Page 1

arXiv:physics/0412180v2 [physics.chem-ph] 8 Apr 2005

On Emerging Field of Quantum

Chemistry at Finite Temperature

Liqiang Wei

Institute for Theoretical Atomic, Molecular and Optical Physics

Harvard University, Cambridge, MA 02138

February 2, 2008

Abstract

In this article, we present an emerging field of quantum chemistry

at finite temperature. We discuss its recent developments on both

theoretical and experimental fronts. We describe and analyze several

experimental investigations related to the temperature effects on the

structure, electronic spectra, or bond rupture forces for molecules.

This includes the study of the temperature impact on the pathway

shifts for the protein unfolding by atomic force microscopy (AFM),

the temperature dependence of the absorption spectra of electrons

in solvents, and temperature influence over the intermolecular forces

measured by the AFM. On the theoretical side, we review a recent

advancement made by the author in the coming fields of quantum

chemistry at finite temperature. Starting from Bloch equation, we

have derived the sets of hierarchy equations for the reduced density

operators in both canonical and grand canonical ensembles. They

provide a law according to which the reduced density operators vary

in temperature for the identical and interacting many-body particles.

By taking the independent particle approximation, we have solved the

equation in the case of a grand canonical ensemble, and obtained an

eigenequation for the molecular orbitals at finite temperature. The

explicit expression for the temperature-dependent Fock operator is

also given. They will form a foundation for the study of the molecular

electronic structures and their interplay with the finite temperature.

1

Page 2

Furthermore, we clarify the physics concerning the temperature ef-

fect on the electronic structure or processes of the molecules which is

crucial for both theoretical understanding and computational study.

Finally, we summarize our discussion and point out the theoretical

and computational issues for the future explorations in the fields of

quantum chemistry at finite temperature.

Keywords Quantum chemistry at finite temperature; temperature depen-

dent; polymers; protein folding; intermolecular forces; solved electrons.

1Introduction

The history for quantum chemistry development is almost synchronous to

that of quantum mechanics itself. It begins with Heitler and London’s study

of electronic structure of H2molecule shortly after the establishment of wave

mechanics for quantum particles [1]. There are two major types of molecu-

lar electronic theories: valence bond approach vs. molecular orbital method

with the latter being the popular one for the present investigation. It has

gone through the stages from the evaluation of molecular integrals via a

semiempirical way to the one by an ab initio method. Correlation issue is

always a bottleneck for the computational quantum chemistry and is un-

der intensive study for over fifty years. For large molecular systems such

as biomolecules and molecular materials, the development of the combined

QM/MM approach, pseudopotential method and linear scaling algorithm

has significantly advanced our understanding of their structure and dynam-

ics. There are about eight Nobel prize laureates whose researches are related

to the molecular electronic structure theory. This not only recognizes the

2

Page 3

most eminent scientists who have made the outstanding contributions to the

fields of quantum chemistry, but more importantly, it indicates the essen-

tial roles the electronic structure theory has been playing in the theoretical

chemistry as well as for the whole areas of molecular sciences. Nowadays,

quantum chemistry has been becoming a maturing science [2, 3].

Nevertheless, the current fields of quantum chemistry are only part of the

story for the molecular electronic structure theory. From the pedagogical

points of view, the quantum mechanics based on which the traditional quan-

tum chemistry is built is a special case of more general quantum statistical

mechanics [4, 5, 6]. In reality, the experimental observations are performed

under the conditions with thermodynamic constraints. Henceforth, there is

a need to extend the current areas of quantum chemistry to the realm of, for

instance, finite temperature [4, 5, 6].

Indeed, many experimental investigations for various fields and for dif-

ferent systems have already shown the temperature or pressure effects on

their microscopic structures [7-30,49-57,63,71-87]. The polymeric molecules

are one of the most interesting systems for this sort of studies [7-16]. The ex-

perimental measurement on the absorption spectra, photoluminescence (PL),

and photoluminescence excitation (PLE), and spectral line narrowing (SLN)

for the PPV and its derivatives all show the same trend of the blue shift

with the increasing temperature [7, 8, 9]. This attributes to the temperature

dependence of their very rich intrinsic structures such as the vibronic cou-

pling [14, 15, 16]. The experimental investigation of the temperature effect

on the biomolecules started in the late nineteenth century [17, 18]. Most re-

cently, it has been extended to the study of folding and unfolding of protein

3

Page 4

or DNA [19, 20, 21]. In addition to the observed patterns for the unfolding

forces with respect to the extension or temperature, it has been proved that

the temperature-induced unfolding is another way for the study of mech-

anisms or pathways of protein folding or unfolding processes [19-23]. The

newest related development is on the AFM measurement made by Lo et al.

of the intermolecular forces for the biotin-avidin system in the temperature

range from 286 to 310K [63]. It has shown that an increase of temperature

will almost linearly decrease the strength of the bond rupture force for the

individual biotin-avidin pair. The study of temperature effect on the absorp-

tion spectra of solvated electron began in the 1950’s and it is still of current

interest. A striking effect is that an increasing temperature will cause the

positions of their maximal absorption red shift [71-85].

In recent papers, we have deduced an eigenequation for the molecular

orbitals [4, 5]. It is the extension from the usual Hartree − Fock equation at

zero temperature to the one at any finite temperature [88, 89]. It opens an

avenue for the study of the temperature effects on the electronic structures

as well as their interplay with the thermodynamic properties. In the third

section, we will present this equation and give the details for its derivation.

In the next section, we will show four major types of experiments related

to the study of the temperature influences over the microscopic structure

of molecular systems. In the final section, we will discuss and analyze our

presentations, and point out both theoretical and computational issues for

the future investigation.

4

Page 5

2 Experimental Development

In this section, we mainly describe the experimental investigations related

to the temperature effect on the bonding, structure and electronic spectra of

molecules. We choose four kinds of the most recent developments in these

fields which are of chemical or biomolecular interests.

