GENERAL ARTICLE

On the Saha Ionization Equation

Sushanta Dattagupta

We revisit the Saha Ionization Equation in order to high-

light the rich interdisciplinary content of the equation that
straddles distinct areas of spectroscopy, thermodynamics and
chemical reactions. In a self-contained discussion, relegated
to an appendix, we delve further into the hidden message of
Sushanta Dattagupta, having
the equation in terms of rate theory. We empoly a pedagogi-
spent more than forty years
cal style appropriate for a course in equilibrium and nonequi- in teaching, research and
librium statistical mechanics. administration in various
institutions and universities
across India, is now a senior
1. Introduction scientist of the Indian
National Science Academy.
In a couple of years, we are going to observe the centenary of He has written extensively, in
a remarkable equation known as the Saha Ionization Equation journals and books, on topics
of condensed matter,
(SIA), also referred to as the Saha Ionization Formula (SIF) writ-
non-equilibrium phenomena,
ten down by Meghnad Saha [1]. At that time Saha was about and more recently Tagore
twenty-six, exactly the same age as another great Bengali physi- Model of education. His
cist and a close friend – Satyendra Nath Bose. About the cele- current physics interests are
in quantum dissipation and
brated equation, Bose is quoted to have remarked to a gathering
stochastic thermodynamics of
of his students: “Do you know what Meghnad did? Sitting at nanoscopic systems.
home, he measured the temperature of stars!” [2]. The originality
and the far-reaching consequences of the SIA are so remarkable
that the said theory is said to be the harbinger of modern astro-
physics [3]. Saha’s work created a stir in then-Calcutta which
was, at that time, arguably the centre of not just physics but all
forms of intellectual pursuits of India. An example of the soci-
etal impact of the famous equation can be discerned in a popular
book on science in Bengali written by Asia’s first Nobel Laureate
Rabindranath Tagore, in the mid nineteen thirties, named Viswa-
Parichay (An Introduction to the Universe). Tagore wrote and we Keywords
translate: “It is expected that the ninety-two elements that have Ionization, astrophysics, spec-
troscopy, chemical reaction, tran-
been found on the Earth must have had their traces in the Sun too.
sition state.
Only thirty-six of these were however discovered in preliminary

RESONANCE | January 2018 41

 GENERAL ARTICLE

Saha moulded and investigations. It is the Bengali scientist Meghnad Saha who has
integrated the unearthed (the pun is mine) the puzzle of the missing elements in
just-evolved quantum
the Sun” [4].
model of Bohr and the
consequent spectroscopy Much has been written about the history of physics preceding
of atoms, the extant
Saha’s pioneering research, the West’s grudging acceptance of
knowledge of
thermodynamics, and his equation and how the latter had influenced and spurred sub-
what is also significant – sequent activities in astrophysics [5]. A detailed biography along
the theory of chemical with a lucid account of the SIA has also been presented in a form
rate processes. readily available to university students [6]. Our aim in this arti-
cle is not to delve into history but to offer a critical assessment
of how Saha moulded and integrated the just-evolved quantum
model of Bohr and the consequent spectroscopy of atoms, the ex-
tant knowledge of thermodynamics, and what is also significant
– the theory of chemical rate processes. Indeed our emphasis is
to put Saha’s work in the modern context of equilibrium and non-
equilibrium statistical mechanics, in a fairly nontechnical format.
Hopefully, our relook at this nearly hundred-year-old equation
would illustrate to today’s young students the importance of inter-
disciplinary research.
An outline of the paper is as follows. In Section 2 we indicate
what spectroscopic considerations led Saha to formulate his sim-
plified model of the ionized stellar atmosphere, and inter alia, in-
troduce an important parameter called the ionization potential de-
noted by U, by way of the specific case of sodium [6]. This part of
the discussion relates to Saha’s use of the Bohr model of the atom
and measurement of the spectral line that yields U. Section 3 is the
core of the paper which describes how Saha ingeniously took re-
course to the notion of thermodynamic equilibrium in the context
of chemical reactions. The fascinating simile that Saha employed
between ionization and chemical reactions allowed him to relate
the concentrations of species involved in the ionization (disso-
ciation) process to the exponential of a parameter (−U/kB T ), kB
being the Boltzmann constant and T the temperature. Now comes
another startling cross-border foray – Saha wrote the said concen-
trations in terms of ratios of partition functions – a key concept of
Gibbs’ statistical mechanics.

