad line shapes of the coupled intermolecular modes and distinguishing
them from intramolecular modes.

The dissertation is organized as follows:

Chapter 1 begins with a short discussion of charge-transfer reactions in
photosynthetic reaction centers and raises the question of the structural origin of the
quantum efficiency of natural photosynthesis electron-transfer processes in reaction-
center protein complexes. It is followed by a detailed narrative on electron-transfer
theory. Finally, the model target systems are presented.

Chapter 2 begins with an explanation of vibrational coherence and wave packet
theory. A comparison of experimental techniques used to examine wave packet motions
is next. This is followed by a detailed description of the laser system used in the

experiments discussed in this dissertation. The chapter concludes with a discussion of

signal processing.

In Chapter 3, the first Observations of vibrational coherence in the 10-220-cm" region
from bacteriochlorophyll a in solution will be presented. The distinction between
bacteriochlorophyll's intramolecular normal modes and intermolecular modes between
the bacteriochlorophyll and the surrounding solvent will be discussed.

In Chapter 4, the rapidly damped vibrational coherence observed at room temperature
in bacteriochlorophyll 0 solutions in pyridine, acetone, 1-propanol and cyclohexane is
compared. The results in this chapter Show for the first time that intermolecular modes
between large electronic Chromophores and the surrounding solvent can be resonance-
Raman active.

In Chapter 5, the rapidly damped vibrational coherence observed at room temperature
in the bacteriochlorophyll—containing proteins B777 and 3820 are compared to the
vibrational coherence observed in bacteriochlorophyll 0 solutions. The structural origin

of the rapidly damped vibrational coherence will be discussed.

CHAPTER 1
INTRODUCTION
1.0 Photosynthetic Reaction Centers
1.0.1 Charge-Transfer Reactions in Photosynthetic Reaction Centers
The work presented in the following chapters is motivated by the persistent question
of the structural origin of the quantum efficiency of natural photosynthesis electron-
transfer processes in reaction-center protein complexes."8 The mechanism of the primary
charge-separation in the photosynthetic reaction center remains a topic of intense interest
despite many years of sustained efforts among several labs."9"0 Over the last twenty
years, covalently linked donor—acceptor supramolecular assembliesl "'4 have been
synthesized by a number of groups with the intention of emulating some of the
photochemical properties of photosynthetic reaction centers. Long-lived charge-

15-17
(1,

separations with relatively high quantum yields have been achieve and some

18-20

impressive progress has been made in connecting light-harvesting arrays to charge-

separation units to make artificial reaction centers that reproduce the essential features of
bacterial photosynthesis.2 I ‘22
The artificial systems lack, however, the protein-derived medium that is exploited by
the natural photosynthetic reaction centers to obtain an effective irreversibility for the
charge-separated states. The protein's folded structure apparently provides more than just
a rigid scaffold that holds the donor—acceptor pairs in a favorable geometry. Protein-
derived vibrational motions are apparently coupled to the forward electron-transfer

reactions so that they occur in a nearly activationless manner,23 and it is likely that

formation of charge-separated products is accompanied by reorganizational motions of

the protein that resemble in some ways the dynamics of molecular solvents.24'28 At
present, the role of the solvent medium in the quantum efficiency of synthetic donor

acceptor assemblies is a topic of interest.

1.0.2 Structure of the Photosynthetic Reaction Center

The reaction centers found in purple bacteria have been studied extensively."8 The
X-ray crystal structures of the reaction centers of Rhodoseudomonas viridis (see figure
1.1) and Rhodobacter sphaeroides have been known for almost fifteen years. The redox-
active Chromophores are arranged along two chromophore branches (L and M), which are
related by a pseudo C 2 symmetry axis between the primary electron donor P and the non-
heme Fe ion, as shown in figure 1.2. The L and M branches serve as electron acceptors
for P. A main interest of current research is focused on understanding of how the initial-
electron transfer reactions occur only along the L branch. There is good reason to believe
that the protein-derived surroundings of P and of the adjacent monomeric
bacteriochlorophyll a macrocycles (BChls), BL and BM, contribute to the breaking of
symmetry that results in the charge separation reaction in a given direction. BChl dimers
may be favored as primary electron donor structures in reaction centers due to their
charge- transfer properties, which might contribute to the high quantum efficiency and
spatial directionality of the initial electron-transfer events. Model systems for artificial
photosynthesis have been developed by Wasielewski and coworkers that function with a

single chlorophyll although not as efficiently as native reaction centers.ll

 

Figure 1.1. The photosynthetic reaction center from Rhodopseudomonas viridis,3
rendered from RCSB Protein Data Bank29 file laij. This view shows the topology of the
cytochrome (green, top), L (blue), M (pink), and H (green, bottom) subunits.

 

Figure 1.2. Chromophores in the reaction center from Rhodopseudomonas viridis.3 The
arrows represent the directionality of the electron transfer reactions.

1.1 Electron-Transfer Theory

30'” the reorganization energy is the free energy required to

In the Marcus theory,
move the atoms of the donor-acceptor pair, and to some extent those of the intervening
medium, so that the structure of the product state is assumed in advance of the
instantaneous electron-transfer event (see figure 1.3). The vibrational modes that are said
to be coupled to an electron-transfer reaction are thought to be the ones that make
significant contribution to this energy. The reorganization energy is traditionally
separated into two terms:

’1 = ’h T ’10 (1.1)

The inner-sphere term, 2,, is the change in energy associated with the change of structure

of the donor and acceptor molecules. The outer-sphere term, 10, is the part associated

with the change of structure of the solvent medium.29
For donor—acceptor systems in a molecular solvent, the simplest approach for the

calculation of the outer-sphere reorganization energy employs dielectric-continuum

theory, which treats the solvent as a continuous medium characterized by optical and

static dielectric constants 80 and as:
1 1 1 1 1
A = A 2 —-— —+—-— 1.2
0 ( q) is, as][20 2a, R) ( )

Here the solvent forms a cavity of radius a, and 02 around the donor and acceptor,
respectively; Aq is the charge transferred, and R is the distance from the donor to the
acceptor.34 There are problems with using such a picture to treat long-distance electron-
transfer reactions in a protein. A protein’s folded structure presents a discontinuous,

heterogeneous medium Of varying polarizability,35 and there is reason to suspect that the

Energy

 

 

 

 

\fibrational Displacement

Figure 1.3. Diabatic harmonic potential-energy curves for the reactant (13,) and product

(Eb) of an electron-transfer reaction as a function of the displacement of one of the

normal modes of vibration n; 4,, marks the mode-specific reorganization energy, and

—AE° is the driving force.

bridge between the donor and acceptor is through bonded and hydrogen-bonded
pathways rather than just through space.”43 An opposing view from Dutton and
coworkers“45 holds that long-distance electron transfer in proteins occurs through a
homogeneous barrier that is not necessarily directed through covalent pathways. Recent
work by the Winkler/Gray group supports the former view in that electron transfer
through bonded linkages over a given distance was shown to be much faster than through
nonbonding interactions.46

The magnitude of the reorganization energy plays an important role in determining
the rate of an electron-transfer reaction.32 For non-adiabatic electron transfer in the high-
temperature limit, where the relevant normal modes of vibration can be treated

classically, the reorganization energy and the driving force, —AG° , control the magnitude

Of the Gibbs free energy Of activation, AG'

AG“ = (2+AGO)2/4A (1.3)

For a spontaneous reaction, the driving force is the negative of the Gibbs free-energy

difference between the product and reactant states at their respective equilibrium
geometries; spontaneous reactions (AG0 < 0) have positive driving forces. The

expression for the rate constant,

27: V2 *
k = _ DA CXP(—AG/ ) 1.4
h (47r/ikBT)“ 7' kBT ( )

 

includes in the pre-exponential term an electronic donor—acceptor coupling factor, VDA,

which describes the spatial overlap of the electron density from the donor and acceptor at

the geometry of the transition state. Note that the preexponential and exponential terms
contain factors of [Wand ,1“ , respectively, so one would predict that the highest
electron-transfer rates are Obtained when it is minimized, say, by constructing a rigid
donor—acceptor system and by running the reaction in a non-polar medium. For
spontaneous reactions, however, the rate constants magnitude is sensitively controlled by
how well the driving force is matched to the reorganization energy. The usual expectation

that the rate constant should increase as the driving force increases corresponds to the

normal regime of the Marcus theory (-AGO < X). The maximum rate occurs when

—AG° = ,1 , the condition that produces activationless kinetics (AG. 2 0) . The inverted
regime, where the rate constant decreases as the driving force increases, is encountered

as the driving force is increased past the matching condition (-AGO > ll) .47

1.2 Coupled Modes in Photosynthesis

The reaction centers from purple bacteria exhibit a set of redox-active Chromophores
and surrounding protein structures that apparently optimize the reorganization energy and
driving forces for each electron-transfer step so that the reactions exhibit a temperature-
independent or weakly temperature-dependent profile.48 Vibrational modes derived from

the protein medium in the 80—120-cm'I regime23‘48'51

are apparently coupled in each step
even though a broad range of time scales, a range of intervening protein structures, and

several types of cofactors are involved, as can be seen in figures 1.1 and 1.2:

1. the reduction of the bacteriochlorophyll dimer (special pair) primary electron

donor, P, by the cytochrome subunit,52

10

2. the light-driven primary electron-transfer step from P to the bacteriopheo-

phytin acceptor, BPhL53-56

3. the secondary electron transfer from BPhL to the quinone Qa.57'59

The first and third reactions are ground-state reactions involving relatively slow
electron transfers over long distances. In contrast, the second reaction listed above is a

light-initiated, excited-state reaction that occurs so rapidly (T z 3138) that reactants may

55. 6
d, 5

not be vibrationally equilibrate so the temperature dependence is not as easily

interpreted. Experimental evidence from studies of electron-transfer in photosynthetic
reaction centers strongly support the claim that low-frequency modes, in the 100-cm'l
regime, are associated with the reaction coordinate. Small and coworkers showed that a
Franck-Condon progression along low—frequency normal-mode coordinates accounts for
a substantial fraction of the optical reorganization energy using photochemical hole-

53,54

burning spectroscopy at cryogenic temperatures. Vos, Martin and coworkers

observed vibrational wave packet motion in pump-probe transients from reaction center

60-64

preparations using femtosecond pump-probe spectroscopy. The work detected several

low frequency modes under 160 cm’1 .60‘62‘63’65‘66 Similar low-frequency modes were
observed by Stanley and Boxer67 in modulations of the spontaneous fluorescence from
P*. The damping time for wave-packet motion is comparable to the electron-transfer
time scale,60 so the possibility that the active modes are coupled tO the electron-transfer

69-72

reaction coordinate has been raised by Vos et al.,68 Shuvalov and coworkers, and

limb and coworkers.”74

11

1.3 Assignments for the Coupled Modes

Vos, Martin, and coworkers detected several active modes in the vibrational
coherence from the stimulated emission from P*. The strongest features were observed
in the 90-16O-cm'l range.75 The lowest frequencies (10—30 cm'l) in the observed set were
assigned to collective protein-derived motions."O This suggestion was made in
consideration of the peak observed in the 50-cm'l region of the vibrational density of

states obtained from molecular dynamics simulations in small proteins;76

a similar feature
is observed in non-resonant Raman spectra from small protein crystals.77 Small and
coworkers assigned the 30-cm'l modes they detected in hole-buming experiments to
protein phonons.53’54

Vos, Martin, and coworkers then showed that the vibrational coherence from P*
exhibits frequency shifts and intensity changes in response to a range of point mutations
that alter the hydrogen-bonding pattern between the pair of bacteriochlorophyll (BChl)
macrocycles and the surrounding protein subunits.78 The active modes in the
90—160-cm'l range were assigned not to the BChl macrocycles themselves but to the
protein structure itself because mutations that affected the electronic configuration of P
did not cause frequency shifts.75‘78 This conclusion was supported by Struve et al.75 and
Chachisvilis and coworkers;76 both reporting that BChl in solution does not exhibit
vibrational coherence.

These results lead us to suggest that the coupled modes arise from intermolecular

modes between the chromophores and the surrounding protein medium rather than

involving collective motions of the bridging protein between the donor and acceptor. A

12

tuning of the electronic properties of the embedded chromophores through predominantly
solvatochromic effects would be mediated by these modes, which would resemble those
between a chromophore and its clustered solvent molecules. A change in the electron
configuration of the donor or acceptor might be expected to displace weak, formally non-
bonding interactions with protein-derived groups such as hydrogen bonds, polar solvent
(dipole-dipole) interactions, or perhaps even non-polar solvent (London dispersion)
interactions. These interactions are known to exhibit frequencies in the right range to be
the coupled modes whose frequencies were identified experimentally. In molecular
liquids, intermolecular mode frequencies detected in nonresonant Optical Kerr effect
experiments lie in the 10—200-cm'l range;79 collective motions of liquids are noted in the

lO-cm'l regime,80 but pair-wise interactions are found in the 60—100-cm'l range.8|

1.4 Resonance Raman Activity in Intermolecular Modes

Electron-transfer reactions coupled to protein—chromophore intermolecular modes
would be mediated by a partial delocalization of the chromophores electron density onto
the protein-derived group. Therefore, it is probable that the relevant modes would be
active in the resonance Raman spectrum. Mathies and coworkers are among the few to
detect intermolecular modes in condensed phases using resonance Raman spectroscopy in
the frequency domain. Resonance-Raman active modes were detected from alcohol82 and
water83 solutions of the solvated electron. The solvated electron is a diffuse structure that
makes a significant ground-to-excited-state change in size and shape;84'88 it apparently

interacts strongly enough with a number of solvent molecules in the first shell that the

absorption transition is vibronically coupled to a number of solvent intramolecular

l3

modes.”83 The lowest frequency Raman-active modes observed in water, at 110 cm",
are the intermolecular hindered translational modes between the electron and molecules
in the first solvation shell.83

In contrast, Waterland and Kelley89 did not observe features from intermolecular
modes in their UV resonance Raman studies of the nitrate ion in several solvents, perhaps
because the melectron density of the nitrate ion is highly localized. Barbara and
coworkers90 noted significant solvent dependences in the intervalence charge-transfer rate
and in the damping of the ground-state vibrational coherence in the femtosecond pump-
probe studies of the cyanide-bridged Ru(ll,III) mixed-valence complex, but no solvent-
dependent components were directly observed. As in the case of the nitration ion, a
highly localized n—electron density is probably responsible for the lack of detection of
intermolecular modes.

Chromophores with spatially delocalized n—electron density owing to conjugation are
much more likely to exhibit vibronic coupling to intermolecular modes. We suggest that
time-domain resonance Raman methods, such as the pump—probe method called dynamic
absorption spectroscopy that was described by Shank, Mathies and coworkers,”95 are
very likely to be successful in the study of intermolecular modes in delocalized
n—electron systems. In the frequency domain, broad low-frequency line shapes arising
from inhomogeneously broadened features are not easily discerned from the baseline. In
the time domain, however, the same line shapes result in sharp, rapidly damped features
in the vibrational coherence, and the broad line shapes can be resolved into components
on the basis of their distinct phases. Resonance Raman activity should be exhibited by

intermolecular chromophore—protein or chromophore—solvent modes that are displaced

l4

owing to the transitions of the chromophore; similar mode displacements would be
activated by electron-transfer reactions since the redox-active electrons are associated

with the frontier highest-occupied and lowest-unoccupied molecular orbitals.

1.5 Research Design

In the preceding narrative, we have raised the hypothesis that electron-transfer
reactions in condensed phases and in the photosynthetic reaction center are coupled to
intermolecular vibrational modes between the redox-active chromophores and the
clustered solvent molecules in the first solvation shell. We focus the content of this thesis
on understanding the structural origin of the vibrational coherence from the special pair,
P, in the photosynthetic reaction center. We test the hypothesis that at least some of the
components in the vibrational coherence arise from intermolecular modes between the
BChI macrocycles in P and the surrounding protein medium. Our approach is to start
with monomeric bacteriochlorophyll in solution, to determine whether the clustered
solvent molecules in the first solvation shell contribute to the vibrational coherence. We
then progress to a consideration of the vibrational coherence in BChl-containing proteins,
B820 and B777. Here, we focus exclusively on how the protein structure Of B820
impacts the character Of the low-frequency vibrational coherence. We compare the
vibrational coherence of the monomeric B777 system, which exposes the BChl
macrocycle to the surrounding detergent environment, to that of the B820 system, where
the BChl macrocycles are essentially completely protected from the surrounding solvent

by the protein structure.

15

1.6 References

(1) Parson, W. W. In Photosynthesis Wew Comprehensive Biochemistry, Volume
15); Amesz, 1., Ed.; Elsevier: New York, 1987, p 43—61.

(2) Deisenhofer, J.: Epp. 0.; Miki. K.; Huber, R.; Michel, H. J. Mol. Biol. 1984,
180, 385—398.

(3) Deisenhofer, .I.; Epp, 0.; Miki, K.; Huber, R.; Michel, H. Nature 1985, 318,
618—624.

(4) Allen, J. P.; Feher, G.; Yeates, T. O.; Rees, D. C.; Deisenhofer, J.; Michel, H.;
Huber, R. Proc. Natl. Acad. Sci. USA 1986, 83, 8589—8593.

(5) Allen, .I. P.; Feher, G.; Yeates, T. 0.; Komiya, H.; Rees, D. C. Proc. Natl. Acad.
Sci. USA 1987. 84, 5730—5734.

(6) Rhee, K.-H.; Morris, E. P.; Barber. 1.; Kiihlbrandt, W. Nature 1998, 396, 283-
286.

(7) Zouni, A.; Witt. H.-T.; Kern, 1.; Fromme, P.; KrauB, N.; Saenger, W.; Orth, P.
Nature 2001, 409, 739—742.

(8) Jordan, P.; Fromme, P.; Witt, H. T.; Klukas, O.; Saenger, W.; KrauB, N. Nature
2001, 411, 909—917.

(9) Boxer, S. G.; Goldstein, R. A.; Lockhart, D. J.; Middendorf, T. R.; Takiff, L.
J. Phys. Chem. 1989, 93, 8280—8294.

(10) Friesner, R. A.; Won, Y. Biochim. Biophys. Acta 1989, 977, 99—122.
(1 1) Wasielewski, M. R. Chem. Rev. 1992, 92, 436—461.

(12) Wasielewski, M. R.; Wiederrecht, G. P.; Svec, W. a.; Niemczyk, M. P. Solar
Energy Mat. Solar Cells 1995, 38, 127—134.

(13) Redmore, N. P.; Rubtsov, l. V.; Therien, M. .l. J. Am. Chem. Soc. 2003, 125,
8769—8778.

(14) Splan, K. E.; Stem, C. L.; Hupp, J. T. Inorg. Chim. Acta 2004, 357, 4005—4014.

(15) Gust, D.; Moore, T. A.; Makings, L. R.; Liddell, P. A.; Nemeth, G. A.; Moore,
A. L. J. Am. Chem. Soc. 1986, 108, 8028—8031.

16

(16) Gust, D.; Moore, T. A. Science 1989, 244, 35—141.

(17) Gust, D.; Moore, T. A.; Moore, A. L.; Lee, S.-J.; Bittersmann, E.; Luttrull, D.
K.; Rehms, A. A.; DeGarazianO, J. M.; Ma, X. C.; GaO, F.; Belford, R. E.; Trier,
T. T. Science 1990, 248, 199—201.

(18) Prathapan, S.; Johnson, T. E.; Lindsey, J. S. J. Am. Chem. Soc. 1993, 115, 7519—
7520.

(19) Hsiao, J.-S.; Krueger, B. P.; Wagner, R. W.; Johnson, T. E.; Delayen, J. K.;
Mauzerall, D. C.; Fleming, G. R.; Lindsey, J. S.; Bocian, D. F .; Donohoe, R. J.
J. Am. Chem. Soc. 1996,118,11181—11193.

(20) Loewe, R. S.; Lammi, R. K.; Diers, J. R.; Kirrnaier, C.; Bocian, D. F .; Holten,
D.; Lindsey, J. S. J. Mater. Chem. 2002, 12, 1530—1552.

(21) Stenberg-Yfrach, G.; Rigaud, J.-L.; Durantini, E. N.; Moore, A. L.; Gust, D.;
Moore, T. A. Nature 1998, 392, 479—482.

(22) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res 2001, 34, 40—48.

(23) Bixon, M.; Jortner, J. J. Phys. Chem. 1986, 90, 3795—3800.

(24) Maroncelli, M.; Maclnnis, J.; Fleming, G. R. Science 1989, 243, 1674—1681.
(25) Simon, J. D. Pure Appl. Chem. 1990, 62, 2243—2250.

(26) Barbara, P. F.; Walker, G. C.; Smith, T. P. Science 1992, 256, 975—981.

