Exam Survival Guide Physical Chemistry - Jochen Vogt - 2017 - Springer International Publishing, Cham - Anna's Archive
Exam Survival Guide Physical Chemistry - Jochen Vogt - 2017 - Springer International Publishing, Cham - Anna's Archive
Exam Survival Guide Physical Chemistry - Jochen Vogt - 2017 - Springer International Publishing, Cham - Anna's Archive
This work is subject to copyright. All rights are reserved by the Publisher,
whether the whole or part of the material is concerned, specifically the rights
of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The publisher, the authors and the editors are safe to assume that the advice
and information in this book are believed to be true and accurate at the date
of publication. Neither the publisher nor the authors or the editors give a
warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
References
2.2 Problems
Reference
3 Changes of State
3.1 Systems
3.2.1 Problems
3.3.5 Problems
3.4.3 Problems
References
4 Thermochemistry
4.2 Problems
5 Chemical Equilibrium
5.2 Problems
6 Chemical Kinetics
6.2 Problems
Reference
7 Kinetic Theory
7.1.2 Pressure
7.2 Problems
8 Statistical Thermodynamics
8.3 Problems
References
9.2 Problems
References
10 Spectroscopy
10.2 Problems
References
Appendix A
A.1 Physical Constants
A.3.3 Logarithms
A.3.5 Derivatives
A.3.16 Matrices
Index
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_1
Abstract
This introductory chapter develops and discusses a concept of
mathematically oriented problem-solving in physical chemistry. Based on a
definition of the scientific discipline physical chemistry, the basic skills
needed for successful problem-solving are identified. The concept of
problem-solving is exemplified using a sample problem text. Finally, an
overview of the problems in the various chapters is given, along with
comments on the level of difficulty and thematic cross-links among the
various topics.
(c) Analyze the problem text with regard to special key words.
(c) Make a sketch that collects and illustrates the important facts.
(b) If crucial details are lacking, reflect again on the essential issues
that might be missing.
(c) If you think that essential quantities are undefined in the problem
text, will these quantities be cancelled out at the stage of the
mathematical solution?
(e) Be critical: are you convinced that you have found the correct
approach?
(a) Write down the key equations on the basis of the essential issues
identified.
The order of topics in this workbook roughly follows the way in which
physical chemistry is presented in contemporary textbooks. Stoichiometry
(Chap. 2) is the natural starting point of any quantitative treatment in general
chemistry. Moreover, stoichiometry is a prerequisite for the understanding of
fields such as chemical equilibrium (Chap. 5) and chemical kinetics (Chap.
6). The attentive reader will notice that certain concepts such as the extent of
reaction introduced in Chap. 2, are systematically used in the subsequent
chapters.4 In this sense, the arrangement of problems has an intrinsic order.
But this should by no means prevent the reader from entering into the
problems at an arbitrary point. In a few cases where the solution of a
problem assumes that the reader has dealt with the preliminary contents of
other problems, this is explicitly noted.
Concerning the complexity of the problems, the level of difficulty
gradually increases from chapter to chapter, not only from a mathematical,
but also from a conceptual point of view. Concerning mathematics and the
methods of solution, the attentive reader will notice interesting parallels. A
prime example is the set of problems dealing with oscillating chemical
reactions (Problems 6.6 and 6.7) in the chapter on reaction kinetics on the
one hand, and the set of problems dealing with LASER operation in Chap. 10
(spectroscopy) on the other.5 Seemingly an accidental mathematical
conformity at first sight, these similarities reveal a hidden relationship with
regard to interaction in complex systems that the reader might discover.6
In Chap. 9 dealing with quantum mechanics, problems highlighting some
rather abstract aspects of quantum mechanics, such as operator algebra, were
included, for several reasons. First, the interpretation of quantum mechanics
raises interesting discussions in seminars. Second, operator algebra in
quantum mechanics is a powerful method of producing results with
sometimes surprisingly sparse efforts.7 Third, graduate students starting to
listen to specialized conference talks, e.g., in spectroscopy, will experience
the necessity of being familiar with these methods for their future scientific
work.
References
1. Cullerne JP, Machacek A (2008) The language of physics—a foundation for university study. Oxford
University Press, Oxford
2. Bodner GM, Herron JD (2003) Problem solving in chemistry. In: Gilbert JK, De Jong O, Justi R,
Treagust DF, Van Driel JH (eds) Chemical education: towards research-based training. Springer,
Berlin
2 In fact, mathematics has been called the language of physics [1]. A mathematical formulation of a
problem combines exactness with the complete refinement to the essential facts in a quantitative
manner.
4 In my experience, many students are reserved in using the extent of reaction in concrete problems.
Not appearing explicitly in any fundamental laws such as the law of mass action, it seems somehow
dispensable. In fact, it is possible to work out a correct solution without using it explicitly. However,
this requires an intellectual effort that unconsciously achieves the same purpose as the conscious use
of this concept would do systematically.
5 In fact, for the numerical solution of the laser equations in Problem 10.11, you can use the computer
code of Problem 6.7, with only small modifications.
6 The present book can, of course, only draw the reader’s attention to such points without analyzing the
relationships in full detail, as has been done by Hermann Haken [3].
7 An instructive example is the solution of the quantum double well problem (Problem 9.17) for which
the energy levels can be calculated with arbitrary precision without solving one single integral
explicitly.
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_2
Abstract
Stoichiometry deals quantitatively with the conversion of substances in the
course of a chemical reaction. In this short chapter, we make ourselves
familiar with the definition of some important quantities concerning chemical
reactions. We use them throughout this book, as stoichiometric considerations
are applied in virtually all problems dealing with chemical reactions such as
thermochemistry (Chap. 4) or chemical kinetics (Chap. 6). The problems
presented in this chapter deal with elementary applications of stoichiometry.
(2.6)
Here, n J (t) is the amount of X J at time t, and n J 0 is its initial value.
Note that because the definition of the stoichiometric numbers of the reactants
is to be negative, the amount of reactant decreases, whereas the amount of
product increases in a chemical reaction. Furthermore, as the number of
moles of any substance is always positive, the reaction comes to an end, if
one of the reactants | ν J ξ | reaches n J 0. In general, all but one of the
reactants is present in excess, and one reactant is the limiting reactant .
The definition of ξ does not refer to a specific substance: a differential
change in the extent of reaction is given by
(2.7)
Another important quantity that characterizes the composition of a system,
is the mole fraction x i :
(2.8)
Note that the sum of all mole fractions in a system is 1. Similarly, the
concentration c i of a species is defined as
(2.9)
where V is the system volume.
(2.11)
The molar volumes of substances in the solid and in the liquid state are
generally neglected in favor of the molar volumes of substances in the
gaseous state. The relation among pressure p, volume V, and temperature T is
established by the equation of state. For gases, the equation of state of the
perfect gas,
(2.12)
is in many problems a reasonable approximation. R = 8. 314462 J K−1 mol−1
is the molar gas constant . At atmospheric pressure and a temperature of 298
K, the molar volume of a perfect gas is about 24.8 l.
2.2 Problems
An additional problem directly related to stoichiometry is Problem 3.13 on
page 64.
Solution 2.1
The solution to this problem illustrates the amount of substance of a well-
known material: sodium chloride. At first we shall determine its molar mass
and its molar volume. From the periodic table of elements we take the atomic
weights of sodium and chlorine, M Na = 22. 990 g mol−1 and M Cl = 35. 453 g
mol−1, respectively. Thus, the molar mass of NaCl is the sum of these atomic
weights: M NaCl = M Na + M Cl = 58. 443 g mol−1.
To determine the molar volume v of NaCl, we use the definition of the
density as mass per volume (see Eq. (2.11)):
Hence,
(2.13)
Thus, according to our result, one mole of NaCl has the volume of a cube
of about 3.0 cm in edge length.
Next, from our result on v, we shall determine the nearest neighbor
distance of Na and Cl in the rocksalt lattice shown in Fig. 2.1. The crystal
structure of NaCl is well-known from X-ray crystallography experiments,
and the nearest-neighbor distance can be measured with great precision.
However, just with our result of the molar volume of NaCl, we can
determine d: one mole of NaCl corresponds to N A = 6. 022 × 1023 NaCl
formula units. Hence, the volume occupied by one formula unit NaCl is
(2.14)
There are several possibilities for relating the volume per formula unit to
the crystal structure shown in Fig. 2.1. The shaded cube has a volume V cube
= d 3. This cube has four chloride ions and four sodium ions at its corners.
However, we must bear in mind that each ion is shared by eight neighboring
cubes joining at the respective ionic site. Thus, each of the cubes contains
sodium ions, and chloride ions, i.e., 0.5 NaCl formula units.
Hence, , and thus
(2.15)
This result is very close to the result obtained in X-ray diffraction
experiments [1].
Fig. 2.2 Cards for the elements vanadium and oxygen in the periodic system of the elements (see
Sect. A.4)
Solution 2.2
A simple experiment that does not require a complex setup is the
determination of the stoichiometric composition of a metal oxide after it has
been formed by combustion in an oxygen atmosphere. All that is needed is a
laboratory weighing scale. In our problem, the mass of the reactant, pure
vanadium, is m V = 10. 000 g. The mass of the reaction product, vanadium
oxide, is m oxide = 17. 852 g. We determine its stoichiometric formula and the
corresponding reaction. We start with the general reaction
(2.16)
and we need to determine x and y. First, we evaluate the amount of vanadium
we have used. According to the PSE, the molar mass of vanadium is 50.942
g mol−1. Thus, mol (Fig. 2.2).
The reaction product has a higher mass than the reactant. The reason is
the oxygen, which has been incorporated during the combustion. The mass of
the oxygen is therefore merely the difference m oxide − m V. Hence, the amount
of oxygen is
(2.17)
The ratio between oxygen and vanadium is thus n O: n V = 0. 4908: 0.
1963 = 5: 2. Therefore x = 2 and y = 5 and the reaction sought is
(2.18)
Solution 2.3
This exercise deals with a chemical reaction that comes to an end when one
of the educts is completely consumed. Initially, at a temperature T = 1,000 K
and a pressure of p 0 = 5 MPa, the tank with the volume V 1 = 0. 5 m3 contains
TiCl4(g), the other tank with the volume V 2 = 1 m3 contains CH4(g). We need
to formulate the conversion reaction to TiC(s) and HCl(g):
(2.19)
Next, we calculate the initial mole numbers of these reactants assuming
perfect gas behavior (Eq. (2.12)):
(2.20)
(2.21)
As there is an excess of methane, TiCl4 limits the reaction. What does
this mean? In our problem, the final extent of reaction according to Eq. (2.6)
is when all TiCl4 is consumed. The situation is illustrated in Fig.
2.3. The amounts of CH4, HCl, and TiC can be obtained at ξ final using Eq.
(2.6): There will be mol of methane left in the
reaction volume V 1 + V 2. Moreover, n HCl final = 0 + 4ξ final = 1, 202. 8 mol,
and n TiC = 0 +ξ final = 300. 7 mol.
We calculate the final pressure
(2.22)
The molar mass of TiC is M TiC = 59. 9 g mol−1. Hence, at the end of the
reaction, the mass of the solid reaction product TiC is m TiC final = M TiC × n
final = 18. 0 kg.
TiC
Reference
1. Bragg WH, Bragg WL (1915) X rays and crystal structure. Bell, London
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_3
3. Changes of State
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Thermodynamics is introduced as a quantitative method of characterizing the
changes of state of systems. The topic is subdivided in three categories.
Problems dealing with thermal state variables and equations of state are
found in Sect. 3.2, along with a compact summary of essential theory.
Problems focusing on the caloric state variables are discussed in Sect. 3.3,
again preceded by a summary of basic concepts. Finally, a set of problems
dealing with heterogeneous systems, phase transitions, and mixtures is
presented in Sect. 3.4.
This chapter deals with a field in physical chemistry that offers a direct
approach from our every-day viewpoint: changes in state. A walk through a
winter landscape may stir deep feelings in us about the beauty of nature in its
entirety, but it may also be a good starting point for developing conceptions
about processes in nature and their origin. If we look, for example, at a foggy
lake in winter with ice cakes floating downstream, we see water in its
different forms: water as a liquid, as vapor, or as ice. The melting of a snow-
flake on a warm surface, or the vaporization of a rain drop are concrete
examples of changes of state. But even a change in pressure, temperature, or
volume is a change of state. Thermodynamics is the result of human reflection
about such processes, and it provides the necessary concepts for
understanding the general principles behind them, such as the phase diagram
of a substance, which relates its states of aggregation to pressure and
temperature (Fig. 3.1).
Fig. 3.1 Schematic phase diagram of a pure substance near the triple point (T) with coexistence lines
of solid and liquid, gas and liquid, gas and solid. C is the critical point. SC marks the supercritical phase
3.1 Systems
For the analysis of processes in nature, it is indispensable to subdivide
the considered totality of interacting matter into parts. Typically, we are only
interested in the evolution of a certain amount of matter, clearly distinguished
by the environment, the surroundings. Usually, the properties of the
surroundings are not well-known, but neglecting them completely would be
too crude an approximation. Therefore, the concept of the system is
introduced, which can be subdivided into subsystems, separated by well-
defined boundaries. A system that exchanges neither matter nor heat with its
surroundings is called an isolated system . A system that only exchanges
heat with its surroundings is called a closed system . An open system
exchanges both heat and matter with its surroundings. A chemical system
usually contains a very large number of atoms or molecules, to the order of
1023. On a macroscopic scale, one is primarily interested in only a few state
variables, which result as the average of the movement of all the interacting
atoms and molecules constituting the system. Intensive quantities
characterizing a system do not depend on the size of the system, whereas
extensive quantities do. Extensive properties of subsystems are additive.
Note that for extensive quantities such as the volume V, capital letters are
used in general. Lower case letters are reserved for the related (intensive)
molar quantity, e.g., the molar volume v.
(3.3)
It contains two model parameters, a and b, which can be fitted for each
substance to experimental p–V –T data. To some extent, the van der Waals
model involves the existence of the critical point and a possible coexistence
of condensed phase and gas phase, but it is of limited accuracy (see also
Problem 3.4).
The virial equation for 1 mole of substance relates pressure p,
temperature T, and the molar volume v in the following way:
(3.4)
It has the advantage of directly linking the p–V –T behavior of a
substance to intermolecular interaction. The second virial coefficient B is a
temperature-dependent quantity, which is related to pair interaction between
molecules, the third virial coefficient C is related to interaction among three
molecules, etc.
In the course of a change of state of the system, the thermal state variables
are subject to changes. A change in volume, for example, results as a
consequence of changes in pressure and temperature:
(3.5)
Given a specific equation of state, the differential quotients themselves
can be determined. Moreover, especially in the case of condensed phases,
they can be expressed by important material properties such as the isobaric
thermal expansion coefficient ,
(3.6)
or the isothermal compressibility
(3.7)
Thermodynamics is able to derive relations between material properties
such as α and κ without assuming any microscopic theory of matter (see
Problem 3.1). Note that a material property in general depends on the
temperature or pressure.
3.2.1 Problems
Problem 3.1 (Thermal State Variables)
(3.8)
(3.10)
(3.11)
(3.12)
(3.13)
we have thus shown Eq. (3.8). For the proof of Eq. (3.9) in subproblem (b),
we start again with the total differential of the volume (Eq. (3.5)), and divide
by V:
(3.14)
(3.15)
With the definition of α (Eq. (3.6)) and κ (Eq. (3.7)), we have
(3.16)
(3.17)
which is what was to be shown.
Solution 3.2
Taking liquid water as an example of a condensed phase, we examine the
effects of thermal expansion and compressibility in comparison with perfect
gas behavior.
In subproblem (a) we calculate the change in volume of water, heated at
an atmospheric pressure from T 1 = 298. 15 to T 2 = 323. 15 K. This is the
situation in which we heat water in an open cooking pot, as shown in Fig.
3.2a. In our experience, the effect of thermal expansion is small. We adopt
Eq. (3.9) for the case of constant pressure (dp = 0) and obtain
We obtain
We take these derivatives and obtain from Eqs. (3.6) and (3.7) after
resubstitution of
and
Thus, for a perfect gas, the expansion coefficient is α = 3. 35 × 10−3 K−1
at 298.15 K, and the compressibility is κ = 1. 0 × 10−5 Pa−1 at 100,000 Pa.
Gases have a much larger thermal expansion coefficient than condensed
phases, which, in addition, strongly depend on temperature. Gases also have
a much higher compressibility than condensed phases: if you keep the outlet
of a bicycle tire inflator shut, you can compress the air a small amount.
However, human forces are not able to do the same with liquid water.
(3.18)
or 1. 445 × 1021 particles per cm3. Note that the gas density is simply the
inverse of the molar volume.
In subproblem (b) we use the virial equation Eq. (3.4), but we only
consider the second virial coefficient , which is related to pair interactions
between molecules. Hence, we obtain
The solution with the negative sign yields a vanishing molar volume in
the limit B → 0. It is thus not meaningful in the sense of our problem. The
solution with the positive sign yields a molar volume of 368.0 cm3 mol−1.
Hence, the gas density under consideration of the second virial coefficient is
2. 717 × 10−3 mol cm−3 or 1. 636 × 1021 particles per cm3.
To discover whether or not the repulsive or attractive nature of the
intermolecular interaction is more important under the given conditions, we
compare the molar volumes that we have calculated: if molecular
interactions are neglected, i.e., in the approximation of the perfect gas, we
obtain v = 415. 8 cm3 mol−1 from Eq. (3.18). If we take molecular interactions
into account by considering the second virial coefficient, we obtain a smaller
value of 368.0 cm3 mol−1. Imagine for a moment that we were able to switch
the molecular interaction on and off. If we were to switch it off, a constant
number of molecules would fill a larger volume. Then, if we were to switch
it on again, the molar volume would shrink, i.e., the average distance
between the molecules would be smaller. Thus, we conclude that under the
chosen conditions, the molecular interaction among methane molecules is
attractive, not repulsive. It is convenient to define the so-called compression
factor , which is less than 1 if attractive interactions are dominant,
exactly 1 for vanishing interactions, and greater than 1 if repulsive
interactions are dominant.
In subproblem (c) we include the third virial coefficient in the
calculation:
(3.19)
Note that the sum in the bracket is simply the compression factor z.
Because an analytic solution is tedious, we determine the molar volume
iteratively. Systematically, this could be done using Newton’s method (see
appendix Sect. A.3.19). But even a trial and error procedure starting from
the best guess value v = 368 cm3 mol−1 yields the result with just five
functional evaluations of Eq. (3.19), as demonstrated in Table 3.1.
Table 3.1 Iterative determination of the molar volume of methane based on the virial equation
Hence, the molar volume of methane is 376 cm3 mol−1, determined with
an accuracy of 1 cm3 mol−1. Our final result for the gas density is 2. 660 ×
10−3 mol cm−3. It is instructive to compare our results with experimental
values for methane under these conditions. The true experimental molar
volume of methane at 300 K and 0.6 MPa is 376.2 cm3 mol−1 [1], and the
experimental compression factor is 0.90496. This shows that at a moderate
pressure of 6 MPa, the inclusion of the third virial coefficient is sufficient for
an accurate description of the p–V –T behavior of methane. Omission of the
third virial coefficient, however, consistent with the neglect of three body
interactions, gives a result that deviates by about 2% from the experimental
value.
a. Plot the van der Waals isotherms of nitrogen at 100, 126, 150,
and 300 K.
Solution 3.4
This exercise deals with the van der Waals model of real gases, which to
some extent is capable of explaining the coexistence of the gas phase with a
condensed phase and the critical point (see Fig. 3.1). How good is the van
der Waals model in predicting such properties in a concrete example? Before
we move on to the solution, we recall the qualitative behavior of the pressure
as a function of volume under isothermal conditions below the critical
temperature where gas liquefaction is possible. Consider a gas in a sealed
vessel (Fig. 3.3). By means of a moveable piston, the gas is more and more
compressed, and the pressure increases. If the volume goes below a certain
value, the gas is partially liquefied, and coexistence of the gas and the liquid
is observed. Under these conditions, the pressure within the vessel is the
vapor pressure p v of the substance. If the gas is completely liquefied, the
pressure increases considerably, because a condensed phase is barely
compressible (see Problem 3.2).
Fig. 3.3 The different stages of gas liquefaction at constant temperature
(3.20)
for 1 mol of nitrogen, for which we shall plot isotherms at 100, 126, 150, and
300 K. At a constant temperature, Eq. (3.20) is the mathematical
representation of a van der Waals isotherm. It has a pole for v = b, and thus
involves a finite volume of the molecules. The two parameters a and b need
to be determined from the critical data of nitrogen, the critical temperature T
c = 126. 1 K, and the critical pressure provided p c = 3. 5 MPa (Fig. 3.1).
Above the critical temperature, no coexistence of the gas phase and the liquid
phase is possible, and a supercritical phase is formed. The relation between
the critical data and the van der Waals parameters is obtained from the
analysis of the critical isotherm, . The textbook result is
(3.21)
(3.22)
Division of these equations yields
and
Fig. 3.4 Van der Waals isotherms of N 2 at various temperatures. C is the critical point
The van der Waals isotherms, computed with these values for the
temperatures in question are shown in Fig. 3.4. The 126 K isotherm is the
critical isotherm. It has the critical point as an inflection point.1 The
isotherms at 150 and 300 K show monotonic behavior. The 100 K isotherm
has a minimum and a maximum, which does not reflect the real behavior of a
gas described above.
To obtain a more realistic description within the van der Waals model,
the Maxwell construction shall be applied in subproblem (b). It is
illustrated in Fig. 3.5, where the 100 K isotherm is again shown. The
horizontal line intersecting the isotherm at the three points A, B, and C
indicates a certain constant pressure , which within the model will be
interpreted as the vapor pressure p v at the given temperature, if the enclosed
area between A and B has the same absolute value as the enclosed area
between B and C. Mathematically, these areas are related to the integrals
Fig. 3.5 Maxwell construction of the vapor pressure of nitrogen at 100 K. The curved line is the van
der Waals isotherm at this temperature, the horizontal dashed line p = const. marks the coexistence of
liquid and gas, if the enclosed areas between the intersecting points A and B, and B and C take the same
value
(3.23)
and
(3.24)
The sum W = W AB + W BC is the work done in a cycle starting at A along
the van der Waals isotherm to the point C, and back on the constant pressure
line to point A. For reasonable values of , W AB takes negative
values, whereas W AB is positive, and our task is to find , which is
consistent with vanishing work done in this virtual thermodynamic cycle. The
problem is complicated by the fact that the determination of the intersection
points v A , v B , and v C for a given value cannot be done analytically.
Using an iterative procedure, we obtain MPa. The related
intersection points are v A = 0. 57509 × 10−4 m3 mol−1, v B = 0. 13679 × 10−3
m3 mol−1, and v C = 0. 49053 × 10−3 m3 mol−1. W AB is − 0. 1086 × 103 J, and
W BC = 0. 1086 × 103 J. Thus, for a temperature of 100 K, the van der Waals
model predicts the vapor pressure of nitrogen to be p v = 1. 28 MPa. It is
instructive to compare this value with the experimental value taken from the
literature [2], which is 0.76 MPa. We conclude that the van der Waals model
only gives an approximate quantitative prediction of the vapor pressure of
nitrogen.
(3.26)
The derivative
(3.27)
is called the constant volume heat capacity of the system. The derivative
(3.28)
is called the internal pressure . As a result of the assumption of non-
interacting point masses, the model of the perfect gas does not involve any
dependence of the internal energy on system volume. Hence, its internal
pressure is zero and internal energy depends only on temperature. In contrast,
there is nonzero internal pressure in the model of van der Waals .
Based on the mechanical definition of work related to force F and
distance s, dW = Fds, in addition to the definition of pressure as the
quotient of force F acting on an area A, the work done on a system upon a
change in volume from V 1 to V 2 is defined as
(3.29)
If W is negative, work is done by the system at the expense of its internal
energy or the transfer of heat according to Eq. (3.25). If W is positive, work
is done to the system.
As the internal energy is the sum of all kinetic and potential energy among
the atoms and molecules that constitute a system, the amount of energy needed
to create this system at a certain temperature is U. If the system has a certain
volume V, an additional amount of work pV is necessary to give it room by
displacing the surroundings at an external pressure p. Hence, the work
necessary to create a system and give it room is the enthalpy ,
(3.30)
The differential
(3.31)
the isobaric heat capacity
(3.32)
is derived as a further material property.
(3.33)
where δ Q rev is the infinitesimal heat transferred reversibly to a system at a
temperature T.
According to the Second law of thermodynamics , the entropy in an
isolated system tends toward a maximum:
(3.34)
In the limiting case of a reversible process, Δ S = 0. At this point, it is
worth noting that the calculation of Δ S in an irreversible process requires
the consideration of a reversible equivalent thermodynamic process. For n
moles of a perfect gas undergoing a change of state either from a temperature
T 1 to T 2, from a volume V 1 to V 2, or from pressure p 1 to p 2, the change in
entropy is
(3.35)
where c v is the constant volume molar heat capacity, and c p is the isobaric
molar heat capacity of the perfect gas. A similar equation also holds for a
van der Waals gas (see Problem 3.8). The inspection of Eq. (3.35), although
strictly only valid for a perfect gas, is the key to a general understanding of
the direction of irreversible processes: the entropy of a gas increases, if V 2
> V 1, because in this case . This is the explanation for the above-
mentioned spontaneous free expansion of gases.
(3.38)
where is the heat capacity ratio . Equation (3.37) will be derived in
Problem 3.9.
(3.43)
(3.44)
(3.45)
(3.46)
3.3.5 Problems
Problem 3.5 (Molar Heat Capacities of a van der Waals Gas )
Solution 3.5
This exercise is an instructive example of how relations between material
properties can be derived using the thermodynamic schemes of calculation.
Before we start, we recall the textbook result for the difference in molar heat
capacities c p and c v of a perfect gas :
(3.47)
For a van der Waals gas, we expect a similar expression to hold that contains
the two van der Waals parameters a and b. Moreover, we expect that the
sought expression is identical to Eq. (3.47) in the limit a → 0 and b → 0. We
start our derivation considering the first law (Eq. (3.25)) for 1 mol of a
substance, according to which δ q p = du + p dv is the transferred heat at a
constant pressure.3 Using the total differential for the molar internal energy
(cf. Eq. (3.26)), this can be written as
The first term on the right-hand side contains the constant volume heat
capacity (see Eq. (3.27)). Hence,
(3.48)
The expression in square brackets containing the internal pressure can be
simplified by considering:
(3.49)
Here, we have considered once more the total differential for the molar
internal energy using the expression from Table 3.2 with the molar entropy s
and the molar volume v as natural variables. Therefore,
(3.50)
follows, where we have made use of one of the Maxwell relations. Thus,
Eq. (3.48) simplifies to
(3.51)
So far, our intermediate result (3.51) is general, as we have not yet
specified an equation of state to replace the derivatives of the thermal state
variables. Using Eq. (3.20), it is straightforward to get . However,
cannot be evaluated directly, as the van der Waals equation (3.20)
cannot be solved for v. The trick is to consider the total differential dp,
which is zero at a constant pressure: . Therefore,
(3.52)
follows, and moreover
(3.53)
As expected above, our result agrees with the expression for a perfect
gas, if we set the van der Waals parameters at zero. Moreover, c p − c v → R
for v → ∞, i.e., for a dilute van der Waals gas, the difference c p − c v is the
same as for a perfect gas.
Solution 3.6
This problem deals with the case of an isothermal change of state of a system
of two perfect gases, separated from each other. The initial situation is
depicted in Fig. 3.7. In subproblem (a) we explain why the system is not in
mechanical equilibrium. The latter is established, if the net force acting on
the piston is zero. The force F 2 = −F 2 e z is directed downward (in a
negative z-direction), F 1 = +F 1 e z points upward in a positive z-direction.
Because the initial pressures are different, p Ar 0 ≠ p Ne 0, we can prove that
there is a net force acting on the piston with area A:
(3.54)
Hence, mechanical equilibrium is not established, and the piston moves
in the direction that increases the volume filled with neon, until the pressure
in both volumes is the same:
(3.55)
This is the condition for mechanical equilibrium that we use to calculate
the final volumes of the two perfect gases:
We make use of the fact that the total volume V = V Ar + V Ne of the gases
is unchanged if the piston moves, and, moreover, n Ne = 3n Ar. Hence the
condition for mechanical equilibrium is simplified to
Consistent with the compression of argon, the work done on argon, W Ar,
is positive. Accordingly, W Ne is negative.
At the end of this problem it is worth reflecting on the significance of
mechanical equilibrium in the context of thermodynamics. Although we are
dealing with changes of state, we usually characterize these changes by an
initial state, a final state, and perhaps intermediate states. Even if these states
are not states of thermodynamic equilibrium, they may still be states of
mechanical equilibrium in which the mechanical forces are balanced exactly.
We deal with an example in Problem 3.7. Furthermore, mechanical
equilibrium is a precondition for thermodynamic equilibrium. 4 Therefore,
states of mechanical equilibrium are, for example, important for the
discussion of reversible and irreversible changes of state. An example is
presented in Problem 3.9.
b. Calculate the work and the change in entropy for both gases.
Solution 3.7
This problem is an extension of Problem 3.6. Starting from the same initial
conditions (see Fig. 3.7), we calculate the volume of the two gases, neon and
argon, separated by a piston, if the latter moves reversibly and adiabatically
instead of isothermally into its equilibrium position. What is different? If no
heat exchange occurs, neon does expansion work at the expense of losing
internal energy, and cools down. Conversely, argon, which is compressed,
heats up. Hence, the final temperature of both gases is different. Furthermore,
we cannot assume that the final equilibrium pressure on both sides of the
piston will be the average of the initial values, as in Problem 3.6. Our
solution is again based on the condition for mechanical equilibrium, Eq.
(3.55), but we use Eq. (3.36) for reversible adiabatic changes of state:
(3.56)
(3.57)
is the heat capacity ratio. Thus, by using Eq. (3.55), the
condition is
the same pressure is obtained for p Ne. Note that the equilibrium pressure
differs from our result for the isothermal case (Problem 3.6). Finally, we
calculate the temperature of both gases using the equation of state for the
perfect gas, and the initial temperature T 0 = 298 K:
As expected, neon has cooled down after its expansion, and the
compressed argon has heated up markedly.
In subproblem (b) we shall calculate the work and the change in entropy
for both gases. For an adiabatic change of state, the work can be calculated
from the first law (Eq. (3.25) ) with the assumption δ Q = 0. Therefore,
and
Fig. 3.8 Free expansion of neon and argon after removal of the piston
(3.58)
We note that the number of moles of neon and argon can be calculated
from the initial conditions, mol, and n Ne = 3n Ar,
because p Ne 0 = 3p Ar 0. According to Eq. (3.58) T f is 15.9 K smaller than T
0,the initial temperature of the two gases, before the piston went into its
position of mechanical equilibrium. Did we expect this? To check our result,
we calculate T f in a second way, using an argument of energy conservation.
Between the initial state of the gases at temperature T 0 = 298 K and the final
state at T f , the above calculated work involved with the movement of the
piston was W = 16 J. No further was done, because, after the piston was
removed, the expansion of the gases was a free expansion. As there was no
heat transfer with the surroundings, we did indeed expect a cooling of the
gases in the final state by
(3.59)
and thus T f = T 0 −Δ T = 282. 1 K, i.e., the same result as above. Finally,
we can calculate the entropy change in the gases using Eq. (3.35):
(3.60)
(3.61)
As expected, the mixing of the gases leads to an increase in their entropy,
emphasizing the irreversible nature of this process.
(3.62)
Solution 3.8
Sometimes it is quite difficult to deal with entropy changes in the correct
way. An instructive case is the free expansion of a perfect gas . The
movement of individual gas particles is completely uncorrelated, and in a
process of diffusion,6 the entirely accessible volume rapidly filled by the
gas. This is an example of an irreversible process . From the atomistic point
of view, the uncorrelated movement of non-interacting particles leaves a
probability of finding all the particles back in the initial volume, which is so
small, that it never happens in practice. Students frequently think that they
have figured out a contradiction between the irreversible nature of the free
expansion and the fact that there is no heat transfer with the surroundings.
They argue that according to the Clausius equation, Eq. (3.33), the entropy
change should then be zero, and thus the expansion should be reversible. Of
course, they overlook the fact that the application of Eq. (3.33) assumes that
the heat is exchanged in a reversible process, which is not the case. The
reason for the zero heat transfer in the case of a free expansion of a perfect
gas is the absence of intermolecular interaction, equivalent to the fact that the
internal energy of a perfect gas is only a function of temperature, not of
volume.
An even more puzzling case is the free expansion of a van der Waals gas,
which we deal with in this problem. Here, we have to take molecular
interaction into account, which, during expansion must be overcome. Thus,
under isothermal conditions, if the gas is in thermal contact with the
surroundings, there will be a small heat transfer; thus, there will also be a
change in entropy of the surroundings.
In subproblem (a) we derive a general formula for the molar entropy
change in a van der Waals gas undergoing a change of state. Equation (3.62)
is analogous to Eq. (3.35), which is valid for a perfect gas. Our derivation
starts with Eq. (3.26). If we consider 1 mol of substance,
(3.63)
where Π is the internal pressure. Equating the last expression to du = T ds −
p dv, we obtain
(3.64)
Exploiting Eq. (3.50) in Problem 3.5 and the van der Waals equation of
state Eq. (3.3), we can evaluate the bracket term and obtain
(3.65)
Finally, integration of the latter equation proves Eq. (3.62):
(3.66)
(3.67)
This equation is very similar to the corresponding expression for a
perfect gas (Eq. (3.35)). Interestingly, the change in molar entropy of a van
der Waals gas does not depend on the van der Waals parameter a or on the
internal pressure Π, which according to Eqs. (3.50) and (3.3), is given by
(3.68)
Equation (3.66) can be generalized for arbitrary mole numbers:
(3.69)
In subproblem (b) we calculate the change in entropy Δ S for n = 1 mol
xenon with a = 4. 250 dm6 bar mol−2 and b = 0. 0511 dm3 mol−1. The initial
and final volume is V 1 = 0. 5 dm3 and V 2 = 1 dm3 respectively. In SI units, a
= 0. 4250 m6 Pa mol−2 and b = 5. 11 × 10−5 m3 mol−1. Using Eq. (3.69), the
entropy change in the gas due to the isothermal expansion is
(3.70)
(3.71)
(3.72)
As expected, the entropy of the gas increases upon expansion, for a van
der Waals gas as well. For comparison, the corresponding result for a perfect
gas (Eq. (3.35)) would have been + 5. 763 J K−1. Now, we are interested in
the entropy change in the surroundings associated with this process, Δ S surr.
As discussed above, there is a low heat transfer from the surroundings to the
gas upon expansion, resulting in a negative Δ s surr. However, is it correct to
use the Clausius equation (3.33) and write
(3.73)
at this point? Above, we have mentioned the significance of a heat transfer in
a reversible process as a requirement for using the Clausius equation, but
here, we assume the expansion of our van der Waals gas to be irreversible.
Therefore, how can we justify Eq. (3.73)? Consider the surroundings as a
system that is so large that the small heat transfer Δ Q surr does not change the
system temperature. The entropy change in the surroundings only depends on
the amount of heat transferred. We could construct an alternative reversible
thermodynamic process in a system in contact with the surroundings, leading
to just the same heat transfer and thus to the same change in entropy of the
surroundings. Hence, regarding the entropy change of the surroundings, it
does not matter if the van der Waals gas undergoes a reversible or an
irreversible change of state, as long as only the same heat is transferred.
Next, we calculate Δ Q surr = −Δ Q gas on the basis of the first law Eq. (3.25).
Because in a free expansion, no work is done by the gas, Δ Q gas corresponds
to the change in its internal energy, Δ Q gas = Δ U gas. As T = const.,
(3.74)
After a step of integration, we obtain
(3.75)
(3.76)
(3.77)
Thus, following our above argument,
(3.78)
Hence, the total entropy change in the total system constituted by the
surroundings and the van der Waals gas is Δ S = Δ S gas +Δ S surr = +4. 798 J
K−1.
b.
Calculate the final temperature and volume in the case of a
reversible expansion.
Solution 3.9
This exercise deals with adiabatic changes of state of a perfect gas and the
differences between reversible and irreversible processes. Initially locked, a
moveable piston limits the volume of the gas to V 1 = 0. 01 m3. The perfect
gas within the cylinder is thermally isolated from the surroundings, i.e., no
heat transfer is possible. Using the equation of state (Eq. (3.2)), we can
calculate the initial pressure within the cylinder,
(3.79)
(3.80)
(3.81)
As this initial pressure is larger than the external pressure of p ex = 105
Pa, the gas expands adiabatically as soon as the piston is unlocked. Hence,
work is done upon expansion, and, according to the first law (Eq. (3.25)) and
the adiabatic condition δ Q = 0, the amount of work done by the gas is
balanced by a reduction in its internal energy, which in turn is a function of
temperature. Thus, we expect a cooling of the gas during expansion.
Fig. 3.9 Adiabatic irreversible expansion of a perfect gas, initiated by removing a mass from the
piston. Upon expansion, the temperature of the gas is reduced
(3.83)
(3.84)
(3.85)
The final volume is
(3.86)
As expected, the gas has cooled down from 320 to about 255 K during
expansion. The work is best calculated from this temperature difference,
(3.87)
Before discussing these results, we move on to subproblem (b), where
reversible adiabatic expansion is assumed. How could we at least
approximately realize this experimentally? We can obtain quasi reversible
processing in a step-by-step procedure, by removing small pieces of mass
one by one from the piston, as illustrated in Fig. 3.10. With the limit of
arbitrarily small differential masses dm, we would obtain reversible
expansion. We check this by recalling the difference between an irreversible
and a reversible change of state: if a process is irreversible, then the initial
state can only be re-established by doing work. If we look at Fig. 3.9, this
corresponds to lifting the single mass upward back onto the piston by doing
linear work. The piston then moves down until the initial state is reached. In
contrast, as illustrated in Fig. 3.10, we could simply establish the initial state
by putting the small mass pieces back on the piston one by one—without
doing linear work. A second intuitive criterion of reversibility is time
invariance . If we record a movie of the quasi-reversible expansion in Fig.
3.10 and run the movie in reverse, we could simply see a physically
meaningful process: pieces of mass are taken back onto the piston, and the
piston moves gradually down step by step, i.e., the reversible compression of
the gas. In contrast, a movie of the process illustrated in Fig. 3.9 run
backward would show a physically absurd scene: a gas does not
spontaneously reduce its volume and heat up!
Fig. 3.10 Adiabatic quasi reversible expansion of a perfect gas by removing several small pieces of
mass from the piston. In each of the intermediate steps, the gas takes different values of pressure and
temperature
(3.88)
After integration,
(3.89)
and consideration of Eq. (3.47), one of Poisson’s equations follows:
(3.90)
(3.91)
(3.92)
Finally, the work done by the gas is
(3.93)
In subproblem (c) we deal with entropies and calculate them explicitly.
We expect Δ S rev = 0 for the case of the reversible adiabatic expansion, and
Δ S irrev > 0 for the irreversible adiabatic expansion, consistent with the
second law and Eq. (3.34). We bear in mind that the entropy of a gas
increases upon expansion, but it decreases if its temperature is reduced.
Thus, there are two competing effects that govern the total change in entropy.
