100% found this document useful (1 vote)
6 views

Getting Started on Time-Resolved Molecular Spectroscopy Jeffrey A. Cina instant download

The document is an introduction and preface to the book 'Getting Started on Time-Resolved Molecular Spectroscopy' by Jeffrey A. Cina, which aims to provide an accessible treatment of ultrafast spectroscopy for graduate-level students. It discusses the author's journey and motivations for writing the book, emphasizing the importance of understanding molecular dynamics in ultrafast signals. The text is designed to be user-friendly, with exercises and illustrative signal calculations to enhance comprehension of the underlying theory.

Uploaded by

awwalutogu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
6 views

Getting Started on Time-Resolved Molecular Spectroscopy Jeffrey A. Cina instant download

The document is an introduction and preface to the book 'Getting Started on Time-Resolved Molecular Spectroscopy' by Jeffrey A. Cina, which aims to provide an accessible treatment of ultrafast spectroscopy for graduate-level students. It discusses the author's journey and motivations for writing the book, emphasizing the importance of understanding molecular dynamics in ultrafast signals. The text is designed to be user-friendly, with exercises and illustrative signal calculations to enhance comprehension of the underlying theory.

Uploaded by

awwalutogu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 78

Download the full version and explore a variety of ebooks

or textbooks at https://ebookmass.com

Getting Started on Time-Resolved Molecular


Spectroscopy Jeffrey A. Cina

_____ Tap the link below to start your download _____

https://ebookmass.com/product/getting-started-on-time-
resolved-molecular-spectroscopy-jeffrey-a-cina/

Find ebooks or textbooks at ebookmass.com today!


Here are some recommended products for you. Click the link to
download, or explore more at ebookmass.com

Resolving Spectral Mixtures With Applications from


Ultrafast Time-Resolved Spectroscopy to Super-Resolution
Imaging 1st Edition Cyril Ruckebusch (Eds.)
https://ebookmass.com/product/resolving-spectral-mixtures-with-
applications-from-ultrafast-time-resolved-spectroscopy-to-super-
resolution-imaging-1st-edition-cyril-ruckebusch-eds/

Programming Arduino: Getting Started with Sketches (Tab)


Monk

https://ebookmass.com/product/programming-arduino-getting-started-
with-sketches-tab-monk/

Programming Arduino: Getting Started with Sketches, Third


Edition Simon Monk

https://ebookmass.com/product/programming-arduino-getting-started-
with-sketches-third-edition-simon-monk/

Programming Arduino: Getting Started with Sketches 3rd


Edition Simon Monk

https://ebookmass.com/product/programming-arduino-getting-started-
with-sketches-3rd-edition-simon-monk/
Programming the Photon: Getting Started with the Internet
of Things Rush Christopher.

https://ebookmass.com/product/programming-the-photon-getting-started-
with-the-internet-of-things-rush-christopher/

Programming the Raspberry Pi, Third Edition: Getting


Started with Python Simon Monk

https://ebookmass.com/product/programming-the-raspberry-pi-third-
edition-getting-started-with-python-simon-monk/

Getting Started with Angular: Create and Deploy Angular


Applications 1st Edition Victor Hugo Garcia

https://ebookmass.com/product/getting-started-with-angular-create-and-
deploy-angular-applications-1st-edition-victor-hugo-garcia/

Getting Started with SQL and Databases: Managing and


Manipulating Data with SQL Mark Simon

https://ebookmass.com/product/getting-started-with-sql-and-databases-
managing-and-manipulating-data-with-sql-mark-simon/

Programming the Intel Galileo: Getting Started with the


Arduino -Compatible Development Board Christopher Rush

https://ebookmass.com/product/programming-the-intel-galileo-getting-
started-with-the-arduino-compatible-development-board-christopher-
rush/
GETTING STARTED ON TIME-RESOLVED MOLECULAR
SPECTROSCOPY
Getting Started on Time-Resolved
Molecular Spectroscopy
Jeffrey A. Cina
University of Oregon

3
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Jeffrey A. Cina 2022
The moral rights of the author have been asserted
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2021945727
ISBN 978–0–19–959031–5
DOI: 10.1093/oso/9780199590315.001.0001
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
For Barbara, of course.
Preface

Well, it has been quite an adventure so far. It was my good fortune to be drawn
into the theory of ultrafast spectroscopy early in the femtosecond era, due in large
part to the prior influence of Rick Heller’s wave-packet descriptions of continuous-
wave spectroscopies (thank you, Laurie!); Bob Silbey’s coaching during the 1980’s,
when multi-pulse optical-phase-controlled picosecond measurements were first under
consideration; and the pleasure and long-term benefit of collaboration with Norbert
Scherer, Stuart Rice, Graham Fleming, and other co-workers at Chicago. At that
stage, one had to find one’s own way, and for me that way started with ordinary
time-dependent perturbation theory as I’d learned it in Bob Harris’s (personally life-
changing) quantum mechanics classes at Berkeley, coupled with a desire to illuminate
the dynamics underlying optical measurements in terms of the evolving nuclear wave
packets that accompany each molecular electronic state.
It hadn’t occurred to me to write this book, or any other, but in 2011 I was in a bit
of a hiatus that came at the end of a multi-year group-reading project with co-workers
at Oregon on L&L’s Electrodynamics of Continuous Media. At Bob Mazo’s suggestion,
Sonke Adlung from Oxford University Press called sometime that year and asked if I
might like to write a book. It happened that I was running a fever at the time and
said, “Sure!”
So, here you have it, the result of the succeeding decade of puzzlement, formulation,
and reformulation—aided by the patience and helpful advice of the numerous collab-
orators and colleagues whose names are among those listed in my Acknowledgements,
hopefully without any inadvertent omissions.
There are valuable treatises already available on the principles of nonlinear optical
spectroscopy of molecular and material systems, notably those by Shaul Mukamel
and Minhaeng Cho.1,2,3,4 The coverage of those works goes beyond what is presented
here. So why write another? I believe that the ultrafast community could benefit
from a deliberately accessible, stepwise treatment (which doesn’t mean an easy one),
more of a textbook than a comprehensive exposition, which can serve as a bridge
between the graduate-level training in quantum mechanics that’s standard in Physical
Chemistry programs and the advanced formulations that serve as guidebooks for those
practicing in the field. The standard approach is to start from the equilibrium density

1 S. Mukamel, Principles of Nonlinear Optical Spectroscopy (Oxford University Press, New York,
1999).
2 M. Cho, Two-Dimensional Optical Spectroscopy (CRC Press, Boca Raton, 2009).
3 D. J. Tannor, Introduction to Quantum Mechanics. A Time-Dependent Perspective (University
Science Books, Sausalito, 2007).
4 J. Yuen-Zhou, J. J. Krich, I. Kassal, A. Johnson, and A. Aspuru-Guzik, Ultrafast Spectroscopy:
Quantum Information and Wavepackets, (IOP Publishing, Bristol, 2014).
viii Preface

matrix of the target molecule and express nonlinear optical signals as a convolution
of the appropriate nonlinear optical response functions with the electric field of the
incident laser pulses. In addition, informed by very early analyses of multi-wave mixing
in nonlinear optical crystals, conventional descriptions are often couched in terms of
optical wave propagation in extended media. Everything in the present text is (or
should be!) physically consistent with the widely applied existing theoretical analyses.
Applications of existing descriptions sometimes tend, in my view, to lose track of
the molecular-level dynamics underlying ultrafast signals, in part because those de-
scriptions compel, or at least encourage one to think about both bra- and ket-sides of
the density operator simultaneously. On the other hand, informed by the superposition
principle, framing things at least initially as a sum of terms in Hilbert space having
various orders in the external-field strengths makes it easier to think about the perti-
nent contributions to the molecular state one at a time. What is developed, analyzed,
and interpreted in terms of Hilbert-space wave functions can easily be converted to a
density-matrix description through a manipulation that takes just a couple of lines.
The wave-propagation picture of nonlinear optical response, while important for
experimental purposes, tends to obscure the fact that, even in an extended sample, each
molecule can often be regarded as undergoing absorption, fluorescence, and Raman
scattering all by itself, as well as acquiring the nonlinear induced dipole moments that
give rise to time-resolved signals. One goal of this text is to formulate as much as
possible in terms of nonlinear induced molecular dipoles—these are ultimately to be
expressed as quantum mechanical overlaps between pairs of multi-pulse wave packets—
and to derive any necessary macroscopic wave-propagation aspects, such as wave-
vector matching, from that microscopic starting point.
The strength of treatments based on nonlinear optical response functions is their
generality. Nonlinear optical signals are to be derived from that fundamental under-
pinning by convolving the relevant response functions with the actual form of the in-
cident laser pulses by multidimensional integration over time variables. It is the case,
though, that such an approach encodes much quantum dynamical information that is
not of immediate use in calculating or interpreting the actual laser-driven dynamics
under consideration. Those strategies sometimes tend, in practice, to invite simplified
comprehensive but phenomenological descriptions that do not attach directly to the
specific, perhaps non-generic form of the regions of the molecular potential energy
surfaces that govern the measured signal.
The treatment put forward here reverses the order of operation between quantum
mechanical averaging and integration over time. It carries out the latter first, with
the use of pulse propagators, operators analogous to those sometimes used in magnetic
resonance spectroscopy, which encapsulate the influence of a nonzero-duration laser
pulse within an instantaneously acting quantum mechanical operator. The reduced
pulse propagators that transfer nuclear wave packets between electronic potential en-
ergy surfaces under the influence of the external fields, reshaping them in the process,
capture the elements of coherent control that are inherent in short-pulse optical spec-
troscopy. This approach puts the focus on the motion of nuclear wave packets in the
optically accessed regions of the electronic potential energy surfaces on which they
evolve. It declines the frequent practice in work-ups based on nonlinear response func-
Preface ix

tions of idealizing the characteristics of the experimentally available light pulses and,
in effect, regarding those response functions themselves as constituting signals, rather
than signal transducers. As mentioned above, the working expressions arrived at here
present signal contributions as overlaps between well characterized individual multi-
pulse nuclear wave packets.
For ease of use and at the cost of some redundancy, I’ve tried to make each chapter
of this book usable and comprehensible on its own. There are some slight inconsisten-
cies of notation between chapters, so unless the reader is advised to do so, applying a
formula from one chapter to an equation in another should be done with caution.
In lieu of end-of-chapter problems, there are many boxed exercises embedded in
the text. By and large, these represent the type of derivation or physical analysis that
I work through to consolidate my own grasp of the ideas at hand. It’s my guess that
readers who excuse themselves from these exercises will be sacrificing something in
the depth of their understanding.
The final portion of most chapters consists of illustrative signal calculations. Car-
rying these out or interpreting their form is both our reward for slogging through the
underlying theory and a pale substitute for the experiential satisfaction of performing
actual measurements. The molecular models that are targeted in these calculations
are chosen to facilitate a thorough interpretation in terms of the underlying molecu-
lar dynamics. It is hoped that these illustrations will help inform what may be the
less complete interpretations that are possible with more complicated experimental
targets.
There is a smattering of references throughout the text but no comprehensive
bibliography. The works cited are just a few of those I found helpful or inspiring.
Many others could undoubtedly have served as well or better; the ones I mention can
perhaps serve as points of entry in the search for other relevant examples.
With help from the many colleagues who have read portions of this text, I’ve tried
hard to root out conceptual misdirection and outright errors. From experience, I know
that there’s no such thing as a small mistake in a scientific text. My sincere apology
for any that may remain.
Personal taste has a significant influence in science (gasp!), and this textbook
reflects my own. I hope, though, that the treatment given here will be to the liking of
at least some readers. I presume further to hope that, after working their way through
this book, those who persevere in doing so will be equipped to make the best use of
more sophisticated methodologies, gaining as much physical insight as possible and
avoiding some of the pitfalls to which applications of those approaches occasionally
give rise.

Jeff Cina
University of Oregon
2021
Acknowledgements
My heartfelt thanks to the following individuals and organizations: y’all know what for.
John Adamovics, Paul Alivisatos, Rise Ando, Ara Apkarian, Sonke Adlung, Natasha
Aristov, Alan Aspuru-Guzik, Bridgette Barry, Mark Berg, Berkeley Chemists for
Peace, Jason Biggs, Sandra Bigtree, Isabella Bischel, Eric Bittner, Harry Bonham,
Stephen Bradforth, Bill Braunlin, Paul Brumer, Irene Burghardt, Carlos Bustamante,
Laurie Butler, Carl Bybee, Jianshu Cao, Roger Carlson, Howard Carmichael, Sister
Carola FSPA, Craig Chapman, Xiaolu Cheng, Minhaeng Cho, Gerri Hoffman Cina,
Merrill T. Cina, Zoë Cina-Sklar, Rob Coalson, Devin Daniels, Joshua Daniels, Julia
Daniels, Rebecca Daniels, Peter Dardi, Brent Davidson, Jahan Dawlaty, Peter Dewey,
Jenni Dobbins, Mike and Carol Drake, Camille and Henry Dreyfus Foundation, Tom
Dyke, Ellen Epstein, my many friends at Espresso Roma, Elaine Finley Giannone,
Graham Fleming, Joan Florsheim, Ignacio Franco, David Frank, Karl Freed, Sarah
Fuchs, Rosemary Garrison, Ibrahim Gassama, Diane Gerth, Michelle Golden, Evi
Goldfield, Eoghan Gormley, John Simon Guggenheim Memorial Foundation, John
Hardwick, Alex Harris, Charles Harris, Christine Harris, Jerry Harris, Katherine Har-
ris, Robert A. Harris (my former PhD advisor, an exemplar of science, and a lifelong
friend), Ray Heath, Rick Heller, Eric Hiller, Maddie Holst, Travis Humble, Katharine
Hunt, Heide Ibrahim, Deb Jackson, Suggy Jang, Truus Jansen, David Jonas, Tara
Jones, Taiha Joo, Chanelle Jumper, Joe Kao, Mike Kellman, Alexis Kiessling, Dmitri
Kilin, Young-Kee Kim, Phil Kovac, Peter Kovach, Cindy Larson, Larry LeSueur, Sis-
ter Leora FSPA, Susan Levine-Friedman, Mark Limont, Katja Lindenberg, Giulia Lip-
parini, Andy Marcus, Alex Matro, Karin Matsumoto, Jack Maurer, Bob Mazo, David
McCamant, Alden Mead, Horia Metiu, Stan Micklavzina, Coleen Miller, Florabelle
Moses, Shaul Mukamel, Sid Nagel, US National Science Foundation, Keith Nelson,
Kenji Ohmori, Takeshi Oka, Janine O’Guinn, Colleen O’Leary, April Oleson, Mrs.
Ruth Patterson, Barbara Perry, Fred Perry, David Picconi, David Pratt, Jim Prell,
Mike Raymer, Tom Record, Stuart Rice, Philip Richardson, Mary Rohrdanz, Victor
Romero-Rochin, Sandy Rosenthal, Sister Rosilda FSPA, Penny Salus, Marsha Saxton,
Rich Saykally, Norbert Scherer, Greg Scholes, Moshe Shapiro, Mike Sheahan, Yu-Chen
Shen, Nancy Shows, Robert J. Silbey (my postdoctoral advisor, a gem of a human be-
ing), Michael Sipe, Barbara Sklar, Martin Smith, Tim Smith, Peter Straton, Joe
Subotnik, Susie Cina Sullivan, David Tannor, Judithe Thompson, Nacho Tinoco, Dr.
Brandy Todd, Emma Tran, Daniel Turner, Rob Tycko, David Tyler, Lowell Ungar, my
fellow members of Oregon’s faculty labor union United Academics, Marian Valentine,
Steven van Enk, Hailin Wang, John Waugh, Bob Weiss, Julia Widom, Kevin Wiles,
Cathy Wong, Claude Woods, Duane C. Wrenn and Leticia Steuer of Energetic Soul,
John Wright, Joel Yuen-Zhou, Larry Ziegler.
Contents

1 Short-pulse electronic absorption 1


1.1 Basic set-up 1
1.2 Energy changes of molecule and field 2
1.3 Expectation values 3
1.4 The Heller formula 5
1.5 Systems starting in thermal equilibrium 6
1.6 Example absorption calculations 7
2 Adiabatic approximation 11
2.1 Molecular Hamiltonian 11
2.2 Molecular eigenstates 12
3 Transient-absorption spectroscopy: Making ultrashort pulses
worthwhile 17
3.1 Model Hamiltonian and signal expression 17
3.2 Transient-absorption dipole 20
3.3 Exemplary calculations 23
4 How fissors works: Femtosecond stimulated Raman spectroscopy
as a probe of conformational change 34
4.1 Basic idea 34
4.2 Signal formation 36
4.3 Fissors dipole 38
4.4 Example signal calculation 45
5 Transient-absorption reprise: Taking advantage of vibrational
adiabaticity 53
5.1 Transient-absorption signal under vibrational adiabaticity 53
5.2 Calculated transient-absorption signals 55
6 Two and a half approaches to two-dimensional wave-packet
interferometry 61
6.1 Introduction 61
6.2 Measured quantities 62
6.3 Quantum mechanical aspects 76
6.4 Example signals 85
7 Two-dimensional wave-packet interferometry for an electronic
energy-transfer dimer 96
7.1 Energy-transfer dimer 96
7.2 Whoopee signal 99
xii Contents

7.3 Illustrative calculations 105


Appendix A Electromagnetic energy change due to light ab-
sorption 129
Appendix B Delay regions for doubly excited-state-visiting over-
laps in the difference-phased singly excited-state populations 131
Appendix C Delay regions for overlaps contributing to the difference-
phased doubly excited-state population 135
Index 137
1
Short-pulse electronic absorption

1.1 Basic set-up


This chapter treats the interaction between an individual molecule, whose fixed lo-
cation is regarded as the spatial origin, and a short pulse of light. The molecule
may be isolated or immersed in a condensed-phase environment, but is assumed to
have an energetically isolated electronic transition at or near resonance with the in-
cident laser pulse. The time-dependent Hamiltonian of such a system can be written
H(t) = H + V (t), where

H = |giHg hg| + |ei( + He )he| , (1.1)

and
V (t) = −m̂ · E(t) . (1.2)
Hg and He are the nuclear Hamiltonians governing all relevant intra- and intermolec-
ular degrees of freedom in the ground and excited electronic states, respectively, and
 is the “bare” electronic transition energy. The electric dipole-moment operator1 is

m̂ = m |eihg| + |gihe| , (1.3)

and the electric field of the pulse is assumed to take the form

E(t) = eEf (t) cos(Ωt + ϕ) , (1.4)

at the molecule’s location, with polarization e, pulse envelope f (t) peaked at t = 0


with approximate temporal width σ, carrier frequency Ω, and unspecified optical phase
ϕ.
The molecular state evolves according to
d 1
|Ψ(t)i = H(t)|Ψ(t)i , (1.5)
dt i~
with the initial condition

|Ψ(t  0)i = [t]|gi|ng i = |gi[t]gg |ng i = |gi|ng ie−itng /~ . (1.6)

We have adopted the notation


1 The electronic transition moment m may depend on nuclear coordinates in general, but we make
a “Condon approximation” by neglecting this possibility for the present.
2 Short-pulse electronic absorption

[t] ≡ exp{−iHt/~} and [t]gg ≡ hg| exp{−iHt/~}|gi = exp{−iHg t/~} , (1.7)


and |ng i is a nuclear eigenket in the electronic ground state obeying Hg |ng i = ng |ng i.
We shall seek a perturbative solution through second order in V , which can be written
abstractly as
|Ψ(t)i = |0i + | ↑i + | ↑↓i . (1.8)
Here |0i = [t]|gi|ng i for all times; explicit formulas for | ↑i = |eihe|Ψ(t)i and | ↑↓i ∼
∼ =
|gihg|Ψ(t)i − |0i will be developed shortly.