(27) Stratt, R. M.; Maroncelli, M. J. Phys. Chem. 1996, 100, 12981—12996.

(28) Lampa-Pastirk, 8.; Beck, W. F. J. Phys. Chem. B 2004, 108, 16288-16294.

(29) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.;
Shindyalov, I. N.; Boume, P. E. Nucl. Acids Res. 2000, 28, 235—242.

(30) Marcus, R. A. J. Chem. Phys. 1956, 24, 966—989.

(31) Marcus, R. A. J. Chem. Phys. 1965, 43, 679—701.

(32) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265—322.
(33) Marcus, R. A. Isr. J. Chem. 1988, 28, 205—213.

(34) Walker, G. C.; Beratan, D. N. In Encyclopedia of Chemical Physics and
Physical Chemistry; Moore, J. H., Spencer, N. D., Eds.; Institute of Physics

17

Publishing: Bristol, England, 2001, p 2657—2680.

(35) Cohen, B. E.; McAnaney, T. 8.; Park, E. S.; Jan, Y. N.; Boxer, S. G.; Jan, L. Y.
Science 2002, 296, 1700—1703.

(36) Beratan, D. N.; Onuchic, J. N.; Hopfield, J. J. J. Chem. Phys. 1987, 86, 4488—
4498.

(37) Onuchic, J. N.; Beratan, D. N. J. Chem. Phys. 1990, 92, 722—733.

(38) Beratan, D. N.; Onuchic, J. N.; Betts, J. N.; Bowler, B. B; Gray, H. B. J. Am.
Chem. Soc. 1990, 112, 7915—7921.

(39) Onuchic, J. N.; de Andrade, P. C. P.; Beratan, D. N. J. Chem. Phys. 1991, 95,
1131.

(40) Betts, J. N.; Beratan, D. N.; Onuchic, J. N. J. Am. Chem. Soc. 1992, 114, 4043-
4046.

(41) Beratan. D. N.; Betts, J. N.; Onuchic, J. N. J. Am. Chem. Soc. 1992, 96, 2852—
2855. '

(42) Schulten, K. Science 2000, 290, 61—62.
(43) Balabin, I. A.; Onuchic, J. N. Science 2000, 290, 114—1 17.

(44) Moser, C. C.; Keske, J. M.; Warncke, K.; Farid, R. S.; Dutton, P. L. Nature
1992, 355, 796—802.

(45) Page, C. C.; Moser, C. C.; Chen, X.; Dutton, P. L. Nature 1999, 402, 47—52.

(46) Wenger, O. S.; Leigh, B. S.; Villaherrnosa, R. M.; Gray, H. B.; Winkler, J. R.
Science 2005, 307, 99—102.

(47) Closs, G. L.; Miller, J. R. Science 1988, 240, 440—447.
(48) Bixon, M.; Jortner, J. Chem. Phys. Lett. 1989, 159, 17—20.

(49) Kirmaier, C.; Holten, D.; Parson, W. W. Biochim. Biophys. Acta 1985, 810, 33—
48.

(50) Fleming, G. R.; Martin, J.-L.; Breton, J. Nature 1988, 333, 190—192.

(51) Breton, J.; Martin, J.-L.; Fleming, G. R.; Lambry, J.-C. Biochemistry 1988, 27,
8276.

18

(52) Hopfield, J. J. Proc Natl Acad Sci U S A 1974, 71, 3640.

(53) Johnson, S. G.; Tang, D.; Jankowiak, R.; Hayes, J. M.; Small, G. J.; Tiede, D.
M. J. Phys. Chem. 1990, 94, 5849—5855.

(54) Johnson, S. G.; Tang, D.; Jankowiak, R.; Hayes, J. M.; Small, G. H. J. Phys.
Chem. 1989, 93, 5953—5957.

(55) Lin, 8.; Taguchi, A. K. W.; Woodbury, N. W. J. Phys. Chem. 1996, 100, 17067—
17078.

(56) Haffa, A. L. M.; Lin, 8.; Katilius, E.; Williams, J. C.; Taguchi, A. K. W.; Allen,
J. P.; Woodbury, N. W. J. Phys. Chem. B. 2002, 106, 7376—7384.

(57) Gunner, M. R.; Robertson, D. E.; Dutton, P. L. J. Phys. Chem. 1986, 90, 3783—
3795.

(58) Gunner, M. R.; Dutton, P. L. J. Am. Chem. Soc. 1989, 111, 3400—3412.

(59) Feher, G.; Okamura, M.; Kleinfeld, D. In Protein Structure: Molecular and
Electronic Reactivity; Austin, R., Buhks, E., Chance, B., De Vault, D., Dutton,
P. L., F rauenfelder, H., Gol'danskii, V., Eds.; Springer-Verlag: New York,
1987, p 399—421.

(60) V05, M. H.; Rappaport, F .; Lambry, J.-C.; Breton, J .; Martin, J.-L. Nature 1993,
363, 320—325.

(61) Vos, M. H.; Jones, M. R.; Breton. J.; Lambry, J.-C.; Martin, J.-L. Biochemistry
1996, 35, 2687—2692.

(62) V08, M. H.; Jones, M. R.; Hunter, C. N.; Breton, J.; Lambry, J.-C.; Martin, J.-L.
Biochemistry 1994, 33, 6750—6757.

(63) V05, M. H.; Jones, M. R.; Hunter, C. N.; Breton, J.; Martin, .I.-L. Proc. Nat.
Acad. Sci. USA 1994, 91, 12701—12705.

(64) Vos, M. H.; Jones, M. R.; McGlynn, P.; Hunter, C. N.; Breton, J.; Martin, J.-L.
Biochim. Biophys. Acta 1994, 1186, 117—122.

(65) V05, M. H.; Lambry, J. C.; Robles, S. J.; Youvan, D. C.; Breton, J.; Martin, J. L.
Proc Natl Acad Sci USA 1991, 88, 8885-8889.

(66) V05, M. H.; Lambry, J. C.; Robles, S. J.; Youvan, D. C.; Breton, J.; Martin, J. L.
Proc. Nat. Acad. Sci. USA 1992, 89, 613—617.

(67) Stanley, R. J.; Boxer. S. G. J. Phys. Chem. 1995, 99, 859—863.

19

(68) V03, M. H.; Jones, M. R.; Martin, J. L. Chem. Phys. 1998, 233, 179—190.
(69) Yakovlev, A.; Shkuropatov, A.; Shuvalov, V. FEBS Lett. 2000, 466, 209—212.

(70) Yakovlev, A.; Shkuropatov, A.; Shuvalov, V. Biochemistry 2002, 41, 2667—
2674.

(71) Yakovlev, A.; Shkuropatov, A.; Shuvalov. V. Biochemistry 2002, 41, 14019—
14027.

(72) Shuvalov, V.; Yakovlev, A. FEBS Lett. 2003, 540, 26-34.
(73) SpOrlein, S.; Zinth, W.; Wachtveitl, J. J. Phys. Chem. B 1998, 102, 7492-7496.

(74) Huppmann, P.; SpOrlein, S.; Bibikova, M.; Oesterhelt, D.; Wachtveitl, J.; Zinth,
W. J. Phys. Chem. A. 2003, 107, 8302—8309.

(75) V03, M. H.; Martin, J.-L. Biochim. Biophys. Acta 1999, 1411, 1—20.

(76) Brooks III, C. L.; Karplus, M.; Pettitt, B. M. Proteins: 0 Theoretical Perspective
of Dynamics, Structure, and Thermodynamics; John Wiley and Sons: New York,
1988.

(77) Peticolas, W. L. Meth. Enzymol. 1978, 61, 425—458.

(78) Rischel, C.; Spiedel, D.; Ridge, J. P.; Jones, M. R.; Breton, J.; Lambry, J.-C.;
Martin, J.-L.; VOS, M. H. Proc. Natl. Acad. Sci. USA 1998, 95, 12306—1231 1.

(79) Lotshaw, W. T.; McMorrow, D.; Thantu, N.; Melinger, J. S.; Kitchenbaum, R.
.1. Raman Spect. 1995, 26, 571—583.

(80) McMorrow, D.; Lotshaw, W. T. Chem. Phys. Lett. 1993, 201, 369—376.
(81) McMorrow, D.; Lotshaw, W. T. J. Phys. Chem. 1991, 95, 10395—10406.

(82) Tauber, M. J.; Stuart, C. M.; Mathies. R. A. J. Am. Chem. Soc. 2004, 126, 3414—
3415.

(83) Tauber, M. J.; Mathies, R. A. Chem. Phys. Lett. 2002, 354, 518—526.

(84) Alfano, J. C.; Walhout, P. K.; Kimura, Y.; Barbara, P. F. J. Chem. Phys. 1993,
98, 5996-5998.

(85) Kimura, Y.; Alfano, J. C.; Walhout, P. K.; Barbara, P. F. J. Phys. Chem. 1994,
98, 3450—3458.

20

(86) Reid, P. J.; Silva, C.; Walhout, P. K.; Barbara, P. F. Chem. Phys. Lett. 1994,
228, 658—664.

(87) Walhout, P. K.; Alfano, J. C.; Kimura, Y.; Silva, C.; Reid, P. J.; Barbara, P. F.
Chem. Phys. Lett. 1995, 232, 135—140.

(88) Silva, C.; Walhout, P. K.; Reid, P. J.; Barbara, P. F. J. Phys. Chem. A 1998, 102,
5701-5707.

(89) Waterland, M.; Myers Kelley, A. J. Chem. Phys. 2000, 113, 6760—6773.

(90) Kambhampati, P.; Son, D. H.; Kee, T. W.; Barbara, P. F. J. Phys. Chem. A 2000,
104, 10637—10644.

(91) Fragnito, H. L.; Bigot, J.-Y.; Becker, P. C.; Shank, C. V. Chem. Phys. Lett.
1989, 160, 101—104.

(92) Pollard, W. T.; Dexheimer, S. L.; Wang, Q.; Peteanu, L. A.; Shank, C. V.;
Mathies, R. A. J. Phys. Chem. 1992, 96, 6147—6158.

(93) Peteanu, L. A.; Schoenlein, R. W.; Wang, Q.; Mathies, R. A.; Shank, C. V.
Proc. Natl. Acad. Sci. USA 1993, 90, 11762—11766.

(94) Schoenlein, R. W.; Peteanu, L. A.; Wang, Q.; Mathies, R. A.; Shank, C. V.
J. Phys. Chem. 1993, 97. 12087—12092.

(95) Wang, Q.; Schoenlein, R. W.; Peteanu, L. A.; Mathies, R. A.; Shank, C. V.
Science 1994, 266, 422—424.

21

CHAPTER 2

THEORY AND EXPERIMENTAL METHODS FOR THE STUDY OF
VIBRATIONAL COHERENCE

2.0 Introduction

As reviewed in Chapter l,it is likely that low-frequency vibrational modes play an
important role in the electron-transfer reactions in photosythesis."2 Several groups have
employed ultrafast pump-probe methods to characterize vibrational coherence from a
variety of systems including the photosynthetic reaction center.2'6 In the following, we
discuss the theory and experimental techniques necessary to test the hypothesis discussed

in Chapter 1.

2.1 Theory
2.1.1 Wave Packets

In the following, we describe a classical description, based on the Franck-Condon
principle, on vibrational wave packets following the treatment of Lee and Heller.8"0
Wave packets have been described quantum mechanically as well as classically;
however, the classical description is easier to visualize conceptually.

As introduced in the previous chapter, the BChl-containing target systems exhibit
strong absorption bands in the visible and UV that are assigned to a It —-> 7f transition.
The movement Of an electron from a bonding orbital to an antibonding orbital results in
the lengthening of one or more bonds in the excited state, causing the equilibrium
structure of the excited state to be displaced from that of the ground state. This structural

change offsets horizontally the excited-state surface from that of the ground-state surface,

22

as illustrated in figure 2.1.

An excited-state wave packet originates when an ensemble of molecules in the
sample interact with the pump field from the incident light. The interaction vertically
transfers the equilibrium probability density from the ground-state to the excited-state
surface in the form of a localized wave packet (F1 in figure 2.1). The wave packet is
acted upon by the forces of the excited-state potential surface. This gives rise to a
moving wave-packet (t=3); the wave packet moves towards the new equilibrium position
in an attempt to minimize its energy with respect to the excited-state geometry.

A second interaction with the pump field transfers part of the excited-state wave
packet back to the ground-state potential surface. Because the wave packet was displaced
in the excited state, it is no longer at equilibrium with respect to the ground-state
geometry. This creates a ground-state wave packet that evolves in time according to the
forces arising from the ground-state potential-energy surface (i=3). The formation of the
ground-state wave packet is also known as resonant impulsive stimulated Raman

scattering, or RISRS.7

2.1.2 Resonance Raman Spectroscopy

Raman spectroscopy can be used to examine ground-state and excited-state
vibrational dynamics of complex systems. The vibrational frequencies give specific
information about geometry and electronic structure of the chromophore in the ground
state, while the intensities of resonance Raman lines give information about the
symmetry, equilibrium geometry, and dynamics of the chromophore in the excited state.ll
Work by Mathies and coworkers” established the use of resonance Raman intensities to

Obtain information on the reaction dynamics of P in the bacterial reaction center. The Lee

23

 

 

 

 

 

 

\fibrational Displacement

Figure 2.1. Excited-state and ground-state coherent wave-packet motion in the dynamic
absorption experiment. The excited-state and ground-state potential-energy surfaces are
drawn as parabolas that are displaced with respect to a generalized multimode coordinate;
rg and re mark the equilibrium ground-state and excited-state geometries, respectively.
Thick arrows represent the resonant pump-laser field; thin arrows show the direction that
the wave packets evolve during the first passage on the two surfaces. The numbers
indicate event times, starting with ground-state probability density, creation of the
excited-state wave packet by the pump field, creation Of the ground-state wave packet by

the pump field, and evolution during the first vibration.

24

and Heller wave packet picture discussed aboveg‘lo‘lz’l3 describes resonance Raman
intensity as the time-dependent spatial overlap of the excited-state wave packet prepared
by the absorption of a photon with the final wave packet prepared on the ground-state
surface following spontaneous emission of a photon. The intensity of the overlap integral
is governed by the time evolution of the wave packet on the excited-state surface before
the transition to the ground state. H

Ground-state and excited-state vibrational coherences in pump-probe experiments
arise from structural displacement of the excited-state and the ground-state surface along

”'17 The same modulation components and frequencies would

normal coordinate modes.
be expected from the excited-state and ground-state wave packets for a bound electronic
state with a comparable surface shape as the ground-state. If an excited-state wave
packet moves away from the Franck-Condon region and crosses to a product state
surface, new modulation frequencies might be observed. In this case, the Franck-Condon

active modes can promote a crossing to the product-state surface.18

2.1.3 Dynamic-Absorption Spectroscopy

Shank, Mathies and co-workersl‘m'26

showed more than ten years ago that resonant
impulsive excitation of large electronic chromophores in condensed phases and in
proteins results in a nonstationary absorption/stimulated-emission spectrum. When
monitored with a short, delayed probe pulse that is resonant with the ground-state
absorption spectrum, the intensity of the transmitted probe light is modulated at the
normal-mode frequencies that would appear in the resonance Raman spectrum. If the

probe pulse is resonant with the stimulated-emission spectrum, the modulation

frequencies are those of the excited electronic state. This so-called dynamic-absorption

25

spectroscopy was described by Pollard and MathiesM'l7 in terms of coherent motion of
vibronic wave packets (vibrational coherence) on the resonant electronic excited-state
and ground-state potential-energy surfaces, the latter produced by stimulated-Raman
transitions.

As discussed by Pollard and Mathies,l7 most of the prior literature on the pump-probe
absorption technique treated the probe pulse as being monochromatic. The absorbance of
the sample at the probe wavelength is measured in terms of the loss of energy of the
transmitted probe pulse. Since the absorbance of the sample is averaged over the actual
probe pulse spectrum, the measured spectra are broadened by the spectral width of the
pulses. Alternatively, dynamic-absorption spectroscopy directly monitors the spectrum
of the probe pulse resulting in high resolution in both the time and frequency
domains.'6"7 The transmitted probe beam is spectrally dispersed onto a detector after
passing through the sample. The transient absorption spectrum of a sample is obtained
by measuring how the spectrum of the probe pulse is changed by transmission though the
sample.17 The pump-probe signal is displayed as the ratio of the intensity of the probe

pulse in the presence and absence of the pump pulse.

2.1.4 Frequency Domain vs. Time Domain

The main limitation Of resonance Raman spectroscopy is that low-frequency modes
are difficult to detect due to Rayleigh scattering at the excitation wavelength. Whereas
dynamic-absorption spectroscopy is not affected by Rayleigh scattering, low-frequency
vibrational modes can easily be detected, as long as the time record of the data is of

sufficient length. Higher frequency modes are obtained easier through resonance Raman

26

spectroscopy. The frequency information returned from dynamic-absorption Spectroscopy
is affected by the pulse duration. In order to observe wave-packet motion at a given
frequency, the excitation pulse has to be much shorter than the vibrational period. This
means that the duration of the pulses used can be adjusted to prevent the observation of
higher frequency modes, if one is only interested in the low-frequency vibrational modes

of a system.

2.2 Experimental: Femtosecond Pump-Probe Spectrometer Design and Characterization
Dynamic-absorption experiments were performed with the femtosecond pump-probe
spectrometer described in the following sections. The spectrometer consists of a
femtosecond tizsapphire oscillator, a prism pair pulse compressor, and a rapid-scanning
Mach-Zehnder interferometer. The construction of the tizsapphire oscillator and the
design of the rapid-scanning Mach-Zehnder interferometer will be discussed in sequence.

27.28

An early version of the spectrometer was discussed previously, and the current

version was described in more recent manuscripts.29'30

2.2.1 Ti:Sapphire Oscillator

The oscillator shown in figure 2.2 is a Mira Optima 900-F Obtained from Coherent.
The oscillator utilizes titanium sapphire as the gain medium and is equipped with X-wave
broad-tuning-range cavity optics, allowing for wide tunability from 700 nm to 980 nm.
The oscillator is pumped by a Verdi (Coherent) 5-W laser. The output pulses were used
at the natural pulse-repetition rate (75 MHz). Figure 2.3 shows the output spectrum
obtained from the mode-locked oscillator. This spectral bandwidth results in 60-fs

(sechz) pulses shown in figure 2.4.

27

11
M1
1:: ourpur
COUPLER
sur — —
M3
M5 STARTER
BRF
BPl { M2

 

    
 

   
   

 

I.-
'—

----------4

M7
8P2

PUMP BEAM

Figure 2.2. The Mira 900-F tizsapphire laser oscillator and pulse compressor

employed in the dynamic-absorption spectroscopy experiments. Symbols: M1, output
coupler; BP1-BP2, Brewster prisms; M2-M3, flat cavity mirrors; BRF, birefringent filter;
L1, focusing lens; M4-M5, curved mirrors; TiAle3, titanium: sapphire crystal; M6, flat

mirror; M7, flat end mirrors; p1-p2, SFIO prisms.

28

Relative Intensity

 

 

 

.0
c:
l

 

I I I I I I
750 760 770 780 790 800 810

Wavelength (nm)

Figure 2.3. Output intensity spectrum of 60-fs (sechz; l2-nm (fwhm) spectral width)

pulses from the tizsapphire oscillator.

29

Relative Intensity

 

 

 

 

I l I l
-400 -200 O 200 400

Time (is)

Figure 2.4. Zero-background autocorrelation trace of a 60-fs (sechz) pulse obtained from

the tizsapphire oscillator.

30

2.2.2 Extracavity Pulse Compression

Extracavity pulse compression and group-velocity-dispersion precompensation was
accomplished with a double-passed pair of SFlO prisms. The prism separation was
adjusted to minimize the pump—probe autocorrelation width at the position of the
reference nonlinear crystal (vide infra), which was confocal with the sample position.

The design of the extracavity compressor is shown in figure 2.5.

2.2.3 Rapid-Scanning Modified Mach—Zehnder Interferometer

Femtosecond spectroscopy experiments employing multiple pulses require one arm of
the interferometer to be scanned slowly. This motion is typically Obtained step-wise with
a delay stage. This technique can take seconds to acquire a single data point. A single
scan can take minutes to obtain, resulting in large errors due to laser instability.
Substituting a rapid scanning translation stage and retroreflector into the design increases
signal-tO-noise ratios without broadening the instrument response function.“ The
interferometer also uses a UZ-retarding photoelastic modulator as a lOO-kHz amplitude
modulator. The design of the interferometer has been described by previous group
members,27‘28 but some significant changes have been made to improve sensitivity and
reproducibility. Figures 2.6 and 2.7 show the design of the rapid-scanning modified

Mach—Zehnder interferometer.
2.2.3.1 Layout of Interferometer

The horizontal polarization of the incident pulse train is rotated to vertical using a

periscope arrangement with mirrors of s and p polarizations oriented for 45° reflections.