42 RESONANCE | January 2018

 GENERAL ARTICLE

As we mentioned at the outset, our objective here is not merely a The key hidden idea of
recounting of what Saha did but, more importantly perhaps, what Saha, and indeed also of
Eyring, is that the rate is
he had anticipated and what he had implied but did not spell out!
actually a ‘steady-state
Recall that the enunciation of the SIA had preceded by almost rate’, and hence, the
fifteen years Eyring’s ‘Transition State Theory’ of chemical reac- underlying ‘chemical
tions, which very craftily connected the rate of chemical reactions equilibrium’, in point of
to the statistical mechanics of partition sums (see [8] and accom- fact pertains to a
stationary or a steady
panying articles on p.11 and p.103 in this issue). While making state which must be the
this analysis of post-Saha developments we wish to point out that culmination of a
a concept like a rate that connotes to ‘a change of something with dynamical process that
respect to time’, is outside the realm of thermodynamics and sta- describes the approach
to equilibrium.
tistical mechanics, as the latter deals with phenomena in thermal
equilibrium. The key hidden idea of Saha, and indeed also of
Eyring, is that the rate is actually a ‘steady-state rate’, and hence,
the underlying ‘chemical equilibrium’, in point of fact pertains
to a stationary or a steady state which must be the culmination
of a dynamical process that describes the approach to equilib-
rium. The latter process has to be described by going beyond
equilibrium statistical mechanics and by writing time-dependent
rate equations that explicitly violate time-reversal invariance and
incorporate dissipation – ingredients which are outside the scope
of the formulation of Gibbs and indeed that of Saha. The steady-
state rate has to be extracted from the asymptotic (i.e., time t go-
ing to infinity) limit of the rate equations in terms of ratio of parti-
tion functions, as anticipated by Saha. Such a theory is attributed
to the path-breaking work of Kramers which, in the language of
chemical reactions – the language employed by Saha – analyzes
a diffusion equation of a Fokker–Planck structure that builds-in,
through the so-called ‘drift’ term, a thermodynamic ‘free’ energy
as a function of reaction coordinates [9]. How the Kramers rate
gets linked to the ratio of partition functions of the ‘products’ and
‘reactants’, as in the SIA, is relegated to the Appendix (which
is intended for the more advanced and curious reader). Finally,
Section 4 offers a few concluding remarks.

RESONANCE | January 2018 43

 GENERAL ARTICLE

2. The Ionization Potential

Saha’s basic premise was Saha’s basic premise was that the stellar atmosphere is so hot
that the stellar and at such a low density that the constituent atoms are in their
atmosphere is so hot and
gaseous state and they behaved like a dilute classical gas, as in
at such a low density that
the constituent atoms are Maxwell’s kinetic theory. That is to say, the atoms do not inter-
in their gaseous state and act with each other except during collisions which are viewed as
they behaved like a instantaneous. Furthermore, because of high temperatures, the
dilute classical gas, as in atoms can be stripped of their valence or outer-shell electrons and
Maxwell’s kinetic
theory. be ionized, leading to a plasma of ions and electrons. In a dilute
gas the ions and the electrons are so far apart that the so-called
Debye screening due to electrostatic Coulomb interactions can
be ignored. For the sake of definiteness, we exemplify Saha’s
treatment by citing the case of sodium (Na) [7, 10]. The latter
has a single electron in the outermost shell in the 3s state, that
is to say, the principal quantum number n in the Bohr–Rydberg
sense equals 3, and the orbital angular momentum l equals 0. The
ionization process involves Na losing its 3s electron yielding a
singly-ionized Na+ in a process of dissociation that can be sym-
bolically designated:

Na → (Na+ ) + e− . (1)

Clearly, the emitted electron would possess a kinetic energy that

would be governed by the temperature of the gas via the clas-
sical equipartition theorem. The reverse process of association,
in which an ionized sodium atom captures an itinerant electron,
would yield a neutral atom but in an excited state, concomitant
with the release of energy in the form of emission of a photon of
frequency ν:

(Na+ ) + e− → Nanorm + hν , (2)

h being the Planck constant. Within the theories of Eyring and

Kramers the process in (2) can be diagrammatically represented
by Figure 1, in which the transition state is an activated ‘complex’
– Naexc , in the present case. It is interesting also to write the