1.2 Energy changes of molecule and field


In pulsed absorption, the molecular system experiences a change in energy
∆E = hΨ(t  0)|H|Ψ(t  0)i − hΨ(t  0)|H|Ψ(t  0)i

= h↑ |H| ↑i + h0|H| ↑↓i + h↑↓ |H|0i
= h↑ |H| ↑i + 2Reh0|H| ↑↓i . (1.9)
By energy conservation, it must be that ∆E + ∆U = 0, where ∆U is the accompa-
nying energy change of the electromagnetic field, calculated to the same order in V .
The electromagnetic energy change results from interference between the propagating
incident field and the electric field E(t) radiated by the oscillating dipole it induces in
the molecule:
Z
1
d3 s (E(t) + E(t))2 − E 2 (t)
 
∆U =

Z
∼ 1
= d3 s E(t) · E(t) ; (1.10)

E(t) in this expression is obtained from that in eqn (1.4) by replacing t with t − u · s/c,
where u is a unit vector in the propagation direction and s = sn locates the field-point.
The radiated dipolar field is given by
1
m̈(t − sc ) × n × n

E(t) = 2
c s
1
= − 2 (1 − nn) · m̈(t − sc ) ; (1.11)
c s
the double dots denote a second derivative with respect to time, and integration in
eqn (1.10) is over the saucer-like spatial region where the two fields overlap, depicted
in Fig. 1.1. The time-dependent dipole in eqn (1.11) is
m(t) = hΨ(t)|m̂|Ψ(t)i = 2Reh0|m̂| ↑i , (1.12)
through first order in the external field. If we take z to be the direction of incidence of
the laser pulse and x to be its polarization direction, then the energy change becomes
Z ∞
∆U = −EΩ dτ f (τ ) sin(Ωτ + ϕ)mx (τ ) ; (1.13)
−∞

this conversion from spatial to temporal integration is carried out in Appendix A.


Expectation values 3

x,y

~ct

~cσ

Fig. 1.1 Disk-shaped region of overlap between the propagating laser pulse and the induced
dipolar field from a molecular source at the origin. Electromagnetic interference between these
fields accounts for the light pulse’s energy change due to its interaction with the molecule.

As hinted by the physically suggestive form of eqn (1.13), there is a direct route
to this formula by way of a basic one for the rate of change of the molecular energy.
In general, one has

dE d
= hΨ(t)|H(t)|Ψ(t)i = h dΨ dΨ dH
dt |H(t)|Ψi + hΨ|H(t)| dt i + hΨ| dt |Ψi . (1.14)
dt dt

Using | dΨ 1
dt i = i~ H(t)|Ψi and H(t) = H + V (t), this becomes dE/dt = hΨ|dV /dt|Ψi;
the well-known statement that the time derivative of the expectation value of the
Hamiltonian equals the expectation value of the Hamiltonian’s time derivative. In the
present instance, we have

dE dE(t)
=− · hΨ(t)|m̂|Ψ(t)i ∼
= EΩf (t) sin(Ωt + ϕ)mx (t) , (1.15)
dt dt
which, together with energy conservation, returns eqn (1.13) upon integration, irre-
spective of the order in E to which the induced dipole is approximated.

1.3 Expectation values


In order to relate ∆E and ∆U to the underlying quantum mechanical molecular dynam-
ics, we need formulas for the zeroth-, first-, and second-order contributions to the time-
dependent state (1.8). The corresponding interaction-picture ket, |Ψ̃(t)i ≡ [−t]|Ψ(t)i,
obeys
d 1  1
|Ψ̃(t)i = [−t] − H + H + V (t) |Ψ(t)i = Ṽ (t)|Ψ̃(t)i , (1.16)
dt i~ i~
with Ṽ (t) = [−t]V (t)[t], subject to the initial condition |Ψ̃(t  0)i = |gi|ng i. The
1
Rt
formal solution, |Ψ̃(t)i = |gi|ng i + i~ −∞
dτ Ṽ (τ )|Ψ̃(τ )i, can be iterated to obtain a
second-order approximation,

|Ψ̃(t)i = 1 + P (t; τ ) + P (t; τ )P (τ ; τ̄ ) |gi|ng i , (1.17)
4 Short-pulse electronic absorption

1
Rt
where P (t; τ ) ≡ i~ −∞
dτ Ṽ (τ ).2
For an electronically resonant or near-resonant field, it is a good approximation
to neglect highly oscillatory terms in P ’s integrand (i.e., to make a rotating-wave
approximation) and write

P (t; τ ) = iF |eihg|e−iϕ p(eg) (t; τ ) + |gihe|eiϕ p(ge) (t; τ ) ,



(1.18)

where F ≡ Eσe · m/2~, and


Z t

p(eg)
(t; τ ) ≡ [−τ ]ee [τ ]gg f (τ )e−iΩτ (1.19)
−∞ σ

is a reduced pulse propagator ; here [t]ee = he|[t]|ei = exp{−i( + He )t/~}. It follows


from this definition that p(ge) = (p(eg) )† .
Using this notation in eqn (1.17) and reverting to the Schrödinger picture yield

|Ψ(t)i = |gi[t]gg 1 − F 2 p(ge) (t; τ )p(eg) (τ ; τ̄ ) |ng i




+ |eiiF e−iϕ [t]ee p(eg) (t; τ )|ng i , (1.20)

whence |0i = |gi[t]gg |ng i (as stated above),

| ↑i = |eiiF e−iϕ [t]ee p(eg) (t; τ )|ng i , (1.21)

and
| ↑↓i = −|giF 2 [t]gg p(ge) (t; τ )p(eg) (τ ; τ̄ )|ng i . (1.22)

The reduced pulse propagator p(eg) is evidently responsible for shaping the e-state
nuclear wave packet generated from |ng i in the e ← g transition, while the nested
combination p(ge) p(eg) sculpts the second-order wave packet in the g-state.
Upon substituting these forms in eqn (1.9), we find

∆E = F 2 hng | p(ge) (∞; τ )( + He ) p(eg) (∞; τ̄ )|ng i


− 2F 2 Rehng |Hg p(ge) (∞; τ ) p(eg) (τ ; τ̄ )|ng i
= 2F 2 Rehng | p(ge) (∞; τ )( + He ) − Hg p(ge) (∞; τ ) p(eg) (τ ; τ̄ )|ng i ;

(1.23)

the observation time has been set to infinity because the energy increment stops chang-
ing after the end of the laser pulse. Notice that the molecular energy change is indepen-
dent of the optical phase ϕ, as we should expect under the conditions considered here.
The first term in the last member of eqn (1.23) accounts for the energy gained by pop-
ulation transfer to the electronic excited state, while the second tracks the energetic
consequence of ground-state “bleaching” (i.e., the loss of g-state population).

2 The τ appearing after the semicolon in P (t; τ ) is not an argument of the pulse propagator,
but simply identifies its integration variable. This notation proves useful in writing nested pulse
propagators like those appearing in the second-order term of eqn (1.17).
The Heller formula 5

For the change in electromagnetic energy (1.13) on the other hand, we have
Z ∞
∆U = −EΩmx dτ f (τ ) sin(Ωτ + ϕ)
−∞

× 2Re iF e−iϕ hng |[−τ ]gg [τ ]ee p(eg) (τ ; τ̄ )|ng i .



(1.24)

Within the rotating-wave approximation, this formula reduces to

∆U = −2F 2 ~Ω Rehng |p(ge) (∞; τ )p(eg) (τ ; τ̄ )|ng i


= −F 2 ~Ω hng |p(ge) (∞; τ )p(eg) (∞; τ̄ )|ng i , (1.25)

which relates the field’s energy loss to the center frequency of the laser pulse and the
squared norm of the nuclear wave packet it generates in the excited electronic state.
R∞
Use p(ge) (∞; τ )( + He ) − Hg p(ge) (∞; τ ) = −∞ dτ d
 iΩτ
σ i~ dτ [−τ ]gg [τ ]ee e f (τ ) along
with integration by parts to verify once again, acknowledging approximations, that
∆E = −∆U, as required by energy conservation.

1.4 The Heller formula


We can explore the effect of pulse duration by adopting a specific, Gaussian form
f (τ ) = exp{−τ 2 /2σ 2 } for the envelope function. If we write
Z ∞
dτ ∞ dτ̄
Z
i
∆E = ~ΩF 2 f (τ )f (τ̄ )hng | e− ~ (+He −ng −~Ω)(τ −τ̄ ) |ng i , (1.26)
−∞ σ −∞ σ

and introduce new integration variables T = (τ + τ̄ )/2 and t = τ − τ̄ , we then find


Z ∞
√ dt iΩt i
∆E = ~ΩF π2
e hng | e− ~ (+He −ng )t |ng i exp{−t2 /4σ 2 } . (1.27)
−∞ σ

In the continuous-wave (cw) limit (σ longer than the inverse of the frequency spacing
between adjacent vibronic levels), eqn (1.27) becomes Heller’s celebrated expression
for the molecular electronic absorption spectrum as a Fourier transform of the time-
dependent overlap between a propagating e-state nuclear wave packet and the g-state
wave function from which it originates.3
Equation (1.27) can be formally evaluated for shorter pulses (lower spectral reso-
lution) by using e-state nuclear eigenkets obeying He |ne i = ne |ne i to obtain
2
X
∆E = 2π~ΩF 2 |hne |ng i|2 exp{− σ~2 ( + ne − ~Ω − ng )2 } . (1.28)
ne

As in the cw case, however, it is often more fruitful to evaluate ∆E by treating the


wave-packet overlap in Heller’s formula under short-time, semiclassical, or system-
bath-decomposition approximations. Reduced pulse propagators themselves can often
3 Eric J. Heller, The Semiclassical Way to Dynamics and Spectroscopy (Princeton University Press,
Princeton, 2018).
6 Short-pulse electronic absorption

be handled with a variety of approximations relying on the brevity of ultrashort laser


pulses relative to the timescale of nuclear dynamics.

1. Derive eqn (1.28) from eqn (1.27) by inserting the stated vibronic completeness
relation and carrying out the resulting Gaussian integration.
2. Repeat this derivation,
P Pstarting instead
R ∞ from the final form in eqn (1.25) and
i τ~ (+ne −~Ω−ng )
using p(eg) (∞, τ ) = ne ng |ne ihng | −∞ dτ
σ f (τ )e .

1.5 Systems starting in thermal equilibrium


In the treatment given so far, the initial state of the target system together with its
environment is described “microcanonically,” taking it to be of a specified-energy form
proportional to |gi|ng i. In this approach, the pulsed-absorption signal, along with the
nonlinear optical signal expressions to be developed in subsequent chapters, is given
as hng |Ô|ng i, where Ô is a specified nuclear coordinate operator expressible in terms
of wave-packet-reshaping reduced pulse propagators and episodes of evolution under
the nuclear Hamiltonians associated with one or more electronic states; eqns (1.23)
and (1.25) both have this structure.
It is to be appreciated that the signal formulas derived in this manner are much
more general than the microcanonical set-up might seem to suggest. For the initial
nuclear state |ng i is assumed to be an eigenstate of Hg (not necessarily the ground
state) and this Hamiltonian can be regarded as governing both “system” and “envi-
ronmental” nuclear degrees of freedom: the former would comprise the translational,
orientational, and internal vibrational modes of a target chromophore and may also
include some nuclear degrees of freedom belonging to its immediate surroundings,
while the latter would consist of the nuclear degrees of freedom of the remainder of
the molecular environment.
If the system-environment boundary is defined spatially, for instance, and is set
far enough away from the target chromophore that sound waves emanating from this
molecule due to pulse-induced electronic transitions and the accompanying nuclear
dynamics cannot propagate to the boundary and back on the overall timescale of
the measurement (perhaps several tens of picoseconds), then the relevant portion of
the nuclear operator Ô = Ôsys ⊗ Ienv pertains to the system alone. Reduced pulse
(eg)
propagators, in particular, take simplified forms such as p(eg) = psys ⊗ Ienv to an
excellent approximation, as the same condition automatically applies to the duration
of any incident laser pulse. The signal can then be written as
hng |Ô|ng i = Trsys [Ôsys ρ̂sys ] , (1.29)
where the quantum mechanical trace is taken over system degrees of freedom and
ρ̂sys = Trenv [ |ng ihng | ] is the system density operator expressed as a trace over envi-
ronmental degrees of freedom of the microcanonical distribution.
If the “environment” is sufficiently large and sufficiently weakly coupled to the
system that it becomes equivalent, as regards the statistical properties of the system,
to a heat bath at some absolute temperature T, then the initial density operator of
(sys)
the system will take the canonical form ρ̂sys = Z −1 exp{−Hg /kB T }. Here Z =
Example absorption calculations 7

(sys)
Trsys [e−Hg /kB T ] is the partition function of the system and kB is the Boltzmann
constant. In this way, any of the optical signal expressions investigated here can be
readily translated into the corresponding finite-temperature Boltzmann-weighted form.
A handy shortcut is to regard the |ng i themselves as nuclear eigenstates of a
subsystem in weak contact with a temperature reservoir. Then the thermally weighted
sum of individual signal contributions hng |Ô|ng i reduces to Tr[Ôρ̂], with
ρ̂ = Z −1 ng |ng ie−ng /kB T hng |, or Tr[Ôρ̂] = Z −1 ng e−ng /kB T hng |Ô|ng i.
P P

1.6 Example absorption calculations


We can illustrate some basic features of pulsed linear absorption by investigating a
model one-dimensional system with potential-energy curves reminiscent of a diatomic
molecule embedded in a solid medium. The nuclear Hamiltonians in eqn (1.1) are
p2x
taken to be Hg(e) = 2m + Vg(e) (x), where

mω 2 2
Vg (x) = x − αx3 + βx4 (1.30)
2
and Ve (x) = Vg (x − xe ). Here α = ~ω/32x3rms and β = ~ω/860x4rms , with xrms =
p
~/2mω. We set the wave number of the harmonic vibration to 200 cm−1 , which
corresponds to a period of τx = 2π/ω = 166.8 fs. The spatial shift of the e-state
potential is assigned the value xe = 2.626xrms , so the classical inner turning point at
an energy 2e lies at x = 0, directly above the g-state minimum. It is not necessary to
specify a value for the vibrational reduced mass m. Fig. 1.2 shows the two potential-
energy curves. The bare electronic transition energy  in eqn (1.1) is given an arbitrary
wave number of 20, 000 cm−1 .

1200

1000
V(x) (cm-1 )

800
Vg(x) Ve(x)
600

400

200

0
0 5 10 15
x/xrms

Fig. 1.2 Potential curves governing nuclear motion in the g- and e-states of the one-dimen-
sional model Hamiltonian. Ve does not include the bare electronic transition energy.

For this simple system, it is a straightforward numerical exercise to solve for the
eigenvalues and eigenstates of both nuclear Hamiltonians and to obtain position repre-
sentations of the latter, using a harmonic-oscillator basis for instance. Matrix elements
of the reduced pulse propagators can also be easily determined. The absorption spec-
trum can then be evaluated as a function of the carrier frequency for any incident pulse
8 Short-pulse electronic absorption

duration, using either eqn (1.25) or eqn (1.28). The spectrum for σ = 5τx starting from
the initial state |g|0g , shown in Fig. 1.3, is of sufficient resolution clearly to discern
individual vibronic transitions. The absolute value of the corresponding Heller kernel
appearing in the integrand of eqn (1.27) is plotted in the same figure, and reflects the
complicated wave-packet dynamics that ensues upon the abrupt promotion of |0g  to
the e-state potential. It exhibits quasi-harmonic motion for a couple vibrational pe-
riods, after which wave-packet spreading in Ve (x) born of anharmonicity gives rise to
less regular temporal evolution in the overlap. The kernel’s overall amplitude decays
on a timescale set by the pulse-duration parameter σ.

~ (✏+He ✏0g )t |0g i]


#%! #%!

!%( !%(

!%' i !%'
Abs[h0g | e
!%& !%&

!%$ !%$
2
t2 /4

!%! !%!
e

! "!! #!!! #"!! $!!! ! " #! #" $!

ℏΩ!ϵ ")*!! # !!τ!

Fig. 1.3 On the left is the absorption spectrum for the 1D model, calculated for the case of
narrow-bandwidth σ = 5τx pulses of continuously variable center frequency. The dimension-
less field-strength parameter F is chosen so that the highest (1e ← 0g ) peak has unit intensity.
The right panel shows the absolute value of the temporal overlap function, which provides a
wave-packet-dynamical description of the absorption process.