31

M5 M4

 

 

 

p2

M6

 

M7

M2

 

M3

 

 

M1

Figure 2.5. Extracavity compressor employed in the dynamic-absorption spectroscopy
experiments. Symbols: Ml-M6, silver mirrors; M7, half silver mirror; p1-p2, SFIO

Brewster-angled prisms.

32

To Sample

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Position
, M2 M. In ..
0167 1:: 02 M7
1:: M2
P1 M6
cal/2
EQP2
FEM-90 M5
M1 BS1 U M8 -
M4 M9

Figure 2.6. Modified Mach-Zehnder rapid-scanning interferometer employed in the
dynamic-absorption spectroscopy experiments. Symbols: M1-M9, silver mirrors; BS1-
BS2, beam splitters; C1-C2, compensation optics; M2, half-wave plates; P1-P2, Glan-

laser calcite polarizers; PEM-90, photoelastic modulator.

33

[E]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A I.
U
M6 Acton Research
SpectraPro 150
it
4»
:: <3
i 5 M4
55 M5
5: PMT P1
variable-density I i _.
filter 1;
441—» X
ii <-> sample
ii <_>
ii 01
i i M2 55:14 \ BS1
%'::::::33.22335?
M3 M1
M

 

 

Figure 2.7. Sample and detection portion Of the Modified Mach-Zehnder rapid-scanning

interferometer employed in the dynamic-absOrption spectroscopy experiments. Symbols:

M1-M6, silver mirrors; BSI, beam splitters; Cl, compensation optics; P1, Glan-laser

calcite polarizers; x, B-barium borate crystal; PMT, photomultiplier tube; PD,

photodiode.

34

The beam is split into pump and probe arms by a 50% beamsplitter. The probe arm
employs a manually adjusted translation stage (Newport) and two silver mirrors (New
Focus 5013) to balance the path length of the pump and probe arms of the interferometer.

A galvanometer-driven translation stage and retroreflector (Clark-MXR ODL-150) is
used to scan the pump-pulse time-of-flight delay at a 1.3-Hz repetition rate over a 12-ps
range. The pump beam is amplitude modulated at 100 kHz by a M2-retarding fused-silica
photoelastic modulator (Hinds Instruments) and a Glan-laser calcite polarizer (Karl
Lambrecht). The fused silica plate in the photoelastic modulator is normal to the pulse
train and 45° with respect to the Optical table. The polarizer allows only vertically
polarized light to pass. A mica zero-order M2 plate (Karl Lambrecht) prior to the
polarizer is used to adjust the pulse train for 100% depth modulation.

The probe beam’s plane of polarization is analyzed at 45° with respect to the pump
beam’s plane of polarization by a Glan-laser calcite polarizer; a mica zero-order M2 plate
(Karl Lambrecht) is used to control the probe intensity. The polarizers were rotated to
obtain a 4:1 pump-probe power ratio. The probe arm contains two compensation optics,
one matching the beam splitter, the other matching the quartz plate in the photoelastic
modulator, to exactly equate the material passed by the beam in both the pump and probe
arms. This allows a single pair of prisms to compensate for group-delay dispersion.

A 10-cm focal-length BK7 lens was used to focus the pump (125 pJ/pulse, 10 mW)
and probe beam (30 pJ/pulse, 2.5 mW) onto the sample’s position. A 25-p.m pinhole was
used initially to ensure spatial overlap of the pump and probe beams. Afier transmitting
through the sample and after recollimation, the probe beam was analyzed at 63.44° with

respect to the pump polarization by another Glan-laser calcite polarizer in order to obtain

35

dichroism-free signals.5 The transmitted probe beam was then passed through a
monochromator (Acton Research SpectraPro 150, 2-nm spectral band pass) and detected
by an amplified silicon photodiode (Thorlabs PDA520). The photodiode signal was
demodulated by a digital lock-in amplifier (SRS 750, Stanford Research Systems), which
was referenced to the lOO-kHz pump modulation frequency.

The lock-in amplifier was Operated in a differential detection mode, with the
transmitted-probe photodiode signal balanced by that from a second, reference
photodiode. The reference photodiode was used to detect a variable sum of a portion of
the pump and probe beams. both of which were split from the main beams prior to the
sample-focusing lens. The intensity of the pump-beam fraction on the reference
photodiode was adjusted with a variable-density neutral-density filter to cancel the post-
time-zero bleaching signal at a closer point in the probe-delay scan; the unmodulated
probe beam was used to bias the reference photodiode into the operating intensity regime
used by the sample-beam’s photodiode. This practice improved the signal/noise ratio
significantly by allowing a partial nulling of the noise arising from the oscillator and
pump laser.

The pump and probe beams were sampled prior to the focusing lens to provide a pair
of reference pulses for autocorrelation analysis during data acquisition. A B-barium
borate crystal (Cleveland Crystals, lOO-pm thickness, cut for type I sum-frequency
generation with 800-nm input) was placed at the focus of a 10-cm focal-length BK7 lens;
the second-harmonic beam was detected by a photomultiplier tube and lock-in amplifier
(LIA-MV-ZOO-H, Femto Messtechnik). The group-delay dispersion in the auto-

correlation arm of the interferometer was the same as in the sample arm; identical beam

36

splitters, compensator plates, and focusing optics were employed. A sample/hold
amplifier and digitizer (National Instruments SC-2040 S/H and PC1-6024E, respectively)
simultaneously acquired at a 17-kHz sampling rate the analog output of the pump-probe
and autocorrelation lock-in amplifiers, the photodiode intensity, and the galvanometer’s
position signal, resulting in 7000 points per scan; the effective dwell time per acquired
data point was 1.8 fs. The signal-averaging system and monochromator were controlled
by LabVIEW (National Instruments) routines. We eliminated scan-to-scan delay drift by
employing the autocorrelation signal as a zero-delay time reference pulse for each scan;
in previously reported rapid-scanning experiments, a modest broadening of the

instrument-response function was Observed in long averaging runs.27’32

2.3 Data Acquisition

A sample/hold amplifier and digitizer (National Instruments, SC-2040S/H and
PC1-6024E) was used to record the analog outputs of the two lock-in amplifiers (for the
pump—probe signal and autocorrelation signal), the transmitted-probe photodiode signal,
and the position Of the galvanometer (delay reference) signal. Figure 2.8 shows the
multiple traces that are obtained during dynamic-absorption experiments. Two complete
scans of the rapid-scanning delay stage are shown as a function of real (laboratory) time.

The pump-probe delay is calculated from the position of the rapid-scanning delay
stage. The position of the stage is calibrated by adjusting the distance the probe pulse
must travel and measuring the movement of the zero-background autocorrelation signal
relative to the position of the delay stage. The change in distance traveled by the probe

pulse is converted to time and delay-stage position is measured in volts. A conversion

37

Pump Probe
Delay (fs)
0)
O
O
O

 

Autocorrelation
Intensity

 

 

 

 

 

 

 

 

 

 

0‘ 1‘1

 

 

 

A T/ T (relative units)

 

_

Time (s)

Figure 2.8. Multiple traces Obtained simultaneously during dynamic-absorption
experiments of bacteriochlorophyll a in pyridine. The probe beam was detected at 757
nm. Two complete scans of the rapid-scanning delay stage are shown as a function of real
time. Top: Pump-probe delay determined from the position of the rapid-scanning delay
stage. Middle: Zero-background autocorrelation. Bottom: Single-wavelength dynamic-

absorption transient.

38

factor, with the units of fs/V, can be calculated and used to convert the position of the
delay stage to pump-probe delay. In most experiments, the conversion factor was 19300
fsN.

For each scan of the delay stage, the pump and probe pulses interact twice. As
Observed in figure 2.8, two scans of the delay stage give four zero-background
autocorrelation and dynamic-absorption signals. Only data from one half of a complete
scan is digitized in practice.

The scan range is adjusted to visualize only one pump-probe interaction, as shown in
figure 2.9. The delay axis is scanned at 1.3 Hz over a 12-ps delay range and 7000 data
points are acquired with a dwell time of 1.8 fs. The digitization frequency is 17 kHz.
LabVIEW routines control the signal-averaging system and the monochromator. Multiple
scans are averaged for a single-wavelength measurement. The autocorrelation signal is
used as a zero-delay-time reference to eliminate scan-tO-scan delay drift. Figure 2.10
shows a scan-averaged single-wavelength dynamic-absorption transient as a function of

increasing probe delay.

39

 

 

 

 

 

 

 

 

 

a, 10000—
_Q A
9% 8000-
0. (U
E a 40004
50 2000—
01
-2000—
C
.9
<15 2
q, ._
§ §
‘5
<
0 ll
1.5
'E
3
Q)
.2
E
e k z _ M
I—
r— 0—
Q l
o 1

Time (s)

Figure 2.9. Zoomed traces from figure 2.8. Obtained from bacteriochlorophyll a in
pyridine. The probe beam was detected at 757 nm. A portion of one scan of the pump-
probe delay is shown as a function of real time. Top: Probe delay as determined from the
position of the rapid-scanning delay stage. Middle: Zero-background autocorrelation.

Bottom: Single-wavelength dynamic-absorption transient.

40

L

AT I T (relative units)

1

 

 

 

I I I I I I
O 2000 4000 6000 8000 10000

Probe Delay (fs)

Figure 2.10. Single-wavelength dynamic-absorption transient of bacteriochlorophyll a
in pyridine solvent displayed as a function of probe delay. The probe beam was detected

at 757 nm.

41

2.4 References

(1) Wang, Q.; Schoenlein, R. W.; Peteanu, L. A.; Mathies, R. A.; Shank, C. V.
Science 1994, 266, 422—424.

(2) Vos, M. H.; Rappaport, F .; Lambry, J .-C.; Breton, J.; Martin, J.-L. Nature 1993,
363, 320—325.

(3) Vos, M. H.; Jones, M. R.; Hunter, C. N.; Breton. J.; Lambry, J.-C.; Martin, J.-L.
Biochemistry 1994, 33, 6750—6757.

(4) Vos, M. H.; Jones, M. R.; Hunter, C. N.; Breton, J.; Martin, J.-L. Proc. Nat.
Acad. Sci. USA 1994, 91, 12701—12705.

(5) VOS, M. H.; Jones, M. R.; McGlynn, P.; Hunter, C. N.; Breton, J.; Martin, J.-L.
Biochim. Biophys. Acta 1994, 1186, 1 17—122.

(6) Amett, D. C.; Moser, C. C.; Dutton, P. L.; Scherer, N. F. J. Phys. Chem. B 1999,
103, 2014—2032.

(7) Jonas, D. M.; Bradforth. S. E.; Passino. S. A.; Fleming, G. R. J. Phys. Chem.
1995, 99, 2594—2608.

(8) Heller, E. J. Acc. Chem. Res. 1981, 14, 368—375.

(9) Heller, E. J.; Sundberg. R. L.; Tannor, D. J. Phys. Chem. 1982, 86, 1822—1833.

(10) Lee, S.-Y.; Heller, E. J. J. Chem. Phys. 1979, 71, 4777—4788.

(1 1) Myers, A. B.; Mathies, R. A. In Biological Applications of Raman Spectroscopy;
Spiro, T. G., Ed.; Wiley—lnterscience: New York, 1987; Vol. 2, Resonance
Raman Spectra of Polyenes and Aromatics, p 1—58.

(12) Tannor, D. J.; Heller, E. J. J. Chem. Phys. 1982, 77, 202—218.

(13) Williams, S. 0.; lmre. D. G. J. Phys. Chem. 1988, 92, 3363—3374.

42

(14) Pollard, W. T.; Fragnito, H. L.; Bigot, J.-Y.; Shank, C. V.; Mathies, R. A. Chem.
Phys. Letters 1990, 168, 239—245.

(15) Pollard, W. T.; Lee. S.-Y.; Mathies. R. A. J. Chem. Phys. 1990, 92, 4012—4029.

(16) Pollard, W. T.; Dexheimer, S. L.; Wang, Q.; Peteanu, L. A.; Shank, C. V.;
Mathies, R. A. J. Phys. Chem. 1992, 96, 6147—6158.

(17) Pollard, W. T.; Mathies. R. A. Annu. Rev. Phys. Chem. 1992, 43, 497—523.
(18) Zhu, L.; Sage, J. T.; Champion. P. M. Science 1994, 266, 629—632.

(19) Peteanu, L. A.; Schoenlein, R. W.; Wang, Q.; Mathies, R. A.; Shank, C. V. Proc.
Natl. Acad. Sci. USA 1993, 90, 11762—11766.

(20) Schoenlein, R. W.; Peteanu, L. A.; Mathies. R. A.; Shank, C. V. Science 1991,
254, 412—415.

(21) Fragnito, H. L.; Bigot. J.-Y.; Becker, P. C.; Shank, C. V. Chem. Phys. Lett. 1989,
160,101—104.

(22) Brito Cruz, C. H.; Fork, R. L.; Knox, W. H.; Shank, C. V. Chem. Phys. Lett.
1986, 132, 341—345.

(23) Brito Cruz, C. H.; Gordon, J. P.; Becker, P. C.; Fork, R. L.; Shank, C. V. IEEE
J. Quant. Elect. 1988, 24, 261—266.

(24) Mathies, R. A.; Brito Cruz. C. H.; Pollard, W. T.; Shank, C. V. Science 1988,
240, 777—779.

(25) Pollard, W. T.; Brito Cruz, C. H.; Shank. C. V.; Mathies, R. A. J. Chem. Phys.
1989, 90, 199—208.

(26) Dexheimer, S. L.; Wang, Q.; Peteanu, L. A.; Pollard, W. T.; Mathies, R. A.;
Shank, C. V. Chem. Phys. Lett. 1992, 188, 61—66.

(27) Diffey, W. M.; Beck, W. F. Rev. Sci. Instrum. 1997, 68, 3296—3300.

43

(28) Diffey, W. M.; Homoelle, B. J.; Edington, M. D.; Beck, W. F. J. Phys. Chem. B
1998, 102, 2776—2786.

(29) Carson, E. A.; Diffey, W. M.; Shelly, K. R.; Lampa-Pastirk, S.; Dillman, K. L.;
Schleicher, J. M.; Beck, W. F. J. Phys. Chem. A 2004, 108, 1489—1500.

(30) Shelly, K. R.; Carson, E. A.; Beck, W. F. J. Am. Chem. Soc. 2003, 125, 11810—
11811.

(31) Edelstein, D. C.; Romney, R. B.; Scheuermann, M. Rev. Sci. Instrum. 1991, 62,
579—583.

(32) Feldstein, M. J.; VOhringer, P.; Scherer, N. F. J. Opt. Soc. Am. B 1995, 12, 1500—
1510.

44

CHAPTER 3

VIBRATIONAL COHERENCE FROM THE DIPYRIDINE COMPLEX OF
BACTERIOCHLOROPHYLL a: INTRAMOLECULAR MODES IN THE 10—220—cm"
REGIME, INTERMOLECULAR SOLVENT MODES, AND RELEVENCE TO
PHOTOSYNTHESIS'

3.0 Introduction

When impulsive (sub-vibrational-period) pump pulses are employed in experiments
with reaction centers from purple bacteria, the stimulated emission and fluorescence
signals from P exhibit intensity modulations at several frequencies over the 10—200-cm'l
regime that persist as long as the 3-ps time scale for electron transfer to the
bacteriopheophytin acceptor, BPhL.L2 The modulations arise from coherent wave-packet
motions (vibrational coherence) on the excited-state potential-energy surface of P, but the
nature of the normal modes that contribute to the signal is still debated. Because previous
attempts to observe underdamped vibrational coherence from monomeric BChl in
solution were unsuccessful}4 and because changes in the observed modulation
frequencies occur in mutants,S an assignment to BChl—protein intermolecular modes or to
protein-derived modes was suggesteds’6 Resonance-Raman spectra obtained from
reaction centers with isotopically labeled cofactors and nonnal-coordinate analyses
suggest, in contrast, that intramolecular BChl modes are involved.7'8 Several of these
modes have an out-of-plane character that would modulate the strong electronic coupling
and intramolecular charge-transfer properties of P.7'l3 There is the additional possibility

that P exhibits a collective, intrapair mode that does not occur in BChl monomers.”l5

 

I The material in this chapter has been published in another form: Shelly. K. R.; Carson, E. A.; Beck, W. F.
J. Am. Chem. Soc. 2003, 125, 11810—11811.

45

The first observations of vibrational coherence in the 10-220-cm'I region from
bacteriochlorophyll a (BChl) in solution are presented in the following sections. A
distinction can be made for the first time between BChl’s intramolecular normal modes
and intermolecular modes between BChl and solvent. The results discussed in this
chapter show that the low-frequency vibrations that accompany the initial electron-
transfer reaction from the paired-BChl primary electron donor, P, in photosynthetic

8 arise predominantly from intramolecular modes of histidine-ligated

reaction centersw"
BChl macrocycles. The results also suggest that polar solvent interactions can

significantly perturb the electronic properties of BChl in a manner that might have

important functional consequences.

3.1 Experimental
3.1.1 Sample Preparation

Synthetically prepared bacteriochlorophyll a (Frontier Scientific) was dissolved in
anhydrous pyridine (spectrophotometric grade, acquired from Jade Scientific) to obtain
an absorbance of 0.6 at 780 nm for a l-mm path length. The sample was manipulated
under a dry nitrogen atmosphere. It was held at room temperature (22 °C) in a fused-
silica flow cuvette; the path length was 1 mm. The bacteriochlorophyll a solution was
passed through a 0.22-pm filter prior to use. BChl samples for femtosecond spectroscopy
were held in a 1-mm, fused-silica flow cuvette. Samples were circulated using a
peristaltic pump through the cuvette at a rate of 6 mL/min. The absorption spectrum of
the sample was monitored throughout the duration Of the experiment for changes arising

from photochemistry or permanent photobleaching.

46

3.1.2 Continuous-Wave Spectroscopy

Continuous-wave absorption spectra for all samples were Obtained with a Hitachi
U-2000 spectrometer at 22 °C with 2-nm spectral band pass. Fluorescence spectra were
acquired at room temperature with a Hitachi F -4500 fluorescence spectrophotometer
(5-nm excitation and emission band pass). The fluorescence intensities were corrected
using a calibration curve obtained from a standard lamp. When presented as a function of
wavenumber, the fluorescence intensities are multiplied by the square of the wavelength
in order to compensate for the fixed (in wavelength units) spectral band pass of the

. . 7
emrssron spectrometer.‘0‘2'

3.1.3 F emtosecond Spectroscopy

Dynamic-absorption spectroscopy was performed with the femtosecond pump-probe
spectrometer described in Chapter 2, figures 2.2 and 2.5-2.7. In the experiments
described in this chapter, the output spectrum emitted by the oscillator exhibited a
bandwidth of 12 nm, 250 cm'I (fwhm), centered at 780 nm. The extracavity SF 10 prisms
were adjusted to produce 50-fs pulses (sechz). The transmitted probe light was detected

at 787 nm (2-nm band pass).

3.2 Results

3.2.1 Continuous-Wave Spectroscopy

Figure 3.1 shows the absorption and fluorescence Spectra exhibited by BChl in
pyridine solvent at room temperature. The spectra are plotted as A(V)/ V and F (v)/v3 ,

respectively, with normalization to unit area. The integrals of these quantities report the

47

Wavelength (nm)

 

 

 

 

 

 

820 800 780 760 740
l 1
Probe ’r‘
1‘16: f, “
. . 1?. ' ‘. Laser
'5 ' '1
8’ ‘.
9 ‘.
if) 3 .
2 F/v
__3_ A/v ‘I‘
Q |
1
0 ‘ ' 4 " ‘
I l I
12.0 12.5 13.0 13.5

Energy (103 cm“)

Figure 3.1. Continuous-wave absorption and fluorescence spectra from

bacteriochlorophyll a in pyridine plotted as relative dipole strengths, A(V)/V

and F (v) / v3 , respectively. The spectra are normalized to unit area. The fluorescence

spectrum was excited at 790 nm (12700 cm").

48

dipole strength; the square of the transition-dipole moment, which is proportional to the
Einstein coefficients for absorption and stimulated emission, Bab and Bba,
respectively.”26 As plotted, the sum of the spectra can be compared to the time-
integrated dynamic-absorption spectrum in the weak-field limit.27 The 0—0 vibronic
transition is found in the vicinity of 12700 cm'l (790 nm), where the two normalized
spectra cross. Also shown in figure 3.1 is the intensity spectrum from the tizsapphire laser
that was the source of pump and probe pulses in the dynamic-absorption experiments.
The shaded region in figure 3.1 coincides with the portion of the transmitted probe
spectrum that was passed by the monochromator to the photodiode.