44 RESONANCE | January 2018

 GENERAL ARTICLE

Figure 1. Energy versus re-

action coordinate, in which
State 1 corresponds to the
‘reactants’ – in this case Na
ion and electron, and State
2 corresponds to the ‘prod-
ucts’, that is normal Na. The
transition state is a saddle
point that describes an acti-
vated complex in the form of
Na in the excited state, in the
present scenario.

energy balance akin to the so-called ‘photo-ionization’ process,

in the light of Figure 1:

Nanorm + hν = Na+ + e− + hνK , (3)

where hνK is the energy corresponding to the kinetic energy of

the ejected electron. Photo-ionization is then exactly akin to the
photoelectric effect in which an electron is emitted by a metal
upon being impinged by an incident electron. Evidently,

U = h(ν − νK ) . (4)

In the context of photoelectric effect, U is called the work func-

tion of the metal.
The SIA, in this case of sodium, can then be simply written as:

[(x+ ) · (x− )]/x = [2πme (kB T )/h2 ]3/2 · exp[−(U/kB T )] , (5)

RESONANCE | January 2018 45

 GENERAL ARTICLE

U being the ionization potential. The lower case x’s denote the
respective concentrations of positive ions, electrons, and neutral
atoms. The left-hand side of (5) is often referred to as ‘equilib-
rium constant’ or ‘chemical constant’, in the literature on chem-
ical reactions. Saha further expressed the left hand side of (5)
in terms of dimensionless concentrations and the pressure of the
stellar gas with the aid of the ideal gas law. The derivation of (5) is
provided below, in Section 3, and more elaborately, in Appendix
1.
If the excited state corresponds to the outer electron being in the
4p state (n = 4, l = 1) with the total angular momentum j = 1/2, the
photon would have a wavelength of 5896 Angstrom, while a 4p -
excited state (n = 4, l = 1) with j = 3/2 would lead to a photon of
wavelength 5890 Angstrom, as illustrated in [1]. These two spec-
tral lines form the familiar doublet, with a wavelength difference
of 6 Angstrom due to spin-orbit interaction, that can be viewed in
a sodium vapour lamp in an undergraduate physics laboratory. Of
course, depending on the kinetic energy of the electron prior to its
capture, the excited state is not just limited by the ones associated
with the familiar doublet but indeed is one of a whole lot of states
of the S-, P-, D-series (and n = 3, 4, 5,... and others, going into
the continuum). Besides, there can be multiple ionizations in-
volving more than one electron. Now comes the crucial link with
the Bohr quantum theory of atom. Saha surmised that the emitted
photon must be the result of a transition from one discrete level
to the ‘normal’ 3s-level (lower in energy) such that the energy of
the photon would be related to the ionization potential U by (4).
However, it may be pointed out that the transition from a discrete
level to the ground level can have an energy less that U - e.g., in
hydrogen, 2p to 1s transition would yield 10 eV as opposed to
13.6 eV for U. As we had explained, there would be not just one
U but a series of them, which were all measured by Saha. He also
looked at the spectrograph of numerous other stellar gases data
from Harvard University and extracted U for all of them. Note
that quantum mechanics enters into consideration only with re-
gard to the identification of U while the rest of the theory simply

46 RESONANCE | January 2018

 GENERAL ARTICLE

hinges on the classical kinetic theory of gases.

In order to get an idea of the numbers involved, it is best to look at
a simpler dissociation process, that of neutral hydrogen into a hy-
drogen ion and an electron, like in (1), and treated in [11]. In this
case, there is a solitary ionization potential U whose magnitude is
13.6 eV, of the order of the Coulomb energy, if we take the atomic
radius to be about 1 Angstrom. Converting it to wavelength λ we
obtain λ ∼ 103 Angstroms, and a temperature T ∼ 1.6 × 105
K. Hence, the transition state theory works well for interstellar
temperatures even somewhat higher than 104 K, and of course at
lower temperatures.

