LABORATORY REPORT ON

RESIDENCE TIME AND NON-IDEAL FLOW

Sean Donovan (17769534); Michael Allan Ramos (19242176); Dennis Uluitoga (15238236)
Dr. Richard Browner (Unit Coordinator)

1. INTRODUCTION

In the preliminary design of chemical reactors, ideality is always assumed to simplify the design
calculations. However, in reality, non-ideal behaviors exist. Thus, the reactor efficiency observed
in real setting will always be lower compared to that of the efficiency derived from the design
equations of ideal reactors.

One important cause of non-ideality is due to the actual flow pattern inside the reactor - as this will
affect the time provided for the species to react and their corresponding contacting pattern. In an
ideal back-mixed reactor, it is assumed that the reacting mixture is perfectly mixed throughout the
reactor volume and the mean residence time is equal to that of the reactor’s space time. In an
actual back-mixed reactor, however, some of the introduced particles may exit the reactor
immediately due to short-circuiting and some may spend a long period of time inside the reactor
due to existence of stagnant regions (Figure 1).

Figure 1. Non-ideal flow patterns which may exist in

mixed-flow reactor (Levenspiel, 1999).

In order to account for these non-idealities, analysis of the reactor behavior with respect to the
quality of mixing and residence time distribution can be conducted. One common method is to
perform a stimulus-response experiment where an amount of tracer is instantaneously introduced
to the system, say in the reactor inlet or vessel, as a stimulated disturbance. Consequently, the
response of the reactor is determined by measuring the amount of tracer leaving the reactor with
respect to time. A residence time distribution (RTD) curve can therefore be obtained from the data,
which in turn provides a way to analyze the behavior of the reactor. Schematic of an RTD
measurement using a pulse input is shown in Figure 2.
 Figure 2. RTD measurement using a pulse input (Fogler, 2006).

Comparison of the outlet tracer concentration profiles in an ideal and actual reactor may propose
several probable causes of non-ideality. Levenspiel (1999) provides different cases of non-ideal
flow in back-mixed reactors determined from a pulse stimulus-response test (Figure 3).

Figure 3. Misbehaving mixed-flow reactors (Levenspiel, 1999).

In this experiment, a stimulus-response test using a pulse tracer input will be used to evaluate the
flow behavior of a certain CSTR and to identify possible causes of non-ideality in the set-up.
2. EXPERIMENTAL PROCEDURE

In this experiment, acetic acid was used as a tracer which was introduced to the reactor system as
a pulse input. The response of the system over time was determined thru pH measurement in the
reactor effluent. The experimental procedure was composed of two parts: (1) construction of pH-
concentration calibration curve, and (2) pulse stimulus-response test as detailed below.

2.1 Construction of pH-Concentration Calibration Curve for Acetic Acid

1. Using a stock solution of 5 M acetic acid, appropriate dilutions with the use of tap water
were made to prepare the following solution concentrations: 0.05 M, 0.005 M, 0.0005M.
2. Corresponding pH measurements were made for the solutions prepared in (1) and the tap
water.

2.2 Pulse Stimulus-Response Test

The pulse stimulus-response test is conducted using two different reactor set-ups: (1) use of a
single back-mixed reactor, and (2) use of two back-mixed reactors connected in series.

Single Reactor
1. The volume of the reactor was first determined by filling it fully with tap water of known
volume using a graduated cylinder.
2. The agitator and pump were started to give a volumetric flow rate of 200 ml/min. Exact
flow rate measurements were conducted by collecting an outlet volume over a recorded
time using a graduated cylinder.
3. Upon establishment of a stable volumetric flow rate, a 20-ml volume of 5 M acetic acid
was instantaneously introduced at the base of the reactor vessel thru syringe injection.
4. Collection of sufficient sample volumes (for pH measurement) from the reactor effluent
were made at several pre-determined times (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30 min).
5. Finally, the pH of the collected samples were measured and recorded.

Two Reactors in Series

1. The two reactors were treated to be equally-sized reactors. Therefore, the total volume was
calculated by doubling the determined reactor volume in the single reactor set-up.
2. The agitator and pump were started to give a volumetric flow rate of 400 ml/min. Exact
flow rate measurements were conducted by collecting an outlet volume over a recorded
time using a graduated cylinder.
3. Upon establishment of a stable volumetric flow rate, a 20-ml volume of 5 M acetic acid
was instantaneously introduced at the base of the first reactor vessel thru syringe injection.
4. Collection of sufficient sample volumes (for pH measurement) from the effluent of the
second reactor were made at several pre-determined times (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15,
20, 25, 30 min).
5. Finally, the pH of the collected samples were measured and recorded.
3. RESULTS AND DISCUSSION

3.1 Construction of pH-Concentration Calibration Curve for Acetic Acid

The results of pH measurement of tap water and different concentrations of acetic acid solution (0,
0.05, 0.005, 0.0005 M) is presented in Table 1 below.