In the dynamic-absorption experiments, the titanium sapphire laser's spectrum
(12600—13000 cm", with peak intensity at 12800 cm!) excites BChl over a spectral range
that starts just to the red of the absorption maximum, passes through the 0—0 transition,
and proceeds on to cover the entire peak of the absorption band. The laser spectrum also
spans a large fraction of the blue tail of the fluorescence spectrum. Accordingly, as used
in the dynamic-absorption experiments as the probe beam, the laser Spectrum covers
some of the stimulated emission spectrum and the ground-state depletion spectrum,

allowing for detection of both ground-state and excited-state vibrational coherence.

3.2.2 Dynamic-Absorption Experiments

The dynamic-absorption transient Obtained from BChl in pyridine is shown in figure
3.2. Following an intense coherence spike near the zero of time, the transient exhibits
two regions Of vibrational coherence. These regions are more easily discernible in the
oscillatory portion of the pump—probe signal. The oscillatory signal is isolated by

subtracting a fitted triple-exponential function from the Signal starting at 200 fs and

49

.3

O

O
I

01
o
l

 

O
s

 

AT I T (relative units)
{a
I

 

 

 

—-i

I l I I
-2000 0 2000 4000 6000 8000

Probe Delay (fs)

Figure 3.2. Pump-probe signal obtained from BChl in pyridine solvent.

50

extending to the end Of the recorded range at 8 ps. The triple exponential fit and
corresponding residual oscillatory fraction can be seen in the top and bottom panels,
respectively, of figure 3.3.

Figure 3.3 shows the 200—7000-fs delay range of the oscillatory portion of the pump—
probe signal obtained from BChl in pyridine with excitation of the Qy transition. The
first region of vibrational coherence is a relatively intense, rapidly damped feature in the
200—600-fs regime. Following the rapidly damped section, a second region Of weaker
underdamped cosinusoidal modulations is observed that persists even to the 8-ps delay
point. Because the probe wavelength lies just to the blue of the fluorescence spectrum,
the modulations mostly arise from the excited-state wave packet in the near-tuming-point
region.24

The Fourier-magnitude spectrum corresponding to the slowly-damped modulations is
shown in figure 3.4. The spectrum reports at least twelve reproducible modulation
components, with center frequencies ranging from 1 1 to 206 cm". Above 30 cm], the
spectrum strongly resembles the resonance-Raman spectra from P and/or the monomeric

7,29-31

species BChlL and BCth in the reaction center in having significant peaks at

138 cm’l and 183 cm", respectively.

3.2.3 Rapidly-Damped Vibration Coherence from Bacteriochlorophyll a and Pyridine

In this section, we consider in more detail how much the pyridine solvent contributes
to the BChl pump—probe Signal in the region of time where the strong, rapidly-damped
oscillation is observed. Figure 3.5 compares the pump—probe signal Obtained at room

temperature from BChl in pyridine solvent with the signal Obtained under the same ex-

51

AT / T (relative units)

 

 

 

I I T I
2000 4000 6000 8000

Probe Delay (fs)

'01
J

'o
l

.0
01
l

 

5':
0'1
1

 

'o
1

AT I T (relative units)

 

 

 

I I I T I l l
0 1 000 2000 3000 4000 5000 6000 7000

Probe Delay (fs)

Figure 3.3. Triple exponential fit (top) and corresponding residual oscillatory portion of

the pump-probe signal from BChl in pyridine (bottom).

52

11

Relative Magnitude
33
49
59
69
80
103
1 16
138
169
183
206

 

 

O I I I I
0 50 100 150 200
Frequency (cm")

Figure 3.4. Hanning-windowed Fourier-magnitude spectrum of the 200-8000-fs range.

53

perimental conditions from neat pyridine solvent. Note that the y axes of the two graphs
are set up so that the relative intensities of the two signals can be determined by
inspection. The discussion that follows shows that we can neglect any contribution of the
nonresonant oscillatory signal from pyridine to the signals Shown in figures 3.2-3.3 and
3.7.

The BChl pump—probe signal shown in figure 3.5 is the same one used in figures 3.2-
3.3 and 3.7. The most intense feature is the bipolar pump—probe coherence spike near the
zero of time. This spike is superimposed on the pump-induced change in transmittance
that can be attributed to the change in ground-state population arising from the 7: —) 75‘
transition; the intensity change from the baseline prior to time zero and nondecaying
(y >1 ns) fraction corresponds to a AT/T of 10". As stated in the caption to figure 3.5, the

full range of the oscillatory signal used in figure 3.2 and in the subsequent analysis of the
components of the vibrational coherence is perhaps ten percent as large, a AT/T of 105.
In terms of the relative-intensity units used in figures 3.2 and 3.5, there is a full-scale
modulation of about 2.5 units.

The pump—probe Signal observed in neat pyridine arises from the optical-Kerr effect
(OKE). It exhibits a spike near the zero of time that arises from the nonresonant electric
hyperpolarizability, a weak beating modulation near the 150-fs time point, and a
subsequent exponential decay that arises from diffusive reorientational relaxation. The
signal is very similar in shape to that reported in the past by McMorrow and Lotshaw,33
but the laser pulses used in our experiment are shorter in duration. The signal/noise ratio
of our signal is poor compared to that of the McMorrow and Lotshaw signal, even after

many hours of signal averaging, because we detect the signal against a large transmitted-

54

 

 

 

 

 

 

 

 

 

BChl a in Pyridine
4o—
30— T
20—
.52 10—
S
o 0 41
£- 1 I I I
g -2000 o 2000 4000 6000 3000
L
l— 5—
: 4- PyridineBlank
3....
2'1
1—
o-..__L.J — ——— —- A.;—4 - 1.2.
I I I I I
-2000 0 2000 4000 6000 8000

Probe Delay (fs)

Figure 3.5. Pump-probe signal obtained under the same Optical conditions from BChl in
pyridine solvent and in neat pyridine. The top graph shows the range corresponding to
the population change in the BChl ground state. The y axes employ a consistent relative-

intensity scaling.

55

probe background. (The best technique for detection Of nonresonant OKE signals is to
rotate the probe-analyzing polarizer so that it is crossed (obtaining maximum extinction)
against the probe’s plane of polarization; then, by a small rotation of a M4 retarder plate
or of the analysis polarizer. a small leakage signal from the probe is admitted to obtain
Optical-heterodyne detection (OHD) conditions. SO, the OHD-OKE signal is detected
essentially against a zero background, and the signal/noise ratio is appropriately much
better than that obtained under our conditions. In effect, our analysis polarizer is rotated
way beyond the limit needed to obtain full heterodyning.) Our intention, here, of course,
is not to record the OHD-OKE signal but to obtain a valid estimate of how much the
pyridine solvent itself contributes a nonresonant background signal in the pump—probe
experiment with BChl present, under the same Optical conditions.

Over the time range from 150 fs to 7 ps Shown in figure 3.3, the pyridine blank
exhibits a full-range modulation Of much less than 1 relative-intensity unit. Figure 3.5
shows that the Significant OKE modulations observed in pyridine occur at shorter delay
times and exhibit a different intensity profile than those observed in the BChl/pyridine
sample. Further, the BChl and pyridine-OKE modulation spectra are not the same. The
pyridine OKE spectrum (not shown), obtained using the methods of the next section,
exhibits a peak in the 60—90-cm‘l regime and substantial intensity in the 0—60-cm'l
region; considering the effects of our shorter laser pulses. the spectrum is consistent with
that reported previously by McMorrow and Lotshaw.33

A direct comparison of intensities between the BChl and pyridine-blank sample is not
appropriate because it does not account for the attenuation of the laser intensity in the

BChl sample. In the BChl pump—probe experiment. the absorbance was 0.6 at the center

56

of the laser spectrum. With respect to the background pyridine OKE signal, the presence
of BChl contributes an attenuation of the laser fluence for the pump and probe beams and
it attenuates the exiting third-order polarization signal (the OKE signal) beam to the same
degree. Accordingly, in the presence of BChl, the background pyridine OKE signal is
approximately 1/(10'0‘6)3 = 1.6 x 10'2 = 1/163 as large as that observed in neat pyridine.
Given the very small amplitude of this signal over the range of the BChl signal that is
shown in figure 3.5, it is certainly safe to ignore the presence of the OKE signal in the

analysis.

3.3 Data Analysis
3.3.1 Fourier-Magnitude Spectrum Estimation

The oscillatory fraction of the pump—probe signal (see figure 3.3) was obtained by
subtracting a fitted triple-exponential function from the signal segment starting well
beyond the pump—probe coherence spike, at the ISO-f5 delay point, and extending to the
end of the acquired signal. It was then multiplied by a Hanning (or raised cosine) window

function,38

227k
W(k) —0.5[I—COS(mJ] (3.1)

defined for the point index k in the data segment in terms of the number of data points n.
The window function gradually forces the intensity at the beginning and end of the Signal
segment to zero amplitude. This procedure is required to suppress satellite ripples in the
Fourier-magnitude spectra, which arise from leakage of signal frequencies into adjacent
channels owing to truncation of the time-domain signal at the beginning and end. The

window function also causes a broadening of the observed signal line shape at a given

57

frequency, but this same broadening can be applied in simulation Of the frequency-
domain spectrum to return accurate damping constants. Next, the windowed signals were
zero-padded by a factor of 32 in order to enhance the point density along the frequency
axis in the Fourier-magnitude spectra. Lastly, the spectrum was compensated for the
finite width of the instrument-response function by deconvolution in the frequency
domain: the raw magnitude was scaled by the reciprocal of the Fourier magnitude at
frequency a),- Obtained from the pump-probe autocorrelation signal.39 This analysis is
shown in detail in figure 3.6. The top panel shows an oscillatory signal with a frequency
of 200 cm'I and a damping time of 100 fs. The middle panel shows the signal after
multiplication by the Hanning window function described by equation 3.1, where the
intensity at the beginning and end of the signal go to zero amplitude. The resulting
Fourier magnitude spectrum is shown in the bottom panel of figure 3.6.

The reader should take note that the deconvolution step mentioned above causes the
signal/noise to deteriorate as the frequency increases. This effect arises from the need to
have the laser pulses be much shorter than the period of a given mode’s vibrational
period. With our SO-fs pulses, we find that we cannot reproduce features in the Fourier-
magnitude spectrum above 220 cm". At that point, the power in the pump—probe
autocorrelation function is about 50 percent of that at 0 cm". The resolution Of the
Fourier-magnitude spectrum is determined by the length of the analyzed time-domain
data segment and the windowing function. Using the modulated portion of the BChl
signal shown in figure 3.3, the effective resolution in the calculated Fourier-magnitude

spectrum is 10 cm".

58

 

Intensity

 

 

 

 

 

 

 

 

o _
I I I I l
0 200 400 600 800 1000
Time(fs)
.3:
g o —
E
I I I I I
0 200 400 600 800 1000
Time(fs)
o 1 1
‘O
3
'E
c»
m
2
.3
I!
a?
0 _
I I I I I
0 100 200 300 400 500

Frequency (cm")

Figure 3.6. Oscillatory signal with a frequency of 200 cm'1 and a damping time of 100 fs
(top panel). Oscillatory signal from top panel multiplied by a Hanning window function
described by equation 3. 1 (middle panel). Hanning-windowed F ourier—magnitude

spectrum of the 200-cm'I oscillation (bottom panel).

59

3.3.2 Modeling ofthe Slowly Damped Components.
The parameters listed in table 3.1 were obtained by a robust frequency-domain
modeling procedure.24 The experimental Fourier-magnitude spectrum (figure 3.4) was

modeled using a sum Of damped cosines,

1(1) = ZAIe—t / 7" cos(a),-t + (15,-) (3.2)

l
The model signal was sampled using the delay times t and segment lengths Obtained from
the experimental signal (see figure 3.2); it was processed by the same sequence of
windowing, zeropadding, and Fourier-magnitude spectrum calculation. The intensities A,,
frequencies 14, and damping times y, were adjusted iteratively to reproduce the intensity,
position, and width of each significant feature in the experimental F ourier-magnitude
spectrum. Note that the phases ¢, are optionally obtained by optimization of the model in
the time domain. The components that were included in the model are those that were

reproduced in frequency and relative intensity in a set Of replicate experiments.

3.3.3 Modeling ofthe Rapidly Damped Component

The broad line shapes that would be expected from the rapidly-damped region are not
observed in the frequency domain due to the Hanning window function used to obtain the
F ourier-magnitude spectra. The Hanning window greatly reduces the intensity of this
time region, so the F curler-magnitude spectra are dominated by the vibrational coherence
Observed on a longer timescale (r >1000 fs). Accordingly, the rapidly damped
modulation component observed over the 200—600-fs time region (see figures 3.2 and

3.3) was modeled using a distribution of cosinusoidal components defined by the sum of

60

Table 3.1. Frequencies, normalized deconvolved intensities,“ and damping constants for
the underdamped oscillatory componentsb Observed in the pump-probe signal from

bacteriochlorophyll a in pyridine solvent

 

 

frequency (cm'l) intensity8 7 (fs)
I I 1.00 800
33 0.09 I 100
49 0.03 1500
59 0.02 1700
68 0.03 1500
80 0.03 1500
103 0.02 1700
l 16 0.04 1500
138 0.26 1000
169 0.12 1200
182 0.18 1200
206 0.16 1300

 

atrelative to the magnitude of the l l-cm'l component after deconvolution in the frequency

domain

bmodeled by a sum Of damped cosinusoids. Z, 4.0—"’8 COSUUJ)

61

two lognormal (asymmetric Gaussian)40 distribution functions, resulting in a model signal
that closely approximates the time-domain signal (see figure 3.3). (This model will be
discussed in more detail in Chapter 4.) The distribution was sampled at 5-cm'l intervals
from 1 cm'I to 1000 cm". The intensity of the distribution at each sampling point was
used as the scaling factor for a damped cosinusoid, as in equation 3.2, above. The
damping factor for each component was arbitrarily fixed at 1 ps; changing this parameter
(say, to 2 ps) had little effect on the time-domain signal obtained by summing over the
entire distribution.

The parameters for the two lognormal components were iteratively adjusted until a
close facsimile of the observed signal was obtained (see the top panel of figure 3.7). The

final parameters are listed in table 3.2.

3.4 Discussion
3.4.1 Observation of Two Damping Time Scales

Two damping time scales are Observed in the ground-state vibrational coherence from

BChl in pyridine. The difference in line shape for the two regions Of vibrational

coherence indicates that they arise from two distinct classes of vibrational modes.

We suggest that the rapidly-damped vibrational components arise from
intermolecular modes between the BChl macrocycle and the surrounding solvent. The
rapidly-damped modes have inhomogenously broadened line shapes that can be described
by a distribution of damped cosinusoids. The distribution arises from the disordered
ensemble of solvated bacteriochlorophyll structures.

Similar broad spectra are observed in nonresonant Optical-Kerr-effect (OKE) signals

62

 

 

AT/ T (relative units)
0
O

 

 

I I
250 300 350 400 450 500
Probe Delay (fs)

 

Relative Magnitude

 

 

I I I I I I
100 150 200 250 300 350

Frequency (cm'1)

Figure 3.7. Expanded view of the rapidly damped oscillation from BChl in pyridine
solvent (see figure 3.5.) superimposed with a model (dashed curve) arising from a

distribution of mode frequencies, as shown in the bottom panel.

63

Table 3.2. Log-normal line shape parameters for the solvent-mode distribution“

 

component amplitude frequency (cm") width (cm") skew
1 0.68 145 35 1.2
2 l 175 47 1.2

 

 

amodeled using a distribution of cosinusoidal components defined by the sum Of two

lognormal (asymmetric Gaussian)40 distribution functions

64

from liquids, which carry modulations arising from librational (hindered reorientational)
motions.32 But the OKE spectrum from pyridine liquid is Observed at a lower frequency,
60 cm'l, with collective modes making a contribution at 15 cm'l.33 The resonant origin

of the present signal requires the involvement of a BChl—solvent-pyridine intermolecular

mode that is displaced by the BChl n—m" transition. The relatively high frequency favors

an assignment to a perpendicular hydrogen-bonding interaction like that present in

34,35

pyrrole, with protons of non-ligated pyridines in the first solvent shell attacking the

BChl rr-electron density.

In addition to the rapidly-damped components. slowly-damped vibrational
components are observed arising from intramolecular motions from the BChl macrocycle
itself. The intramolecular modes exhibit comparable intensities over the 3O—200-cm’l
region, but the l l-cm'l mode and the pyridine-solvent mode yield much stronger signals.
As indicated in table 3.2, the relatively intense I l-cm'l component (3-ps period) exhibits

the broadest line shape (damping constant 7: 800 fs); the higher-frequency components
exhibit sharper line shapes (7: 1.2—1.7 ps). The observation of the sharp line shapes

supports the suggestion7 that the very lowest-frequency components are affected by the
nature of the axial ligand to the Mg(ll) ion. Owing to the long damping times, in the
range of those observed for intramolecular modes of organic molecules in solution,”28
we know that these features arise from intramolecular modes belonging to the BChl—r
dipyridine complex.

These observations prompt the following conclusions about the role that BChl—

protein “solvent” interactions might play in the reaction center:

1. Owing to its low frequency. the l l-cm'l mode involves motion of a large reduced

65

mass along a coordinate with a weak force constant, such as an out-of-plane distortion of
the BChl macrocycle. We suggest that resonance enhancement of this mode involves
coupling to the pyridine—solvent mode, but it is not Obvious why the other low-frequency
modes, several of which have some degree of out-Of-plane character}8 are not as
strongly enhanced. The low-frequency intramolecular modes of BChls in the reaction
center might be comparably affected by mode-specific and highly directional BChl—
protein interactions. In fact. the resonance-Raman spectra from BChlLM and P exhibit

7'8‘29'3 1 that probably arise from the details Of the

distinct low-frequency intensity patterns
binding sites.

2. The broad line shape for the solvent mode in pyridine liquid arises from the
disordered ensemble of solvated BChl structures. The strength of this signal suggests that
some of the changes in the vibrational coherence from P in mutant reaction-center
preparations5 might arise from an ordered solvent interaction, which would exhibit a
narrow line shape, between P and the surrounding protein.

The observation of polar-solvent effects on the resonance-Raman intensities from the
BChl-dipyridine complex imply that significant changes in the electronic structure of the
BChl macrocycle should arise from protein-derived solvent perturbations in the reaction
center. These interactions might contribute to the protein’s control over the reactivity of a
BChl site despite not necessarily causing a large change in the geometry. Some larger
effects on the resonance-Raman spectra and charge-transfer properties of BChlL and

BPhL owing to an introduction of neighboring charges have already been reported.”37

66

3.5 References

(1) Vos, M. H.; Rappaport, F.; Lambry, J.-C.; Breton, J.; Martin, J.-L. Nature 1993,
363, 320—325.

(2) Stanley, R. J.; Boxer, S. G. J. Phys. Chem. 1995, 99, 859—863.

(3) Savikhin, S.; Zhu, Y.; Lin, S.; Blankenship, R. E.; Struve, W. S. In Ultrafast
Phenomena IX; Barbara, P. F ., Knox, W. H., Mourou, G. A., Zewail, A. H., Eds.;
Springer-Verlag: Berlin, 1994, p in press.

(4) Chachisvilis, M.; Fidder. H.; Pullerits, T.; SundstrOm, V. J. Raman Spect. 1995,
26, 513—522.

(5) Rischel, C.; Spiedel, D.; Ridge, J. P.; Jones, M. R.; Breton, J.; Lambry, J.-C.;
Martin, J.-L.; VOS, M. H. Proc. Natl. Acad. Sci. USA 1998, 95, 12306—12311.

(6) VOS, M. H.; Martin, J.-L. Biochim. Biophys. Acta 1999, 1411, 1—20.

(7) Czamecki, K.; Diers, J. R.; Chynwat, V.; Erickson, J. P.; Frank, H. A.; Bocian, D.
F. J. Am. Chem. Soc. 1997, 119, 415—426.

(8) Palaniappan, V.; Aldema. M. A.; Frank. H. A.; Bocian, D. F. Biochemistry 1992,
31,11050—11058.

(9) Warshel, A.; Parson, W. W. J. Am. Chem. Soc. 1987, 109, 6152—6163.

(10) Warshel, A.; Parson, W. W. J. Am. Chem. Soc. 1987, 109, 6143—6152.

(1 1) Thompson, M. A.; Zerner, M. C.; Fajer, J. J. Phys. Chem. 1991, 95, 5693—5700.
(12) Thompson, M. A.; Zerner, M. C.; Fajer, J. J. Phys. Chem. 1990, 94, 3820—3828.
(13) Lathrop, E. J. P.; Friesner, R. A. J. Phys. Chem. 1994, 98, 3056—3066.