We start with Eq. (3.35), first dealing with the case of the irreversible
adiabatic expansion, and use the results for V 2 and T 2 from subproblem (a):
(3.94)
(3.95)
Fig. 3.11 p–V diagram of the adiabatic expansion. Point A is the initial state, B the final state of the
reversible (isentropic) expansion, C is the final state of the irreversible case
In the isentropic, reversible adiabatic case more work is done upon
expansion, consistent with stronger cooling of the gas and a smaller final
volume. Thus, the final state variables of the gas do not coincide. This is also
seen in the p–V diagram of the expansion in Fig. 3.11. The curve marks the
reversible (isentropic) expansion, ending in the final state at point B. For
each point between the initial and final volumes, the pressure of the gas is
precisely defined according to the adiabatic line, which we could reconstruct
by removing a series of infinitesimally small masses from the top of the
piston, as outlined above. In contrast, the irreversible expansion lacks any
intermediate states. In the p–V diagram, this case is only characterized by the
two points A and C of the initial and final state.
(3.96)
(3.97)
(3.98)
(3.99)
The chemical potential μ j of a substance in a special phase is the
amount of Gibbs free energy the system gains, if dn j mol of substance are
added at constant pressure and temperature:
(3.100)
Moreover, using the Euler equation U = TS − pV + ∑ j μ j n j it can be
shown that
(3.101)
i.e., the Gibbs free energy of a system is the sum of the chemical potentials of
all substances in their various phases, weighted with their respective
amounts of substance. The conditions for the direction of a spontaneous
change, Eqs. (3.41) and (3.42), in addition to the above equations, are the
basis for the description of heterogeneous systems, including phase
diagrams and, moreover, the phenomena related to the mixing of substances,
the colligative properties : osmotic pressure, vapor pressure reduction, the
increase in the boiling point, and the lowering of the freezing point.
The coexistence lines in a phase diagram (see Fig. 3.1) are characterized
by a reversible exchange of atoms or molecules between two different
phases, e.g., the liquid phase and the gas phase. Because Δ G = Δ H − T Δ S
= 0 in this case, the molar entropy change involved with such a phase
transition on a point of the coexistence line is
(3.102)
where Δ h tr is the respective molar transition enthalpy, e.g., the molar heat of
vaporization. On the coexistence lines, the chemical potentials in two phases
are equal. This condition is sufficient to deduce the coexistence lines in the
phase diagram of a pure substance (Fig. 3.1). The Clapeyron equation
(3.103)
provides the gradient of a coexistence line on the p–T diagram, where Δ v tr
is the change in the molar volume involved with the phase transition. For the
special cases of a gas-liquid or a gas-solid phase transition, two
approximations are frequently made: (1) The gas-phase is described by the
equation of state for a perfect gas (Eq. (3.2)). (2) The molar volume of the
condensed phase is neglected over the molar volume of the gas phase. In this
case, Eq. (3.103) together with Eq. (3.102) can be used to derive the
Clausius-Clapeyron equation (see Problem 3.10a)
(3.104)
(3.105)
An idea of how this relation can be derived gives the solution to
Problem 3.11b. In the liquid phase, the respective expression is
(3.106)
⋆
where the a A is the activity of substance A in solution, and μ A is the
chemical potential of the pure substance A. The symbol ⋆ now refers to a
new standard state of a pure substance, which is also called Raoult’s law
standard state . The activity of A is defined via its partial pressure p A over
the solution, related to the vapor pressure of the pure substance, p A ⋆:
(3.107)
It is important to understand that the activity is a property of the solution,
but it is determined indirectly by the partial pressure in the gas phase above
the solution in chemical equilibrium, where the chemical potential in the
solution and in the gas phase are equal. For a real solution, the activity may
depend on the composition of the solution in a complicated way. In the
limiting case of an ideal solution, the activity corresponds to the mole
fraction: a A = x A, and the partial pressure of A can be calculated using
Raoult’s law :
(3.108)
In general, the relation between activity and mole fraction is
(3.109)
where γ A is the activity coefficient of A.
3.4.3 Problems
Problem 3.10 (Vapor Pressure of a Pure Substance)
(3.111)
Based on the equality of the chemical potentials, we derive the
coexistence line between vapor and liquid in subproblem (a). Starting with a
point on the coexistence line at a given pressure and temperature, p and T,
we consider the chemical potentials of the vapor and the liquid at p + dp and
T + dT, still on the coexistence line:
(3.112)
We can expand these chemical potentials into a power series and
consider only the linear terms,
(3.113)
Using the definition of the chemical potential Eq. (3.100) and the
relations and , the differential quotients and
are identified as the negative (partial) molar entropy s i and the (partial)
molar volume v i of the substance in phase i respectively. Thus,
and as we may consider the chemical potentials of the vapor and the liquid
phase to be equal at p + dp and T + dT as well,
(3.114)
We have now derived the Clapeyron equation . Next, we make the
approximation Δ v vap ≈ v gas, i.e., we neglect the molar volume of the liquid
over the molar volume of the gas phase, which is reasonable at moderate
pressures. Moreover, assuming perfect gas behavior, we can express
by the equation of state. Inserting this into the Clausius equation, we
obtain:
(3.115)
As a consequence, if the molar heat of vaporization is known and one
point on the coexistence line, e.g., the standard boiling point, the vapor
pressure at any given temperature can be determined. Moreover, because the
standard molar entropy of vaporization is as a special case of
Eq. (3.102), we obtain
(3.116)
Fig. 3.12 Graphical determination of the enthalpy and entropy of vaporization from vapor pressure
variation with temperature according to Eq. (3.116)
We can use this relation to determine Δ h vap ⊖ and Δ s vap ⊖ and the
standard boiling point of ethylamine, as we do in subproblem (b). Therefore,
we plot against the reciprocal temperature, as shown in Fig. 3.12. If Δ
h vap ⊖ and Δ s vap ⊖ are constant over the temperature range of tabulated
data, then a linear behavior can be expected. Inspection of the diagram shows
that this seems to be the case. A linear regression provides the axis intercept
and the inclination of the best-fit line:
Note that in general, Δ h vap and Δ s vap will depend markedly on
temperature, if the examined temperature interval is sufficiently large. The
standard boiling point can be determined in two different ways. The simplest
way is based on Eq. (3.102) and yields
(3.117)
The alternative way is based on the analysis of Fig. 3.12. Because at T b
⊖ the vapor pressure is only p ⊖, the best-fit line intersects the line
at K−1. We therefore obtain T b ⊖ = 289. 2 K. The
discrepancy of 0.1 K can be explained by the statistical and systematic
uncertainties in the experimental vapor pressure data. In subproblem (c) we
calculate the pressure within two identical vessels, which at T = 333 K
contain 5 and 10 g of ethylamine respectively. This question challenges the
student’s ability to assess whether under given conditions a system is a
homogeneous system or a heterogeneous system. The molar mass of
ethylamine is M = 45 g mol−1. Thus, the first vessel contains an amount of
(3.118)
The second vessel is filled with 10 g of ethylamine, which corresponds to
mol. Next, we calculate the vapor pressure of ethylamine at T = 333
K using our results from subproblem (a). Apparently, the vapor pressure is
(3.119)
(3.120)
(3.121)
If ethylamine in the vessel is present as a liquid phase coexisting with a gas
phase, i.e., as a heterogeneous system, then the inner pressure of the vessel
will be this vapor pressure of about 5 bar. Here, however, we have also take
the possibility into account that the amount of ethylamine might be insufficient
to establish the coexistence of liquid and vapor. If we assume that gaseous
ethylamine is a perfect gas, then the nominal gas phase pressure at a given
amount of substance and temperature according to the equation of state is
(3.122)
As a consequence, the 5 g of ethylamine in the first vessel is present in the
gaseous state. This is also illustrated in Fig. 3.13, where the coexistence line
between vapor and liquid is shown in a p–T diagram. The point p 1 resides in
the area below the coexistence line, i.e., in the area of gaseous ethylamine.
The pressure within the vessel is thus p 1 = 3. 08 bar
Fig. 3.13 Coexistence line p vap(T) of gaseous and liquid ethylamine (solid line) according to Eq.
(3.119). The dashed lines indicate the nominal gas pressure of 5 g (p 1) and 10 g (p 2) ethylamine in
the vessel according to the perfect gas equation of state. At T = 333 K, p 2 > p vap > p 1
(3.123)
as also shown in Fig. 3.13. Therefore, a fraction of ethylamine condenses and
forms a liquid phase coexisting with gaseous ethylamine. In this case, the
pressure within the vessel is thus the vapor pressure p vap = 5. 04 bar.
a. For both forms of carbon, calculate the molar Gibbs free energy
of formation at a pressure of 10 bar and 298 K. Assume that
diamond and graphite are incompressible solids.
Solution 3.11
In this exercise, we deal with the relative stability of two forms of carbon,
diamond and graphite. We deal with the relationship between the Gibbs free
energy of a substance as a function of pressure, its structure, and the phase
diagram.
Graphite is the stable form of carbon under normal atmospheric pressure.
As an element in its standard state , by definition, its standard molar Gibbs
free energy of formation is zero.8 Diamond, however, can be formed if a high
pressure is exerted on carbon. Its crystal structure is more compact and thus
its density is higher than that of graphite.
In subproblem (a), we seek the function g( p) of diamond and graphite
and we treat them as incompressible solids. From Eq. (3.99),
(3.124)
where v is the molar volume, which is related to the density ρ and the molar
mass M: . Integration yields
(3.125)
For graphite with g( p ⊖) = 0 and ρ = 2300 kg m−3, the molar free Gibbs
energy takes a value of
(3.126)
For diamond with g( p ⊖) = 2900 J mol−1, we obtain a value of 2903.1 J
mol−1.
In subproblem (b) we consider the function g( p) for CO2 treated as a
perfect gas:
(3.127)
Using the equation of state, .
(3.128)
For CO2 with g ⊖ = −394. 4 kJ mol−1, we obtain a value of g(106 Pa) =
−388. 7 kJ mol−1. In Table 3.4, all results are summarized:
Table 3.4 Molar Gibbs free energies of graphite, diamond, and CO2 at standard pressure and at p =
106 Pa
Substance State g( p ⊖ ) g(106 Pa) Δg
C (Graphite) Solid 0 + 4. 7 × 10−3 + 4. 7 × 10−3
C (Diamond) Solid +2.9000 +2.9031 + 3. 1 × 10−3
CO2 Gaseous −394.4 −388.7 +5.7
Although the molar Gibbs free energy of solids changes only weakly to the
order of a few J mol−1 under a moderate change in pressure, the Gibbs free
energy of a perfect gas changes markedly to the order of several kJ mol−1.
This general trend allows a different treatment of gases and condensed
phases in the thermodynamic characterization of the chemical equilibrium in
Chap. 4
In subproblem (c) we calculate the pressure at which diamond becomes
the stable form of carbon. Looking at our tabulated results, we recognize that
the molar Gibbs free energy of diamond increases slightly more weakly than
that of graphite, if the pressure is increased. Thus, at a certain pressure p eq,
the molar Gibbs free energies of diamond and graphite are equal. For p > p
eq, diamond becomes the stable form of carbon, indicated by the lower value
of g( p). The situation is illustrated in the schematic phase diagram in Fig.
3.14. To evaluate p eq, we use our above results and consider the transition
Fig. 3.14 Schematic phase diagram of carbon with the phase boundary between graphite and diamond
and our result for p eq is 1.62 GPa. Experimentally, the transition from
graphite to diamond is observed at pressures above 2 GPa.
The main components of liquid petroleum gas (LPG) are propane and
butane in seasonally varying compositions. Assuming that propane and
butane constitute an ideal mixture , calculate the maximum acceptable
mole fraction of propane, for which at a temperature of 50 ∘C the
internal pressure of a tank containing LPG does not exceed a value of
12 bar. The molar standard heats of vaporization of pure propane and
butane are 19.0 and 22.4 kJ mol−1 respectively. The standard boiling
points of propane and butane are 231.1 and 272.7 K respectively.
Fig. 3.15 A tank containing a binary solution of a more volatile substance indicated by white balls, and
a less volatile species (blue balls), which is in excess. In the gas phase, the more volatile component is
enriched
Solution 3.12
In middle Europe, the summer composition of LPG fuels is about 40 mass-%
propane and 60 mass-% n-butane, and vice versa in winter. At the filling
station, a compressor has to work against the tank internal pressure. Hence,
to guarantee successful filling, the internal pressure should not exceed the
maximum pressure of the compressor, p max = 12 bar in this problem. In
summer, the temperature of a car tank may easily reach 50 ∘C. At this
temperature, the vapor pressure of pure butane and propane is calculated
using the Clausius Clapeyron law (Eq. (3.115)). For butane, with Δ h vap,
⊖ = 22. 4 kJ mol−1 and T ⊖ = 272. 7 K
butane b, butane
For propane with Δ h vap, butane ⊖ = 19. 0 kJ mol−1 and T b, butane ⊖ = 231.
1 K, a value of 16.94 bar results. Consistent with the higher enthalpy of
vaporization and the lower standard boiling temperature, propane is more
volatile than butane. The situation is illustrated in Fig. 3.15. Propane (C3H8)
and n-butane (C4H10) are hydrocarbons with similar chemical properties.
The assumption of an ideal mixture is thus reasonable and justifies the
application of Raoult’s law (Eq. (3.108)), by which we calculate the total
pressure in the gas phase
(3.130)
For p = p max and x B = 1 − x P , we obtain an expression for the
maximum acceptable mole fraction of propane:
Hence, at least in summer, the fraction of propane should not exceed
60%. From this result, and the molar masses of propane and butane (44.1 and
58.1 g mol−1), the maximum acceptable mass fraction of propane is 53%.
Solution 3.13
The lowering of the vapor pressure of a solvent in the mixture with another
substance is one of four colligative properties .9 These are used in analytics
to determine the molar mass of a solute. Combined with results from an
elementary analysis providing the relative abundances of chemical elements,
the molecular formula of an unknown compound can be determined. In this
problem, we consider m X = 20 g of an unknown compound (denoted X),
which is dissolved in m D = 1000 g diethyl ether. This causes a lowering of
the vapor pressure of diethyl ether from p D ∗ = 586 hPa to p D = 583 hPa.
Moreover, elementary analysis of X yields mass fractions f C = 0. 414 for
carbon, f H = 0. 055 for hydrogen, f N = 0. 096 for nitrogen, and f O = 0. 438
for oxygen. We determine the molar mass and the molecular formula. As
stated in the text of the problem, the unknown substance X and the solvent are
assumed to constitute an ideal binary mixture. We can thus use Raoult’s law
Eq. (3.108) to obtain the mole fraction of diethyl ether:
(3.131)
The mole fraction for X results from the condition
(3.132)
Thus,
(3.133)
Using the definition of the mole fraction in Eq. (2.8) at page 11 we obtain
(3.134)
where n X and n D are the amounts of X and diethyl ether respectively. We
can solve this equation for the value of X:
(3.135)
Here, we have expressed n D using the given mass of the solvent and its
molar mass, M D = 74 g mol−1, determined from the given molecular formula
of diethyl ether. If we combine
(3.136)
with Eq. (3.135), we obtain an expression for the molar mass of X:
(3.137)
The given mass fractions of the elements can now be exploited to determine
the molecular formula of the unknown substance: if Z C is the number of
carbon atoms in X and M C = 12 g mol−1 is the elemental atomic weight of
carbon, the following relation holds:
(3.138)
As a consequence, the number of carbons is
(3.139)
In the same way, we obtain
(3.140)
(3.141)
(3.142)
We therefore conclude that the molecular formula of the unknown compound
is C10H16N2O8.
Solution 3.14
According to the second law, a change of state in an isolated system is
spontaneous, if the total entropy change is positive. To apply the second law,
we calculate the entropy change involved with the freezing of supercooled
water and thereby show that this process is indeed spontaneous. This can be
quite tedious, because upon the process of freezing, the supercooled water
exchanges heat with the surroundings. Our analysis of entropy changes thus
has to include not only the entropy change of the water, but also the entropy
change of the surroundings:
(3.143)
Moreover, the application of the Clausius formula Eq. (3.33) requires
reversible heat transfers: for the calculation of Δ S, we therefore have to
replace the direct freezing at 250 K by (a) the heating of the supercooled
water to the standard temperature of fusion T f ⊖, (b) the reversible freezing
of the water at T f ⊖, and (c) the cooling of the frozen ice from T f ⊖ to 250
K, as outlined in Fig. 3.16. The property of entropy to be a state function
guarantees that the sum of entropy changes in the steps (a), (b), and (c) in
Fig. 3.16 corresponds to the entropy change of the entire irreversible
process, indicated as the dashed line in Fig. 3.16. For n = 1 mol, the entropy
change of water is
(3.144)
The first integral corresponds to the virtual entropy change of heating the
water from the temperature T 1 = 250 K to T f ⊖. The second term is the
entropy change involved with the reversible freezing of water at T f ⊖,
according to Eq. (3.102). The negative sign takes into account that the latent
heat of freezing is the negative of the given molar heat of fusion, Δ h f ⊖ =
6008 J mol−1. The third term, finally, gives the entropy change of cooling the
frozen ice from T f ⊖ to 250 K. After the evaluation of the integrals,
(3.145)
(3.146)
Fig. 3.16 Temperature dependence of the entropy of liquid water and ice as a function of temperature
(schematic). The entropy change involved with the freezing of supercooled water at 250 K (dashed
line) is the same as the alternative route (solid line) consisting of heating of liquid water to the standard
temperature of fusion at 273.15 K (a), reversible freezing (b), and cooling the ice down to 250 K (c)
As expected from the diagram in Fig. 3.16, Δ S water corresponding to the
difference in S water between point 4 and point 1 is negative. Next, we focus
on the entropy changes in the surroundings involved with steps a, b, and c
respectively. We presume that the surroundings constitute a huge heat
reservoir, so that arbitrary amounts of heat can be exchanged with the
surroundings without a change in its temperature, which is T 1 = 250 K. Then,
we can calculate the entropy changes of the surroundings, as we have done in
Problem 3.8 (Eq. (3.73)) and obtain
(3.147)
(3.148)
The first term corresponds to the entropy change of the surroundings, if
the heat n c p (liq.)(T f ⊖− T 1) necessary to heat the liquid water to T f ⊖ is
transferred from the surroundings to the water in step (a). The second term is
the gain of entropy if the latent heat released by the water upon freezing is
transferred to the surroundings in step (b). The third term, finally, is the gain
of entropy of the surroundings due to the heat released by the frozen ice when
it is cooled back to the lower temperature T 1. Thus, summing up all entropy
changes using Eq. (3.143), we obtain an increase in total entropy of Δ S =
+1. 885 J K−1. Using the second law, we conclude that the cooling of
supercooled water at 250 K is a spontaneous process.
Solution 3.15
Multi-phase systems necessarily have phase boundaries: surfaces. As the
chemical environment at an interface is generally different, the molecules
experience a different bonding at the interface compared with the bulk. As a
consequence, the total energy of a finite piece of matter also depends on its
surface area. The dependency of the thermodynamic potentials on the surface
area is considered via the surface energy γ per unit area, equivalent to a
surface tension . For larger systems, surface effects on the thermodynamic
properties can often be neglected. However, for objects on the nano scale,
such as the very small atmospheric water droplets considered in this
problem, the surface energy may influence, among other things, the ice
crystallization properties of such droplets. From the viewpoint of
thermodynamics, water crystallizes in the form that under given conditions
has the lower Gibbs free energy of formation. Presuming the same molar
entropy of cubic and hexagonal ice, water crystallizes as hexagonal ice if we
neglect the influence of the interface, because, as stated in the text of the
problem, its molar enthalpy is 35 J mol−1 smaller than that of cubic ice. Let
us consider a water droplet with a radius of r, a volume , a surface
area A(r) = 4π r 2, and mole number . M = 18 g mol−1 is the molar
mass of water, and ρ = 0. 93 g cm−3 is the density. The Gibbs free energy of a
droplet of hexagonal ice is
(3.150)
For a droplet of cubic ice of the same radius,
(3.151)
If the entropies of cubic and hexagonal ice are the same as stated above,
(3.152)
(3.153)
We use this expression to determine the critical radius at which cubic
and hexagonal water droplets of the same radius have the same Gibbs free
energy (G cubic(r crit.) − G hexagonal(r crit.) = 0):
(3.154)
(3.155)
(3.156)
The critical radius for a water droplet is only about 15 nm. If r > r crit.,
the Gibbs free energy of hexagonal ice is lower than that of cubic ice, and the
droplet crystallizes in the hexagonal crystal structure. For r < r crit.,
thermodynamics supports crystallization in the cubic form of ice.
References
1. Frenkel M, Marsh KN (eds) (2003) Virial coefficients of pure gases and mixtures. Landolt-
Börnstein, New Series, Group IV, vol 21. Springer, Heidelberg
Footnotes
1 The condition of the inflection point to have vanishing first and second derivatives of the function p(v)
yields the relations between the van der Waals parameters a and b on the one hand, and the critical
data p c , T c , and v c on the other hand.
2 Because there is always an equation of state that relates p, T, and V, it does not make sense to
assume U to be a function of all thermal state variables.
3 For the use of lower letters for molar quantities, see Sect. 3.1.
4 Thermodynamic equilibrium between two systems involves thermal, mechanical, and also chemical
equilibrium.
5 For a rather elementary analysis of the mixing of gases based on statistics, see Problem 8.2.
7 Note that we cannot use Poisson’s equation here, because the latter assumes a reversible adiabatic
change of state.
8 See the definition of the enthalpy of formation in Sect. 4.1.1 at page 71.
9 The other colligative properties are the elevation of the boiling point, the depression of the freezing
point, and osmotic pressure.
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_4
4. Thermochemistry
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Thermochemistry deals with the heat transferred or released by a system
during a change of its state or a chemical reaction. Calorimetry is an
experimental method for measuring such heat transfers. The reaction
enthalpy, the reaction entropy, and the Gibbs free energy of reaction are
defined and related to the molar standard enthalpies of formation and molar
standard entropies of reactants and products. The selection of problems in
this chapter deals with key aspects of thermochemistry, such as the
determination of molar heat of formation. Problem 4.3 exemplifies the use of
the Gibbs free energy of reaction as a criterion for the occurrence of
chemical processes.
Note that this definition implies that the standard enthalpy of the
formation of elements in their standard state is zero. In the same way, the
standard molar Gibbs free energy of formation of substances are defined and
tabulated at a reference temperature of 298.15 K. Tabulated standard molar
entropies, however, are absolute values.2
(4.1)
The molar reaction enthalpy is the change of H r with ξ:
(4.2)
Note that unlike H r , the molar reaction enthalpy Δ h r is an intensive
property (see Sect. 3.1). Tabulated heats of formation refer to the standard
state at the reference temperature 298.15 K. The standard molar reaction
enthalpy is thus
(4.3)
A process or chemical reaction characterized by Δ h r > 0 is called
endothermic . In contrast, a process or reaction characterized by Δ h r < 0 is
called exothermic . Similarly, the molar standard reaction entropy of this
reaction is calculated from the molar standard entropies of the reactants:
(4.4)
The molar standard Gibbs free energy of reaction is calculated from
the standard free enthalpies of formation of the reactants or the values of Δ h
⊖ and Δ s ⊖ at the reference temperature T :
r r r
(4.5)
A process or chemical reaction characterized by Δ g r > 0 is called
endergonic . In contrast, a process or reaction characterized by Δ g r < 0 is
called exergonic .
(4.6)
Similarly, using the relation , the reaction entropy at T is
(4.7)
Note that these equations assume that no phase transition occurs in the
temperature range under consideration between the reference temperatures T
r and T. In Problem 3.14 we dealt with the entropy change in a case in which
a phase transition occurs.
4.2 Problems
Additional problems related to thermochemistry can be found in Chap. 5
(chemical equilibrium).
b. Calculate Δ h x ⊖ at 550 K.
Fig. 4.1 Hydration of propene to isopropanol (2-propanol)
Solution 4.1
Combustion calorimetry offers the possibility of determining reaction
enthalpies indirectly using Hess’s law. In subproblem (a) we determine the
molar standard reaction enthalpy Δ h x ⊖ for the hydration of propene (Eq.
(4.8), Fig. 4.1) from the given molar heats of combustion of the hydrocarbons
propene and isopropanol. Technically, this reaction is conducted in a reactor
containing a suitable catalyst. A direct determination of Δ h x ⊖ is hampered
by the fact that parallel reactions occur, e.g., the reaction to 1-propanol.
Using Eq. (4.3) we write down an expression for Δ h x ⊖,
(4.9)
Next, we express the unknown molar enthalpies of formation occurring
on the right side using the given heats of combustion. Combustion means the
complete reaction of a substance with oxygen (see Problem 2.2 in Chap. 2).
If the reactant is a hydrocarbon, the combustion products are CO2 and H2O.
To exploit the heats of combustion provided for propene and isopropanol (2-
propanol), we must formulate the equation for their combustion. In the case
of propene, we have
(4.10)
For isopropanol, we have
(4.11)
Be aware that combustion reactions are exothermic (Δ h c ⊖ negative), and
the heat released to the surroundings is thus positive. Using Eq. (4.3), we can
set up two equations that relate the molar combustion enthalpies of these
substances to their formation enthalpies. Note that gaseous O2 is an element
in its standard state. Its enthalpy of formation is thus zero.
(4.12)
(4.13)
If we now subtract Eq. (4.13) from Eq. (4.12), we obtain
(4.14)
By comparing this expression with Eq. (4.9), we see that the reaction
enthalpy sought is simply the difference in the combustion enthalpies:
(4.15)
An alternative solution is based on the graphical representation of the
reactions Eqs. (4.8), (4.10), and (4.11) in a Born Haber cycle diagram , as
illustrated in Fig. 4.2. It is obvious that the reaction of propene and water can
be realized by the exothermic combustion of propene followed by the
endothermic synthesis of three CO2 and four H2O molecules to form
isopropanol. According to Hess’s law (see Sect. 4.1.4) the total molar
reaction enthalpy Δ h x is the sum of the molar reaction enthalpies in the
sequence of reaction steps, i.e., Δ h c (P) and −Δ h c (I), as indicated in the
figure.
Fig. 4.2 Born Haber cycle diagram for the hydration of propene (schematic). The reaction of propene
(C3H6) and water to isopropanol (C3H8O) is realized by the combustion of propene followed by the
synthesis of the combustion products to form isopropanol
Moreover, the inspection of the Born Haber cycle in Fig. 4.2 reveals a
typical difficulty of experimental calorimetry: accuracy. To obtain the
standard molar heat of reaction Δ h x ⊖ sought, we must subtract
comparatively large combustion enthalpies and the result Δ h x ⊖ is smaller
orders of magnitude. A precise measurement of heats of combustion is thus
the key to obtaining useful results. The goal is to achieve chemical accuracy
, which means that a heat of formation is determined with a maximum
uncertainty of 1 kcal mol−1.
The value for Δ h x ⊖ is valid for the reference temperature T r = 298. 15
K. In practice, the hydration of propene is conducted at a higher temperature.
In subproblem (b) we determine Δ h x ⊖ at T 2 = 550 K. This is an
application of Kirchhoff’s law (see Sect. 4.1.3). Moreover, this task is
simplified by the fact that the constant pressure molar heat capacities given
are assumed to be constant over the temperature range specified.3 Using Eq.
(4.6) under special consideration of reaction Eq. (4.8), the reaction enthalpy
sought is
(4.16)
Solution 4.2
This problem deals with a very simple case of calorimetry . LiCl is an ionic
crystal and completely dissolves in water according to
(4.17)
In subproblem (a) we determine the standard molar heat of solvation Δ h
⊖ of this reaction. We exploit the temperature change Δ T = +4. 2 K of the
solv
solution and the molar heat capacity of the solvent water given. We notice
that the increase in temperature indicates that the dissolution of LiCl is an
exothermic process. Apparently, we ignore any heat transfer to the
environment, i.e., the amount of heat
(4.18)
involved with the temperature jump in the solution is directly related to the
heat of reaction. The heat capacity C p of the solution is approximately given
by the heat capacity of water. The amount of water is determined using the
density ρ and the molar mass which is 18 g mol−1:
(4.19)
As a consequence, the heat released by the reaction was
(4.20)
To determine the molar heat of solvation, we need to know the amount of 1 g
LiCl. From the periodic system, we take the atomic weights of Li and Cl and
obtain the molar mass M LiCl = 42. 39 g mol−1. Hence, the amount of 1 g LiCl
is mol. The molar standard heat of solvation is thus
kJ mol−1.
In the same way, we can proceed in subproblem (b), where the same
experiment leads to a temperature reduction of 1.1 K for the salt KCl,
indicating an endothermic reaction. The molar mass of KCl is M KCl = 74. 55
g mol−1 and 1 g of KCl thus corresponds to an amount of n KCl = 0. 0134 mol.
A temperature reduction of 1.1 K is consistent with a heat loss of Q = −230. 0
J. The molar heat of solvation for KCl is thus kJ mol−1.
The second part of the subproblem deals with the determination of heats
of formation from the calorimetric results obtained so far. Although it is
principally not possible to determine the absolute heat of formation of an ion
in aqueous solution by performing an experiment,4 we can use Hess’ law and
eliminate the unknown heat of formation of chlorine, which is the anionic
species occurring in both reactions:
This is just the heat of formation sought relative to the value for Li+ ions.
Table 4.1 Thermochemical standard data for selected substances, valid for the reference temperature
298.15 K
Solution 4.3
According to thermodynamics, the stability of a compound against reaction
with other reactants is governed by the Gibbs free energy of reaction, which
depends on temperature (Eq. (4.5)). If Δ g r ⊖ is positive, the compound is
stable. A negative sign, in contrast, favors its reaction. In metallurgy, a
practical tool to estimate the stability of metals and their oxides against
reaction with carbon is the Ellingham diagram . Here, we take the example
of the reduction of nickel oxide to become acquainted with this method. The
nickel oxide in the vessel may be reduced by carbon in two ways, with
carbon monoxide or carbon dioxide as a gaseous reaction product.
In subproblem (a), we use the thermochemical data provided in
Table 4.1 to calculate the molar standard reaction enthalpies (Eq. (4.3)) and
the molar standard reaction entropies (Eq. (4.4)) for the reactions Eqs. (4.23)
and (4.24). For reaction Eq. (4.23), we obtain
(4.28)
where we have already taken into account that the molar standard heat of
formation of Ni(s) and C(s) is zero. Moreover, the molar standard reaction
entropy for this reaction is
(4.29)
Thus, at the reference temperature 298.15 K, the molar Gibbs free energy
of reaction is endergonic,
and thus, not favored by thermodynamics. If we now assume that these values
for the reaction entropy and the reaction enthalpy are constant, in good
approximation even for higher temperatures, we can look for the temperature
where Δ g 1 ⊖ changes its sign, i.e. the temperature above which the reaction
is exergonic :
(4.30)
In the same way in which we determine the molar reaction enthalpy and
entropy for the reaction Eq. (4.24), where NiO is reduced and carbon
monoxide is formed, we obtain:
(4.31)
and
(4.32)
This reaction is also endergonic at the reference temperature 298.15 K (Δ g 2
⊖ = +74. 4 kJ mol−1) and becomes exergonic above
(4.33)
To conclude, thermodynamics predicts that nickel oxide may be reduced by
carbon above T 1 = 441 K under the formation of CO2, and above T 2 = 703 K
under the production of CO.
In subproblem (b), we follow a second method to obtain the same result.
We consider the three oxidation reactions, namely the oxidation of nickel
(Eq. (4.25)), the oxidation of carbon to CO2 (Eq. (4.26)), and to CO (Eq.
(4.27)). For each of these reactions, we can again use the data in Table 4.1 to
determine the molar Gibbs free energy of reaction as a function of
temperature. These plots are straight lines, owing to the functional form Δ g r
⊖(T) = Δ h ⊖− T Δ s ⊖. They are shown in Fig. 4.4.
r r
Fig. 4.4 Ellingham diagram for oxidation reactions Eqs. (4.25), (4.26), and (4.27). The intersection
points P1 and P2 of the lines are at the temperatures T 1 and T 2 obtained in subproblem (a)
It is striking that the three lines have quite different slopes owing to the
very different reaction entropies. Gaseous species have a much higher
standard entropy than the solid elements and compounds, as can be seen in
Table 4.1. Moreover, the entropies of the three gaseous species are similar.
In a reaction that does not change the number of gaseous species, the reaction
entropy is thus small. This is the case for the dashed line in Fig. 4.4,
representing the oxidation of carbon to carbon dioxide. The oxidation of
carbon to carbon monoxide increases the number of gaseous species;
therefore, the reaction entropy is positive and the standard Gibbs free energy
of reaction has a negative slope (short dashed line). The oxidation of nickel,
on the other hand, reduces the number of gaseous species; Δ g r ⊖(T) has a
positive slope (solid line). The three lines have intersection points with each
other. Two of them are indicated as P1 and P2. The temperatures of the
intersection points, 441 and 703 K, match the temperatures T 1 and T 2 we
have determined in subproblem (a). How can this be explained? The idea
behind the Ellingham diagram is to compare the Gibbs free energies of
reaction of oxidation reactions. If the oxidation of a substance has a higher Δ
g r ⊖ than the oxidation reaction of another species, the former is reduced
rather than oxidized. In this sense, we can set up a sequence of reactions in
which NiO is reduced and carbon is oxidized:
(4.34)
and
(4.35)
The first reaction is reaction Eq. (4.23); the second is reaction Eq. (4.24)
with the stoichiometric numbers multiplied by two. It is obvious that the
conditions Δ g 1 ⊖ = 0 and Δ g 2 ⊖ = 0 are identical to the condition for the
intersection points, i.e., Δ g 4 ⊖ = Δ g 3 ⊖ for the reduction of nickel and
production of CO2, and Δ g 5 ⊖ = Δ g 3 ⊖ for the reduction of nickel and
production of CO. The sense of the Ellingham diagram becomes apparent if
the Gibbs free energies of reaction of a larger number of oxidation reactions
are plotted on the same diagram. A look at such a diagram then provides
quick information if oxidation or reduction is favored in the presence of
another element or its oxides. The weak temperature dependence of the
reaction enthalpy and the reaction entropy can be included in the diagram, in
addition to phase transitions that lead to a characteristic bending of the lines.
However, such diagrams only reflect the thermodynamics of these reactions,
not the kinetics.
Footnotes
1 For the definition of the standard state see Sect. 3.4.1
2 For the calculation of the absolute entropy of monatomic gases based on statistical thermodynamics,
see Problem 8.6.
3 A problem with temperature-dependent heat capacities can be found in Chap. 5 (Problem 5.6).
4 Electrolyte solutions are electrically neutral, requiring at least two different kinds of charged species
in a calorimetric experiment.
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_5
5. Chemical Equilibrium
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Chemical reactions are irreversible processes that reach a state of
equilibrium. Under well-defined conditions, this state of chemical
equilibrium of a system is characterized by a unique composition, defined by
the law of mass action.
Problems dealing with chemical equilibrium and the law of mass action
are among those topics that students consider to be difficult. A general
method based on the equilibrium extent of reaction is presented to tackle such
problems in a systematic way. The selection of problems highlights different
aspects of chemical equilibrium, such as equilibrium in parallel reactions,
equilibrium in open and closed systems, or equilibrium in dilute solutions.
(5.2)
0 eq
where n j is the initial amount of substance X j . Once ξ is known, all the
amounts of substances can be determined by the set of equations Eq. (5.2).
Four things are worth mentioning here. The determination of n j eq j = 1, 2, …
is not a multidimensional problem, as the quantities n j do not vary
independently. As already outlined in Chap. 2, the particle numbers strictly
follow the reaction Eq. (5.1). Thus, the seeming complexity of typical
textbook problems in chemical equilibrium is reduced in most cases to the
determination of only one number, namely ξ eq. Second, important for the
correct identification of ξ eq in many problems is the following inequality:
(5.3)
It is based on the fact that the extent of reaction is by definition a positive
mole number, and, as all n j eq are positive, ξ eq cannot become larger than
the smallest fraction , formed by the initial amounts of all reactants and
(5.4)
(5.5)
2
Here, μ j is the chemical potential of species X j . Depending on the
nature of the chemical system considered in a problem, the chemical
potential is best expressed in terms of partial pressure (Eq. (3.105)) in the
case of perfect gases participating in a reaction, in terms of activity (3.106)
in the case of real solutions, or in terms of concentrations, or mole fractions.
Then, insertion into Eq. (5.4) and application of the minimum condition
following from Eq. (5.5) leads to the law of mass action. Three
common forms of the law of mass action are given in the following
equations.
(5.6)
(5.7)
(5.8)
In these equations a j , p j , and c j denote activity, partial pressure, and
concentration in the state of chemical equilibrium; ν j is the stoichiometric
coefficient of substance X j . K is an equilibrium constant deduced from the
value of the standard molar Gibbs free energy of reaction Δ g r 0. The
standard state to which Δ g r 0 refers is indicated by the superscript 0. This
nomenclature takes into account that, especially for reactions in the liquid
phase, other standard states than that defined in Sect. 4.1.1 might be used. For
gas phase reactions, the standard state is related to standard pressure p ⊖.
The equilibrium constant K c is not directly related to fundamental
thermodynamic data. In problems dealing with gas phase reactions of perfect
gases, the concentrations are easily expressed by the partial pressure,
temperature, and the equation of state, . Thus, it is no difficult task to
get the value of K c from the general equilibrium constant K.
(5.9)
Textbook and examination problems almost always assume that the
standard molar heat of reaction is constant with temperature. This is an
approximation. Integration of Eq. (5.9) then yields the simple relation called
Van’t Hoff reaction isobar :
(5.10)
(5.11)
it is common to write the law of mass action in the following way:
(5.12)
Here, AH is an acid, A− its acetate ion, and it is assumed that the
concentration of water (55.55 mol dm−3) is much higher than the
concentrations of all other substances. In this case, can be considered
constant. By convention, the equilibrium constant and are then combined
and their product is the dissociation constant 3 K a . Moreover, it is common
to write
(5.13)
Here,
(5.14)
is the pH value of the solution,4
(5.15)
Equation (5.13) is also called Henderson-Hasselbalch equation .
5.2 Problems
Problem 5.1 (Br2 Decay)
(5.16)
Solution 5.1
Fig. 5.2 Scheme of a Br atom beam source operating at constant pressure and high temperature
(5.17)
(5.20)
This is consistent with the notion that at room temperature almost no Br2
molecule will decay spontaneously. If the temperature is raised to the much
higher temperature T 2, we can expect that the vibrational degrees of freedom
of the diatomic will be excited, involving an increased probability of
dissociation. We therefore expect a higher equilibrium constant at T 2. We
may assume that the molar heat of reaction is constant between T 1 and T 2.