Table 1. Results of pH measurements.

[HAc], [mol/L] pH Log [HAc]
0.05 3.02 -1.3010
0.005 4.01 -2.3010
0.0005 6.5 -3.3010
0 7.8 -

In order to construct a linearized pH-concentration calibration curve for acetic acid, graph of pH
versus log [HAc] was used (Figure 4). Two linear equations were fitted to represent the curve:
Equation (1) for pH > 4.01 and Equation (2) for pH < 4.01.

Figure 4. pH-concentration calibration curve for acetic acid.

3.2 Pulse Stimulus-Response Test

3.2.1 Single Continuous Stirred-Tank Reactor

§ Calculation of reactor’s space time:

The volume of the continuous stirred-tank reactor was found to be 1,800 L. Therefore, setting
the volumetric flow rate at 200 L/min, the space time for the reactor was calculated as follows:

𝑉 1,800 𝐿
𝑠𝑝𝑎𝑐𝑒 𝑡𝑖𝑚𝑒, 𝜏 = = = 9 𝑚𝑖𝑛
𝜐 200 𝐿/𝑚𝑖𝑛

§ Reactor’s response in a pulse stimulus test:

Upon achieving a steady-state flow of 200 liters of tap water per minute in the reactor, a pulse
disturbance was introduced to the system by injecting 20-ml of 5 M acetic acid solution. The
result of pH measurement of the reactor’s effluent at pre-determined times is summarized in
Table 2. Using this data, corresponding acetic acid concentrations ([HAc]) of each pH
readings were calculated using the developed equations from the pH-concentration calibration
curve in section 3.1.

Table 2. pH and acetic acid concentration in the reactor’s effluent.

Time, [min] pH [HAc], [mol/L]
0
1 2.85 0.0743
2 2.95 0.0588
3 3.04 0.0477
4 3.12 0.0396
5 3.19 0.0337
6 3.26 0.0286
7 3.33 0.0243
8 3.39 0.0211
9 3.44 0.0188
10 3.50 0.0164
15 3.75 0.0092
20 3.98 0.0054
25 4.17 0.0043
30 4.52 0.0031

3.2.2 Two Continuous Stirred-Tank Reactors in Series

§ Calculation of the total space time:

The total volume of the reactors in series is 3,600 L, which was calculated by doubling the
determined volume in the single reactor test. In order to maintain the same resulting space time of
9 min, the volumetric flow rate in the series was set at 400 L/min.

𝑉 3,600 𝐿
𝑠𝑝𝑎𝑐𝑒 𝑡𝑖𝑚𝑒, 𝜏 = = = 9 𝑚𝑖𝑛
𝜐 400 𝐿/𝑚𝑖𝑛

§ Reactors series’ response in a pulse stimulus test:

Upon achieving a steady-state flow of 400 liters of tap water per minute in the reactors series, a
pulse disturbance was introduced to the system by injecting 20-ml of 5 M acetic acid solution in
the vessel of the first reactor. The result of pH measurement of the second reactor’s effluent at pre-
determined times is summarized in Table 3. Using this data, corresponding acetic acid
concentrations ([HAc]) of each pH readings were calculated using the developed equations from
the pH-concentration calibration curve in section 3.1.

Table 3. pH and acetic acid concentration in the reactor’s effluent.

Time, [min] pH [HAc], [mol/L]
0
1 4.01 0.0050
2 3.71 0.0100
3 3.62 0.0124
4 3.59 0.0133
5 3.58 0.0136
6 3.59 0.0133
7 3.60 0.0130
8 3.63 0.0121
9 3.67 0.0110
10 3.70 0.0103
15 3.94 0.0059
20 4.31 0.0038
25 4.84 0.0023
30 6.03 0.0008
3.3 Evaluation of Reactor Efficiency Based on Non-ideality of Flow

3.3.1 Mean Residence Time in Ideal Continuous Stirred-Tank Reactor

Ideally, the mean residence time, 𝑡̅, for back-mixed reactors is taken as equivalent to its space time,
𝜏. However, occurrence of non-ideal flow patterns may cause a deviation of the observed or actual
mean residence time.
A normalized exit age distribution curve representing the time it takes for a certain fraction of
species to exit a reactor upon its addition can be constructed by plotting of CA/CA0 against the
reduced time given by 𝑡/𝑡̅. The relation between these parameters are given by the equations
below.