(14) Warshel, A. Proc. Natl. Acad. Sci. USA 1980, 77, 3105—3109.

(15) Hayes, J. M.; Small, G. J. J. Phys. Chem. 1986, 90, 4928—4931.

(16) Parson, W. W. In Photosynthesis (New Comprehensive Biochemistry, Volume
15); Amesz, J., Ed.; Elsevier: New York, 1987, p 43—61.

(17) Boxer, S. G.; Goldstein, R. A.; Lockhart, D. J.; Middendorf, T. R.; Takiff, L.
J. Phys. Chem. 1989, 93, 8280—8294.

(18) Friesner, R. A.; Won, Y. Biochim. Biophys. Acta 1989, 977, 99—122.

(19) Alden, R. G.; Cheng. W. D.; Lin, S. H. Chem. Phys. Lett. 1992, 194, 318—326.

67

(20) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Second ed.; Kluwer
Academic/Plenum Publishers: New York, 1999.

(21) Lampa-Pastirk, S.; Lafuente, R. C.; Beck, W. F. J. Phys. Chem. B 2004, 108,
12602—12607.

(22) Diffey, W. M.; Beck, W. F. Rev. Sci. Instrum. 1997, 68, 3296—3300.

(23) Diffey, W. M.; Homoelle, B. J.; Edington, M. D.; Beck, W. F. J. Phys. Chem. B
1998, 102, 2776—2786.

(24) Carson, E. A.; Diffey, W. M.; Shelly, K. R.; Lampa-Pastirk, S.; Dillman, K. L.;
Schleicher, J. M.; Beck, W. F. J. Phys. Chem. A 2004, 108, 1489—1500.

(25) Shelly, K. R.; Carson. E. A.; Beck, W. F. J. Am. Chem. Soc. 2003, 125, 11810—
11811.

(26) Cantor, C. R.; Schimmel. P. R. Biophysical Chemistry. Part [1: Techniques for
the Study of Biological Structure and Function; W. H. Freeman and Company:
San Francisco, 1980.

(27) Pollard, W. T.; Mathies, R. A. Annu. Rev. Phys. Chem. 1992, 43, 497—523.
(28) Bardeen, C. J .; Shank, C. V. Chem. Phys. Lett. 1994, 226, 310—316.

(29) Shreve, A. P.; Cherepy. N. J.; Franzen, S.; Boxer, S. G.; Mathies, R. A. Proc.
Natl. Acad. Sci. USA 1991, 88, 1 1207—1 1211.

(30) Cherepy, N. J.; Shreve, A. P.; Moore, L. J.; Franzen, S.; Boxer, 8. G.; Mathies,
R. A. J. Phys. Chem. 1994, 98, 6023—6029.

(31) Cherepy, N. J.; Shreve, A. P.; Moore, L. J.; Boxer, S. G.; Mathies, R. A. J. Phys.
Chem. B 1997, 101, 3250—3260.

(32) Lotshaw, W. T.; McMorrow, D.; Thantu, N.; Melinger, J. S.; Kitchenbaum, R.
J. Raman Spect. 1995, 26, 571—583.

(33) McMorrow, D.; Lotshaw, W. T. Chem. Phys. Lett. 1993, 201, 369—376.
(34) Gamba, Z.; Klein, M. L. J. Chem. Phys. 1990, 92, 6973—6974.
(35) Wynne, K.; Galli, C.; Hochstrasser, R. M. Chem. Phys. Lett. 1992, 193, 17—22.

(36) Cua, A.; Kirmaier, C.; Holten. D.; Bocian, D. F. Biochemistry 1998, 37, 6394--
6401.

(37) Czamecki, K.; Kirmaier, C.; Holten, D.; Bocian, D. F. J. Phys. Chem. A 1999,
103, 2235—2246.

68

(38) Press, W. H.; Flannery, B. P.; Teukolsky, S. A.; Vetterline, W. T. Numerical
Recipes: the Art of Scientific Computing; Cambridge University Press:
Cambridge, 1986.

(39) McMorrow, D.; Lotshaw, W. T. Chem. Phys. Lett. 1990, 174, 85—94.

(40) SianO, D. B.; Metzler. D. E. J. Chem. Phys. 1969, 51, 1856—1861.

69

CHAPTER 4

INTERMOLECULAR VIBRATIONAL COHERENCE BETWEEN
BACTERIOCHLOROPHYLL 0 AND POLAR AND NONPOLAR SOLVENTS

4.0 Introduction

We focus in this chapter on the structural origin of the rapidly damped components of
the vibrational coherence that can be observed in BChl solutions as discussed in chapter
3. The intention Of this work is to test our assignment of the rapidly damped signals to
modes involving the solvent molecules. We compare the rapidly damped vibrational
coherence Observed at room temperature in BChl a solutions in pyridine, acetone, and 1-
propanol. We also present the rapidly damped vibrational coherence observed in BChl a
in a nonpolar solvent, cyclohexane. The results support an assignment of the vibrational
coherence to intermolecular modes between BChl and clustered solvent molecules in the
first solvation shell because the mode frequency scales with the dipole moment of the
polar solvent. The evidence of vibrational coherence in BChl a solutions in cyclohexane
suggests that the intermolecular binding potential also depends on the polarizability of the
solvent. This work shows for the first time that intermolecular modes between large

electronic chromophores and the surrounding solvent can be resonance Raman active.

4.1 Experimental Procedures
4.1.1 Sample Preparation
4.1.1.1 Bacteriochlorophyll a in polar solvents
Synthetically prepared bacteriochlorophyll a was used as received from Frontier

Scientific. Samples for femtosecond spectroscopy were prepared in a glove bag under a

70

dry N2 atmosphere and in minimal light. Pyridine (spectrophotometric grade) was
obtained from Jade Scientific; acetone was obtained from J. T. Baker; l-propanol was
obtained from CCI. The dry BChl powder was dissolved in enough solvent to obtain an
absorbance of 0.6—0.7 at 750 nm for a 1-mm path length. The samples were passed

through a 0.22-pm filter prior to use.

4.1.1.2 Bacteriochlorophyll a in cyclohexane

Bacteriochlorophyll a was used as received from Frontier Scientific. Samples were
prepared under the same conditions as described above. Pyridine (spectrophotometric
grade) was obtained from Jade Scientific; cyclohexane (spectrophotometric grade) was
obtained from Spectrum. To the calculated volume of cyclohexane required to dissolve
1 mg of dry BChl powder at an absorbance Of 0.6-0.7 at 750 nm for a l-mm path length,
a certain number of equivalents of pyridine was added to establish a chosen

BChlzpyridine ratio. The samples were passed through a 0.22-pm filter prior to use.

4.1.2 Continuous-Wave Spectroscopy
Continuous-wave absorption and fluorescence spectra for all samples were acquired
at room temperature with Hitachi U-2000 and Hitachi F -45003pectrometers, respectively.

The fluorescence intensities were corrected as described in chapter 3.

4.1.3 F emtosecond Spectroscopy
Dynamic-absorption spectroscopy was performed with a self-mode-locked tizsapphire
oscillator and a rapid-scanning, modified Mach-Zehnder interferometer. The

spectrometer’s main features have been described in Chapter 2. For these experiments,

71

the tizsapphire oscillator was pumped by a Verdi (Coherent) ND-YVO4 5-W laser. The
oscillator was tuned to obtain a spectrum centered at 750 nm; the bandwidth was 11.5 nm
(fwhm). The width of the zero-background autocorrelation function at the sample
position was 60-fs fwhm (sechz) in these experiments. Probe light was detected at 757
nm (2-nm band pass). The differential probe detection scheme, as described in Chapter
2, was employed in these experiments to partially cancel laser noise.

BChl samples for femtosecond spectroscopy were held at room temperature (22 °C)
in a fused-silica flow cuvette; the path length was 1 mm. Samples were circulated using
a peristaltic pump through the cuvette at a rate of 6 mL/min. The absorption spectrum of
the sample was monitored throughout the duration of the experiment for changes arising

from photochemistry or permanent photobleaching.

4.2 Results

4.2.1 Continuous-Wave Spectroscopy

Figures 4.1 and 4.2 Show the continuous-wave absorption spectra from BChl a in
cyclohexane solvent at various BChlzpyridine ratios and in the three polar solvents;
pyridine, 1-propanol and acetone. The spectra span the three absorption bands arising
from 7: —> n' transitions: with increasing wavelength, the Soret or B band (~420-nm,
from SO —> S, transitions), and the Q, and Q). bands (~6OO nm and ~750 nm, from
SO -> S, transitions).

The spectra from BChl in cyclohexane show the effects of the binding of pyridine to
an axial ligation site on the central Mg(II) ion. The spectra from the 120.6 BChlzpyridine

ratio shows a red shoulder in the 820-nm range that evidences the presence Of BChl-BChl

aggregates in the sample. This red shoulder has largely disappeared in the 1:1.2 sample,

72

 

 

 

 

 

 

C
.9
E
8 B
.0
<1:
0.)
.2
70'
60“
n:
C
0—
I I I I r
300 400 500 600 700 800 900

Wavelength (nm)

Figure 4.1. Continuous-wave absorption spectra at 22 °C from BChl a in cyclohexane
solvent at BChlzpyridine ratios of(a)1:0.6, (b)1:1.2, and (c)1:100. The vertical line is at
580 nm.

73

 

0-1

0-1

Relative Absorption

 

 

 

T l l I l l
300 400 500 600 700 800 900
Wavelength (nm)

Figure 4.2. Continuous-wave absorption spectra at 22 0C from BChl a in three neat

solvents: (a) pyridine,(b) 1-propanol, and (c) acetone.

so this sample mostly contains BChI monomers. The 1:100 sample is red-shifted and
narrowed somewhat due to the presence of excess pyridine in the bulk solvent. All three
samples show a QJr absorption in the 580-nm range, which shows that the central Mg(II)
ion in the BChl macrocycle (see figure 4.3) ligates at most a single pyridine . The
presence of an axially ligated pyridine is necessary to avoid the formation of aggregates
because the peripheral ketone functional groups tend to find the axial binding sites.

Figure 4.2 shows that the Q, bands observed in 1-propanol and acetone are shifted to
the blue from that in pyridine, by 4 and 7 nm, respectively, owing to bulk
solvatochromism. Similar small shifts would be anticipated for the QJr bands, but the Q,
band Observed in acetone is shifted to the blue farther than would be expected by solvato-
chromism alone from the Q, bands Observed in pyridine and 1-propanol. Katz and
coworkers"3 showed that this shift arises from a change in axial coordination for the
Mg(lI) ion bound by the BChl macrocycle. Observation of the Q; band for BChl a in the
610-nm region indicates the presence of two axially ligated solvent molecules, so the
Mg(II) ion is six-coordinate in pyridine and l-propanol, with two solvent molecules
bound axially in each case.2 The 580-nm Q, band observed in acetone, however, indicates
that the Mg(ll) ion coordinates only a single acetone ligand.

The pump—probe experiments reported in this chapter were conducted with excitation
of the Qy absorption band. Figure 4.4 shows the Q,.-band absorption and fluorescence

spectra of BChl a in pyridine at 22 °C. The spectra are plotted with respect to
wavenumber as A(V)/ V and F (v)/v3 . respectively, with normalization to unit area.

Also shown in figure 4.4 is the output spectrum from the tizsapphire oscillator that was

employed in the dynamic-absorption. one-color pump—probe experiments. The shaded

75

Me

  

Melllllrn

Figure 4.3. Structure of bacteriochlorophyll a molecule, the dominant chromophore in
2

purple non-sulfur photosynthetic light-harvesting proteins.

76

Wavelength (nm)

820 800 780 760 74|0 720
l l

 

 

 

 

 

 

5
or
C
:93.
a)
.9
o
.9-
D
O .
I T I I I
12.0 12.5 13.0 13.5 14.0

Energy (103 cm“)

Figure 4.4. Continuous-wave absorption and fluorescence spectra from BChl in pyridine
solvent plotted as the dipole strength, A(V)/ V and F (v)/v3 , respectively, and normalized

to unit area. The dashed curve shows the intensity spectrum emitted by the tizsapphire
oscillator. The shaded region shows the range of the transmitted probe light that was

passed by the monochromator to the detector in the dynamic-absorption experiment.

77

region marks the 4-nm band pass of the transmitted probe spectrum that was passed by
the monochromator to the photodiode. The oscillator was tuned to obtain a spectrum that
was centered at 757 nm in order to move the probe spectrum as far off resonance from
the BChl a stimulated-emission spectrum as possible and yet still have the pump
spectrum on resonance with the ground-state absorption spectrum. The absorption and
fluorescence dipole-strength spectra shown in figure 4.4 can be used to judge the relative
contributions of signals from wave packets on the ground state and excited state,
respectively, in the detected probe band width. The pump spectrum and detected probe
band width specified in figure 4.4 favors detection of ground-state vibrational coherence

prepared by resonant impulsive stimulated Raman scattering (RISRS).4

4.2.2 Dynamic-Absorption Experiments

The dynamic-absorption transient obtained from BChl a in pyridine is shown in
figure 4.5. Following an intense coherence spikes'8 near the zero Of time, the transient
exhibits a short series of damped oscillations, with two strong recurrences prior to the
500-fs delay point. The first intense beat, at 200 fs, corresponds to the rapidly damped
feature that we reported previously; a weaker and more slowly damped set of oscillations
is observed out at least to the 8—ps delay point. The modulations arise from resonant
vibronic excitation Of the BChl a molecules; far weaker modulations arising from non-
resonant optical-Kerr-effect (OKE) excitations are observed in neat pyridine solutions
under the same excitation and detection Optical conditions, as was pointed out in

chapter 3.9

78

X3

w
l

N
l

 

 

1ps

 

1

AT/ T(relative units)

 

l

 

 

 

I I I I
0 2 4 6 8 10
Probe Delay (ps)

—
—

Figure 4.5. Dynamic-absorption transient from BChl in pyridine. The inset shows an
expanded view Of the 0—1 -ps portion of the signal. The ordinate is scaled relative to the
magnitude of the pump-induced change in transmission that follows the initial (0-5 ps)

exponential decay arising from solvation dynamics.

79

In the following, we present a time-domain modeling of the rapidly damped
oscillatory signals from BChl a in pyridine (figures 4.6 and 4.8), acetone (figure 4.9), and
l-propanol (figure 4.10). In each case, the oscillatory signal was isolated from the overall
signal (such as that in figure 4.5) by subtracting a fitted triple-exponential decay function
from the signal starting at 150 fs and extending to the end of the recorded range at 12 ps.

Figure 4.6 shows a series of preliminary models that attempt to describe the
oscillatory signal obtained with BChl a in pyridine with a single modulation component.

Figure 4.6a Shows that a damped 225-cm'l cosinusoid,

[(1) = 10 cos(a)t — (me—H7 (4-1)
is capable of roughly describing only the first beat in the signal; the damping time
(y = 92 fs) defines a very broad Lorentzian line shape in the frequency domain
(Aa)= 362 cm’l).IO The second beat in the model is significantly phase shifted with
respect to that ofthe experimental signal. and a large deviation of the model and signal is
observed at delays >400 fs.

A better description Of the oscillatory signal from BChl a in pyridine is Obtained from

a distribution of damped cosinusoids,
[(1) = J: darL(co, A,a)0.Aa), p)cos(a)t — (the-”7 (4.2)

which can be implemented computationally as a sum of discrete cosinusoids (equation

4.1), as shown in figure 4.7. The intensity L for a cosinusoid of frequency 03 is

determined here by a log-normal distribution, as parameterized by the area (A), center
frequency (to), width (Am). and asymmetry (p 2 1).ll These parameters were specified as

adjustable parameters in a nonlinear regression routine for the pro Fit 6.0 program; the

80

 

 

 

1O3XAT/T

 

 

 

 

200 300 400 500 600
80 —
4o — C
0 .—
-4o .. °'
I I I I I
200 300 400 500 600
Probe Delay (fs)

Figure 4.6. Expanded view of the rapidly damped oscillation observed in the dynamic-
absorption transient from BChl a in pyridine (see figure 4.5) and three optimized single-
modulation-component models employing different line shapes: (a) a discrete 225-cm'l
cosinusoid (equation 4.1) with the damping time y= 92 fs, (b) a Gaussian distribution of
damped cosinusoids with a center frequency of 225 cm'1 and a width (fwhm) of 50 em'l,
and (c) a log-normal distribution of damped cosinusoids with a center frequency of

285 cm’l and a width (fwhm) of 97 cm". The models for (b) and (c) were defined by

equation 4.2, with the distributions L(a)) set as a Gaussian and as a log-normal,

respectively. The scaling of the ordinate is relative to the magnitude of the pump—probe

ground-state depletion signal, as indicated in figure 4.5.

81

£(ar) A /(t) 3

Ac» -
2 ps

y=1.5ps

 

 

 

 

          

   
 

. 1.1/111,. (1.1.34.-.
. ' """"'O'5 \

0'1"

   

     

 

2ps V \7 2ps

Figure 4.7. Modeling the rapidly damped vibrational coherence Observed in BChl a
solutions with a distribution of slowly damped cosinusoids, L(a)), as defined by equation
4.2. A log-normal distribution [.(ar), (a), defined by its center frequency, coo, width, Am,
and asymmetry (or skew), p, is sampled over its full width. Each sample defines the
intensity of a Lorentzian line shape that corresponds in the time domain, (b), to a
cosinusoid at the sampled frequency a) with a homogeneous (or intrinsic) damping time,
7, here set arbitrarily to 1.5 ps. Summing of the cosinusoids obtained by sampling over
the full width of the distribution L(ar), (c), results in a rapidly damped waveform, (d), that

resembles the vibrational coherence observed in BChl solutions.

82

robust (gradient-search) optimization algorithm was employed in this work.12 For figure
4.6b, p was fixed to unity in order to obtain a Gaussian line shape. In figure 4.6c, an
asymmetric line shape was obtained by allowing p to float; in table 4.1, we use the Sign
of p to indicate whether the broader side of the optimized line shape is directed to high
(p > 1) or low frequency (p < l). The tabulated values for the area A report the integral

over the distribution £(w).

A: Edwflw) (4.3)

For multicomponent models. as used below, the A values are normalized so that they sum
to unity for the signal in pyridine: the A values reported for the signals in the other two
solvents are normalized by the same scaling factor used for pyridine, so they report the
relative modulation intensity compared to that of the signal in pyridine.

In the models shown in figure 4.6b and c, and in subsequent models employing
equation 4.2 in a multicomponent model. the intrinsic (or homogeneous) damping time y
was arbitrarily fixed tO l .5 pS, in the range Of the damping times for the slowly damped
modulation features we observed for BChl a in pyridine9 and for the carbocyanine dye -
DTI‘CI in polar solution.4 The optimized fit to the oscillatory signal is not strongly

dependent on the choice of y in any case. Similar line shapes L are obtained when longer
damping times are specified; L gets narrower as y is shortened. The sum Of discrete

cosinusoids (equation 4.2) converges satisfactorily with y = 1.5 ps when A(a)) is sampled
at 5-cm'l intervals over the l—500-cm'I range. Modulation frequencies a) > 500 cm'I are

not impulsively excited to a significant degree by the 60-fs pump pulses employed in the

83

present experiments, so the sum can be truncated at that point without any discernible
impact on the fit.

While the fit Obtained from the Gaussian distribution (figure 4.6b) is not clearly
distinguishable from that provided by the Lorentzian line shape of the discrete cosinusoid
used in figure 4.6a, the fit Obtained from the asymmetric distribution (figure 4.6c)
represents a significant improvement. It does a much betterjob of describing the first and
second beats in the signal than the model employing a Gaussian distribution (figure 4.6b),
but both models fail to provide an adequate description Of the signal’s rising trend before
the first beat, at delays <200 fs. Further, all three models shown in figure 4.6 exhibit a
weaker third beat, centered at the 480-fs delay point, that the experimental signal lacks.
This analysis suggests that a more elaborate model, with at least two independent
components and asymmetric line shapes, is required to describe the signal satisfactorily.
Figure 4.8 shows that an excellent description of the oscillatory signal from BChl a in
pyridine can be Obtained from a sum of two distributions of damped cosinusoids. Both
distributions were defined by equation 4.2; each was allowed to have an independent

phase ((5 and asymmetry p. The optimized log-normal components are centered at 121
cm“I and 205 cm".