3. Theory of Chemical Reactions

As has been stressed earlier, Saha’s goal was to determine the

temperature of stars from spectroscopic data. Clearly, therefore,
he needed a theory in which the ionization energy U and the ther-
mal energy kB T ought to enter on an equal footing. To this end,
we can imagine à la Saha, the association process of (2) to be akin
to a chemical reaction of two species (Na+ ) and (e− ) resulting into
neutral Na [11].
Onset of chemical equilibrium implies that the net free energy in
the above process must vanish. Recall that the differential change
in the free energy F, from thermodynamics, is given by:

dF = −PdV − S dT + μdN , (6)

where P is the pressure, V is the volume, S is the entropy, T is

the temperature, μ is the chemical potential, and N is the number
of species. If the reaction takes place at constant volume and
constant temperature, the first two terms on the right of (6) are
zero, and hence, equilibrium (dF = 0) means the net μ dN must
vanish [12]. Because the magnitude of dN is unity, for all the
three species in (2),

(μ+ ) + (μ− ) = (μ) , (7)

RESONANCE | January 2018 47

 GENERAL ARTICLE

where, by (+) we mean (Na+ ), by (−) we mean (e− ), and μ without

superscript is meant for neutral Na. Combining (4) and (5),

(∂F/∂N)+V,T + (∂F/∂N)−V,T = (∂F/∂N)V,T . (8)

Until now, only thermodynamic considerations have been em-

ployed. But, following Saha, in order to bring-in U, we now need
to express the derivatives in (8) in terms of the partition function
Z of statistical mechanics [12, 13]. Recall that Z and F are con-
nected by [11–13]:

F = −kB T ln Z, (9)

and Z is the product of two components, one for translational

motion of the species (no rotational or vibrational component is
relevant for (2)) and the other for the internal (quantum) variables
[11]:

Z = [(Z)tran (Z)int ]N , (10)

where the presence of the power N hinges on our assumption that

the species N do not interact with one another. Evidently, the
translational component relates to the kinetic energy of motion,
and hence,

[(Z)tran ]N = [1/(N!)]{V(2me πkB T )3/2 /[(h)3 ]}N . (11)

The mass m for Na and Na+ can be taken to be nearly the same
and hence, the respective translational contributions to Z are equal,
leaving only the kinetics of the electron to reckon with, for which
however, there is no internal contribution to Z! The pre-factor
(N!) takes care of the correct ‘Boltzmann counting’ in view of
the indistinguishability of individual species. Using the Stirling
approximation,

ln (N!) ∼ N[ln (N)1],

(7–9) can be combined to yield for the chemical potential μ:

48 RESONANCE | January 2018

 GENERAL ARTICLE

μ = (∂F/∂N)V,T = kB T ln(Z)int + ln[(V/N) · (2me πkB T )3/2 /h3 ].

(12)
Incidentally, (11), in combination with the Stirling formula, yields
the Sackur–Tetrode equation for the entropy of an ideal gas at
high temperatures, that Saha was quite aware of [13]. Balancing
the chemical potentials as in (7) and writing the ratio N/V as the
concentration x, we have:

[ln(Z)int −ln(Z + )int ] = ln(x+ )+ln(x− )−ln(x)−ln{[2πme (kB T )/h2 ]}3/2 ,

(13)
me being the electron mass. Rearranging terms,

[(x+ ) · (x− )]/x = [2πme (kB T )/h2 ]3/2 · [(Zint )]/[(Zint

+
)]. (14)

Using (9) the ratio of the internal partition functions turns out to
be the exponential of the difference of the respective free energy
which, for internal dynamics, is precisely the ionization potential
U. Thus, we are led to the SIA quoted earlier in (5). Couched
though in the simplified scenario of Na atoms, (14), in essence,
is the core thesis behind the SIA. As anticipated in the first para-
graph of this section, what appears in the equation is the dimen-
sionless ratio of the ionization energy and the thermal energy.

4. Epilogue

In conclusion, we go back to where we started. As we have em-

phasized in the introductory section, our motivation to write this This pedagogical essay
pedagogical essay was to take stock, after nearly one hundred was to take stock, after
nearly one hundred
years, of the ionization equation, and to underscore the signifi-
years, of the ionization
cance of interdisciplinary signatures behind Saha’s futuristic and equation, and to
far-reaching thoughts. Today, we have the advantage of hind- underscore the
sight – over nearly a century – but it is amazing to ponder on significance of
how a young Indian scientist, working under adverse conditions interdisciplinary
signatures behind Saha’s
of colonial regime, dared to piece together a variety of phenom- futuristic and
ena, spread over different topics of science, to lay the foundation far-reaching insights.