For a single continuous stirred-tank reactor:

𝐶< @
>? A
= 𝑒 @̅ = 𝑒 (?@C) 𝑬𝑸𝑼𝑨𝑻𝑰𝑶𝑵 (𝟑)
𝐶<=

For multiple continuous stirred-tank reactors in series:

𝐶< 𝑛N (N?R) (?N@C )

= 𝑡Q 𝑒 𝑬𝑸𝑼𝑨𝑻𝑰𝑶𝑵 (𝟒)
𝐶<= (𝑛 − 1)!

where n is the number of reactors connected in series, CA is the concentration of species A in the
exit stream at any time, t, CA0 is the initial concentration of species A in the reactor, 𝑡̅ is the mean
residence time, and tr is the reduced time.

By setting arbitrary values for the reduced time, an exit distribution curve can be obtained by using
equations (3) and (4). Figures 5 to 7 show the ideal profile of the relative concentration of A as a
function of time in a single CSTR, two CSTRs, and multiple CSTRs connected in series,
respectively.
 Figure 5. Relative concentration of A as a function of time in a single CSTR.

Figure 6. Relative concentration of A as a function of time in two CSTRs connected in series.

 Figure 7. Relative concentration of A as a function of time in
multiple CSTRs connected in series; (n = no. of CSTR in series).

To compare the capability of the different reactors set-ups presented above in retaining the species
A within the reactor vessel, the fractions (of the total species A initially introduced to the reactor)
exited and within the vessel were evaluated at time equivalent to the mean residence time of the
ideal reactor set-ups. These fractions can be calculated by setting 𝑡 = 𝑡̅ or 𝑡Q = 1 in equations (3)
or (4) (Table 4). Alternatively, it can be determined graphically as shown in Figure 7.

Table 4. Efficiency of different reactors set-up in holding species A.

No. of CSTRs in series Fraction of A remaining in Fraction of A exited the
the tank at 𝒕 = 𝒕̅ tank at 𝒕 = 𝒕̅
1 0.3679* 0.6321*
2 0.5413 0.4587
5 0.8773 0.1227
10 1.2511 -0.2511
*Calculated using equation (3); otherwise, using equation (4)
 Figure 7. Graphical determination of fraction of A remaining in the tank at
𝑡 = 𝑡̅ for different reactor set-ups; (n = no. of CSTR in series).

3.3.2 Comparison of the Single Reactor with Ideal Reactor

An exit age distribution curve can be constructed for the actual single reactor used in the
experiment thru the manipulation of data obtained from the pulse stimulus-response test (Figure
8). The values of 𝐶< /𝐶<= can be calculated with respect to time from the pH measurement data
by using the calibration curve, or equations (1) and (2), and equation (3). Results of calculation is
summarized in Table 5. The initial concentration of acetic acid in the reactor, [HAc]0, was
calculated as follows:
20 𝑚𝑜𝑙 𝐻𝐴𝑐
𝑚𝑜𝑙𝑒𝑠 𝑜𝑓 𝐻𝐴𝑐 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 >1000 𝐿 𝐻𝐴𝑐 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛A >5 𝐿 𝐻𝐴𝑐 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛A
[𝐻𝐴𝑐]= = =
𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑒𝑎𝑐𝑡𝑜𝑟 1800 𝐿

= 𝟎. 𝟎𝟓𝟓𝟔 𝑴

Table 5. Relative concentration of acetic acid in reactor effluent as function of the reduced time.
Time pH Reduced Time, tr [HAc] [HAc]/[HAc]0
0 0.0000 0 0.0000
1 2.85 0.1111 0.0743 1.3365
2 2.95 0.2222 0.0588 1.0592
3 3.04 0.3333 0.0477 0.8591
4 3.12 0.4444 0.0396 0.7133
5 3.19 0.5556 0.0337 0.6061
 6 3.26 0.6667 0.0286 0.5150
7 3.33 0.7778 0.0243 0.4377
8 3.39 0.8889 0.0211 0.3806
9 3.44 1.0000 0.0188 0.3389
10 3.5 1.1111 0.0164 0.2947
15 3.75 1.6667 0.0092 0.1648
20 3.98 2.2222 0.0054 0.0965
25 4.17 2.7778 0.0043 0.0776
30 4.52 3.3333 0.0031 0.0562

Figure 8. Comparison of exit distribution curve of the actual reactor and ideal reactor.

From Figure 8, it can be seen that the observed mean reduced time for the reactor is actually
about 0.91 of the ideal one. From this, the observed mean residence time or the time it takes for
63.21% of the acetic acid to exit the reactor can be calculated as follows:

𝑡 = h𝑡Q̅ ijk l( 𝑡 ̅ )(0.91)(9 min) = 8.19 𝑚𝑖𝑛

The reduction from the expected residence time of 9 minutes may be due to presence of stagnant
fluid as suggested by Figure 3 taken from Levenspiel (1999). Occurrence of stagnant regions
within the reactor volume may be due to insufficient mixing power and/or improper positioning
of agitator (i.e. agitator is too high with respect to the base of reactor vessel leading to formation
of stagnant regions at the bottom corners of the reactor). Such stagnation will result to existence
of ineffective volumes within the reactor vessel thus lowering the effective volume of the reactor
to hold the reacting mixture. Hence, the reduction in residence time.