A similar analysis was performed with the oscillatory signals obtained from BChl a in
acetone and l-propanol solutions. Preliminary models (not shown) were obtained with
single damped cosinusoids; a 251-cm'l cosinusoid (y= 65 fs; Aa) = 513 cm") provided
the best fit to the signal in acetone while a 208-cm'l cosinusoid (y= 144 fs; Aw = 231
cm'l) was the best fit for the signal in l-propanol. As in pyridine, these models deviate

from the observed signals beyond the first beat. Figure 4.9 shows that a sum of two

84

 

 

 

 

 

 

 

80 1
I...
\
l- 40 —
Q
o
r-
.40 .—
I I I I I
200 300 400 500 600
Probe Delay (fs)
a:
‘O
.3
C
O)
(U
2
a:
.2
E
33’ o
I I I I I
0 1 00 200 300 400 500

Frequency (cm")

Figure 4.8. Expanded view of the rapidly damped oscillation observed in the dynamic-
absorption transient from BChl a in pyridine (see figures 4.5 and 4.6) superimposed with
a model defined by the sum of two independent log-normal distributions [(0) of damped
cosinusoids, each defined by equation 4.2. The scaling of the ordinate is relative to the
magnitude of the pump—probe ground-state depletion signal, as in figure 4.5. Bottom:

Plots of L(a)) for the two components Observed in pyridine and their sum, 911(0)) (thick

curve). The model parameters are provided in table 4.1.

85

 

 

 

 

 

80 .-
'—
\
I'- 40 '-
<1
x 0 -
(O
o
v- -40 _.

I I I I I
200 300 400 500 600
Probe Delay (is)
a:
“O
a
“E
O)
(B
2'
m
.2
E
or
at o
I I I I I
o 100 200 300 400 500

Frequency (cm")

Figure 4.9. Expanded view of the rapidly damped oscillation Observed in the dynamic-
absorption transient from BC hl a in acetone superimposed with a model defined by the

sum of two independent log-normal distributions L(a)) of damped cosinusoids, each

defined by equation 4.2. The scaling of the ordinate is relative to the magnitude of the

pump—probe ground-state depletion signal, as in figure 4.5. Bottom: Plots of £(w) for the
two components observed in acetone and their sum, 914(w) (thick curve). The model

parameters are provided in table 4.1.

86

distributions Of damped cosinusoids, one centered at 108 cm'l and the other at 210 cm'1
provides a very gOOd description Of the signal from BChl a in acetone. This signal differs
from the one observed in pyridine most obviously in having a more intense and wider
second beat. The oscillatory signal observed from BChl a in l-propanol can be
distinguished from the ones observed in the other solvents in having four relatively strong
beats on the <800-fs time scale; the signals observed in pyridine and acetone have only
two strong beats that are confined to the <500-fs regime. Figure 4.10 shows that a sum of
three distributions of damped cosinusoids is required to obtain an adequate description of
the signal from BChl a in l-propanol; a model with only two components (not shown)
describes only the first two beats. Two of the modulation components, with distributions
centered at 66 cm'1 and 171 cm", are comparable to the pair observed in the other
solvents. The third, a very broad distribution centered at 165 cm", is the most intense
component of the three. It is distinct in having significant intensity extending down to 0

cm", and it exhibits almost a It Shift of phase with respect to that Of the other

components.

The signal observed from BC hi (I in cyclohexane differs from the signals obtained in
polar solvents in that there is only one strong, reproducible beat present on the < 300 fs
time scale. Figure 4.1 I shows that a sum of two distributions of damped cosinusoids is
needed to describe the signal from BChl a in cyclohexane. The model signal arises
predominantly from a broad, very low frequency component, centered at 30 cm", that
extends to 0 cm". The weaker, second modulation component is centered at 235 cm".

The model parameters for the rapidly damped vibrational coherence from BChl a

observed in the four solvents are compared in table 4.1. A better sense of how the

87

P- 40..
\
L?

0..
X
'2’:
V40—

 

 

I I I I I I I
200 300 400 500 600 700 800

 

 

 

 

Probe Delay (fs)
0
‘O
E
C
O)
(U
E
(D
.2
:4 fl
0) 0 —___
a: I I I I I
o 100 200 300 400 500

Frequency (cm‘)

Figure 4.10. Expanded view of the rapidly damped oscillation observed in the dynamic-
absorption transient from BChl a in l-propanol superimposed with a model defined by

the sum of three independent log-normal distributions L(a)) of damped cosinusoids, each

defined by equation 4.2. The scaling of the ordinate is relative to the magnitude of the

pump—probe ground-state depletion signal, as in figure 4.5. Bottom: Plots of L(a)) for the
three components observed in 1-propanol and their sum, 914(0)) (thick curve). The model

parameters are provided in table 4.1.

 

 

 

 

 

 

1.. 50‘-1
\ 3 Jr.
*5 o—
X
a -50-
,_
-1 .4
00 I I I I T
200 300 400 500 600
Probe Delay (fs)
o
'D
3
'E
0)
cu
2
.3
E
8 0 ‘
I I I I I
0 100 200 300 400 500

Frequency (cm")

Figure 4.11. Expanded view of the rapidly damped oscillation Observed in the dynamic-
absorption transient from BChl a with a single pyridine axial ligand in cyclohexane
superimposed with a model defined by the sum of two independent log-normal

distributions L(0)) of damped cosinusoids, each defined by equation 4.2. The scaling of

the ordinate is relative to the magnitude of the pump—probe ground-state depletion signal,

as in figure 4.5. Bottom: Plots of I.( 0)) for the two components observed in cyclohexane

and their sum, 911(0)) (thick curve).

89

Table 4.1. Model parameters for the rapidly damped vibrational coherence Observed in

BChl a solutionsa

 

 

 

 

 

 

 

Solvent
Component Parameter” Pyridine Acetone 1-Propanol Cyclohexane
I 02,. cm" 127 108 81 30
A0), cm" 120 109 25 180
p -l.l 1.7 1.2 1.2
¢, rad -0. l 78 0.329 0.398 0.012
A 0.725 0.326 0.095 3.027
2 0),. cm'l 195 210 182 235
A0). cm" 91 92 63 160
p 1.3 1.2 -l .1 1.4
a). rad -0.041 1.713 0.194 0.156
A 0.275 0.0406 0.106 75.46
3 (1)0, cm'l — — 165 —
Arr) , cm'l —— _ 426 _
)0 — —- -2.4 —
¢~ rad — —— 3.036 .—
A —— — 0.263 ——
Sum (m). cm" 146 189 131 123
Z, A, 1.000 0.732 0.465 22.94
Dipole momentC p. Debye 2.19 2.88 1.68 ~0

 

3 Figures 4.8—4.1 l

bSee equations 4.2—4.5 and the text.

cgas phase dipole moment'8

vibrational coherence depends on the choice of solvent can be Obtained from an analysis

of the mean frequency of the fitted distributions,

90

0) = fdwflflwyo

< > fdwflflw)

 

(4.4)

with the integral calculated over the 0) axis with the sum of the distributions L,-(0)) from

the two or three components.

91(0)) = 211,10) (4.5)

Table 4.1 also shows that the dipole moment of the solvent scales with the mean
frequency of the fitted distribution. The trend shows that part of the ((0) dependence is

on the dipole moment. However, the large amplitudes of the individual components as
well as the sum of the distributions for BChl in cyclohexane, compared to those from the
other three solvents, suggest that a large part is apparently independent of the dipole

moment and indicates some dependency on the polarizability of the solvent.

4.3 Discussion

The results presented in this chapter show that resonant impulsive excitation of the
lowest 7: -> If transition of BChl a is followed by a rapidly damped coherent vibrational
motion of the surrounding solvent molecules. As shown in figure 4.4, the tuning of the
pump-laser spectrum and the transmitted probe band width favors detection of coherent
wave-packet motion on the ground-state potential-energy surface. The wave packets are
prepared and launched on the ground-state surface by the resonant impulsive stimulated
Raman scattering (RISRS) mechanism,'3 so we know that the vibrational modes that

contribute to the modulated signal are those with coordinates that are displaced from the

91

l

ground-state equilibrium geometry in the BChl a 7: electronic excited state.14 In short,

we conclude that the rapidly damped vibrational coherence arises from solvent—BChl
intermolecular modes that are coupled to the BChl It ——> ,7‘ transition. To our knowledge,

this work represents the first direct Observation of resonance-Raman activity involving
intermolecular vibrational modes between solvent molecules and large electronic
chromophores. The assignment of the vibrational coherence to intermolecular modes
between BChl and clustered solvent molecules in the first solvation shell is supported by
the fact that the mode frequency scales with the dipole moment of the polar solvent, as

shown in table 4.1.

4.3.1 Structural Assignments
We are aware of only a few attempts to detect intermolecular modes in condensed
phases using resonance-Raman spectroscopy in the frequency domain. Mathies and

[15

coworkers observed resonance-Raman active modes from alcoho and water16 solutions

of the solvated electron. The solvated electron is a diffuse structure that makes a

significant ground-to-excited-state change in size and shape;‘7’21

it apparently interacts
strongly enough with a number of molecules in its first solvation shell that the absorption
transition is vibronically coupled to a number of solvent intramolecular modes.”’16 The
lowest frequency Raman-active modes observed in water, at 1 10 cm", are assigned to the
intermolecular hindered translational modes between the electron and molecules in the
first solvation shell.l6 In contrast, Waterland and Kelley22 did not observe features from

intermolecular modes in their UV resonance-Raman studies of the nitrate ion in several

solvents, but the n—electron density of the nitrate ion is highly localized. A similar

92

localization is probably responsible for the lack of detection of intermolecular modes in
the femtosecond pump-probe studies of the cyanide-bridged Ru(II,III) mixed-valence
complex by Barbara and coworkers.23 Significant solvent dependences were noted in the
intervalence charge—transfer rate and in the damping of the ground-state vibrational
coherence, but no solvent-dependent modulation components were directly observed.

A consideration Of this background makes it seem reasonable to suggest that

It —9 rr‘transitions in chromophores with delocalized rr-electron density, such as the

chlorophylls and porphyrins, might exhibit vibronic coupling to intermolecular modes
with the small number of first-shell solvent molecules that are clustered“:25 with it owing
to specific structural interactions. Displacement of the intermolecular mode would be

mediated by a change in the shape and extent of the n—electron density owing to the
7: —) n‘change in electron configuration. Note that similar mode displacements might be

driven by electron-transfer reactions since the redox-active electrons are usually
associated with the frontier molecular orbitals. The semicontinuum theory for electron

26'” involves a comparable

transfer reactions described by Mikkelsen and Ratner
structural picture: a small number of clustered solvent molecules in the first solvation
shell surrounded by a dielectric continuum.

We suggest that the resonance-Raman active solvent-mode interactions with BChl a

involve an attack by a solvent dipole on the rr-electron density above and below the plane

of the macrocycle. These interactions might resemble those in liquid pyrrole, where the

protons of one pyrrole molecule make a perpindcular attack on the rr-electron density of

an adjacent molecule,29 yielding intermolecular modes in the 100-cm'I region.30

Evidence for this configuration in the BChl-solvent interaction is obtained from

93

molecular mechanics calculations.

Figure 4.12 presents a possible model for the interaction between two non-ligated
pyridine solvent molecules and the n-electron density of the BChl macrocycle in the
dipyridine complex. The model was obtained using a molecular mechanics calculation
with the PCMOdel3' program and the MMX force field. The MMX force field3| adds

c132‘3 3 to permit modeling of inorganic

parameters to the well known MM2 force fiel
complexes such as those containing first-row transition metals. The calculation began
with an optimized structure for a single. isolated BChl molecule. Then, two pyridine
molecules were coordinated to axial binding sited on the central Mg(II) ion. The energy
of the resulting structure was then minimized. Lastly, two additional non-ligated
pyridines were then added at a distant position relative to the BChl macrocycle. Figure

4.12 shows the typical result of minimizing the energy of the resulting system. The non-

ligated pyridines attack the n-electron density on either face of the macrocycle and
apparently n-stack with the axially ligated pyridines. Similar configurations were

obtained with a wide range of starting configurations for the non-ligated pyridines. The
final T-shaped configuration of the non-Ii gated pyridines resemble configurations in neat
. . 29 . . . . . 34,35
pyrIdrne, and there rs a srrnrlar motif 1n benzene crystals.

Figure 4.13 illustrates the possible interaction between acetone solvent dipoles and
the n-electron density of the BChl macrocycle. The interaction was modeled using the
same sequence of molecular mechanics calculations as described above.

The observation of more than just a single mode distribution, however, evidences two

(or three, in l-propanol) configurations or sites for the solvent molecules clustered

around the BChl macrocycle. Owing to doming of the axially coordinated Mg(II) ion out

94

 

 

 

Figure 4.12. A possible model obtained using a molecular mechanics calculation with
the MMX force field and the PCModel31 program for the interaction between two non-
Iigated pyridine solvent molecules and the rc-electron density of the BChl macrocycle in

the dipyridine complex.

95

 

Figure 4.13. A possible model obtained using a molecular mechanics calculation with
the MMX force field and the PCModel31 program for the interaction between acetone
solvent dipoles and the It-electron density of the BChl macrocycle.

96

of the plane of the coordinating pyrrollic nitrogens, the two faces of the BChl macrocycle
present distinct n—electron surfaces that might be assigned to the two solvent-binding
sites. Even in the six-coordinate BChl 0 complexes observed in pyridine and l-propanol,
it is likely that the Mg(ll) ion is still domed so that it favors the attack of one of its two
solvent-derived axial ligands. Consider that the back-side attack of the sixth solvent-
derived axial ligand involves a longer reach to the Mg(ll) ion, so two well-resolved
binding equilibria for the axial ligands are observed.2 A consequence of the doming that
is certainly relevant to the interaction of the non-ligated solvent molecules with the

rr-electron density is the resulting tilting of the dipole moment out of the plane of the

BChl macrocycle.

The model for the signal in l-propanol (see figure 4.10) is dominated by a very broad
feature that extends all the way to 0 cm’| that is not observed in the other two solvents
(see figures 4.8 and 4.9). We suggest that this very low-frequency character arises from

36‘” of l-propanol molecules that are coupled to a given BChl

hydrogen-bonded chains
macrocycle. This assignment is made in analogy to the collective modes in certain liquids
that contribute to OKE signals in the < 20 cm'1 regime.“41 As in the other solvents,

clustered l-propanol molecules in the first solvation shell interact directly with the

rr-electron density of the BChl macrocycle, but the extended tail of the line shape down to

0 cm'1 reflects the presence of hydrogen-bonded chains of various lengths that are linked
to the attacking molecules. It follows that the vibrational coherence observed in pyridine
and acetone solvents lacks such an ultra-low-frequency character because these solvents
lack the ability to form intermolecular hydrogen bonds that would promote formation of

chains of molecules.

97

4.3.2 Line Shapes For Intermolecular and Intramolecular Modes in Solution and Proteins
The time domain resonance-Raman methods used in the present work are uniquely

capable of detecting intermolecular modes in delocalized )r-electron systems because the

line shapes would be expected to be broad owing to structural disorder in the
chromophore—solvent interaction. Similar line shapes are observed in nonresonant OKE
experiments in neat liquids,‘m‘42‘43 but we stress again that the Raman—active modes
detected in the present experiments arise from resonant excitation of the electronic
chromophore. In the frequency domain, broad low-frequency Raman line shapes arising
from inhomogeneously broadened features are not easily discerned from the baseline. In
the time domain, however, the same line shapes result in sharp, rapidly damped features
in the vibrational coherence, and the broad line shapes can be resolved into components
on the basis of their distinct phases. A time-domain analysis, such as the one based on
distribution functions used in the present work, is essential if reasonable models for the
vibrational coherence are sought. A conventional Fourier-transform analysis of the
rapidly damped modulations returns line shapes in the frequency domain that are usually
distorted owing to the truncation of the waveform at zero time, the individual components
are resolved spectrally, so it is difficult to obtain reliable information on the phases of the
individual components. Further, the demonstration in this work that the line shapes for
the intermolecular modes are asymmetric argues against the use of conventional linear—
prediction, singular-value decomposition methods, which are usually based on an
assumption that the signal can be described by a discrete sum of a small number of
damped cosinusoids.

As noted in chapter 3, the rapidly damped vibrational coherence observed in BChl a

98

in pyridine exhibits a much stronger signal in the time domain than is associated with the
slowly damped vibrational coherence that extends to the 8-ps delay point. We assign the
slowly damped signals to vibrational modes that are predominantly intramolecular in
character, involving mostly motions of the BChl macrocycle only. These modes are
essentially homogeneously broadened or less strongly inhomogeneously broadened. The
damping time of these signals is relatively long because the line shape is narrow in the
frequency domain, which reflects a relatively narrow range of geometries for the
associated oscillator. We report here that the slowly damped vibrational coherence was

too weak to be observed in the acetone and l-propanol solutions. The most likely

 

explanation for this observation involves axial coordination of pyridine molecules to the
Mg(II) ion and an associated distortion of the BChl macrocycle. This distortion
apparently mixes some of the strong Raman activity in the plane of the macrocycle into
the low-frequency, out-of—plane modes that are visible in the <300-cm’l regime. Bocian
and coworkers have previously reported that the axial coordination of the Mg(II) ion is
related to the activity of the out-of—plane deformation modes that are observed in the

< 200 cm'l-regime.44‘47 It is likely that the axially-ligated acetone and l-propanol n
molecules cause a much smaller distortion of the BChl macrocycle and a much lower 1 1
degree of mixing of activity into the out-of-plane deformations, rendering the slowly

damped vibrational coherence associated with these modes too weak to be observed

 

under our conditions. '—

99

4.4 References

(1) Katz, J.; Strain, H.; Leussing. D.; Dougherty, R. J. Am. Chem. Soc. 1968, 90,
784—791.

(2) Evans, T.; Katz, J. Biochim. Biophys. Acta 1975, 396, 414—426.

(3) Shipman, L.; Cotton. T. M.; Norris. J. R.; Katz, J. J. Am. Chem. Soc. 1976, 98,
8222—8230.

(4) Carson, E. A.; Diffey, W. M.; Shelly, K. R.; Lampa-Pastirk, S.; Dillman, K. L.;
Schleicher, J. M.; Beck, W. F. J. Phys. Chem. A 2004, 108, 1489—1500.

(5) Balk, M. W.; Fleming, G. R. J. Chem. Phys 1985, 83, 4300—4307.
(6) Engh, R. A.; Petrich, J. W.; Fleming, G. R. J. Phys. Chem. 1985, 89, 618—621.

(7) J00, T.; Jia, Y.; Yu. J.-Y.; Lang. M. J.; Fleming, G. R. J. Chem. Phys. 1996, 104,
6089—6108.

(8) Cong, P.; Deuhl, H. P.; Simon, J. D. Chem. Phys. Lett. 1993, 211, 367—373.

(9) Shelly, K. R.; Carson. E. A.; Beck, W. F. J. Am. Chem. Soc. 2003, 125, 11810—
11811.

(10) McHale, J. L. Molecular Spectroscopy; Prentice Hall: Upper Saddle River, New
Jersey, 1999.

(l l) Siano, D. B.; Metzler. D. E. J. Chem. Phys. 1969, 51, 1856—1861.

(12) pro Fit 6. 0; QuantumSoft: Uetikon am See, Switzerland.

(13) Yan, Y.-X.; Cheng, L.-T.; Nelson, K. A. In Advances in Non-linear
Spectroscopy; Clark, R. J. H., Hester, R. E., Eds.; John Wiley & Sons:
Chichester, England, 1988, p 299—355.

(14) Myers, A. B.; Mathies, R. A. In Biological Applications of Raman
Spectroscopy; Spiro, T. G., Ed.; Wiley—Interscience: New York, 1987; Vol. 2,

Resonance Raman Spectra of Polyenes and Aromatics, p 1—5 8.

(15) Tauber, M. J.; Stuart. C. M.; Mathies. R. A. J. Am. Chem. Soc. 2004, 126, 3414—
3415.

(16) Tauber, M. J.; Mathies, R. A. Chem. Phys. Lett. 2002, 354, 518—526.

100

(17) Alfano, J. C.; Walhout, P. K.; Kimura, Y.; Barbara. P. F. J. Chem. Phys. 1993,
98, 5996-5998.

(18) Kimura, Y.; Alfano, J. C.; Walhout. P. K.; Barbara, P. F. J. Phys. Chem. 1994,
98, 3450—3458.

(19) Reid, P. J.; Silva, C.; Walhout, P. K.; Barbara, P. F. Chem. Phys. Lett. 1994,
228, 658—664.

(20) Walhout, P. K.; Alfano, J. C.; Kimura, Y.; Silva, C.; Reid, P. J.; Barbara, P. F.
Chem. Phys. Lett. 1995, 232, 135—140.

(21) Silva, C.; Walhout, P. K.; Reid. P. J.; Barbara, P. F. J. Phys. Chem. A 1998, 102,
5701-5707.

(22) Waterland, M.; Myers Kelley. A. J. Chem. Phys. 2000, 113, 6760—6773.

(23) Kambhampati, P.; Son. D. H.; Kee, T. W.; Barbara, P. F. J. Phys. Chem. A 2000,
104, 10637—10644.

