Lecture-14: Fundamental Equations for Reactor Engineering

The basis of most of reactor modeling is a set of fundamental ‘balance’ equations. These
equations are mathematical statements of physical laws that require conservation of mass,
energy, momentum, and quantities of particular chemical species. In general, these can be
derived by selecting a control volume and writing a balance on the quantity under consideration
(mass, energy, etc.) over that control volume. The control volume could be an entire reactor, a
catalyst pellet, or an infinitesimally small element of volume within the reactor. This balance
will have the form:

or, more simply, {accumulation = in – out + production}

The separation between transport in and transport out is somewhat artificial, so it may often
make sense to consider transport as a single term:

When we are considering an infinitesimally small control volume, it often makes sense to split
the transport rate into two or more mathematically separate terms. Usually we split it into a
‘convection’ term – transport due to the overall (mass-average) motion of the fluid – and a
‘diffusive term’ – motion of the particular substance under consideration (energy, momentum, or
a chemical species) relative to the overall motion. The diffusive term is driven by gradients of
the substance (and possibly also by gradients of other quantities). Then the general balance is

We will show later that application of this balance to an infinitesimally small control volume
allows us to derive a general set of balance equations that can be integrated and/or simplified to
describe many situations, including the prototypical reactor types (batch, plug flow, CSTR).
First, however, we will apply the balance to some simple macroscopic situations.
The Ideal Batch Reactor:

Consider first an insulated, well-mixed batch reactor of volume V, with some chemical reactions
N
(in general written as   ij Aj  0, i  1,..., M ) going on inside it. First consider a balance on the
j 1

total mass inside the reactor, for which

Net flow rate of mass into reactor = 0

Rate of production of mass within the reactor = 0
d  V 
Rate of accumulation of mass within the reactor =
dt

Where ρ is the (uniform) fluid density and V is the reactor volume, so that ρV is the total mass
inside the reactor. From this trivial example, we then get the balance equation

which simply says that the total mass inside the reactor is constant.

Likewise, we can write a balance for the number of moles of the kth chemical species in the
reactor.
Net flow rate of Ak into reactor = 0
M
Rate of production of Ak within the reactor = V   ik ri
i 1

d VCk  dCk
Rate of accumulation of Ak within the reactor = V (for constant V)
dt dt

This gives us the familiar set of batch reactor rate equations

Something we have not done so far is to also consider the energy balance for the batch reactor.
Applying the balance equation for total enthalpy inside the insulated reactor gives

Net flow rate of enthalpy into reactor = 0 (remember, we said it was well-insulated)
Rate of production of enthalpy within the reactor = 0

Rate of accumulation of enthalpy within the reactor =

d VHˆ  
dt

This gives us another apparently trivial equation, this time for the specific enthalpy ( Ĥ )
 
d VHˆ   0 , from which we see that VHˆ  a constant
dt

so the specific enthalpy ( Ĥ , enthalpy per unit mass) is constant inside the reactor.

However, this is not as trivial as it first appears, since what we are usually interested in is the
temperature rather than the enthalpy. Thus, we could try to write the specific enthalpy in terms
of the temperature. We would quickly see that the relationship between enthalpy and
temperature depends on the composition in the reactor (different species have different
enthalpies, defined relative to some reference species and conditions).

The total enthalpy per unit volume of an ideal mixture of N species can be written as

where H k , and Ck are the partial molar enthalpy and molar concentration of species k.

If we assume zero heat of mixing (as for an ideal solution) and a constant specific heat for each
of the species, this can be written in terms of the temperature as

where Hfk, and Cpk are the molar heat of formation and molar heat capacity of species k, T is the
temperature, and Tref is the reference temperature at which the heats of formation are defined.
The above would be considered drastic simplifications in many analyses of phase equilibria and
other situations where small differences in enthalpy can be of great importance. In general,
however, for reacting systems and particularly gas phase reacting systems, the enthalpy changes
due to reaction are much larger than enthalpies of mixing, errors in enthalpy due to assumption
of a constant specific heat, etc.