The absorption spectrum of the same system at temperature T = 298 K, with the
same pulse duration and field strength, is plotted as a solid line in Fig. 1.4. While the
spectral resolution is unchanged, the Boltzmann-weighted distribution of population
over initial levels with ng = 0, 1, . . . leads to the appearance of new peaks at various
ne ← 1g , 2g , . . . transition frequencies. The resolution is degraded with the use of
wider-bandwidth, σ = 0.5τx pulses. The resulting linear absorption spectrum of the
model system in the ng = 0 state is shown as the dashed line in Fig. 1.4.
The spectra in Figs 1.3 and 1.4 illustrate several key ideas. The quasi-cw absorp-
tion of this relatively simple system provides detailed information on the energy eigen-
spectrum and the Franck–Condon overlaps ne |ng , and is amenable to a dynamical
description through the Heller kernel. In more complicated, multi-mode systems, how-
ever, thermal congestion and the presence of many closely spaced energy levels make it
more difficult to resolve individual vibronic transitions. Short-pulse linear absorption,
on the other hand, pays for its temporally well-localized excitation with a correspond-
ing loss of spectral resolution. In a one-pulse experiment, this loss of resolution in the
frequency domain remains uncompensated by improved temporal resolution resulting
from precise control of the timing within a sequence of individual light-matter interac-
tions. As is explored in subsequent chapters, strategies have been developed to address
Example absorption calculations 9

1.0

0.8

0.6

0.4

0.2

0.0
0 500 1000 1500 2000
- (cm-1 )

Fig. 1.4 Solid line is the absorption spectrum of the model at room temperature, again with
σ = 5.0τx . Dashed curve gives the spectrum starting from ng = 0, with σ = 0.5τx .

this conundrum by systematically varying the delay between excitation and probing
in several different forms of ultrafast nonlinear optical spectroscopy.






 


      



  


     


Fig. 1.5 Nuclear wave packets in the electronic excited state of the model system at t = 0
(top) and 0.5753τx (bottom). The legend identifies the long, intermediate, and short values
of the pulse-duration parameter σ used in these calculations
.

Figure 1.5 illustrates the dynamics initiated by pulses of differing duration in the
model molecule. It shows wave packets in the position representation,

ψe (x, t) ≡ x|e−i(He −2e )t/ p(eg) (∞; τ )|0g  , (1.31)

at t = 0 and 0.5753τx (about half the period of vibrational motion at energy 2e ), with
center frequency Ω = ( + 2e − 0g )/. The upper panel plots the real part of ψe (x, 0).
For the longest pulse-length, σ = 2.0τx , this initial wave packet closely resembles the
10 Short-pulse electronic absorption

stationary wave function hx|2e i. The shortest pulse, with σ = 0.2τx , essentially copies
hx|0d i into the excited electronic state. Each of these wave packets after half a period of
evolution, along with one generated by a pulse of intermediate duration, is illustrated
in the lower panel. The longest-pulse packet remains almost unchanged, while that
generated by the shortest pulse migrates to larger distance and is distorted by motion
in the anharmonic e-state potential. It is ultrashort laser pulses’ capacity to initiate
and interrogate wave-packet dynamics that underlies their usefulness in time-resolved
nonlinear optical spectroscopy.
2
Adiabatic approximation

2.1 Molecular Hamiltonian


The fact that atomic nuclei tend, as a result of their greater mass, to move more slowly
than electrons do undergirds many basic features of molecular structure and dynamics.
In particular, it motivates the ubiquitous Born–Oppenheimer, or adiabatic approxi-
mation, which greatly simplifies the quantum mechanical description of stationary and
nonstationary molecular states. In this chapter, we describe the molecular adiabatic
approximation and spell out the conditions for its validity.1 In succeeding chapters, we
will encounter additional applications of the quantum mechanical adiabatic theorem,
which can be invoked whenever there is a clear timescale separation between relevant
molecular degrees of freedom.
The time-independent Schrödinger for a molecule takes the general form

(TN + Te + Vee + VeN + VN N )|Ψi = E|Ψi , (2.1)

in which the Hamiltonian operator is the sum of five terms:

X P̂α · P̂α
TN =
2Mα
nuclei α

is the nuclear kinetic energy operator,


X p̂i · p̂i
Te =
2m
electrons i

is the electronic kinetic energy,


XX e2
Vee =
i j>i
|r̂i − r̂j |

is the repulsive, electron-electron Coulombic potential energy,


X X (−Zα e)e
VeN =
i α |r̂i − R̂α |

1 Some of the arguments in this section follow those made in Chap. 21 of Gordon Baym, Lectures
on Quantum Mechanics (Westview Press, New York, 1990).
12 Adiabatic approximation

is the attractive, electron-nuclear Coulombic interaction, and


X X (−Zα e)(−Zβ e)
VN N =
α β>α |R̂α − R̂β |

is the repulsive, nucleus-nucleus Coulombic interaction.


The ratio of electron mass to nuclear mass is typically m/M ∼ 10−4 to 10−5 .
Typical electronic momenta are of size h/λdeBroglie ∼ ~/a, where a is the molecu-
lar size, so successive electronic energy levels are separated by about ∼ ~2 /ma2 , or
several electron volts. If a nucleus were displaced by the full molecular size a from
its equilibrium position, the energy would change by an amount on the order of this
electronic transition energy. Regarding the nuclear motion as roughly harmonic, then,
M ω 2 a2 ∼ ~2 /ma2 , which implies
 m 1/2 ~2
~ω ∼ . (2.2)
M ma2
Vibrational excitation energies are thereby found to be a smaller than electronic en-
ergies by a factor (m/M )1/2 ∼ 10−2 . Additionally, since the nuclear kinetic energy is
of size P 2 /2M ∼ ~ω/2, nuclear momenta are roughly
 1/4  1/4
M ~ M
P ∼ ∼ p. (2.3)
m a m

Thus nuclear momenta are about ten times larger than electronic momenta, and nu-
clear speeds P/M are one thousand times smaller than electronic speeds. Finally, the
typical deviation of a nuclear position from its equilibrium average value is given by
M ω 2 δ 2 /2 ∼ ~ω/2, which leads to the estimate (δ/a)2 ∼ (m/M )1/2 , or
 m 1/4
δ∼ a; (2.4)
M
the spatial range of zero-point nuclear motion is about one-tenth the range of electronic
motion (i.e., one-tenth the “molecular size”). These ballpark figures support a physical
picture of a molecule as a well-defined configuration of nuclei executing slow, small
amplitude oscillations around their equilibrium positions within a delocalized medium
of rapidly moving electrons.

Make a physical argument estimating molecular rotational excitation energies rela-


tive to vibrational and electronic energies.

2.2 Molecular eigenstates


Since the nuclei are well localized compared with the electrons—and much more slowly
moving—while the nuclear excitation energies are a small fraction of electronic ener-
gies, we are led to regard the nuclear kinetic energy operator TN as a perturbation
and seek solutions to the electronic Schrödinger equation which obey
Molecular eigenstates 13

(Te + Vee + VeN + VN N )|ϕn (R̂)i = En (R̂)|ϕn (R̂)i . (2.5)

The electronic eigenkets are parametrized by the nuclear coordinate operators, which
we symbolically denote by R̂ = (R̂1 , R̂2 , . . . ). The electronic eigenenergies En (R̂) like-
wise depend on the locations of the nuclei.
We can relate the nuclear-coordinate-operator-dependent electronic eigenkets to
those at some reference nuclear arrangement—say the equilibrium configuration in
the electronic ground state—by a unitary tranformation:

|ϕn (R̂)i = U (R̂)|ϕn i . (2.6)

The electronic unitary operator U (R̂) is actually a large collection of different unitary
operators parametrized by R̂. Since the reference eigenkets |ϕn i constitute a complete
orthonormal set in the electronic state-space, we could P seek molecular eigenkets as
expansions in this fixed electronic basis of the form n |ϕn i|Θn i, where the |Θn i are
nuclear states. But it proves more useful by far to look for energy eigenstates of the
combined electronic and nuclear degrees of freedom as expansions in the adiabatic
electronic basis: X X
|Ψi = |ϕn (R̂)i|Φn i = U (R̂) |ϕn i|Φn i . (2.7)
n n

Justification for the adiabatic electronic basis as the practical choice emerges when
we substitute |Ψi in the molecular Schrödinger equation (2.1) and take account of the
small ratio of electronic to nuclear mass.
Making this substitution and availing ourselves of the electronic Schrödinger equa-
tion (2.5) lead to
Xh i X
TN + Em (R̂) |ϕm (R̂)i|Φm i = E |ϕm (R̂)i|Φm i . (2.8)
m m

Taking the inner product of eqn (2.8) with |ϕn (R̂)i yields a set of coupled equations
for the nuclear states {|Φn i}:
h i X
hϕn (R̂)|TN |ϕn (R̂)i + En (R̂) |Φn i + hϕn (R̂)|TN |ϕm (R̂)i|Φm i = E|Φn i . (2.9)
m6=n

Let us examine the operator coupling nuclear kets associated with different electronic
states, which can be rewritten as

X P̂α · P̂α
hϕn (R̂)|TN |ϕm (R̂)i = hϕn |U † (R̂) U (R̂)|ϕm i
α
2Mα
X1
= hϕn | U † (R̂)P̂α U (R̂) · U † (R̂)P̂α U (R̂)|ϕm i
α
2M α
X 1 h i h i
= hϕn | P̂α + Aα (R̂) · P̂α + Aα (R̂) |ϕm i , (2.10)
α
2Mα
14 Adiabatic approximation

where the induced vector potential Aα (R̂) = −i~U † (R̂)∇α U (R̂) is an operator in the
electronic state-space parametrized by the nuclear coordinate operators.2 Now
X 1
hϕn (R̂)|TN |ϕm (R̂)i = hϕn |ϕm i P̂α · P̂α
α
2Mα
X 1  
+ P̂α · hϕn |Aα (R̂)|ϕm i + hϕn |Aα (R̂)|ϕm i · P̂α
α
2Mα
X 1
+ hϕn |Aα (R̂) · Aα (R̂)|ϕm i . (2.11)
α
2Mα

The inner product hϕn |ϕm i equals δnm . The matrix elements of the vector potential,
hϕn |Aα (R̂)|ϕm i = −i~hϕn (R̂)|∇α ϕm (R̂)i, are of estimated size ~/a as one would have
to move the αth nucleus a molecule-sized distance to produce near unit overlap between
a given adiabatic electronic state and another to which it had been orthogonal. The
second sum on the right-hand side of eqn (2.11) therefore contains terms of order
3/4
p2

~ m
Pα ∼ ,
Mα a m Mα
a thousand times smaller than electronic excitation energies. Terms in the third sum
are of similar or smaller size,
1 1 X
hϕn |Aα (R̂) · Aα (R̂)|ϕm i = hϕn |Aα (R̂)|ϕl i · hϕl |Aα (R̂)|ϕm i
2Mα 2Mα
l
 2
1 X ~ m p2
∼ ∼ 10 ,
2Mα a Mα 2m
l

assuming that ten states contribute significantly.


We see that the portions of TN coupling different electronic states are about one
one-thousandth of a standard electronic energy splitting. The corresponding pertur-
bative corrections to a molecular eigenstate3 go as
matrix element
∼ 10−3 ,
electronic energy difference
while corrections to the eigenenergy are of size
(matrix element)2 ~2
∼ 10−6 .
energy difference ma2
Except in the presence of degenerate or unusually closely spaced electronic energies,
both of these corrections should be negligibly small.
2 The ∂ ∂ ∂
gradient with respect to the αth nuclear coordinate is ∇α = i +j +k . In
∂Xα ∂Yα ∂Zα
evaluating ∇α U (R̂) it is to be understood that we first calculate ∇α U (R1 , R2 , · · · ) and then replace
(R1 , R2 , · · · ) with (R̂1 , R̂2 , · · · ).
3 See, for example, Chap. 5 of J. J. Sakurai, Modern Quantum Mechanics, Revised Edition
(Addison-Wesley, Reading, 1994).
Molecular eigenstates 15

We are led to adopt the Born–Oppenheimer approximation, under which the ex-
pansion in eqn (2.7) is reduced to a single term,4

|Ψi ∼
= |ϕn (R̂)i|Φn i . (2.12)

In this approximation, the nuclear state that accompanies |ϕn (R̂)i obeys the equation
h i
hϕn (R̂)|TN |ϕn (R̂)i + En (R̂) |Φn i = E|Φn i , (2.13)

whose eigenvalues are the approximate molecular eigenenergies.


Using eqn (2.11) with m set equal to n and writing

hϕn |Aα (R̂) · Aα (R̂)|ϕn i = hϕn |Aα (R̂)|ϕn i · hϕn |Aα (R̂)|ϕn i
X
+ hϕn |Aα (R̂)|ϕm i · hϕm |Aα (R̂)|ϕn i ,
m6=n

we can rewrite eqn (2.13) as


( )
X 1 h i h i
(n) (n) 0
P̂α + Aα (R̂) · P̂α + Aα (R̂) + En (R̂) |Φn i = E|Φn i , (2.14)
α
2Mα

where

A(n)
α (R̂) = −i~hϕn |U (R̂)∇α U (R̂)|ϕn i = −i~hϕn (R̂)|∇α ϕn (R̂)i (2.15)

is the adiabatic vector potential experienced by the αth nucleus in the nth electronic
state, and
X 1 X
En0 (R̂) = En (R̂) + hϕn |Aα (R̂)|ϕm i · hϕm |Aα (R̂)|ϕn i . (2.16)
α
2Mα
m6=n

The nonadiabatic correction to En (R̂) is typically minuscule, as outlined above for


hϕn |Aα (R̂) · Aα (R̂)|ϕm i/2Mα . The adiabatic vector potentials can often—but not
(n)
always5 —be made to vanish entirely: Aα (R̂) must be real, for

hϕ|∇ϕi = ∇ hϕ|ϕi −h∇ϕ|ϕi = −hϕ|∇ϕi∗


| {z }
1

is imaginary. If the electronic wave function hr1 , r2 · · · |ϕn (R̂)i can be chosen real,
(n)
as it most often can be, then hϕn (R̂)|∇α ϕn (R̂)i will be real. Hence Aα (R̂) is also
imaginary and therefore zero.
4 The Born–Oppenheimer-approximated state |Ψi is not, in general, a tensor product. Rather it is
a “sum” of products d3 R1 d3 R2 · · · |ϕn (R1 , R2 , · · · )i|R1 , R2 , · · · ihR1 , R2 , · · · |Φn i, in which the
R R
electronic state is correlated with the arrangement of the nuclei and weighted by the probability
amplitude, Φn (R1 , R2 , · · · ) = hR1 , R2 , · · · |Φn i, for the nuclei to have a particular arrangement.
5 M. V. Berry, “Quantal phase factors accompanying adiabatic changes,” Proc. R. Soc. Lond.
A392, 45–57 (1984).
16 Adiabatic approximation

When the dust settles, it is often sufficient to combine a solution of the simplified
nuclear wave equation, h i
TN + En (R̂) |Φn i = E|Φn i , (2.17)

in which the electronic eigenenergy as a function of nuclear coordinates acts as a


potential energy surface for the nuclear motion, with the adiabatic electronic state
|ϕn (R̂)i to obtain an approximate molecular eigenket with total energy (electronic
plus nuclear) equal to E.

1. How much can you figure out about the eigenstates and eigenenergies of the
model Hamiltonian
P12 + P22 M ω2 2
H= + (Q1 + Q22 ) + κQ1 σx + κQ2 σy ,
2M 2
where (Q1 , Q2 ) and (P1 , P2 ) are vibrational coordinates and momenta, respectively,

σx = |eihg| + |gihe| ,

σy = −i|eihg| + i|gihe| ,
and
σz = |eihe| − |gihe| ,

are Pauli operators in a two-dimensional electronic state-space, and κ = k ~M ω 3 ,
with a dimensionless constant k, is a coupling coefficient between electronic and
nuclear degrees of freedom?
2. In addition to determining the stationary nuclear wave functions, the approximate
Hamiltonian appearing in eqn (2.17) governs nuclear dynamics. Consider a one-
dimensional vibrational Hamiltonian,

P2 M ω2 2
T + E(X) = + X ,
2M 2
where [X, P ] = i~. Investigate the time-dependent expectation value and mean-
square deviation of X and P in a Glauber coherent state specified by the initial
condition |Φi = e−iP̂ X◦ /~ |0i, where X◦ is some initial displacement and |0i is the
ground state of this harmonic nuclear Hamiltonian.
Two references pertinent to the first of these exercises are given below.6

6 H. C. Longuet-Higgins, U. Öpik, M. H. L. Pryce, and R. A. Sack, “Studies of the Jahn-Teller


effect. II. The dynamical problem,” Proc. R. Soc. Lond. A 244, 1–16 (1958); J. C. Slonczewski and
V. L. Moruzzi, “Excited states in the dynamic Jahn-Teller effect,” Physics 3, 237–254 (1967).
3
Transient-absorption spectroscopy:
Making ultrashort pulses worthwhile

This chapter develops a basic description of ultrafast transient-absorption spectroscopy


and presents some calculated signals. By driving a sample with electronically resonant
pump and delayed probe pulses, both of short duration on the timescale of atomic mo-
tion, and spectrally decomposing the transmitted probe beam, this widely applied tech-
nique overcomes short-pulse absorption’s lack of spectral resolution in two respects:
it gains high temporal resolution due to a precisely controlled time-delay between
the pump and probe pulses, and it monitors spectral dynamics through frequency-
by-frequency observation of the time-varying pump-induced intensity change of the
transmitted probe. Transient-absorption spectroscopy provides direct information on
wave-packet dynamics in more extended regions of the ground and excited electronic
potential energy surfaces than linear absorption and, because it involves two pulses
which may have different spectral composition, enables selective probing of multiple
higher-lying states.

3.1 Model Hamiltonian and signal expression


We treat a molecule having three relevant electronic states, with the Hamiltonian

H = |giHg hg| + |ei(He + e )he| + |f i(Hf + f )hf | , (3.1)

the potential energy surfaces of whose nuclear Hamiltonians take the value zero at
their respective minima, and whose bare electronic energies obey f − e < e < f .
The molecule’s fixed location identifies the spatial origin. Its interaction with two
ultrashort laser pulses is governed by

V (t) = −m̂E(t) ; (3.2)


m̂ = meg (|eihg| + |gihe|) + mf e (|f ihe| + |eihf |) ;
E(t) = Eu fu (t) cos(Λu t + ϕu ) + Er fr (t − td ) cos[Λr (t − td ) + ϕr ] .