(24) Jortner, J. Z. Phys. D — Atoms, Molecules, and Clusters 1992, 24, 247—275.

(25) Spence, T. G.; Trotter. B. T.; Burns, T. D.; Posey, L. A. J. Phys. Chem. A 1998,
102, 6101—6106.

(26) Mikkelsen, K. V.; Ratner. M. A. J. Chem. Phys. 1989, 90, 4237—4247.
(27) Mikkelsen, K. V.; Ratner. M. A. J. Phys. Chem. 1989, 93, 1759—1770.

(28) Billing, G. D.; Mikkelsen, K. V. Advanced Molecular Dynamics and Chemical
Kinetics; John Wiley & Sons: New York, 1997.

(29) Gamba, Z.; Klein, M. L. J. Chem. Phys. 1990, 92, 6973—6974.

(30) Wynne, K.; Galli, C.; Hochstrasser. R. M. Chem. Phys. Lett. 1992, 193, 17—22.
(31) PCMODEL 8.0; Serena Software: Bloomington, IN.

(32) Allinger, N. J. J. Am. Chem. Soc. 1977, 99, 8127-8134.

(33) Allinger, N. J.; Kok, R. A.; lrnarn, M. R. J. Comp. Chem. 1988, 9, 591-595.
(34) Cox, E.; Cruickshank, D.; Smith. J. Proc. R. Soc. A 1958, 247, 1—21.

(35) Bacon, G.; Curry, N.; Wilson, S. Proc. R. Soc. A 1964, 279. 98—110.

101

(36) Garg, S. K.; Smyth. C. P. J. Phys. Chem. 1965, 69, 1294—1301.

(37) Jorgensen, W. L. J. Am. Chem. Soc. 1980, 102, 543—549.

(38) Bertolini, D.; Cassettari. M.; Salvetti. G. J. Chem. Phys. 1983, 78, 365—372.
(39) Sumi, H.; Marcus. R. A. J. Chem. Phys. 1986, 84, 4272—4276.

(40) Lotshaw, W. T.; McMorrow. D.; Thantu, N.; Melinger, J. S.; Kitchenbaum, R.
J. Raman Spect. 1995, 26. 571—583.

(41) Lotshaw, W. T.; Staver. P. R.; McMorrow, D.; Thantu, N.; Melinger, J. S. In
Ultrafast Phenomena 1X ; Barbara. P. F ., Knox, W. H., Mourou, G. A., Zewail,
A. H., Eds.; Springer-Verlag: Berlin, 1994, pp 91--92.

(42) McMorrow, D.; Lotshaw, W. '1‘. Chem. Phys. Lett. 1993, 201, 369—376.

(43) McMorrow, D.; Lotshaw. W. T. J. Phys. Chem. 1991, 95, 10395—10406.

(44) Palaniappan, V.; Aldema. M. A.; Frank, H. A.; Bocian, D. F. Biochemistry
1992, 31,11050—11058.

(45) Czamecki, K.; Diers. J. R.; Chynwat, V.; Erickson, J. P.; Frank, H. A.; Bocian,
D. F. J. Am. Chem. Soc. 1997, 119. 415—426.

(46) Thompson, M. A.; Zerner. M. C.; Fajer, J. J. Phys. Chem. 1990, 94, 3820—3828.

(47) Thompson. M. A.; Zerner. M. C.; Fajer, J. J. Phys. Chem. 1991, 95, 5693—5700.

102

CHAPTER 5

VIBRATIONAL COHERENCE FROM BACTERIOCHLOROPHYLL IN THE LIGHT
HARVESTING SUBUNITS 8777 AND 8820

5.0 Introduction

The photosynthetic purple non-sulfur bacteria usually contain two types of integral
membrane complexes, a core light harvesting complex and a peripheral light harvesting
complex.1 The light-harvesting complex 1 (LH 1) is known as the core complex that
surrounds a single reaction center. A model of LHl is shown in figure 5.1.2’3 LHl is
surrounded by several copies of a ring-shaped peripheral complex known as light-
harvesting complex 2 (LH2). figure 522‘3 Energy absorbed by LH2 is transferred to the
adjacent LHI complex and then to the reaction center. Loach et al. 4’5 and van Grondelle
and coworker56’7'” have shown that LHI can be reversibly dissociated into its subunits
by detergent solubilization. The 8820 subunit consists of a pair of a helices sandwiching
a strongly coupled pair of histidine-Ii gated BChl molecules'z‘l3 Van Grondelle and
coworkers6'” have provided evidence that the strongly coupled pair of BChl 0 molecules
held by the B820 subunit are arranged analogously to the pair of BChl in the LH2 8
structure.‘4

In this work, we employ 8820 as a limited model for the primary electron donor, P, in

 

the photosynthetic reaction center. 8820 allows us to study the electronic properties of a
bacteriochlorophyll dimer in the absence of energy transfer to or from adjacent subunits
or electron transfer to adjacent acceptors. as in the photosynthetic reaction center. A

model for 8820, as shown in figure 5.3. was obtained by Hu and Schultenl5 using

103

 

Figure 5.1. Model of the LHI light-harvesting complex from the calculations of
Schulten and coworkers.” Only the chromophores are shown here. (Refer to figure 5.3
for the details of how each pair of BChl are organized by a pair of membrane-separating

a helices.)

104

 

Figure 5.2. Crystal structure of the LH2 complex from Rhodopseudomonas acidophila
as reported by Caffrey and coworkers.16 The (X. and B units are shown in red and blue,

respectively, the 8800 pigments in purple, the B850 pigments in green, and the

 

carotenoids in yellow. The structure can be found under the accession number 2FKW in
the RC SB Protein Data Bank.

105

parameters from the LHI projection map obtained from electron microscopy” and from
the LH2 crystal structure. The distance between the Mg(11) ions is 9.2 A; the plane-to-
plane distance between the two macrocycles is 3.6 A.'8 In comparison, the pair of BChl
in P is more closely spaced; the distance the Mg(Il) ions is 7.6 A; the plane-to-plane
distance between the two macrocycles is 3.3 A.'9

Van Grondelle and coworkersb'” have extensively characterized the spectroscopic
properties of the pair of BChl 0 molecules held by the B820 subunit. At high detergent
concentrations, the B820 complex reversibly falls apart and produces the subunit B777,

shown in figure 5.4.5"1 8777 consists of one (1 helix ligated to a single BChl molecule.

We used 8777 to study the protein-bound monomeric chlorophyll system in order to

discern effects from the formation of the BChl dimer in 8820.

5.1 Experimental Procedures
5.1.1 Sample Preparation

Rhodospirillum rubrum G9 was obtained as a lyophilized pellet from the American
Type Culture Collection. It was grown in three-liter carboys under low-intensity
incandescent light (one 60 W bulb at a distance of 1 foot) in the presence of 5%
COz/95% N2 in modified Hutner's medium, as described by Cohen-Bazire et al.20 Cells
were harvested by centrifugation at 5000 x g for 5 min in a Sorvall GSA rotor; the cells
were stored at -10° C in the dark as pellets in the growth medium. Chromatophores
were isolated from whole cells by a series of steps involving sonication and
ultracentrifugation as described by Loach and coworkersfw the benzene extraction step

was not required because G9 is a carotenoidless mutant. The chromatophores were

106

   

Figure 5.3. Model of the B820 subunit from LHl from the calculations of

Schulten and coworkers.15

107

   

Figure 5.4. Possible model for the B777 subunit formed by solubilization of the B820
subunit of LHl. LHl is the result of dissociating 8820 into two polypeptides, each with
a single ligated BChl macrocycle, so there are two possible 8777 structures in a given

sample.

108

stored in 0.1 M HEPES (4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid) buffer

solution at pH 7.5 at -20° C for up to 1 week prior to use.

5.1.1.1 Isolation and Preparation of 8820 Subunits

Initial experiments were performed with 8820 subunits prepared using the detergent
n-octyl-rac-2,3-dipropyl sulfoxide (ODPS, Bachem) as described by Visschers and
coworkers.7 However. as shown in figure 5.5, the isolation of 8820 using ODPS was
somewhat unsuccessful in this laboratory. Figure 5.5 shows the absorption spectra of
chromatophores isolated from R. rubrum G9 following the addition of various ODPS
detergent concentrations. A 1.0% (w/v) solution of ODPS was added to a suspension of
chromatophores in order to prepare 8820 subunits. Figure 5.6 shows that, while a large
amount of 8820 is present, as judged from the absorption band at 820 nm, but there is
still a substantial amount of undissociated chromatophores present, which exhibit an
absorption maximum at 873 nm. In addition to the formation of 8820, a small shoulder
can be observed at 777 nm due to the initial formation of 8777 subunits.

Subsequent experiments employed preparations of 8820 that were prepared
according to the procedure of Miller et 01.4 8820 subunits were isolated from the
chromatophore preparations by solubilization with the detergent octyl
[3 -D-g1ucopyranoside (OG, Bachem). To a suspension of chromatophores, OG was
added to establish a ~0.75% (w/v) solution while monitoring the absorbance at 820 nm.
Figure 5.6 shows the gradual shift of the absorbance maximum from 873 nm to 820 nm
that occurs over a period of 17 hours. Small additions of OG solution were then added to

ensure that no further change in absorbance at 820 nm was detected. Unlike

109

 

 

 

1.2 — , 0.5% ODPS ‘3
pt, -- 1.0% ODPS ,
1.0— l;.- I — 3.0% ODPS ,
§ 0.8 — -' I
3
o 0.6—- 1
m
.0
< 0.4—
0.2 —
0.0
300 400 500 600 700 800 900 1000
Wavelength (nm)
1-2 ~ 0.5% ODPS
-- 1.0% ODPS
1-0 - -— 3.0% ODPS ,
a) / \.
O
C
m
-E
O
In
.0
<

 

 

 

650 700 750 800 850 900 950 1000
Wavelength (nm)

Figure 5.5. Absorption spectra illustrating the formation of 8777 subunits from
detergent-solubilized chromatophores isolated from R. rubrum G9, top panel. At low
concentrations of n-octyl-rac-2.3-dipropyl sulfoxide (ODPS), only the chromatophores
are present, as indicated by the absorption maximum at 873 nm. At intermediate
detergent concentrations, chromatophores, 8820 subunits, and 8777 subunits are
simultaneously present within the sample. The 8820 population is indicated by the
formation of the absorption peak at 820 nm. B777 subunits are exclusively present in the
presence of an increased concentration of ODPS detergent (~3.0% (w/v)). The bottom
panel shows an expanded view of the Qy absorption band so that the conversion of the

initial 870 nm peak to the 777 nm peak can be more easily observed.

110

 

 

 

 

   

 

 

 

0.5 .. 0.75% B-OG after 10 min
,3}, -- 0.75% 0-06 after 1 hr _,. -,
0.5 - 7 1‘... — 0.75% B-OG after 17 hr
8 0.4 -«
E
3 0.3 - -
2 0.2 -
0.1 --1
0.0
300 400 500 600 700 800 900 1000
Wavelength (nm)
0-6 -* 0.75% [3-06 after 10 min
0 _ -- 0.75% B-OG aftert hr
'5 — 0.75% B-OG after 22 hr .
§ 0.4 —
m / .f ‘ '.
E 0 3 .0 ‘ ."/ \. i.
8 ' / . \." \ t.
.o ------- . '1
< 0.2 — \2‘
\"x
0.1 ~‘:.'.:'::;:.':v_-:_=.°='—I-""’” "\‘. ,
0'0 I I I I I I I
650 700 750 800 850 900 950 1000 q..-

Wavelength (nm)

Figure 5.6. Absorption spectra illustrating the formation of 8820 subunits from
detergent-solubilized chromatophores isolated from R. rubrum G9 over a period of 17

hours, top panel. The chromatophores were solubilized with 0.75% (w/v) octyl [3 -D-

 

glucopyranoside (0G). The bottom panel shows an expanded view of the Qy absorption i.

band. No chromatophores are present in the sample 22 hours after the addition of 0G

solution.

111

chromatophores treated with ODPS detergent, where an equilibrium between 8777 and
8820 is maintained even at fairly high detergent concentrations, chromatophores treated

with OG yield only 8820 subunits.

5.1.1.2 Isolation and Preparation of 8777 Subunits

8777 subunits were prepared by the addition of ODPS to a suspension of
chromatophores until the concentration of ODPS was ~3.0 % (w/v). Additional ODPS
was added until no change in absorbance at 777 nm occurred and no visible peak at 820
from unsolubilized 8820 was observed. The gradual formation of 8777 from
chromatophores is shown in figure 5.5. ODPS, rather than OG, was used in the
preparation of 8777 because it was not possible to drive the equilibrium completion to
the 8777 form in the presence of OG under our conditions. Again, all steps were

conducted in the dark at room temperature (22 OC).

5.1.2 Continuous-Wave Spectroscopy
Continuous-wave absorption and fluorescence spectra for all samples were acquired
at room temperature with Hitachi U-2000 and Hitachi F-4500 spectrometers,

respectively. The fluorescence intensities were corrected as described in chapter 3.

5.1.3 F emtosecond Spectroscopy

Dynamic-absorption spectroscopy was performed with the femtosecond pump-probe
spectrometer described in chapter 2, consisting of a self-mode-locked tizsapphire
oscillator and a rapid-scanning Mach-Zehnder interferometer. The output spectrum

emitted by the oscillator exhibited a band width of 12 nm. The spectrum was centered at

112

 

J

 

818 nm for experiments performed on 8820 in 0G and 750 nm for experiments
performed on 8777 in ODPS. Extracavity compression employed a pair of SF10 prisms,
which were adjusted to produce 60-fs pulses (sechz) in all experiments. Probe light for
experiments done on 8820 was detected at 818 nm (2-nm band pass) and

757 nm (2-nm band pass) for experiments done on 8777.

8820 and 8777 samples for femtosecond spectroscopy were held at room
temperature (22 0C) in a l-mm fused-silica flow cuvette. Samples were circulated using
a peristaltic pump through the cuvette at a rate of 5 mL/min. The absorption spectrum of
the sample was monitored throughout the duration of the experiment for changes arising

from photochemistry or permanent photobleaching.

5.2 Results

5.2.1 Continuous-Wave Spectroscopy
Figures 5.7 and 5.8 show the absorption and fluorescence spectra exhibited by 8820

in 0G and 8777 in ODPS, respectively, at room temperature. The spectra are plotted as
A(V)/V and F(v)/v3 . respectively. with normalization to unit area. As plotted, the sum

of the spectra can be compared to the time-integrated dynamic-absorption spectrum in the
weak-field limit.22 The 0—0 vibronic transition in 8820 occurs in the vicinity of

11900 cm"I (840 nm), where the two normalized spectra cross. Also shown in figure 5.7
is the intensity spectrum from the tizsapphire laser that was the source of pump and probe
pulses in the dynamic-absorption experiments. The intensity spectrum, centered at

~820 nm, was tuned away from the stimulated emission to favor detection of ground-state
vibrational coherence prepared by resonant impulsive stimulated Raman scattering. The

shaded region coincides with the portion of the transmitted probe spectrum that was

113

 

 

Wavelength (nm)

 

 

 

 

 

 

820 800 780 760 740 720
l l l l I l

5

g F/v3

2

(I)

.9 NV

.8

D

0— ------- I
I I I I
12.0 12.5 13.0 13.5 14.0

Energy (103 cm'1)

Figure 5.7. Continuous-wave absorption and fluorescence spectra from 8777 in ODPS
plotted as the dipole strength, A(v)/ v and F(v) / v3 , respectively, and normalized to unit

area. The dashed curve shows the intensity spectrum emitted by the tizsapphire oscillator.
The shaded region shows the range of the transmitted probe light that was passed by the

monochromator to the detector in the dynamic-absorption experiment.

114

 

 

Wavelength (nm)

940 900 860 820 780 740
I I l l l I

 

Flv3

 

Dipole Strength

 

 

 

 

#33: a _.
0 __ - . .. - A
I I F I I I

10.5 11.0 11.5 12.0 12.5 13.0 13.5 14.0
Energy (103 cm“)

 

Figure 5.8. Continuous-wave absorption and fluorescence spectra from 8820 in OG
plotted as the dipole strength, A(v)/ v and F (v)/v3 , respectively, and normalized to unit E

area. The dashed curve shows the intensity spectrum emitted by the tizsapphire oscillator.
The shaded region shows the range of the transmitted probe light that was passed by the

monochromator to the detector in the dynamic-absorption experiment.

 

115

passed by the monochromator to the photodiode. The 0—0 vibronic transition in 8777 in
ODPS occurs in the vicinity of 12800 cm‘1 (785 nm), as can be seen in figure 5.8. The
intensity spectrum from the tizsapphire laser was centered at 750 run. As in the
experiments described in Chapter 4, the choice of pump spectrum and detected probe
band width favors detection of ground-state vibrational coherence. The range of the

transmitted probe light can be seen in figure 5.7.

5.2.2 Dynamic-Absorption Experiments

Figure 5.9 shows the dynamic-absorption transient obtained from 8777 in ODPS.
Following the intense spike near the zero of time, a short series of damped oscillations
are observed in the transient prior to 1 ps, as can be seen in the inset of figure 5.9. These
strong recurrences correspond to the rapidly damped features discussed in the previous
chapters. The dynamic-absorption transient obtained from 8820 in 0G is shown in figure
5.10. The zoomed inset in figure 5.10 shows the rapidly damped vibrational coherence
exhibited by 8820, with at least two recurrences lasting as long as 2 ps.

Figures 5.1 1 and 5.12 present a time-domain modeling of the rapidly damped

oscillatory signals from 8820 in OG and 8777 in ODPS, respectively. In each case, the

:‘ tinge!

oscillatory signal was isolated by subtracting a fitted triple-exponential function from the

signal starting at 150 fs and extending to the end of the recorded range at 12 ps. The

 

oscillatory residual can be fit to a model consisting of a sum of one or more distributions II
of damped cosinusoids. As in the time-domain modeling of the signals obtained from the

BChl solutions described in Chapter 4. the distributions are defined by,

1(1) = J: de(0). A, (00.A(r). ,0) COS(a)l — ma” 7 (5.1)

116

 

 

 

 

 

 

 

 

x 3

a 3 -
'E
:3
.3 2 "
% F ‘1 DS
is 1 —
|._
2
Q o _._______‘

I I I I I l

O 2 4 6 8 10

Probe Delay (ps)
Figure 5.9. Pump-probe signal obtained from 8777 in ODPS at room temperature. In

 

117

01
l
Tr..-

 

 

 

 

 

 

 

A I
.‘L’ .'
.g t
m 1.0—‘ 1: ii
.2 J
‘66 i"
I: .
ta
0‘0“ d 1 2 ps
I I I I I I
0 2 4 6 8 10
Probe Delay (ps)
Figure 5.10. Pump-probe signal obtained from 8820 in OG at room temperature. r...

 

118

The intensity I. for a cosinusoid of frequency to is determined here by a log-normal dis-

tribution, as parameterized by the area (A), center frequency (0)), width (Aw), and
asymmetry (p 2 1).23 The values for the area A report the integral over the distribution

11(0)),

A = I: (1011(0)) (5.2)

The distribution was sampled at 5 cm'l intervals from 1 cm’1 to 500 cm". The intrinsic
damping time, 1’ . was arbitrarily fixed to 1.5 ps.
Figure 5.1 1 Shows that the sum of two distributions of damped cosinusoids is required

to fit the early time oscillatory signal from 8777 in ODPS. Both distributions were

 

defined by equation 4.2; each was allowed to have an independent phase ¢ and
asymmetry p. The lognormal components are centered at 147 cm'1 and 184 cm". A
comparison of figure 5.1 1 and 5.12 shows that the rapidly damped vibrational coherence
exhibited by 8820 is more than ten times weaker in intensity that that of 8777. A model
arising from three lognormal distribution functions provides a good description of the

signal observed from 8820 in 0G. The intensity distributions for the three individual

r.
components are shown in the lower panel of figure 5.1 1. Two of the components have :
fairly narrow line shapes and maxima at 28 cm'l and 102 cm". A third component with a

broader line shape is observed at 48 cm". Note that the three components in the signal

from 8820 are centered at lower frequencies than observed in 8777 or in the BChl 2,1"

 

solutions described in Chapter 4.
An additional high frequency modulation can also be seen in the signal in figure 5.12

from 8820. These features were not considered in the modeling of the rapidly damped

119

 

 

 

 

80 —
l.—
l— 40 -
<1
X
n 0 —
o
‘-
-40..
l l l T l
200 300 400 500 600
Probe Delay (fs)
o
'D
3
'E
O)
(U
2
G)
.2
E
0 ——'/\
(I O-—

 

l l T l l
0 100 200 300 400 500

Frequency (cm")

Figure 5.11. Expanded view of the rapidly damped oscillation observed in the dynamic-
absorption transient from B777 in ODPS (see figure 5.9) superimposed with a model
defined by the sum of two independent log-normal distributions L(a)) of damped
cosinusoids, each defined by equation 5.2. The scaling of the ordinate is relative to the
magnitude of the pump—probe ground-state depletion signal, as in figure 5.9. Bottom:

Plots of L(a)) for the two components observed in ODPS and their sum, MM2) (thick

curve). The model parameters are provided in table 5.1.