The sum of the terms involving the heat capacities gives the specific heat of the whole mixture,
which we’ll call Cˆ p . Taking it outside the sum gives

We can take the time derivative to put this into differential form, and we will further assume that
the mixture specific heat ( Cˆ p ) is independent of both temperature and composition (another big
assumption!), so it is constant in time. Then
 dCk
where the expression for derived above has been used. Reversing the order of the
dt
summations gives

Finally, the sum of the heats of formation times the stoichiometric coefficients is recognized as
the definition of the heat of reaction, so

From the enthalpy balance, we know that this is equal to zero, so we finally have an equation for
the temperature

which simply says that the temperature in the reactor will rise or fall due to heat released or taken
up by chemical reaction.

The above derivation was for an adiabatic (well-insulated) reactor. If there were some total net
heat flux QH into the reactor, then the energy balance would have been

d VHˆ  QH
dt
and the equation for the reactor temperature would have been

If this heat flux was due to heat transfer across an area A from a thermal reservoir at temperature
Tr with heat transfer coefficient hf then this would be

and so on.
The Ideal Continuous Stirred Tank Reactor (CSTR):

If, instead of a batch reactor, we had a constant volume continuous stirred tank reactor (CSTR)
with volume V, we could write the balances in the same way.
Mass flow rate into reactor = ρoQo
Mass flow rate out of reactor = ρQ
Rate of production of mass within the reactor = 0
d  V 
Rate of accumulation of mass within the reactor =
dt

Where ρo and Qo are the density and volumetric flow rate of the feed stream coming into the
reactor, while ρ is the density of the mixture within the reactor (which is equal to the density of
the stream coming out) and Q is the volumetric flow rate out of the reactor. So, the overall mass
balance for the reactor is

If the reactor volume is fixed, this becomes

and, if we can also assume that the density of the fluid inside the reactor is equal to the feed
density and is constant in time, then this is just Q = Qo.

A mole balance on the kth species is

Flow rate of Ak into reactor = Qo Cko
Flow rate of Ak out of reactor = QCk
M
Rate of production of Ak within the reactor = V   ik ri
i 1

d VCk  dCk
Rate of accumulation of Ak within the reactor = V (for constant V)
dt dt

So we have

If the density in the reactor is the same as that of the feed, then we can define a residence time as
τ = V/Qo = V/Q, and the equation becomes
If the CSTR is non-isothermal and adiabatic, then the enthalpy balance is

Flow rate of enthalpy into reactor =

Flow rate of enthalpy out of reactor =
Rate of production of enthalpy within the reactor = 0

Rate of accumulation of enthalpy within the reactor =


d VHˆ 
dt
There is no heat flow via heat transfer through the reactor walls (that’s what we mean by
adiabatic), but of course the flows coming in and going out have a certain enthalpy content, and
this is reflected in the inflow and outflow terms given above. We can re-write the accumulation
term as we did for the batch reactor so that temperature is the dependent variable rather than
enthalpy, and we can do essentially the same thing for the two flow terms. This makes these
terms

Putting this all together gives

Or, assuming constant volume, density, heat capacity, etc.

dCk
Substituting in the expression for derived above, this becomes
dt
 N

 C
k 1
ko  Ck  H fk
The term appears on both sides of the equation and therefore cancels out. We

N M M
saw above that  H fk   ik ri   H i ri , and substituting this finally gives us
k 1 i 1 i 1

If the reactor were not adiabatic, but instead had a total heat flux QH into it, that flux would just
be added to the equation as

Applying the Balances to an Infinitesimal Control Volume:

Having regained a certain comfort level with writing and manipulating these balance equations,
we will now apply them to an infinitesimally small control volume (with volume dV).

First, let’s look at the term for net inflow. If, at some point in space, we have a flux of some
quantity X that is given by NX then the net outflow from an infinitesimally small volume centered
at that point is given by the divergence of that flux (d(outflow)=    N X  dV ). This is a
differential form of Gauss' divergence theorem that we all learned in calculus class. Note that
the flux is a vector quantity – it has both magnitude and direction. It could be written in terms of
its three Cartesian coordinates and unit vectors along the three axes
as N X  N X x x  N X y y  N X z z . The divergence of N X in Cartesian coordinates is then given by
NXx NXy N Xk
  NX    .
dx dy dk

Definitions of the divergence operator in other coordinate systems are given in the back of Bird,
Stewart, and Lightfoot, and in many other places.