The pump (u) and probe (r ) pulses have envelopes fu (t) and fr (t − td ), respectively,
of temporal width ∼ σu and ∼ σr ; their arrival times are separated by td . The two
field strengths (Eu and Er ), center frequencies (Λu and Λr ), and uncontrolled optical
phases (ϕu and ϕr ) are all independent.
The measured quantity in a transient-absorption experiment is the change in the
spectrally resolved loss of energy from the probe pulse due to the prior action of
18 Transient-absorption spectroscopy

the pump. To derive a formula for such a signal, we could calculate the change in
electromagnetic energy—in a narrow frequency range of width δω centered at ω̄—due
to interference between the spatially propagating probe field and the field radiated by
a portion of the system’s induced dipole moment,
Z
1 n
d3 R Erω̄ (R, t) + Eu2 rω̄ (R, t) · Erω̄ (R, t) + Eu2 rω̄ (R, t)
  
∆Uω̄ =

o
− Er2ω̄ (R, t)
Z
∼ 1
= d3 R Erω̄ (R, t) · Eu2 rω̄ (R, t) . (3.3)

In this equation, Eu2 r is the contribution to the induced dipolar field proportional
to pump intensity and the probe field strength. The subscript ω̄ denotes a spectrally
filtered field component; for instance, the filtered probe is given by
Z ω̄+ δω2 dω
Erω̄ (R, t) = e−iωt Ẽr (R, ω) + c.c. , (3.4)
ω̄− δω
2

R∞
where Ẽr (R, ω) = −∞ dt eiωt Er (R, t).1
The integration over space in eqn (3.3) could be evaluated by a procedure analo-
gous to that in Appendix A.2 As an alternative, we pursue a direct derivation of the
transient-absorption signal as a time integral for the contribution to the change in
molecular energy, ∆Eω̄ = −∆Uω̄ , that is bilinear in the pulse intensities. As in eqn
(3.2), we assume for simplicity that both transition dipole moments and both field
polarizations are oriented along the same (say x -) axis. Temporarily writing m(t) for
mu2 r (t) and E(t) for Er (t), we can break up the probe field into pieces which would
be detected by idealized pixels sensitive to different small frequency ranges:
X
E(t) = Eω̄ (t) , (3.5)
ω̄>0

and the transient-absorption dipole could be similarly decomposed as


X
m(t) = mω̄ (t) . (3.6)
ω̄>0

In terms of the latter two quantities, the transient-absorption signal would be written
Z ∞
dEω̄ (t)
∆Eω̄ = − dt mω̄ (t) , (3.7)
−∞ dt
in which both probe-field and the dipole-moment components belong to the spectral
segment centered at ω̄ (compare eqn (1.15)). Now,
1 Equations (3.3) and (3.4) differ from eqn (3.2) in making explicit the vector nature of the incident
and radiated fields.
2 See also J. A. Cina, P. A. Kovac, C. C. Jumper, J. C. Dean, and G. D. Scholes, “Ultrafast
transient absorption revisited: Phase-flips, spectral fingers, and other dynamical features,” J. Chem.
Phys. 144, 175102/1–18 (2016).
Model Hamiltonian and signal expression 19

Z ω̄+ δω
2 dω −iωt
Eω̄ (t) = e Ẽ(ω) + c.c. , (3.8)
ω̄− δω
2

where Z ∞ Z ∞
Ẽ(ω) = dt eiωt E(t) = eiωtd dτ eiωτ E(τ + td ) . (3.9)
−∞ −∞

Since the probe field E(τ + td ) has a temporal width ∼ σ  2π/δω about τ = 0, for
frequencies in the range ω̄ − δω δω
2 < ω < ω̄ + 2 we have
Z ∞
Ẽ(ω) ∼
= eiωtd dτ eiω̄τ E(τ + td ) = ei(ω−ω̄)td Ẽ(ω̄) . (3.10)
−∞

As a result, from eqn (3.8),

sin δω
2 (t − td )
Eω̄ (t) ∼
= 2 Re Ẽ(ω̄)e−iω̄t

. (3.11)
π(t − td )
The molecular energy acquisition involving a particular spectral segment of the
probe field and the full induced dipole moment is
Z ∞
dEω̄ (t) δω
− dt m(t) ∼
= iω̄ Ẽ(ω̄)m̃∗ (ω̄) + c.c. (3.12)
−∞ dt 2π

Since the right-hand side of eqn (3.12) involves the ω̄ Fourier components of both
the probe field and the induced dipole, it clearly corresponds to ∆Eω̄ of eqn (3.7),
the transient-absorption signal at the given frequency. Reintroducing the appropriate
subscripts, the signal may therefore be written
Z ∞
dErω̄ (t)
∆Eω̄ = − dt mu2 r (t) ; (3.13)
−∞ dt

the transient-absorption dipole in its entirety appears in the integrand, rather than
just its ω̄ element.
For a probe pulse of the form given in eqn (3.2), the Fourier components are

Er iωtd −iϕr ∞
Z

Ẽr (ω > 0) = e dt ei(ω−Λr )(t−td ) fr (t − td )
2 −∞
Er iωtd −iϕr ˜
= e fr (ω − Λr ) , (3.14)
2
whence3
sin δω
2 (t − td )
Erω̄ (t) ∼
= Er f˜r (ω̄ − Λr ) cos[ω̄(t − td ) + ϕr ] . (3.15)
π(t − td )
Inserting eqn (3.15) in eqn (3.13) gives a working formula for the transient-absorption
signal,
3 We specialize to symmetric pulse envelopes, so f˜r (ω̄ − Λr ) is real-valued.
20 Transient-absorption spectroscopy


sin δω
2 (t − td )
Z
∆Eω̄ = ω̄Er f˜r (ω̄ − Λr ) dt mu2 r (t) sin[ω̄(t − td ) + ϕr ] . (3.16)
−∞ π(t − td )

The dynamics of the targeted system enters the signal through the interpulse delay-
dependence and spectral content of the induced dipole moment.

Fill in the steps that lead to eqns (3.11), (3.12), and (3.15). Show that eqn (3.11)
is consistent with eqn (3.15) for a probe pulse of the form given in eqn (3.2).

3.2 Transient-absorption dipole


3.2.1 Contributions to mu2 r
In order to calculate mu2 r , the portion of hΨ(t)|m̂|Ψ(t)i that is of second order in
the pump field and first order in the probe, we need a perturbative expansion of the
molecular state including all relevant terms through that same order. In a system with
three electronic states of the variety under study, this expansion can be written

|Ψi ∼
= |0i + | ↑u i + | ↑r i + | ↑u ↑u i + | ↑u ↓u i + | ↑u ↑r i + | ↑u ↓r i + | ↑r ↑u i + | ↑r ↓u i
+ | ↑r ↑r i + | ↑r ↓r i + | ↑u ↑u ↓r i + | ↑u ↓u ↑r i + | ↑u ↑r ↓u i + | ↑u ↓r ↑u i
+ | ↑r ↑u ↓u i + | ↑r ↓u ↑u i . (3.17)

Here |0i is the time-dependent molecular state unperturbed by either laser pulse. The
arrows in the other kets signify energetically upward (g-to-e or e-to-f ) or downward
(e-to-g or f -to-e) electronic transitions driven by the pump or the probe, as indicated
by the subscript. These transitions are required to occur in the order listed from left
to right. Kets in which the pulses act “out of order” (probe before pump) vanish for
interpulse delays significantly longer than the pulse durations. From the expectation
value of the dipole-moment operator, we can isolate

mu2 r (t) = h0|m̂| ↑u ↑u ↓r i +h↑u ↑r ↓u |m̂|0i + h↑r ↑u ↓u |m̂|0i


| {z }
exp{−2iϕu }

+ h↑u ↑r |m̂| ↑u i + h↑r ↑u |m̂| ↑u i + h↑r |m̂| ↑u ↑u i


ESA | {z }
exp{−2iϕu }

+ h↑u ↓u ↑r |m̂|0i + h0|m̂| ↑u ↓r ↑u i +h↑r ↓u ↑u |m̂|0i


GSB | {z }
exp{−2iϕu }

+ h↑r |m̂| ↑u ↓u i + h↑u |m̂| ↑u ↓r i + h↑r ↓u |m̂| ↑u i + c.c. (3.18)


ISRS SE | {z }
exp{−2iϕu }

The four under-braced matrix elements will be seen to carry an uncontrolled optical
phase factor exp{−2iϕu }. Over the many laser shots needed to acquire a transient-
absorption signal, the phase 2ϕu randomly samples the range from zero to 2π, leading
to a negligible net contribution from these elements.
Transient-absorption dipole 21

In the terms in eqn (3.18) labeled excited-state absorption (ESA), ground-state


bleach (GSB), impulsive stimulated Raman scattering (ISRS), and stimulated emis-
sion (SE), none of the pulse actions listed separately within the bra or the ket occur
outside their nominal order (pump before probe). Only these terms survive when td
significantly exceeds the pulse durations. The four remaining elements can contribute
only at shorter delays. Two of these correspond to variations of the ESA and GSB
elements in which the probe interchanges roles with one of the pump actions; in the
absence of different spectral content in the pump and probe, the molecule itself would
not distinguish these terms from their conventionally interpreted counterparts. The
same relationship exists between the SE and ISRS terms. The two other short-td
elements are optical-phase-controlled counterparts of the stricken, unstable element
h0|m̂| ↑u ↑u ↓r i.

Provide physical explanations for each of the interpretive labels in eqn (3.18).

3.2.2 Multi-pulse bras and kets


Time-dependent perturbation theory provides formulas for the various multi-pulse
contributions to the molecular state in eqn (3.17) which enter  the induced dipole

of eqn (3.18). We seek a solution to i~ ∂t |Ψ(t)i = H + V (t) |Ψ(t)i with the initial
condition |Ψ(t  0)i = [t]|gi|ng i, where Hg |ng i = ng |ng i (consult Chapter 1 for
notation). Through third order in V , we find
 X
|Ψ(t)i = [t] + i [t − tj ]Pj (t − tj ; τ )[tj ] (3.19)
j=u,r
X
+ i2 [t − tj ]Pj (t − tj ; τ )[tj − tk ]Pk (τ + tj − tk ; τ̄ )[tk ]
j,k
X 
+ i3 [t − tj ]Pj (t−tj ; τ )[tj −tk ]Pk (τ +tj −tk ; τ̄ )[tk − tl ]Pl (τ̄ +tk −tl ; τ̄¯)[tl ] |gi|ng i ,
j,k,l

where tu = 0, tr = td , and

(eg) −iϕj (eg) (f e) −iϕj (f e)


Pj (t; τ ) = Fj e |eihg| pj (t; τ ) + Fj e |f ihe| pj (t; τ ) + H.c. , (3.20)

after making a rotating-wave approximation. In eqn (3.20), we have used reduced pulse
propagators of the form
Z t
(eg) dτ
pj (t; τ ) = [−τ ]ee [τ ]gg e−iΛj τ fj (τ ) , (3.21)
−∞ σj

(eg)
for example, along with Fj = Ej meg σj /2~.
Rather than generate explicit expressions for all the multi-pulse kets and bras ap-
pearing in eqn (3.18), some of which may become insignificant with particular choices
22 Transient-absorption spectroscopy

of pulse center frequencies and bandwidths, or certain ranges of interpulse delay time,
it is more instructive simply to write out a few examples. We have, for instance,

|0i = [t]|gi|ng i = |gi[t]gg |ng i = |gi|ng ie−ing t/~ ; (3.22)

| ↑u i = |eiiFu(eg) e−iϕu [t]ee p(eg)


u (t; τ )|ng i ; (3.23)

| ↑r i = |eiiFr(eg) e−iϕr [t − td ]ee p(eg)


r (t − td ; τ )[td ]gg |ng i ; (3.24)

and

| ↑u ↑r i = |f ii2 Fu(eg)Fr(f e) e−iϕu −iϕr [t-td ]ff p(f


r
e)
(t-td ; τ )[td ]ee p(eg)
u (τ +td ; τ̄ )|ng i . (3.25)

Gain some practice by writing expressions analogous to eqns (3.22)–(3.25) for a few
more multi-pulse kets.
From expressions of the sort just considered, we can obtain the dipole matrix
elements. For the ESA element, for example, we find
2
h↑u ↑r |m̂| ↑u i = −imf e Fu(eg) Fr(f e) eiϕr hng |p(ge) (ef )
u (τ + td ; τ̄ )[−td ]ee pr (t − td ; τ )
(eg)

× [−t + td ]ff [t − td ]ee [td ]ee pu (t − td ) + td ; τ̄¯ |ng i . (3.26)

If we define a filtered-probe reduced pulse propagator,


dτ sin δω
2 τ
Z
(f e)
prω̄ (∞; τ ) = δω
[−τ ]ff [τ ]ee e−iω̄τ , (3.27)
−∞ σ̄ 2 τ

with σ̄ ≡ 2π/δω, write sin[ω̄(t − td ) + ϕr ] = 2i exp{− iω̄(t − td ) − iϕr } + c.c. in eqn


(3.16), and neglect terms in its integrand that oscillate at optical frequencies, we then
find the contribution of excited-state absorption to the transient-absorption signal,

f˜r (ω̄ − Λr ) (eg) 2 (f e) 2


∆Eω̄ESA = ~ω̄ Fu Fr hng |p(ge) (ef )
u (τ + td ; τ̄ )[−td ]ee pr (t; τ )
σr
(f e)
× prω̄ (∞; t)[td ]ee p(eg) ¯
u (t + td ; τ̄ )|ng i + c.c. , (3.28)
R∞
where f˜r (ω) = −∞ dt eiωt fr (t), as defined in eqn (3.14). Notice is to be taken of the
t-nesting that occurs because the reduced propagator for the spectrally filtered probe
pulse is the last-acting operator. Similar expressions for the other transient-absorption
signal contributions can readily be found by analogy with eqn (3.28).

Carry out this exercise for the contributions of ground-state bleach, impulsive stim-
ulated Raman scattering, and stimulated emission to transient absorption. You can
check your answers by consulting Section 3.3.2.
Exemplary calculations 23

3.3 Exemplary calculations


3.3.1 2D model and absorption spectrum
In order to carry out illustrative calculations of transient-absorption signals, we frame
a model Hamiltonian in the form of eqn (3.1) for a molecular system consisting of
one slower and one faster nuclear degree of freedom, which could be thought of, re-
spectively, as a slow “conformational” coordinate and an internal molecular vibration
whose equilibrium average value is slightly displaced upon electronic excitation. The
three nuclear Hamiltonians are
P2 p2
Hj = + + Vj (X, x) , (3.29)
2 2
for j = g, e, f . The g-state potential is

Ω2 ωg2
Vg (X, x) = (X − Xg )2 + (x − xg )2 . (3.30)
2 2
The slow-mode frequency Ω serves as a reference, set to 40 cm−1 for plotting purposes,
while ωg = 13.5pΩ; the ground-statepminimum lies at (Xg , xg ) = (−1.8 Xrms , 0 xrms ),
where Xrms = ~/2Ω and xrms = ~/2ωg . The slow mode serves as a photochemical
reaction coordinate in the e-state, whose potential function is
 
3
Ve (X, x) = αe (X − Xe )4 + 4Xe (X − Xe )3 + 3Xe2 (X − Xe )2
2
2
ω
+ e (x − xe )2 + βe (X − Xe )(x − xe ) , (3.31)
2
4
with αe = 0.01613 ~Ω/Xrms , Xe = 4.0 Xrms , ωe = 11.0 Ω, xe = 1.5 xrms , and βe =
−0.4550 ~Ω/Xrms xrms . We set the bare electronic energy of this state to e = 600.0 ~Ω.
The f -state potential function is of similar form,
 
3 4 3 2 2
Vf (X, x) = αg (X − Xf ) + 4Xf (X − Xf ) + 3Xf (X − Xf )
2
ωf2
+ (x − xf )2 + βf (X − Xf )(x − xf ) , (3.32)
2
4
in which αf = 0.02419 ~Ω/Xrms , Xf = 3.0 Xrms , ωf = 9.0 Ω, xf = 0.9 xrms , and
βf = −0.3344 ~Ω/Xrms xrms . The bare electronic energy of the f -state is assigned
the value f = 900.0 ~Ω. Contour diagrams of the three electronic potential energy
surfaces are displayed in Fig. 3.1.
The continuous-wave electronic absorption spectrum provides a first acquaintance
with the spectroscopy and dynamics of this system. The model Hamiltonian is simple
enough to be diagonalizable (we used a discrete position basis), and the spectrum can
readily be calculated from eqn (1.28) once the transition energies and Franck–Condon
overlaps are determined. As seen in Fig. 3.2, with the incident frequency denoted as Λ
and line-widths determined by the finite duration σ = 1.6 × 2π/Ω of a Gaussian pulse,
24 Transient-absorption spectroscopy

"

#
!!!"#
!!!
!!

!#

!! !" !# % !! !" !# % !! !" !#

! !!
!!!"# ! !!
!!!"# ! !!
!!!"#
$ # " ! $ # " ! $ # " ! %

!! !! !!

Fig. 3.1 Vg , Ve , and Vf are shown from left to right. Contours are separated by 2 Ω.

1.0

0.8

0.6

0.4

0.2

0.0
0 200 400 600 800

ℏΛ -ϵϵe (cm-1 )

Fig. 3.2 Linear absorption of the two-dimensional model system in arbitrary units versus
wave-number displacement from the e ← g bare electronic transition energy.

the spectrum exhibits interesting progressions in both the low-frequency, reaction-


coordinate mode and the higher-frequency vibration. The transition at e − 34.4 cm−1
from the initially populated lowest level in the g-state to the photochemical “product
state” corresponding to the lowest e-state level—localized in the potential-energy well
at (Xe , xe )—is barely visible due to the tiny Franck–Condon overlap between the as-
sociated nuclear wave functions.

Can you rationalize the value e − 34.4 cm−1 of the energy for the very weak zero-
zero absorption transition through a consideration of the zero-point energies in the
ground and excited electronic states?
The wave-packet dynamics underlying linear absorption in this system is brought
to light by the Heller overlap,

C(t) = 0g |eitHg / e−it(e +He )/ |0g , (3.33)


Exemplary calculations 25

appearing in eqn (1.27). In this formula, |0g i denotes the two-dimensional vibrational
ground state. As explained in Chapter 1, C(t) is the complex-valued overlap between
the initial, stationary vibrational wave function and the moving wave packet engen-
dered by its transfer to the electronic excited state. Fourier transformation of eqn (3.33)
times the temporal window exp{−t2 /4σ 2 } (and some constant pre-factors) yields the
spectrum. The absolute value of the overlap times the window function is plotted in
Fig. 3.3. It decays rapidly as the 2D wave packet gains slow-mode momentum P and

1.0

0.8
|exp{-t2 /4σ2 } C(t)|

0.6

0.4

0.2

0.0
0 1 2 3 4 5 6
t⨯Ω/2π

Fig. 3.3 The magnitude of the Heller overlap times the temporal window versus time in
g-state slow-mode periods. Values taken at several small-integer multiples of 5/2 times the
fast-mode vibrational period in the e-state are marked with dots.

is pushed to larger X values, with a high-frequency sawtooth structure resulting from


small-amplitude oscillatory motion in the x coordinate. Subsequent broad recurrences
arise from the wave packet’s unfocused partial return to the Franck–Condon region,
where it regains overlap with the overall vibrational ground state. Some values of the
plotted quantity at intervals of 2.5 × 2π/ωe are highlighted with black dots.
Figure 3.4 samples the wave-packet dynamics in the excited electronic state. The
real part of the evolving two-dimensional e-state nuclear wave function is shown at
the six equally spaced instants highlighted in Fig. 3.3. The overall phase is referenced
to that of the lowest-lying vibronic level in the excited electronic state. The interval
between the snapshots is a small fraction, 0.2273 × 2π/Ω, of the slow mode’s g-state
period.
The time-evolving average values of X and x are displayed in Fig. 3.4. Both expec-
tation values are seen to move away from their equilibrium values in the ground state
and to oscillate about the e-state product well. Episodes of small hXi coincide with
the times of relatively large overlap between ground- and excited-state wave packets
observed in Fig. 3.3.