Therefore, we can use the Van’t Hoff reaction isobar (Eq. (5.10)) and obtain
(5.21)
(5.22)
Therefore, the equilibrium constant at 1,500 K is K 2 = 0. 0551 and is thus
many orders of magnitude greater than the room temperature value. A note
here about calculus: an error often seen in examinations is the addition or
division of terms with incompatible physical units. In this case, we need to
pay attention to the correct insertion5 of the heat of reaction Δ h r ⊖ = 192. 9
kJ mol−1, which is 192,900 J mol−1. Having obtained the equilibrium constant
at 1500 K, we predict the mole fractions of bromine atoms and bromine
molecules at this temperature in subproblem (c). The total pressure is fixed
at p = 10 mbar. Taking this into account, we can introduce the mole fractions
and x Br and write
(5.23)
Insertion of these relations into the law of mass action (Eq. (5.17)) yields
(5.24)
which contains the two unknown mole fractions. To determine the latter, we
need additional equations. The strategy is, as outlined in Sect. 5.1, to
determine the equilibrium extent of reaction ξ eq and then to re-express the
mole fractions sought by ξ eq. We have (see Eq. (5.2))
(5.25)
where we exploit the information that initially only molecular bromine is
present in the vessel with an amount . Using the definition of the mole
fraction in Eq. (2.8) we can write
(5.26)
If we insert these expressions in Eq. (5.24), we obtain
(5.27)
The latter equation can be simplified using the third binomial formula
(Eq. (A.3)):
(5.28)
This is an equation with only one unknown, ξ eq. We can solve for ξ eq
and obtain
(5.29)
With the value of K 2 from subproblem (b) and p ⊖ = 100, 000 Pa the root
takes the value 0.7612. If we finally reinsert this result into the above
expressions for the mole fractions Eq. (5.26), we obtain
(5.30)
The partial pressure of atomic bromine is thus p Br = x Br p = 8. 6 mbar,
and the partial pressure of Br2 is 1.4 mbar, accordingly. Hence, 86% of the
particles in the vessel are atomic bromine, whereas only 14% of the particles
are Br2 molecules.
Subproblem (d) deals with the question how the mole fraction of atomic
bromine can be further increased. Concerning temperature, our results of
subproblem (b) provide a clear answer: if we further increase the
temperature, the equilibrium constant further increases because of the
endothermic nature of the dissociation reaction, and hence the fraction of
bromine atoms then increases further. But what about total pressure? Let us
check the results of subproblem (c): the fraction of bromine is increased if
the equilibrium extent of reaction ξ eq is increased. We therefore analyze Eq.
(5.29):
(5.31)
At the limit of vanishing total pressure all bromine molecules dissociate.
Reduction of total pressure thus increases the fraction of bromine atoms. As a
consequence, we should operate the effusion source for atomic bromine at a
high temperature and low total pressure.
(5.32)
(5.33)
(5.34)
a. Show that K 3 = K 1 × K 2.
b. Clean air normally contains 78% N2 and 21% O2. Calculate the
mole fraction and the concentration of NO and NO2 in clean air
(total pressure 100,000 Pa, 298.15 K), assuming conditions of
chemical equilibrium and K 1 = 1. 69 × 1012, K 2 = 4. 2 × 10−32.
Solution 5.2
This problem deals with nitrogen oxide, NO x . These molecules occur as
trace gases in the atmosphere. Although atmospheric processes, strictly
speaking, never reach a state of equilibrium, we may ask if the observed
abundance of these trace gases is roughly in accordance with the prediction
from the law of mass action. For simplicity, only the molecules NO and NO2
are considered. The reaction laws Eqs. (5.33) and (5.34) describe the
formation of these oxides from the elements N2 and O2 in their standard
states. Equation (5.32) describes the conversion of NO into NO2 and vice
versa. If we consider the chemical equilibrium for each of these parallel
reactions, there must be relations among the equilibrium constants. In
subproblem (a) we show that K 3 = K 1 × K 2. We write down the law of
mass action for each of these reactions and assume perfect gas behavior:
(5.35)
Therefore,
(5.36)
(5.37)
Hence, the laws of mass action can be expressed in terms of mole
fractions:
(5.38)
In the next step, we notice a considerable simplification of the problem if
we recognize that there is excess nitrogen and oxygen. Hence, their mole
fractions can thus be assumed to be constant. Therefore, and x NO are
best calculated using K 2 and K 3:
(5.39)
and
(5.40)
Here, we have used the relation K 3 = K 1 × K 2 shown in subproblem (a).
Finally, we calculate concentrations. Using the equation of state for a perfect
gas (Eq. (2.12)), the definition of concentration (Eq. (2.9)), and Eq. (5.37),
we obtain
(5.41)
and
(5.42)
The long-term concentration of NO2 in the USA at the beginning of the
twenty-first century6 is about 50 ppb, (1 ppb = 10−9). Hence, our result for
nitrogen dioxide underestimates the real concentration by a factor of 25.
(5.43)
The molar heats of reaction of A → B and A → C are −12, and −16 kJ
mol−1 respectively (see Fig. 5.3). For simplicity, assume that entropy
differences between the conformers are negligible. Ignore all possible
intermolecular interactions.
a.
Write down the laws of mass action for both reactions and
calculate the equilibrium constants both for room temperature
(T 1) and for T 2 = 700 K.
b. Derive expressions for the mole fractions of each the three
conformers in chemical equilibrium. Calculate x A, x B, and x C
for both temperatures T 1 and T 2. Under which condition is the
abundance of the intermediate conformer A insignificantly
small?
Solution 5.3
In Problem 5.2 we have dealt with a set of parallel reactions, in which
equilibrium abundances of all substances could be easily calculated because
there were excess amounts of several reactants. This led to a decoupling of
the problem. But how do we proceed in the general case of coupled parallel
reactions? In this problem, we shall consider three conformers A, B, and C
of one substance in chemical equilibrium. An example could be a
monosubstituted cyclohexane:
(5.44)
The molecule has two different stable chair conformations. In the axial
conformation, the substituent group X is oriented perpendicularly with regard
to the seat of the chair, whereas in the equatorial conformation, X is within
this plane. Both geometries have a slightly different total energy. A transition
from the axial into the equatorial conformation is caused by an internal
flipping of the molecule into its energetically less favorable boat
configuration. This brings the group X into the equatorial position. Then, by a
second flipping at the opposite end, the molecule comes back into the
equatorial chair position.
In subproblem (a), we write down the laws of mass action and calculate
equilibrium constants for both reactions. We use the amounts of substances n
A, n B, and n C to characterize the equilibrium:
(5.45)
If the molar entropies of A, B, and C are equal, the molar reaction
entropy is zero and thus the molar Gibbs free energies of reaction are Δ g 1 =
Δ h 1 = −12 kJ mol−1 and Δ g 2 = Δ h 2 = −16 kJ mol−1 respectively. The room
temperature equilibrium constants are thus:
(5.46)
and
(5.47)
Equilibrium constants at T 2 = 700 K are calculated using Eq. (5.10). We
obtain
(5.48)
and thus K 1(T 2) = 7. 86. In the same way, we obtain K 2(T 2) = 15. 63.
In subproblem (b), we seek formulas predicting the abundances of the
three conformers in chemical equilibrium. We assume that initially, all
molecules are present in the conformer A with an amount n A 0. Because A
undergoes two different types of chemical reactions, we generalize Eq. (5.2)
by introducing two different numbers, ξ 1 and ξ 2, representing the extent of
reaction in these two reaction channels. Hence, in chemical equilibrium, the
amounts of the conformers are:
(5.49)
Insertion of these expressions in Eq. (5.45) yields
(5.50)
These two equations can be rearranged in a system of equation
with A being a 2 × 2 matrix:
(5.51)
We use Cramer’s rule (see Appendix Sect. A.3.17) to solve this system.
By evaluating the determinant of the coefficient matrix A and the
determinants of A 1 and A 2 with the first and second columns replaced by the
’vector’ y,
(5.52)
(5.53)
(5.54)
we obtain the solutions
(5.55)
and
(5.56)
Insertion of these results into Eq. (5.50) yields the amounts of the three
conformers. Recognizing that the total amount of the substance is always n A
0, we can compute the mole fractions sought:
(5.57)
(5.58)
(5.59)
Using these formulae and the above calculated values for the equilibrium
constants K 1 and K 2, the mole fractions for both temperatures can be
determined. These are given in Table 5.1.
Table 5.1 Mole fractions of the three conformers A, B, and C in chemical equilibrium at 298 and at
700 K
Temperature (K) x A x B xC
298 0.0013 0.1659 0.8328
700 0.0408 0.3209 0.6382
(5.60)
is widely used to reduce the content of carbon monoxide in hydrogen
gas.
a.
Write down the law of mass action for the given reaction. From
thermochemical data found in Table 5.2, calculate the
equilibrium constants at 298.15 and at 400 K. Does a catalyst
influence the chemical equilibrium?
Table 5.2 Standard molar heats of formation and standard Gibbs free energy of formation of some
compounds at 298.15 K
Solution 5.4
Some examination problems for chemical equilibrium require the
consideration of extra constraints apart from the law of mass action and the
set of equations given in Eq. (5.2). Such an extra constraint may be the
request for a special composition of the system in the state of equilibrium. In
this case, the obvious task is to determine the initial composition of the
system. This problem offers an example in which we deal with the water gas
shift reaction Eq. (5.60), which is used in industrial scale production of high
purity syngas7 . In subproblem (a) we write down the law of mass action
using partial pressures:
(5.61)
Now, we determine the equilibrium constant from the thermochemical
data in Table 5.2. The standard molar heat of reaction is:
(5.62)
The standard molar Gibbs free energy of reaction is:
(5.63)
Hence, the equilibrium constant at the reference temperature T 1 = 298.
15 K is
(5.64)
The equilibrium constant at T 2 = 400 K is calculated using Van’t Hoff’s
reaction isobar (Eq. (5.10)):
(5.65)
Hence, at 400 K the equilibrium constant is K(T 2) = 1, 488 and thus
considerably smaller compared with the room temperature value. This is
consistent with the fact that the water gas shift reaction is exothermic. Like
many reactions of high technical relevance, the water gas shift reaction is
conducted on the surface of a catalyst . We answer the question, does a
catalyst influence the equilibrium, with a no. The Gibbs free energy of
reaction, from which the equilibrium constant is calculated, does not depend
on the reaction pathway. The presence and the nature of the catalyst only
influence reaction rates, and not the composition of the system in chemical
equilibrium.
In subproblem (b), we determine how much water vapor needs to be
added to a 0.1:0.9 mixture of CO and H2 to achieve a reduction of the CO
content to 1% as the equilibrium is established. We first define the initial
ratio
(5.66)
expressed by unknown initial amounts n CO 0 and , which are not given.
Do we need to know them? No, because, with regard to the composition, it
should not matter if we are dealing with 1 mol or with 100 t of gas mixture. In
the same way, we define the equilibrium ratio
(5.67)
Finally, we define a third ratio
(5.68)
It is the initial ratio of water vapor and hydrogen sought. Now, we introduce
the equilibrium extent of reaction ξ eq. Using Eq. (5.2) and recognizing that
initially there is no carbon dioxide present in the vessel, we write
(5.69)
We insert these relations into Eq. (5.67) and obtain
(5.70)
Solving for ξ, we obtain, after some rearrangements:
(5.71)
This equation relates the extent of the reaction to the quantities given and the
unknown . The next step is to exploit the law of mass action. We assume
perfect gas behavior and therefore a proportionality of the partial pressure
and amount of the substance. As a consequence,
(5.72)
(5.74)
About more than 10% water vapor has to be added to the gas mixture and
nearly all the water is converted to hydrogen gas, whereas carbon monoxide
is oxidized. Further analysis would show that greater purification, e.g.,
CO:H2 = 1: 999, would require a disproportionate addition of water (B = 0.
11, R x = 0. 18) and subsequent removal of the excess water.
(5.75)
a. Write down the law of mass action for the given reaction.
Calculate the equilibrium constants at 298.15 and 920 K from
the thermochemical data given in Table 5.3.
Solution 5.5
Here, we have another problem of chemical equilibrium that requires the
consideration of additional conditions. In this problem, the total pressure is
constant, whereas the volume may change as the reaction proceeds. As we
are dealing with a dissociation reaction where the number of molecules
increases, we expect that the volume increases by an upward movement of
the piston (see Fig. 5.4). Thus, work is done during this process.
(5.77)
The standard molar Gibbs free energy of reaction at this temperature is thus:
(5.78)
Therefore, the reaction is endothermic, and also endergonic at this
temperature, consistent with a very small equilibrium constant, which is:
(5.79)
The equilibrium constant at the higher temperature T 2 = 920 K is
calculated using Eq. (5.10) under the assumption that Δ h r ⊖ is constant
between T 1 and T 2:
(5.80)
Hence, the equilibrium constant at 920 K is K(T 2) = 3. 01 and the reaction is
exergonic. For the calculation of the partial pressures in subproblem (b) we
thus expect that most of the methanol is dissociated in the state of chemical
equilibrium, whereas the volume of the reaction vessel is essentially filled
with hydrogen and formaldehyde in equal amounts (see Fig. 5.4). Introducing
the partial pressures p F, p H, p M of formaldehyde, hydrogen, and methanol
respectively, we write down the law of mass action for the reaction:
(5.81)
Here, we have expressed partial pressures by the respective amounts of
substances using the perfect gas equation of state. Following the basic
procedure outlined above in Sect. 5.1, we write down the set of equations
describing the change of the amounts of reactants and products (Eq. (5.2)).
We take into account that at first only methanol, with an initial amount of n M
0, is present in the reactor, whereas the amounts of the products formaldehyde
and hydrogen are zero:
(5.82)
0 0
We calculate n M from the initial volume V and pressure p ⊖ using the
state of equation of a perfect gas:
(5.83)
We recognize that not only the amounts of methanol, formaldehyde, and
hydrogen depend on the extent of reaction, but also the volumes: because of
Dalton’s law of additivity of the partial pressures we can write:
(5.84)
Solving for V we obtain:
(5.85)
Now, we insert the last equation and Eq. (5.82) into the law of mass action
(Eq. (5.81)) and obtain:
(5.86)
We can solve this equation for the extent of reaction:
(5.87)
Using Eq. (5.85)we calculate the volume in the state of equilibrium and
obtain V = 1. 87 × 10−3 m3. At constant pressure the work is thus:
(5.88)
The partial pressure of hydrogen and formaldehyde is computed using the
equation of state and the equilibrium amounts of these substances:
(5.89)
Methanol has an equilibrium partial pressure of:
(5.90)
We can check our result by summing the partial pressures. Our results
confirm our expectation that most of the methanol is converted to
formaldehyde and hydrogen. Discussing the results in a seminar group, one of
the students was astonished about the fact that work is done in this reaction,
even though it is endothermic. Is this really a contradiction? What would you
have answered? The last question of the problem is about reversibility. The
process is irreversible. If a state of chemical equilibrium is reached, it
cannot be reversed without changing the external conditions.
Table 5.4 Polynomial representation of constant pressure molar heat capacities of various gases, valid
in the temperature range between 298 and 800 K
Molecule a −1 mol−1 ) a −2 −1 −3 −1 −4 −1
0 (J K 1 (J K mol ) a 2 (J K mol ) a 3 (J K mol )
CO(g) 31.08 − 1. 452 × 10−2 3. 1415 × 10−5 − 1. 4973 × 10−8
Molecule a −1 mol−1 ) a −2 −1 −3 −1 −4 −1
0 (J K 1 (J K mol ) a 2 (J K mol ) a 3 (J K mol )
CO2(g) 18.86 7. 937 × 10−2 − 6. 7834 × 10−5 2. 4426 × 10−8
H2(g) 22.66 4. 381 × 10−2 − 1. 0835 × 10−4 1. 1710 × 10−7
H2O(g) 33.80 − 0. 795 × 10−2 2. 8228 × 10−5 − 1. 3115 × 10−8
Solution 5.6
As stated in Sect. 5.1.2, the assumption of a constant reaction enthalpy can be
a crude approximation if larger temperature intervals are considered. In the
concrete case of the water-gas shift reaction , we check the difference
between an approximative and the more precise treatment. In subproblem
(a), the goal is the generalized form of the Van’t Hoff reaction isobar with a
temperature-dependent heat of reaction, Δ h r (T). Integration of Eq. (5.9)
yields:
(5.91)
According to Kirchhoff’s law (see Eq. (4.6), on page 73),
(5.92)
where Δ h r (T) is the value of the heat of reaction at the reference
temperature T 1, and Δ c p (τ) = ∑ i ν i c p; i (τ). The left side of Eq. (5.91) is
easily integrated. Inserting Kirchhoff’s law (Eq. (5.92)), we obtain:
(5.93)
Using a polynomial representation of the molar heat Δ c p (τ) = ∑ k Δ a k τ k
with Δ a k = ∑ i ν i a k; i , we can write:
(5.94)
Sorting the terms with regard to the dependence on T, we obtain:
(5.95)
The integrals are now evaluated using the integral table in the appendix
(Sect. A.3.7):
(5.96)
(5.97)
and, in the same way, Δ a 1 = 0. 14565 J K−2 mol−1,
Δ a 2 = −2. 35827 × 10−4
J K−3 mol−1, and Δ a 3 = 1. 69614 × 10−7 J K−4 mol−1.
Fig. 5.5 Equilibrium constant for the water gas shift reaction as a function of temperature. Note the
logarithmic scaling
Using Eq. (5.96), we can now calculate for arbitrary T 2. The results are
depicted in Fig. 5.5 (solid line). For comparison, the dashed line represents
the value of the equilibrium constant according to Eq. (5.10). In particular,
the corrected value of the equilibrium constant at 400 K is 1,433. The
uncorrected value, used in Problem 5.4, is 1,488. Therefore, the systematic
error in this case is about 4%. If the temperature intervals become larger, of
course, the deviations become more significant.
(5.99)
Determine the molar heat of reaction Δ h r and the reaction
enthalpy Δ s r of the reaction. You may assume that both Δ h r
and Δ s r are constant within the temperature range 498 to 623
K.
Solution 5.7
Fig. 5.6 Graphical representation of Eq. (5.99) in the temperature range between 625 and 500 K
(5.100)
and, moreover, the relation:
(5.101)
Hence,
(5.102)
Comparison with Eq. (5.99) reveals
(5.103)
and
(5.104)
In subproblem (b), we shall calculate the partial pressures of dimethyl ether,
water, and methanol in the state of chemical equilibrium at 523.15 K. Using
Eq. (5.99), we calculate the equilibrium constant for this temperature and
obtain K(523.15 K) = 10. 624. The following procedure is similar to our
solution of Problem 5.5b. If p D, p W, and p M are the equilibrium partial
pressures of dimethyl ether, water, and methanol respectively, the law of
mass action is
(5.105)
where n D, n W, and n M are the amounts of substances. Introducing the
equilibrium extent of reaction, ξ (we omit the superscript ’eq’), we can
write:
(5.106)
where n M 0 is the initial amount of methanol in the vessel. The molar mass of
methanol is M M = 31. 05 g mol−1 and its initial mass is m = 2 g. Thus, the
initial amount is mol. We insert Eq. (5.106) into Eq. (5.105)
and obtain the conditional equation for ξ:
(5.107)
Solving for ξ we obtain a quadratic equation:
(5.108)
with the two solutions
(5.109)
(5.111)
It is advisable to insert these partial pressures again in Eq. (5.105) to test if
our results are correct. The fraction is 10.621, in very good agreement
with the equilibrium constant K = 10. 624 on which our calculation was
based. The small deviation is a contaminant from round-off errors during the
calculation.
(5.112)
(5.113)
(5.114)
Solution 5.8
This simple application of the law of mass action deals with the acidity of
rain water. Even in clean air—free from atmospheric trace gases such as SO2
or NO x —carbon dioxide is present and is thus dissolved in rain water
forming carbonic acid (see Fig. 5.7). The equilibrium between gaseous CO2
and dissolved CO2 is described by the so-called Ostwald solubility
coefficient , which we assume to be K 0 = 1. A fraction of the dissolved CO2
reacts to carbonic acid, H2CO3. After deprotonation HCO3 − is formed (Eq.
(5.112)), which causes the production of hydronium ions and thus a decrease
in the pH value. For simplicity, we ignore the autoprotolysis of the solvent
water and, moreover, the fact that additional H3O+ ions are formed by the
reaction HCO3 −(aq) + H2O(l) ⇌ CO3 2−(aq) + H3O+(aq). This is because the
equilibrium constant for this reaction is orders of magnitude smaller than K 2.
We write down the law of mass action for the first reaction:
(5.115)
(5.116)
The dissociation of carbonic acid is characterized by the dissociation
constant K a2, which is defined above in Eq. (5.12) and is more commonly
used to describe the dissociation of acids in dilute solutions:
(5.117)
If we solve Eq. (5.115) for and insert this concentration into Eq.
(5.116), we obtain
(5.118)
Because no other acids are present (perfectly clean water) and
autoprotolysis is ignored, . The concentration of the solvent
water, which is largely in excess, is 55.555 mol l−1. Therefore,
insertion of Eq. (5.118) into Eq. (5.117) yields:
(5.119)
With the definition of the pH value (Eq. (5.14)), we obtain pH = 5. 6 for
clean air rain water. Hence, because of the presence of atmospheric natural
CO2, rain water is already acid. Recapitulating our method of solving this
problem it is worth mentioning the following: in this problem the abundance
of atmospheric CO2 takes a constant value regardless of how much CO2 is
dissolved in water. Thus, the initial concentration and the equilibrium
concentration of carbon dioxide are identical. The same is true for the
concentration of water. Therefore, the calculation of can be carried out
in a straightforward way. When deciding whether we have to balance
floating concentrations using the concept of the extent of reaction (Eq. (5.2))
or whether it is justified to assume constant concentrations, the following
question may also be helpful: is the system under consideration an open
system or a closed system? In the present problem concerning atmospheric
CO2 we have an open system—the carbon dioxide of the entire atmosphere is
in equilibrium with the dissolved carbon dioxide in a small rain drop. Now,
consider the more complicated case of a sealed bottle of mineral water
where the gas phase in the small volume between the water surface and the
crown seal is in equilibrium with the dissolved gas (see Fig. 5.8). In this
more complicated case, we have a closed system and we must distinguish
between initial concentration and equilibrium concentration.
Fig. 5.8 Chemical equilibrium between gaseous and dissolved CO2 in a sealed bottle of water
Solution 5.9
This problem again picks up on the problem of carbon dioxide solvation in
water (see Fig. 5.7). As stated in the discussion of the solution to
Problem 5.8, the solvation of a finite amount of CO2 in a closed system
requires a treatment using floating concentrations, which is the method we
must follow now. Moreover, it is a good idea to use a representation of the
laws of mass action (Eqs. (5.115)–(5.117)) involving amounts of substances.
With the volumes of the gas phase and the water, v g = 0. 05 l and v l = 0. 95
l, we can introduce new equilibrium constants κ 0, κ 1, and κ 3, and write:
(5.120)
(5.121)
and
(5.122)
where , l mol−1 × 55. 555 mol l−1 = 1.
667 × 10−3, and κ 2 = v l K a2 = 0. 95 l × 2. 5 × 10−4 mol l−1 = 2. 375 × 10−4
mol. The mole numbers in Eqs. (5.120)–(5.122) are understood to be
equilibrium mole numbers. These depend on the equilibrium extents of
reaction for
(5.127)
In the next step, we insert Eqs. (5.123)–(5.126) into Eqs. (5.120)–(5.122)
and obtain the following set of three equations for the three unknowns ξ 0, ξ 1,
and ξ 2:
(5.128)
The strategy is to eliminate two of the unknowns and obtain an equation for
one extent of reaction. Therefore, we take the first equation and solve for ξ 0:
(5.129)
Insertion into the middle expression of Eq. (5.128) yields
(5.130)
(5.131)
(5.132)
This expression can be written as a quadratic equation for ξ 2,
(5.133)
We obtain the roots of the quadratic equation using Eq. (A.4) and obtain
(5.134)
As C = 3. 75447 × 10−7 mol is positive, we can exclude the solution with the
minus sign in front of the root. Hence,
(5.135)
Having obtained this result, we use Eq. (5.131) and obtain ξ 1 = 4. 50471
× 10−5 mol. Insertion of this result into Eq. (5.129) then yields ξ 0 = 3. 83455
× 10−3 mol. Using Eqs. (5.123)–(5.126), we can now calculate all mole
numbers:
(5.136)
Solution 5.10
This problem is a typical application of chemical equilibrium in dilute
solutions (compare Sect. 5.1.3). Although strong acids such as HCl
dissociate completely, a weak acid such as acetic acid (CH3COOH) is only
partially dissociated. In TCA, the hydrogens of the methyl group are replaced
by the more electronegative chlorine, which affects the bonding in the
carboxyl group. As a consequence, TCA has a higher dissociation constant
than acetic acid. Given K a = 0. 3, we determine the equilibrium
concentrations and the degree of dissociation of TCA in subproblem (a)
according to the reaction:
(5.137)
Analogous to Eq. (5.12), we can write
(5.138)
We can assume that initially no H3O+ and TCA− ions are present in the
solution. Hence, . To determine the initial concentration of
TCA we have to consider its molar mass, M TCA = 163. 38 g mol−1. If m = 5 g
of TCA are dissolved in 1 dm3 water, then mol
dm−3. In the next step we have to set up the equations for the equilibrium
concentrations analogous to Eq. (5.2):
(5.139)
(5.140)
Note that the concentration of the solvent water is taken as a constant
contained in the dissociation constant K a . The concentration is
related to the equilibrium extent of reaction. Insertion of these equations into
Eq. (5.138) yields
(5.141)
If we solve for we obtain the quadratic equation
(5.142)
with the roots
(5.143)
(5.144)
(5.145)
Thus, more than 90% of the TCA is dissociated.
Moving to subproblem (b) we deal with the question of how the degree
of dissociation is affected by the initial concentration of TCA; more
precisely, we study the case α(c TCA 0) as c TCA 0 → 0. We have already
dealt with the analogous problem for a gas phase reaction in Problem 5.1. A
superficial inspection of Eq. (5.144) shows that and thus in
the limit c TCA 0 → 0. However, this does not mean that the degree of
dissociation reaches zero. We have to consider the
(5.146)
(5.147)
Thus, at the limit of infinite dilution of the acid, the degree of
dissociation reaches a value of 1, which means that all TCA molecules are
deprotonated. The limiting behavior of α(c TCA 0) is illustrated in Fig. 5.9.
Fig. 5.9 Degree of dissociation of trichloroacetic acid as a function of initial concentration. The
dashed lines indicate the situation treated in subproblem (a)
Footnotes
1 See also the discussion of the limiting reactant in Problem 2.3.
2 For the definition of the chemical potential see Eq. (3.100) at page 52.
3 Whether the dissociation constant or the equilibrium constant is given in a concrete problem can be
decided by a consideration of its physical dimension: although K in Eq. (5.11) is dimensionless, K a in
Eq. (5.12) has the dimension of a concentration.
4 Strictly speaking, the definition of the pH value is based on activities rather than on concentrations.
5 A frequently occurring error of novices is to add quantities with different units or, as in this case, to
misapply the factor 103 hidden in the unit kilojoule (kJ).
6 According to the United States Environmental Protection Agency (http://www.epa.gov).
6. Chemical Kinetics
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Reaction kinetics deals with the question how fast a chemical reaction
proceeds. From a mathematical point of view, problems in reaction kinetics
are formulated as rate equations, i.e., differential equations in time. In the
limit t → ∞ the results of reaction kinetics and thermodynamics (chemical
equilibrium) must coincide. We consider this in detail in Problem 6.4. At the
end of the chapter we deal with the fascinating field of autocatalysis and
oscillating chemical reactions.
(6.1)
where is the concentration of species J in a given volume V. Hence,
a reaction rate can be measured, for example, by observing the decay of the
concentration of an educt. But how does the reaction rate depend on the
conditions under which a reaction proceeds?
6.1.2 Reaction Rate Laws
Obviously, the reaction rate r depend on the educt concentrations. To classify
chemical kinetics, it is common to consider a rate law and to introduce
reaction orders 1:
(6.2)
Here, k is the rate constant , and α i is the order of the reaction with
regard to educt J i , the latter being present with the concentration [J i ]. The
sum of the specific orders α i is the integral order or simply the order of the
reaction. It is important to realize that the coefficients α i are not the
stoichiometric numbers of the reaction. Only in the special case of an
elementary reaction can the order be deduced by a simple inspection of the
reaction. The determination of the reaction orders, either experimentally, or
by deduction from a suitable reaction mechanism, is a frequent task of
reaction kinetics.
If we combine the rate law Eq. (6.2) with Eq. (6.1) for one of the educts
or products, we obtain a differential equation for the concentration of this
species as a function of time, which may be solved by integration observing
the initial conditions.
6.1.2.2 Molecularity
What kinds of elementary reactions are possible? Fortunately, there are
only three relevant types, differing in the number of educt molecules
participating in the elementary reaction step. This number is called the
molecularity of the reaction. Molecularity one is an unimolecular reaction
A ⟶ Products, which is a decay reaction. Molecularity two is called a
bimolecular reaction , either of type 2A ⟶ Products, or the somewhat more
complicated type A + B ⟶ Products. Molecularity three is called a
termolecular reaction , in its most general case A + B + C ⟶ Products. It is
obvious that the probability of a simultaneous impact of three molecules is
much smaller in comparison with a collision of only two molecules.
However, depending on the number of product molecules, the laws of energy
and momentum conservation sometimes require the presence of a third
collision element. An elementary reaction of molecularity four, however, is
so unlikely that we do not even have to consider it. Thus, a net chemical
reaction may be mapped to only these few relevant types of elementary steps,
and, of course, their variations.
(6.3)
A plot of lnr 0 against ln[J 1]0 should give a straight line with a slope α 1.
Thus, by means of systematic variation of the initial concentration of all
educts, all reaction orders can be determined.
(6.4)
E a is called the energy of activation , A is the pre-exponential factor .
A frequent task is to determine the activation energy and the pre-exponential
factor from experimental data. The determination of an activation energy
within a reaction mechanism provides information about the energy barrier
that has to be overcome for a certain reaction to occur. In theoretical terms,
the energy barrier depends on the reaction path on the potential energy
surface characterizing a reaction (see Fig. 6.1).
Fig. 6.1 Schematic illustration of a potential energy surface in one dimension for a hypothetical
reaction A + B → AB. E a is the activation energy as the height of the energy barrier relative to the
educts A and B. The maximum energy characterizes the transition state A B ‡ leading to bond
formation
(6.5)
For example, a consideration of the number of collisions in a gas of two
different species based on the kinetic theory of gases would predict2 in
Eq. (6.5).
6.2 Problems
Problem 6.1 (Reaction Order and Half Life)
a.
Consider the reaction A P with the rate constant k. The
initial concentration of the educt A is c A(0). If α > 1 is the
order of the reaction, show that the half-life is
(6.6)
Solution 6.1
The half-life of an educt is the time after which its concentration decays to
half of its initial value:
(6.7)
In subproblem (a), we shall prove the relationship between , the rate
constant, and the initial concentration c A(0) given by Eq. (6.6). This relation
is valid if the order of the reaction α is greater than 1. The rate law (Eq.
(6.2)) for this reaction is:
which, on the other hand, can be expressed by the change of c A with time
(Eq. (6.1)):
(6.8)
The separation of variables
and thus
Note that we could solve now for c A(t) and obtain the explicit
expression for the concentration of the educt as a function of time, always
provided that α > 1, as stated above. In the next step, we use the definition of
the half-life (Eq. (6.7)): after the time , the concentration c A has decayed
to half of its initial value. Inserting this in the last expression, we obtain:
By considering the natural logarithm and the rules for logarithms (see
appendix Sect. A.3.3), we can solve for α:
(6.9)
Hence, the sought reaction order is about . Inserting this result into Eq.
(6.7) together with one data pair enables us to solve for the rate constant k:
(6.10)
T (∘ C) 20 25 30 35 40 45
b.
A salve containing 10% by weight BPO was stored at a
temperature of 25 ∘C. Determine the mass fraction after a
storage period of 60 days.
Solution 6.2
As temperature has an enormous impact on reaction rates (and other
degradation processes), the room-temperature storage of pharmaceuticals
may be an important issue. Here, we consider the stability of a typical
radical former assuming first-order kinetics.
In subproblem (a), we determine the energy of activation and the pre-
exponential factor from a suitable plot of rate constants measured at different
temperatures. Starting from the Arrhenius law Eq. (6.4), we interpret
(6.12)
a. Write down the rate law of this reaction assuming that reaction
Eq. (6.12) is an elementary reaction.
b.
If the initial concentrations of methane and hydroxyl radicals
are 1 × 10−9 mol m−3 and 4 × 10−12 mol m−3, what is the half-
life of OH? Assume a rate constant of k = 4. 75 × 103 m3 mol−1
s−1.
Solution 6.3
This problem deals with the decay of methane and hydroxyl in a bimolecular
reaction according to Eq. (6.12). The exercise is instructive as it
demonstrates two different views of a classical problem in chemical
kinetics.
In subproblem (a), we write down the rate law for this elementary
reaction of type A + P → Products, which is the first order for methane, and
also first order for OH. Thus, the reaction rate is:
(6.13)
In subproblem (b), we shall determine the time t 50, after which the
concentration of hydroxyl decays to 50% of its initial value [OH]0 = 4 ×
10−12 mol m−3. As mentioned above, there are two different ways to solve
this problem. This is possible because there is an excess of the other
reactant, methane: [CH4]0 = 1 × 10−9 mol m−3. Note that after a complete
decay of OH, the concentration of methane is still 9. 96 × 10−10 mol m−3 ≈
[CH4]0. Under these conditions, we may treat the problem as a first-order
decay of OH with an effective rate constant:
(6.14)
(6.15)
The well-known textbook result for the half-life is
(6.16)
The second view of the problem is the exact treatment of the second
order reaction. Ready to use analytical expressions can be found in
textbooks, but it is instructive to recapitulate the details. If more than one
reactant is present, then the integration of the rate law is more complicated
than in the cases we have already dealt with in Problem 6.1. Focusing on the
extent of reaction ξ we can write
(6.17)
where V is the volume. The amounts of methane and hydroxyl are (see Eq. (2.
6))
(6.18)
and
(6.19)
Insertion of these two expressions into Eq. (6.17) yields after separation of
the variables and integration
(6.20)
Integration using Eq. (A.38) in the appendix yields
(6.21)
(6.22)
(6.23)
This equation is the basis for the determination of the half-life of the hydroxyl
radicals. After the time t 50, the concentrations of the reactants are
(6.24)
and4
(6.25)
Insertion of these expressions into Eq. (6.23) yields
(6.26)
(6.27)
and assume that both forward and backward reactions follow a first-
order kinetics with the rate constants k 1 and k 2 respectively.
b.
If n A 0 and n B 0 are the initial amounts of A and B in a state far
from chemical equilibrium, show that the extent of reaction is
given by
(6.28)
Solution 6.4
Fig. 6.3 Distortion and relaxation back into chemical equilibrium of two coexisting molecular
conformers A and B
(6.31)
Here, and are the equilibrium concentrations of A and B. On
the other hand, the latter are related by the law of mass action to the
equilibrium constant:
(6.32)
Comparison of the last two equations thus establishes the important
relation sought:
(6.33)
Similar relations between the rate constants of the forward and backward
reactions and K could also be derived for the general case where the reaction
orders differ from unity.
In subproblem (b), we shall investigate in detail how the system comes
to equilibrium. The goal is to integrate the coupled differential equations Eq.
(6.29) and (6.30). We are advised to introduce the extent of reaction ξ and to
prove Eq. (6.28). In doing so, we learn how to use the extent of reaction to
solve more complicated examples of chemical kinetics problems. Using Eq.
(2.6) we can express the amount of substances of A and B as a function of
time:
(6.34)
(6.35)
Within a constant volume we can set up a differential equation for n A
based on Eq. (6.29),
(6.36)
in which we insert the last two expressions. Because , we obtain
(6.37)
We separate the variables and integrate with the proper boundaries t = 0,
ξ(0) = 0, and ξ(t) at an arbitrary time t:
(6.38)
Using the integral Table in the appendix (Eq. (A.37)) we can solve the
left integral. After insertion of the integration boundaries and application of
Eq. (A.6) we obtain:
(6.39)
If this equation is solved for the extent of reaction ξ(t),
(6.40)
is obtained, which was demonstrated. Together with Eqs. (6.34) and (6.35),
this equation is the basis for calculating the amounts of both species A and B
at an arbitrary time starting from t = 0, and the respective initial values n A 0
and n B 0.
It is worth inspecting this expression in more detail. It is striking that
chemical equilibrium is established with an effective rate constant k 1 + k 2.
In those cases, however, where equilibrium largely favors either A or B, one
of the two rate constants can be ignored. Moreover, we briefly consider the
special case where n A 0 and n B 0 are already the equilibrium amounts of A
and B respectively. Because of n B 0 = Kn A 0 and , the prefactor in
Eq. (6.28) is zero in this case and therefore ξ = 0.
Fig. 6.4 Determination of the rate constants k 1 and k 2 for the forward and reverse reactions from
the relaxation dynamics back to chemical equilibrium. The excess amount of conformer A due to
perturbation is δ; the characteristic relaxation time is τ
where τ is the relaxation time defined as the time after which half of the
excess amount δ is decayed. From the figure, we estimate a relaxation time of
11,000 s. This is consistent with an effective rate constant
Using Eq. (6.33), we can substitute k 2 and obtain k 1 = 5. 7 × 10−5 s−1 for the
forward reaction. The rate constant of the backward reaction is thus:
(6.41)
(6.42)
(6.43)
c.
Based on Eq. (6.43), derive a relation for the reaction half-life.
Test your relation by considering the limits k 1 ≫ 2k 2[A]0 and
2k 2[A]0 ≫ k 1 for which the half-life should take the forms of
first-order and second-order reactions respectively.
Solution 6.5
Sometimes, there are different possible reaction channels for a substance to
decay. For example, the triphenylimidazolyl radical (TPI, see Fig. 6.5) can
react with the solvent chloroform in a pseudo first-order reaction via
hydrogen capture [1]
(6.44)
However, TPI can also undergo a dimerization in a second-order
reaction:
(6.45)
We thus have the reaction scheme outlined in Eqs. (6.41) and (6.42). The
objective is of course to find an analytic solution for the concentration of the
reactants as a function of time. The first step is to set up the rate equations,
which we do in subproblem (a). If we add up the reaction rates for the first-
order decay according to Eq. (6.41) on the one hand, and the second-order
reaction (6.42) on the other hand, we obtain
(6.46)
The rate equation has a nonlinear term. At first sight, an analytic solution
seems to be tedious. In subproblem (b), we simply prove that such an
expression exists. Equation (6.43) contains the reciprocal concentration.
Intuitively, it may thus be a good starting point to check whether the
introduction of the reciprocal concentration might simplify the problem:
(6.47)
In mathematical terms, our problem is a special case of the Bernoulli
differential equation (see Sect. A.3.20). For t = 0, . Expressions
for the solution of this differential equation are given in the appendix Sect.
A.3.18. If we introduce the integrating factor
(6.48)
This result is simply Eq. (6.43). If the rate constants and [A]0 had been
given, we would be able to calculate the concentration.
In subproblem (c), however, we deal with the reaction half-life
resulting from Eq. (6.43). We proceed in the same way as in Problem 6.1.
Following the definition for the half-life given there (Eq. (6.7)), we can write
Hence,
(6.49)
We check whether this expression is correct by testing the limiting cases
of k 1 ≫ 2k 2[A]0 and 2k 2[A]0 ≫ k 1. In the first case, the fraction
becomes a very small number and thus:
(6.50)
(6.51)
(6.52)
The reactant A is either continuously supplied to the system, or is in
excess; thus, a constant concentration [A]0 can be assumed: [A]0 = 1
mol dm−3, k 1 = 2. 8 dm3 mol−1 s−1, k 2 = 142 dm3 mol−1 s−1, and k 3 = 3.
9 × 10−3 s−1.
c.
Assume small variations of concentrations of B and C around
the respective stationary state concentrations, i.e., [B](t) = [B] s
+β(t) and [C](t) = [C] s +γ(t). Based on the rate equations from
subproblem (b), show that the differential equations for β and γ
are:
(6.53)
(6.54)
Solution 6.6
Chemical reactions during which the concentration of some reactants
oscillate with time are a fascinating phenomenon. Under special conditions,
such a reaction, in principle, does not reach chemical equilibrium. Instead,
the concentrations oscillate around their steady-state value. In this exercise,
we investigate these oscillations in detail. The reaction mechanism defined
by Eqs. (6.50)–(6.52) is the well-known Lotka-Volterra kinetic model .