RESONANCE | January 2018 49

 GENERAL ARTICLE

of a completely novel area – that of stellar atmosphere. Through

this essay we wish to pay tribute to this great man, an institution-
builder, a champion of research in universities, a planner of post-
independent India – especially with regard to atomic energy and
the river system [5, 6] – who is alas largely forgotten in the hum-
drum of today’s pursuit of cocooned, specialized research.

5. Acknowledgments

While carrying out this work I have benefitted from useful in-
puts by Profs. Mustansir Barma, Jayanta Bhattacharjee, Kumar
Chitre, Naresh Dadhich, Partha Ghose, Harihara Subra Mani, and
Narayanasami Sathyamurthy, to whom I am indebted. I am par-
ticularly grateful to Profs. Barma, Chitre, and Sathyamurthy for
critical reading of the manuscript and invaluable comments which
have led to improvement in the presentation. I am also thank-
ful to the Senior Scientist scheme of the Indian National Science
Academy, New Delhi, for supporting my research work at the
Bose Institute. Finally, I record my gratitude to a referee for rais-
ing pertinent issues that have helped enhance the clarity of the
paper.

Appendix 1. Kinetics – Ionization Equation in a Post-Saha Perspective

A moment’s thought would reveal that the derivation of the ionization equation, as sketched in Section 3,
is based on purely equilibrium considerations as extant in thermodynamics and exploits the fact that in
equilibrium, the net chemical potential of the reacting species vanishes. By the same token, a non-zero
net chemical potential ought to be the driving force to take the system out of equilibrium. We may then
invert the question and ask: Can (14) be arrived at, as the end-product of a dynamic process when all time-
dependent variables have reached their constant, equilibrium values? In other words, can we relate the left
hand side of (14) to a rate constant, though Saha actually bypassed the concept of rate in his original papers?
To find the answer to this poser the rate of production of (Na) from (Na+ ) and (e− ) can be formulated in
terms of a first order rate equation of the form:

(d/dt)[x(t)] = γ1 (T )[x+ (t)] · [x− (t)] − γ2 (T )[x(t)], (A.1)

Continued

50 RESONANCE | January 2018

 GENERAL ARTICLE

Appendix 1. Continued.

where the γ’s are temperature-dependent terms which have the dimension of frequency [(t)−1 ] and are appro-
priately called rate constants. Such rate equations are frequently encountered in a variety of situations, such
as radioactive decay, population dynamics of predators and preys, to name a few. Innocuous though they
look, the rate equations of the kind (A.1) disobey time-symmetry as the left hand side changes sign when t is
reversed to −t, an attribute that is absent in Newton’s classical mechanics, Schrödinger’s quantum mechan-
ics or Gibbs’ statistical mechanics! While (A.1) provides a description for a system out-of-equilibrium, a
straightforward way of handling equilibrium would be to set the left side of (A.1) to zero, when [x] reaches
a constant value. In that case,
{[x+ ] · [x− ]}/[x] = [γ2 (T )]/[γ1 (T )]. (A.2)

The condition (A.2) is often referred to as ‘detailed balance’. If the right hand side of (A.2) is designated as
a temperature-dependent constant Ω(T), the latter evidently equals, in the case of association/dissociation
of sodium vapour, the right hand side of the SIA (5). As a matter of fact, Saha did introduce the concept of
‘reaction rate’ without however going through the subtleties discussed in this section [1, 7]. An exponential
temperature dependence of Ω(T) on a parameter, like (-U)/kB T ), is not uncommon in evaporation of liquid
into vapour, or thermal emission of an electron from a filament in a diode [14], etc. Just as evaporation
of water, for instance, can take place at temperatures much lower than the normal boiling point – and at
thermal energies smaller than the latent heat – so can ionization occur at temperatures substantially below
that corresponding to the ionization potential [15]. Indeed, the entire validity of the concept of transition
state hinges on the barrier height (Figure 1) being ‘larger’ than the thermal energy [16].