2.1 Temperature effects on geometric structure and

UV-visible electronic spectra of polymers

The first important systems where the important issues related to the tem-

perature effect on the geometric structure and electronic spectra are the

polymeric molecules. Many experimental investigations and some theoreti-

cal work already exist in the literature [7-16]. However, how the temperature

changes the microscopic structures of the polymers are still not completely

understood and there are many unresolved issues in interpreting their elec-

tronic spectra. We list here a few very interesting experimental investigations

for the purpose of demonstration.

The poly(p-phenylenevinylene)(PPV ) is one of the prototype polymeric

systems for the study of their various mechanical, electronic, and optical

properties. The impact from the temperature on the absorption spectra,

photoluminescence (PL), and photoluminescence excitation (PLE) of the

PPV have also been investigated both experimentally and theoretically [7,

8, 9]. In an experiment carried out by Yu et al., the absorption spectra are

measured for the PPV sample from the temperature 10 to 330K. The details

of the experiment are given in their paper [9]. The resulting spectra for the

absorption at T = 80 and 300K are shown in Figure 1 of that paper. We

5

Page 6

see that there is a pronounced change in the spectra when increasing the

temperature. They also study the PL and PLE spectra for the PPV. The

measured PL spectra at two temperatures: 77 and 300K are demonstrated

in the Figure 3, and the PLE spectra at those temperatures are depicted in

the Figure 4 of the paper [9]. They both show the dramatic changes of the

band blue shift when the temperature is increased. Similar studies have also

been performed before by the other groups [7, 8]. They observed the similar

behaviors.

Another interesting investigation is related to the temperature effect on

the spectral line narrowing (SLN) of the poly(2-methoxy-5-(2

′-ethylhexyloxy)-

1,4-phenylenevinylene)(MEH − PPV) spin-coated from either THF or CB

solvents [11]. In the experiment done by Sheridan et al., the SLN is mea-

sured together with the absorption and PL as shown in Fig. 1 of their paper.

It is found that the same trend of the SLN blue shift is observed as that for

the absorption and PL with an increasing temperature. They attribute this

to the same reason of the electronic structure modification resulting from the

variation of the temperature.

2.2 Temperature effects on structure, dynamics, and

folding/unfolding of biomolecules

Biomolecules are complex systems, featuring a large molecular size, a hetero-

geneity of atomic constitutes and a variety of conformations or configurations.

Their energy landscape thereby exhibits multiple substates and multiple en-

ergy barrier, and varies in size for the barrier heights [24, 25, 26, 27, 28]. The

temperature should have a strong effect on their structure and dynamics in-

6

Page 7

cluding the folding or unfolding [17-57]. This effect could be either from the

fluctuation of thermal motion of molecules or due to the redistribution of

electronic charge as we will discuss in the next section.

The experimental observation of the temperature impact on the micro-

scopic structure of biological systems dates back to the very early days. One

focus, for example, is on the measurement of the elastic properties of the hu-

man red blood cell membrane as a function of temperature [17, 18]. Another

related study is about the influence over the thermal structural transition of

the young or unfractionated red blood cells due to the involvement of protein

spectrin which might modified the spectrin-membrane interaction [29, 30].

Most recently, the atomic force microscopy (AFM) has been used to detect

the impact from the variation of the temperature on the spectrin protein

unfolding force as well as on the bond rupture force for the biotin-avidin

system [19, 20, 21, 63].

The AFM is a surface imaging technique with an atomic-scale resolution

capable of measuring any types of the forces as small as 10−18N. It combines

the principle of scanning tunnelling microscopes (STM) and stylus profilome-

ter, and therefore can probe the surfaces of both conducting and nonconduct-

ing samples [31, 32]. The imaging on soft materials such as biomolecules with

the AFM has been performed beginning in the 1980’s [33, 34, 35]. Recently, it

has been applied to measure the adhesive forces and energies between the bi-

otin and avidin pair also as we will show in the next subsection [36, 37, 38, 39].

Unlike other experimental techniques, the AFM features a high precision and

sensitivity to probe the surface with a molecular resolution, and can be done

in physiological environments.

7

Page 8

In an AFM investigation of mechanical unfolding of titin protein, for ex-

ample, the restoring forces all show a sawtooth like pattern with a definite

periodicity. It reveals much information about the mechanism of unfold-

ing processes [42, 43]. The observed pattern, in addition to fit a worm-like

chain model, has also been verified by the steered molecular dynamics or

Monte carlo simulations [44, 45]. Similar study has been extended to other

systems [46, 47, 48].

The same kind of experiments has also been performed by varying the

temperature. In the experiment carried out by Spider and Discher et al. [21],

the spectrin protein is chosen for the AFM study at different temperatures.

Thousands of tip-to-surface contacts are performed for a given temperature

because of the statistical nature of the AFM measurement. The observed

curve for the relation between the unfolding force and extension shows the

similar sawtooth pattern for all temperatures. In addition, the tandem repeat

unfolding events are more favored at lower temperature as demonstrated in

the unfolding length histograms. Most striking is that the unfolding forces

show a dramatically nonlinear decreasing relation as the temperature T ap-

proaches the transition temperature Tm. This is shown in the Figure 3B of

paper [21].

Similar behaviors regarding the force-temperature dependence have also

been observed via either AFM or optical tweezers for the forced overstretch-

ing transition for the individual double-stranded DNA molecules [19, 20, 49,

50].

Some other interesting experiments which illustrate the effect of temper-

ature on the microscopic structure of biomolecules have also been performed

8

Page 9

even though the detailed physical origins of the effect (from either the elec-

trons or the molecules) have not been specified [52, 53, 54, 55, 56, 57]. In

a circular dichroism (CD) spectra and high resolution NMR study, for in-

stance, it shows that the secondary structure of the Alzheimer β (12-28)

peptide is temperature-dependent with an extended left-handed 31helix in-

terconverting with a flexible random coil conformation [52]. Another example

is related to the analysis of the temperature-dependent interaction of the pro-

tein Ssh10b with DNA which influences the DNA topology [53, 54, 55]. The

study from the heteronuclear NMR and site-directed mutagenesis indicates

that the Ssh10b exists as a dimer: T form and C form. Their ratio is deter-

mined by the Leu61−Pro62peptide bond of the Ssh10b which is sensitive to

the temperature.