In subproblem (a), we comment on the term autocatalytic reaction used,
e.g., for Eq. (6.50). A reaction is called autocatalytic if at least one reactant
is also a product. This is the case for Eq. (6.50), where species B in a
bimolecular reaction with reactant A is reproduced. As a consequence, the
number of B molecules would increase continuously if species A were held
constant. Moreover, Eq. (6.51) falls under the category autocatalytic reaction,
in which species C is reproduced.
The formulation of the rate equations is our first task in subproblem (b).
We have to consider that the concentration of B increases with rate constant k
1 in reaction Eq. (6.50),
and decreases with rate constant k 2 in reaction Eq.
(6.51). Application of Eqs. (6.1) and (6.2) for the special case of a
bimolecular reaction gives us:
(6.55)
In a similar way, we obtain:
(6.56)
where we consider the first-order decay of species C with rate constant k 3
and its production in reaction Eq. (6.51) with rate constant k 2. We note that
the concentration of product D continuously increases if the system is fed
with A. In that sense, a steady state, which we consider, is a dynamic
equilibrium characterized by constant concentrations of the intermediates,
[B] s and [C] s respectively. We find these steady-state concentrations
assuming zero reaction rates for B and C in Eqs. (6.55) and (6.56). Thus,
and
(6.57)
and
(6.58)
Inserting the given values for the rate constants and [A]0, we obtain [B] s
= 2. 75 × 10−5 mol dm−3 and [C] s = 0. 02 mol dm−3 as the stationary
concentrations of the reaction intermediates. It is worth mentioning that a
second steady state exists: [B] s = [C] s = 0. As soon as [B](t) becomes zero,
there is no further production of species C and the latter decays to zero
according to Eq. (6.52).
We continue with subproblem (c), where we seek to find nonstationary
solutions for our system. Analytic solutions, however, are difficult, because
the coupled equations contain nonlinear terms. Therefore, we restrict our
analysis to small perturbations β(t) and γ(t) of the nonzero stationary state
values. In this case, terms containing the product β γ can be ignored and the
differential equations become linearized. To derive these equations we insert
[B](t) = [B] s +β and [C](t) = [C] s +γ in Eq. (6.55) and obtain:
Insertion of Eqs. (6.57) and (6.58) proves that all but the last term on the
right-hand side cancel each other out. This proves Eq. (6.53). In a similar
manner, we proceed with Eq. (6.53), where we obtain:
Where, again, all but the last term cancel each other out, which proves Eq.
(6.54).
In subproblem (d) we deal with nonzero solutions of Eqs. (6.53) and
(6.54), which constitute a coupled linear system with constant coefficients.
The rigorous mathematical treatment of such a system is sketched in the
appendix or described in mathematics textbooks. Here, we take a more direct
way and take the derivative of Eq. (6.53)
(6.59)
with the angular frequency
(6.60)
Equation (6.59) has harmonic solutions of the form β(t) = β 0cos(ω t),
which is easily shown by insertion. Because the derivative of cos(ω t) is
−ωsin(ω t), we obtain the related harmonic expression for γ from Eq. (6.53):
(6.61)
Hence, the concentrations of the reaction intermediates oscillate
harmonically around their nonzero steady-state values. One period takes
s. There is a phase shift of 90∘ between the perturbations β and γ. The
oscillations and the phase shift are consequences of the coupling between the
concentrations of the two intermediates. According to Eq. (6.54) γ grows if
the concentration of species B is above its steady-state value. However, as
the concentration of species C exceeds its average value and thus γ reaches a
positive level, β starts to decay again and the concentration of species B thus
diminishes.
In subproblem (e), we plot the oscillating concentrations of the reaction
intermediates. At t = 0, [B] = 3. 75 × 10−5 mol dm−3, i.e., the amplitude β 0 of
the oscillation around the stationary state value [B] s is 10−5 mol dm−3. A plot
of the intermediate concentrations can be found in the left-hand diagram of
Fig. 6.7. In the right-hand diagram [C] is plotted against [B]. The shape of the
closed trajectory is elliptic, and its center is the stationary point. Let us
discuss what would happen if the amplitude β 0 became larger so that the
trajectory would cross the line [B] = 0. In this case, there would be no further
production of species C according to the autocatalytic reaction Eq. (6.51)
and [C], from that time on, would decay exponentially with the rate constant
k 3. On the other hand, [C] would reach a value of zero, and [B] would grow
from that time on.
Fig. 6.7 Concentration of the reaction intermediates B and C oscillating around their stationary state
values in harmonic approximation. Note the break and the different scaling of the ordinate axis. Left:
Concentrations as a function of time. Right: Elliptic trajectory around the stationary state point S
(6.62)
(6.63)
(6.64)
The reactant A is either continuously supplied to the system, or is in
excess; thus, a constant concentration [A]0 can be assumed: [A]0 = 1
mol dm−3, k 1 = 2. 8 dm3 mol−1 s−1, k 2 = 142 dm3 mol−1 s−1, and k 3 = 3.
9 × 10−3 s−1.
a. The rate equations and the stationary state of the given reaction
mechanism have been discussed already in Problem 6.6 (Eqs.
(6.55)–(6.58)). Show that the dynamics of the system conserves
the quantity
(6.65)
b. Use numeric methods to solve the rate Eqs. (6.55) and (6.56).
Assume [B](t = 0) = 3. 75 × 10−5 mol dm−3 and for [C] the
steady-state concentration. Plot the resulting concentrations of
the reaction intermediates [B] and [C] as a function of time.
Also, make a plot [C] against [B] and compare these results
with the curves obtained in Problem 6.6d. Determine the
quantity V (t) defined in subproblem (a) to test the numerical
stability of your calculation.
Using Eqs. (6.65), (6.55)–(6.56), and the basic differentiation rule Eq.
(A.23), we obtain
All terms on the right-hand side cancel each other out. Thus, V is constant
over time.
In subproblem (b), we seek numerical solutions for the rate equations
(6.66)
(6.67)
starting with the same initial concentrations as in Problem 6.6d. This gives us
the possibility of checking the validity of the harmonic approximation
assumed in that exercise. In mathematical terms, this is an initial value
problem. Nowadays, there are a number of mathematical software packages
available to students that tackle the numerical solution of initial value
problems. A minimum amount of coding is necessary to obtain numerical
solutions. Other students who are eager to gain a deeper insight in the
technical details of numerics use a programming language and code the
problem on their own. Following the latter case, we have to choose a
suitable numerical scheme and an adequate step length h = t n+1 − t n (n = 1,
2, …) to calculate the concentrations [B](t n+1) and [C](t n+1) from [B](t n )
and [C](t n ), starting from the initial concentrations, [B](t 0 = 0) and [C](t 0
= 0). Some popular methods with different accuracies can be found in the
appendix (Sect. A.3.21). The fourth order Runge-Kutta method , with a time
interval h = 10−3 s, yields the results shown in Fig. 6.8.
Fig. 6.8 Concentration of the reaction intermediates B and C oscillating around their stationary state
values. Note the break and the different scaling of the ordinate axis. Left: Concentrations as a function
of time. Right: Egg-shaped trajectory around the stationary state point S. The dashed line indicates the
approximative trajectory from Problem 6.6d
Fig. 6.9 Concentration of the reaction intermediates B and C oscillating around their stationary state
values. Note the break and the different scaling of the ordinate axis. Left: Concentrations as a function
of time. Right: Trajectory around the stationary state point S. Point P indicates the starting point of the
simulation
Looking back on our solution, we have used numerical methods to solve
the rate equations for a reaction mechanism featuring oscillatory behavior. It
is worth mentioning that in Chap. 10 we deal with the LASER5 in
Problems 10.10 and 10.11 in a similar manner. As it turns out, the
mathematical description of a lasing system, consisting of an electromagnetic
field within a resonator and an active medium in a dynamic equilibrium, is
similar to the description of the kinetic model of the Lotka-Volterra type.
Reference
1. Lavabre D, Pemienta V, Levy G, Micheau JC (1993) J Phys Chem 97:5321
[CrossRef]
Footnotes
1 Note that there are cases where the most general form r = k f([J 1], [J 2], …) is more appropriate.
2 According to kinetic theory (Eq. (7.3)), the collision rate between N A particles of a species A and N
4
According to Eq. (6.19), the extent of reaction at time t 50 is .
7. Kinetic Theory
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Kinetic theory provides the basis for macroscopic state variables, such as
pressure and temperature, by developing a microscopic picture of particles
with kinetic energy, momentum, and, moreover, internal degrees of freedom.
Thus, it follows an atomistic model. To treat a system of typically 1023
particles, kinetic theory introduces the concept of probability and statistics
into science. (The probability concept is, moreover, key to the description of
quantum mechanics. However, we must clearly distinguish between
thermodynamic probability introduced by Ludwig Boltzmann and quantum
mechanical probability resulting from Max Born’s interpretation of the wave
function (see Sect. 9.1.2 on page 216).) Among the probability density
functions, the Maxwell-Boltzmann distribution for the particle velocity is the
most prominent, and we deal with it in several problems. The characteristics
of a molecular beam leaving an effusion cell, in addition to film growth, are
further examples that demonstrate the practical use of kinetic theory.
(7.1)
where N tot is the total number of particles, and k B = 1. 3806488 × 10−23 J
K−1 is the Boltzmann constant (see Sect. A.1). This distribution follows from
very general assumptions, such as the homogeneity and the isotropy of the
system. Note that the Maxwell-Boltzmann distribution assumes thermal
equilibrium, which in many situations may not be established.
7.1.2 Pressure
The microscopic picture (Fig. 7.1) gives the relationship between the
velocity of the particles and pressure p. The assumption of elastic collisions
of particles with the system walls and momentum conservation yields:
(7.2)
where is the number of particles per volume, m is their mass, and 〈v 2〉
their mean square velocity resulting from Eq. (7.1).
Fig. 7.2 Collisions between particles in a mixture of perfect gases of types A and B, treated within the
hard sphere model. Collisions of the particle in the center occur with all particles within a cylindrical
volume with a cross-sectional area π d AB 2
(7.3)
where
(7.4)
is the effective mass . In the simplest case of elastic collisions of rigid
spherical particles with diameters d A and d B, the cross-sectional area is
that of a disk of radius (see Fig. 7.2):
(7.5)
Equation (7.3) is based on the summation of the z AB individual
collisions of N A particles of type A with type B. To calculate z AB, it is
tentatively assumed that all particles are at rest, except for one of type A
sweeping through the gas with the mean velocity 〈v A〉. A collision with a
particle of type B occurs if it is passed by the moving particle at a distance
less than or equal to d AB (see Fig. 7.2), i.e., within the cross-sectional area
σ AB. The latter defines the basal plane of the cylindrical volume σ AB〈v A〉,
which the moving particle passes per unit of time. Given the particle density
of type B particles, the moving particle collides with
particles of type B. The picture, however, is incorrect because all particles
are in permanent movement. Therefore, the mean velocity of particle A is
replaced by the mean relative velocity 〈v AB〉 of the particles. Therefore, the
collision frequency of one particle of type A with particles of type B is
(7.6)
The somewhat tedious calculation of 〈v AB〉 is treated in Problem 7.3.
The result is
(7.7)
From the collision frequency of a particle of type A, its mean free path
λ A between two collisions results:
(7.8)
Here z is the number of collisions the particle experiences per unit of
time.
(7.10)
7.2 Problems
Problem 7.1 (Maxwell-Boltzmann Distribution I)
(7.11)
c. Provide expressions for the average speed 〈v〉, the mean square
speed 〈v 2〉, and the velocity at which the distribution takes its
maximum value. Give explicit values of these quantities for
helium gas at 300 K and at 1,000 K.
Solution 7.1
In this problem, we familiarize ourselves with the Maxwell-Boltzmann
distribution function for the velocity of particles in a gas. The velocity of a
single particle is a vector. In the absence of conditions that would define a
preferred direction, however, the gas is an isotropic system. This must also
be reflected in the various forms of the Maxwell-Boltzmann distribution
using different sets of coordinates.
In subproblem (a), we give expressions for the distribution using
different choices of coordinates. In Cartesian coordinates, the velocity
squared is v 2 = v x 2 + v y 2 + v z 2, where v x , v y , and v z are the velocity in
x-, y-, and z-direction respectively. Hence,
(7.12)
Apparently, the distribution function is separable, i.e.,
(7.13)
In many problems, however, spherical coordinates (see Sect. A.3.11)
are the suitable choice. The velocity vector is characterized by its scalar
value v, and the direction expressed by a polar angle θ and the azimuthal
angle ϕ. We obtain:
(7.14)
Fig. 7.3 Maxwell-Boltzmann velocity distribution for helium near room temperature (solid line) and at
1,000 K (dashed line). As shown for T = 300 K, the maximum of the distribution v m , the average
speed 〈v〉, and the root mean square velocity take different values because the distribution is
asymmetric
The integral is evaluated using the integral table in the appendix (Eq.
(A.48)). We obtain:
(7.16)
(7.17)
At a temperature of 300 K, the mean square velocity of helium is thus 1.
87 × 106 m2 s−2; the root mean square velocity is 1368 m s−1. At a
(7.18)
This expression is related to the internal energy of a monatomic perfect
gas with its three translational degrees of freedom. Moreover, the result for
〈v 2〉 inserted into Eq. (7.2) yields the equation of state of the perfect gas : pV
= N tot k B T. Finally, we seek the most probable velocity, i.e., the maximum
of the Maxwell-Boltzmann distribution. The necessary condition for the
maximum is:
Insertion of Eq. (7.14) yields
(7.19)
At 300 K, the maximum value of helium is 1,117 m s−1; at 1,000 K v m =
2, 039 m s−1 is obtained. Our results for v m , 〈v〉, and are different for a
given temperature, which is also illustrated in Fig. 7.3 for T = 300 K. This is
caused by the asymmetry of the distribution functions.
(7.20)
where .
(7.21)
(7.22)
Solution 7.2
In this exercise, we deal with the kinetic energy distribution of a gas.
Classical mechanics relates the kinetic energy of a particle of mass m to its
velocity v by means of:
(7.23)
The derivation of the probability density distribution of the particles having a
kinetic energy between E and E + dE should thus be straightforward. In
subproblem (a), we start from the Maxwell-Boltzmann velocity distribution
(7.24)
We make a substitution in the exponent according to Eq. (7.23).
Moreover, we consider dE = mv dv = p dv and the relation
between momentum p and kinetic energy E. We obtain:
Therefore,
(7.25)
The last equation shows that the energy distribution is independent of the
particle mass and only depends on temperature, as expected for a perfect gas
(compare also Eq. (7.18) in Problem 7.1). The exponent suggests a special
scaling of energy in units of the thermal energy k B T at an arbitrary given
temperature. Thus, we define the dimensionless quantity .
Accordingly, we use dE = k B T d ε and obtain:
The last line proves Eq. (7.20). The resulting probability density function
(7.26)
Fig. 7.4 Probability density function for the kinetic energy ε of a perfect gas scaled to the thermal
energy according to Eq. (7.26)
(7.27)
which gives the fraction of molecules with an energy between zero and ε ∗.
We write the exponential function as a power series (see appendix Eq.
(A.55)) and obtain:
Then, the strategy is to make the integration piecewise for every member
of the sum:
We thus obtain:
(7.28)
which proves Eq. (7.21). Is this equation applicable? This depends on the
convergence of the series.
In subproblem (d), we check this by experience. We set up a simple
computer routine that, starting from the first term for n = 0, repeatedly
calculates the next member of the series and adds it to the sum, until a certain
convergence limit κ is reached. For a given index n with , the
next member of the sequence is easily calculated using the following method:
. Proceeding in this way avoids the repeated expensive
calculation of factorials n! . If we were to choose a convergence criterion of
κ = 10−9, convergence would be reached after the summation up to n = 12 for
ε ∗ = 1. For larger values of ε ∗ more terms would have to be added. For ε ∗ =
10 we would need to add 40 terms to reach convergence.
Fig. 7.5 Cumulative probability for particles with a kinetic energy below a threshold value ε ∗ (W c (0,
ε ∗), dashed curve), or above ε ∗ (W c (ε ∗, ∞), solid curve)
For values of ε ∗ up to 10, the results obtained with Eq. (7.28) are
depicted in Fig. 7.5 (red curve). Starting from zero, the W c (0, ε ∗) raises and
reaches a saturation value of 1, corresponding to the expected normalization
of the probability density distribution w(ε). The solid curve in Fig. 7.5
illustrates the reversed cumulative probability of finding a particle with
kinetic energy above the threshold value, W c (ε ∗, ∞) = 1 − W c (0, ε ∗). It is
unity for ε ∗ = 0 and reaches a saturation value of zero. It can be shown that W
∗
c (ε , ∞) is obtained using Eq. (7.22). The complimentary error function
erfc(x) = 1 −erf(x) used in this equation is implemented by standard
mathematical software packages. It is calculated numerically by a power
series (see Eq. (A.60) in the appendix). Using the function W c (ε ∗, ∞), we
can determine the sought probabilities of particles having energies above
numerous threshold values. These probabilities are summarized in Table 7.1.
More than 26% of the particles in a perfect gas have a kinetic energy that is
twice the thermal energy at equilibrium. The energy of nearly 2% of the
particles exceeds the thermal energy fivefold.
Table 7.1 Cumulative probability W c (ε ∗, ∞) of particles with energies above several multiples of the
thermal energy, ε ∗
ε∗ 2 3 4 5 10
(7.29)
Solution 7.3
This problem is related to interatomic or intermolecular collision rates
(Sect. 7.1.3, Eqs. (7.3) and (7.6) respectively). In subproblem (a), we use
classical mechanics and vector calculus to prove that the kinetic energy of
two particles A and B moving with the velocity vectors v A and v B can be
written as the sum of the kinetic energy of their center of mass (COM),
(7.30)
and a second kinetic energy term containing the relative velocity squared of
the two particles and an effective mass μ. The situation is illustrated in Fig.
7.6. The center of mass is located between the two particles. Its velocity
vector is obtained by differentiation of Eq. (7.30):
(7.31)
The total kinetic energy of the two particles, on the one hand, is the sum
of the individual kinetic energies of the particles, i.e.,
(7.32)
We could start from this last expression by ingeniously expanding both
fractions, but it seems easier to start from the right-hand side of Eq. (7.29):
The two terms containing the scalar product v A v B cancel each other
out. We obtain:
It is worth mentioning that Eq. (7.29), which we have just proven, is used
extensively in the treatment of two-body problems.1 Usually, the origin in
Fig. 7.6 in the center of mass is chosen and thus a separation from the kinetic
energy of the center of mass from the energy of the internal degrees of
freedom.
In subproblem (b) the mean relative velocity 〈v AB〉 is computed under
the assumption of the Maxwell-Boltzmann velocity distribution Eq. (7.1) for
both types of particles. The relative velocity is:
(7.33)
The expectation value is thus obtained:
(7.34)
where the triple integral over all Cartesian velocity components of the
particles A and B is from −∞ to + ∞ respectively. In the next step, we make a
transformation from the six Cartesian velocity components of particle A and
particle B to the three velocity components of the center of mass, v COM, x , v
COM, y , v COM, z , and the three relative velocity components v rel, x , v rel, y ,
and v rel, z :
(7.35)
(7.36)
The apparent advantage is that the root in Eq. (7.34) only contains the
three relative velocity components. Based on the results of subproblem (a),
the Boltzmann factors can be transformed in:
(7.37)
(7.38)
For the other components, we obtain the same results and thus the entire
Jacobian is unity. Therefore, we can separate the triple integral for the center
of mass velocity components and obtain:
(7.39)
The boundaries of the triple integrals are again −∞ and ∞ for each
component. The first triple integral for the center of mass components can be
further separated into Gaussian integrals (see appendix Sect. A.3.7) for the
Cartesian components:
Thus, we obtain
(7.40)
The triple integral for the relative velocity components can be separated
if we transform from Cartesian to spherical coordinates v AB, θ, and ϕ.
Because and dv AB, x dv AB, y dv AB, z = v AB 2 dv
AB dθdϕ
(7.41)
which is identical to Eq. (7.7).
Assume the perfect gas behavior of helium and xenon and calculate
both the individual collision rates and the mean free path of both
species for a 50:50 mixture at room temperature with a total pressure
of 1 × 10−8, 0.1, and 105 Pa. The van der Waals radii of helium and
xenon are 140 pm and 216 pm respectively.
Solution 7.4
Picking up the example of a gas mixture of helium and xenon, this exercise
gives us an idea of the collision rates occurring in a gas at different
pressures, and, moreover, the mean free paths. Such quantities are of
importance, for example, for the characterization of transport processes in
the gas, which depend significantly on pressure. If we assume perfect gas
behavior for both species, we can relate the particle density to the total
pressure by means of:
Note that both species are equally abundant and that their partial
pressures are thus half the total pressure. At p tot = 1 × 10−8 Pa we have the
conditions of ultrahigh vacuum (UHV). We obtain
m−3. For p tot 0.1 Pa, we have the conditions of
medium vacuum, where the particle densities take a value of 1. 22 × 1019
m−3. Under the conditions of atmospheric pressure, we have
m−3.
Equations for the collision rates and the mean free path were provided in
Sect. 7.1.3. Collisions may occur between helium and helium, xenon and
xenon, and helium and xenon. The cross-sections for these collisions can be
estimated using the hard sphere model (Fig. 7.2) and the given van der Waals
radii. If d AB is the sum of the radii of two collision partners A and B, then
the hard sphere collision cross-section is σ AB = π d AB 2. Thus,
where is the particle density of the collision partner. Inserting the cross-
sections, mean relative velocities and particle densities, we obtain z He He =
5. 33 × 10−4 s−1 under ultrahigh vacuum conditions, 5. 33 × 103 s−1 under
medium vacuum, and 5. 33 × 109 s−1 at atmospheric pressure. Similarly, for
collisions among xenon atoms, we have z Xe Xe = 2. 22 × 10−4 s−1, 2. 22 × 103
s−1, and 2. 22 × 109 s−1 for UHV, medium vacuum, and atmospheric pressure
respectively. For collisions of helium and xenon we obtain z He Xe = 6. 19 ×
10−4 s−1, 6. 19 × 103 s−1, and 6. 19 × 109 s−1 for UHV, medium vacuum, and
atmospheric pressure respectively.
Having obtained these collision rates, we can determine the mean free
path for both species under the various conditions according to Eq. (7.8):
(7.42)
(7.43)
The mean velocities of helium and xenon are calculated using Eq. (7.16) and
the respective atomic masses: 〈v He〉 = 1256 m s−1, 〈v Xe〉 = 219 m s−1. The
resulting values for λ He and λ Xe at various pressures can be found in
Table 7.2.
Table 7.2 Mean free path of helium and xenon in a 50:50 mixture at room temperature and various
pressure conditions
(7.44)
is used to characterize the flow within vacuum components of extension l by
comparison with the mean free path. Under ambient pressure conditions (105
Pa), the mean free path is within the submicrometer range. Under such
conditions, collisions between particles markedly influence spectroscopic
line profiles (collisional broadening) . Another application is related to the
above-mentioned transport processes, e.g., heat conduction . To be efficient,
heat conduction requires many collisions between gas particles. In an
ultrahigh vacuum where the mean free path exceeds the typical extensions of
vacuum equipment by orders of magnitude, heat conduction mediated by
particle collisions is thus largely suppressed. Under these conditions, heat
transfer based on radiation is the only effective process. In Problem 9.3 this
situation is considered in more detail.
Solution 7.5
In this exercise, we work out the principle of a simple method of molecular
beam formation: effusion. The so-called Knudsen effusion cells are widely
used in molecular beam epitaxy to grow thin films on surfaces or to
investigate molecular beam reactive scattering in surface catalysis.
In subproblem (a), our aim is to establish the velocity distribution of the
particles emitted through the orifice. Equation (7.45) is not identical to the
Maxwell-Boltzmann velocity distribution, as would perhaps be expected at
first sight. To see this, we have to recapitulate the derivation of the total
impingement rate on a surface area element dA, which is part of the orifice
area. All particles impinging on this area leave the compartment if the shutter
is opened. Because the mean free path is larger than the orifice extension, we
can assume that there will be no collisions among the particles. Hence, a
particle with velocity between v and v + dv coming from a direction
characterized by a polar angle between θ and θ + d θ and an azimuthal angle
between ϕ and ϕ + d ϕ leaves the orifice with only these values. Using results
from Problem 7.1, we use the Maxwell-Boltzmann distribution in spherical
coordinates:
(7.46)
This gives us the effective number of particles with only this particular
velocity and direction. If we put the origin in the center of the surface
element dA as illustrated in Fig. 7.8, we can narrow down the number of
particles that hit the surface element within the time dt. First, the polar angle
θ must take values between 0 (perpendicular impingement) and (extreme
grazing incidence). Second, dependent on its velocity, the distance a particle
moves within a time interval between t and t + dt is v dt. Hence, a particle
with this speed only hits the surface element if it is within the prism volume
shown in Fig. 7.8. Its volume is dV = vcos(θ) dt dA. The number of particles
with speed v and direction within this prism is therefore:
(7.47)
which is in accordance with Eq. (7.9). If we omit in the last step the
integration over the velocity, we obtain the impingement rate of all particles
with velocities between v and v + dv:
(7.48)
Fig. 7.8 Derivation of the number of particles with velocity between v and v + dv impinging on a
surface area element dA within the time interval between t and t + dt from a polar angle between θ and
θ + d θ and an azimuthal angle between ϕ and ϕ + d ϕ
Thus,
(7.50)
We now scale the energy to the thermal energy introducing . With
dE = k B T d ε we obtain the simple expression
(7.51)
Apparently, the kinetic energy distribution of molecular beam effusing
through the orifice only depends on the temperature, and not on the particle
mass. It is instructive to compare w(ε) with the corresponding function from
the Maxwell-Boltzmann distribution (Eq. (7.20)) from Problem 7.2, i.e., with
the kinetic energy distribution within the compartment.
Fig. 7.9 Comparison between the kinetic energy distribution function of a molecular beam leaving the
effusion cell (solid line) and the kinetic energy distribution within the compartment (dashed line, see
Problem 7.2)
(7.52)
The negative sign compensates for the fact that an increase in velocity by
a positive dv involves a negative dt. Defining a parameter , we use
the last expression to obtain the rate at which particles arrive at the detector
at a given time t:
(7.53)
The detector counts the incoming particles per period of time; thus, we
assume that the detector signal is proportional to this rate.2 Thus, the detector
signal, S(t), is:
(7.54)
Fig. 7.10 Time of flight distribution of a group of particles at different distances from the detector
where c is a constant. S(t) is plotted in Fig. 7.10 for the case of neon
(atomic mass m = 20. 18 amu), a temperature of 400 K, and three different
distances. With increasing distance, the TOF profile becomes broader and its
maximum moves, as expected, to longer arrival times. We could look up the
respective peak values or obtain an expression for the position t max of the
TOF peak at which the first derivative of S(t) becomes zero:
(7.55)
We thus obtain:
(7.56)
With α = 574. 12 for the given conditions, we thus locate the TOF peaks
at 0.55 ms for d = 0. 5 m, 1.10 ms for d = 1. 0 m, and 1.65 ms for d = 1. 5 m,
corresponding to a velocity of 909 m s−1. Finally, it is instructive to check
whether gravitational effects have a significant influence on the trajectories
of the particles between shutter and detector. Given the gravity acceleration
of g = 9. 81 m s−2, the vertical shift after the time t is given by . If we
insert the above values for the TOF, we obtain a vertical shift of 1. 5 × 10−6
m at distance of 0.5 m, 5. 9 × 10−6 m at d = 1 m, and 1. 3 × 10−5 m at d = 1. 5
m. Gravitational effects are therefore negligible under these conditions.
Fig. 7.11 Water molecules impinging on a surface form a film of amorphous ice
Solution 7.6
In this exercise, we deal with the growth of an amorphous ice film on a cold
surface. Within kinetic theory, it is possible to include time in considerations
of changes of state—a point that is excluded in an analysis based on
thermodynamics in Chap. 3 The goal is to calculate the time t to grow an ice
film 1 μm thick on a cold surface, given the gas phase partial pressure p and
temperature T. The problem is a simple application of the relation of the
impingement rate Eq. (7.10), which gives the relation between the number of
particles impinging on a surface per second and per m 2, and the gas
temperature and partial pressure of the particles. We may assume that each
particle that hits the surface sticks to it—an assumption that in the case of ice
is valid at cryogenic temperature below 100 K. Under these conditions H2O
freezes as amorphous ice (see Fig. 7.11). At first, we calculate the number of
H2O molecules forming a d = 1 μm thick ice film with an area of, say, A = 1
m2. Because the density is
(7.57)
this film has the mass m = ρ A d, corresponding to an amount
(7.58)
and thus a particle number of
(7.59)
where is the molar mass of water and N A is the Avogadro
constant. Given an impingement rate
(7.60)
the time sought is
(7.61)
With a mass kg of the water molecule and ρ = 940 kg m−3
we obtain the result t = 87, 712 s, i.e., under the given conditions a film
thickness of 1 μm is reached after more than 24 h.
Footnotes
1 Text book examples of two-body problems are, for example, the hydrogen problem or the rigid
rotator.
2 In practice, the detector sensitivity is a function of the velocity, with the tendency to decrease as v
increases.
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_8
8. Statistical Thermodynamics
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Using the concept of statistics and probability, statistical thermodynamics
relates the state variables of macroscopic systems to the properties of its
microscopic constituents. The state variables can be attributed to one single
quantity: the partition function of the system.
The selection of problems in this chapter focuses primarily on the basic
principles of probability calculus, as it is the prerequisite for a deeper
understanding of statistical thermodynamics. The correct treatment of
factorials (see Sect. A.3.9 in the appendix) is a frequently occurring technical
issue that is addressed, for example, in Problem 8.3. Using the probability
calculus, we see how the Boltzmann distribution can be motivated and how
the phenomenon of diffusion can be understood. A second focus is on
problems dealing with concrete cases of partition functions and their relation
to thermodynamic quantities such as entropy or molar heat capacities.
(8.2)
called statistical weight is the number of ways in which a macrostate can be
realized, N is the number of particles, and N i is the number of particles in the
ith range of possible position and momentum.
(8.3)
with the partition function
(8.4)
The goal is to determine macroscopic state variables such as internal
energy U, the heat capacity C V , or its entropy S from the system partition
function Q.
For the statistical treatment of the system the ensemble concept is key, which
was introduced by J.W. Gibbs. An ensemble is a collection of identical
copies of the system. The ensemble concept allows the treatment of systems
comprising only a few molecules, or in an extreme case, only one single
molecule. Moreover, by choosing a large number of copies, the
application of Stirling’s approximation to handle factorials is justified. If
temperature T, volume V, and particle number N are constant in each of these
copies, then the ensemble is called a canonical ensemble, and the partition
function is called canonical partition function . Such a canonical ensemble
is illustrated in Fig. 8.1 for small . The diathermic walls in the ensemble
guarantee constant temperature in all copies of the system, but the energy
fluctuates around the most probable energy value. At very large the
probability distribution has a very sharp peak. If q is the partition function of
one copy of the system, then the partition function Q is obtained as a product
over all copies. Depending on whether or not these copies are
distinguishable, the partition function is
(8.5)
(8.6)
The classification into distinguishable and indistinguishable units is
frequently based on quantum mechanics in which the indistinguishability of
identical particles is established. It has been pointed out, however, that in his
early analysis of the Gibbs paradox , Gibbs argued on the basis of
operational distinguishability on purely classical grounds [1]. It can be
shown that the internal energy of the system, expressed using the canonical
partition function, is
(8.7)
The entropy is
(8.8)
(8.11)
The quantum mechanical result for the translational part of the partition
function obtained from the particle in a box model is, as textbooks show,
(8.12)
where h is the Planck constant (see Sect. A.1), Λ is a thermal wavelength ,
and V is the system volume.
The rotational partition function of a heteronuclear diatomic is
(8.13)
where J is the rotational quantum number and B is the rotational constant (see
Sect. 10.1.2). Based on the harmonic oscillator model (see Sect. 10.1.3), the
vibrational partition function of the diatomic takes a compact form:
(8.14)
Here, ν is the vibrational frequency of the stretch vibration. In the case of
polyatomic molecules, or in all cases beyond the harmonic oscillator model,
the expressions are more complicated. Moreover, in symmetric molecules
like H2 or CO2, nuclear spin statistics must also be considered when the
molecular partition function is determined.
8.3 Problems
A further problem related to statistical thermodynamics is Problem 9.1 at
page 227.
Solution 8.1
Fig. 8.2 Myoglobin protein structure
Thus,
(8.16)
An astonishing fact about the molecule is that under cellular conditions the
molecule only occurs in a few well-defined conformational structures
characterized by a few special sequences out of 6150. For simplicity, we
assume that there is only one such functional form of the molecule. If the
molecule had to find this unique structure in a random-walk process where
each conformational change takes only a few picoseconds, this process
would take a very long time. In contrast, real folding times are less than 1 s
[2]. This discrepancy is called Levinthal paradox . Moreover, if the folding
process were driven by the minimization of the Gibbs free energy Δ G = Δ H
− T Δ S of the molecule, then this unique functional form would have to be
characterized by a steep minimum in enthalpy that balances the large entropic
contribution − T Δ S to Δ G, where Δ S is the conformational entropy. It is
our task to calculate this entropy using Boltzmann’s statistical approach to
entropy. We assume two macrostates: (1) the functional state, which is
realized by W 1 = 1 configuration of the molecule, and (2) the dysfunctional
state, realized by all remaining W 2 = N − 1 conformational sequences. As N
is so large, we can safely assume W 2 = N. Thus,
(8.17)
Using N A k B = R = 8. 3145 J K−1 mol−1 and lnW 1 = ln1 = 0, we obtain
Fig. 8.3 Two compartments filled with different gases, separated by a baffle
Solution 8.2
This exercise gives a rather elementary demonstration of the mixing of gases
and how it is related to the basic principles of probability.3 The initially
separated gases in the two compartments I and II comprise only four
particles. This enables us to write down all possibilities regarding how the
particles can be distributed in the compartments after the baffle is removed.
In our solution, we use the following notation to indicate a special
configuration of particles: we use the number 1 to indicate when a particle is
found in compartment I, and the number 0 if it is in compartment II. The
initial configuration as shown in Fig. 8.3 is:
i.e., the first four numbers indicate the positions of the species B (the dark
balls) with label 1, 2, 3, and 4. The next four numbers give the positions of
the particles 1, 2, 3, and 4 of species G (white balls).
In subproblem (a), we give the probability that all particles, no matter
whether they belong to species G or species B, are found in compartment I.
This is the configuration:
In a perfect gas, the particles do not interact; thus, their movements are
independent, and the probability of finding one special particle, say B1 (dark
ball with label 1), is . Either it is found in compartment I, or, with the
same probability , it is found in compartment II. As the particles move
independently, the presence of particles in the compartment does not
influence the probability of finding the other particles in the compartment.
All particles have the same probability , and the joint probability for this
configuration is thus:
(8.18)
The probability is small, but not too small. Statistically, 1 in 256
measurements would yield this configuration. However, if the total number of
particles were increased to 2N = 100, the probability of finding all the
particles in compartment I would be within the range 10−30. For realistic
particle numbers N ≈ 1023, the probability of this event is de facto zero.
In subproblem (b), we give the probability of finding all particles of
species B in compartment I and all particles of species G in compartment II.
This event is in fact the initial configuration. It has the same probability as
any of the possible 22N = 256 configurations. The probability is, therefore:
(8.19)
Thus, the event of a perfect separation of the gases is rare; when limited to
realistic particle numbers, it is unimaginably small.
In subproblem (c), we generate a probability distribution of finding n
particles of species B in compartment 1, where n is 0, 1, 2, 3, or 4. These
events represent in general several possible configurations. We can write
them down explicitly:
(8.20)
Fig. 8.4 Probability distribution for the event of finding n particles of species B in compartment I
(8.21)
With this relation, it is straightforward to generate the second probability
distribution of finding n particles of whatever species in compartment I. With
the total number of 2N = 8 particles, the probabilities are:
(8.22)
The resulting distribution is shown in Fig. 8.5.
Fig. 8.5 Probability distribution for the event of finding n particles of whatever species in compartment
I
(8.23)
(8.24)
and give an interpretation of this result. Hint: Use Stirling’s
approximation (see Eq. (A.63))
and the Taylor expansion Eq. (A.56).
Solution 8.3
Fig. 8.6 Discrete movement of a molecule ”M” in one dimension. After an eight-step hopping
sequence, LRRLLLRL, the molecule is situated at x = −2a with a being the step length
for this special 8-step sequence is . For a special N-step sequence, the
joint probability is . This special sequence puts the molecule in its final
position at x = −2a. However, this is not the only sequence that produces this
result. For example, the sequence RLRLLLRL would also put the molecule in
this final position. Clearly, the total number of steps to the right (R) and steps
to the left (L) is decisive with regard to the final position. If there are n steps
to the right, then we have N − n moves to the left. Among the 2 N different
sequences with N steps, there are
(8.25)
sequences with n steps to the right. We have to add up the probabilities of
each sequence among this subset and obtain the probability that the molecules
move n steps to the right:
(8.26)
In the next step, we have to analyze at which position the molecule is placed
after a N-step sequence with n steps to the right and N − n steps to the left.
Each step to the right increases the position index m by one, while m is
reduced by one in a step to the left. As a consequence, after N steps, the
position index is
(8.27)
Hence, the number of steps to the right leading to the special position index m
is:
(8.28)
If we insert this result into Eq. (8.26) we obtain:
(8.29)
which is simply the expression Eq. (8.23). It is worth inspecting Eq. (8.27)
again: if we have an even total number of steps (N = 2K, K = 0, 1, 2, …), then
m = 2n − 2K = 2(n − K) will also be even. Therefore, an even total number of
steps places the molecule at an even position, m = 0, ±2, ±4, …. Conversely,
if the total number of steps is odd, then m will also be odd and the molecule
reaches m = ±1, ±3, …. This notion is important for the correct interpretation
of the results in subproblem (c), but also for the plotting of the probabilities
in subproblem (b) for N = 4, 5, 8. For N = 4, the possible end points are m =
±2, ±4. Evaluation of the associated probabilities using Eq. (8.23) yields:
Moreover, we have:
In the same way, we can calculate the possible values for the requested
distributions P 5 and P 8. The results are illustrated in Fig. 8.7. Inspection of
the plots reveals that the probability of finding the molecule in the center
(low values of m) is highest, whereas high absolute values of m have
decreasing probability. Moreover, the bell-shaped profile of the distribution
is especially apparent in the case of P 8. This suggests that for a large total
number of steps, the distribution could take the Gaussian shape of a normal
distribution as illustrated in the appendix in Fig. A.2. This limiting case is
examined in subproblem (c). We prove that Eq. (8.23) is identical to Eq.
(8.24) when limited to high N. This task is instructive because we learn how
to cope with factorials using Stirling’s approximation . This sort of calculus
is typical for problems occurring in statistical thermodynamics. For large N,
the factorial N! is5
(8.30)
and thus
(8.31)
Using Eq. (8.31) and the rules for logarithms in the appendix (Sect.
