Ch E 441 - Chemical Kinetics and Reaction Engineering

Residence Time Distributions

in Chemical Reactors
Residence Time Distributions
• The assumption of a “perfectly mixed” reactor
often falls short of reality.
• Residence time distributions are used to model
the imperfect mixing behavior of real reactors.
– Cumulative age, F(t)
– External age, E(t)
– Internal age, I(t)
• Gas-liquid CSTR (A(g) + B(l) C(l))
– Reaction occurs at gas-liquid interface
– Liquid phase is perfectly mixed
– Rate is proportional to bubble surface area
– Residence time of gas bubble in reactor
is proportional to bubble volume
• Larger bubble escape rapidly
• Smaller bubbles may remain in reactor until consumed
– Understanding of RTDs is necessary for analysis
• PBR
– Sections of the catalyst bed may offer less resistance
to flow, resulting in a preferred pathway through
the bed.
– Molecules flowing through the “channel” do
not spend as much time in the PBR as those
taking another path.
– Consequently, there is a distribution of
residence time for the PBR.
• CSTR
– Short-circuiting may occur (the direct movement
of material from inlet to outlet.
– Dead zones may exist (regions with a minimum
of mixing and thus virtually no reaction takes
place).
• Concepts that must be addressed in approaching
a solution to such problems:
– distribution of residence times occurs
– quality of mixing varies with position in reactor
– a model must used to describe the phenomenon
• Accounting for nonideality requires
– knowledge of macromixing (RTD)
– application of the RTD to a reactor (micromixing) to
predict reactor performance.
RTD Functions
• In any reactor, the RTD can affect performance
– Ideal Plug Flow and Batch Reactors
• Every atom leaving reactor is assumed to have resided in
the reactor for exactly the same duration. No axial
mixing.
– Ideal CSTR
• Some atoms leave almost immediately, others remain
almost forever. Many leave after spending a period of time
near the mean residence time. Perfect mixing.
• RTD is characteristic of mixing in a reactor.
• RTDs are not unique to reactor type.
RTD Functions
Different reactor types can have the same RTD.
Measurement of RTD
• RTD is measured experimentally by use of an
inert tracer injected into the reactor at t = 0.
Tracer concentration is measured at effluent as a
function of time.
• Tracer must be non-reactive and non-absorbing
on reactor walls/internals.
• Tracer is typically colored or radioactive to allow
detection and quantification.
• Common methods of injection are pulse and step
inputs.
Pulse Input RTD Measurement
• An amount of tracer No is suddenly (all at once)
injected into the feed of a reactor vessel with
flow at a steady state.
• Outlet concentration is measured as a function of
time.
feed effluent
reactor

injection detection
 pulse injection
pulse response

C C

- 0 + - 0 +
t t

feed effluent
reactor
injection detection
• Injection pulse in system of single-input and
single-output, where only flow (no dispersion)
carries tracer material across system
boundaries.
• The amount of tracer material N leaving the
reactor between t and t+ t for a volumetric
flowrate of is N C t t
where t is sufficiently small that the
concentration of tracer C(t) is essentially constant
over the time interval.
• Dividing by total amount of tracer injected, No
yields the fraction of material that has a
residence time between t and t+ t:

• where E(t) represents the residence-time

distribution function.
C t
N t Et t Et Ct
0 C t dt
No No
Step Input RTD Measurement
step injection
step response

C C

t t

feed effluent
reactor
Step Input RTD Measurement
injection detection
Step Input RTD Measurement
• In general, the output concentration from a
vessel is related to the input function by the
convolution integral (Levenspiel):
t
Cout t Cin t' E t' dt
t
0

where the inlet concentration takes the form of

either a perfect pulse input (Dirac delta
function), imperfect pulse injection, or a step
input.
Step Input RTD Measurement
Step Input RTD Measurement
• Considering a step input in tracer
concentration for a system of constant :
0 t 0
Co t
Co t 0 constant can be brought
outside the integral

Cout t t E
Cin 0 t' t' dt
Step Input RTD Measurement
t
Co E t' dt
0
 t
• Divide by Co Cout
E t' dt' Ft
Co step 0