2.3 Temperature effects on intermolecular forces

The study of the general issues related to the temperature effects on the

microscopic structure has been most recently extended to the realm of in-

termolecular forces. Since the usual intermolecular forces such as hydrogen

bond, van der Waals force, ionic bond, and hydrophobic interaction are weak

and typically of the order of 0.1 eV or 4.0 kT at the physiological temper-

ature, the variation of temperature will thereby have a very strong influence

over the strength of these forces.

The first experimental investigation on the temperature-dependent inter-

molecular forces is for the biotin-avidin system and by an AFM measure-

ment [63]. The biotin-avidin complex is a prototypical receptor and ligand

system with the biotin binding strongly up to four avidin protein [58, 59, 60,

9

Page 10

61, 62]. They have an extremely high binding affinity, and therefore serves

as a model system for various experimental investigations. In the experi-

ment carried out in Beebe’s group [63], the receptor avidin is attached to the

AFM tip and linked to the agarose bead functionalized with the biotin. The

temperature of the entire AFM apparatus is varied at a range from 286 to

310K. In addition, the loading rate is kept very slow so that the thermal

equilibrium for the biotin-avidin pairs is assumed. The forces expected to

be determined is the rupture force Fibetween the individual biotin-avidin

pair which is defined as the maximum restoring force [63]. In actual AFM

experiment, however, the total adhesive force between the tip and substrate

is measured. It is a sum of finite number n of the interactions between each

biotin and avidin pair. To extract the individual and average bond rupture,

a statistical method has been developed in Beebe’s group [39, 40, 64, 65, 66].

They assume a Poisson distribution for the number n of the discrete rupture

forces or linkages from multiple measurements, and have obtained the single

force Fiat different temperatures. The result is shown in the Figure 3 of

the paper [63]. We see that the individual rupture force Fifor the biotin-

avidin pair is decreased by about five-fold in strength when the temperature

is increased from 286 to 310K.

To interpret the observed temperature impact on the biotin-avidin forces,

Peebe’s group has performed a thermodynamic analysis [63]. Based on the

simple models and arguments [67, 68], they have come out an equation that

connects the square of the single bond-rupture force Fito the absolute tem-

perature T as follows,

F2

i= 2∆E‡kbond− 2kBTkbondln

?τR

τD

?

(1)

10

Page 11

where the kbondis the force constant of the individual biotin-avidin pair, and

the time τRis the characteristic time needed to break n pairs of those forces.

The E‡is the energy required to remove the biotin from avidin’s strongest

binding site and the corresponding time is τD. More detailed on this analysis

can be found in the paper [63]. The relation between the square of the force

Fiand temperature T is also plotted as Figure 5 in that paper. Therefore,

from the relation (1) and that Figure, the information about the stiffness of

the ligand and receptor bond and the critical binding energy, etc. can be

obtained. Obviously, what we need at present is a microscopic theory which

can account for all these relations and properties.

2.4 Temperature effects on absorption spectra of elec-

trons in solvents

The structure and dynamics of solute in solvent is one of the most impor-

tant fields in chemistry since most of the chemical reactions occur in solution

phases. In the meantime, it is also one of the most challenging fields in the-

oretical chemistry with many unsettled issues. The variation of temperature

in the measurement of absorption spectra of solvated electron in various sol-

vents has proved to be a useful means for the understanding of the solvation

processes [71-85].

There are several experimental techniques available for this type of stud-

ies with the pulse radiolysis being the most commonly used one. There are

also several research groups doing the similar experimental investigations

and obtaining the consistent results relating to the temperature effects on

the optical absorption spectra of solvated electron in solvents. In a recent

11

Page 12

experiment carried out in Katsumura’s group, for example, the pulse radi-

olysis technique is employed to study the optical absorption spectra of the

solvated electron in the ethylene glycol at different temperatures from 290

to 598K at a fixed pressure of 100 atm. In addition to the faster decay of

absorptions, it is found that, their maximal positions shift to the red with

the increasing temperature as shown in the Figures 1 to 3 of the paper [81].

This is in contrast to the situation for the electronic spectra of the polymers.

They also point out the need to quantify the change of the density in the

experiment in order to really understand the observed results.

The same type of experiment has been extended to the study of the optical

absorption spectra for Ag0and Ag+

2 in water by varying the temperature,

and similar results have been obtained [80].

3 Theoretical Development

Having presented four different types of experiments above, we have seen

that the temperature effect on the microscopic structure of molecules is a

very interesting and sophisticated field. Many more needs to be understood

and probed. Even though the experimental study has been for a long time,

very limited number of the related theoretical work is available, especially

at the first-principle level. In other words, the quantum chemistry at finite

temperature is not a well-defined or well-established field [4, 5, 6].

It is true that the influence of temperature on the microscopic structure is

a complicated phenomenon. There exists different functioning mechanisms.

One consideration is that the variation of the temperature, according to

12

Page 13

the Fermi-Dirac statistics, will change the thermal probability distribution

of single-particle states for a free electron gas. It is expected that similar

situation should occur for an interacting electron system, and therefore its

microscopic structure will be correspondingly altered. Another consideration

is that, for molecules or solids, the thermal excitation will cause the change of

the time scales for the molecular motions. This will most likely bring about

the transitions of the electronic states, and therefore lead to the breakdown

of the Born-Oppenheimer approximation. Electron-phonon interaction is a

fundamental topic in solid state physics and its temperature dependence is

well-known. As a result, the temperature could change the strength of the

coupling between the electronic and molecular motions. Nevertheless, we

tackle the issues in a simpler way. We treat only an interacting identical

fermion system, or neglect the coupling of the electronic motion with those

of the nucleus in molecules or solids. We expect that some sort of the general

conclusions will come out from this study. As a matter of fact, this is also the

approach usually adopted in a non-adiabatic molecular dynamics, in which

purely solving the eigenequation for the electrons will provide the reference

states for the investigation of the coupling motions between the electrons and

the nucleus of the molecules.