120

 

 

103x AT/T
o
L

 

 

T l l l T l l 1
400 600 800 1000 1200 1400 1600 1800 2000

Probe Delay (fs)

 

 

 

 

a)
3 4:
'E ,_.
O)

m
E
m
.2
7.6
83’ o
l l l I
0 50 100 150 200
Frequency (cm‘)
Figure 5.12. Expanded view of the rapidly damped oscillation observed in the dynamic- l"

absorption transient from B820 in 00 (see figure 5.10) superimposed with a model

defined by the sum of two independent log-normal distributions L(a)) of damped

 

cosinusoids, each defined by equation 5.2. The scaling of the ordinate is relative to the

magnitude of the pump—probe ground-state depletion signal, as in figure 5.10. Bottom:

 

Plots of L(a)) for the two components observed in CO and their sum, MM2) (thick curve).

The model parameters are provided in table 5. l.

121

Table 5.1. Model parameters for the rapidly damped vibrational coherence observed in

BChl-containing proteinsa

 

 

 

 

 

Component Parameterb B777 B820
1 (0,. cm" 147 28
A0) . cm'l 79 20
p -l .3 1.4
(15. rad 1.514 1.303
A 0.190 0.081
2 w, , cm" 184 102
Am, cm'l 202 29
p 1.1 -l .5
¢. rad 1.558 -0.035
A 4.985 0.072
3 (00 , cm'l — 48
A0) , cm'l — 100
p — -1.8
¢. rad — 0.5484
A — 0.020
Sum (w), cm’ 191 41
Z, A, 1.508 0.084

 

aFigures 5.11—5.12

bSee equations 5.3—5.5 and the text.

122

 

-'_-1.-) malt-11

 

oscillations. The damping time for the components is in the range of those assigned to
intramolecular vibrations in BChl solutions. see chapter 3.

The model parameters for the rapidly damped vibrational coherence observed in the
BChl proteins are compared in table 5.1. As described in chapter 4, an analysis of the
mean frequency of the fitted distributions was performed on the signals obtained from

B777 and B820. The mean frequency is defined by,

deM(w)w
(m) = J; (5.3)
Edmflmw)

 

with the integral calculated over the (0 axis with the sum of the distributions 12(0)) from

the two or three components,

911(0)) = Z L,-(w) (5.4)

The results are compared in table 5.1. The integrated areas for B777 and B820 are

relative to those observed from BC hl in neat pyridine, discussed in chapter 4.

5.3 Discussion

5.3.] Comparison of Line Shapes and Components in the Vibrational Coherence From

B777 and B820
As in the BChl solutions (see chapter 4). the rapidly damped vibrational coherence
observed in B777 and B820 probably arises from formally intermolecular interactions
between groups in the surrounding solvent or protein and the BChI macrocycle. This
assignment is based on the comparable line shapes and frequencies observed in the two

classes of systems. The changes in the line shape and intensity that occur in moving from

123

 

 

B777 to B820 are consistent with their structures (see figures 5.3 and 5.4). The main
issue involves the number and degree oforder of the surrounding molecules that attack
the BChl macrocycles in the two systems. One face of the BChl macrocycle in B777 is
apparently exposed to a large number of mobile dipoles near the aqueous boundary of its
detergent micelle, producing a broad component in the vibrational coherence that is very
similar to that observed from BChl in the organic solvents.

In contrast, the two BC 111 macrocycles in B820 are largely protected from solvent
molecules, as can be seen in figure 5.13. This space-filling rendering of the 3820 system

shows that very little of the n-electron density of either BChl macrocycle can be

approached by solvent molecules owing to coverage by the a helix on one side and by

 

the adjacent BChl macrocycle on the other. We suggest that this protection leads to the
loss of the large component at 184 cm" in the B777 signal. The detected components in
the vibrational coherence in 8820 probably arise from a small number of relatively
ordered interactions with protein-derived groups, so the intensity is of the modulation in
the time domain is smaller and the line shapes in the frequency domain are narrower.
Additionally, the strongly-coupled electronic character of the 8820 system impacts the r
observed intensity of the vibrational coherence. Delocalization of the n-electron density
over both macrocycles in a given B820 molecule would be expected to attenuate the

displacement of a given intermolecular interaction between a solvent molecule and the

 

adjacent n-electron density of a given BC hl macrocycle. Because the resonance Raman

. . . . 7
cross sectlon scales W1th the square of the d1splacement,24‘“5

the intensity of the ground-
state vibrational coherenceZZ‘J’Q8 observed in B820 should be intrinsically smaller than

observed in the monomeric B777 system.

 

Figure 5.13. Space-filling (CPK) representation of the BChl dimer in the B820 subunit

 

from LHl based on the calculations of Schulten and coworkers.15

125

5.3.2 Assignments for Components in the Vibrational Coherence in B777 and 8820

The vibrational coherence from B777 is dominated by a large, broad component at
~l 84 cm". The relatively high frequency ofthis component is similar to those observed
in BChl solutions. It is reasonable to assign the 184-cm"I vibrational mode to an ion-
dipole interaction arising from the interaction between the BChl macrocycle and the
charged end of the detergent molecules in the surrounding micelle. These interactions
might resemble those discussed in the previous chapter where dipoles from the first-shell
solvent molecules attack the n-electron density above and below the plane of the BChl a
macrocycle. The protein. however, introduces the added complexity of a possible
interaction with the polypeptide chain as well as the surrounding solvent environment.

A lower frequency component, around 140 cm'l, is also observed in the vibrational
coherence from B777. This component is similar in both frequency and line shape to the
l 10-cm'I mode observed in the vibrational coherence from B820. The narrowed line
shapes of these components suggest a more ordered interaction. It is possible, then, that
these mode interactions involve an attack on the n—electron density of the BChl monomer
or dimer by specific polar residues from the surrounding protein medium. The presence
of a large, non-zero intensity component in 8820, in the 0—15-cm'l regime, suggests that
a significant fraction of the signal also arises from a collective interaction between the
BChl dimer and polar groups attached to the polypeptide backbone.

We can compare the vibrations in B820 to those in B777 without much consideration
of the strong electronic coupling between the two BChl macrocycles in B820 because the
experiment was designed to detect mostly ground-state wave packet motion, and only the

lower exciton state in B820 is pumped. It has been long suspected,29'3' however, that an

 

I g ‘ J“

 

interdimer mode, in which the two macrocycles vibrate with respect to each other, should
be observed in the resonance Raman spectrum in the < 100 cm'l region. In the strong
electronic coupling limit, the n-electron density is delocalized over both macrocycles and

. 9-
effect1vely produce a supermolecule. Warshel and coworkers2 3 I

predicted the
interdimer mode to be at about 100 cm'1 for the special pair in the reaction center of
Rhodobacter sphaeroides. Bocian and coworkers,32 in contrast, have speculated that this
mode should be observed around 30 cm". The interaction between the two BChls in
B820 is weaker than that in the reaction center, so the interdimer mode might be observed

at a somewhat lower frequency. The 30 cm’l mode observed in 8820 might arise from the

interaction between the BChl and the polypeptide chain, but a similar feature is not

 

observed in the vibrational coherence from B777. Therefore, it is attractive to assign the

30 cm"I feature in B820 to an interdimer vibrational.

5.3.3 BChl—Protein Interactions in 8820 and in the Photosynthetic Reaction Center
As discussed in the last section. the components in the rapidly-damped vibrational
coherence in 8820 and B777 that have relatively narrow line shapes may arise from
interactions with polar amino acid residues in the surrounding protein structure. A F
similar argument might be applied to the vibrational coherence from P in the 'i-
photosynthetic reaction center, where site-directed mutations were observed to cause

frequency shifts.33 In the following, we consider structural assignments for some of these

 

interactions.
In B820, only a small number ofthe polar residues are close enough, within 6 A of

the BChl dimer, to be capable of interacting directly with the n-electron density. Figure

5.14 and table 5.2 show these possible residues. The tabulated distances were estimated

127

 

Figure 5.14. Model of the BChl dimer in the B820 subunit from LHI based on the
calculations of Schulten and coworkers.” The polar residues in the vicinity of the dimer
are shown as line structures; tryptophan 41, tryptophan 43, tryptophan 48, and tyrosine
41. These residues are possible candidates for the interaction between the n-electron
density of the dimer and the protein medium. In order to simplify the figure, the axially
coordinated histidines, H32 and H39, are not shown.

128

Table 5.2. Potential residue candidates for the interaction between the BChl dimer and

the surrounding protein medium in B820a

 

Residue Distance (Eb

 

 

Y4] 5.51
W43 4.44
W45 3.24
W48 5.53

 

a Model of the B820 subunit from LH] based on the calculations of Schulten and

coworkers. ‘5

b The distances were estimated using Pymol.“

 

using Pymol34 from the B820 model developed by Hu and Schulten.l5 In the reaction
center from Rhodobacter sphaeroides, the residues within 6 A of P are indicated in figure
5.15 and table 5.3. There are several residues in the reaction center of R. sphaeroides that
are at comparable distances from the BChl dimer as those in 8820. V05, Martin, and
coworkers showed that the vibrational coherence from P* exhibits frequency shifts and
intensity changes in response to a range of point mutations that altered the hydrogen-
bonding pattern between the pair of BChl macrocycles and the surrounding protein 1
structure. Specifically, the removal of the hydrogen bond between H168 (see figure 5.15)

and the protein resulted in significant changes to the vibrational spectrum.33

 

5.4 Conclusions and Future Work

The work described in this thesis provides a new perspective on the structural origin
of the vibrational modes that control the dynamics of electron transfer in purple bacterial
reaction centers. In solution, the intermolecular modes between BChl and its clustered
polar solvent molecules make a dominant contribution to the low-frequency vibrational
coherence, but the frequency dependence of the intermolecular mode as a function of the
solvent’s dipole moment shows that the London dispersion interaction, arising from the
polarizability of the surrounding molecules. makes a significant contribution. In a

hydrophobic region, such as in the transmembrane region of the reaction center, the

 

dispersion interaction would be expected to be especially important. The presence of a
few polar side chains in the neighborhood, as discussed in the previous section, would
afford a tuning of the coupled mode frequency so that the reorganization energy and the

driving force could be precisely matched. Reorganization of the protein medium, through

 

Figure 5.15. The primary electron donor, P, from the reaction center from Rhodobacter
sphaeroides. The polar residues in the vicinity of the special pair are shown as line
structures; serine 190, serine 205, serine 224, tryptophan 156, asparagine 187, threonine
186, and hi stidine 168. These residues are possible candidates for the interaction
between the n—electron density of the special pair and the protein medium. In order to

simplify the figure, the axially coordinated histidines, H173 and H202, are not shown.

131

 

Table 5.3. Potential residue candidates for the interaction between the BChl dimer and

the surrounding protein medium in Rhodobacter sphaeroidesa

 

 

Residue Distance (A)b
W156 4.64
H168 3.46
T186 3.65
N187 5.07
8190 4.42
S205 4.62
S244 4.25

 

3 Structure obtained from the RCSB Protein Data Bank.35

b The distances were estimated using Pymol.3‘1

b)

i\)

 

 

. . . 6-
polar and nonpolar (v1scoelast1c) solvat1on responses,3 38

would be expected to play an
additional, energy dissipative role in the dynamics that affords control over the
reversibility of the charge-separation reactions. These ideas can be tested in experiments
that probe the nature of the line shape and the mode frequency for the intermolecular
mode in protein hosts.

A project that can potentially provide a more sophisticated understanding of the
solvent-derived intermolecular modes would involve studies of low-frequency vibrational
coherence in free-base and Zn(II)-substituted porphyrin systems in solution. Porphyrin

compounds are frequently used as components in artificial light-harvesting arrays 3944

4:47
and donor-acceptor. ’

An advantage to using porphyrins as a target system is that they
can be studies as isolated chromophores and as parts of charge-transfer complexes. The
Zn(ll) porphyrins are structurally comparable to the BChl systems. As a result, the
intermolecular modes between porphyrins and the cluster solvent molecules should be

comparable to the ones we have studied in the bacteriochlorophyll systems as described

in this thesis.

133

. ' :9: 2.1-1.5.2:. taut-IE,“

 

5 .5 References

(1) Zuber, H. In The Photosynthetic Reaction Center; Deisenhofer, J., Ed.;
Academic Press, lnc.: San Diego. 1993; Vol. 1, p 43~100.

(2) Kiihlbrandt, W.; Wang, D. N. Nature 1991, 350, 130—134.
(3) Kiihlbrandt, W.; Wang, D. N.; Fujiyoshi, Y. Nature 1994, 367, 614—621.

(4) Miller, J. F.; Hinchigeri, S. B.; Parkes-Loach, P. S.; Callahan, P. M.; Sprinkle,
J. R.; Riccobono, J. R.; Loach, P. A. Biochemistry 1987, 26, 5055—5062.

(5) Parkes-Loach. P. S.; Sprinkle. J. R.; Loach, P. A. Biochemistry 1988, 27, 2718—
2727.

(6) van Mourik, F .; van der Oord, C. .1. R.; Visscher, K. J.; Parkes-Loach, P. S.;
Loach, P. A.; Visschers, R. W.; van Grondelle, R. Biochim. Biophys. Acta
1991,1059, 111—119.

(7) Visschers, R. W.; Nunn, R.; Calkoen. F.; van Mourik, F.; Hunter, C. N.; Rice,
D. W.; van Grondelle, R. Biochim. Biophys. Acta 1992, 1100, 259—266.

 

(8) Visschers, R. W.; van Mourik, F.; l\rlonshouwer, R.; van Grondelle, R. Biochim.
Biophys. Acta 1993, 1141, 238—244.

(9) Visschers, R. W.; van Grondelle, R.; Robert, B. Biochim. Biophys. Acta 1993,
1183, 369—373.

(10) Pullerits, T.; van Mourik, F.; Monshouwer, R.; Visschers, R. W.; van
Grondelle, R. J. Lumin. 1994, 58. 168—171.

(11) Visser, H. M.; Somsen, O. J. G.: van Mourik, F.; Lin, 8.; van Stokkum, I. H.
M.; van Grondelle, R. Biophys. J. 1995, 69, 1083—1099.

(12) Pullerits, T.; Sundstrom, V. Acc. Chem. Res. 1996, 29, 381—389.
(13) Monshouwer, R.; van Grondelle, R. Biochim. Biophys. Acta 1996, 1275, 70—75.

(14) Creemers, T. M. H.; De Caro, C. A.; Visschers, R. W.; van Grondelle, R.;
Véilker, S. J. Phys. Chem. B 1999. 103, 9770—9776.

 

(15) Hu, X.; Schulten, K. Biophys. J. 1998. 75, 683—694.

(16) McDermott, 0.; Prince, S. M.; Freer, A. A.; Hawthomthwaite-Lawless, A. M.;
Papiz, M. Z.; Cogdell, R. J.; Isaccs. N. W. Nature 1995, 374, 517—521.

(17) Karrasch, S.; Bullough, P. A.; Ghosh, R. EMBOJ. 1995, 14, 631—638.

134

(18) Jordan, P.; Fromme, P.; Witt, H. T.; Klukas, 0.; Saenger, W.; KrauB, N. Nature
2001, 411, 909—917.

(19) Rhee, K.-H.: Morris. E. P.; Barber. J.; Kiihlbrandt, W. Nature 1998, 396, 283—
286.

(20) Cohen-Bazire. (3.; Sistrom, W. R.; Stanier, R. Y. J. Cell. Comp. Physiol. 1957,
49, 25—43.

(21) Loach, P. A.; Androes, G. M.; Maksim. A. F.; Calvin, M. Photochem.
Photobiol. 1964, 2, 443—454.

(22) Pollard, W. T.; Mathies, R. A. Annu. Rev. Phys. Chem. 1992, 43, 497—523.
(23) Siano, D. B.; Metzler. D. E. J. Chem. Phys. 1969, 51, 1856-1861.

(24) Tang, J.; Albrecht, A. C. 1n Raman Spectroscopy: Theory and Practice;
Szymanski, H. A., Ed.; Plenum: New York, 1970; Vol. 2, p 33—68.

(25) Myers, A. B.; Mathies, R. A. In Biological Applications of Raman
Spectroscopy; Spiro. T. 0., Ed.; Wiley—lnterscience: New York, 1987; Vol. 2,
Resonance Ramon Spectra ofPolt'enes and Aromatics, p 1—5 8.

(26) Pollard, W. T.; Lee, S.-Y.; Mathies. R. A. J. Chem. Phys. 1990, 92, 4012—4029.

(27) Pollard, W. T.; Dexheimer, S. L.; Wang, Q.; Peteanu, L. A.; Shank, C. V.;
Mathies, R. A. J. Phys. Chem. 1992. 96, 6147—6158.

(28) Johnson, A. E.; Myers. A. B. J. (Them. Phys. 1996, 104, 2497—2507.
(29) Warshel, A. Proc. Natl. Acad. Sci. (SA 1980, 7 7, 3105—3109.
(30) Warshel, A.; Parson, W. W. J. Am. ("hem Soc. 1987, 109, 6143—6152.

(31) Warshel, A.; Creighton, S.; Parson. W. W. J. Phys. Chem. 1988, 92, 2696—
2701.

(32) Czamecki, K.; Diers, J. R.; Chynu-‘at. V.; Erickson, J. P.; Frank, H. A.; Bocian,
D. F. J. Am. Chem. Soc. 1997. 119, 415—426.

(33) Rischel, C.; Spiedel. D.; Ridge. J. P.; Jones. M. R.; Breton, J.; Lambry, J.-C.;
Martin, J .-L.; Vos, M. H. Proc. A'atl. Acad. Sci. USA 1998, 95, 12306—1231 1.

(34) Pymol; Delano Scientific l..l.C: San Francisco, CA.

(35) Stowell, M. H., McPhillips. T.M.. Recs, D.C., Soltis, S.M., Abresch, E.,
Feher, G. Science 1997, 2.79. 812-816.

 

 

(36) Lampa-Pastirk, S.; Lafuente, R. C; Beck, W. F. J. Phys. Chem. 82004, 108,
12602—12607.

(37) Berg, M. J. Phys. Chem. A 1998, 102. 17—30.
(38) Berg, M. J. Chem. Phys. 1999. 110. 8577—8588.

(39) Prathapan, 8.; Johnson, T. E.; Lindsey. J. S. J. Am. Chem. Soc. 1993, 115,
7519—7520.

(40) Hsiao, J .-S.; Krueger, B. P.; Wagner. R. W.; Johnson, T. E.; Delayen, J. K.;
Mauzerall, D. C.; Fleming, G. R.; Lindsey, J. S.; Bocian, D. F.; Donohoe, R. J.
J. Am. Chem. Soc. 1996, 118, 11181—11193.

(41) Loewe, R. S.; Lammi, R. K.; Diers. .l. R.; Kirmaier, C.; Bocian, D. F.; Holten,
D.; Lindsey, J. S. J. Mater. Chem. 2002, 12, 1530—1552.

(42) Bothner-By, A. A.; Dadok, J.; Johnson. T. E.; Lindsey, J. S. J. Phys. Chem.
1996, 100, 17551—17557.

 

(43) Strachan, J.-P.; Gentemann, 8.; Seth, J.; Kalsbeck, W. A.; Lindsey, J. S.;
Holten, D.; Bocian, D. F. J. Am. Chem. Soc. 1997, 119, 11191—1 1201.

(44) Yatskou, M. M.; Koehorst, R. B. M.; van Hoek, A.; Donker, H.; Schaafsma, T.
J.; Gobets, B.; van Stokkum. 1.; van Grondelle, R. J. Phys. Chem. A. 2001, 105,
11432—11440.

(45) Wagner, R. W.; Lindsey, .l. S. J. Am. Chem. Soc. 1994, 116, 9759—9760.

(46) Wagner, R. W.; Lindsey. J. 8.; Seth. J.; Palaniappan, V.; Bocian, D. F. J. Am.
Chem. Soc. 1996, 118, 3996—3997.

(47) Prathapan, S.; Yang, S. 1.; Seth. J.; Miller, M. A.; Bocian, D. F.; Holten, D.;
Lindsey, J. S. J. Phys. (f‘hem. B 2001. 105, 8237—8248.

 

I36

llllllllllllllllllllll LIBRARIES

ll11111111111111lllllzllllllllllllllllllllllll