3.3.2 One-color transient-absorption signals


Here, we consider “one-color” transient-absorption signals resulting from the response
of our model system to degenerate pump and probe pulses, both of which are resonant
26 Transient-absorption spectroscopy

4 4 4

2 2 2
x/xrms

x/xrms

x/xrms
0 0 0

-2 -2 -2

t=0 0.227 2π/Ω 0.455 2π/Ω


-6 -4 -2 0 2 4 6 8 -6 -4 -2 0 2 4 6 8 -6 -4 -2 0 2 4 6 8

X/Xrms X/Xrms X/Xrms


4 4 4

2 2 2
x/xrms

x/xrms

x/xrms
0 0 0

-2 -2 -2

0.682 2π/Ω 0. 909 2π/Ω 1.136 2π/Ω


-6 -4 -2 0 2 4 6 8 -6 -4 -2 0 2 4 6 8 -6 -4 -2 0 2 4 6 8
X/Xrms X/Xrms X/Xrms

Fig. 3.4 Real part of nuclear wave packet in the excited electronic state at t = 0, 1, ..., 5
times 2.5 (2π/ωe ), the times marked in the plot of Heller’s absorption kernel.

2
〈X〉/Xrms

-1

-2
2.0

1.5
〈x〉/xrms

1.0

0.5

0.0
0 1 2 3 4 5 6
t⨯Ω/2π

Fig. 3.5 hXi and hxi versus time during wave-packet motion in the e-state.
Exemplary calculations 27

with the g-to-e electronic transition.4 Because these pulses are non-resonant with
f ↔ e, all of the overlaps in the transient-absorption dipole of eqn (3.18) involving
amplitude transfer to or from the f -state become negligibly small, and it simplifies to

mu2 r (t) ∼
= h↑u ↓u ↑r |m̂|0i+h↑r ↓u ↑u |m̂|0i+h↑r |m̂| ↑u ↓u i+h↑u |m̂| ↑u ↓r i + c.c. (3.34)

According to the pattern established in Section 3.2.2, the first of these terms gives rise
to a ground-state-bleach signal contribution,

f˜r (ω̄ − Λr ) (eg) 2 (eg) 2


∆Eω̄GSB = −~ω̄ Fu Fr h0g |p(ge) ¯ (eg)
u (τ̄ ; τ̄ )pu (τ + td ; τ̄ )[−td ]gg
σr
(eg)
× p(ge)
r (t; τ )prω̄ (∞; t)[td ]gg |0g i + c.c. (3.35)

(compare eqn (3.28)). The second term in eqn (3.34) requires the probe to act before
the pump and will be neglected here. The third term generates the impulsive Raman
signal contribution,

f˜r (ω̄ − Λr ) (eg) 2 (eg) 2 (eg)


∆Eω̄ISRS = −~ω̄ Fu Fr h0g |[−td ]gg p(ge)
r (t; τ̄¯)prω̄ (∞; t)
σr
× [td ]gg p(ge) (eg)
u (t + td ; τ )pu (τ ; τ̄ )|0g i + c.c. (3.36)

The fourth term produces the stimulated-emission signal,

f˜r (ω̄ − Λr ) (eg) 2 (eg) 2 (eg)


∆Eω̄SE = −~ω̄ Fu Fr h0g |p(ge) ¯
u (t + td ; τ̄ )[−td ]ee pr ω̄ (∞; t)
σr
× p(ge)
r (t; τ )[td ]ee p(eg)
u (τ + td ; τ̄ )|0g i + c.c. (3.37)

The minus in front of these signals corresponds to the fact that in each case the pump
pulse tends to decrease probe absorption. In both GSB and ISRS contributions, wave-
packet evolution during the interpulse delay occurs in the electronic ground state,
whereas in SE it is in the excited state.
Adopting a simplification that affects the calculated signal only at delay times as
short as the pulse durations, we neglect the truncation of pump action by the probe.
In the stimulated-emission contribution, for instance, we make the replacement,

p(ge) (t; τ )[td ]ee p(eg) ∼ (ge) (t; τ )[td ]ee p(eg) (∞; τ̄ ) .
r u (τ + td ; τ̄ ) = pr u (3.38)

This approximation ceases to matter once td exceeds the pulse durations and the ma-
(eg)
trix elements of pu take constant values (see below). Detailed attention must be paid
to the implementation of this approximation in the case of a pump-pulse propagator
whose upper limit involves the integration variable of a reduced propagator for the
spectrally filtered probe pulse. For example, the impulsive stimulated Raman signal of
(ge) (eg) (ge)
eqn (3.36) includes the combination pr (t; τ̄¯)prω̄ (∞; t)[td ]gg pu (t + td ; τ ), in which
4 We postpone to Chapter 5 the calculation and interpretation of two-color transient-absorption
signals from a similar model system in which a pump-pulse resonant with g-to-e is accompanied by
a probe-pulse resonant with e-to-f.
28 Transient-absorption spectroscopy

the pump is truncated by the integration variable of the (temporally elongated) fil-
(ge)
tered probe. The left-most term, pr (t; τ̄¯), only takes non-negligible values once t
approaches or exceeds td ; and we are seeking an accurate treatment only for td signif-
icantly longer than the pulse durations. In this delay range, t + td in the argument of
(ge)
pu is always larger than σu , and we can write,

p(ge)
(eg)
(t;τ̄¯)prω̄ (∞;t)[td ]gg p(ge) ∼ (ge) (t;τ̄¯)prω̄ (∞;t)[td ]gg p(ge) (∞;τ ) . (3.39)
(eg)
r u (t+td ;τ ) = pr u

(ge) (eg)
Show that the same reasoning can be applied to pu (t + td ; τ̄¯)[−td ]ee prω̄ (∞; t)
(ge)
× pr (t; τ ) appearing in eqn (3.37) for the stimulated-emission signal.
Under these approximations, the three relevant contributions (3.35)-(3.37) to the one-
color transient-absorption signal take the simpler forms

f˜r (ω̄ − Λr ) (eg) 2 (eg) 2


∆Eω̄GSB = −~ω̄ Fu Fr h0g |p(ge) ¯ (eg)
u (τ̄ ; τ̄ )pu (∞; τ̄ )[−td ]gg
σr
(eg)
× p(ge)
r (t; τ )prω̄ (∞; t)[td ]gg |0g i + c.c. , (3.40)

f˜r (ω̄ − Λr ) (eg) 2 (eg) 2 (eg)


∆Eω̄ISRS = −~ω̄ Fu Fr h0g |[−td ]gg p(ge)
r (t; τ̄¯)prω̄ (∞; t)
σr
× [td ]gg p(ge) (eg)
u (∞; τ )pu (τ ; τ̄ )|0g i + c.c. , (3.41)

and
f˜r (ω̄ − Λr ) (eg) 2 (eg) 2 (eg)
∆Eω̄SE = −~ω̄ Fu Fr h0g |p(ge) ¯
u (∞; τ̄ )[−td ]ee pr ω̄ (∞; t)
σr
× p(ge)
r (t; τ )[td ]ee p(eg)
u (∞; τ̄ )|0g i + c.c. , (3.42)

respectively.
With a choice of Gaussian pulses having envelopes fj (t) = exp{−t2 /2σj2 } for j =
u or r, the relevant first-order reduced pulse propagators have forms like
Z t
(eg) dτ  τ2 iτ 
hne |pj (t; τ )|0g i = hne |0g i exp − 2 + e + ne − 0g − ~Λj
σ
−∞ j 2σ j ~
q
π −s2 /2
  1 t 
= hne |0g i 2 e 1 + erf √2 σj − is , (3.43)
σ
where s = ~j (e + ne − 0g − ~Λj ) (i.e., the dimensionless offset of ne ← 0g from
Rx 2
resonance) and erf(x) = √2π 0 dy e−y . This quantity is plotted in Fig. 3.6 with respect
to its arguments s and t. Notice that the real part of the reduced propagator begins
to take non-negligible values as the light pulse turns on and remains non-negligible
thereafter, whereas the imaginary part is substantial only during the short episode of
pulse action. Both portions become small for transitions displaced from the carrier
frequency by more than ∼ 2π/σj .
Exemplary calculations 29

(eg)
Fig. 3.6 Real part (left) and imaginary part (right) of pj (t; τ ) divided by the Franck—
Condon overlap hne |0g i, as functions of s = (e + ne − 0g − ~Λj )σj /~ and t/σj .

The ground-state bleach signal of eqn (3.40) and the impulsive Raman signal of eqn
(ge) (eg)
(3.41) both call for the nested double pump-pulse propagator, pu (∞; τ )pu (τ ; τ̄ ).
Nuclear matrix elementsPof this operator are easily obtained by inserting a com-
pleteness relation 1 =
√ ne |ne ihn
√e
| and changing the variables of integration to
x = (τ + τ̄ )/ 2 and y = (τ − τ̄ )/ 2. This procedure leads to

2
σu 2 X 2
− 4~2 (ng −0g )
hng |p(ge) (eg)
u (∞; τ )pu (τ ; τ̄ )|0g i = πe hng |ne ihne |0g ie−ξ [1 + erf(iξ)],
ne
(3.44)
with ξ = [~Λu + (ng +0g )/2 − e − ne ] σu /~. The summand in eqn (3.44) is a prod-
uct of a term that depends only on the nuclear eigenenergies and terms that depend
2 2 2
on the form of the corresponding eigenfunctions. The overall factor e−σu (ng −0g ) /4~
enforces a requirement that the initial and final nuclear states lie within the spectral
2
bandwidth of the pump pulse. The factor e−ξ [1 + erf(iξ)] might appear similarly to
suggest that e + ne − (ng + 0g )/2 must equal ~Λu within ±2π~/σu in order for
a given e-state vibrational level to contribute to the double pulse-propagator; that
this is emphatically not the case is illustrated by Fig. 3.7, which plots the real and
2
imaginary parts of this factor as a function of ξ. While the real part is just e−ξ , the
2 ξ 2
imaginary part is e−ξ √2π 0 dz ez , which is seen to take its values of largest magni-
R

tude at ξ ∼
= ±0.924 and to fall off much more slowly with increasing |ξ| than the real
part.5 This behavior of the second-order pulse propagator accounts for the efficacy of

2 2 2 Rξ 2
5 Forlarge positive ξ, we have √2π 0ξ dz ez −ξ ∼ = √2π e−2ξ 0 dz e2ξz = √1πξ (1 − e−2ξ ), so the
R

imaginary part of the double pulse-propagator falls off inversely as the resonance offset.
30 Transient-absorption spectroscopy

preresonant impulsive stimulated Raman excitation.6,7

1.0

0.5
e - (1+erf[i ])
2

0.0

-0.5

-10 -5 0 5 10


2
Fig. 3.7 Real (solid) and imaginary (dashed) parts of e−ξ [1 + erf(iξ)].

Verify the derivations of (3.43) and (3.44).

Under the current neglect of pump-probe overlap effects, the remaining second-
order pulse propagators are those involving the probe pulse nested inside the spectrally
filtered probe, with matrix elements such as

 ∞
dt sin δω
 t
2 t −iω̄t dτ − 2σ
τ2
(eg) +iΛr τ
ne |prω̄ (∞; t)p(ge)
r (t; τ )|n̄e  = δω
e e r2
ng −∞ σ̄ 2 t −∞ σr

× ne |[−t]ee [t]gg |ng ng |[−τ ]gg [τ ]ee |n̄e  . (3.45)

The outer, t-integration takes the form of a windowed Fourier transformation; it could
easily be carried out numerically in the present calculations. But the sinc-like form
of its prolonged temporal window is just an artifact of our idealized treatment of
spectral filtration.8 It is therefore expedient—and little affects the outcome—to make
the replacement
 δω 2
sin δω
2 t ∼
 
1 δω 3
 
∼ −1
δω 2 t
δω t 2 t − 6 2 t =e 6 ,
2
δω = (3.46)
2 t

6 L. Dhar, J. A. Rogers, and K. A. Nelson, “Time-resolved vibrational spectroscopy in the impulsive


limit,” Chem. Rev. 94, 157–193 (1994); T. J. Smith and J. A. Cina, “Toward preresonant impulsive
Raman preparation of large amplitude vibrational motion,” J. Chem. Phys. 104, 1272–1292 (1996).
2 2 2
the limit of very short σu , both e−σu (ng −0g ) /4 and e−ξ [1 + erf(iξ)] approach unity,
2
7 In

and we have ng |pu (∞; τ )pu (τ ; τ̄ )|0g  ∼
(ge) (eg)
= π ne ng |ne ne |0g  = πng |0g ; vibronic resonance is
ensured regardless of the precise value of Λu , and the second-order action of the pump pulse simply
accounts for partial evacuation of the initially populated nuclear level.
8 Radiative or non-radiative decay of excited states often occur on timescales competitive with
2π/δω, and optical dephasing in multimode systems typically occurs more quickly still. These dy-
namical features diminish the importance of the precise form of the filtered-probe’s temporal envelope.
Exemplary calculations 31

which is correct through second order in t and vanishes for |t| → ∞. With this sub-
stitution, it becomes possible to evaluate eqn (3.45) explicitly, yielding
√ X 6λ2 λ̄2 σr2
 
(eg)
hne |prω̄ (∞; t)p(ge)
r (t; τ )|n̄ e i = 3 hn |n
e g ihn |n̄
g e i exp − −
ng
δω 2 2
i 12λ + λ̄σr2 δω 2
  
× 1 + erf p , (3.47)
δω 24 + 2σr2 δω 2

where ~λ = e + ne − ng − ~ω̄ and ~λ̄ = e + n̄e − ng − ~Λr . The summand in
eqn (3.47) is a product of contributing Franck–Condon overlaps with a factor that
depends on the offset of the corresponding transition frequencies from ω̄ and Λr . We
now have at our disposal representative examples of matrix elements of all the reduced
pulse-propagator combinations needed to calculate transient-absorption signals from
our model system.

1. Derive eqn (3.47) from√ eqns (3.45) √ and (3.46) by making successive variable
changes to (t̄ = tδω/2√ 6,p τ̄ = τ /σr 2), then (U = t̄ sinp θ + τ̄ cos θ, u = t̄ cos θ −
τ̄ sin θ), with cos θ = 2 3/ 12 + σr2 δω 2 and sin θ = σr δω/ 12 + σr2 δω 2 . As a check
on your√derivation (and mine) you may consider the unrealistic situation in which
δω = 2 3/σr and ω̄ = Λr , and compare the result to eqn (3.44).
(ge) (eg)
2. Obtain an approximate formula for hng |prω̄ (∞; t)pr (t; τ )|n̄g i, which is needed
for the GSB and ISRS contributions, by analogy with eqn (3.47).
Now we can take a look at the one-color transient-absorption spectrum of our model
system. The components given by eqns (3.40)–(3.42) were calculated with identical
pump and probe pulses having center frequencies Λu = Λr matching e /~ = 600 Ω and
durations σu = σr = 0.20(2π/ωe ). The dectector slit-width is set to δω = 3.33 Ω, so
that σ̄ = 250 fs remains sufficiently short to resolve motion on the 833-fs timescale of
the 40-cm−1 mode.
The calculated contributions along with the net transient-absorption signal are
shown in Figs 3.8 and 3.9. Each of these mostly negative quantities is plotted as
the hyperbolic tangent of the corresponding component divided by about 87% of its
largest magnitude in order to accentuate the lower-amplitude features.9 Those max-
imum magnitudes take relative proportions 1 : 1.07 : 1.98 : 3.70 for the ground-state
bleach, impulsive stimulated Raman, stimulated emission, and total signals, respec-
tively. Interpulse delay is reckoned in periods of the slow ground-state vibration, and
the exhibited spectral range is nearly twice the full width at half maximum of the
pulses’ power spectrum.
The central stripes in the GSB and ISRS contributions undulate at frequency
Ω, while sideband oscillations occur at the higher frequency ωg . In the stimulated-
emission portion, e-state dynamics at the round-trip time of the slower motion and
the vibrational frequency ωe both become manifest. Because the pump pulse is much
shorter than 2π/Ω, impulsive Raman excitation of the slower mode becomes inefficient,
9 The plotted quantities are independent of E , E , and the transition dipole moment m , so we
u r eg
need not specify values for these quantities.
32 Transient-absorption spectroscopy

Fig. 3.8 Ground-state bleach (left) and impulsive stimulated Raman (right) components
of the transient-absorption signal from the two-dimensional model system with degenerate,
electronically resonant pump and probe pulses. Hyperbolic-tangent scaling is used here and
in the next figure to spread the most closely spaced contours.