A.3.3), we write
The three highlighted terms cancel each other out. Further evaluation yields:
Highlighted terms cancel each other out. Further simplification using the
rules for logarithms is not possible. However, the terms can be
expanded in a power series according to Eq. (A.56) in the appendix:
(8.32)
For large N, the power series can be truncated. However, at this point it is
important to collect all terms up to the second order. If we were to truncate
the expansion after the first term, then we would lose some quadratic terms in
m. Therefore, we insert the expansion truncated after the second term and
obtain:
For large N, the last term becomes small and can be ignored. Rearranging
then yields
(8.33)
and thus
(8.34)
which is simply the expression that was to be proven.
(8.35)
which is in fact a normal distribution. Note that the continuum limit requires
that a → 0 and τ → 0 in a way that is constant. Then, the variance of the
distribution function,
(8.36)
increases with time, consistent with an irreversible diffusion process of the
molecule. It is common to introduce the diffusion constant D:
(8.37)
Hence the probability density function takes the form:
(8.38)
This probability density function is a solution of Fick’s second law :
(8.39)
A generalization of the diffusion of the molecule in three dimensions is
straightforward. The random walk approach to diffusion we have worked out
in this problem is instructive in conjunction with the origin of irreversibility :
even though each elementary step of the molecule is in principle a reversible
operation, a sequence of such steps statistically introduces irreversibility.
Hint: This problem presumes that you have dealt with Problem 8.3.
Solution 8.4
Here, we deal with a concrete example of the random walk approach to
diffusion on surfaces . On the atomic scale, single crystals have a periodic
structure. At the interface, the periodicity is canceled in the dimension
perpendicular to the surface, but the periodicity in the other two lateral
dimensions is often maintained on mesoscopic length scales. This is
demonstrated, for example, in atomic resolution scanning tunneling
spectroscopy (STM) experiments, or, in the case of insulating materials, in
scanning force microscopy (SFM) experiments. Nowadays, these surface
science techniques have reached a resolution that detects single molecules
and maps their movements from scan to scan [3]. Molecules can be bound to
a surface, either by chemisorption (covalent bonding, binding energies > 50
kJ mol−1), or by physisorption (weak van der Waals forces, binding energies
< 50 kJ mol−1). Decisive for the molecule’s mobility - quite often the
prerequisite for surface chemical reactions to occur - is the depth of the
lateral potential the molecule faces. For simplicity, we identify the well
depth of the lateral potential with the activation energy for diffusion, E a = 8.
5 kJ mol−1. Attempts of the molecule to move to a neighboring adsorption site
are lateral frustrated translations with a restoring force that is mediated by
the lateral potential. In our example, the frequency of this degree of freedom
is ν 0 = 1 THz.
In subproblem (a), we derive an expression for the diffusion coefficient
D introduced in the discussion of Problem 8.3(c). According to Eqs. (8.36)
and (8.37), the diffusion constant is given by
(8.40)
where τ is the time for one elementary step of the molecule. It is obvious that
the hopping of the molecule from one site to the next site is itself a statistical
process and τ has the nature of an average time that is influenced by
temperature. If the number of hopping attempts is ν 0, this frequency has to be
multiplied with a Boltzmann factor to obtain the number of
successful attempts:
(8.41)
As a consequence, our expression for the diffusion constant as a function of
temperature is:
(8.42)
At a temperature of 80 K, we obtain
In subproblem (b), we answer how far the molecule will come
statistically after a 1 min random walk, assuming the continuum
approximation Eq. (8.35). It has been stressed in the solution of Problem 8.3
that the probability density distribution
(8.43)
is normalized, i.e.,
(8.44)
The expectation value 〈x〉, also called the first moment of W is:
(8.45)
This is because W(x, t) is symmetrical with regard to the variable x: W(x, t)
= W(−x, t), while at the same time the integral xW(x, t) is asymmetrical.
Therefore, we must look at the second moment of the distribution, 〈x 2〉. As
W(x, t) is symmetrical:
Thus, 〈x 2〉 is the spatial variance σ 2 of W(x, t) at time t (see Eq. (8.36)), and
its standard deviation is m. This corresponds to 13,000
unit cells. Hence, even at 80 K, the molecule is quite mobile on the surface.
The corresponding profile of the probability of finding the molecule between
location x and x + dx after this time is illustrated in Fig. 8.9.
In subproblem (c), we determine the probability of finding the molecule
after 1 min at its original site, i.e., between and . Note that x = 0,
the center of the distribution, has the highest probability density. The
probability sought is a cumulative probability, similar to what we discussed
in Problem 7.2 or in Problem 9.14c in connection with quantum mechanical
probability. The cumulative probability is the definite integral:
Fig. 8.9 Probability density function W(x, t) of a molecule after a 60 s random walk
Again, we can use the fact that W(x, t) is symmetrical with regard to x.
Therefore,
(8.46)
The function p(t) is shown in Fig. 8.10 for arbitrary time t between 0 and 60
s, calculated with the given value of a = 6 × 10−10 m and the diffusion
constant obtained above: D = 5. 08 × 10−13 m2 s−1. As expected, p(0) equals 1
and reaches within 1 s a value of 2 × 10−4. After 60 s the probability is p(60
s) = 3. 1 × 10−5. This means that statistically only 1 in 30,000 experiments
would detect the molecule at its initial site.
Fig. 8.10 Probability as a function of time to find the molecule at its initial adsorption site. Note the
logarithmic scaling
a. If the systems are separated from each other so that they can be
considered decoupled, are they still in equilibrium and
maintaining their initial temperature?
b.
Write down an expression for the total energy E of both
systems. If the systems are decoupled from each other, which
condition follows for the probability p(E) of measuring the
total energy E?
c. Show that p(E) is the Boltzmann distribution Eq. (8.3). Hint:
Set up a differential equation for the probability distribution.
Fig. 8.11 Two systems A and B initially in contact at temperature T stay in equilibrium after their
separation, where they can be considered decoupled
Solution 8.5
In this exercise, it is our goal to derive the Boltzmann distribution Eq. (8.3)
for a system composed of two identical subsystems from very general
assumptions. Among the various derivations found in textbooks, this one is
quite compact; thus, it can be reproduced in oral examinations. The two
systems are initially in contact with each other and in equilibrium. The
temperature in both subsystems is constant, but the energies E A and E B in
the systems fluctuate around a mean value.
In subproblem (a), we answer the question whether both systems stay in
equilibrium if they are separated from each other. This is in fact the case. If
two systems at different temperature are brought into contact, they equalize
their temperature until equilibrium is established. If they are then decoupled,
they maintain this temperature, at least if adhesion effects can be ignored.
This idea, first introduced by J.W. Gibbs, guarantees the equilibrium of two
perfectly decoupled systems at a constant temperature.
In subproblem (b), we focus on the consequences of this situation: the
total energy E is additive, i.e.,
(8.47)
If the probability of finding an energy E A in system A is p A(E A), and the
probability of measuring E B in system B is p B(E B), then we can set up a
joint probability for the total system to measure the energy E. Because the
systems are now independent of each other, the joint probability is obtained
by multiplication of the probability distributions p A and p B:
(8.48)
In subproblem (c), we show that p(E) is the Boltzmann distribution Eq. (8.3).
In fact, the additivity of the energies in the subsystems (Eq. (8.47)) and their
mutual independence (Eq. (8.48)) unambiguously determine the mathematical
form of the probability distribution. But how can we show this? The first step
may be the notion that the condition Eq. (8.48) is special, and that it is
satisfied by the exponential function:
(8.49)
(8.50)
on the one hand, and on the other hand
(8.51)
(8.52)
(8.53)
Note that this condition holds for arbitrary E A and E B. As a
consequence, these expressions must be constant. Because p A, p B and p are
probability densities, they must be positive. Their gradients, as mentioned
above, should be negative. Therefore, the constant must be a negative
number:
(8.54)
The differential equation, e.g., for p A is obtained by rearrangement of
this expression:
(8.55)
Its solution is:
(8.56)
which is the Boltzmann distribution. The constant C must follow from the
normalization of the distribution function. Its determination would require
further assumptions about the nature of the system, e.g. if the spectrum of
possible energy values is continuous or if it is discrete. C is closely related
to the partition function. The assignment of β to the temperature of the system
is achieved by comparison with a concrete case, e.g., the Maxwell-
Boltzmann velocity distribution of a monatomic perfect gas. In this case, the
average energy is (see Eq. (7.18) in Problem 7.1). By comparison, it
can be shown that .
a. Prove that the molar entropy of a monatomic perfect gas with mass
m and volume V is:
(8.57)
b.
Calculate the standard molar entropies of gaseous neon, gaseous
zinc, and gaseous lithium at T = 298. 15 K. Compare the results
with experimental values s ⊖(Ne(g)) = 146. 3 J mol−1 K−1, s
⊖(Zn(g)) = 161. 0 J mol−1 K−1, and s ⊖(Li(g)) = 138. 8 J mol−1 K−1.
Solution 8.6
Is statistical thermodynamics able to correctly predict the molar entropy of a
monatomic gas? In this exercise, we analyze this in more detail. In
subproblem (a), we show that the molar entropy of a monatomic perfect gas
is given by Eq. (8.57), which is called a Sackur-Tetrode equation in the
literature. It is obvious that our starting point for calculating the entropy is
Eq. (8.8)
which traces back the entropy to the system partition function, Q. The
partition function in turn can be written as a product of the N single particle
partition functions, where N is the number of particles, q. In our case N = N A
. The proper form of the latter is that of indistinguishable particles (see Eq.
(8.6)); thus, the expression for the molar entropy becomes
(8.58)
Next, we apply Stirling’s approximation for the factorial N A ! = N A lnN A −
N A (see Eq. (A.62)) and we use the product rule Eq. (A.15) to form the
derivative with regard to temperature:
(8.59)
By means of Eq. (A.23) and N A k B = R, this can be transformed to
(8.60)
Now, we must make an explicit assumption about the single particle
partition function: a monatomic particle has only three translational degrees
of freedom. Its partition function is given in Eq. (8.12). In particular, we
have:
(8.61)
and
(8.62)
Insertion of these results into Eq. (8.60) yields the expression
(8.63)
which is equivalent to Eq. (8.57). A remarkable point is the occurrence of the
Planck constant in the expression. This suggests that a quantum mechanical
treatment of the gas might be necessary for the correct determination of the
entropy.8 Another remarkable point is that the only element-specific quantity
in Eq. (8.57) is the particle mass. It is thus interesting to see how entropies
calculated using the Sackur-Tetrode formula compare with experimental
values for the standard molar entropies of monatomic gases.
In subproblem (b) we consider two examples, the rare gas neon, Ne(g),
and zinc vapor, Zn(g). The atomic weights are m Ne = 20. 18 m u and m Zn =
65. 41 m u respectively. Assuming standard conditions and room temperature,
the gas volume is:
(8.64)
By taking the values h, k B , N A , and m u from the appendix (Sect. A.1), we
obtain the values shown in Table 8.1.
Table 8.1 Standard molar entropies of monatomic gases at T = 298. 15 K, calculated using the Sackur-
Tetrode equation (8.57) in comparison with experimental values. All values in J K−1 mol−1
Element s ⊖s ⊖
calc exp
Ne(g) 146.329 146.3
Zn(g) 160.995 161.0
Li(g) 133.017 138.8
For neon and zinc, the agreement between calculated and experimental
standard entropies is excellent. In the case of lithium, however, a deviation
of more than 5 J K−1 mol−1 is recognized. Can we understand and correct this
discrepancy in detail? It is related to the internal degrees of freedom of the
gas particles. The Sackur-Tetrode equation only takes into account the
translational entropy of the gas, as we considered only the translational part
of the partition function. Thus, we have overlooked one or more degrees of
freedom, which are present in lithium, but apparently not in zinc or in xenon:
we need to consider the electronic structure of these particles. Lithium
(atomic number Z = 3) has the electron configuration 1s 2 2s 1. Its electronic
ground state is characterized by the term symbol . The superscript ”2” is
the multiplicity of the electronic state.9 The electronic partition function is
obtained from Eq. (8.11), where the degeneracy g i of the ith electronic state
is given by its multiplicity 2S i + 1 with S i being the total electronic spin in
this state:
(8.65)
The first excited electronic states of lithium have the term symbols and
. Their energy measured relative to the ground state energy is 178.5 kJ
mol−1. As a consequence, these states are not occupied under room
temperature conditions. They are thus negligible for the calculation of q el..
By taking E 0 = 0,
(8.66)
and thus
(8.67)
This modification leads to a contribution to the entropy of Rln2 = 5. 763 J K−1
mol−1. The corrected value for the standard room temperature entropy of
lithium is thus 138.780 J K−1 mol−1, in good agreement with the experimental
value in Table 8.1. The ground states of atomic zinc and neon both have the
term symbols1 S 0. There is thus no significant contribution of the electronic
degrees of freedom to their room temperature entropy. Note that contributions
resulting from the nuclear spin multiplicity and mixing of different isotopes
are not included in the experimental value of the standard molar entropy. This
is consistent with the third law of thermodynamics, i.e., the statement that s
⊖(T = 0 K) of an element in the state of ideal crystallization is zero [4]. In
(8.68)
b.
Apply Eq. (8.68) to the rotational contribution to the heat
capacity of molecular hydrogen, H2(g). The energy levels are
given by (see Eq. (10.11) in Chap. 10):
(8.69)
where J is the rotational quantum number and I = 4. 6 × 10−48
kg m2 is the molecule’s moment of inertia. If nuclear spin
symmetry is ignored, J can take even and odd values J = 0, 1, 2,
… and a given rotational state is (2J + 1)-fold degenerate. Plot
c V (T) in the range between 0 and 500 K. Use at least ten
rotational levels for your calculation. Interpret the behavior of
c V (T) in the limit T → 0 and T → ∞. Can you explain the
overshooting of the heat capacity in the temperature range
between 50 and 100 K?
Solution 8.7
In this problem, we deal with the rotational degrees of freedom of the
diatomic molecule H2 and how they contribute to its molar heat capacity.
Historically, the correct interpretation of the low temperature molar heat
capacity of molecular hydrogen was a problem that stimulated the
development of early quantum mechanics [5]. The evaluation of the rotational
contribution to the constant volume molar heat capacity is based on the
rotational partition function, q rot. For the latter, however, no closed
expression can be given, as in the case of the translational partition function
or the harmonic oscillator vibrational partition function. Thus, we must fall
back on Eq. (8.11).
In subproblem (a), we show that Eq. (8.68) is the correct expression for
the molar heat capacity of a system with L discrete energy levels E i with a
degree of degeneracy g i respectively. We assume N indistinguishable
subunits. Using Eq. (8.6), we have:
(8.70)
where q is the partition function of an ensemble subunit, e.g., a single
molecule. The internal energy is obtained from Eq. (8.7), i.e.,
(8.71)
where . Using Stirling’s formula Eq. (A.62), we obtain:
(8.72)
Then, application of the chain rule (Eq. (A.17)) yields
(8.73)
as the proper expression for the internal energy. The constant volume heat
capacity for N particles is the derivative with regard to temperature:
(8.74)
We must apply the product rule for differentiation in combination with the
chain rule to obtain:
(8.75)
In fact, we can write this in a more symmetrical way by reducing the terms to
a common denominator, but we have to pay attention to the correct indexing:
(8.76)
If we now factor out k B from the numerator and assume that the particle
number is Avogadro’s number N A , we obtain the molar heat capacity
(8.77)
Fig. 8.12 Rotational energy levels of the H2 molecule in the model of the rigid rotator (schematic).
If at least L = 10 levels are considered, the heat capacity follows the bold
line. At T = 0 K, the heat capacity is zero. Then, between 20 and 50 K it
increases steeply and reaches a maximum at about 70 K. Above this
temperature, the molar heat capacity decreases gradually toward a saturation
value of 1R. This high temperature limit is the classical value for two
rotational degrees of freedom of the molecule contributing to c V . The
fact that c V, rot → 0 for T → 0 shows that the rotational degrees of freedom
are frozen at a very low temperature, where only the rotational level J = 0 is
occupied. It is instructive to analyze the overshooting of the molar heat
capacity exceeding the classical value of 1R by about 10% at about 70 K.
The characteristics of c V, rot in the case of a two-level system (L = 2) is also
shown in Fig. 8.13. The steep increase above 20 K is caused by the
population of the J = 1 level. Because the population of this excited state
reaches a state of saturation at a higher temperature, the system can no longer
absorb further energy; thus, the heat capacity decreases. If three levels are
considered (L = 3), the peak in the heat capacity curve is well reproduced.
The overshooting is thus caused by the population of the levels J = 1 and J =
2. The decrease in the heat capacity appears to be a consequence of the fact
that the level J = 3 has too much in energy to be reached below 100 K.
Although a similar trend in the rotational molar heat capacity is observed
in deuterated hydrogen (HD), ordinary H2 shows a different behavior. As
proposed for the first time in 1929 by Bonhoeffer and Harteck [6], H2 has
two spin-isomers, ortho-H2 with a total nuclear spin of I = 1 (multiplicity 3)
and para-H2 with a total nuclear spin of I = 0 (multiplicity 1). Nuclear spin
symmetry and the Pauli principle state that ortho-H2 can only occupy the
rotational states with odd J, whereas para-H2 populates only states with even
J.
Fig. 8.14 Rotational contribution to the low temperature heat capacity of ortho H2 (solid) and para H2
(short dashed) based on Eq. (8.68). Also shown (dashed) is the curve for normal H2, i.e., a mixture of
ortho and para H2 3:1. c V, rot is plotted in units of the gas constant R
a. Use Eq. (8.68) in Problem 8.7 and show that the contribution to
the molar heat capacity of a two-level system is
(8.79)
(8.80)
Lithium doped potassium chloride, Li:KCl, exhibits a Schottky
anomaly with a peak value near 0.4 K. Determine the energy
difference Δ. Can you give an interpretation of the peak in the
heat capacity vs temperature curve?
Solution 8.8
The appearance of a peak in the heat capacity vs temperature curves of
materials is associated with the term Schottky anomaly. It appears, for
example, in paramagnetic materials, but also in crystals of the alkali halide
KCl, which are doped with the smaller lithium ion. The multilevel nature of
this system is discussed in detail in Problem 9.17. In subproblem (a), we
lead back to Eq. (8.79), describing the heat capacity of a two-level system
for the more general expressions for multilevel systems found in
Problem 8.7. For L = 2 and g i = 1, i = 1, 2 we can write:
(8.81)
Introducing the energy difference Δ between the two levels, we can make the
substitution E 2 = E 1 +Δ and obtain
(8.82)
(8.83)
Extrema of the heat capacity will thus be characterized by a special value of
α m . The necessary condition for the maximum in the heat capacity vs
temperature curve is:
(8.84)
Application of the product rule for differentiation (Eq. (A.15)) yields:
(8.85)
Additional simplifications yield a transcendental equation:
(8.86)
or
(8.87)
which cannot be further simplified. A graphical solution of the last equation
is shown in Fig. 8.15, according to which the extremum is characterized by α
m = 2. 3994 ≈ 2. 4. We have thus justified Eq. (8.80). In subproblem (c), we
deal with the example of lithium doped KCl. Low-temperature measurements
of the heat capacity show that this material exhibits a Schottky anomaly with
a peak temperature of only T Peak ≈ 0. 4 K. This corresponds to a small
energy splitting of:
(8.88)
Below, in Problem 9.17, we deal with a simplified model of the Li:KCl
system using quantum mechanics.
Fig. 8.15 Graphical solution of the transcendental equation (8.87). Function f 1(α) = α is a line through
Fig. 8.16 Schottky anomaly in the system Li:KCl according to Eq. (8.79)
(8.89)
(8.90)
Let us consider the change in the occupation probability of the ground
state:
(8.91)
From Eq. (8.83) we obtain:
(8.92)
As a consequence, the change in occupation probability of the ground
state is proportional to c Schottky(T):
(8.93)
Therefore, the peak in the heat capacity can be associated with the maximum
decrease in occupation probability of the ground state. Obviously, because of
Eq. (8.90) and thus
(8.94)
the population gain of level 2 also reaches a maximum at the peak
temperature T Peak. The change in occupation probability of the excited state
as a function of temperature is also shown in Fig. 8.17.
Fig. 8.17 Population change of the excited state in the two-level system related to the Schottky
anomaly in Li:KCl. The maximum gain of p 2 is found at the peak temperature T Peak (Eq. (8.80))
References
1. Jaynes ET (1992) The Gibbs paradox. In: Smith CR, Erickson GJ, Neudorfer PO (eds) Maximum
entropy and Bayesian methods. Kluwer, Dordrecht, pp 1–22
[CrossRef]
2. Levinthal C (1969) How to fold graciously. In: Debrunner P, Tsibris J, Münck E (eds) Mössbauer
spectroscopy in biological systems. University of Illinois Press, Urbana (IL), p 22
3. Davies, PR, Roberts MW (2008) Atom resolved surface reactions. RSC Publishing, Cambridge
5. Gearhart CA (2010) ”Astonishing Successes” and ”Bitter Disappointment”: the specific heat of
hydrogen in quantum theory. Arch Hist Exact Sci 64:113
Footnotes
1 Strictly speaking, rotational and vibrational degrees of freedom are coupled in a molecule. In many
cases, however, the treatment of a molecule as a rigid rotator with entirely harmonic vibrational
modes is a useful approximation. It allows a separation of rotation and vibrational degrees of
freedom.
2 The switching of a molecule between different conformations has been treated in Problem 5.3.
3 We have dealt with a thermodynamical description of gas mixing in Chap. 3, Problem 3.7.
4
The dominant event would have the probability . Consideration of
5 Note that in some textbooks you might find a less precise version of Stirling’s formula: lnN! ≈ NlnN −
N which, in the present problem would not yield the correct prefactor containing .
7
The differential dx is thus .
8 The calculation of the internal energy of the perfect gas based on Eq. (8.7) and the partition function
Eq. (8.61) would in fact reproduce the classical result .
9 The multiplicity 2S + 1 of an electronic state is related to the total electronic spin S, which is
determined in the case of the ground state by the single electron in the 2s atomic orbital. The
electronic structure of lithium is treated in more detail in Problem 9.5.
© Springer International Publishing AG 2017
Jochen Vogt, Exam Survival Guide: Physical Chemistry, DOI 10.1007/978-3-319-49810-2_9
Abstract
Quantum mechanics provides the basis for the description of matter on the
atomic scale. Developed in conjunction with progress in spectroscopy, it
explains phenomena such as the photoelectric effect, molecular spectra,
electronic structure, and the chemical bond. Challenging and at the same time
fascinating, the predictions of quantum mechanics are beyond direct everyday
perception. (To some extent this is in contrast to thermodynamics and
changes of state (Chap. 3) where experience from everyday perception is a
reasonable criterion for assessing the soundness of predictions.) The basic
concepts introduce the postulates of quantum mechanics. Aside from
problems dealing with black body radiation, wave-packet propagation, and
the hydrogen atom, applications of operator calculus and the variational
principle are highlights that show perspectives with regard to tackling
advanced problems in this field.
(9.1)
From the slope of the function U(ν), the fundamental constant h can be
determined, which is called Planck’s constant.
If the frequency of the light irradiating the cathode is below the threshold
frequency ν 0, no electrons are detected, no matter how high the light
intensity. The interpretation of the photoelectric effect given by Albert
Einstein is that the energy of light is mediated in discrete portions, quanta,
with the energy
(9.2)
The associated particle with zero mass called photon moves at the speed
of light c, and has the momentum
(9.3)
The resulting wave-particle duality was extended by Louis-Victor de
Broglie to electrons and other massive particles. In this context, Eq. (9.3) is
called the de Broglie equation and λ de Broglie wavelength .
(9.4)
Here, 〈E osc〉 is the average oscillator energy of an atomic oscillator
with a frequency ν. The classical equipartition theorem predicts 〈E osc〉 = k B
T. Thus, the spectral energy would diverge to infinity if ν → ∞. Max Planck
has corrected the expression for 〈E osc〉 based on his hypothesis of the
quantization of the oscillator energy:
(9.5)
In Problem 9.1, we show that Planck’s assumption leads to the correct
expression for the energy density of a black body:
(9.6)
The power P emitted by a black body with a total area A depends only on
its temperature and is given by the Stefan-Boltzmann law :
(9.7)
Here,
(9.8)
is the Stefan-Boltzmann constant.
(9.9)
Specifically,
(9.10)
is called the norm of the wave function.
Comment: Dirac’s notation calls 〈ψ | a bra-vector
Postulate 4: A Hermitian operator is one for which
(9.11)
Observables, i.e., measurable quantities with real eigenvalues, are
represented by Hermitian operators. Comment: If is an arbitrary
operator, then is called its adjoint operator , and
(9.12)
As a consequence, a Hermitian operator is called self adjoint
operator : .
Postulate 5: If is an arbitrary operator, then | ψ n 〉 is called
eigenfunction or eigenstate of , and λ n eigenvalue, if the following
equation holds:
(9.13)
If there are several possible eigenstates, then the eigenstates are
orthonormal2:
(9.14)
Comment: This requires the eigenstates to be normalized. The
spectrum of possible eigenstates constitutes a complete basis, and an
arbitrary state function can be written:
Since
(9.16)
Comment: The operator is the Hamiltonian whose eigenvalue is
the total energy (see Problem 9.6).
9.1.2.2 Commutators
Application of quantum mechanics to special systems and their discussion
frequently makes use of the concept of a commutator between two operators
and :
(9.17)
If then the two operators commute with each other. In this case, the
order of these operators does not influence the result. An important
consequence is that commuting operators share a common orthonormal
system of eigenfunctions.
(9.18)
The problem can be treated exactly and the solutions can be found in
textbooks. The energy eigenvalues are:
(9.19)
where the angular frequency ω of the motion is related to the force constant k
(9.20)
Fig. 9.1 Energy levels and wave functions of the harmonic oscillator. The amplitude of the wave
functions was scaled by a factor of 0.5
(9.22)
is the Rydberg energy. As can be seen in Eq. (9.21), there is a degeneracy of
the energy levels with regard to different values of l. In the absence of
external fields a further degeneracy with regard to the magnetic quantum
number m occurs, which takes integer values between − l and + l. The wave
functions for the hydrogen problem are a solution of the Schrödinger equation
(9.23)
where the Laplacian operator Δ is best expressed in spherical coordinates r,
θ and ϕ (see Eq. (A.68) in Sect. A.3.11). The wave function separates into
two parts:4 a radial function R nl (r) and a spherical harmonic Y lm (θ, ϕ):
(9.24)
The eigenvalue equation for the angular part is:
(9.25)
with being the operator of the angular momentum squared, and l is the
angular momentum quantum number. A state with l = 0 is called an s-state, l
= 1 is called a p-state, and l = 2 a d-state. Explicit expressions for the first
spherical harmonics are given in the Appendix in Sect. A.3.14. The
eigenvalue equation of the radial function is:
(9.26)
Its solutions are tabulated in the Appendix Sect. A.3.15 for hydrogen (Z = 1).
The spatial shape of some orbital wave functions is shown in Fig. 9.2.
Fig. 9.2 Spacial geometry of some orbitals of the hydrogen atom. The surface of the constant contour
value ± 0. 05 of the orbital wave functions is shown. The inset demonstrates the construction of (real)
2p x , 2p y and 2p z functions from radial functions and (complex) spherical harmonics
9.2 Problems
Problem 9.1 (Derivation of the Average Oscillator Energy)
(9.29)
derive Planck’s equation for the average oscillator energy
(9.30)
E 0 is an arbitrary constant energy that is determined in Problem 9.2.
Fig. 9.3 Thermal equilibrium between the electromagnetic radiation field within a cavity coupled with
absorption and emission to atomic oscillators at the cavity walls (schematic)
Solution 9.1
In this exercise, we deal with the average oscillator energy introduced in
Sect. 9.1.1.2. Doing so, we understand an essential part of Planck’s law for
the radiation density of a black body. Before entering the calculation, it is
useful to reflect on the nature of the oscillators for which we calculate the
average energy. Following Planck, we can assume a thermal equilibrium
between the electromagnetic radiation field within a cavity and the atoms at
the cavity walls. The equilibrium is established by absorption and emission
processes of quanta of energy h ν. The scenario is sketched schematically in
Fig. 9.3. Experimentally, it has been found that the spectral energy density of
a black body is independent of its composition, chemical nature, and
structure. The simple model representing the atoms of the cavity wall is thus
an ensemble of so-called Lorentz-oscillators , spring mass systems vibrating
at a frequency ν. We have seen how such an ensemble of oscillators can be
treated by means of statistical thermodynamics in Sect. 8.1.2
According to Planck’s assumption, the energy values of an oscillator of
frequency ν is:
(9.31)
Here, n counts the number of quanta by which the oscillator is excited.
Next, the concept of probability and statistics is introduced: we express the
expectation value of the oscillator energy, 〈E osc〉, by the average number of
quanta, 〈n〉:
(9.32)
For 〈n〉 we use a Boltzmann-Ansatz:
(9.33)
Note that the factor α compares the quantum energy of the oscillator with
the thermal energy at a given temperature T. The rest is clever mathematics.
The sums in the last equation can be expressed by a geometrical series (see
Eq. (A.57) in the Appendix):
(9.34)
and
(9.35)
Putting the last three equations together, we obtain:
(9.36)
Hence, by resubstitution, the average energy of a harmonic oscillator at a
temperature of T is
(9.37)
leading to the correct expression for the spectral energy density of black
body radiation, Eq. (9.6). Interestingly, Max Planck went a slightly different
route, focusing on the average oscillator entropy [1].
Solution 9.2
Although the zero point energy of an oscillator would follow from a rigorous
quantum mechanical treatment of the harmonic oscillator based on the
Schrödinger equation, the same result has already been obtained from
consideration of the so-called correspondence principle , proposed by Niels
Bohr [2]: Each quantum system should behave like a classical system in the
limit where classical mechanics applies. Classically, one expects an
ensemble of oscillators at a finite temperature to have an average oscillator
energy of k B T. This value should be reproduced by the quantum mechanical
treatment; if high quantum numbers n are occupied, i.e., if the average energy
of the oscillator consists of so many energy portions h ν that the discrete
nature of the excitation spectrum becomes negligible: this is the limit h ν ≪ k
B T. Therefore, we start with the oscillator energy in Eq. (9.30) and make use
of the power series expansion Eq. (A.58) found in the Appendix,
Fig. 9.4 Principle of electron impact heating. U acc is the high voltage, by which emitted electrons are
accelerated toward the copper sample
Solution 9.3
Cathode rays discovered by J.J. Thomson have practical applications in
science and technology. The impact of electrons is a very effective way of
heating a sample to a high temperature.5 Under ultrahigh vacuum conditions
at a pressure below 10−5 Pa, heat transfer from the sample to the
surroundings is predominantly radiative; if heat losses across the sample
mounting can be ignored.6 Assuming the copper sample to be a black body,
the total radiation power emitted can be estimated using the Stefan Boltzmann
law (Eq. (9.7)). If we additionally account for the temperature of the
surroundings T surr., the net thermal power emitted by the sample at a given
temperature T sample is
where A is the total area of the copper sample, i.e., twice the cross-sectional
area. To determine the necessary acceleration voltage, we equate this thermal
power with the electrical power as the product of U acc and the electron
current I e impinging on the sample. Hence,
(9.39)
and solving for U acc we obtain
Solution 9.4
Fig. 9.5 Experimental setup to measure the photoelectric effect from cathode materials. The inset on
the left illustrates the photo effect in the energy band model of a metal . The work function e ϕ is the
difference in energy of the Fermi energy E F and the vacuum potential level
b. Figure 9.6 shows the energy levels of the Li atom (L.J. Radziemski
et al., Phys. Rev. A 52 (1995), 4462). Use the result of
subproblem (a) to calculate the energy levels of the lithium atom.
Compare the result with the experimental data and describe the
differences.
c.
An improved approximative model for hydrogen-like atoms makes
use of the so-called quantum defects δ l (l = s, p, …), which modify
the formula for the energy levels according to:
(9.41)
Use the data in Fig. 9.6 and determine the parameters δ l for
lithium.
Solution 9.5
Although the quantum mechanical models give a very concise picture for
atoms and ions with one single electron, systems with more than one electron
can only be treated in an approximate fashion. Here, we shall consider the
lithium atom with three electrons and the ground state configuration 1s22s1. If
the atom undergoes a process of ionization, the 2s electron is removed. The
situation is shown in Fig. 9.7. In the simple picture of the Bohr model, the
two 1s electrons lead to a shielding of nuclear charge. If the shielding were
complete, the 2s electron would see an effective charge of + e. For all
transitions in which only the 2s electron is involved, an approximative
treatment on the basis of an adapted model for the hydrogen atom can be
applied.
Fig. 9.7 Simple picture of the lithium atom within the Bohr model. The charge of the nucleus is
partially shielded by the 1s electrons
(9.42)
If the model is reasonable, we expect Z eff close to 1. In the ionization
process (n i = 2 and n f → ∞), the last equation yields:
and thus
The result is in fact close to 1, but it also indicates that the shielding of
the nuclear charge by the 1s electrons is not complete.
Further, in subproblem (b), we deal with the lithium atom and refine the
treatment of its electronic structure to some extent. Alkali metals (Li, Na, K,
…) have a single valence electron and the electronic excitations of the atom
are governed by transitions of this valence electron in excited states. The
energy levels of such states are given in Fig. 9.6. At first, we investigate how
well the effective nuclear charge Z eff = 1. 26 can reproduce this term
diagram. The energy levels depend only on the quantum number n and are
calculated using:
(9.43)
We obtain E 2 = −4. 28 eV, E 3 = −1. 90 eV, E 4 = −1. 07 eV, and E 5 = −0.
69 eV. We notice that the lithium energy levels E nl depend markedly on the
angular momentum quantum number l, which is not reproduced by our simple
formula. Our result for E 2 lies between the 2s level (−5.39 eV), and the 2p
level (−3.54 eV). Similarly, our result for E 3 falls between 3s (−2.02 eV)
and 3d (−1.51 eV). Our value E 4 is fairly close to the energy of the 4s state
(−1.05 eV), the same is true for E 5, which is close to the energy of the 5s
state (−0.64 eV). In summary, we can conclude that the higher the principal
quantum number n, the better the prediction based on the effective nuclear
charge. For n = 2, the agreement is worse and the model cannot explain the
observed splitting between 2s and 2p, for example. Is an improved treatment
possible?
In subproblem (c), we follow an alternative treatment that goes back to
the work of Rydberg and Schrödinger [3], who introduced the quantum defect
δ l . We use Eq. (9.41) and determine δ l from the energy levels in Fig. 9.6.
We could carry out a trial and error procedure or nonlinear fitting to
determine δ s from the values for 2s, 3s, 4s, and 5s. Alternatively, we solve
for δ l and obtain:
(9.44)
We obtain δ s = 0. 412 (2s), 0.405 (3s), 0.401 (4s), and 0.390 (5s). All
values are close to each other. We take the average value of 0.402 δ s and
obtain the energy levels shown in Table 9.1. With the same method, we
obtain δ p = 0. 040 and δ d = 0. 000. The agreement between experimental
and calculated energies is good.
Table 9.1 Comparison of energy levels in the lithium atom (see Fig. 9.6) and calculated energies based
on the quantum defect approximation with Eq. (9.41)
Because of its simplicity the quantum defect model has been popular and
is still used to describe highly excited atoms, the so-called Rydberg atoms .
(9.45)
where H is the total energy of the particle, and p its momentum.
Apply the following operators to the free particle wave
function:
(9.46)
In theoretical mechanics there are two functions on which the equations of
motion of a classical mechanical system are based: one is the Hamilton
function , or the Hamiltonian H, which is the total kinetic energy plus the
total potential energy of the system10:
(9.47)
In the case of a single particle moving freely, the potential V takes a constant
value. In general, however, V may vary with the location of the particle. If V
also depended on time, we would be dealing with a nonconservative system .
If the Hamiltonian of a system is known, the equations of motion follow from:
(9.48)
(9.49)
Here, p i is the momentum of the ith particle and x i is its position. In our
case of a one-particle system, the equations of motion are thus:
(9.50)
and
(9.51)
The momentum of the particle is constant, and its position is obtained by
integration. If x(0) is the position at t = 0, then:
(9.52)
where is the constant velocity of the particle. Thus, the classical
trajectory is well-defined by classical mechanics.
In subproblem (b), we switch to a quantum mechanical description using
the one-particle wave function Eq. (9.45). This choice is not mandatory, but
note that – provided that the superposition principle holds in quantum
mechanics – any arbitrary complex wave function can be represented by a
superposition of plane waves. Another strong argument is the analogy with
optics and electromagnetic wave theory, which in a vacuum makes use of a
plane wave representation of the electric and the magnetic field vector,
which is a solution of the wave equation. What sort of wave equation does
Eq. (9.45) satisfy? We check this by taking the partial derivative in time:
(9.53)
Thus, the special operator applied to the wave function reproduces the
latter, and multiplies it with the total energy. In mathematical terms, Eq.
(9.53) is an eigenvalue problem. The plane wave is an eigenfunction of the
special operator . We call this operator the Hamiltonian operator , as
its eigenvalue is simply the total energy (see subproblem (a)). The same
arithmetic can be performed using the other operators. For the partial
derivative in x, we obtain:
(9.54)
Thus, we identify the special operator as the momentum operator
(in x-direction), and the wave function also proves to be an eigenfunction of
, and p is its eigenvalue. It is straightforward now to seek an operator of
the kinetic energy, , which is the third operator we apply. We obtain:
(9.55)
Moreover, we recognize that there is a second way of expressing the
Hamiltonian operator as the sum of the operator of the kinetic energy and the
constant potential:
(9.56)
Thus, if we combine Eq. (9.53) with Eq. (9.56), we obtain the wave equation
sought, which is satisfied by the plane wave:
(9.57)
This is the Schrödinger equation of a massive particle for the special
case of a constant potential. Note that derivatives in space (second order)
and in time (first order) differ, whereas the wave equation describing
electromagnetic waves in a vacuum contains second-order derivatives in
space and time, as the textbooks show. In summary, using the plane wave
description of a particle moving in a constant potential, we have identified
the wave equation describing the motion of the particle. Moreover, we have
seen that the physical observables total energy H, kinetic energy E kin, and
momentum p are associated with special operators. This concept can be
generalized to an arbitrary observable O and its associated operator .
(9.58)
where B( p) is the distribution function for the momentum. Consider the
special distribution function:
a.
Give the expression for the probability density w(x) = | ψ(x) | 2
to find the particle within x and x + dx. Note that A follows
from the normalization of the wave function. Sketch the
probability density for an electron with average kinetic energy
E 0 = 100 eV and Δ p = 2. 5 × 10−3 p 0.
Solution 9.7
According to the superposition principle in quantum mechanics (see Sect.
9.1.2), an arbitrary function can be written as a superposition of plane
waves. Although a single plane wave has a well-defined momentum and
energy, a wave packet constructed by the superposition of many plane waves
is characterized by a momentum distribution, thus involving the statistical
aspect of the wave function. Quantum mechanics assumes that the wave
function ψ(x, t) contains all the information about the particle, and | ψ(x) | 2 =
ψ ∗(x)ψ(x) dx is the probability of finding the particle between x and x + dx.
Note that the rules dealing with complex numbers are summarized in the
Appendix, Sect. A.3.4.
In subproblem (a), we calculate the probability density w(x) = | ψ(x) | 2
for the wave packet Eq. (9.58). In the first step, we seek the spatial
representation of the wave function:
(9.59)
Here, we have inserted the special momentum distribution function B( p)
which states that within a certain interval 2Δ p around the average momentum
p 0, all values of p have the same weight characterized by the quantity A,
which we will determine later. The integral can be solved using the integral
table in the Appendix. We obtain:
(9.60)
(9.61)
(9.62)
We determine A from the normalization of w(x):
(9.63)
Solving for A, we get
(9.64)
and thus the probability density function is
(9.65)
Note that the point x = 0 requires special treatment because of the pole.