In the following, we will present a self-consistent equation within the

framework of equilibrium statistical mechanics which decides the molecular

orbitals at a given temperature.

13

Page 14

3.1 Hierarchy Bloch equations for reduced density op-

erators in canonical ensemble

We consider an identical and interacting N-particle system. In a canonical

ensemble, its Nth-order density operator takes the form

DN= exp(−βHN), (2)

and satisfies the Bloch equation [90, 91]

−

∂

∂βDN= HNDN, (3)

where

HN=

N

?

i=1

h(i) +

N

?

i<j

g(i,j), (4)

is the Hamiltonian for the N particle system composed of the one-particle

operator h and two-body operator g. The β is the inverse of the product of

Boltzmann constant kBand absolute temperature T.

Since the Hamiltonian (4) can be written as a reduced two-body operator

form, the second-order reduced density operator suffices to describe its N (≥

2) particle quantum states. A pth-order reduced density operator is generally

defined by [92, 93]

Dp= Lp

N(DN),(5)

where Lp

Nis the contraction operator acting on an Nth-order tensor in the

N-particle Hilbert space VN. The trace of the Dpgives the partition function,

Tr(Dp) = Z(β,V,N). (6)

Rewrite the Hamiltonian in a form

HN= Hp

1+

N

?

j=p+1

h(i) +

p

?

i=1

N

?

j=p+1

g(i,j) +

N

?

i<j(i≥p+1)

g(i,j), (7)

14

Page 15

where

Hp

1=

p

?

i=1

h(i) +

p

?

i<j

g(i,j), (8)

and apply the contraction operator Lp

Non both sides of the Eq. (3), we

develop an equation that the pth-order density operator satisfies [4]

−

∂

∂βDp

= Hp

1Dp+ (N − p)Lp

p+1

?

h(p + 1)Dp+1?

g(p + 1,p + 2)Dp+2?

+ (N − p)Lp

p+1

?p

i=1

?

g(i,p + 1)Dp+1

?

+

+

?

N − p

2

?

Lp

p+2

?

. (9)

It provides a law according to which the reduced density operators vary in

terms of the change of temperature.

3.2 Hierarchy Bloch equations for reduced density op-

erators in grand canonical ensemble

The above scheme for deducing the equations for the reduced operators can

be readily extended to the case of a grand canonical ensemble, which is a

more general one with a fluctuating particle number N. In this ensemble,

the density operator is defined in the entire Fock space

F =

∞

?

N=0

⊕VN,

and is written as the direct sum of the density operators DG(N) associated

with the N-particle Hilbert space VN,

DG=

∞

?

N=0

⊕DG(N), (10)

where

DG(N) = exp[−β(H − µN)],

= exp(−β¯H),(11)

15

Page 16

and

¯H = H − µN, (12)

is called the grand Hamiltonian on VN. The form of the Hamiltonian H has

been given by Eq. (4) and the µ is the chemical potential. The corresponding

pth-order reduced density operator is therefore defined as

Dp

G=

∞

?

N=p

⊕

?

N

p

?

Lp

N[DG(N)], (13)

with the trace given by

Tr(Dp

G) =

??

N

p

??

D0

G

(14)

and

D0

G= Ξ(β,µ,V ).(15)

The Ξ(β,µ,V ) is the grand partition function.

In a similar manner, we can also derive the hierarchy equations that the

reduced density operators in the grand canonical ensemble obey [5]

−

∂

∂βDp

=

¯Hp

1Dp+ (p + 1)Lp

p+1

?¯h(p + 1)Dp+1?

g(p + 1,p + 2)Dp+2?

+ (p + 1)Lp

p+1

?p

i=1

?

g(i,p + 1)Dp+1

?

+

+

?

p + 2

2

?

Lp

p+2

?

, (16)

where

¯Hp

1=

p

?

i=1

¯h(i) +

p

?

i<j

g(i,j), (17)

and

¯h(i) = h(i) − µ.(18)

It gives us a law with which the reduced density operators in the grand

canonical ensemble vary in temperature.

16

Page 17

3.3 Orbital approximation and Hartree-Fock equation

at finite temperature

The Eqs. (9) and (16) define a set of hierarchy equations that establish the

relation among the reduced density operators Dp, Dp+1, and Dp+2. They

can be solved either in an exact scheme or by an approximate method. The

previous study of N electrons with an independent particle approximation to

the Schr¨ odinger equation for their pure states has lead to the Hartree − Fock

equation for the molecular orbitals [94-100]. We thereby expect that the same

approximate scheme to the reduced Bloch equations (9) or (16), which hold

for more general mixed states, will yield more generic eigenequations than

the usual Hartree − Fock equation for the molecular orbitals.

We consider the case of a grand canonical ensemble. When p = 1, Eq.

(16) reads

−∂

∂βD1=¯H1D1+Tr(¯hD1)

D0

D1−1

D0D1¯hD1+2L1

2

?

g(1,2)D2?

+3L1

3

?

g(2,3)D3?

.

(19)

Under the orbital approximation, the above second-order and third-order

reduced density operators for the electrons can be written as

D3= D1∧ D1∧ D1/(D0)2

(20)

and

D2= D1∧ D1/D0. (21)

These are the special situations for the statement that a pth-order reduced

density matrix can be expressed as a p-fold Grassmann product of its first-

order reduced density matrices. With these approximations, the last two

17

Page 18

terms of Eq. (19) can be evaluated in a straightforward way as follows

2L1

2

?

g(1,2)D2?

= (J − K)D1, (22)

and

3L1

3

?

g(2,3)D3?

=Tr(gD2)

D0

−

1

D0D1(J − K)D1, (23)

where

J = Tr2

?

g · D1(2;2)

?

/D0, (24)

and

K = Tr2

?

g · (2,3) · D1(2;2)

?

/D0, (25)

are called the Coulomb and exchange operators, respectively. With the (2,3)

being the exchange between the particle 2 and 3, the action of the K on the

reduced density operator is

K · D1(3;3) = Tr2

?

?

g · (2,3) · D1(2;2)

?

/D0· D1(3;3)

= Tr2

g · D1(3;2) · D1(2;3)

?