Fig. 3.9 Stimulated emission (left) and total signal (right) from the 2D model.

and the low-frequency features in GSB and ISRS are less prominent than the Franck–
Condon displacement-driven slow motion in the SE signal. The blue sidebands in GSB
and the red ones in ISRS are seen to be out of phase with each other. While sideband
structure is present on both blue and red sides of the SE spectrum, the latter appears
more complicated. The two sides of the SE signal more obviously exhibit anti-phasing
Other documents randomly have
different content
l'amusait et l'intéressait; elle vit jouer Andromaque, qui lui fit
répandre plus de six larmes; le Médecin malgré lui l'a fait pâmer de
rire, le Tartuffe l'intéressa [80]. Et tout cela ne l'empêche nullement de
remplir exactement ses devoirs de religion, et de demander à sa fille
toutes les fois qu'elle communie [81].
Les affaires, les divertissements et les festins ne faisaient pas oublier
les jeux d'esprit, passés en habitude dans la haute société de cette
époque. «Lavardin et des Chapelles ont rempli des bouts-rimés que
je leur ai donnés; ils sont jolis, je vous les enverrai [82].»
Madame de Sévigné, entraînée elle-même par la nécessité de
paraître aux états d'une manière conforme à son rang et à la
réception qu'on lui faisait, se pare d'un luxe qu'elle ne pouvait avoir
à la cour et à Paris, mais qui dans sa province était convenable et de
bon goût. Ainsi, quand elle rendait des visites dans ses environs, ou
quand elle allait à Vitré, elle faisait atteler six chevaux à sa voiture;
et elle témoigne naïvement à sa fille que son bel attelage et la
rapidité de ses chevaux lui plaisent beaucoup [83].
Pendant le temps que durèrent les assises des états, elle se rendait
à Vitré le moins souvent qu'elle pouvait, et préférait se tenir à la
campagne; mais elle n'était pas toujours maîtresse de suivre en cela
sa volonté. D'ailleurs on ne la laissait jamais jouir en paix de ses
champs et de ses bois; et la dépense que lui occasionnaient les
visiteurs était pour elle un motif puissant pour céder aux instances
qui lui étaient faites de sortir des Rochers.
Elle écrit de Vitré, le 12 août, à madame de Grignan [84]:
«Enfin, ma chère fille, me voilà en pleins états; sans cela, les états
seraient en pleins Rochers. Dimanche dernier, aussitôt que j'eus
cacheté mes lettres, je vis entrer quatre carrosses à six chevaux
dans ma cour, avec cinquante gardes à cheval, plusieurs chevaux de
main et plusieurs pages à cheval: c'étaient M. de Chaulnes, M. de
Lavardin [85], MM. de Coëtlogon [86], de Locmaria, le baron de Guais,
les évêques de Rennes, de Saint-Malo, les messieurs d'Argouges [87],
et huit ou dix autres que je ne connais point; j'oublie M. d'Harouïs,
qui ne vaut pas la peine d'être nommé. Je reçois tout cela. On dit et
on répondit beaucoup de choses. Enfin, après une promenade dont
ils furent fort contents, une collation, très-bonne et très-galante,
sortit d'un des bouts du mail, et surtout du vin de Bourgogne, qui
passa comme de l'eau de Forges: on fut persuadé que cela s'était
fait avec un coup de baguette. M. de Chaulnes me pria instamment
d'aller à Vitré. J'y vins donc lundi au soir.»
Quatre jours après, elle écrit de nouveau de Vitré [88]: «Je suis
encore ici; M. et madame de Chaulnes font de leur mieux pour m'y
retenir; ce sont sans cesse des distinctions peut-être peu sensibles
pour nous, mais qui me font admirer la bonté des dames de ce pays-
ci; je ne m'en accommoderais pas comme elles, avec toute ma
civilité et ma douceur. Vous croyez bien aussi que sans cela je ne
demeurerais pas à Vitré, où je n'ai que faire. Les comédiens nous
ont amusés, les passe-pieds nous ont divertis, la promenade nous a
tenu lieu des Rochers. Nous fîmes hier de grandes dévotions... Je
meurs d'envie d'être dans mon mail. La Mousse et Marphise ont
grand besoin de ma présence.»
Les lettres que madame de Sévigné recevait de sa fille lui
apprenaient que la Provence ne se montrait pas aussi facile que la
Bretagne. «Vous me ferez aimer, lui dit-elle, l'amusement de nos
Bretons plutôt que l'indolence parfumée de vos Provençaux [89]»; et
elle mande à sa fille que M. d'Harouïs souhaite que les états de
Provence donnent à madame de Grignan autant que ceux de
Bretagne ont donné à madame de Chaulnes [90]. En effet, les états de
Bretagne firent à la duchesse de Chaulnes présent de deux mille
louis d'or, qui lui furent envoyés par une députation composée de
dix-huit membres, à la tête desquels étaient les évêques de Quimper
et de Nantes, chargés de la complimenter [91].
Madame de Sévigné parle de ces dons avec un ton ironique qui
décèle sa pensée: «On a donné cent mille écus de gratifications,
deux mille pistoles à M. de Lavardin, autant à M. de Molac, à M.
Boucherat, au premier président, au lieutenant du roi; deux mille
écus au comte des Chapelles, autant au petit Coëtlogon; enfin des
magnificences. Voilà une province [92]!» Oui; mais la Bretagne, mal
défendue par ses députés contre les exactions du pouvoir, se révolta
quatre ans après; et la Provence, sous la bénigne administration du
comte de Grignan, qui se ruina en la gouvernant, fut heureuse et
tranquille.
Madame de Sévigné est exacte pour les sommes données à
Lavardin, premier lieutenant général, pour des Chapelles et
Coëtlogon; mais elle se trompe pour M. de Molac, second lieutenant
général, qui n'eut que 25,000 liv. Le marquis de Lavardin eut, en
outre des 25,000 liv., 16,000 liv. pour ses gardes et officiers; le duc
de Chaulnes, gouverneur, eut 100,000 liv., et 20,000 liv. pour ses
gardes et officiers; le duc de Rohan eut 22,000 liv.; l'évêque de
Rennes eut la même somme, et le premier président 20,000 liv. De
Colbert, intendant de Bretagne, reçut 9,000 liv.; le marquis de
Louvois, grand maître et surintendant des forêts, 8,000 liv., et tous
les autres à proportion [93].
En accordant tout ce qui leur était demandé, les états firent des
remontrances tendant à faire révoquer plusieurs édits nuisibles à la
province; mais les réponses furent faites aux états tenus deux ans
après, en 1673: elles prouvent que ces remontrances furent
illusoires. Cependant quelques-unes sont des espèces de
protestations contre certaines dispositions des édits royaux, qu'on
affirme être contraires aux coutumes de la province. Pour toutes les
demandes de cette nature, le roi promet de se faire informer de ces
coutumes: il semble ainsi reconnaître qu'il veut les respecter [94].
Les assises des états furent terminées le 5 septembre. Madame de
Sévigné, en annonçant à sa fille cette fin dans sa lettre datée de
Vitré le lendemain, s'exprime ainsi [95]: «Les états finirent à minuit;
j'y fus avec madame de Chaulnes et d'autres femmes. C'est une
très-belle, très-grande et très-magnifique assemblée. M. de Chaulnes
a parlé à tutti quanti avec beaucoup de dignité, et en termes fort
convenables à ce qu'il avait à dire. Après dîner, chacun s'en va de
son côté. Je serai ravie de retrouver mes Rochers. J'ai fait plaisir à
plusieurs personnes; j'ai fait un député, un pensionnaire; j'ai parlé
pour des misérables, et de Caron pas un mot [96], c'est-à-dire, rien
pour moi; car je ne sais point demander sans raison.»
On voit que madame de Sévigné désapprouvait les prodigalités des
états; mais son texte, pour ce qui la concerne, a besoin d'une
explication, qui n'a jamais été donnée.
La terre de Sévigné [97] avait été démembrée, ou avait depuis
longtemps cessé d'être la principale possession de la famille de ce
nom [98]. Cette famille possédait la seigneurie des Rochers depuis le
milieu du XVe siècle, par le mariage d'Anne de Mathefelon, fille et
héritière de Guillaume de Mathefelon, seigneur des Rochers, avec
Guillaume de Sévigné. Mais il restait à la famille de Sévigné de ses
anciennes possessions une terre de Sévigné près de Rennes, dans la
commune de Gevezé, consistant en deux métairies, en moulins et
quelques fiefs, dont la valeur totale est estimée par le fils de
madame de Sévigné à 18,000 livres (36,000 fr.), tandis qu'il porte le
prix de la terre des Rochers à 120,000 liv. (240,000 fr.) [99]. Parmi les
fiefs restés à la famille de Sévigné, était, dans la ville de Vitré, une
maison avec cour et jardin, qu'on appelait la Tour de Sévigné. Cette
maison était un fief qui relevait du duc de la Trémouille, baron de
Vitré [100]. Par acte passé le 2 septembre 1671 (trois jours avant la fin
des états), madame de Sévigné fit une rente de cent francs aux
bénédictins de Vitré, et hypothéqua cette rente ou pension sur la
Tour de Sévigné [101].
Ce don fut sans doute fait en reconnaissance des réparations
exécutées aux frais de la province à la grosse tour qui donnait son
nom à la maison de Vitré. Voilà pourquoi elle dit, «J'ai fait un
pensionnaire,» et qu'en même temps elle avance qu'elle n'a rien
demandé, parce que la demande qu'elle avait formée ne pouvait
souffrir aucune difficulté, puisque cette grosse tour était engagée
dans les fortifications de la ville, et en faisait partie. M. le duc de
Chaulnes, qui voulait faire venir à Vitré madame de Sévigné, prit ce
prétexte pour la forcer à quitter son château des Rochers: il fit la
plaisanterie de l'envoyer chercher par ses gardes, en lui écrivant
qu'elle était nécessaire à Vitré pour le service du roi, attendu qu'il
fallait qu'elle donnât des explications sur la demande qu'elle faisait
aux états; et qu'en conséquence madame de Chaulnes l'attendait à
souper [102].
C'est dans cet hôtel de la Tour de Sévigné que demeurait la brillante
marquise lorsqu'elle restait à Vitré. Cette année, elle en laissa la
jouissance à son fils, qui y donnait à souper à ses amis [103]. C'est
aussi dans cette maison qu'allèrent loger, lorsqu'ils arrivèrent à Vitré
pour la tenue des états, de Chesières, l'oncle de madame de
Sévigné, son parent d'Harouïs, et un député nommé de Fourche [104].
Lorsqu'elle y restait, elle était accablée de visites. «Hier, dit-elle, je
reçus toute la Bretagne à ma Tour de Sévigné [105].» Mais lorsque les
états furent terminés, que le duc de Chaulnes fut parti, elle n'alla
plus à Vitré. Son fils l'avait quittée depuis longtemps, et bien avant la
fin des états, où son âge ne lui permettait pas d'être admis. Quoique
l'été fût constamment froid et pluvieux [106], madame de Sévigné
resta aux Rochers, pour que l'abbé de Coulanges pût surveiller les
travaux de la chapelle [107], et pour avoir le temps de terminer les
embellissements de son parc [108]. Elle avait envie d'aller visiter une
autre terre qu'elle possédait en Bretagne, près de Nantes, nommée
le Buron; mais, dit-elle, «notre abbé ne peut quitter sa chapelle; le
désert du Buron et l'ennui de Nantes ne conviennent guère à son
humeur agissante [109].» Madame de Sévigné se soumet, et ne va pas
au Buron.
Cette terre, à quatre lieues de Nantes, avait un château ancien, mais
bien bâti [110]. Le marquis de Sévigné en fit abattre les arbres
séculaires qui en faisaient tout l'agrément [111], et ce beau domaine
fut ensuite dégradé et ruiné par un administrateur infidèle ou
inintelligent, et par un fermier de mauvaise foi [112].
Il n'en fut pas ainsi des Rochers, que madame de Sévigné ne cessa
jamais d'accroître et d'embellir, et qu'elle vint si souvent habiter [113].
La construction de la chapelle, de forme octogone, surmontée d'une
coupole, située au bout du château et isolée, fut achevée en cette
année 1671; mais ce ne fut qu'après quatre ans que l'intérieur fut
entièrement en état, et qu'on put enfin y célébrer la messe, pour la
première fois, le 15 décembre 1675. A cette époque si froide de
l'année, madame de Sévigné se promenait avec plaisir dans ses bois,
plus verts que ceux de Livry, et augmentés de six allées charmantes,
que madame de Grignan ne connaissait point [114]. Depuis, ce
nombre d'allées fut presque doublé [115].
Madame de Sévigné avait multiplié dans son parc les inscriptions
morales, religieuses et autres, presque toujours tirées de l'italien.
Sur deux arbres voisins elle avait inscrit deux maximes contraires:
sur l'un, La lontananza ogni gran piaga salda (L'absence guérit les
plus fortes blessures); sur l'autre, Piaga d'amor non si sana mai
(Blessure d'amour jamais ne se guérit). Une des plus heureuses
inscriptions fut sans doute ce vers du Pastor fido, qu'elle avait fait
graver au-dessus d'une petite fabrique placée au bout de l'allée de
l'Infini, afin de se garantir de la pluie:
Di nembi il cielo s'oscura indarno [116].
Une autre allée, nommée la Solitaire, longue de douze cents pas, fut
plantée plus tard, et madame de Sévigné s'en enorgueillit comme de
la plus belle [117]. Elle avait fait construire dans différents endroits du
parc un assez grand nombre de petites cabanes qu'elle appelle des
brandebourgs [118], pour lire, causer et écrire à son aise, à l'abri du
soleil, du serein, et surtout de la pluie. Quant à son mail, dont elle
parle si souvent, c'est pour elle une belle et grande galerie, au bout
de laquelle on trouvait la place Madame, d'où, comme d'un grand
belvéder, la campagne s'étendait à trois lieues, vers une forêt de M.
de la Trémouille (la forêt du Pertre). Elle n'est pas moins engouée de
son labyrinthe, que son fils aimait par-dessus tout, et où nous
apprenons qu'il se retirait souvent avec sa mère pour lire ensemble
l'Histoire des variations de l'Église protestante, de Bossuet [119]. Mais
ce fut seulement vingt-sept ans après avoir été commencé, vers la
fin de l'année 1695, que madame de Sévigné, alors à Grignan, apprit
de son fils, qui était aux Rochers, que Pilois avait enfin terminé le
labyrinthe. Ainsi, les Rochers furent pour madame de Sévigné,
comme ses lettres, l'occupation de toute sa vie [120].
De son antique manoir, des constructions qu'elle avait ajoutées, des
ombrages qu'elle avait formés, il ne reste plus rien que la
chapelle [121], où le Christ est toujours invoqué, et l'écho de la place
de Coulanges, qui répète encore le nom de madame de Sévigné [122].
CHAPITRE II.
1671.

Bohémienne qui ressemble à madame de Grignan.—Ce que madame de


Sévigné fait pour elle.—Portrait de madame de Grignan en bohémienne.—
Madame de Grignan accouche d'un fils.—Il est tenu sur les fonts de
baptême par la Provence.—M. et madame de Grignan vont habiter le
château de Grignan.—Description de ce château.—Des personnes, parents
et amis de M. et de madame de Grignan, qui fréquentaient ce château.—
De la comtesse d'Harcourt.—Seigneur Corbeau.—L'archevêque d'Arles.—
L'évêque d'Uzès.—Le bel abbé.—Le chevalier Adhémar.—Le grand
chevalier.—Claire d'Angennes, fille aînée de madame de Grignan, se retire
au couvent, et fait don de son bien à son père.—Mademoiselle d'Alérac, sa
fille cadette, se marie.—Des sœurs de M. de Grignan.—La religieuse
d'Aubenas.—La marquise de Saint-Andiol.—La comtesse de Rochebonne.
—Du chevalier comte de la Garde, parent de M. de Grignan.—Madame de
Sévigné prête au comte de la Garde le portrait de sa fille.—De madame
du Puy du Fou.—Du personnel de la maison de madame de Grignan.—
Mademoiselle Deville, la femme de chambre.—Mademoiselle de
Montgobert, demoiselle de compagnie.—Ripert, intendant.—Madame de
Grignan faisait la mode en Provence.—Ses nombreuses réunions et son
luxe à Aix.—Se retirait quelquefois au couvent des Filles de Sainte-Marie.
—N'avait pas le même goût que sa mère pour la solitude et la campagne.
—Aime à primer.—Le maréchal de Bellefonds veut céder sa place de
premier maître d'hôtel du roi.—Le comte de Grignan se dispose à
l'acheter.—Madame de Grignan s'y oppose.—Plaintes de madame de
Sévigné à ce sujet.