We will first derive a balance equation for total mass. The local flux of mass is simply the
density times the velocity. So, the terms in the balance are

Net flow rate of mass into the control volume =       v   dV

Rate of production of mass within the control volume = 0
d    dV  
Rate of accumulation of mass within the control volume =
dt

The volume of the control volume is constant (because we select it to be constant) so when we
write the balance equation, we can factor it out of the derivative and it cancels out. Then we
have

This is the familiar continuity equation for total mass, as we would find it in any fluid mechanics
or transport phenomena book. It is a differential form of the statement that mass is conserved
(neither created nor destroyed, as implied by setting the production rate = 0).

In the same way, we can write a balance equation for the number of moles of a particular
chemical species in the control volume. For the moment, we will just denote the local molar flux
of species k by NCk. This flux has units of moles per area per time and includes both convective
and diffusive transport of the chemical species.
   dV
Net flow rate of species k into the control volume =    N Ck
M
Rate of production of species k within the control volume = dV   ik ri
i 1

d  Ck  dV  
Rate of accumulation of species k within the control volume =
dt

Again, the volume of the control volume is constant, and we can factor it out and cancel it to get

Now, all that remains is to evaluate the molar flux NCk in terms of more convenient variables.
Most commonly, it is split into convective and diffusive fluxes, where the convective flux is
defined as transport due to the overall, mass-average velocity of the fluid, and the diffusive flux
describes any motion of a particular species relative to the overall mass average velocity.

The convective flux is simply the mass average velocity multiplied by the concentration

Convective flux = vCk

th
The diffusive flux can be defined in a number of ways. If we denote the velocity of the k
species by Uk, its velocity relative to the mass average velocity of the fluid by Vk and the
diffusive flux by Jk, then we have the relationships

Diffusive flux = Jk = VkCk = (Uk – v)Ck

If species k diffuses in the mixture only due to concentration gradients, and it obeys Fick’s law,
then its diffusive flux is given by
 J k   Dk Ck

where Dk is the diffusion coefficient of species k in the overall mixture. A problem with this
expression is that, unless the diffusion coefficients of the species are concentration dependent
and carefully constructed to be self-consistent, it will violate the overall continuity equation.
That is, summing all of the species velocities computed using this expression, weighted by the
species mass fractions, would not give the mass average velocity. For gases, a more rigorous
approach would show that the flux of a particular species depends on gradients of all of the other
species. This can be expressed in terms of the Stefan-Maxwell equations, which are given
Froment and Bischoff, as well as in Bird, Stewart, and Lightfoot and many other places. We
won’t consider them here. There may also be diffusive fluxes due to temperature gradients
(Stefan diffusion or thermal diffusion) or other gradients in the fluid. Both these
multicomponent diffusion effects and thermal diffusion effects are important in some cases (like
CVD reactor modeling or flame modeling). However, we won’t consider them in detail here.
For most classical reactor modeling problems, an approach that uses an effective Fickian
diffusion coefficient Dk is satisfactory. If we had laminar flow through an empty (no catalyst)
reactor, then Dk would be related to true molecular diffusion coefficients. In general, for
turbulent flow or flow through a bed of catalyst particles, Dk will be a dispersion coefficient that
includes the effects of mixing due to the turbulent eddies or complex flow patterns around the
catalyst particles. In many cases, this dispersion coefficient will be much larger than a typical
molecular diffusion coefficient, and will be approximately equal for all species in the mixture,
since they all experience the same turbulent eddies and flow patterns.

In any case, after splitting the species flux into diffusive and convective components, the balance
equation for species k becomes
M
dCk
    vCk     J k    ik ri
dt i 1

If we use the Fick’s law expression for the diffusive flux, this is
M
dCk
    vCk      Dk Ck     ik ri
dt i 1

If we can assume that the density of the mixture is constant, then the convective term can be
simplified. From the overall mass balance (continuity) equation

If the density is constant (in both time and space), then this simply becomes
 v  0

Using an identity from vector calculus, we can rewrite the convective term in the species balance
equation as
   vCk   v Ck     v  Ck
So, for a constant density system we have

If the diffusion or dispersion coefficient is also constant then

This equation is a common starting point for many mathematical models of chemical reactors.
From the derivation presented here, we can see some of the assumptions that are made in using
it. We have assumed constant density, Fickian diffusion with constant diffusion coefficients, etc.