Aside from these, it can be noticed that a sharp early peak of acetic acid concentration (with
magnitude of about 1.335[HAc]0) was recorded as early as a reduced time of 0.20. Such surge of
acetic acid concentration at the reactor effluent may suggest that a short-circuiting of tracer to the
reactor exit stream may have occurred. Higher apparent [HAc]0, compared to that of the computed
one, due to reduction in effective volume caused by existence of stagnant regions may have also
contributed to this high peak of acetic acid concentration in the effluent.

3.3.2 Comparison of the Two-Reactors in Series with Ideal Reactors

The same manipulation of experimental data, as done in the single reactor, can be made in order
to construct an exit age distribution curve (Figure 9) for this two-reactor set-up. However, equation
(4) with n=2, instead of equation (3), was used for the calculation of CA/CA0. Results of such
manipulation is summarized in Table 6. Also, the total reactor volume for the series will be twice
that of the single reactor and hence [HAc]0 was calculated as:

20 𝑚𝑜𝑙 𝐻𝐴𝑐
𝑚𝑜𝑙𝑒𝑠 𝑜𝑓 𝐻𝐴𝑐 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 >1000 𝐿 𝐻𝐴𝑐 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛A >5 𝐿 𝐻𝐴𝑐 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛A
[𝐻𝐴𝑐]= = =
𝑡𝑜𝑡𝑎𝑙 𝑣𝑜𝑙𝑢𝑚𝑒 (1800 𝐿)(2)

= 𝟎. 𝟎𝟐𝟕𝟖 𝑴

Table 6. Relative concentration of acetic acid in the effluent of the

second reactor as function of the reduced time.
Time pH Reduced Time, tr [HAc] [HAc]/[HAc]0
0 0.0000 0.0000
1 4.01 0.1111 0.0050 0.1800
2 3.71 0.2222 0.0100 0.3617
3 3.62 0.3333 0.0124 0.4459
4 3.59 0.4444 0.0133 0.4781
5 3.58 0.5556 0.0136 0.4894
6 3.59 0.6667 0.0133 0.4781
7 3.60 0.7778 0.0130 0.4671
8 3.63 0.8889 0.0121 0.4356
9 3.67 1.0000 0.0110 0.3969
10 3.70 1.1111 0.0103 0.3702
15 3.94 1.6667 0.0059 0.2118
20 4.31 2.2222 0.0038 0.1364
25 4.84 2.7778 0.0023 0.0835
30 6.03 3.3333 0.0008 0.0278
 Figure 9. Comparison of exit distribution curve of the actual and ideal two reactors in series.

Although the 𝑡Q̅ ijk cannot be evaluated in the same manner as done in the single-reactor data due
to low peak of the actual curve, the fraction of species A remaining in the tank at one mean
residence time for the actual (~40%) and ideal (54.13%) case can be compared. From these values,
the effect of non-ideality of flow in the actual set-up that lowers the efficiency of the reactor can
be realized. One possible reason for the occurrence of low peak is that the apparent [HAc]0 may
be lower due to existence of unaccounted volume of the connecting pipe within the two reactors
in series (incorrect volume) – which may have also caused the late mean if for example at two
mean residence time as the reference. Also, a short lag in the concentration profile was observed
for the actual set-up as compared to the ideal one during the start of the sampling. This may be
caused by the connecting pipes and outlet pipe which may have acted as a plug-flow reactor
connected within the reactor series. Nonetheless, at one mean residence time, employing two
reactors in series still gave a higher fraction of species A retained within the system compared to
that of a single reactor.
4. CONCLUSIONS

• The efficiency of reactors in holding or retaining a certain fraction of species within the
vessel increases with increasing number of reactors in series.
• Non-ideality of flow patterns will decrease the efficiency of the reactor due to decrease in
mean residence time. These non-idealities in CSTRs may be caused by short-circuiting of
fluids and/or existence of stagnant region within the reactor volume due to insufficient
mixing caused by inadequate mixing power and/or improper positioning of the agitator.

5. REFERENCES

Fogler, H. S. (2006). Elements of chemical reaction engineering / H. Scott Fogler (4th ed.. ed.).
Upper Saddle River, NJ: Upper Saddle River, NJ : Prentice Hall PTR.

Levenspiel, O. (1999). Chemical reaction engineering / Octave Levenspiel (3rd ed.. ed.). New
York: New York : Wiley.