/D0. (26)

Substitution of Eqs. (22) and (23) into Eq. (19) yields the Bloch equation

for the first-order reduced density matrix of N interacting electrons under

orbital approximation,

−

∂

∂βD1= (F − µ)D1+

?Tr¯hD1

D0

+TrD2

D0

?

D1−

1

D0D1(F − µ)D1, (27)

where

F = h + J − K, (28)

is called the Fock operator at finite temperature. Redefine the normalized

first-order reduced density operator

ρ1= D1/D0, (29)

18

Page 19

we can simplify above equation into

−

∂

∂βρ1= (F − µ)ρ1− ρ1(F − µ)ρ1. (30)

Furthermore, from Eq. (30) and its conjugate, we get

Fρ1− ρ1F = 0, (31)

which means that the Fock operator F and the first-order reduced density

matrix ρ1commute. They are also Hermitian, and therefore have common

eigenvectors {|φi?}. These vectors are determined by the following eigen

equation for the Fock operator,

F|φi? = ǫi|φi?.(32)

The first-order reduced density operator is correspondingly expressed as

ρ1=

?

i

ω(β,µ,ǫi)|φi??φi|, (33)

where ω(β,µ,ǫi) is the thermal probability that the orbital is found to be

in the state {|φi?} at finite temperature T. Substituting Eq. (33) into Eq.

(30), we can obtain the equation this thermal probability ω(β,µ,ǫi) satisfies,

−

∂

∂βω(β,µ,ǫi) = (ǫi− µ)ω(β,µ,ǫi) − (ǫi− µ)ω2(β,µ,ǫi). (34)

Its solution has the same usual form of the Fermi − Dirac statistics for the

free electron gas as follows,

ω(β,µ,ǫi) =

1

1 + eβ(ǫi−µ), (35)

with the energy levels {ǫi} determined by Eq. (32).

19

Page 20

4 Discussion, Summary and Outlook

In this paper, we have presented a description of both experimental and the-

oretical developments related to the temperature impact on the microscopic

structure and processes for the molecules.

In the theoretical part of this paper, we have depicted the sets of hier-

archy Bloch equations for the reduced statistical density operators in both

a canonical and a grand canonical ensembles for the identical fermion sys-

tem with two-body interaction. We have solved the equations in the latter

case under a single-orbital approximation and obtained an eigen-equation

for the single-particle states. It is the extension of usual commonly used

Hartree − Fock equation at the absolute zero temperature to the situation

at any finite temperature. The average occupation number formula for each

single-particle state is also obtained, which has the same analytical form as

that for the free electron gas with the single-particle state energy determined

by the Hartree − Fock equation at finite temperature (32).

From Eqs. (24), (25) and (28), we see that the Coulomb operator J,

the exchange operator K, and therefore the Fock operator F are both the

coherent and the incoherent superpositions of single-particle states. They

are all temperature-dependent through an incoherent superposition factor,

the Fermi − Dirac distribution, ω(β,µ,ǫi). Therefore, the mean force or the

force field, and the corresponding microscopic structure are temperature-

dependent.

We have expounded the physics relating to the temperature effects on

the electronic structure or processes of the molecules. This is very critical for

our understanding and studying the temperature influence over the molecular

20

Page 21

structure. From this analysis, for example, we could see that the temperature

should have a stronger effect on the molecular transition states and therefore

their chemical reactivity. Starting from Eqs. (9) or (16), it will be a very

significant work to establish a corresponding multireference theory for the

molecular orbitals at finite temperature [96, 101, 102].

On the experimental sides, we have exposed four major fields of investi-

gations of chemical or biomolecular interests, which show the temperature

impact on their structures, spectra, or bond rupture forces.

The complete determination of the geometric structure and electronic

spectra of the polymeric molecules is a very difficult task.As has been

stated in papers [15, 16], there are many different components contributing

to the change of the spectra. At present, we focus on the study of the effect

from the temperature. We have demonstrated that it can alter both the

shapes and positions of the absorption and other spectra for the PPV and

its derivatives. As has been analyzed, the increase of the temperature will

cause the excitation of the vibrational, rotational and liberal motions which

might also lead to the electronic transition. The Huang − Rhys parameter

has been introduced to describe the strength of the coupling between the

electronic ground- and excited-state geometries. Furthermore, it has been

found that this factor is an increasing function of the temperature [7, 8,

9]. Obviously, a more detailed study of the electronic structure, excitation

and spectroscopic signature at the first-principle level which includes the

temperature-dependent force field is expected.

The temperature has proved to be a big player in both experimental and

theoretical study of the structure and dynamics of biomolecules including

21

Page 22

their folding or unfolding. At a first glance, the energy gap between the

HOMO and LUMO for the biomolecules should be small or comparable to

the Boltzmann thermal energy kBT because of their very large molecular

size. Therefore a change of the temperature should have a strong influence

over their electronic states, and consequently, the energy landscape and the

related dynamics including the folding or unfolding, etc. The experimental

investigation with the AFM and other techniques of the temperature effect on

the shift of their unfolding pathways might have verified this sort of thermal

deformation of potential energy landscape [19, 20, 21]. This is in contrast to

the tilt and deformation of energy landscape of biomolecules including their

transition states resulting from the applied mechanical forces [69].

Unfolding proteins by temperature is not just one of the classical exper-

imental techniques for the study of the structure, dynamics and energetics

of the biomolecules. It has also been employed, for example, in the molecu-

lar dynamic simulation to study the structure of transition states of CI2 in

water at two different temperatures: 298 K and 498 K [22]. The later high

temperature is required in order to destabilize the native state for monitoring

the unfolding as done in the real experiments. In another recent molecular

dynamics simulation [23], Karplus’s group has compared the temperature-

induced unfolding with the force-stretching unfolding for two β-sandwich

proteins and two α-helical proteins. They have found that there are the sig-

nificant differences in the unfolding pathways for two approaches. Neverthe-

less, in order to get more reliable results, temperature-dependent force fields

need to be developed and included in the molecular dynamics simulations.