Now, we will write an enthalpy balance on the differential control volume. The terms are
Net flow rate of enthalpy into the control volume =     N Ĥ  dV
Rate of production of enthalpy within the control volume = 0

Rate of accumulation of enthalpy within the control volume =


d  Hˆ  dV  
dt

As before, this gives us

 
d  Hˆ
   N Hˆ
dt

We once again will separate the flux into diffusive and convective components. The convective
flux is
Convective flux =

We will call the diffusive flux q – the usual symbol for a heat flux. We can further divide it into
an ordinary heat flux due to conduction according to Fourier’s law of conduction, and a
component due to the diffusion of the chemical species (which each have enthalpies different
from the average enthalpy). So, we write the diffusive heat flux as
N
q   T   J k H k
k 1

where λ is the thermal conductivity. This will be an effective thermal conductivity for turbulent
flow or flow through a bed of catalyst particles, rather than the actual thermal conductivity of the
mixture, in the same way that we used an effective diffusion coefficient (or dispersion
coefficient) for these situations.

Substituting these fluxes into the enthalpy balance equation gives

We will now rewrite the enthalpy in terms of the temperature as we did for the batch and stirred
tank reactors.

This makes the overall balance equation

We can now write the molar enthalpy of the species in terms of their heats of formation and heat
capacity

And we can write the diffusive flux of the species in the Fick’s law form used above, to get

dCk
Substituting in the expressions for , this becomes
dt

The terms
N N

 H fk    Dk Ck  and
k 1
 H     vC  
k 1
fk k

appear on both sides of the equation and can be cancelled. As we saw before, the sum

can be rewritten in terms of the heats of reaction as

 .

Making these changes leaves us with

If we can assume constant density, thermal conductivity, specific heat, and diffusion coefficients,
and apply the product rule for differentiation, this becomes

For constant density, the continuity equation tells us that   v  0

This leaves us with

In this derivation, we have neglected several potential contributions to the enthalpy balance such
as heat fluxes due to radiation (important in high temperature systems), heat generation by
viscous dissipation (important in high shear flows of high viscosity fluids), and cooling due to
expansion (important in high speed gas flows).

A slight variant on the above equation, resulting from slightly different assumptions, is presented
on page 301 of Froment and Bischoff.
Simplification of the 'General' Differential Balances to get Ideal Reactor Equations:

Before leaving this section, we will see how these ‘general’ balance equations in differential
form can be simplified to give the batch, CSTR, and plug flow reactor equations.

In a well-mixed batch reactor, no spatial dimensions need to be considered – all of the spatial
gradients are zero. Simply eliminating all of the terms that contain gradients or divergences from
the balance equations leaves us with

which are, of course, the same batch reactor equations that we derived before.

It is similarly straightforward to simplify the ‘general’ energy and species mole balances to get
the equations describing the steady-state plug-flow tubular reactor. The general equations (for
constant properties) were

If we assume that there are spatial variations in only one direction, say x, the direction along the
reactor axis, then these equations assume a 1-dimensional form.

These are the equations for an unsteady-state plug-flow reactor with axial diffusion. The first
equation (overall continuity equation) implies that the axial velocity is constant for constant
density one-dimensional flow.

If the reactor operates in steady-state, then the time derivatives are zero, and we have
These are the equations for a steady-state plug-flow tubular reactor with axial diffusion.
Finally, if the axial velocity is sufficiently high that transport by axial diffusion and conduction
are negligible compared to convective transport, then these equations become

These are the standard equations for an ideal plug flow tubular reactor. They are identical to the
equations for the batch reactor, except that the time derivative has been replaced by the product
of the velocity and a spatial derivative.

Obtaining the equations for a CSTR from the general equations is somewhat less direct. Starting
from the general equations in differential form, we would integrate over the reactor volume.
Everywhere within the reactor, the gradients are zero, so that all of the terms containing
gradients disappear. However, at the reactor inlet and outlet, we must apply boundary conditions
that specify the transfer of heat, mass, and chemical species in and out of the reactor. These
inflows and outflows are exactly those that we used when we derived the CSTR equations from
macroscopic balances on the reactor. So, formally, this would allow us to derive the CSTR
equations from the ‘general’ balance equations. However, it seems both more intuitive and
sensible to derive these equations from a macroscopic balance as we did near the beginning of
this section.

After completing your study of these lecture notes you should be able to:

(1) Understand the derivations presented

(2) Derive the appropriate balance equations for an arbitrary reactor configuration using similar
methods