Les constructions, les plantations dont s'occupait madame de


Sévigné, ne pouvaient calmer les inquiétudes toujours croissantes
que lui faisaient éprouver les approches du terme de la grossesse de
sa fille, encore moins diminuer la peine quelle ressentait de s'en être
séparée. Le tumulte des états, les grandes réunions, les visites
reçues et rendues, les festins, les spectacles, la musique, les danses,
avaient encore moins de pouvoir [123]. Le plus souvent ces moyens de
distraction produisaient un effet contraire. Dans une des fêtes
données à Vitré pour l'amusement de la société qui s'y trouvait
rassemblée, on fit danser une troupe de bohémiens. Ils dégoûtèrent
d'abord madame de Sévigné par leur saleté [124]. Mais dans le
nombre des femmes qui faisaient partie de cette troupe, elle en vit
une plus proprement et plus élégamment vêtue. Cette fille la frappa
par sa ressemblance avec madame de Grignan. Les beaux yeux, les
belles dents, l'élégance de la taille de la bayadère, et surtout la
grâce avec laquelle elle dansait, rappelaient mademoiselle de
Sévigné dans les ballets du roi. La pauvre mère en fut émue; elle fit
approcher la jeune fille, la traita avec amitié; et celle-ci, encouragée
par cet accueil, pria sa nouvelle protectrice de vouloir bien écrire en
Provence pour son grand-père.—«Et où est votre grand-père?» lui
demanda madame de Sévigné.—«Il est à Marseille, madame»,
répondit d'un ton doux et triste la bohémienne.—Madame de
Sévigné devina; elle promit d'écrire, et écrivit en effet à M. de
Vivonne, général des galères, en faveur du galérien grand-père de la
bohémienne.—Ah! madame de Grignan! cette lettre si touchante, si
joviale, vous fut envoyée; elle fut soumise à votre censure; c'est
vous qui fûtes chargée de la remettre au gros crevé: pourquoi n'en
avez-vous pas conservé de copie? Pourquoi ne pouvons-nous la lire
comme toutes celles qui vous furent écrites, et connaître les
résultats de votre ressemblance avec la petite-fille du forçat,
«capitaine bohème d'un mérite singulier [125]?»—Ces résultats furent
heureux: non-seulement madame de Grignan remit la lettre, mais
elle intercéda pour le vieux forçat, mais elle parvint à briser ses fers,
mais elle fit un sort à cette bohémienne, assez belle danseuse pour
qu'elle fût elle-même glorieuse de lui ressembler.—Aucune tradition
ne nous apprend cela; cela n'a pas été dit, cela n'est écrit nulle part:
mais pouvons-nous en douter, lorsque nous apprenons, d'après un
ancien inventaire du château de Grignan, «que l'appartement
qu'occupait madame de Sévigné, quand elle était dans ce château,
se composait de deux pièces; que l'une se nommait chambre de la
Tour, et l'autre chambre de la Bohémienne, parce qu'au-dessus du
chambranle de la cheminée était un portrait de madame de Grignan,
costumée en bohémienne [126]?»
Madame de Grignan avait offert à sa mère des consolations un peu
subtiles aux tourments de l'absence; comme de se promener en
imagination dans son cœur, où elle trouverait mille tendresses.
Madame de Sévigné répond: «Je fais quelquefois cette promenade;
je la trouve belle et agréable pour moi.... Mais, mon Dieu, cela ne
fait point le bonheur de la vie; il y a de certaines grossièretés solides
dont on ne peut se passer [127].»
Cependant le motif de ses craintes et de ses inquiétudes disparut;
elle fut enfin délivrée du gros caillou qu'elle avait sur le cœur [128].
Elle se préparait à quitter les Rochers et à retourner à Paris, quand
elle apprit l'heureuse nouvelle que madame de Grignan était
accouchée d'un fils, blond comme sa mère, et qu'elle avait donné à
M. de Grignan, qui n'avait eu jusqu'ici que des filles de toutes ses
femmes, un héritier. Madame de Sévigné avait prédit à madame de
Grignan que cette fois elle aurait un fils; et l'on peut juger de ce
qu'elle ressentit en apprenant que ses prédictions et ses espérances
s'étaient réalisées [129], que tous ses conseils maternels avaient eu un
plein succès [130]. «Que pensez-vous, dit-elle, qu'on fasse dans ces
excès de joie? Le cœur se serre, et l'on pleure sans pouvoir s'en
empêcher. C'est ce que j'ai fait, ma très-belle, avec beaucoup de
plaisir: ce sont des larmes d'une douceur qu'on ne peut comparer à
rien, pas même aux joies les plus brillantes [131].» Elle fut pourtant
très-flattée d'apprendre que son petit-fils avait été baptisé par la
Provence. En effet, les états étaient encore assemblés à Lambesc
lorsque madame de Grignan y accoucha le 17 novembre. Le
lendemain, le comte de Grignan se rendit dans l'assemblée, et «vint
offrir, dit le procès-verbal de cette séance, le fils qu'il a plu à Dieu de
lui donner dès le jour d'hier, et de vouloir bien lui faire la faveur de le
tenir au nom de toute la province sur les fonts du baptême, et de lui
donner tel nom qu'il lui plaira.... Sur quoi l'assemblée a délibéré que
messieurs les procureurs généraux du pays témoigneront à
monseigneur le comte de Grignan et à madame sa femme la joie de
toute la province, et particulièrement de l'assemblée, sur la
naissance de ce premier mâle dans sa famille, et lui feront de très-
humbles remercîments de l'honneur qu'il avait fait à la province, de
le faire tenir de sa part pour recevoir les saintes eaux du baptême,
avec tous les sentiments d'amour et de reconnaissance possibles. Et
l'assemblée a délibéré que les frais en seront supportés par le pays,
suivant le rôle qui en sera tenu par le sieur Pontèves, trésorier des
états [132].»
Ainsi naquit et fut nommé Louis de Provence, marquis de Grignan,
dont Saint-Simon, son ami de collége, déplore la perte prématurée,
loue la brillante valeur et l'excellent caractère [133].
Madame de Grignan accoucha facilement, et aussitôt après la fin des
états elle alla avec son mari habiter le château de Grignan, qu'elle
avait quitté pour se rendre à Lambesc. Ce séjour lui était favorable
pour le rétablissement de sa santé. La petite ville de Grignan,
aujourd'hui chef-lieu de canton, à quatorze kilomètres de
Montélimar, se penche sur le revers méridional d'un coteau escarpé:
ornement assez beau d'un bassin arrosé par les petites rivières de
Berre et de Lez, couvert cependant, sur plusieurs points, de rochers
stériles. Les maisons de la ville sont mal bâties; mais l'église se fait
remarquer par un air de magnificence et par ses arceaux gothiques,
témoignages de l'antiquité de sa construction [134]. Au-dessus de
cette église, et de niveau avec son faîte, est un plateau qui domine
toute la ville, et dont la vue s'étend sur le pays d'alentour. C'est sur
ce plateau que s'élevait le château de Grignan. Isolé de toutes parts,
ce noble et grand édifice semblait suspendu dans l'air, comme le
palais magique construit par l'enchanteur Appollidon [135], auquel
madame de Sévigné le compare: sa position, ses murs élevés, ses
tourelles, le faisaient ressembler à un ancien château fort; car, à
l'époque dont nous nous occupons, la façade moderne, construite et
jamais achevée [136] aux frais d'un des beaux-frères de madame de
Grignan, l'évêque de Carcassonne, n'existait pas encore. Ce château,
le plus beau de toute la province, manquait d'ombrage; le territoire
qui l'entoure est en général maigre et sablonneux. Les vents du nord
y sont impétueux et fréquents, et y détruisent presque annuellement
la majeure partie des récoltes [137]; et jusque dans ces derniers
temps, à cause du mauvais état des routes, il était d'un accès
difficile.
Cependant, ce séjour convenait mieux à l'indolence naturelle et aux
susceptibilités de madame de Grignan que celui d'Aix ou de
Lambesc. La nécessité d'ouvrir son salon à toutes les notabilités, les
visites à rendre et à souffrir, les exigences cérémonieuses des dames
de Provence, lui étaient insupportables [138]. Elle était moins exposée
à ce genre d'ennui dans son château; mais elle ne pouvait s'y
dérober entièrement. Dans sa position surtout, il ne lui était pas
facile de rompre ces amitiés du monde, dont «la dissimulation est le
lien, et l'intérêt le fondement [139].» Encore moins pouvait-elle se
soustraire aux devoirs de parenté. Ainsi, il lui fallait recevoir
fréquemment, et avec toutes les démonstrations d'une satisfaction
sincère, cette comtesse d'Harcourt, née Ornano, tante du comte de
Grignan, mère du prince d'Harcourt et de cette demoiselle d'Harcourt
qui fut mariée au prince de Cadaval au commencement de l'année
1671, et dont les noces, honorées de la présence du roi et de la
reine, donnèrent lieu à cette belle fête à laquelle assista madame de
Sévigné [140]. Il faut se garder de confondre cette comtesse
d'Harcourt avec la princesse d'Harcourt, fille de Brancas le distrait,
liée aussi avec madame de Grignan, qui lui trouvait peu d'esprit.
Cette princesse d'Harcourt, dont nous avons déjà parlé [141], fut
nommée dame du palais, et chargée avec son mari, Henri de
Lorraine, prince d'Harcourt, de conduire, en 1679, la reine d'Espagne
à son époux. Le prince d'Harcourt était cousin germain de M. de
Grignan [142]. La comtesse d'Harcourt, sa tante, habitait le Pont-Saint-
Esprit, et se trouvait ainsi peu éloignée du château de Grignan, où
elle allait fréquemment rendre visite. Madame de Sévigné plaint
souvent sa fille d'être obligée de supporter un tel fardeau; elle
souhaite d'être à Grignan, pour la débarrasser de cette vieille tante.
«Après cette marque d'amitié, ajoute-t-elle, ne m'en demandez pas
davantage, car je hais l'ennui à la mort: vous seule au monde seriez
capable de me faire avaler ce poison, et j'aimerais fort à rire avec
vous, Vardes, et le seigneur Corbeau [143].» C'est par ce dernier nom
que, à cause de son teint basané, madame de Grignan appelait son
beau-frère l'évêque de Claudiopolis, coadjuteur de l'archevêque
d'Arles son oncle [144]. Il était alors à Grignan, où il avait passé l'été,
tout le temps de la tenue des états: homme du monde, aimable
auprès des femmes, souvent à Paris et à la cour [145], prudent,
spirituel, fort attaché à son frère, et zélé pour la gloire de la maison
de Grignan, il fut très-goûté de madame de Sévigné et de sa fille. Il
ne pouvait souffrir qu'elles lui donnassent du monseigneur:
«Appelez-moi plutôt Pierrot ou seigneur Corbeau,» disait-il. Il parlait
et écrivait avec facilité, mais il n'aimait pas à écrire; et madame de
Sévigné lui défendait toujours de répondre à ses lettres, par la
crainte que, s'il se croyait obligé de le faire, il ne la prit en
déplaisance: elle voulait qu'il réservât sa main droite pour jouer au
brelan. Elle le raillait aussi sur son penchant pour la bonne chère, et
elle attribuait à cela les attaques de goutte qu'il commençait déjà à
ressentir: il n'avait alors que trente-trois ans [146]. Par la suite il excita
l'indignation de madame de Sévigné, à cause de son ingratitude
envers son oncle l'archevêque d'Arles, auquel il devait succéder [147].
Des deux oncles paternels du comte de Grignan, le plus élevé en
dignité, l'archevêque d'Arles, était un homme excellent, et aimé de
toute la Provence [148]. Il avait, pendant les troubles de la Fronde,
apaisé les émeutes populaires à Arles et à Marseille, et empêché que
ces deux villes ne se révoltassent contre le gouvernement. En 1660,
lors du voyage de la cour en Provence, Louis XIV logea chez lui; et
ce fut alors qu'il le nomma commandeur de ses ordres [149].
Bienfaiteur de sa famille, l'archevêque d'Arles en était tendrement
chéri; mais cependant il augmentait les embarras du gouverneur,
parce qu'il était toujours opposé à l'évêque de Marseille pour les
affaires ecclésiastiques, comme le comte de Grignan l'était pour les
affaires civiles [150]; ce qui contribuait à accroître l'animosité de ce
prélat hautain, mais habile, qui avait acquis un grand ascendant sur
l'assemblée des états, une place élevée dans l'estime des ministres.
Madame de Sévigné parvint à l'adoucir et à le rendre moins hostile,
et elle neutralisa les effets de son influence contre les Grignan par le
moyen de ses amis, de Pomponne et de le Camus, premier président
de la cour des aides [151]. Pour toutes ces négociations elle se servait
utilement de l'autre oncle de M. de Grignan, évêque et comte
d'Uzès [152], homme sage et prudent; plus souvent à la cour et dans
son abbaye d'Angers que dans son diocèse; plein d'affection pour
madame de Grignan, et très-zélé pour les intérêts de son frère;
toujours empressé à faire auprès des ministres les démarches que
lui demandait madame de Sévigné. Comme elle, il agissait aussi
directement sur l'évêque de Marseille; et s'il ne parvenait pas à lui
inspirer des sentiments de concorde et d'amitié, il l'empêchait au
moins de se montrer adversaire violent [153].
M. de Grignan avait un autre frère dans l'état ecclésiastique, très-
différent de seigneur Corbeau par sa figure, car il était d'une beauté
remarquable [154]: on l'avait surnommé le bel abbé. A l'époque dont
nous traitons, âgé seulement de vingt-huit ans, il n'avait pas encore
soutenu sa thèse en Sorbonne. Doué de capacité et ambitieux, il fut
successivement agent général du clergé, abbé de Saint-Hilaire,
nommé évêque d'Évreux, mais non confirmé comme tel [155]. Il fut
sacré évêque de Carcassonne dans l'église de Grignan. Son faste et
sa prodigalité contrariaient madame de Sévigné, qui aurait voulu
qu'une partie de ses riches revenus ecclésiastiques fussent employés
à faire du bien à ses frères [156], et particulièrement au moins riche
de tous, le chevalier de Grignan, Adhémar.
Plein de courage et animé d'une noble ambition, Adhémar [157]
parvint, par de beaux faits d'armes, au grade de maréchal de camp,
lorsque son frère aîné épousa mademoiselle de Sévigné. Quoique
bien jeune encore, il obtint le commandement du régiment qui
portait le nom de Grignan [158]; et, à cette occasion, madame de
Sévigné prit le soin de lui donner une devise: c'était une fusée
poussée à une grande élévation, avec ces mots italiens: Che peri,
purchè s'innalzi [159], «Qu'elle périsse, pourvu qu'elle s'élève.» Le plus
jeune de tous les Grignan, il n'avait point cette morgue de famille qui
faisait dire à M. de Guilleragues que tous les Grignan étaient des
glorieux. Lorsqu'on lui opposait l'exemple du chevalier Adhémar [160],
«Celui-là, disait-il pour ne pas se rétracter, n'est que glorioset.» Ce
singulier sobriquet de petit Glorieux resta au chevalier Adhémar [161].
De tous ses frères, il était le plus attentif et le plus complaisant pour
madame de Grignan; il lui servait de secrétaire lorsque quelque
indisposition l'empêchait de tenir la plume [162]. Ce fut là sans doute
ce qui valut à madame de Grignan les malins vaudevilles et les
épigrammes que l'on composa sur elle [163], moins cependant à
propos d'Adhémar qu'au sujet du frère de celui-ci, nommé, à cause
de sa taille, le grand chevalier. Il se trouvait alors au château de
Grignan, et mourut l'année suivante à Paris, de la petite vérole, chez
son oncle l'évêque d'Uzès [164]. C'est à ce chevalier de Grignan que
madame de Sévigné défendait de monter à cheval en présence de sa
fille [165], tant le souvenir de la fausse couche qu'il avait occasionnée
par sa chute faisait d'impression sur elle. Tels étaient dans la famille
de Grignan les hommes qui se réunissaient au château de Grignan,
et en composaient la société. Les filles que le comte de Grignan
avait eues de son premier mariage avec Angélique-Claire d'Angennes
étaient encore trop jeunes pour y figurer [166]. L'aînée n'avait que dix
ans, et la cadette seulement sept ans, lorsque leur père se remaria
avec mademoiselle de Sévigné [167]. Le duc de Montausier, leur oncle
par alliance, puisqu'il avait épousé Julie d'Angennes, s'opposait à ce
qu'elles allassent demeurer chez leur belle-mère, craignant que
celle-ci ne se prévalût de l'innocence de leur jeune âge, et ne leur
inspirât prématurément de l'inclination pour la vie religieuse:
cependant il finit par céder aux instances de madame de Grignan, et
s'aperçut bientôt qu'il ne s'était pas trompé dans ses prévisions [168].
Louise-Catherine-Adhémar, l'aînée des deux filles de M. de Grignan
et de Claire d'Angennes, excitée par sa belle-mère, ses oncles et
toute sa famille, dans son penchant à la dévotion, voulut entrer aux
Carmélites; mais la délicatesse de sa santé ne lui permit pas de
soutenir les austérités de l'ordre: elle ne put achever son noviciat;
elle se retira comme pensionnaire dans un couvent, et y vécut avec
autant de régularité et de piété que la religieuse cloîtrée la plus
attachée à ses devoirs. Sur le bien de sa mère, il lui revenait
quarante mille écus; elle en fit don à son père; et madame de
Grignan ne déguise pas qu'elle se servit de l'influence qu'elle avait
acquise sur cette jeune fille, pour la déterminer à prendre cette
résolution. Bussy profite de cette occasion pour lancer un sarcasme
piquant, mais juste [169], contre madame de Grignan; et madame de
Sévigné, au contraire, chez qui la tendresse pour sa fille, et sa
continuelle préoccupation pour tout ce qui concernait ses intérêts et
sa grandeur, étouffaient tout autre sentiment, la félicite d'avoir «fait
merveille», et exprime, par les termes les plus énergiques, son
admiration pour Catherine-Adhémar, qu'elle appelle une fille céleste,
par opposition à sa sœur cadette, qui est pour elle la fille
terrestre [170]. En effet, celle-ci, Françoise-Julie, qu'on nommait
ordinairement mademoiselle d'Alérac [171], quoique soumise à la
même éducation et aux mêmes influences que sa sœur, eut des
goûts très-différents: elle aimait le monde, et elle se plaisait
beaucoup dans la société de madame de Sévigné, qui la trouvait
aimable [172]. Jolie et faite pour plaire [173], elle fut recherchée en
mariage par le chevalier de Polignac et M. de Belesbat. Ces deux
mariages se rompirent, non par le fait de madame de Grignan.
Pourtant le défaut d'accord entre la belle-mère et la belle-fille fut tel,
que celle-ci abandonna brusquement la maison paternelle, et se
retira chez son oncle par alliance, le duc de Montausier, et ensuite au
couvent des Feuillantines [174]. Elle se maria enfin avec le marquis de
Vibraye, sans la participation et aussi sans l'opposition de sa
famille [175].
Des trois sœurs qu'avait le comte de Grignan, une seule doit nous
occuper, puisque celle qui se fit religieuse à Aubenas [176], et celle qui
se maria au marquis de Saint-Andiol (en 1661) [177], ne sont
mentionnées que deux ou trois fois dans la correspondance de
madame de Sévigné. Il n'en est pas de même de Thérèse-Adhémar
de Monteil; celle-ci épousa le comte de Rochebonne [178], qui
commanda longtemps à Lyon pour le roi. La comtesse de
Rochebonne ressemblait beaucoup à son frère, le comte de Grignan:
c'est dire assez qu'elle n'était pas belle; aussi est-ce par antiphrase
et en plaisantant que madame de Sévigné la qualifie de jolie
femme [179]. Sa laideur, et la surdité dont elle était affligée, étaient
rachetées par le plus heureux caractère. Elle s'était liée d'amitié avec
madame de Grignan, et l'affection que celle-ci avait pour elle
s'étendait jusqu'à ses enfants. Elle en avait un grand nombre;
presque tous étaient remarquables par leur esprit précoce, leurs
jolies figures, la fraîcheur de leur teint et leurs grâces enfantines [180].
Un des parents du comte de Grignan, que madame de Sévigné
aimait le mieux, était le chevalier comte de la Garde, qui avait été
gouverneur de la ville de Furnes et lieutenant des gardes du corps
de la reine mère [181]. Sa baronnie de la Garde était voisine du comté
de Grignan, et il allait fréquemment au château. Lorsqu'il échoua
dans le projet de mariage qu'il avait conçu, on était presque certain
qu'il resterait célibataire [182]; et comme la forte pension dont il
jouissait le rendait riche, on croyait qu'il avantagerait le comte de
Grignan. Dans cet espoir, madame de Sévigné avait pour lui de
grands égards; il fut la seule personne à laquelle elle permit de faire
copier le portrait de sa fille, peint par Mignard [183]: elle avait refusé
rabutinement, comme elle le dit, cette faveur à ses plus intimes
amis, au bel abbé, l'évêque de Carcassonne, à l'abbesse de
Fontevrault, sœur de madame de Montespan, enfin même à
MADEMOISELLE [184]. Le chevalier de la Garde ne put rien faire pour son
cousin, le comte de Grignan; la riche pension de 18,000 livres dont il
jouissait (36,000 fr.) fut supprimée, et il fut presque entièrement
ruiné [185].
A toutes ces personnes que le mariage de mademoiselle de Sévigné
avec le comte de Grignan avait placées dans des rapports de famille
et d'intimité tant avec elle qu'avec madame de Sévigné, il faut
joindre la marquise du Puy du Fou, mère de la seconde femme du
comte de Grignan [186]. Elle avait peu d'esprit, mais sa bonté la faisait
chérir. Comme elle demeurait à Paris, madame de Sévigné la voyait
souvent, et même la recherchait, à cause de l'attachement qu'elle
avait conservé pour celui qui avait été son gendre, et de l'amitié
qu'elle avait pour madame de Grignan. Madame de Sévigné passait
des heures entières avec madame du Puy du Fou, et lui confiait sa
petite-fille Marie-Blanche, et madame du Puy du Fou en avait soin
comme de son propre enfant [187].
Les Simiane étaient aussi cousins des Grignan [188]; et, parmi les
nouvelles connaissances que son séjour en Provence procura à
madame de Grignan, on remarque la marquise de Simiane, dont le
fils épousa celle à qui nous devons la publication des Lettres.
Madame de Sévigné avait eu occasion de rencontrer dans le monde
madame de Simiane, et elle félicite sa fille d'avoir en elle une
compagnie agréable [189]. Elle fait l'éloge de son amabilité, mais elle
ne lui reconnaît pas une excellente tête; elle la blâme de vouloir se
séparer de son mari, à cause des fréquentes infidélités qu'il lui
faisait, ajoutant assez lestement: «Quelle folie! Je lui aurais conseillé
de faire quitte à quitte avec lui [190].»
La maison de madame de Grignan se composait d'un nombreux
personnel, conforme au rang qu'elle tenait en Provence; et ceux qui
en faisaient partie paraissent avoir été bien choisis pour la soulager
dans les devoirs qu'elle avait à remplir, et la distraire de ce qu'ils
pouvaient avoir de pénible. Deux femmes de chambre étaient
attachées à son service; et l'une d'elles, nommée Deville, fille de son
maître d'hôtel [191], en savait assez pour l'aider, et, au besoin, pour la
suppléer dans ses correspondances. Une demoiselle de Montgobert,
pieuse mais enjouée, d'un esprit original, plaisait beaucoup à
madame de Sévigné [192]; elle était demoiselle de compagnie; et
Ripert [193], l'intendant des Grignan, était un homme d'esprit et d'une
société agréable: il avait sa chambre au château de Grignan, à côté
de celle des deux pages [194].
Madame de Sévigné instruisait avec grand soin madame de Grignan
des variations de la mode. Elle savait que sa fille, par sa beauté et
par son rang, avait en Provence le privilége d'être le patron sur
lequel les femmes se réglaient [195]; et c'est à la cour de Louis XIV
qu'alors la mode avait, pour toute l'Europe, établi le siége de son
empire. Les lettres de madame de Sévigné fourniraient d'exacts et
nombreux détails à celui qui voudrait nous retracer les lois absolues
et les bizarres volontés de cette capricieuse reine du monde élégant.
C'est surtout lorsqu'elle était à Aix [196], que madame de Grignan
avait ses plus fréquentes réunions et étalait le plus de luxe. Madame
de Sévigné faisait fréquemment à sa fille des cadeaux de modes
nouvelles, et lui envoyait des cravates, des éventails, et autres petits
objets; mais madame de Grignan ayant écrit à sa mère qu'elle se
proposait de se faire peindre et de lui faire présent de son portrait,
madame de Sévigné lui envoya un tour de perles de douze mille
écus, acheté à la vente de l'ambassadeur de Venise. Elle lui écrivait
en même temps: «On l'a admiré ici: si vous l'approuvez, qu'il ne
vous tienne point au cou; il sera suivi de quelques autres [197].»
Cependant, même à Aix, madame de Grignan pouvait se soustraire
au monde et à la dissipation; et elle n'y manquait pas aux époques
où la religion lui en faisait un devoir. Elle se retirait alors dans le
couvent des sœurs de Sainte-Marie, où par un privilége spécial, et à
cause de son aïeule la bienheureuse Chantal, elle était admise
temporairement sur le pied de religieuse, et avait sa cellule
particulière [198].
Le séjour de madame de Grignan chez les sœurs de Sainte-Marie
n'était jamais bien long, et n'avait lieu qu'à de grands intervalles.
Cette existence calme et reposée pouvait lui plaire pendant quelques
jours, par son contraste avec l'agitation de sa vie habituelle; mais
elle n'avait pas, comme sa mère, le goût de la retraite et de la
campagne, les rêveries d'une âme profondément émue, les
palpitations d'un cœur avide de tendresse et d'amour. Ces troubles
intérieurs, qui étaient à la fois pour madame de Sévigné une source
intarissable de jouissances et de tourments, lui étaient inconnus. Elle
savait que sa réputation de beauté, de savoir, de raison, de
prudence, s'était accrue à la cour depuis son départ, et par son
année de séjour en Provence [199]. Le rang qu'elle tenait dans ce pays
flattait son orgueil: là, peut-être, elle se félicitait de n'être pas
éclipsée par sa mère comme à Versailles, comme à Paris. C'eût été
pour elle déchoir que de cesser d'être la première, que de se
retrouver sur un degré d'infériorité ou même d'égalité. M. de
Grignan ne pensait pas ainsi: il aurait mieux aimé être auprès du
monarque, que d'avoir l'honneur de le représenter dans une
province lointaine; les peines et les soins du gouvernement lui
étaient à charge. Ayant appris que le maréchal de Bellefonds voulait
quitter sa place de premier maître d'hôtel du roi, il était disposé à
acquérir cette charge; mais madame de Grignan s'y opposa, et le fit
rester en Provence. On peut juger combien cette résolution affligea
sa mère, qui n'osa s'en plaindre que bien doucement. «Ma chère
enfant, lui dit-elle, cette grande paresse de ne vouloir pas seulement
sortir un moment d'où vous êtes, me blesse le cœur. Je trouve les
pensées de M. de Grignan bien plus raisonnables. Celle qu'il avait
pour la charge du maréchal de Bellefonds, en cas qu'il l'eût quittée,
était tout à fait de mon goût; vous aurez vu comme la chose a
tourné: mais j'aimerais assez que le désir de vous rapprocher ne
vous quittât point quand il arrive des occasions; et M. d'Uzès aurait
fort bonne grâce à témoigner au roi qu'il est impossible de le servir
si loin de sa personne sans beaucoup de chagrin, surtout quand on a
passé la plus grande partie de sa vie auprès de lui [200].»
CHAPITRE III.
1671-1672.