This is also the case in the theoretical investigation of protein folding since

22

Page 23

an accurate simulation of protein folding pathways requires better stochastic

or temperature-dependent potentials which have become the bottleneck in

structure prediction [24, 25, 26, 27, 28]. From structural points of view, the

variation of temperature leads to the change of the mean force or the energy

landscape, and therefore provides a vast variety of possibilities, for instance,

in the protein design and engineering.

The intermolecular forces are ubiquitous in nature. They are also ex-

tremely important for the biological systems and for the existence of life.

The intermolecular forces have the specificity which are responsible for the

molecular recognition between receptor and ligand, antibody and antigen,

and complementary strands of DNA, and therefore for the regulation of com-

plex organization of life [103]. For these reasons, the experiment carried out

in Beebe’s group has an immediate significance. It has demonstrated that

the temperature can be an important factor for changing the specificity of

the intermolecular forces and therefore the function of life [63]. Nevertheless,

how the charge redistribution occurs due to the variation of the tempera-

ture has not been interpreted, and a microscopic theory for quantifying the

temperature influence on the intermolecular forces is still lacking. Since the

delicate study of the intermolecular forces provides the insight into complex

mechanisms of ligand-receptor binding and unbinding processes or pathways,

a paramount future research is to establish the links between the intermolec-

ular forces and the temperature within the quantum many-body theory.

The theoretical study of the temperature effects on the optical absorp-

tion spectra of solvated electron is still in very early stage and few published

works are available [82, 83, 84, 85]. One of the earliest studies by Jortner

23

Page 24

used a cavity model to simulate the solvated electron where the electron is

confined to the cavity surrounded by the dielectric continuum solvent [82].

However, his study is not of fully microscopic in nature since he assumed

a temperature dependence of phenomenological dielectric constants which

were also obtained from the available experimental data. In addition, the

model used is too simplified and, for instance, it neglects the intrinsic struc-

ture of the solvent molecules. There are a few recent investigations on the

temperature effects on the absorption spectra of solvated electrons. They all

cannot catch the full features of the experimental observations. One reason

is that the physical nature for the process is not totally understood which

might leads to incorrect models used for the simulation. The other is to

utilize the crude models which might have omitted some important physical

effects. For example, in an analysis by Brodsky and Tsarevsky [83], they have

concluded a temperature-dependence relation for the spectra which is, how-

ever, in contradiction with the experimental findings at high temperature.

The quantum path-integral molecular dynamics simulation cannot produce

those temperature-dependence relations observed in the experiments [84]. In

a recent quantum-classical molecular-dynamics study by Nicolas et al, even

though the temperature-dependent features of optical absorption spectra for

the solvated electron in water have been recovered [85], however, they claim

that the red shifts of absorption spectra with increasing temperature observed

in both experiments and calculation are due to the density effect instead of

temperature. This might cast the doubt of usefulness of our current theoret-

ical work in this area. However, after examining their work, we observe that

they actually have not included any temperature effect on the electron in

24

Page 25

their theoretical model. This effect might be either from the Fermi − Dirac

distribution for individual electrons or due to the electronic excitation caused

by the thermal excitation of the solvent, as we have discussed in paper [4].

Obviously, much finer theoretical work or more experimental investigations

in this area are expected to resolve this dispute.

In addition to the systems discussed above, there are other types which

also show the temperature impact on their microscopic structures. Either

theoretical or experimental work have been done or are in progress. Examples

include the study of the temperature dependence of the Coulumb gap and

the density of states for the Coulomb glass, the experimental investigation of

the temperature effects on the band-edge transition of ZnCdBeSe, and the

theoretical description of the influence from the temperature on the polaron

band narrowing in the oligo-acene crystals [104, 105, 106].

To sum up, the quantum chemistry at finite temperature is a new and

exciting field. With the combination of the techniques from current quantum

chemistry with those developed in statistical or solid state physics, it will

provide us with a myriad number of opportunities for the exploration.

References

[1] W. Heitler, F. London, Z. Physik 1927, 44, 455-472.

[2] L. Wei, arXiv 2003, physics/0307156.

[3] L. Wei, C. C. Sun, Z. H. Zeng, Ann. Physics 2004, 313, 1-15.

25

Page 26

[4] L. Wei, C. C. Sun, Physica A 2004, 334, 144-150.

[5] L. Wei, C. C. Sun, Physica A 2004, 334, 151-160.

[6] P.-O. L¨ owdin, Int. J. Quantum Chem. 1986, 29, 1651-1683.

[7] T. W. Hagler, K. Pakbaz, K. F. Voss, A. J. Heeger, Phys. Rev. B 1991,

44, 8652-8666.

[8] K. Pichler, D. A. Halliday, D. D. C. Bradley, P. L. Burn, R. H. Friend,

A. B. Holmes, J. Phys.: Condens. Matter 1993, 5, 7155-7172.

[9] J. Yu, M. Hayashi, S. H. Lin, K.-K. Liang, J. H. Hsu, W. S. Fann, C.-I.

Chao, K.-R. Chuang, S.-A. Chen, Synth. Met. 1996, 82, 159-166.

[10] Ch. Spiegelberg, N. Peyghambarian, B. Kippelen, App. Phys. Lett. 1999,

75, 748-750.

[11] A. K. Sheridan, J. M. Lupton, I. D. W. Samuel, D. D. C. Bradley, Chem.

Phys. Lett. 2000, 322, 51-56.

[12] E. Peeters, A. M. Ramos, S. C. J. Meskers, R. A. J. Janssen, J. Chem.

Phys. 2000, 112, 9445-9454.

[13] J. Gierschner, H.-G. Mack, L. L¨ uer, D. Oelkru, J. Chem. Phys. 2002,

116, 8596-8609.

[14] J. Yu, W. S. Fann, S. H. Lin, Theor. Chem. Acc. 2000, 103, 374-379.

[15] S. P. Kwasniewski, J. P. Francois, M. S. Deleuze, Inter. J. Quantum

Chem. 2001, 85, 557-568.

26

Page 27

[16] S. P. Kwasniewski, J. P. Francois, M. S. Deleuze J. Phys. Chem. A 2003,

107, 5168-5180.