Using the rule of l’Hôspital, we verify that . Next, we consider the
concrete example of an electron with an energy of 100 eV. This is a typical
energy measurement of electrons in low-energy electron diffraction (LEED)
experiments probing a crystalline surface. An ideal electron gun emits a
monochromatic electron beam. Using Eq. (9.46), the momentum of such a
perfectly monochromatic electron wave would be:
(9.66)
Fig. 9.8 Probability density function w(x) = | ψ(x) | 2 of a particle with boxcar momentum distribution.
The blue line marks the half width 2Δ x
(9.67)
and
(9.68)
which can be written in the following way:
(9.69)
(9.70)
This can be done graphically by seeking the intersection point, which is at α
≈ 1. 39 (see Fig. 9.9). We therefore find the following relation between the
uncertainties in momentum and position:
(9.71)
Note that our result is consistent with Heisenberg’s uncertainty principle [4]:
(9.72)
The uncertainties in position and momentum are mutually related. An
increase in Δ p leads to a decrease in Δ x and vice versa. Applied to the
concrete example of diffraction experiments, this has concrete implications:
a highly monochromatic gun with small Δ p guarantees a large Δ x, which is
advantageous in diffraction experiments where constructive and destructive
interference of large wave trains is needed. An electron gun with a large
energy spread and thus large Δ p, in contrast, would result in small values of
Δ x and interference between atomic layers at a crystal surface would be
hampered. Note that typical atomic layer distances in a crystal are within the
range of a few tenths of a nanometer.
Fig. 9.9 Graphical solution of the nonlinear equation
(9.73)
is a solution of the time-dependent Schrödinger-equation of a
free particle of mass m. N is a normalization constant, the
parameter α is the initial variance σ x 2 of the wave packet,
, , the group velocity .
b.
Determine the normalization constant N in Eq. (9.73).
c. Give an interpretation of Eq. (9.73) by plotting the probability
density ψ ⋆ ψ for t = 0 s, t = 2 × 10−14 s, and t = 5 × 10−14 s for a
ballistic electron with a group velocity of 106 m s−1. The
electron is initially located within 10−9 m at x(t = 0) = 0.
Solution 9.8
Few textbooks deal with explicit solutions of the time-dependent
Schrödinger equation. Here, we consider the ballistic movement of a free
particle in one spatial dimension with time. The probability of locating the
particle within a certain interval between x and x + dx is determined by the
absolute square of the wave function, | ψ(x, t) | 2.
In subproblem (a), we prove that Eq. (9.73) is a solution of the time-
dependent Schrödinger equation. The straightforward, but somewhat tedious
way is to calculate the derivatives of ψ(x, t) in x and t and to show that
equals . Making use of the chain rule, the first
derivative in x is
and thus
To form the second derivative, we use the chain rule on the last equation
again and obtain
(9.74)
As
and
The integral can be evaluated using a definite integral from the integral
table (see Eq. (A.46)):
After the substitution u = x − v g t and we obtain
(9.75)
(9.76)
Fig. 9.10 Movement and spreading of the Gaussian wave packet with time. The blue dashed line
marks the classical trajectory of the particle according to Eq. (9.52) moving with the group velocity v g .
The center of the Gaussian wave packet follows the classical trajectory
If | ψ(t 0)〉 is the initial state of a quantum system at time t 0 and its
time-independent Hamiltonian, the state of the system at time t 0 +τ is:
(9.77)
b.
Using Eq. (9.77), prove that if ψ(t) is normalized at t = t 0, then
it is normalized at arbitrary t.
Solution 9.9
The probability interpretation of quantum mechanics is frequently doubted by
students, perhaps as often as the validity of the second law of
thermodynamics. The interpretation of | ψ | 2 as a probability density in fact
requires the total probability, i.e., the norm, to be conserved. It is our task in
this problem to rigorously show this on the basis of the postulates of quantum
mechanics (Sect. 9.1.2). Apart from the Hamiltonian and the wave
function at t = t 0, no further specifications are made about the system. The
problem is an occasion to deal with the numerous postulates of quantum
mechanics found in Sect. 9.1.2.1.
In subproblem (a), we consider a special operator acting on the wave
function, namely the propagator
(9.78)
We show that applied to wave function at t 0 yields the wave function at a
later time t > t 0. As a hint, it was recommended to expand and | ψ(t)〉. We
thus start with a Taylor series expansion of the wave function (see Eq. (A.54)
in the Appendix):
(9.79)
The sum on the right-hand side, however, can be written in a compact form
using an exponential operator acting on the state | ψ(t)〉 at time t = t 0:
(9.80)
This exponential operator thus shifts the wave function forward in time by a
value τ, which justifies the terminus propagator :
(9.81)
What is left is to replace the time derivative with the Hamiltonian. Using
(see Eq. (9.16)), we find
(9.82)
which proves Eq. (9.77).
As an application of the postulates of quantum mechanics in Sect. 9.1.2.1,
it is worth looking at the special properties of : Because of
As a consequence,
(9.83)
An operator satisfying Eq. (9.83) is called unitary operator , and the
transformation Eq. (9.82) is thus called a unitary transformation . This
leads us to subproblem (b), where we prove that the wave function, if
normalized at t = t 0, is also normalized at t > t 0, e.g., at t = t 0 +τ. The norm
is the scalar product
(9.84)
Because of , we can thus write:
(9.85)
which was to be proven. The conservation of the norm of the wave function
is thus the consequence of the unitary transformation of the wave function
using the propagator . In Problem 9.8 the conservation of the probability
was shown for the special case of a wave packet. The wave packet depicted
in Fig. 9.10 broadens over time, but the total probability density remains
constant. In this exercise, we have seen that this property of the wave
function is general. The conservation of the norm would hold even in a more
complex situation, e.g., if a wave packet would be partially reflected at a
wall.
(9.86)
(9.87)
are introduced, where and are the operators of position and
momentum of the particle of mass m. The particle moves in a harmonic
potential at an angular frequency ω.
(9.88)
(9.89)
(9.90)
(9.91)
where the number operator is defined by means of N = a a. +
(9.92)
c.
If | n〉 is the state function of the harmonic oscillator in its nth
excited state, prove and interpret the following relations:
(9.93)
(9.94)
Solution 9.10
In this exercise, we become more familiarized with operator calculus in
quantum mechanics. We acquire an insight into how this can be
advantageously used for the description of the harmonic oscillator model,
which is of key importance, e.g., in vibrational spectroscopy.
The momentum operator of a particle moving in one dimension has
already been introduced in Problem 9.6:
(9.95)
We use together with the position operator to formulate the operators a +
and a according to Eqs. (9.86) and (9.87). These are called creation and
annihilation operator and it is our goal to clarify their significance. In
subproblem (a), we prove commutation relations for a, a +, and a third
operator called the number operator :
(9.96)
The commutator of two operators has been defined in Eq. (9.17). The
significance of the commutator becomes clear if we realize that two
mathematical objects do not necessarily obey the commutative law. If the
order of two operators acting on a wave function does influence the result,
then they have a non-zero commutator. An example is the operators of
position of momentum, where the order does in fact matter: if we consider
the expressions
and
we notice that
(9.97)
which is the first of the commutator expressions that was meant to be shown.
It is straightforward to apply Eq. (9.17) to a + and a:
(9.98)
This was shown. Note that after expansion of the terms in the first line of Eq.
(9.98), the terms containing and cancel each other out. To prove Eq.
(9.90), we can use Eq. (9.89):
(9.99)
In the same fashion, Eq. (9.91) is shown:
(9.100)
To understand the meaning of these operators we move on to subproblem
(b), where we show that the Hamiltonian of the harmonic oscillator problem
can be rewritten in a simpler form containing the number operator. From the
Schrödinger equation of the harmonic oscillator Eqs. (9.18) and (9.20) we
may extract the Hamiltonian in the form:
(9.101)
where we have considered Eq. (9.95). In this form, the relation to the classic
Hamiltonian H = E kin + V is obvious. To prove Eq. (9.92) it is best to show
that the right-hand side of this equation is equivalent to Eq. (9.101):
(9.102)
Here, it is worth comparing this special form of the Hamiltonian with the
expression for the energies of the harmonic oscillator, Eq. (9.19). It is
obvious that and thus the number operator and the Hamiltonian
operator are commuting operators, meaning that they share the same spectrum
of eigenfunctions. In the form
(9.103)
the significance of N now becomes very clear: it yields the number of quanta,
n, by which the oscillator is excited. In a notation with ket vectors,
(9.104)
Moving on to subproblem (c), we show further relations revealing the
significance of the operators a and a +. Employing the commutator relations
Eqs. (9.90) and (9.91) we first consider the expression:
(9.105)
which proves Eq. (9.93)
Using the same trick with the commutator relation Eq. (9.90) we obtain:
(9.106)
Our interpretation of these expressions is based on the number operator:
If it is applied to the state a + | n〉, it yields n + 1 quanta. Moreover, a state a |
n〉 contains n − 1 quanta. That is, a + increases n by 1; it creates a quantum,
whereas a decreases n by 1, it annihilates a quantum. Further analysis would
in fact show that
(9.107)
and
(9.108)
At the end of this exercise, we reflect on the use of this concept, for example,
in conjunction with spectroscopy. A very important application is the
derivation of selection rules for the harmonic oscillator,11 as these operators
describe transitions between harmonic oscillator states. In addition, if the
harmonic oscillator model is applied to the case of electromagnetic waves
and the photon picture of light, the raising and lowering operators describe
the creation and annihilation of photons in elementary interaction processes
with matter. Moreover, we use these operators in Problem 9.17 to solve the
quantum double well.
(9.109)
are introduced.
a. Verify that the Pauli spin matrices satisfy the commutation rules
for the angular momentum apart from a constant factor.
(9.110)
Use the Pauli spin matrices to construct ladder operators σ +
and σ − with the following properties:
(9.111)
c.
What is the effect of σ z acting on ground and excited states?
Solution 9.11
This second exercise on operator calculus focuses on Pauli spin matrices and
their use in conjunction with the description of spin and two-level systems in
general. In subproblem (a), we verify that the Pauli spin matrices defined in
Eq. (9.109) satisfy the angular momentum commutation rules apart from a
constant factor. Here, we must recapitulate at first these rules for
examinations in spectroscopy or advanced physical chemistry with which we
should be very familiar. Given a system with angular momentum vector J = (J
x , J y , J z ) and the related angular momentum operators , , and , the
three components of the angular momentum cannot be measured at the same
time with arbitrary precision, because they do not commute:
(9.112)
(9.113)
(9.114)
The square of the angular momentum, J 2=J 2 2 2
x + J y + J z and the z-
component is usually chosen to define the state of angular momentum. The
respective operators commute with each other:
(9.115)
Matrices (see Sect. A.3.16 in the Appendix) are mathematical objects that do
not necessarily follow the commutative law. It is straightforward to check
this for the Pauli spin matrices:
(9.116)
and
(9.118)
(9.120)
with I 2 the (2,2) unit matrix. It is obvious that each of the Pauli spin matrices
commutes with the unit matrix. Thus, the commutation rule Eq. (9.115) also
finds its analogy in:
(9.121)
Fig. 9.11 A quantum mechanical two-level system of a ground and an excited state with an energy
difference ℏ ω. Ladder operators σ + and σ − mediate transitions between the two states
In subproblem (b), we focus on the use of the Pauli spin matrices for the
description of a quantum mechanical two-level system . This may describe,
for example, a proton or an electron with spin in an external magnetic
field. Depending on the orientation, the spin with regard to the magnetic
field, the energy of the particle splits into two states (Zeeman effect ). Two-
level systems, however, are also important in spectroscopy for the treatment
of the fundamental interaction between light and matter: absorption,
emission, and induced emission (see Chap. 10). Here, the two-level system
represents an atom or molecule with an initial state | i〉 and a final state | f〉,
whereas the electromagnetic field is described in the model of a harmonic
oscillator with quantum energy ℏ ω matching the energy difference between
the states | i〉 and | f〉. In the literature, such a model is known as a Jaynes
Cummings model . The latter makes use of Pauli spin matrices to describe the
transitions of the two-level system. In the present problem, we consider the
ground state denoted | g〉 and an excited state | e〉, as illustrated in Fig. 9.11. A
general mixed state of the system is given according to Eq. (9.110) where the
absolute values of the two complex numbers α and β give the probability of
finding the system in the pure states | e〉 and | g〉 respectively. This two-level
system can be represented mathematically by vectors where the ground and
the excited state are given as:
(9.122)
Our task is to find ladder operators describing the transitions between these
two states. According to the problem text, these ladder operators can be
constructed from the Pauli spin matrices. There are two possibilities of
finding these operators: guessing is of course possible for the student, with
mathematical intuition. We follow the second, systematic way. Starting with σ
+, we assume that this operator can be written as a linear combination of the
Pauli spin matrices:
(9.123)
According to Eq. (9.111), we can set up two equations to determine the
coefficients a, b, and c:
(9.124)
and
(9.125)
From these equations, it is immediately clear that c = 0. For a and b we have
two equations:
(9.126)
and
(9.127)
The addition of Eqs. (9.126) and (9.127) yields the result , whereas
subtraction gives . We reinsert these values in Eq. (9.123) and obtain
the result:
(9.128)
For the second ladder operator σ − we can proceed in exactly the same way.
The result is:
(9.129)
In subproblem (c), we examine the effect of σ z on the states of the two-
level system. We have:
(9.130)
and
(9.131)
We can interpret this in the following way: in a two-level system with spin, σ
z projects out the spin orientation. In subproblem (a), we have seen that the
Pauli matrices need to be multiplied by a factor of to obtain the spin
operators. Thus,
(9.132)
and
(9.133)
In a general two-level system, σ z can be used to determine the energy of the
state. If the energy of the ground state is and that of the excited state
(see Fig. 9.11), then the Hamiltonian of this system is apparently:
(9.134)
and
(9.135)
In summary, we have seen how a special type of matrix as a concrete
representation of operators can be used to describe transitions in a two-level
system. We were able to formulate the Hamiltonian of such a system, which,
as mentioned above, is the basis for the Jaynes Cummings model used in
optics and spectroscopy.
Fig. 9.12 The model of an electron, which is free to move on a ring with a radius of r
An electron may freely move on a ring with a radius r (see Fig. 9.12).
The Schrödinger equation for this problem is:
(9.136)
where the azimuthal angle ϕ characterizes the position of the electron.
Solution 9.12
The electron on the ring is an instructive example of how quantization results
as a consequence of boundary conditions. An application of this simple
model to electronic excitations in the benzene molecule is shown below in
Problem 9.13. In subproblem (a), we derive the Schrödinger equation Eq.
(9.136) for the wave function ψ(ϕ). As stated in the problem text, the electron
moves freely, i.e., it moves in a constant potential that can be set to zero. The
general form of the Schrödinger equation is thus:
(9.138)
As can be seen in Fig. 9.12, the azimuthal angle ϕ is the only degree of
freedom left to the electron forced onto the ring. The movement of the
electron is best described in spherical coordinates in which the Laplace
operator in the general Schrödinger equation reads (see Eq. (A.68)):
(9.139)
We choose the origin of the coordinate system in the center of the ring,
and we can assume that the ring plane is aligned to the equator . Hence,
sinθ = 1. Because θ and the radial coordinate r are constant, the first two
terms in Eq. (9.139) are zero and the Laplacian reduces to:
Therefore, the Schrödinger equation for the electron on the ring is given
by Eq. (9.136).
In subproblem (b), we seek general solutions by inserting the ansatz
ψ(ϕ) = Ae iM ϕ in Eq. (9.136). As
and therefore
(9.140)
A discrete spectrum of energy levels is a consequence of the special
boundary conditions. Finally, we exploit the normalization condition. The
total probability of finding the electron on the ring is 1. Therefore,
(9.141)
The real part of the complex wave function is illustrated in Fig. 9.13 for
M = 2 and M = 4. Note that according to Eq. (9.141) the ground state wave
function (M = 0) has a constant value across the ring, whereas excited states
exhibit a harmonic behavior, where the number of full oscillations across the
ring is an integer.
Fig. 9.13 Illustration of the solutions for the electron on the ring in a quantum mechanical treatment.
The real part of the complex wave functions is shown for the quantum numbers M = 2 and M = 4
Problem 9.13 (Electronic Excitation of the Benzene Molecule)
Hint: This problem assumes that you have dealt with Problem 9.12.
Solution 9.13
Electronic excitations involving delocalized π-electrons, so-called π → π ∗
transitions, are highly important in relation to chromophores , i.e., groups of
atoms within a material responsible for its color. Benzene (Fig. 9.14) is one
of the prime examples of a molecule with delocalized π electrons. In this
problem, we apply the simple model of the electron on the ring
(Problem 9.12) to the electronic properties of this molecule. We assume that
the six π electrons of benzene move independently from each other across the
ring. Moreover, any interaction with atomic cores is neglected altogether.
The crudeness of this approximation has the advantage of simplicity.
Fig. 9.14 The benzene molecule C6H6 with its cyclic structure. The six π electrons are delocalized
over the hexagonal ring. The radius of the ring (blue) is approximately identical to the carbon-carbon
distance of 140 pm
a. Show that the ground state wave function of the hydrogen atom,
(9.142)
is normalized.
Solution 9.14
This problem is a straightforward application of quantum mechanical rules
and integration. We deal with the ground state solution of the hydrogen
problem and prove in subproblem (a) that the 1s wave function is
normalized, i.e., we must show that:
(9.143)
In spherical coordinates (see Sect. A.3.11 in the Appendix), the integral
is:
Thus, the 1s wave function is indeed normalized.
In subproblem (b), we calculate the expectation value for the radial
coordinate in this state (see postulate 6 in Sect. 9.1.2.1),
(9.144)
The definite integral Eq. (A.45), can be used for the
calculation of integrals with arbitrary powers of the radial coordinate. The
integration over the angles yields the same result as in subproblem (a).
Hence,
(9.145)
Fig. 9.16 Probability density function of the radial coordinate in the 1s ground state of the hydrogen
atom. The maximum and thus the most probable distance of the electron is at r = a 0 ≈ 0. 529 × 10−10
m. The red line marks the expectation value calculated in subproblem (b)
This result may look astonishing at first sight. We must not confuse it with
the most probable radius of the electron, which is given by the maximum of
the probability density function w(r) = | ψ 1s | 2 r 2 shown in Fig. 9.16. Its
maximum is at r = a 0, i.e., the most probable distance of the electron in the
ground state is identical to the Bohr radius. The expectation value for r we
have just computed is at 1. 5a 0. As can be seen in Fig. 9.16, the probability
density decreases steeply for large distances, and also for small distances of
the electron from the nucleus. In subproblem (c), we calculate the
probability of finding the electron within the nucleus. The notion of finding
the electron in the nucleus may seem strange at first. However, such a
calculation is of importance for the estimation of the probability of a certain
type of elementary particle reaction predicted in the year 1935 by Hideki
Yukawa :
(9.146)
The so-called K-capture of some isotopes such as18 37Ar is the reaction of an
electron of the electron shell (most likely the K-shell) with a proton of the
nucleus, converting the latter into a neutron and an electron neutrino ν e .
Therefore, what is the probability of finding the electron within the hydrogen
nucleus? We assume a radius of the nucleus of r N = 10−15 m. Obviously, this
probability is:
(9.147)
The difference from our calculation in subproblem (a) is that we have to
integrate from r = 0 to r = r N . From the integral table in the Appendix we use
Eq. (A.44) and obtain:
(9.148)
Hence, the probability of finding the electron within the nucleus is only
about 10−14. Given one mole of hydrogen atoms, there are, nevertheless,
about 6 billion atoms in which the electron is located within the nucleus.
Problem 9.15 (Hydrogen Problem Applied to Semiconductor
Technology)
Solution 9.15
This exercise includes an adaption of Bohr’s atom model to a special
problem in semiconductor physics. We characterize the spectrum of energy
levels of an n-doped Si crystal. Such a simple treatment already provides
essential features without becoming too involved in the details of solid state
physics. Silicon is a semiconductor , i.e., under ambient conditions it is
neither an electric insulator nor is it a real conductor. The reason for this can
be explained using a simplified band model as shown in Fig. 9.17 (left-hand
diagram). Electric conduction requires free charge carriers in the conduction
band. To excite an electron from the filled valence band to the conduction
band, however, the energy of the band gap E g = 1. 12 eV is needed at least.
At room temperature, the thermal energy k B T is only 26 meV. Hence, a
thermal excitation is unlikely and free charge carriers are not available under
room temperature conditions.
In subproblem (a) we consider an isolated arsenic atom embedded in the
silicon crystal lattice. The latter is described as a continuum with a dielectric
constant of ε = 11. 7. Compared with silicon with electron configuration
[Ne]3s23p2, arsenic with electron configuration [Ar]4s24p3 has an excess
valence electron. We treat this excess electron using the hydrogen problem
(Eq. (9.21)) with the following modifications: (1) we assume an effective
charge of the ion core12 of Z eff e = 1e. (2) we assume an effective mass 13 of
the excess electron m ∗ = 0. 3. Taken together, the modified Rydberg formula
for the energy levels of the excess electron is:
(9.149)
with Ry being 13.606 eV (see Eq. (9.22)). The first three energy levels are
thus E 1 = −29. 8 meV, E 2 = −7. 5 meV, E 3 = −3. 3 meV relative to the lower
edge of the conduction band. Note that ionization of the arsenic atom (n → ∞)
causes the generation of one free charge carrier. The ionization energy for
this process is just thus 29.8 meV, which is close to the thermal energy at
room temperature. As a consequence, according to this simple model, doping
of silicon crystals with arsenic increases its room temperature conductivity.
(9.150)
with a 0 = 0. 5292 × 10−10 m. The radius r 1 of the excess electron in its
ground state is thus 21 × 10−8 cm. A sphere with this radius has a volume of
cm3. Using the given density of ρ = 2. 34 g cm−3 and the
molar mass of silicon M Si = 28. 09 g mol−1, the number of silicon atoms
within this sphere is 2000. Hence, the continuum treatment of the
silicon crystal as a dielectricum seems plausible.
Finally, in subproblem (c), we consider the case in which the arsenic
defects are no longer isolated. In our simple model, we estimate the critical
concentration of arsenic atoms where the orbits of the excess electrons
penetrate each other. If we consider a cube with an edge length 2r 1 = 42 ×
10−8 cm, the orbits of neighboring arsenic defects would touch each other.
This cube has a volume of (2r 1)3 = 7. 4 × 10−20 cm3 with one arsenic defect
at each of its corners, contributing of its sphere. Hence, within such a cube,
there is arsenic defect and the critical defect density is thus:
(9.151)
The consequences for the band structure are shown in Fig. 9.18. Although the
energy levels of isolated excess electrons are sharply defined, the mutual
perturbation of neighboring defects leads to a splitting of these levels into a
quasi-continuum of possible defect states and thus the formation of a band
tail located beneath the lower edge of the conduction band.
(9.152)
| ψ〉 is an appropriate test function depending on one or more
optimization parameters that are varied during the minimization
process. With this test function is the best approximation of the
ground state energy that can be obtained.
a.
Consider the Hamiltonian of the hydrogen atom and use the
Gaussian function
(9.153)
with the normalization constant N and the variational parameter
α as a test function to determine E 0 as an estimation of the
ground state energy. Compare your result with the exact ground
state energy. Hint: It is convenient to introduce atomic units.
(9.154)
where α 1 = 2. 87 and α 2 = 0. 28 are fixed exponents and c 1
and c 2 are the variational parameters. Hints: Write the energy
functional in the form
(9.155)
Transform the equations for the minimum condition in a
system of equations for the coefficients c 1 and c 2 . Exploit
the condition that the denominator D is 1 if ψ is normalized.
Solution 9.16
In this exercise, we deal with approximative methods for the solution of the
time-independent Schrödinger equation. The variational principle is the basis
of many methods in quantum chemistry. Concepts such as the effective charge
we have dealt with in Problem 9.5 may be quite useful. However, there is no
systematic way of improving the results. Of high value in atomic and
molecular structure theory are concepts that may be systematically improved
– even if the improvement leads to an increase in computational costs.
Starting from a suitable test function with one or more adjustable parameters,
the energy functional is calculated and minimized by variation of the
parameter set. The minimum provides the best estimation for the ground state
energy with the chosen test function. Here, we apply a method to the
hydrogen atom, a system for which we know the exact result of the ground
state energy, the negative Rydberg constant (see Eq. (9.22)), or E exact = −0. 5
E h in atomic units.
In subproblem (a), we consider a single Gaussian to be a test function
given in Eq. (9.153). At first sight, it may seem artificial to treat the hydrogen
problem using a Gaussian function. However, because of their properties,
Gaussian functions are widely used in quantum chemistry to construct wave
functions. Hence, our problem is instructive. We start by writing down the
Hamiltonian of the hydrogen problem. In the ground state (1s state, l = 0), the
effective Hamiltonian in atomic units (Sect. 9.1.4) is:
(9.156)
where is the operator of the kinetic energy of the electron. V is its potential
energy, i.e., the attractive Coulomb potential with the core. Our goal is to
calculate Eq. (9.152). We start with the denominator, where we assume that
our test function is normalized (see Problem 9.14), i.e.
(9.157)
Thus,
(9.158)
and the normalization constant is therefore:
(9.159)
The numerator splits into two parts, one with the kinetic energy, and one with
the potential energy:
(9.160)
The second term is:
(9.161)
Because
(9.162)
and
(9.163)
the kinetic energy term is:
(9.164)
where we have again used the integral table in the Appendix (Eq. (A.49)).
This expression reduces to:
(9.165)
Using Eqs. (9.160), (9.161), and (9.165), we obtain the expression for
the energy functional:
(9.166)
Now, we can minimize the functional by setting the first derivative to zero:
(9.167)
Solving for α min, we obtain:
(9.168)
Reinserting the optimal exponent into Eq. (9.166), we obtain an
estimation of the ground state energy:
(9.169)
Fig. 9.19 Comparison between the exact 1s wave function of the hydrogen atom (solid line) and a
Gaussian test function with exponent α = 0. 2829 (dashed line)
The use of a single Gaussian as a test function reaches about 85% of the
true ground state energy E exact. The optimized test function is shown in Fig.
9.19 together with the exact ground state wave function.14 Both functions are
quite different, both in the core region and at larger distances from the core,
where the exact wave function decays more rapidly than the Gaussian
function. As mentioned above, the variational principle offers the possibility
of arriving at an improved result if we use a more flexible test function. This
is examined in subproblem (b), which, however, is quite tedious. Our test
function has now two fixed primitive Gaussian functions. The exponent α 2 =
0. 28 is essentially the optimal value obtained in subproblem (a), whereas
the Gaussian function with α 1 = 2. 87 is more pronounced in the near-core
region. We begin by constructing the energy functional, which should be
written in the form given in Eq. (9.155). Using Eq. (9.154) the denominator
is:
(9.170)
The coefficients S 11, S 12, and S 22 depend only on the given exponents α 1 =
2. 87 and α 2 = 0. 28. Their values are S 11 = 0. 40491, S 12 = 0. 99600, and S
22 = 13. 28749. Later, we exploit the fact that the expression Eq. (9.170),
which is the norm of the test wave function, is unity. The numerator of the
energy functional consists of two terms, the kinetic energy and the
potential energy. For the latter, we have:
(9.171)
(9.172)
Here, we have made use of Eqs. (9.162) and (9.163). Further evaluation
yields:
(9.173)
The integrals are again evaluated using Eq. (A.49) in the Appendix:
(9.174)
If we introduce
(9.175)
then the coefficients are
(9.176)
(9.177)
(9.178)
The numerator of the energy functional is thus:
(9.179)
where H 11 = 0. 64850, H 12 = −1. 23239, and H 22 = −5. 63923. We have now
constructed the energy functional in the form Eq. (9.155). In the next step,
must be minimized by variation of the coefficients c 1 and c 2, i.e., by taking
the derivatives with regard to c 1 and c 2:
(9.180)
A direct evaluation of the resulting equations is tedious, because they can
barely be solved for these coefficients. If we consider N as the numerator of
the energy functional and D as its denominator, we use the product rule:
(9.181)
The denominator is nonzero. Thus,
(9.182)
With
(9.184)
These equations constitute a system of equations for the coefficients c 1 and c
2:
(9.185)
Nontrivial solutions are obtained if the secular determinant is zero:
(9.186)
This is a quadratic equation for the energy functional with the two roots:
(9.187)
where A = S 11 S 22 − S 12 2, B = 2H 12 S 12 − H 11 S 22 − H 22 S 11, and C = H 11
H 22 − H 12 2. The two roots are , and . The
first root apparently corresponds to a maximum energy, the second to the
minimum energy sought. If we compare it with the result from subproblem
(a), −0.4244 E h , we conclude that the addition of a second Gaussian
improves the estimate for the ground state energy to about 95% of the exact
result, −0.5 E h . The results are summarized in Table 9.2.
Table 9.2 Estimated ground state energies of the hydrogen problem based on the variational principle in
comparison with the exact result
Having obtained the optimal value for , our next goal is the
determination of the variational parameters c 1 and c 2 for the negative
solution. We insert in Eq. (9.183). This yields:
(9.188)
The second condition is the normalization of the test wave function:
(9.189)
Substituting c 2 and solving for c 1 we obtain:
(9.190)
Fig. 9.20 Comparison of the exact 1s wave function of the hydrogen atom (solid line) with the
optimized test functions with two Gaussian functions (dashed line)
Fig. 9.22 Linear chain of K+ and Cl− ions with a vacancy as a model of a lithium doped KCl crystal.
Due to the quantum mechanical tunnel effect, the Li+ ion can switch between the two stable positions
(9.191)
with the two minima at x = ±x 0 (x 0 = 1. 38 a 0) and a local maximum V
(0) = η = 5 × 10−4 E h (1 E h = 27. 12 eV, a 0 = 0. 5292 Å).
(9.192)
Write the Hamiltonian in the form:
(9.193)
where is the Hamiltonian of the harmonic oscillator, and
U(x) is an effective potential of the particle.
c.
Show that the solution of the Schrödinger equation is equivalent
to the solution of the eigenvalue problem
(9.194)
where
(9.195)
Solution 9.17
The context of this quantum mechanical problem is related to solid-state
chemistry and physics. More precisely, we deal with a simplified model of
lithium-doped potassium chloride (Li:KCl). We have already considered the
latter as an example of a system that exhibits a Schottky anomaly (see
Problem 8.8). Figure 9.22 shows a linear chain of K+ and Cl− ions. Near x =
0 one potassium is replaced by a Li+ cation. As the latter has a smaller ionic
radius, it has two stable equivalent positions and the ion can switch between
these two positions, either by thermal motion or by means of the quantum
mechanical tunnel effect . In the latter case, the lithium ion tunnels through
the potential barrier between the two positions. The potential of the lithium
in the vacancy, Eq. (9.191), is a symmetric double-well with minima at ± x 0.
It is plotted in Fig. 9.23 using the parameters η and x 0 given in the problem
text. The use of atomic units simplifies the calculations in the following.
Distances are thus defined in multiples of the Bohr radius a 0, energies in
multiples of hartree energy E h . The form of the potential can be explained
by strong repulsive interaction between Li+ and Cl− for | x | > 2 a 0, and
electrostatic interaction between Li+ and all ions in the chain, which,
compared with the interaction at x = 0, is slightly more attractive if the Li+
moves closer to the neighboring Cl−.
Fig. 9.23 Double-well model potential of the lithium ion in the vacancy (solid line). Also shown
(dashed line) is the reference harmonic oscillator potential used for the solution of the quantum double
well problem
(9.197)
where are the known energy eigenvalues of the harmonic
oscillator with an angular frequency ω. The term in brackets in Eq. (9.196) is
the Hamiltonian of the quantum double-well.
Following the instruction given in subproblem (b), we rewrite the
Hamiltonian as a sum of the harmonic oscillator Hamiltonian plus an
effective potential U(x):
(9.198)
The new potential function U(x) is simply the difference between the
double-well potential and the harmonic oscillator potential, where we are
free to choose a suitable value for ω.
In subproblem (c), we use this special form to show that it leads to an
eigenvalue problem. Switching to bra-ket notation, the expansion of the
unknown wave function in harmonic oscillator states | n〉 is
(9.199)
The Schrödinger equation can thus be written
(9.200)
The trick is to multiply from the left with 〈ψ | = ∑ m c m ∗〈m | :
(9.201)
We thus obtain the eigenvalue equation Eq. (9.194), and because
, the matrix elements H mn are given by:
(9.202)
which is simply Eq. (9.195) in bra-ket notation. The determination of the
matrix elements H mn is our next task in subproblem (d). We could solve the
integrals in Eq. (9.195) numerically using explicit forms of the harmonic
oscillator wave functions in the Appendix Eq. (A.76) in combination with the
recurrence relation Eq. (A.77). Here, however, we can benefit from Dirac’s
abstract bra-ket notation. We solve the integrals without explicit usage of the
harmonic oscillator wave functions, merely by using harmonic oscillator
algebra based on creation and annihilation operators introduced in
Problem 9.10, Eqs. (9.86) and (9.87). Using these equations, it can be shown
that the position operator can be written
(9.203)
The position operator can be inserted into the expression containing the
effective potential:
(9.204)
(9.205)
Using Eq. (9.203), we obtain:
(9.206)
and
(9.207)
Note that the order of the operators in the various terms matters. If we
analyze all 20 operator terms in this way, we obtain:
(9.209)
(9.210)
The resulting potential of the reference harmonic oscillator is shown in
Fig. 9.23 (blue dotted line). A numerical procedure to diagonalize the
Hamiltonian matrix (Eq. (9.209)) can be implemented in various ways, for
example, by using mathematical software or a programming language. In the
latter case, powerful linear algebra libraries are available.16
Fig. 9.25 The first six energy levels of the quantum double-well, as obtained from the solution of Eq.
(9.194). The splitting between the ground state (n = 0) and the first excited state (n = 1) is very small (Δ
= 3. 4 × 10−6 E h )
n E (10 −4 E )
n h
0 1.8996
1 1.9339
2 4.7630
3 5.5326
4 7.4947
5 9.4820
For the visualization of the wave functions, we use Eq. (9.192) and the
eigenvectors of the various states. We need the general form of the harmonic
oscillator wave functions Eq. (A.76) and the recurrence relation Eq. (A.77)
for the Hermite polynomials to construct the double-well wave functions.
The first six solutions are shown in Fig. 9.26.
Fig. 9.26 Wave functions in the double-well potential (short dashed) of the ground state (n = 0) and
the first excited states. The wave functions are alternately either symmetrical (n = 0, 2, 4) or
asymmetrical (n = 1, 3, 5)
It is striking that states with even n are also even functions, i.e. ψ n (x) =
ψ(−x) for n = 0, 2, 4, whereas the states with odd n are asymmetrical: ψ n
(−x) = −ψ(x) n for n = 1, 3, 5. The ground state wave function and the solution
for n = 1 have a high amplitude and thus high probability17near the stable
sites of the Li+ at x = ±x 0. However, for the ground state in particular, there
is also a nonzero probability of finding the Li+ within the potential barrier,
and this causes the possibility of tunneling through the barrier. In contrast, the
states n ≥ 2 with energies comparable with the energy barrier or higher have
a considerable probability in the region of the barrier, although, by symmetry,
odd states have | ψ n (0) | 2 = 0.
It is worth mentioning that the model of the one-dimensional atomic chain
in Fig. 9.22 can only yield qualitative results for the description of the
Li:KCl system. A K+ vacancy in a KCl crystal has a three-dimensional
structure and the potential of a Li+ ion has not two, but eight local minima.
The simple one-dimensional model, however, contains the essential physics
to understand the behavior of such multiwell systems. In summary, we have
learned how harmonic oscillator wave functions can be used as a complete
orthogonal set of functions to solve the Schrödinger equation for the special
case of a quantum double-well potential. The method is general and can be
used for other potentials, e.g., the asymmetric quantum double-well .
(9.213)
as a function of the internuclear distance R. Hint: Introduce the
confocal elliptical coordinates
(9.214)
(9.215)
and the angle ϕ ∈ [0, 2π] defined in Fig. 9.27. The volume
element expressed in these coordinates is
.
(9.216)
and the exchange integral
(9.217)
as a function of R.
d.
Insert the trial function Eq. (9.212) into the equation
(9.218)
for the variational principle and derive an eigenvalue equation
for the electron energy. Show that one of the two solutions
yields to a bonding of the two nuclei. Determine c 1 and c 2.
Fig. 9.27 Geometry of the H2 + molecular ion. The origin of the molecule-fixed coordinate system is
in the middle of the connecting line; nuclei 1 and 2 are at positions . The electron at position
r and the two nuclei define a plane that is tilted by an angle ϕ with regard to the x-axis
Solution 9.18
Perhaps the most important application of quantum mechanics in chemistry is
the explanation of chemical bonds from first principles. Contemporary
methods in quantum chemistry provide accurate descriptions of the electronic
structure of molecules and solids. However, these methods typically require
great numerical effort. Paper and pencil approaches are possible for only the
simplest systems, and, moreover, in an approximate manner. Nevertheless,
they allow a qualitative understanding of the nature of the chemical bond.
Here, we deal with the simplest possible molecular system, the H2 +
molecular ion. The geometry is shown schematically in Fig. 9.27. The goal of
the exercise is to determine the potential between the two nuclei as a function
of the internuclear distance R. In the electronic ground state, we expect the
potential energy curve
(9.219)
to exhibit a minimum. In atomic units18, which we use in the following,
is the electrostatic repulsion of the two positively charged nuclei. E
el is the electronic energy acting as an effective potential, calculated at a
given bond distance R. The Schrödinger equation of the electron moving in
the Coulomb potential of the two nuclei is:
(9.220)
where is the distance of the electron from nucleus 1, and
is the distance of the electron from nucleus 2 respectively, and
e z is the unit vector in the z-direction.
The Hamiltonian sought in subproblem (a) is thus:
(9.221)
It is worth noting that this quantum mechanical problem can be solved
precisely by introducing confocal elliptical coordinates [5]. The approximate
solution based on the variational principle, however, shows much of the
spirit of a quantum chemical treatment. The variational principle was already
used in Problem 9.16. The strategy is to use a suitable test function
containing a number of adjustable parameters and to minimize the total
energy. Improvement of the test function, for example, by increasing the
number of parameters, is a systematic approach to the exact solution. In this
problem, the test function is a linear combination of hydrogen 1s wave
functions centered at the sites of the two nuclei:
(9.222)
Apparently, the coefficients c 1 and c 2 are the two adjustable parameters.
Note that in the limiting case R = 0 where the two nuclei constitute a helium
core with a nuclear charge Ze = 2e, this test wave function does not coincide
with the exact ground state solution of the He+ problem, .
In subproblem (b) and subproblem (c), we deal with three different
integrals involved in the treatment of the H2 + ion. The first integral is the
overlap integral
(9.223)
For the solution we switch to a representation using the confocal elliptical
coordinates μ and ν introduced in Eqs. (9.214) and (9.215), in addition to the
angle ϕ. It is the angle of the plane defined by the sites of the nuclei and the
electron relative to the xz-plane. The significance of the coordinate μ
becomes clearer if we imagine that the electron resides on an ellipse, whose
two focal points are the sites of the nuclei. As each point on an ellipse leaves
constant the sum of distances to the focal points, and because R μ = r 1 + r 2,
the coordinate μ has an analog in the radial distance of the electron in the
hydrogen problem formulated with spherical coordinates: the curve of
constant μ is an ellipse, as the sphere is an area of constant radial distance.