Madame de Sévigné quitte les Rochers, et retourne à Paris.—Elle y prend un


nouveau logement.—Elle désirait revenir à Paris pour être utile à son
gendre.—Ce qu'était l'administration des provinces sous Louis XIV.—
L'évêque de Marseille, Janson, cherche à desservir M. de Grignan. Le
marquis de Charost le défend.—Les affaires des états de Provence se
terminent bien.—Louis XIV se prépare à la guerre contre les Hollandais, et
ne semble occupé que des choses de la paix.—Dans cette persuasion,
Boileau écrit son Discours au roi.—Négociations de Louis XIV.—Il accorde
sa confiance à Condé.—Fête donnée, à Chantilly, au roi et à toute sa cour.
—Création de l'Académie d'architecture.—Le roi propose un prix pour
l'invention d'un nouvel ordre d'architecture, qui serait nommé l'ordre
français.—Nouvelles constructions à Versailles, à Compiègne, à
Fontainebleau.—On joue la Comtesse d'Escarbagnas et Psyché.—La
Fontaine publie un nouveau Recueil de fables et un nouveau Recueil de
contes et de poésies diverses.—Ouverture du jubilé.—Le roi touche 1,200
malades.—Publication de Poésies chrétiennes et diverses.—Belle ode de
Pomponne qui s'y trouve insérée.—Molière fait jouer les Femmes
savantes.—Effet produit par cette pièce.—Elle anéantit la réputation de
Cotin et le règne des précieuses.—De leur heureuse influence sur les
mœurs et sur la littérature.—Julie d'Angennes n'existait plus lors de la
première représentation des Femmes savantes.—Son admirable conduite
et son courage.—Elle devient dame d'honneur de la reine.—Ses remords,
ses chagrins à la cour.—Un fantôme lui apparaît.—Elle tombe malade, et
meurt.—Madame de Richelieu est nommée à sa place dame d'honneur.—
Ce que dit madame de Sévigné de cette nomination, attribuée à
l'influence de madame Scarron sur madame de Montespan.—Madame de
Sévigné soupe souvent avec madame Scarron.—Conduite de celle-ci.—Ses
entretiens avec madame de Montespan déplaisent au roi.—Ses
sentiments, et conduite de madame de Montespan en matière de religion.
—Madame Scarron, d'après l'ordre du roi, se charge d'élever les enfants
qu'il aura avec madame de Montespan.—Conduite admirable de madame
Scarron.—Les enfants que le roi a eus de madame de Montespan lui
inspirent une grande tendresse.—Le roi augmente sa pension.—Il lui
donne un carrosse et des gens, et l'appelle à la cour pour rester auprès
des enfants de madame de Montespan.

Décembre commençait; le froid était piquant; mais le ciel était bleu,


et la lumière du soleil éclatait sur les bois dépouillés de verdure de la
vaste campagne des Rochers, lorsque madame de Sévigné quitta sa
solitude avec un regret dont elle était, disait-elle, épouvantée [201].
Elle se mit en route avec deux calèches attelées chacune de quatre
chevaux, pour elle et sa suite; et afin d'éviter le mauvais pavé de
Laval, elle prit d'abord la route de Cossé, alla coucher chez une
parente de madame de Grignan, à Loresse, château situé dans la
commune de Montjean. Le second jour de son voyage, elle coucha à
Meslay; le troisième, à Malicorne, chez la marquise de Lavardin. En
faisant ainsi environ dix lieues par jour, elle se trouva le dixième jour
transportée à Paris, satisfaite de s'être rapprochée de sa fille, et
ayant pris la résolution d'aller la rejoindre aussitôt que les frimas de
l'hiver auraient disparu, et que le retour de la belle saison lui
permettrait d'entreprendre ce long voyage.
Elle avait quitté, après le mariage de sa fille, le logement qu'elle
occupait à Paris rue du Temple, et transporté son domicile rue de
Thorigny, où elle ne devait pas rester longtemps; car, tandis qu'elle
était encore aux Rochers, des ordres avaient été donnés par elle de
louer, rue Sainte-Anastase, une autre maison près de celle du comte
et de la comtesse de Guitaut [202].
On sait que le château d'Époisses en Bourgogne, qui appartenait au
comte de Guitaut, n'était pas éloigné de Bourbilly, terre de madame
de Sévigné. Aussi dit-elle, en écrivant alors au comte de Guitaut,
qu'il est de sa destinée d'être partout sa voisine [203]. Madame de
Sévigné, aidée de son oncle Saint-Aubin, employa une partie de
l'hiver à faire arranger sa nouvelle demeure, fort rapprochée de celle
qu'elle habitait, et elle y coucha pour la première fois le 7 mai
1672 [204]. Elle y avait fait préparer un appartement pour sa fille et
pour son gendre, et, quoique petite, cette maison lui suffisait; elle
s'y trouvait commodément et agréablement [205]. Nous dirons
comment depuis elle changea encore de logement, et occupa le bel
hôtel Carnavalet [206]. Mais il est remarquable que, malgré son intime
liaison avec madame de la Fayette et le duc de la Rochefoucauld, qui
l'appelait si souvent à l'autre extrémité de la ville, elle ne quitta
jamais le Marais ou le quartier du Temple et le quartier Saint-
Antoine; et, dans ses divers changements, elle se rapprocha de plus
en plus de la place Royale, où elle était née. Elle y trouvait
l'avantage d'être près de toutes les connaissances de sa jeunesse,
qui furent aussi les mêmes que celles de son âge avancé. Elle
pouvait fréquenter toujours les mêmes églises, les Minimes, l'église
Saint-Paul, celle des Jésuites, le couvent des Filles de Sainte-Marie
du faubourg Saint-Antoine, dont les religieuses étaient ses bonnes
amies, et qui avaient parmi elles sa nièce Diane-Charlotte, la fille
aînée du comte de Bussy, dont l'esprit la charmait, dont la piété lui
faisait envie [207].
Un motif plus puissant que celui de se rapprocher de sa fille avait
engagé madame de Sévigné à quitter les Rochers: c'était les services
qu'elle pouvait rendre à son gendre depuis que Pomponne avait été
nommé ministre. Lorsqu'elle revint à Paris, les états de Bretagne
étaient depuis longtemps terminés; mais il n'en était pas de même
de ceux de Provence: ouverts à Lambesc au mois de septembre, ils
se prolongèrent jusqu'en janvier 1672 [208]. Ainsi que nous l'avons dit,
madame de Sévigné aida puissamment à ce que le lieutenant
général gouverneur ne fût pas contraint de s'aliéner la population de
la Provence en déployant contre ses représentants les rigueurs du
pouvoir; aussi M. de Grignan appelait-il sa belle-mère son petit
ministre [209]. Louis XIV travaillait alors à régulariser l'administration
de son royaume; et comme les efforts des provinces, des villes et
des départements pour conserver ce que la hache de Richelieu
n'avait pu abattre de leurs priviléges et de leurs libertés, faisaient
obstacle aux ordonnances du monarque absolu, il mettait tous ses
soins à les anéantir ou à les comprimer. Ce fut surtout durant le
cours des années 1671 et 1672 qu'il obtint les plus grands
résultats [210]. Ce que le comte de Grignan faisait en Provence, le duc
de Chaulnes l'exécutait en Bretagne, le prince de Condé en
Bourgogne [211], le duc de Verneuil [212] en Languedoc. Les députés
des états de cette dernière province avaient décidé qu'à l'avenir on
commencerait les délibérations dans un ordre inverse de celui qui
avait été en usage jusqu'alors; c'est-à-dire qu'on voterait d'abord les
subsides, ou les dons gratuits à offrir au roi, avant de s'occuper des
affaires particulières de la province [213]. Une telle résolution enlevait
nécessairement à l'assemblée des états tout moyen de résister aux
exigences de l'autorité. Aussi cette mesure fut-elle bien accueillie par
Louis XIV, et on s'en entretenait beaucoup à la cour. Sur quoi
madame de Sévigné raconte à sa fille l'anecdote suivante: «L'autre
jour, on parlait, devant le roi, de Languedoc et puis de Provence, et
puis enfin de M. de Grignan, et on en disait beaucoup de bien: M. de
Janson [l'évêque de Marseille] en dit aussi; et puis il parla de sa
paresse naturelle: là-dessus le marquis de Charost [214] le releva de
sentinelle d'un très-bon ton, et lui dit: «Monsieur, M. de Grignan
n'est point paresseux quand il est question du service du roi, et
personne ne peut mieux faire qu'il a fait dans cette dernière
assemblée: j'en suis fort bien instruit.» Voilà de ces gens que je
trouve toujours qu'il faut aimer et instruire [215].»
Madame de Sévigné était d'autant plus heureuse de l'issue favorable
des affaires de Provence, qu'elle avait longtemps craint des résultats
tout différents. Elle était naturellement enchantée de Louis XIV et de
ses ministres, qui se montraient satisfaits des services de son
gendre. Elle écrivait à sa fille: «J'ai tremblé depuis les pieds jusqu'à
la tête; je croyais que tout fût perdu: il se trouve que vous avez
attendu votre courrier, et que vous avez bu à la santé du roi votre
maître. J'ai respiré, et approuvé votre zèle. En vérité, on ne saurait
trop louer le roi: il s'est perfectionné depuis un an. Les poëtes ont
commencé à la cour; mais j'aime bien autant la prose, depuis que
tout le monde en sait faire pour conter et chanter ses louanges [216].»
En effet, depuis la paix d'Aix-la-Chapelle [217], Louis XIV semblait
avoir renoncé aux projets ambitieux qu'il avait manifestés dans les
commencements de son règne; et c'est alors qu'il négociait pour
rompre l'alliance qu'avaient contractée entre elles la Hollande,
l'Angleterre et la Suède [218], qu'il s'apprêtait à punir l'ingratitude et
l'orgueil des Hollandais, tandis qu'on le croyait uniquement occupé
des soins du gouvernement, de la prospérité de son royaume, des
embellissements de Versailles et des plaisirs de sa cour [219].
Ce fut sous l'inspiration de cette croyance générale que Boileau
écrivit son Épître au roi, à laquelle madame de Sévigné fait
principalement allusion, parce que cette épître, d'abord publiée
séparément en 1670, le fut de nouveau au commencement de 1672,
peu de temps avant que la guerre fût déclarée; et quoique le poëte
eût opposé la sagesse pacifique du roi de France à la folie des
monarques conquérants, les éloges donnés dans ses vers pleins de
force, de grâce et de finesse, furent d'autant mieux accueillis, que
Louis XIV craignait de voir toute l'Europe se soulever pour s'opposer
à ses envahissements, et qu'il désirait persuader aux peuples et aux
gouvernements qu'il n'armait que parce qu'il y était contraint pour sa
sûreté et celle des autres États monarchiques, menacés par une
république insolente; que tous ses vœux tendaient à conclure une
paix durable.
Pomponne et Courtin en Suède, le chevalier de Gémonville à Vienne,
le marquis de Ruvigny et Colbert de Croisy à Londres, le marquis de
Villars et Bonzy, archevêque de Toulouse, à Madrid, conduisirent les
négociations qui précédèrent cette guerre avec une activité et une
habileté admirables. Ce ne fut que lorsque Louis XIV, par des traités
secrets [220], eut détaché de la Hollande tous les États qui avaient
intérêt à la soutenir, qu'il eut obtenu le concours des uns et la
neutralité des autres, qu'il fit connaître ses desseins [221]. Les
préparatifs des armements faits par lui, par son ministre Louvois, par
Turenne, par Condé furent dissimulés avec le même soin,
enveloppés du même mystère. Avant d'arrêter son plan de
campagne, il ordonna à Louvois de le soumettre par écrit au prince
de Condé, et voulut avoir l'avis de ce grand capitaine. Pour qu'il ne
fût pas distrait de l'important travail qu'il lui confiait, le roi avait
permis au duc d'Enghien de suppléer son père comme gouverneur
de la Bourgogne [222], ce qui était une manière de lui assurer la
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge


connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and


personal growth every day!

ebookmasss.com

You might also like