– F(t) fraction of molecules that have spent a time

t or less in reactor (Cumulative age)

• Differentiate to obtain RTD function E(t)

d
Et C out Co
d
t
step
• Advantages
– Easier to carry out experimentally than pulse test
– Total amount tracer in feed need not be known
• Disadvantages
– Often difficult to maintain a constant tracer
concentration in feed.
– differentiation of data, often leads to large error.
– Requires large amount of tracer, which in some cases
can be expensive.
RTD Characteristics
• E(t) is sometimes called the exit-age distribution
function.
• If the age of an atom is regarded as the amount
of time it spends in the reactor, E(t) is the age
distribution of the effluent.
• E(t) is the most often used distribution function
for reactor analysis.
Integral Relationships
• Fraction of exit stream that has resided in the reactor
for a period of time shorter than a given value of t:
t
E t dt F t
0
• Fraction of exit stream that has resided in the reactor
for a period of time longer than a given value of t:

E t dt 1 F t
t
Integral Relationships
Mean Residence Time
• The nominal holding time, , is equal to the
mean residence time, tm.
• The mean value of the time is the first
moment of the RTD function, E(t).
tE t dt
tm 0 tE t dt
E t dt 0
0
• can be used to determine reactor volume
Other Moments of the RTD
st
• 1 moment – mean residence time

• 2nd moment – variance (extent of spread of

the RTD) 2
t- E t dt
2
tm
0

• 3rd moment – skewness (extent RTD is skewed

relative to thesmean)
3 2t - t 3 E t dt
m
1
3 0
Normalized RTD Function, E( )
• A normalized RTD is often used to allow
comparison of flow profiles inside reactors of
different sizes, where
t
E Et

for an ideal CSTR

1 t
Et e

E Et e
Internal-Age Distribution, I( )
• Fraction of material inside the reactor that has
been inside for a period of time between and
+
1 1
I E d
0
RTD in a Batch or PFR
• Simplest case
• Spike at t = (or = 1) of infinite height and zero
width with an area of one
x 0 x 0
Et t
x 0
gx
x dx 1 x dx g
RTD in a CSTR
• Effluent concentration is identical to that of
reactor contents.
• A material balance for t > 0 on inert tracer
injected as a pulse at t = 0

in - out acc
0 C V dC t
dt Ct C0 e
• Recall definition of E(t), and substitute:
t t
Ct C 0e e
Et
C t dt C 0e t
dt
0 0

0 C V dC t
dt Ct C0 e
Ideal Reactor Response to Pulse
Batch/PFR CSTR

E E

1
t
Laminar Flow RTD
• Velocity profile in a pipe (cylindrical
coordinates) is parabolic according to:
2 2
r 2 o r
U Umax 1 1
2 R
R R

• Time for passage of an element of fluid is

L R2L 1 1
tr
Ur 2 1 rR2 2 1 rR2
o
• The fraction of total fluid passing between r
and r+dr is d / 0: d U r 2 rdr
0
0

2
4 2 4t2
dt rdr R 2 rdr
R2 1 r R2

1
tr
2 1 rR2
• Combining
d U r 2 rdr
0
0

4 2
dt 2 4t 2
2 rdr rdr
R 1 rR 2 2
R

d L 2 rdr t
0 0
 2
L2 R2
dt 4t 2dt
2t
3 t 0
• The minimum time the fluid will spend in the
reactor is L L R2 V
t
Umax 2Uavg R2 2 2
0

• Therefore, the complete RTD function is

0 t 2 0 0.5
2
Et E 1
t 2 3 0.5
2t3 2
• The RTD appears graphically as

1 0 0.5
E 1
E 3 0.5
2

0.5
RTD of PFR and CSTR in series
• CSTR ( s) followed by PFR ( p)
– CSTR output will be delayed by a time of p

0 t p
p
t
Et e s

t p
s
RTD of PFR and CSTR in series
• PFR ( p) followed by CSTR ( s)
– PFR output will delayed the introduction of the pulse
to the CSTR by a time of p

0 t p
p
t
Et e s

t p
s
Regardless of the order, the RTD is the same. However, the
RTD is not a complete description of structure for a particular
reactor or system of reactors (see Example 13-4).