The other coordinate, ν, is related to the difference in the distances between
the electron and the nuclear sites. If ν is varied at constant μ, the electron
moves on the ellipse. An analogy to the quantity cosθ in spherical
coordinates is thus established. With the volume element given in the
problem text, the overlap integral is thus:
(9.224)
Integration over ϕ is trivial and yields a factor 2π. The separation with
regard to the other coordinates yields:
(9.225)
Because
(9.226)
and
(9.227)
the result for the overlap integral is
(9.228)
Fig. 9.28 Overlap integral S, Coulomb integral C, and exchange integral A as a function of the distance
between the two hydrogen nuclei
S(R) is plotted in Fig. 9.28. The overlap integral diminishes rapidly for
increasing R. The second integral is the Coulomb integral Eq. (9.216),
(9.229)
Because and , the introduction of elliptical
coordinates yields:
(9.230)
Integration over ϕ is again trivial and further simplification is possible by
means of the third binomial formula:
(9.231)
Separation of the integrals yields:
(9.232)
where
(9.233)
(9.234)
(9.235)
With these results, the Coulomb integral is:
(9.236)
This function is shown in Fig. 9.28 (short dashed line). At large R, the
integral diminishes as R −1 and thus has a long range. The calculation of the
third integral, called the exchange integral, uses the same methods:
(9.237)
Introduction of elliptical coordinates and integration over the angle ϕ yields
(9.238)
After insertion of Eqs. (9.227) and (9.233), the result for the exchange
integral is:
(9.239)
It is also plotted in Fig. 9.28. In subproblem (d), we apply the variational
principle. From now on, it is convenient to switch to Dirac’s bra-ket
notation. We exploit the fact that the hydrogen 1s wave function is
normalized, as we have shown in Problem 9.14. Moreover, we assume that
the mixing parameters c 1 and c 2 are real. The denominator of the energy
functional Eq. (9.218) is thus:
(9.240)
For the calculation of the nominator of the energy functional we take into
account that the Hamiltonian Eq. (9.221) acting on either of the two wave
functions yields:
(9.241)
and
(9.242)
where E 0 = −0. 5 E h = −13. 605693 eV is the ground state energy of the
hydrogen atom. Thus, the nominator of the energy functional is19
(9.243)
The necessary conditions for a minimum as required in Eq. (9.218) is:
(9.244)
As shown in Problem 9.16 (Eq. (9.182) on page 269), these conditions
are equivalent to:
(9.245)
and
(9.246)
As a consequence, we arrive at the following system of equations:
(9.247)
(9.248)
With the abbreviations H 11 = E 0 − C and H 12 = E 0 S − A, the system is:
(9.249)
A nontrivial solution requires the secular determinant to be zero:
(9.250)
This is a quadratic equation:
(9.251)
with two solutions (see Eq. (A.4))
(9.252)
(9.253)
and simplified even further to yield the two solutions:
(9.254)
and
(9.255)
The two energy eigenvalues E g and E u of the electron can thus be expressed
by the integrals, S(R), C(R), A(R), and the ground state energy of the
hydrogen atom, E 0. Focusing on the entire system of the H2 + ion, we identify
these solutions as the electronic contribution E el to the total potential energy
according to Eq. (9.219). The resulting potential curves are shown in Fig.
9.29. It is striking that the blue curve related to the energy eigenvalue E g has
a minimum at R min = 2. 5a 0, whereas the red curve has a monotonic behavior
that is thus repulsive at all distances. We have therefore shown that one of the
two solutions leads to bonding of the two nuclei. Furthermore, both potential
curves have the same long-range behavior:
(9.256)
Fig. 9.29 Potential between the hydrogen nuclei in the H2 + molecular ion, obtained from an
approximate solution of the one electron Schrödinger equation using the variational method
This is because the last term compensates for the electrostatic repulsion
of the nuclei at long distances due to the long-range behavior of the Coulomb
integral (see Eq. (9.236)) whereas the overlap and exchange integrals are
negligible at large distances. The value V (R) = −0. 5 at large distances is the
total energy of a neutral hydrogen atom and a hydrogen cation. If we neglect
the zero-point energy, the formation of a stable H2 + molecular ion according
to
(9.257)
involves an energy change V (R min) − E 0 = −0. 065 E h = −1. 76 eV. To
determine the two coefficients, c 1 and c 2, we insert the two solutions for
(Eqs. (9.254) and (9.255)) into Eq. (9.249). For the solution with energy E u ,
this yields:
(9.258)
Moreover, normalization of the total wave function requires:
(9.259)
Therefore, we obtain:
(9.260)
and thus
(9.261)
The wave function is thus:
(9.262)
In a completely analogous way, we treat the solution with the eigenvalue E g .
Here, we find
(9.263)
and, as a consequence, the ground state wave function is:
(9.264)
In the terminology of quantum chemistry, ψ g and ψ u are molecular orbitals ,
constructed by a linear combination of the two atomic orbitals ψ 1 and ψ 2
centered at the sites of the nuclei. Both wave functions are plotted for R = R
min = 2. 5a 0 in Fig. 9.30. Apparently, the ground state ψ g is symmetrical with
regard to an inversion of the molecule, whereas the excited state ψ u is
antisymmetric with regard to inversion.
Fig. 9.30 Approximate solutions for the electronic wave functions of the H2 + ion based on the
variational principle: (a ) ground state ψ g according to Eq. (9.264). (b ) excited state ψ u according to
Eq. (9.262). Using inversion symmetry, ψ u is characterized by zero electronic density in the center of
mass
It is worth comparing ψ g and ψ u with the ground state (n = 0) and the
first excited state (n = 1) in the quantum double-well problem (Problem 9.17,
Fig. 9.26 on page 280). In both problems, the symmetrical solution is
energetically favorable compared with the asymmetrical solution. An
argument for the bonding nature of the symmetrical solution in the H2 + ion is
the nonzero electronic charge density among the hydrogen nuclei. In the
asymmetrical solution ψ u , in contrast, the electronic density is repelled from
the center of the molecule by symmetry.
In summary, we have applied the variational principle to the simplest
molecular system, the H2 + ion, containing only one electron. We have shown
that ground state electronic energy overcompensates for the electrostatic
repulsion of the two hydrogen nuclei at a distance of R = 2. 5a 0. It is worth
noting that the construction of molecular orbitals from atomic orbitals using
linear combination is not restricted to 1s orbitals. Extended test wave
functions with additional p orbitals, for example, allow polarization effects
to be included in the model. The method thus allows a systematic
improvement of the solution.
References
1. Gearhart CA (2002) Planck, the quantum, and the historians. Phys Perspect 4:170
[CrossRef]
2. Rud Nielsen J (1976) Niels Bohr collected works, vol 3. The correspondence principle. North-
Holland, Amsterdam
3. Schrödinger E (1921) Versuch zur modellmäßigen Deutung des Terms der scharfen Nebenserien. Z
Phys 4:347
[CrossRef]
5. Grivet JP (2002) The hydrogen molecular ion revisited. J Chem Educ 79:127
[CrossRef]
Footnotes
1 The number of postulates differs from textbook to textbook, although they cover the same content.
2 The Kronecker delta δ mn is defined as follows: δ mn = 1 if m = n, and δ mn = 0 if m ≠ n.
3
If M ≫ m e is the mass of the nucleus, the effective mass is close to m e . In the case of
4 For explicit expressions of R nl (r) and Y lm (θ, ϕ) see Sects. A.3.14 and A.3.15.
5 Sample heating to high temperature, also called tempering , is, for example, needed to obtain well-
defined, clean and crystalline materials.
6 In Problem 7.4 we have seen that heat conduction mediated by particle collisions is largely
suppressed under ultrahigh vacuum conditions.
8 XPS X-ray photoelectron spectroscopy, UPS ultraviolet photoelectron spectroscopy, ESCA electron
spectroscopy for chemical analysis.
13 The justification of an effective mass in solid state physics is based on the fact that the motion of an
electron in the crystal is not free. While maintaining the expression for the kinetic energy,
, the effective mass can be derived from the curvature of electron bands at special
14 We have dealt with the exact solution in Problem 9.14 on page 257. In the literature, atomic orbital
functions with exponential decay in the region far from the core are also called Slater functions.
15 Harmonic oscillator wave functions are illustrated in Fig. 9.1. Explicit formulas are provided in the
Appendix, Sect. A.3.13.
18 See Sect. 9.1.4. Note that in atomic units the unit length is 1a 0, and the unit energy is 1 E h .
10. Spectroscopy
Jochen Vogt1
(1) Chemisches Institut der Universität Magdeburg, Magdeburg, Germany
Abstract
Spectroscopy is an important experimental technique for determining
properties of atoms and molecules, predominantly by means of their
interaction with electromagnetic waves. The foundations of spectroscopy are
closely related to the quantum states of matter; thus, the basic concepts in part
recapitulate the contents of quantum mechanics (Chap. 9). The presented set
of problems deals with different kinds of spectroscopy in the various regions
of the electromagnetic spectrum. One focus is on problems that demonstrate
how quantitative analysis of spectra provides detailed information on
molecular structure and bonding. Other problems deal with the principle and
dynamics of the laser, which is the most important laboratory light source in
contemporary spectroscopy.
(10.5)
and the spectral absorbance
(10.6)
Note that in the literature the absorbance is sometimes based on the natural
logarithm, A e (ν) = −ln T(ν). Moreover, in the presence of light beam
attenuation due to scattering losses, the terminus extinction or attenuance is
used instead of the absorbance.
Moreover, a quantitative description of absorption and emission
processes sketched in Fig. 10.1 can be based on rate equations involving the
spectral energy density of the electromagnetic field, u(ν), and the Einstein
coefficients . The probability per second that an atom or molecule absorbs a
photon of frequency ν is given by:
(10.7)
where B 12 is the Einstein coefficient for induced absorption. Accordingly,
the probability per second for induced emission of an excited atom or
molecule is:
(10.8)
where B 21 is the Einstein coefficient for induced emission. The probability
per second of spontaneous emission does not depend on the energy density of
the electromagnetic field:
(10.9)
A 21 is the respective Einstein coefficient for spontaneous emission. The
lifetime τ of the excited state with regard to spontaneous emission is:
(10.10)
Fig. 10.2 Light scattering processes. Stokes scattering (left) starts from a low-lying initial state, and
the emitted photon has a lower frequency than the absorbed one (red shift). Rayleigh scattering
(middle) involves the same level as the final and initial state. Anti-Stokes scattering (right) starting from
a state of higher energy results in a blue shift of the emitted photon
Higher order processes involve the absorption and emission of more than
one photon. Light scattering processes, for example, are based on the
absorption of a photon under excitation of the scattering atom or molecule on
a virtual intermediate state, followed by the emission of a second photon
and the subsequent transition back to the initial state (Rayleigh scattering), or
to another final state (Raman scattering). Raman scattering is illustrated in
Fig. 10.2 for the case of Stokes scattering and anti-Stokes scattering .
(10.12)
B is related to the rotator’s moment of inertia
(10.13)
with the effective mass μ depending on the masses of the two centers:
(10.14)
As textbooks show, the solutions of the Schrödinger equation for the
dumbbell-shaped rigid rotator are the spherical harmonics Y JM (θ, ϕ) (see
Sect. A.3.14 in the appendix). The two angles θ and ϕ completely define the
orientation of the molecule. Because the energy levels Eq. (10.11) do not
depend on the orientational quantum number M, they are degenerated. The
degeneracy corresponds to the number (2J + 1) of possible M values for a
given J. Rotational transitions mediated by the absorption or emission of
light are bound to selection rules. Pure rotational transitions require a
permanent electric dipole moment and, moreover
(10.15)
As a consequence of the energy levels Eq. (10.11) and these selection rules,
the pure rotation spectrum exhibits equidistant transitions, as shown in Fig.
10.4.
Fig. 10.4 The energy levels of the rigid diatomic rotator and the resulting spectrum of equidistant lines
(10.17)
where n is the vibrational quantum number . The selection rules for
transitions of an harmonic oscillator between two states n and n ′ are:
(10.18)
Moreover, an induced electric dipole moment must be involved with the
vibrational transition. Anharmonicity of the potentials between atomic cores
lowers the energy of a vibrational transition. The generally observed relation
(see Problem 10.6) is
(10.19)
where x e is an anharmonicity constant.
10.2 Problems
Additional problems related to spectroscopy can be found in Chap. 9.
b.
For the following data, provide the missing values of E, ν, λ,
and . Assign the data to the respective range within the
electromagnetic spectrum.
E = 1 eV.
λ = 21 cm (hyperfine splitting of hydrogen)
Solution 10.1
In this exercise, we familiarize ourselves with the basic units used in
spectroscopy. The use of different units in the various spectroscopic
disciplines may seem confusing at first sight. However, depending on the
range of the electromagnetic spectrum (Table 10.1), where transitions of
atoms and molecules occur, some units are indeed more appropriate than
others. In subproblem (a), we give the relationships among energy,
frequency, wavelength, and wave number. The relation between energy and
frequency is established by:
(10.20)
where h = 6. 62606957(29) × 10−34 J s
is the Planck constant (see Sect. A.1
in the appendix). The relation between frequency and wavelength is:
(10.21)
where c = 299, 792, 458 m s−1
is the vacuum speed of light.3 Finally, the
wave number is the reciprocal wavelength:
(10.22)
In subproblem (b), we look at some instructive examples. The first example
is the energy E = 1 eV, i.e., the kinetic energy, that a particle with charge e
gains if it passes a potential difference of 1 V. This is the unit electron volt
that is common in many spectroscopic disciplines. In SI units, this energy is
E = 1. 60218 × 10−19 J, and the frequency is:
(10.23)
Note that it is more common to express the wave number in cm−1 than in m−1.
Fig. 10.5 Survey over the electromagnetic spectrum, classified according to the spectral ranges given
in Table 10.1. Also indicated are the positions of characteristic spectroscopic transitions of
Problem 10.1b
We assign these values to the infrared range (IR); more precisely, the
near infrared (NIR) range, as also indicated in Fig. 10.5 along with the other
examples of this subproblem. The second example is the hyperfine splitting
of hydrogen with its characteristic transition at λ = 21 cm, resulting from the
flipping of the electron spin relative to the orientation of the nuclear spin. It
is within the microwave spectral range of the electromagnetic spectrum. In
radio astronomy the 21 cm line is key for the measurement of hydrogen in
space. Conversion to wave number yields cm−1, an energy of E =
5. 9 μeV, and a frequency of ν = 1. 4 GHz. The third example is not related to
a special transition, it is the thermal energy at room temperature,
(10.24)
where v x is the velocity component of the atom or molecule in line
with the wave field, and v x ≪ c (see Fig. 10.6).
a. Assume a Maxwell-Boltzmann velocity distribution of the gas
particles and show that the Doppler effect causes a Gaussian
line profile with half width
(10.25)
where T is the temperature, and m the mass of the gas particles.
Solution 10.2
Fig. 10.6 Absorption spectroscopy probing atoms or molecules in a gas cell (schematic)
(10.27)
Next, we take the motion of the molecules, which causes the Doppler
effect, into account. If an emitter of a wave is in motion, the frequency of the
wave is higher if the emitter moves toward the observer, and it is lower if the
emitter goes in the opposite direction. Usually, textbooks show that this
Doppler effect is a consequence of wave theory, but Eq. (10.24) can also be
shown in a rigorous way in the photon picture of light, if we take into account
the recoil of a photon emitting atom or molecule [1]. In absorption
spectroscopy, however, the light source can be considered at rest, whereas
the gas particle takes the role of the ”observer” In this case, the same
relation, Eq. (10.24) holds:
(10.28)
Note that v x is the velocity component of a certain molecule in line with
the wave vector of the electromagnetic field. The fact that the field probes
many molecules, each moving with a different velocity, is taken into account
by assuming a Maxwell–Boltzmann distribution of the particle velocity. In
Problem 7.1, we have seen that the velocity distribution for the velocity
component in the x-direction is a normalized Gaussian distribution. Thus, the
number density of particles with the velocity v x is:
(10.29)
Here, we have introduced the most probable velocity v m derived in Eq.
(7.19). We annotate that:
(10.30)
The decisive step now is the transformation of this velocity distribution
into a distribution of the number density as a function of resonance frequency.
Dependent on v x , the atom or molecule absorbs radiation at a slightly
different frequency. Solving Eq. (10.24) for v x , we can make the following
substitutions:
(10.31)
and
(10.32)
Equation (10.30) may be written as:
(10.33)
(10.34)
In comparison with the last equation with the normal distribution Eq.
(A.64) found in the appendix, we can identify its standard deviation
(10.35)
The half-width of the Doppler-broadened spectral line is thus obtained using
Eq. (A.65), which relates the half-width to the standard deviation:
(10.36)
If we replace v m using Eq. (7.19), we finally obtain Eq. (10.25), which
was shown.
In subproblem (b), we calculate the expected Doppler width for a
special rovibrational transition of the CO molecule at room temperature at
cm−1. Using the relation between wave number and frequency,
, we obtain the expression for the Doppler width in wave numbers,
which is:
(10.37)
Hence,
(10.38)
Therefore, the line width sought is:
(10.39)
corresponding to Δ ν = 48 MHz. This result is in fact close to the observed
line width at a low pressure in the gas cell. Note that in typical
undergraduate spectroscopy laboratories the resolution is usually not
sufficient to determine this line width. In these cases, the observed line width
is mostly determined by the spectrometer itself.
In subproblem (c), we deal with another example, the well-known
sodium D-line doublet in the visible part of the electromagnetic spectrum.
From the periodic table in the appendix, we take the atomic mass of sodium,
22.99 m u . Thus,
(10.40)
300 K 430 K
300 K 430 K
Δ ν(D1) 1.315 GHz 1.575 GHz
Δ ν(D2) 1.316 GHz 1.576 GHz
Using Eq. (10.25) and the given transition frequencies 508.332466 THz
(D1 line) and 598.848717 THz (D2 line) respectively we obtain the values
for the Doppler width shown in Table 10.2. The values are within the GHz
range and are thus much smaller than for the splitting of this doublet.
Experimentally, both line profiles would have a shoulder because the
hyperfine structure splitting of the Na 32S ground state of 1.772 GHz is
close to the experimental line width. The hyperfine structure of the excited
state, in contrast, is usually not resolved with simple absorption
spectroscopy. Measuring line widths has interesting technical applications:
for example, the temperature dependence of the Doppler width (Eq. (10.25))
opens up the possibility of determining the temperature of flames from a
spectroscopic measurement.
Solution 10.3
This exercise deals with a concrete application of optical absorption
spectroscopy in chemical analytics. =In Problem 9.13, we have dealt with
the UV absorption of benzene due to electronic excitation of its system of π
electrons. Amino acids with rings such as Trp and Tyr (see Fig. 10.7) absorb
light within the ultraviolet spectral range at a wave length of 280 nm.
Cysteine also weakly absorbs light at this wave length. As these molecules
are found in the amino acid sequences of larger protein structures, their UV
absorption can be used for a quantitative analysis. In subproblem (a), we
determine the concentration of glutamate dehydrogenase in a sample that
absorbs 30% of the incoming light at a wave length of 280 nm. According to
Eqs. (10.5) and (10.6), the absorbance is:
(10.42)
Fig. 10.7 Molecular structure of the three amino acids cysteine (Cys), tyrosine (Tyr), and tryptophan
(Trp)
(10.43)
If z Cys = 6, z Tyr = 18, and z Trp = 4 are the numbers of Cys, Tyr, and Trp units
in glutamate dehydrogenase and c is the sought concentration of this enzyme,
we have
(10.44)
and thus
(10.45)
In subproblem (b), the absorbance of a sample of 1 g oxytocin in 1 l
solution is calculated. Thus, we must determine the molar mass of this
molecule. From the formula C43H66N12O12S2, we obtain the molar mass M =
1007 g mol−1. Thus, the concentration of the sample is c = 9. 9 × 10−4 M. As
the molecule contains 2 units of Cys and 1 unit of Tyr, the sought absorbance
of a sample is:
(10.46)
which is a large value. The transmitted intensity would be only I = I 0 10−1. 5
= 0. 03 I 0.
Solution 10.4
In this problem, we deal with the structure determination of a linear triatomic
molecule: HCN (Fig. 10.8). The goal is to determine the bond lengths of the
C-H and the C-N bond by the measurement of the rotational constants from at
least two different isotopomers. These have different nuclear masses, but it
can be expected that the bond lengths, which are only determined by
electronic structure,7 are the same. This is a consequence of the Born-
Oppenheimer approximation , according to which the total wave function of a
molecule (electronic and nuclear degrees of freedom) can be separated into a
nuclear part and an electronic part.
In subproblem (a) we begin by searching expressions for the positions of
the three nuclei in the center of mass frame of reference. Note that for
different isotopomers the center of mass is at different positions on the
molecular axis of symmetry. We need them to form the moments of inertia,
which in turn are related to the rotational constants (see Eq. (10.12)). The
center of mass, indicated as a dot in Fig. 10.8, lies between the heavier
carbon and nitrogen cores. It is thus reasonable to tentatively set the origin at
the carbon site. This choice sets , , . The position of
the center of mass in this system is:
(10.47)
(10.48)
(10.49)
In subproblem (b), we write down the moment of inertia in the center of
mass frame of reference. The moments of inertia of the molecule are I xx = 0,
I yy = I zz = I where
(10.50)
The goal is to write this as an expression that contains only the known
masses of the nuclei and the bond lengths r CH and r CN. Using Eq. (10.48),
we substitute the position of the carbon, x C. After some algebra, we obtain:
(10.51)
This equation gives the relation of the moment of inertia with the bond
lengths and the masses. On the other hand, I is related to the rotational
constant :
(10.52)
With two different rotational constants for two different sets of nuclear
masses, we have two equations for the two unknown bond lengths:
(10.53)
(10.54)
where , , and , i = 1, 2 are
effective masses.
Unfortunately, these equations are not linear in the bond lengths. Thus, we
have to seek the solutions numerically. This is tackled in subproblem (c). In
the appendix, Sect. A.3.19, Newton’s method for the solution of a nonlinear
system of equations is described. A computer code is easily set up that
implements Newton’s method for this problem. The output from such a
computer code is shown below. Starting from a reasonable initial guess of
the bond lengths of 1 Å for both bonds, Newton’s method converges within
about 10 iterations toward r CN = 1. 1569 Å and r CH = 1. 0689 Å. As can be
seen in the output, the exchange of12C with13C changes the nuclei’s positions
relative to the center of mass by about 0.02 Å. The calculation is for a
perfectly rigid rotator. Even in the vibrational ground state, however, a slight
correction of the rotational constants originating from rotation vibration
coupling is present. Such corrections are considered in Problem 10.7. As
shown there, these corrections change the bond lengths obtained to the order
of 10−3 Å.
Looking back on our solution, we have gained an insight into the structure
determination of a rigid linear polyatomic molecule based on rotational
constants of isotopomers. Owing to different nuclear masses, these
isotopomers have different moments of inertia, but identical bond lengths. In
fact, isotopic substitution is an often applied method in spectroscopy (see
also Problem 10.6). To obtain results we had to resort to a numerical
solution. It is worth drawing attention to the problems arising if more
complex molecules are investigated: (1) the more complex case of nonlinear
molecules is dealt with in Problem 10.5. (2) The case of polyatomic
molecules with more than just three nuclear sites can indeed be tackled by
systematic isotopic substitution to increase experimental data sets. This case,
however, involves another problem: such molecules are more flexible, i.e.,
they appear in different conformations. In addition, nonrigid or floppy
molecules change their structure depending on their rotational state, which
makes the correct interpretation of rotation or rotation-vibration spectra very
puzzling.
The water molecule (H2 16O) is an asymmetric top molecule . Its bond
length is r OH = 0. 9584 Å, the angle enclosed by the OH bonds is
104.45∘.
(10.55)
ranging from − 1 to + 1. A value of κ = −1 would indicate the
limiting case of prolate symmetric top molecule; a value of + 1,
in contrast, would indicate the case of an oblate symmetric top.
Calculate the asymmetry parameter for water.
c.
Energy levels of the rotational states of asymmetric tops can be
written in the form:
(10.56)
where J is the rotational quantum number and τ = −J, …, +J is
an index. Note that in this notation energy levels are expressed
in frequency units of Hertz. The function E J τ (κ) is found in
Table 10.3. Calculate the energy levels of the H2O molecule for
J = 1, 2. Also, calculate the energy levels of water by treating
the molecule formally as an prolate or an oblate symmetric top
molecule. Plot the energy levels in a term diagram. Can you
explain the special labeling of asymmetric top rotator states
in Table 10.3?
Table 10.3 Algebraic relations for asymmetric rotor energy levels according to H.W. Kroto,
Molecular Rotation Spectra, Wiley, London, 1975
τ E J τ (κ)
000 0 0
110 −1 κ + 1
111 0 0
101 +1 κ − 1
220 −2
221 −1 κ + 3
211 0 4κ
212 +1 κ − 3
202 +2
Solution 10.5
This exercise deals with the rotational energy levels of symmetric and
asymmetric top molecules. We familiarize ourselves with the quantum
numbers involved in the description of their rotational states, and the special
treatment of asymmetric top rigid rotators. Although the rotation spectra of
diatomics are quite regular, it is complicated to interpret the asymmetric top
rotational spectra. The water molecule is perhaps the most prominent
example of an asymmetric rotor with a well-known geometry.
(10.57)
where m i are the masses of the atoms at position r i . The oxygen has an
atomic weight of m O = 16 amu; both hydrogens have a mass of 1 amu. Thus,
(10.58)
The center of mass is thus 0.065 Å displaced from the oxygen center in a
negative z-direction toward the hydrogens (red point in Fig. 10.9). Relative
to the center of mass, the oxygen is at position , whereas
the hydrogens take the positions and
respectively. With these coordinates, we are
ready to write down the moment of inertia tensor . The general definition is
(10.59)
As our choice of coordinates was such that the coordinate axes already
coincide with the principal axes of inertia, the off-diagonal elements are all
zero. The diagonal elements need to be calculated. For I xx we obtain:
(10.60)
and thus8 I xx = 1. 0175 × 10−47 kg m2. The second principal moment is:
(10.61)
or I yy = 2. 9232 × 10−47 kg m2. The third principal moment is:
(10.62)
With these results, we can calculate the three rotational constants:
(10.63)
(10.64)
(10.65)
The constants are ordered according to A ≧ B ≧ C. In subproblem (b) we
determine the asymmetry parameter κ defined in Eq. (10.55). This parameter
is highly useful for the interpretation of asymmetric top rotation spectra. By
insertion of the values for A, B, and C we obtain a value of κ = −0. 4300. In
the limiting case in which two of the three moments of inertia are identical,
the molecule becomes a symmetric top molecule . One distinguishes
between more cigar-shaped prolate symmetric tops with A > B = C and more
disk-shaped oblate symmetric tops with A = B > C respectively. The energy
levels of symmetric top rigid rotators are found in the textbooks:
(10.66)
(10.67)
If the energy levels are expressed as a frequency in Hertz, these
expressions simply read:
(10.68)
and
(10.69)
Compared with the treatment of the diatomic rigid rotator, a new quantum
number, K = 0, ±1, …, ±J occurs. The energy levels are degenerated because
of the K 2 dependence. The wave functions for the symmetric top rotator are
the Wigner rotation functions D MK J (θ, ϕ, ψ). They are generalized
spherical harmonics depending on the three Euler angles describing the
orientation of a molecule in space [2]. Although the symmetric top rigid
rotator problem can be treated exactly in a straightforward fashion, the
asymmetric rotor problem is tedious, although analytic expressions for the
energy levels exist. They can be found if for a given quantum number J the
wave function is written as a series over all possible Wigner functions:
(10.70)
The resulting expressions for J = 0, 1, 2 are given in the text of subproblem
(c) (Eq. (10.56) and Table 10.3). It is our task to calculate these first
rotational levels for the water molecule. Like the quantum number K in the
symmetric top case, there is an index τ = −J, …, +J for the (2J + 1) energy
levels. The case J = 0 is trivial; we find E(0, 0) = 0. Starting with J = 1, we
take Eq. (10.56) and pick the appropriate function E 1, −1(κ) from Table 10.3.
Looking at the definition of the asymmetry parameter κ we notice that
simplifications are possible:
(10.71)
In a similar fashion, we obtain:
(10.72)
and
(10.73)
For J = 2, we can simplify the expressions for τ = −1, 0, 2 in an analogous
way and obtain:
(10.74)
(10.75)
(10.76)
The cases J = 2, τ = ±2 are more complicated. The expression from
Table 10.3 is and by insertion of the proper value of
κ the results are E(2, −2) = 4, 063, 909 MHz and E(2, +2) = 2, 144, 767 MHz.
As requested, we plot all these values in a term diagram, which is depicted
in Fig. 10.10 (red levels). The labeling of the various levels is in accordance
with the first column of Table 10.3, in the form . At first sight, this
notation seems to be quite opaque. In the rest of this exercise we try to obtain
a deeper understanding of this notation, which is common in the literature.
Therefore, we formally treat the water molecule as a symmetric top rigid
rotator and determine the energy levels. We start with the prolate case (κ =
−1), in which the rotational constants B and C are identical. We use Eq.
(10.68) to calculate the energy levels for J = 0, 1, 2 and the possible values
for the quantum number K. The results are depicted in Fig. 10.10 on the left-
hand side at κ = −1. Next, we consider the oblate case (κ = +1, B = A) and
proceed in an analogous way. In the term diagram, these energy levels are
shown on the right-hand side at κ = +1. A close look at the dashed lines
connecting certain prolate and oblate energy levels reveals the meaning of
the special labeling of asymmetric top rotator states plotted at the
appropriate κ = −0. 43. Consider, for example, the asymmetric top level 212.
It is exactly on the intersection line between the prolate 2,±1 state and the
oblate 2,±2 state. Most of the levels match with an appropriate intersection
line.9 Figure 10.10 is called a correlation diagram . It reveals the location
of energy levels of the water molecule if the asymmetry parameter κ changes
between the limiting cases − 1 (prolate) to + 1 (oblate). Although the
numbers K are only quantum numbers for these limiting cases of symmetric
top rotors, they are also used as projection quantum numbers in the case of an
asymmetric top rotator.
Fig. 10.10 Correlation diagram of the first H2 16O rotational levels (red lines). The special labeling of
the levels is in accordance with Table 10.3. Also shown are the energy levels for water treated formally
as a prolate symmetric top (κ = −1, rotational constant B = C), and as an oblate symmetric top (κ = +1,
rotational constant B = A)
and
Note that because of the irregular spacing between the rotational transitions;
transitions from higher excited molecules (J > 2) certainly agglomerate in the
vicinity of these two transitions. This complicates the interpretation of
microwave spectra. Useful for the assignment of transitions is the fact that
states with higher J have a lower occupation probability because of the
Boltzmann statistics. Therefore, transitions from these states have a lower
intensity in the microwave spectrum. Such a spectrum is shown in Fig. 10.11.
Note the transition with the lowest frequency at 22.2 GHz. It is the 616 ↔ 523
transition. The frequency of many kitchen microwave ovens is aligned with
this transition.
Fig. 10.11 Microwave pure rotational stick spectrum of H2 16O according to data taken from F.C. de
Lucia, P. Helminger, W.H. Kirchhoff, Microwave spectra of molecules of astrophysical interest V:
water vapor, J. Phys. Chem. Ref. Data, 2 (1974), 211
Problem 10.6 (IR Spectra of Diatomics I)
(10.78)
and the energy states of the Morse oscillator are given by:
(10.79)
where ν is the frequency of the stretch mode, D e = 7. 41 × 10−19 J
is the depth of the potential, and R e = 1. 275 Å is its minimum.
For the parameter a describing the steepness of the potential the
following relation holds:
(10.80)
How many vibrational states has the HCl molecule?
e.
An anharmonic oscillator allows overtone transitions . Calculate
the wave number of the transition n = 0 → 2. At which wave
number would the hot band n = 1 → 2 be observed? On the basis
of occupation probabilities, judge if the hot band is observed at T
= 300 K.
Solution 10.6
(10.82)
Hence, the force constant sought is
(10.83)
The effective mass for DCl is (Eq. (10.82)) . Taking the
effective masses for both isotopomers and the relation between frequency
and wave number ( ), we obtain:
(10.84)
Thus, the stretch mode of DCl is expected at:
In subproblem (d), we take anharmonicity into account and deal with the
number of vibrational states in the HCl molecule. When reading the problem
text for the first time we get stuck and ask ourselves what this means: the
number of vibrational states. We must be aware that, depending on the shape
of the potential, a quantum system may have a finite number of discrete bound
states. Although for example the particle in a box with infinite walls and also
the harmonic oscillator have an infinite number of discrete states, a particle
in a finite potential well is an example of a system that has only a finite
number of discrete states. For energies exceeding the potential well depth,
the energy spectrum becomes continuous and the particle may leave the box.
Extrapolated to the case of a vibrating diatomic molecule with an interatomic
potential allowing for dissociation, the number of bound states should be
finite. A simple and frequently used model for the anharmonic potential
between two atomic sites is the Morse potential Eq. (10.78). It has only three
parameters a, D e , and R e , and it has the advantage that the energy levels
are well-defined (Eq. (10.79)):
Fig. 10.13 The anharmonic potential in the HCl molecule, based on the Morse oscillator model. Also
shown are the resulting energy levels of the Morse oscillator
The first term is identical to the harmonic oscillator expression for the
energy in state n. The second term can be interpreted as a correction resulting
from anharmonicity. The minus sign is important, i.e., the correction leads to
shrinking of the energy differences between adjacent states. The situation is
illustrated in Fig. 10.13, where the potential is plotted with the given values
for the potential depth D e , the potential minimum R e , and the parameter a
describing the steepness of the potential,
(10.85)
Also shown are the energy levels resulting from Eq. (10.79) and the
quantum energy J = 0. 37 eV. In Fig. 10.14, the
energies E n are plotted against the quantum number n. As the dissociation
energy is gradually reached, the energy levels are expected to approach one
another, which is indeed the case for quantum numbers up to n = 24. For n >
24, however, Eq. (10.79) predicts that E n become smaller with increasing n.
This is unphysical. Therefore, the number of vibrational states of the HCl
molecule is 25 (n = 0–24).
Fig. 10.14 Vibrational energies of the HCl molecule as a function of quantum number n, calculated
with Eq. (10.79). Realistic energy levels are those between n = 0 and 24
(10.86)
(10.87)
(10.88)
Therefore, the fundamental n = 0 → 1 transition is observed at:
(10.89)
which means a red-shift of 120 cm−1, in comparison with the pure harmonic
wave number. The transition is shown in Fig. 10.15 in another term scheme
of the first few vibrational levels. Also shown in this figure are additional
transitions that we deal with in subproblem (e). A consequence of
anharmonicity is the breaking of selection rules of the harmonic oscillator.
Therefore, the overtone transition n = 0 → 2 is no longer forbidden. Taking
the energy difference between the levels n = 2 and the ground state we find:
(10.90)
Fig. 10.15 Term scheme and vibrational excitations in the region of the first excited states of the HCl
molecule
Note that overtone transitions and combination bands are always much
less intense than the corresponding fundamentals. Hot bands are another type
of IR transitions. In absorption spectroscopy, they result from the excitation
of a molecule in an excited vibrational state. In our case, we consider the
transition n = 1 → 2. By taking the difference in energy between these two
states we obtain:
(10.91)
We further investigate whether this transition should be observable under
room temperature conditions using occupation probabilities. What is behind
it? The transition n = 1 → 2 is only observable if the state n = 1 is notably
occupied. The occupation probability p n for a state n follows the Boltzmann
statistics:
(10.92)
Fig. 10.16 Occupation probabilities of the ground state and the first excited vibrational states of the
HCl molecule according to Boltzmann statistics
(10.93)
where V (R e ) is the potential at the equilibrium bond length R e
, ν e is the harmonic frequency of the stretch vibration, x e the
anharmonicity constant, α e is the rotation vibration coupling
constant, is the centrifugal distortion constant. Determine B
e, αe,Re,xe, from a suitable polynomial fit of the
spectroscopic data. Hint: Write down relations for the
transition frequencies of the fundamental and first overtone.
Introduce new parameters J ′ = J + 1 and m = −J ′ (P-branch),
m = J ′ (R-branch).
b.
The centrifugal distortion constant is related to the rotational
constant B e via
(10.94)
Does the fitted result for satisfy this condition?
(10.95)
(10.96)
and
(10.97)
Table 10.4 Measured wave numbers of gas-phase rovibrational transitions of carbon monoxide
n=0 →1 n=0 →2
J P R P R
0 2,139.427 2,147.082 4,256.216 4,263.837
1 2,135.547 2,150.856 4,252.301 4,267.542
2 2,131.632 2,154.596 4,248.317 4,271.177
3 2,127.683 2,158.300 4,244.262 4,274.740
4 2,123.700 2,161.969 4,240.138 4,278.234
5 2,119.681 2,165.601 4,235.947 4,281.656
6 2,115.629 2,169.198 4,231.685 4,285.009
7 2,111.544 2,172.759 4,227.354 4,288.289
8 2,107.423 2,176.284 4,222.954 4,291.499
9 2,103.270 2,179.772 4,218.486 4,294.638
10 2,099.083 2,183.224 4,213.949 4,297.704
11 2,094.863 2,186.639 4,209.343 4,300.700
12 2,090.609 2,190.018 4,204.668 4,303.623
13 2,086.322 2,193.360 4,199.929 4,306.475
14 2,082.003 2,196.664 4,195.117 4,309.254
15 2,077.650 2,199.931 4,190.239 4,311.961
16 2,073.265 2,203.161 4,185.295 4,314.596
17 2,068.847 2,206.354 4,180.282 4,317.159
18 2,064.397 2,209.509 4,175.203 4,319.649
19 2,059.915 2,212.626 4,170.055 4,322.064
20 2,055.401 2,215.705 4,164.839 4,324.409
21 2,050.855 2,218.746 4,159.560 4,326.681
22 2,046.276 2,221.749 4,154.212 4,328.878
23 2,041.667 2,224.713 4,148.797 4,331.004
24 2,037.026 2,227.639 4,333.053
25 2,230.526 4,335.030
26 2,233.375
27 2,236.185
All values in cm−1. J is the rotational quantum number of the final state for
the P-branch (J + 1 → J), and of the initial state for the R-branch (J → J + 1)
respectively
Fig. 10.17 Rotationally resolved IR spectrum in the region of the first overtone (0 → 2) of the12C16O
stretch vibration
Solution 10.7
With this challenging exercise, we explore the foundations of molecular
spectroscopy beyond the rigid rotator-harmonic oscillator (RRHO)
approximation. In Problem 10.4 we have determined bond lengths of a
molecule from the rotational constants obtained from pure rotation spectra. In
the discussion of the results it was mentioned that even in the vibrational
ground state, the rotational constants are slightly influenced by vibration
rotation coupling. Neglect of such effects involves a slight systematic error in
the determination of bond lengths. In this problem, we focus on such effects
in detail and treat vibration and rotation all together. In gas-phase infrared
absorption spectra, vibrational transitions go along with changes in the
rotational state. The selection rules for a diatomic rotor, Δ J = ±1 (see Eq.