[17] A. L. Rakow and R. Hochmuch, Biophys. J. 1975, 15, 1095-1100.

[18] R. Waugh, E. A. Evans, Biophys. J. 1979, 26, 115-132.

[19] H. Clausen-Schaumann, M. Rief, C. Tolksdorf, H. E. Gaub, Biophys. J.

2001, 78, 1997-2007.

[20] M. C. Williams, J. R. Wenner, I. Rouzina, V. A. Bloomfield, Biophys.

J. 2001, 80, 1932-1939.

[21] R. Law, G. Liao, S. Harper, G. Yang, D. W. Speicher, D. E. Discher,

Biophys. J. 2003, 85, 3286-3293.

[22] A. Li, V. Daggett, J. Mol. Biol. 1996, 257, 412-429.

[23] E. Paci, M. Karplus, Proc. Natl. Acad. Sci. 2000, 97, 6521-6526.

[24] J. D. Brynegelson, P. G. Wolynes, J. Phys. Chem. 1989, 93, 6902-6915.

[25] E. I. Shakhnovich, A. M. Gutin, Biophys. Chem. 1989, 34, 187-199.

[26] M. Karplus, A.˘Sali, Curr. Opin. Struct. Biol. 1995, 5, 58-73.

[27] K. A. Dill et al., Protein Sci. 1995, 4, 561-602.

[28] D. Thirumulai, S. A. Woodson, Acc. Chem. Res. 1996, 29, 433-439.

[29] M. Minetti, M. Ceccarini, A. Maria M. Di Stasi, T. C. Petrucci, V. T.

Marchesi, J. Cell. Biochem. 1986, 30, 361-370.

27

Page 28

[30] P. Grimaldi, M. Minetti, P. Pugliese, G. Isacchi, J. Cell. Biochem. 1989,

41, 25-35.

[31] G. Binnig, C. F. Quate, Ch. Gerber, Phys. Rev. Lett. 1986, 56, 930-933.

[32] G. Binnig, H. Rohrer, Rev. Mod. Phys. 1987, 59, 615-625.

[33] P. K. Hansma, V. B. Ellings, O. Marti, C. E. Bracker, Science 1988, 242,

209-216.

[34] B. Drake, C. B. Prater, A. L. Weisenhorn, S. A. C. Gould, T. R. Al-

brecht, C. F. Quate, D. S. Cannell, H. G. Hansma, P. K. Hansma,

Science 1989, 243, 1586-1589.

[35] M. Radmacher, R. W. Tillmann, M. Fritz, H. E. Gaub, Science 1992,

257, 1900-1905.

[36] V. T. Moy, E.-L. Florin, H. E. Gaub, Colloids Surfaces 1994, 93, 343-348.

[37] E.-L. Florin, V. T. Moy, H. E. Gaub, Science 1994, 264, 415-417.

[38] V. T. Moy, E.-L. Florin, H. E. Gaub, Science 1994, 266, 257-259.

[39] Y.-S. Lo, N. D. Huefner, W. S. Chan, F. Stevens, J. M. Harris, T. P.

Beebe Jr., Langmuir 1999, 15, 1373-1382.

[40] T. Han, J. M. Williams, T. P. Beebe Jr., Anal. Chim. Acta. 1995, 307,

365-376.

[41] R. I. MacDonald, E. V. Pozharski, Biochemistry 2001, 40, 3974-3984.

28

Page 29

[42] M. Rief, M. Gautel, F. Oesterhelt, J. M. Fernandez, H. E. Gaub, Science

1997, 276, 1109-1112.

[43] M. Carrion-Vazquez, A. F. Oberhauser, S. B. Fowler, P. E. Marszalek,

S. E. Broedel, J. Clarke, J. M. Fernandez, Proc. Natl. Acad. Sci. USA

1999, 96, 3694-3699.

[44] H. Lu, B. Isralewitz, A. Krammer, V. Vogel, K. Schulten K., Biophys.

J. 1998, 75, 662-671.

[45] P. Carl, C. H. Kwok, G. Manderson, D. W. Speicher, D. E. Discher,

Proc. Natl. Acad. Sci. USA 2001, 98, 1565-1570.

[46] A. F. Oberhauser, P. E. Marszalek, H. P. Erickson, J. M. Fernandez,

Nature 1998, 393, 181-185.

[47] P.-F. Lenne, A. J. Raae, S. M. Altmann, M. Saraste, J. K. H. Horber,

FEBS Lett. 2000, 476, 124-128.

[48] R. Law, P. Carl, S. Harper, P. Dalhaimer, D. W. Speicher, D. E. Discher,

Biophys. J. 2003, 84, 533-544.

[49] I. Rouzina, V. A. Bloomingfield, Biophys. J. 2001, 80, 882-893.

[50] I. Rouzina, V. A. Bloomingfield, Biophys. J. 2001, 80, 894-900.

[51] M. Grandbios, M. Beyer, M. Rief, H. Clausen-Schaumann, H. E. Gaub,

Science 1999, 283, 1727-1730.

[52] J. Jarvet, P. Damberg, J. Danielsson, I. Johansson, L. E. G. Eriksson,

A. Gr¨ aslund, FEBS Letters 2003, 555, 371-374.

29

Page 30

[53] H. Xue, R. Guo, Y. Wen, D. Liu, L. Huang, J. Bacteriol. 2000, 182,

3929-3933.

[54] Y. Tong, Q. Cui, Y. Feng, L. Huang, J. Wang, J. Biomol. NMR 2002,

22, 385-386.

[55] Q. Cui, Y. Tong, H. Xue, L. Huang, Y. Feng, J. Wang, J. Biol. Chem

2003, 278, 51015-51022.

[56] W. Rist, T. J. D. Jorgensen, P. Roepstorff, B. Bukau, M. P. Mayer, J.

Biol. Chem. 2003, 278, 51415-51421.

[57] D. M. Vu, J. K. Myers, T. G. Oas, R. B. Dyer, Biochemistry 2004, 43,

3582-3589.