We wish to dispose of one red herring in the literature discussion – made almost in the same breath – of
an ‘equilibrium’ state and a ‘stationary’ state. To expound on that issue, consider the familiar problem
of electric current flow in a metallic wire when connected to a battery. The velocity v(t) of the electron
(assumed one-dimensional, for the sake of simplicity) obeys the equation [17]:

me (d/dt)v(t) = −me γ(T )v(t) + |e| · E, (A.3)

where |e| is the magnitude of the electron’s charge and E is the electric field intensity provided by the
battery. The temperature-dependent term γ(T ), like the γ’s occurring in (A.3) is also a rate that incorporates
the collision losses of the velocity due to scattering of the flowing electron. The presence of the term γ(T )
explicitly breaks the invariance of an otherwise time-reversal-invariant Newtonian equation. As before, we
arrive at a time-independent solution of (A.3) as:

(v)0 = |e| · E/[me γ(T )]. (A.4)

Writing (A.4) as a current density and relating it to E via Ohm’s law immediately leads to the familiar Drude
formula for electrical conductivity in which the inverse of γ(T ) is often referred to as the ‘relaxation time’.
What this facile derivation of the Drude law camouflages however is the fact that a ‘dissipative’ parameter
Continued

RESONANCE | January 2018 51

 GENERAL ARTICLE

Appendix 1. Continued.

like γ(T ) shows up in a steady-state property like the conductivity! Now comes the punch line: Equation
(A.4) is the culmination of a transport process in the steady state (such as also in the onset of a terminal
velocity of a ball falling under gravity through a Stokes medium of a viscous liquid); on the other hand, the
dissipative rate parameter, which aids the process of reaching thermal equilibrium, is conspicuously absent
in the end-product of the SIA (5)! It is important to bear in mind this distinction between a ‘driven’ steady
state, such as electric conduction driven by a battery, as opposed to a thermal equilibrium state, such as
prevailing in intrastellar ionization, albeit both are time-independent solutions of underlying rate equations.

With this preamble we now outline the theory of activated rate processes due to Kramers which subsumes
Eyring’s earlier work on the transition state theory. The starting point of Kramers is essentially a diffusion
equation for a probability density P(r, t) in the phase space defined by generalized coordinates {r} (referred
to as reaction coordinates in chemical reactions) with however a drift term [9, 17]. The latter has its
origin in a force field that can be viewed as the gradient of a potential energy V(r), or in the more general
thermodynamic setting, a free energy. It is written as

(∂/∂t)P(r, t) = (1/ω) · {(kB T )(∂2 /∂r2 )P(r, t) + (∂/∂r)[V(r) · P(r, t)}, (A.5)

where ω is a constant that fixes the dimension of both sides of the equation. Note that if drift is absent such
as in the continuum version of an unbiased random walk [18], we recover the ordinary diffusion equation:

(∂/∂t)P(r, t) = D(∂/∂r2 )P(r, t) , (A.6)

where the diffusion constant D is given by the Einstein relation:

D = kB T/ω. (A.7)

Further, in the time-independent state, a form of P(r) that is a solution of (A.5) is given by

Peq (r) ∼ exp[−V(r)/(kB T ), (A.8)

which is the Boltzmann–Gibbs measure (un-normalized) for a system in thermal equilibrium. Thus, the
diffusion equation we wrote in (A.5) builds-in, at the outset, the thermal equilibrium state that can be
reached asymptotically (as t goes to infinity). (Needless to say, if the situation so demands, we can also
employ a diffusion equation of the same form as in (A.5) to describe a steady state that is not necessarily
an equilibrium state, in the senses delineated above.) Like the Schrödinger equation, the diffusion equation
(A.5) can be written as a continuity equation:

(∂/∂t)P(r, t) + (∂/∂r)J(r, t) = 0, (A.9)

where the current density is Continued

52 RESONANCE | January 2018

 GENERAL ARTICLE

Appendix 1. Continued.

J(r, t) = −(1/ω){(∂/∂r)V(r) + kB T (∂/∂r)}P(r, t). (A.10)

As we have elaborated upon earlier, we are interested in the approach to equilibrium, in the quasi-stationary
state, in which P(r, t) is a slowly varying function of time. Correspondingly, as (A.9) suggests, that is also
the state in which J(r, t) is almost independent of r, and can be represented by:

Jqs (t) = −(1/ω){(∂/∂r)V(r) + kB T (∂/∂r)}Pqs (r, t), (A.11)

wherein the right hand side containing the quasi-stationary probability Pqs and other r-dependent terms
must conspire in such a manner as to render Jqs (t) almost r-independent! The quasi-stationarity of Jqs (t)
is indeed the key idea of the transition state, i.e., the latter is a state of ‘no return’ [16]. Once the classical
trajectories reach the saddle point in Figure 1, there is no looking back – the reaction ‘slides down’ to
the reactant-state at a steady rate. Given this background to the form of quasi-stationary current we now
specialize to the ionization reaction as depicted in (1). Adapting to the language of the transition state
theory, the combination of (Na+ ) and (e− ) can be dubbed as the ‘reactant’, localized in ‘State 1’, while (Na)
is the ‘product’, localized in ‘State 2’ (see Figure 1).
State 1 and State 2 are separated by a large barrier in the space of V(r) versus r where the top of the barrier
is actually the transition state. In the more general setting of a molecular reaction, r represents a ‘reaction
coordinate’ which could be a bond direction. Also, V(r) versus r, in that case, is a free energy landscape.
Once the ‘activated complex’ of the (Na+ )-(e− ) pair reaches the transition state it proceeds to the product
state of neutral Na. The rate equation (A.2) can now be transcribed to the form:

(d/dt)p2 (t) = −(d/dt)p1 (t) = Jqs (t), (A.12)

where
p2 (t) = [Na] = [x(t)], (A.13)

and similarly,
p1 (t) = [Na+ ] · [e− ] = [x+ (t)] · [x− (t)]. (A.14)

We now carry out the following steps in order to arrive at the desired structure of the rate equation.
Step 1
Note that for a ‘large’ barrier height the quasi-stationary probability is expected to be ‘peaked’ around r1
and r2 , hence
 r1+dr1     r1+dr1   
V(r1 ) V(r)
p1 (t) = Pqs (r, t)dr ∼ Pqs (r1 , t) exp · dr exp − , (A.15)
r1−dr1 kB T r1−dr1 kB T

Continued

RESONANCE | January 2018 53

 GENERAL ARTICLE

Appendix 1. Continued.

dr1 being a small region around r1 . Here we have used the fact that near equilibrium,

Pqs (r, t)/[exp(−V(r)/kB T ] ∼ Pqs (r1 , t)/[exp(−V(r1 )/kB T ]. (A.16)

Interchanging 1 and 2 we obtain a similar expression for p2 (t).

Step 2
Using (A.15) and a similar one for p2 (t) we plan to express Jqs (t) in terms of p1,2 (t). To this end, we rewrite
Jqs (t) in (A.11) as    
∂ V(r)
Jqs (t) · [exp(V(r)/kB T ] = −D { exp Pqs (r, t)} , (A.17)
∂r kB T
and integrate both sides over r from r1 to r2 to obtain

Jqs (t) = −[1/(DI3 )]{[exp(V(r2 )/kB T ]Pqs (r2 , t) − [exp(V(r1 )/kB T ]Pqs (r1 , t)}, (A.18)

where  r2
I3 = exp[(V(r)/kB T ]dr. (A.19)
r1

Further, from (A.15),

Jqs (t) = −[1/(DI3 )]{p2 (t)/I2 − p1 (t)/I1 }, (A.20)

which was the desired goal of step 2. In (A.8),

 r1+dr1  
V(r)
I1 = exp − , (A.21)
r1−dr1 kB T

and a similar expression holds for I2 . We now have (A.12) in the coveted form of (A.1):

(d/dt)p2 (t) = [1/(DI3 )]{p1 (t)/I1 − p2 (t)/I2 }. (A.22)

Comparing with (A.18) we can identify,

γ1 (T ) = [1/(DI1 I3 )], γ2 (T ) = [1/(DI2 I3 )], γ1 (T )/γ2 (T ) = I2 /I1 . (A.23)

Continued

54 RESONANCE | January 2018

 GENERAL ARTICLE

Appendix 1. Continued

Recalling that I1 refers to the ‘reactant’ state, the integral over the Gibbs–Boltzmann measure in (A.21) is
just the corresponding partition function that yields the Boltzmann factor for the internal state of (Na+ ) ion
apart from the partition function of the free motion of the electron (e− ), whereas I2 is the corresponding
Boltzmann factor associated with the internal (excited) state of neutral (Na). Thus, we recover the Saha
equation (5) in the text! It is interesting to point out that though the Saha equation refers only to the ratio
of the two rates, as in (A.23), and hence the said equation does not involve the rate explicitly, the more
general kinetic formalism of Kramers allows one to estimate the individual rates as well, via the first two
expressions in (A.23). Indeed Kramers calculated I1 , I2 and I3 by the saddle-point approximation around
the transition state [9, 18]. Further, it is fair to stress that the transition state theory of Eyring and Kramers,
following the pioneering work of Wigner, is much richer in structure than what is needed to derive the
ionization equation. Indeed Wigner’s work, which is the earliest in the field, already had considered various
generalizations of the transition state theory, including corrections to the rate when the stipulation of ‘high
barrier-weak thermal energy’ breaks down as well as when quantum effects are important [16]. Also, the
basic assumption of the transition state that it is a point of no return allows the Kramers rate, or more
accurately its inverse, to be related to what is called the ‘mean first passage time’ [19]. It is not surprising
then that the ratio of ‘first passage times’ gets related to the ratio of partition functions, as is often invoked
in sophisticated analyses of the random walk problem [20].