(10.15)), lead to a regular wing-like arrangement of the vibration rotation
transitions as shown in Fig. 10.17 for the first overtone vibration. The
experimental line positions are given in Table 10.4. The P-branch results
from transitions n, J + 1 → n + 1, J, the R-branch from the transitions n, J →
n + 1, J + 1 respectively. J is the usual rotational quantum number. The Q-
branch n, J → n + 1, J is left out.11 In subproblem (a), we use these data to
determine spectroscopic constants included in Eq. (10.93). It is useful to
analyze this equation, which is the result of a tedious perturbation treatment
of the rigid-rotator harmonic-oscillator (RRHO) model, in detail. The first
three terms are the vibration-rotation energy of the unperturbed RRHO
model. The following three terms are the first corrections from the
perturbation treatment, which typically lead to a decrease in the energy
levels. Anharmonicity—the first correction term—has already been
introduced in Problem 10.6. The second correction is the rotation vibration
coupling, whereas the last correction stems from the centrifugal distortion of
the molecule depending on its rotational state. For the comparison with the
experimental data we must set up expressions for the transition frequencies in
the fundamental and overtone region. We start with writing down the energies
for the ground and the first vibrationally excited state. Following the hints
given, we introduce the new parameter J ′ = J + 1:
(10.98)
(10.99)
(10.100)
(10.101)
Moreover, we have introduced the vibrationally corrected rotational
constants
(10.102)
Before we consider energy differences, this auxiliary calculation is
useful:
If we consider the difference between these two expressions, all terms but
the cubic ones cancel each other out. As a consequence, we can form the
transition frequencies for the P-branch of the fundamental lines:
(10.103)
(10.104)
Inspection of these expressions reveals that they can be combined if we
introduce the new parameter m = −J ′ for the P-branch and m = J ′ for the R-
branch. Then, the transition frequencies for the region n = 0 → 1 are:
(10.105)
The corresponding wave numbers are obtained after division by the speed of
light, c = 2. 99792458 × 1010 cm s−1. If we plot the transition wave numbers
against m, we obtain a curved line. Such a plot is shown in Fig. 10.18. A
cubic fit of the form:
(10.106)
with these data provide the values of ν e (1 − 2x e ), B 1 + B 0, B 1 − B 0, and
. The result of the cubic fit is shown in Table 10.5. Before we further
evaluate these results, we must first consider the transitions of the first
overtone.
Fig. 10.18 Polynomial plot of the experimental data of rovibrational transitions in the region n = 0 → 1.
For the definition of the parameter m, see the text of subproblem (a)
Table 10.5 Results of the polynomial fit of the experimental data put into graphs in Fig. 10.18 and in
Fig. 10.19
Vib. transition a 0 a1 a2 a3
0→1 2143.27144(7) 3.82754(1) − 0. 0175 − 2. 443(2) × 10−5
0→2 4260.0617(2) 3.81006(2) − 0. 035 − 2. 449(4) × 10−5
(10.107)
(10.108)
which can be combined to form:
(10.109)
A plot of the data of the overtone region against m is shown in Fig. 10.19.
The results of the polynomial fit are shown in Table 10.5.
Fig. 10.19 Polynomial plot of the experimental data of rovibrational transitions in the region n = 0 → 2.
For the definition of the parameter m, see the text of subproblem (a):
We are now ready to exploit the data in Table 10.5 to obtain the
quantities sought. First, we focus on the coefficients a 0 containing the
harmonic wave number and the anharmonicity constant. From
we obtain
(10.110)
and
(10.111)
Focusing on the coefficients a 1 and a 2, we can determine the rotational
constants. Because
we obtain12
(10.112)
and
(10.113)
In frequency units, we thus obtain B 1 = 57, 111. 1(3) MHz and B 0 = 57, 635.
7(3) MHz. The difference in the rotational constants is the vibration-rotation
constant α e :
(10.114)
The equilibrium rotational constant B e follows from (Eq. (10.102))
(10.115)
or B e = 57898. 0(6) MHz. Having obtained the rotational constants, we can
determine the C-O bond length. The effective mass for the CO rotor is:
(10.116)
The relation between the rotational constant and the moment of inertia I
or the bond length R e is
(10.117)
and thus
(10.118)
or 1.1282 Å. For comparison, the bond length R 0 evaluated with the
rotational constant in the vibrational ground state, B 0, is 1.131 Å. Finally, we
determine the centrifugal distortion constant contained in the parameter a 3.
The results from the two data sets are similar. From the data set of the 0 → 1
transitions, we obtain MHz. From the overtone
MHz for is consistent with the data. The value for the centrifugal
distortion constant is further scrutinized in subproblem (b). According to
Eq. (10.94) it can be related to B e . This alternative way of computing is
another test of the integrity of the data sets. With
MHz the centrifugal distortion constant is
This value is in good agreement with the result from subproblem (a) and the
condition Eq. (10.94) is reflected by the data.
Starting with subproblem (c) we focus on the anharmonic potential
between the carbon and the oxygen core. The Morse potential model was
introduced in Problem 10.6 and we apply it here once more. Comparison of
Eq. (10.79) with Eq. (10.93) reveals the relationship between the
anharmonicity constant x e and the depth of the Morse potential, D e :
(10.119)
In the Morse oscillator model, the bond dissociation energy D 0 of CO is
the difference between D e and the zero point vibration energy, E n = 0, J = 0,
which can be calculated according to Eq. (10.93):
Hence,
(10.120)
The force constant k is calculated using Eq. (10.16):
(10.121)
The Morse parameter a follows from Eq. (10.80):
(10.122)
The third parameter is simply the equilibrium distance R e = 1. 1282 Å. A
plot of the Morse potential for CO is depicted in Fig. 10.20. A second way of
analyzing the anharmonic potential in the CO molecule is to consider cubic
and quartic force constants. They are related to the spectroscopic parameters
that we have already determined in subproblem (a). We rearrange Eqs.
(10.96) and (10.97) in subproblem (d):
(10.123)
Fig. 10.20 Modeling of the potential between carbon and oxygen. The solid line is the Morse
potential derived from spectroscopic constants. The dashed line is the anharmonic potential derived
from the cubic and quartic force constants
(10.124)
With these anharmonic force constants the carbon-oxygen potential
(10.125)
takes the form shown as the dashed line in Fig. 10.20. For r < R e , the
potential is very close to the Morse potential. For r > R e , however, the
potentials are in line only up to about a bond length of 1.5 Å. Larger
elongations are not adequately described by the cubic and quartic force
constants. Higher order force constants would be needed to describe such
situations. Looking back on the exercise, we have learned how to cope with
larger sets of spectroscopic data and how the fundamental molecular
properties such as the bond length and the bond dissociation energy can be
extracted from these data. Doing so we went beyond the models of the rigid
rotator and the harmonic oscillator.
C 2v
A1 1 1 1 1 z x 2,y 2,z 2
A2 1 1 −1 −1 Rz xy
B1 1 −1 1 −1 x, R y xz
B2 1 −1 −1 1 y, R x yz
The character table of the point group breaks down the different
transformation behavior of the various symmetry types. It is shown in
Table 10.6. Without a detailed harmonic analysis, for example, based on a
quantum chemical method, it is possible to evaluate the symmetry types of the
six vibrational degrees of freedom. Therefore, we seek the reducible
representation Γ red of the nuclei’s movement and we must decompose it into
irreducible representations Γ i according to:
(10.126)
The coefficients a i are obtained from the reduction formula
(10.127)
Here, is the character of one of the symmetry operations taken
from the character table and χ red is the respective character of the reducible
representation for this symmetry operation, and h is the order of the group,
which is 4 in this case.
For each symmetry operation , we consider the transformation
(10.128)
where T is the transformation matrix of the symmetry operation acting on the
”vector” X containing the positions of all nuclei. The character of the
symmetry operation sought is simply the trace14 .
The identity operation is represented by a (12 × 12) unit matrix:
(10.129)
(10.130)
(10.131)
(10.132)
We therefore obtain . It is useful to summarize these results
in an additional line in the character table, as shown in Table 10.7. With the
help of Eq. (10.127), we can determine the coefficients a i
C 2v
A1 1 1 1 1 z x 2,y 2,z 2
A2 1 1 −1 −1 Rz xy
B1 1 −1 1 −1 x, R y xz
B2 1 −1 −1 1 y, R x yz
Γ red 12 − 2 4 2
(10.133)
(10.134)
(10.135)
(10.136)
As a consequence, the reducible representation is reduced into irreducible
representations according to:
(10.137)
However, the movements we have analyzed also contain translations and
rotations of the molecule as a whole. The character table reveals the
symmetry types of the translations x, y, z, in addition to rotations R x , R y , R z
in the last column. The representation of the translations and rotations are
(10.138)
(10.139)
By subtraction of the latter we obtain the result for the vibrations:
(10.140)
We come to the conclusion that three normal modes have the symmetry
type A 1, two modes are of type B 1, and one mode has symmetry type B 2. In
fact, an elaborate harmonic analysis of the molecule based on quantum
chemical methods would confirm this result. Why might this information be
useful? Just one application of this result is the prediction of infrared activity
in subproblem (d). Infrared active modes and the translations x, y, and z
share the same symmetry type. The information is found in the character
tables. All six of the normal modes are infrared active. Moreover, a
vibrational mode is Raman active if its symmetry type corresponds to one of
the products x 2, y 2, z 2, xy,xz, or yz. The latter information can also be taken
from the character table. We conclude that all modes of formaldehyde are
also Raman active.
The six normal modes of H2CO, as they are obtained from a quantum
chemical vibrational analysis, are shown in Fig. 10.23. The modes are
ordered according to increasing vibrational frequencies. The mode with the
lowest frequency ν 1 is a bending mode of symmetry type B 2. The two B 1
modes are a rocking mode ν 2 and the mode ν 6 at highest frequency, which is
a asymmetric C-H stretch mode. The three modes of symmetry type A 1 are C-
O stretch movements combined with in-phase and out-of-phase scissoring
motion of the C-H bonds (ν 3 and ν 4), and the symmetric C-H stretch mode ν
5.
Problem 10.9 (Influence of Nuclear Spin Statistics)
Fig. 10.24 The CO2 molecule and its asymmetric stretch mode ν 3. If the oxygens O1 and O2 belong
to the same isotopic species, then additional selection rules from nuclear spin statistics take effect. One
point symmetry element of the molecule is its inversion center i. Arrows indicate schematically the
motion of the ν 3 vibrational mode
D∞h
Σg+ 1 1 1 1 1 1 x 2 + y 2, z 2
D∞h
Σg− 1 1 −1 1 1 −1 Rz
Σu− 1 1 −1 −1 −1 1
Solution 10.9
In this exercise, our attention is drawn to nuclear spin statistics and its
influence on the spectra of molecules with symmetric structure. The
prominent example is12C16O2 where, as it turns out, every second
rovibrational transition is forbidden in the region of the asymmetric stretch
band ν 3 = 0 → 1.
In subproblem (a), we identify the subset of transitions that are forbidden
owing to nuclear spin statistics. For this, we need to bring together the given
facts and consider the consequences for the symmetry of the molecule’s wave
function. Nuclei of the same isotopic species are indistinguishable quantum
mechanical objects. In12C16O2, we have a two-particle system of16O nuclei,
which we could label 1 and 2. Now, consider an exchange operator acting
on the nuclei’s wave function with the result that the particles are exchanged:
(10.141)
If the operator acts twice, then the original situation is re-established. This
requires a condition for the eigenvalue τ:
(10.142)
(10.143)
(10.145)
This condition is satisfied if even values of J are combined with even
values of n, and, moreover, if odd values of J are combined with odd values
of n.
Fig. 10.25 Term scheme (schematic) of the first allowed (solid lines) and forbidden rovibrational
transitions (dashed lines) of the asymmetric stretch vibration for the12C16O2 molecule
(10.146)
Note the factor in front of the double sum preventing double counting. It
is common practice to introduce a rotational constant
(10.147)
for the vibrational state, which is the generalization of the constants B i (i =
0, 1, 2) we have used in the diatomic case in Problem 10.7. It is clear that
with the sparse information given in this problem, we are not able to
determine the equilibrium rotational constant B e and thus the equilibrium C-
O bond distance. This would require determination of the rotation vibration
constants α e, i for all normal modes. Here we concentrate on the
determination of B 000, the ground state rotational constant from which the
bond length in the ground state can be determined. We must figure out how
this is possible with the information of only two line positions in the P- and
the R-branches of the transition17 000 → 001. At first, the notion is important
that for the calculation of a transition frequency between the ground state and
the singly excited asymmetric stretch mode many of the terms are unchanged
in these two states and thus cancel each other out. With d 1 = d 3 = 1 and d 2 =
2, we obtain for the P-branch (J ′ = J + 1 → J, see the introduction of the
number J ′ in Problem 10.7)
(10.148)
and for the R-branch (J → J ′ = J + 1)
(10.149)
In the problem text, the wave numbers of the first observable transitions
of these branches are given. According to Fig. 10.25 the first observable
transition for the P-branch has J ′ = 2, whereas the R-branch starts with J ′ =
1. We therefore obtain two equations
(10.150)
and
(10.151)
for the frequencies of these two transitions. By taking the difference between
these frequencies and multiplying with the speed of light, we arrive at:
(10.152)
With cm−1 and cm−1, the ground state
rotational constant is:
(10.153)
It is related via Eq. (10.12) to the moment of inertia of the molecule,
(10.154)
and thus to the ground state bond length sought r CO:
(10.155)
Our result for the C-O bond lengths within the CO2 molecule in its
vibrational ground state is thus 1.162 Å.
(10.156)
(10.157)
R is the pump rate, i.e., the number of particles per volume and
time being excited into the upper level, τ is the lifetime of | 2〉 in
the zero field including nonradiative relaxation. The relaxation
time τ r of the photon density accounts for radiative losses in the
cavity, including transmission through one of the mirrors where
the laser radiation is emitted, and c is the speed of light.
Determine the steady-state population of the upper level | 2〉 in
addition to the minimum pump rate R crit necessary for
operation with a He-Ne laser (τ = 1 × 10−8 s, τ r = 3 × 10−8 s, σ
= 3 × 10−13 cm2) and for a Nd:YAG laser (τ = 2. 5 × 10−4 s, τ r
= 7 × 10−10 s, σ = 8 × 10−19 cm2)
c.
Consider small distortions from the steady-state values N 2s and
Φ s , N 2(t) = N 2s + n, and Φ(t) = Φ s +ϕ. Show that the
population distortions can be described by a damped oscillator
(10.158)
where δ is a damping parameter and ω is the angular frequency
of the relaxation oscillations. Compare the relaxation behavior
of the He-Ne laser and the Nd:YAG laser using the parameters
given above and a pumping rate of R = 2R crit. Which laser
shows relaxation oscillations?
Fig. 10.26 Principle of laser operation (schematic). The active medium is represented by a four-level
quantum system within a resonator of d in length. The laser transition is between levels | 2〉 and | 1〉. At
least one of the mirrors (M) has nonzero transmission
Solution 10.10
Lasers 18 are an indispensable tool of spectroscopy, as they provide
monochromatic coherent light of high intensity. Here, we deal with the
principles of laser operation using the model sketched in Fig. 10.26. Within
an optical resonator, there is an active medium that is represented by a four-
level quantum system. Under operating conditions, the level | 2〉 has a higher
population than the lower level | 1〉. This inversion is crucial, as it allows
stimulated emission of quanta with the energy h ν = E 2 − E 1 and thus a
build-up of radiation within the resonator of this frequency. In the zero field
the population inversion would decay owing to spontaneous emission (decay
time T 21) or because of other nonradiative processes, and reach the
population given by the Boltzmann distribution at a certain temperature. An
effective mechanism of inversion is achieved by a suitable pump mechanism
involving the transition from the ground state | 0〉 to a higher state | 3〉 with the
rate R 03. Owing to rapid relaxation, the upper laser level | 2〉 is then
substantially populated. In subproblem (a), we assess the efficiency of an
active medium regarding the decay time of the lower laser level | 1〉. If N 1
and N 2 are the densities of the lower and the upper levels respectively, the
degree of inversion is simply the difference N 2 − N 1. Now, if the decay time
T 10 were long compared with T 21, the lower level | 1〉 would already be
populated because of spontaneous emission processes. Thus, the inversion
would decrease owing to a flooding of the lower level | 1〉. Therefore, for
good laser operation, we need the opposite situation, i.e., a very short
lifetime of the level | 1〉, so that ideally, N 1 = 0 holds. The answer is thus that
T 10 ≪ T 21 is needed for laser operation. Moreover, simple models like ours
assume that the pumping mechanism yields a constant rate by which the upper
level | 2〉 is populated. This requires a constant population of the ground
state, N 0. Under these conditions, the four-level laser system can be
described by the two rate equations Eqs. (10.156) and (10.157). The
equation for the time-dependent population density of the upper laser level |
2〉, N 2(t) can be motivated primarily by the rate equations for induced and
spontaneous emission (Eqs. (10.8) and (10.9)). The photon density Φ can be
related to the spectral energy density u(ν) introduced in Sect. 10.1.1 via u(ν)
= h ν Φ. One relevant parameter is the lifetime τ of the level | 2〉, which
includes spontaneous decay in addition to nonradiative processes that
depopulate this level. Moreover, the time τ r describes the decay of the
photon density caused by imperfect reflectivity of the mirrors, dissipation,
and, of course, losses due to coupling out of the radiation because of
transmission through at least one of the mirrors. The first terms on the right-
hand side of Eqs. (10.156) and (10.157) describe these loss mechanisms of N
2 and Φ respectively. However, the photon density is fed by means of
stimulated emission with the cross section σ, represented by the second term
in Eq. (10.157). This gain is proportional to N 2, which in turn decreases by
the same amount (second term in Eq. (10.156)). The constant pumping rate R
increases N 2 (third term in Eq. (10.156)). Apparently, we do not account for
details such as an index of refraction that is different from unity. We assume
that electromagnetic waves propagate through the resonator at the vacuum
speed of light c. Here, it is worth mentioning that these laser equations are
similar to those describing the Lotka-Volterra kinetic model of oscillating
chemical reactions that we have dealt with in Problems 6.6 and 6.7 Hence, in
the following, we benefit from what we have worked out in chemical
kinetics. It is not easy to solve these coupled equations without numerical
methods.
The first point we examine in subproblem (b) are the stationary
solutions. These are characterized by a constant population N 2s of the upper
laser level, and a constant photon density Φ s . The first derivatives of these
quantities are thus zero. Taking Eq. (10.157), we obtain:
(10.159)
which yields the stationary state population
(10.160)
Interestingly, within this simple model N 2s does not depend on the pump rate
R; it depends only on the cross section for stimulated emission and the decay
time of Φ. For a He-Ne gas laser with τ r = 3 × 10−8 s and σ = 3 × 10−13 cm2
we obtain:
(10.161)
For a Nd:YAG solid-state laser the steady-state density is:
(10.162)
Note that τ r does not depend on the active material, but on the quality of the
resonator. Next, we determine the critical pump rate above which laser
operation is observed. The latter requires Φ s > 0. Insertion of Eq. (10.160)
in Eq. (10.156) yields an expression for the photon density as a function of
the pump rate:
(10.163)
From the condition , we obtain the critical pump rate
(10.164)
For the He-Ne gas laser, we obtain:
(10.165)
For the Nd:YAG laser, the critical pump rate is orders of magnitude higher:
(10.166)
On the one hand, the Nd:YAG laser has a much longer lifetime of the upper
laser level, which reduces R crit, but this is overcompensated for by the much
smaller cross section for stimulated emission. Different types of lasers may
also differ in their dynamic behavior, which we investigate in subproblem
(c). We examine how the system behaves if the steady-state values of N 2 and
Φ s are distorted by small deviations n and ϕ respectively. We show that the
systems behave like a damped oscillator . If we insert N 2(t) = N 2s + n(t) in
Eq. (10.156) we obtain
(10.167)
and thus
(10.168)
The nonlinear term containing the product n(t)ϕ(t) is ignored, as n and ϕ are
considered to be small distortions. In the same way, we treat Eq. (10.157)
and obtain
(10.169)
(10.171)
This differential equation has the form Eq. (10.158). It describes damped
oscillations with angular frequency
(10.172)
and damping parameter
(10.173)
For the analysis of the dynamic behavior we assume a pump rate that is twice
the critical pump rate for laser operation. With R = 2R crit the angular
relaxation, frequency is
(10.174)
and the damping parameter is
(10.175)
Now we solve Eq. (10.158), which is a homogeneous differential equation
with constant coefficients. The general solution is:
(10.176)
where the constants c 1 and c 2 follow from the initial conditions. The
parameters λ 1 and λ 2 are the roots of the characteristic polynomial
(10.177)
which we obtain using Eq. (A.4):
(10.178)
Now we need to consider two cases regarding the sign of the discriminant δ 2
−ω 2. If it is positive then λ 1 and λ 2 are real and the solution for the initial
condition n(t = 0) = n 0 is19
(10.179)
If, however, δ −ω 2 2 is negative, then λ 1 and λ 2 are complex conjugate
numbers, and . Insertion in Eq.
(10.176) reveals the damped oscillatory behavior of n(t) in this case20:
(10.180)
Table 10.9 Parameters influencing the dynamic behavior of lasers concerning small deviations from
the stationary state
Fig. 10.27 Relaxation dynamics for the Nd:YAG laser, calculated with Eq. (10.180) and the
parameters given in the text. Small distortions decay in a damped oscillation within the time scale of 1
ms
Fig. 10.28 Relaxation dynamics for the He-Ne laser, calculated using Eq. (10.179) and the parameters
given in the text. Small distortions n(t) are damped out within the time range of 1 μs
With the parameters τ and τ r given for the He-Ne laser and the Nd:YAG
laser we can calculate ω and δ according to Eqs. (10.174) and (10.175). The
results are given in Table 10.9. For the Nd:YAG laser ω largely exceeds the
damping parameter δ for the assumed pump rate. This laser type thus exhibits
relaxation oscillations, as shown in Fig. 10.27. The oscillations have a
frequency of:
(10.181)
and decay within 1 ms. In contrast, in the case of the He-Ne laser, small
distortions from the stationary state are damped out within 1 μs without
oscillating behavior, as illustrated in Fig. 10.28.
In the next problem, a numerical solution of the laser equations is
examined that confirm the oscillatory behavior around the stationary state
seen in the case of the Nd:YAG laser.
This problems assumes that you have dealt with Problem 10.10.
Consider the four-level laser model introduced in Problem 10.10 (Fig.
10.26). Solve the laser equations
(10.182)
(10.183)
for the population of the upper laser level N 2 and the photon density in
the oscillator Φ numerically either by using mathematical software or
by writing a computer code based, for example, on the Runge-Kutta
scheme of integration (see appendix Sect. A.3.18). Examine the switch-
on behavior of the Nd:YAG laser by selecting the initial conditions N 2
= 0 and Φ(t = 0) = ε where ε is a small positive number. Use
parameters given in Problem 10.10, simulate on a time scale that
covers a few milliseconds, and assume a pump rate R = 5R crit. Why do
you need ε > 0? Can you explain the complex switch-on behavior?
Fig. 10.29 Switch-on behavior of an Nd:YAG laser obtained from a numerical simulation based on
Eqs. (10.182) and (10.183). The population of the upper laser level as a function of time is shown. The
dashed line is the behavior according to Eq. (10.186)
Solution 10.11
This exercise assumes that you have already dealt with Problem 10.10,
where the principle of laser light sources was investigated and a stationary
solution for the rate equations Eqs. (10.182) and (10.183) was determined. In
addition, small deviations from the stationary state values N 2s and Φ s could
be treated by neglecting the nonlinearity in these equations. A full solution of
these equations requires, however, numerical methods. As for the Lotka-
Volterra model for oscillating chemical reactions in Problem 6.7 we can use
the Runge-Kutta method (Sect. A.3.18) to integrate Eqs. (10.182) and
(10.183) into a discrete grid with a time step interval h = t n+1 − t n (n = 1, 2,
…). Starting from the initial values N 2(0) = 0 and Φ(0) = 0, we investigate
the switch-on behavior of the Nd:YAG laser. In Problem 10.10 we have seen
that this type of laser shows the tendency toward relaxation oscillations. If
the pumping is switched on at t = 0, the laser is far from its stationary state
and it is interesting to examine how the stationary state is established. There
is a hint given that we assume a small initial value of the photon density
Φ(0) = ε. The reason becomes clear if we inspect Eq. (10.183): the gain of
the photon density described by the second term is proportional to Φ. Hence,
if Φ = 0 then the photon density stays exactly zero for t > 0. We have thus to
start from a small photon density ε, say Φ(0) = 0. 1, resulting, for example,
from noise in the resonator. An analogous problem does not occur for N 2
because the gain term of N 2 is the constant pump rate R. For the simulation,
we assume a pump rate of R = 5R crit. Let us calculate the expected stationary
state values N 2s and Φ s . The stationary population density N 2s is
independent of R and takes the value 5. 96 × 1016 cm−3 (see Eq. (10.162)). By
means of Eq. (10.163) we obtain:
(10.184)
where we have used the parameters σ = 8 × 10−19 cm2 and τ = 2. 5 × 10−4 s−1.
For the simulation, it is crucial that the integration of Eqs. (10.182) and
(10.183) is performed on a fine grid, e.g., h = 10−11 s. Results of the
simulation are shown in Figs. 10.29 and 10.30. N 2 starts from zero and
continuously grows to reach a value of 1. 2 × 1017 cm−3. During this time
period, the photon density in the resonator is close to zero. At t ≈ 0. 13 ms,
however, Φ exhibits a sharp spike and N 2 is suddenly reduced to about 2. 5
× 1016 cm−3, increases again until at 0.21 ms, the population of the upper
laser level is again suddenly reduced, accompanied by another spike in the
photon density. This pattern is repeated in the following, with the tendency
that the time between consecutive spikes of Φ and setbacks of N 2
respectively, becomes shorter and shorter. In Fig. 10.31, a snapshot of the
simulation between 0.50 and 0.52 ms is shown where the sawtooth-like
variation of N 2 and the spikes of the photon density are seen in more detail.
Fig. 10.30 Switch-on behavior of an Nd:YAG laser obtained from a numerical simulation based on
Eqs. (10.182) and (10.183). The photon density within the resonator is shown to be a function of time.
Note the logarithmic scale for Φ
Fig. 10.31 Population of the upper laser level N 2 and photon density Φ as a function of time.
Snapshot of the simulation between 0.5 and 0.52 ms.
Fig. 10.32 Population of the upper laser level N 2 and photon density Φ as a function of time.
Snapshot of the simulation between 0.8 and 0.82 ms
Figure 10.32 shows the behavior between 0.80 and 0.82 ms. The
repetition rate of spikes has become higher, and they are wider. In Fig. 10.33,
where a snapshot of the simulation between 1.20 and 1.22 ms is shown, N 2
and Φ exhibit a nearly harmonic oscillatory behavior around the stationary
state values calculated above.
Fig. 10.33 Population of the upper laser level N 2 and photon density Φ as a function of time.
Snapshot of the simulation between 1.2 and 1.22 ms
Looking back on our results we were able to numerically solve the laser
equations and to simulate the switch-on behavior of the Nd:YAG laser
including the appearance of spikes and the relaxation from a highly nonlinear
behavior into the stationary state. Oscillations around the stationary state
treated in Problem 10.11 are confirmed by the simulation.
References
1. Louisell WH (1973) Quantum statistical properties of radiation. Wiley, New York
Footnotes
1 We have dealt with a quantum mechanical description of a two-level system in Problem 9.11.
4 The exact result, determined using E = 13. 605693 eV would be 109,737 cm−1.
5 Conservation of energy and momentum requires the generation of two photons. Each photon has the
energy E = m e c 2 = 9. 10938 × 10−31 kg × (299, 792, 458 m s−1)2 ≈ 511 keV.
6 At high concentrations the effects of scattering including multiple scattering events need to be taken
into account.
7 See Problem 9.18 dealing with chemical bonding in the simplest possible molecular system, the H +
2
molecule ion.
9 The deviations are due to the fact that the intersection lines have a curvature.
11 The Q-branch would require Δ J = 0 which is forbidden. Infrared active modes of polyatomic
molecules, however, have a Q-branch if the induced electric dipole moment is oriented perpendicular
to the symmetry axis of a molecule. A Q-branch is also observed in the vibration rotation spectrum of
the paramagnetic NO molecule.
12 Values of rotational constants obtained from infrared spectroscopy are often presented in units of
cm−1. The numbers in brackets indicate the error of the fit in units of the last significant digit.
13 Apart from the point symmetry elements mentioned, the inversion center, horizontal mirror planes
and improper axes are possible point symmetry elements.
15 The parity operation is an inversion with the result (x, y, z) → (−x, −y, −z) or (r, θ, ϕ) → (r, π −θ, ϕ
+π).
16 This follows from the fact that harmonic oscillator wave functions (Eq. (A.76), see also Fig. 9.1),
have even parity for even quantum numbers n = 0, 2, 4, …, and odd parity for odd quantum numbers n
= 1, 3, 5, ….
17 Note that there are at least three different methods for the notation of vibrational states of the CO2
molecule in use in textbooks and in the scientific literature. The notation used in this problem is
sufficient here, but it is not the most general one.
18 The word laser is the abbreviation for light amplification by stimulated emission of radiation.
19
Note that the hyperbolic cosine function is defined .
20
Note that (see Eq. (A.11) in the appendix).
Appendix A
A.1 Physical Constants
In the following table the values of the fundamental physical constants are
collected (Tables A.1 ).
Table A.1 Derived constants contain experimental uncertainties in brackets in units of the last digit
Source: P.J. Mohr, B.N. Taylor, D.B. Newell, CODATA recommended values
of the fundamental physical constants: 2010 , Rev. Mod. Phys. 84 (2012),
1527
Rydberg Ry
J
(A.4)
where Δ = b 2 − 4 ac is called discriminant . If Δ = 0, then the quadratic
equation has only one root; if Δ > 0, it has two real solutions, if Δ < 0 the
solutions are complex.
A.3.3 Logarithms
(A.5)
(A.6)
(A.7)
A.3.5 Derivatives
A.3.5.1 Basic Differentiation Rules
The sum rule
(A.14)
The product rule
(A.15)
The quotient rule
(A.16)
The chain rule
(A.17)
A.3.5.2 Basic Derivatives
(A.18)
(A.19)
(A.20)
(A.21)
(A.22)
(A.23)
(A.24)
(A.25)
(A.26)
A.3.5.3 Partial Derivatives and Total Derivatives
Consider a function f ( t , x , y ) depending on several variables x , y , and t .
As in the case of a function with one variable, the partial derivative with
regard to x is defined by the difference quotient in the limit h → 0:
(A.27)
The rules for partial differentiation follow the above rules for the
differentiation of a function with only one variable; the other variables are
held constant. Example: consider f ( t , x , y ) = 3 x 2 y − t . The partial
derivative with regard to x is:
(A.28)
Consider a case in which the variables x ( t ) and y ( t ) themselves depend
on t . In this case, the total derivative of a function f ( t , x , y ) with regard
to time is:
(A.29)
(A.30)
Sum rule
(A.31)
Integration by parts
(A.32)
(A.33)
with the Jacobian
(A.34)
The method can be generalized to functions with more than two
variables.
(A.47)
(A.48)
(A.49)
(A.50)
(A.51)
Dirac’s delta function :
(A.52)
For an arbitrary function f ( x ), the following equation holds:
(A.53)
(A.59)
(A.60)
(A.63)
(A.64)
μ is the average of the random variable x , σ 2 its variance , and σ its standard
deviation (Fig. A.2 ). The half width or full width at half maximum F W H
M is
(A.65)
(A.68)
(A.69)
(A.71)
(A.72)
(A.73)
(A.74)
(A.75)
The formula for general n is:
(A.76)
where H n ( y ) is a Hermite polynomial . To obtain wave functions with
higher n , the recurrence relation
(A.77)
can be used starting from H 0 ( y ) = 1 and H 1 ( y ) = 2 y .
(A.78)
(A.79)
(A.80)
(A.81)
(A.82)
(A.83)
For general l , m :
(A.84)
with the associated Legendre polynomial
(A.85)
A.3.15 Radial Wave Functions of the Hydrogen
Problem
Explicit radial wave functions R nl ( r ) for the n = 1, 2, 3, l = 0, … , n − 1
(Bohr radius ):
(A.86)
(A.87)
(A.88)
(A.89)
(A.90)
(A.91)
For general n , l , the radial wave function is:
(A.92)
A.3.16 Matrices
A ( n , m ) matrix A is an arrangement of numbers in n rows and m columns:
(A.93)
(A.97)
The inverse A −1 of a ( n , n ) square matrix A is defined by
(A.98)
where is I n is a ( n , n ) unit matrix.
(A.99)
in which the i -th column is replaced by the elements of the vector y . Then,
the solution vector x is given by the n elements
(A.101)
(A.102)
has the analytic solution
(A.103)
C is an integration constant defined by the boundary conditions, and M is
the integrating factor :
(A.104)
(A.105)
(A.107)
(A.109)
is called a Bernoulli equation. A Bernoulli equation can be transformed into
a linear differential equation, if we introduce:
(A.110)
Then, we obtain:
(A.111)
which can be treated as described in Sect. A.3.18 .
(A.114)
(A.115)
(A.116)
Fourth-order Runge-Kutta method (error O ( h 5 ))
(A.117)
(A.118)
(A.119)
(A.120)
(A.121)
Greek symbols:
Index
Absorbance
Absorption process
Activity
Activity coefficient
Anharmonicity constant
Annihilation operator
Anti-Stokes scattering
Arrhenius equation
Associated Legendre Polynomial
Asymmetric top molecule
Asymmetry parameter
Atomic units
Attenuance
Autocatalytic reaction
Avogadro constant
Avogadro, Amedeo
Band model
metal
semi-conductor
Basis set
Bernoulli numbers
Bimolecular reaction
Black body
Black body radiation
Bohr radius
Boltzmann constant
Boltzmann, Ludwig
Bond dissociation energy
Born Haber cycle
Born, Max
Born-Oppenheimer approximation
Bra-vector
Calorimetry
Catalyst
Cathode rays
Center of mass
Character table
Characteristic polynomial
Chemical accuracy
Chemical potential
Chemical reaction
oscillating
Chemisorption
Chromophor
Clapeyron equation
Clausius-Clapeyron equation
Colligative properties
Collisional broadening
Commutator
Complex number
Complimentary error function
Compressibility
Compression factor
Concentration
Correlation diagram
Correspondence principle
Creation operator
Critical point
Cubic force constant
Cumulative probability
Cylindrical coordinates
Cysteine
Dalton, John
de Broglie wavelength
de Broglie, Louis-Victor
Degeneracy
Delta function
Derivative
partial
total
Diffusion
Diffusion constant
Discriminant
Dissociation constant
Dissociation reaction
Doppler effect
Double well problem
Educt
Effective mass
Effusion
Einstein coefficients
Einstein, Albert
Electron impact heating
Elementary reaction
Ellingham diagram
Emission process
induced
spontaneous
Endergonic process
Endothermic process
Energy
Energy of activation
Ensemble
Enthalpy
Enthalpy of formation
Enthalpy of reaction
Entropy
conformational
monatomic gas
Entropy of reaction
Equation of state
perfect gas
van der Waals gas
virial
Equilibrium
chemical
mechanical
thermodynamic
Equilibrium constant
Equipartition theorem
Error function
Euler angles
Euler equation
Euler’s formula
Exchange operator
Exergonic process
Exergonic reaction
Exothermic process
Extensive quantity
Extent of reaction
Extinction
Factorial
Fick’s second law
First law of thermodynamics
Fraunhofer, Joseph
Free expansion
perfect gas
van der Waals gas
FWHM
Gamma spectroscopy
Gaussian
Gibbs free energy
Gibbs paradox
Glutamate dehydrogenase
Group theory
Haken, Hermann
Half width
Half-life
Hamilton function
Hard sphere model
Harmonic oscillator
Hartree
Heat
Heat capacity
isobaric
isochoric
Heat capacity ratio
Heat conduction
Helmholtz free energy
Henderson-Hasselbalch equation
Hermite polynomial
Hess’s law
Hot band
Hyperfine structure splitting
hydrogen
sodium
Ice
Ideal mixture
Imaginary unit
Impingement rate
Intensive quantity
Internal energy
Internal pressure
Inversion
Irreducible representation
Irreversibility
Irreversible process
Isotopic substitution
Jacobian
Jaynes Cummings model
K-capture
Ket-vector
Kirchoff’s law
Knudsen number
Kronecker delta
Lambert Beer law
LASER
Law of mass action
Levinthal paradox
Limiting reactant
Liquid petroleum gas
Lotka-Volterra model
Mößbauer spectroscopy
Macrostate
Matrix
Maxwell construction
Maxwell relations
Maxwell-Boltzmann velocity distribution
Microstate
Molar gas constant
Molar mass
Molar volume
Mole fraction
Molecular orbital
Molecularity
Moment of inertia
Moment of inertia tensor
Morse potential
Multiplicity
Normal distribution
Nuclear spin statistics
Number operator
Operator
adjoint
annihilation
creation
exchange
Hermitian
Laplace
momentum
number
parity
self adjoint
unitary
Optical depth
Ortho hydrogen
Oscillator
damped
Lorentz
Ostwald solubility coefficient
Overtone transition
P-branch
Para hydrogen
Parity operation
Partial derivative
Partition function
canonical
molecular
Pauli principle
Pauli spin matrices
Perfect gas
Permutation
Phase diagram
Photoelectron spectroscopy
Physical constants
Physisorption
Planck constant
Planck, Max
Point group
Point symmetry
Positron emission tomography
Pre-exponential factor
Pressure
Probability
Propagator
Q-branch
Quantum chemistry
Quantum defect
Quantum double-well
asymmetric
symmetric
Quantum number
magnetic
orbital angular momentum
orientational
principal
rotational
vibrational
Quartic force constant
R-branch
Radio astronomy
Raman scattering
Random walk
Raoult’s law
Rate constant
Reactant
Reaction rate
Representation
irreducible
reducible
Rigid rotator
Rotational constant
Runge-Kutta method
Rutherford, Ernest
Rydberg atoms
Sackur-Tetrode equation
Scanning force microscopy
Scanning tunneling spectroscopy
Schottky anomaly
Schrödinger equation
Second law of thermodynamics
Semiconductor
Sodium D-line
Spectroscopy
rotational
vibrational
Spherical harmonics
Spin-statistics theorem
Standard deviation
Standard state
Statistical weight
Stefan-Boltzmann law
Stirling approximation
Stirling formula
Stoichiometric number
Stokes scattering
Supercritical phase
Superposition principle
Surface science
Surface tension
Symmetric top molecule
Symmetry operation
Symmetry type
Syngas
Synthesis reaction
System
closed
heterogeneous
isolated
nonconservative
open
Tempering
Term symbol
Termolecular reaction
Thermal expansion coefficient
Thermal wavelength
Thomson, Joseph John
Time invariance
Total derivative
Transition state
Transmittance
Tryptophan
Tunnel effect
Two-level system
Tyrosine
Uncertainty principle
Unimolecular reaction
Unitary transformation
van der Waals equation
van der Waals model
Van’t Hoff reaction isobar
Variance
Virial coefficient
second
third
Virial equation
Water gas shift reaction
Wave packet
Wave-particle duality
Wigner rotation functions
X-rays
Yukawa, Hideki
Zeeman effect
Zeeman, Pieter
Footnotes
1 Source: W.H. Press, S. Teukolsky, W.T. Vetterling, B.P. Flannery Numerical recipes in C: the art of
scientific computing , Cambridge University Press, Cambridge, 1988