[58] N. M. Green, Adv. Prot. Chem. 1975, 29, 85-134.

[59] S. A. Darst et al., Biophys. J. 1991, 59, 387-396.

[60] O. Livnah, E. A. Bayer, M. Wilchek, J. L. Sussman, Proc. Natl. Acad.

Sci. USA 1993, 90, 5076-5080.

[61] L. Pugliese, A. Coda, M. Malcovati, M. Bolognesi, J. Mol. Biol. 1993,

231, 698-710.

[62] L. Pugliese, A. Coda, M. Malcovati, M. Bolognesi, J. Mol. Biol. 1994,

235, 42-46.

[63] Y.-S. Lo, J. Simons, T. P. Beebe Jr., J. Phys. Chem. B 2002, 106, 9847-

9852.

[64] Williams J. M.; Han T.; Beebe T. P. Jr. Langmuir 1996, 12, 1291-1295.

30

Page 31

[65] L. A. Wenzler, G. L. Moyes, L. G. Olson, J. M. Harris, T. P. Beebe Jr.

Anal. Chem. 1997, 69, 2855-2861.

[66] Z. Q. Wei, C. Wang, C. F. Zhu, C. Q. Zhou, B. Xu, C. L. Bai, Surf. Sci.

2000, 459, 401-412.

[67] H. Grubm¨ uller, B. Heymann, P. Tavan, Science 1996, 271, 997-999.

[68] S. Izrailev, S. Stepaniants, M. Balsera, Y. Oono, K. Schulten, Biophys.

J. 1997, 72, 1568-1581.

[69] E. Evans, K. Ritchie, Biophys. J. 1997, 72, 1541-1555.

[70] E. Evans, Annu. Rev. Biophys. Biomol. Struct. 2001, 30, 105-128.

[71] H. Blades, J. W. Hodgins, Can. J. Chem. 1955, 33, 411-425.

[72] W. C. Gottschall, E. J. Hart, J. Phys. Chem. 1967, 71, 2102-2106.

[73] B. D. Michael, E. J. Hart, K. H. Schmidt, J. Phys. Chem. 1971, 75,

2798-2805.

[74] F.-Y. Jou, G. R. Freeman, J. Phys. Chem. 1979, 83, 2383-2387.

[75] H. Christensen, K. Sehested, J. Phys. Chem. 1986, 90, 186-190.

[76] H. Shiraishi, Y. Katsumura, D. Hiroishi, K. Ishigure, M. Washio, J.

Phys. Chem. 1988, 92, 3011-3017.

[77] V. Herrmann, P. Krebs, J. Phys. Chem. 1995, 99, 6794-6800.

[78] N. Chandrasekhar, P. Krebs, J. Chem. Phys. 2000, 112, 5910-5914.

31

Page 32

[79] G. Wu, Y. Katsumura, Y. Muroya, X. Li, Y. Terada, Chem. Phys. Lett.

2000, 325, 531-536.

[80] M. Mostafavi, M. Lin, G. Wu, Y. Katsumura, Y. Muroya, J. Phys.

Chem. A 2002, 106, 3123-3127.

[81] M. Mostafavi, M. Lin, H. He, Y. Muroya, Y. Katsumura, Chem. Phys.

Lett. 2004, 384, 52-55.

[82] J. Jortner, J. Chem. Phys. 1959, 30, 839-846.

[83] A. M. Brodsky, A. V. Tsarevsky, J. Phys. Chem. 1984, 88, 3790-3799.

[84] A. Wallqvist, G. Martyna, B. J. Berne, J. Phys. Chem. 1988, 92, 1721-

1730.

[85] C. Nicolas, A. Boutin, B. L˙ evy, D. Borgis, J. Chem. Phys. 2000, 118,

9689-9696.

[86] H. G. Drickamer, C. W. Franck, Electronic Transitions and the High

Pressure Chemistry and Physics of Solids (London, Chapman and Hall:

New York, NY, 1973).

[87] H. G. Drickamer, Annu. Rev. Phys. Chem. 1982, 33, 25-47.

[88] D. R. Hartree, Proc. Camb. Phil. Soc. 1928, 24, 111-132.

[89] V. Fock, Zeit. Physick 1930, 61, 126-148.

[90] F. Bloch, Zeits. f. Physick 1932, 74, 295-335.

[91] J. G. Kirkwood, Phys. Rev. 1933, 44, 31-37.

32

Page 33

[92] H. Kummer, J. Math. Phys. 1967, 8, 2063-2081.

[93] J. E. Harriman, Phys. Rev. A 1978, 17, 1257-1268.

[94] L. Cohen, C. Frishberg, Phys. Rev. A 1976, 13, 927-930.

[95] H. Nakatsuji, Phys. Rev. A 1976, 14, 41-50.

[96] H. Schlosser, Phys. Rev. A 1977, 15, 1349-1351.

[97] J. E. Harriman, Phys. Rev. A 1979, 19, 1893-1895.

[98] C. Valdemoro, Phys. Rev. A 1992, 45, 4462-4467.

[99] D. A. Mazziotti, Phys. Rev. A 1998, 57, 4219-4234.

[100] A. J. Coleman, Intern. J. Quantum Chem. 2001, 85, 196-203.

[101] A. C. Wahl, G. Das, J. Chem. Phys. 1972, 56, 1769-1775.

[102] T. L. Gilbert, J. Chem. Phys. 1974, 60, 3835-3844.

[103] H. Frauenfelder, F. Parak, R. D. Young, Annu. Rev. Biophys. Biophys.

Chem. 17, 451-479 (1988).

[104] M. Sarvestani, M. Schreiber, T. Vojta, Phys. Rev. B 52, R3820-3823

(1995).

[105] C. H. Hsien, Y. S. Huang, C. H. Ho, K. K. Tiong, M. Munoz, O.

Maksimov, M. C. Tamargo, Japn. J. App. Phys. Part 1 2004, 43, 459-

466.

[106] K. Hannewald, V. M. Stojanovic, P. A. Bobbert, J. Phys.: Condens.

Matt. 2004, 16, 2023-2032.

33