Suggested Reading
[1] M N Saha, Ionization in the Solar Chromosphere, Phil. Mag. Ser., 6, 40 (238):
472 (1920), and On a Physical Theory of Stellar Spectra, Proc. Royal Society,
Vol.99, No.697, p.135, 1921.
[2] Partha Ghose, Private Communication.
[3] H Bradt, Astrophysical Processes, Cambridge University Press, London, 2008.
[4] Rabindranath Tagore, Visva-Parichay, A publication of Visva-Bharati, Sa-
tiniketan, 1937.
[5] For historical accounts, the reader may refer to some excellent reminiscences
in, M N Saha in Historical Perspective, edited by J Gupta, Thema, Calcutta,
1994, and in, From Astronomy to Astrophysics, in a special volume on the
occasion of the birth centenary celebration of M N Saha, edited by B Sinha
and M Bhattacharya, Saha Institute of Nuclear Physics, Calcutta (1995).
[6] G Venkataraman, Saha and His Formula, Universities Press, Hyderabad, 1995.
[7] M N Saha and B N Srivastava, A Treatise on Heat, The Indian Press Pvt. Ltd.,
Allahabad, 1931, Chapter XVIII.
[8] H Eyring, The Activated Complex in Chemical Reactions, Jour. Chem. Phys.,
Vol.3, No.2, p.107, 1935; for a comprehensive treatment of the Transition State
Theory, see also P W Atkins and J De Paula, Physical Chemistry, XII th Edition,
Oxford University Press, 2014.

RESONANCE | January 2018 55

 GENERAL ARTICLE

[9] H A Kramers, Brownian Motion in a Field of Force and the Diffusion Model of
Chemical Reactions, Physica, (Utrecht), Vol.7, p.284, 1940.
[10] M N Saha and N K Saha, Treatise on Modern Physics, Indian Press, Allahabad,
1934, Chapter XII.
[11] A more detailed treatment for varying concentrations can be found in, J K
Bhattacharjee and Dhruba Banerjee, Intermediate Statistical Mechanics A
Handbook, World Scientific Press, Singapore, 2017, Chap 4.11.
[12] D Chandler, Introduction to Modern Statistical Mechanics, Oxford University
Press, 1987, Chap 4.10.
[13] K Huang, Statistical Mechanics, Chap 7.6, John Wiley & Sons, Inc., New York,
1967.
[14] David Halliday, Robert Resnick and Jearl Walker, Fundamentals of Physics,
Extended 10th Edition, Wiley, 2014.
[15] B W Carroll and D A J Ostille, An Introduction to Astrophysics, Pearson New
International Edition, 2014.
[16] E Wigner, Z Phys. Chem., Vol.B19, 203, 1932; an overview of Wigner’s seminal
work is done by B C Garrett, Perspective on ‘The Transition Sate Method’,
Theo. Chem. Acc., 103, 200, Springer-Verlag, Berlin, 2000.
Address for Correspondence [17] N W Ashcroft and N D Mermin, Solid State Physics, Holt, Rinehart and Win-
Sushanta Dattagupta ston, N.Y. 1976.
Uniworld City [18] See, for instance, S Dattagupta, Diffusion Formalism and Applications, Taylor
“Heights” Tower 6 & Francis 2014.
Flat 002 [19] S Dattagupta and S R Shenoy, Decay of Metastable States – Kramers, First-
New Town (Action Area III) Passage Time and Variational Approaches, Lect. Notes in Physics, G S Agarwal
Rajarhat, Kolkata 700 156. and S Dattagupta, eds., Springer-Verlag, Berlin, 184, 61, 1984.
Email: sushantad@gmail.com [20] M Barma and R Ramaswamy, Jour. Stat. Phys., Vol.43, p.561, 1986.

56 RESONANCE | January 2018