Safety

Characterize
Reactive Chemicals
with Calorimetry
Daniel A. Crowl In designing and operating chemical manufacturing
Michigan Technological Univ.
Jason M. Keith
processes, the reactivity of chemicals must be
Mississippi State Univ. identified to avoid hazards. Calorimetry
experiments combined with analytical methods
provide a straightforward technique to identify
these reactions and their corresponding hazards.

T
he reaction of raw materials is central to many prod- discusses the apparatus used to obtain the necessary data,
ucts manufactured by chemical companies. Many analysis methods to determine key parameters from the
of these reactions are accompanied by potential data, and how to use these parameters to answer the initial
hazards. For example, some reactions generate heat, which, questions. While the analyses presented here are typically
depending on the rate of reaction, may be released over a performed by company experts or outside consultants,
very short period of time; some reactions produce hazard- every chemical engineer should have a basic understand-
ous products or byproducts. Furthermore, if the energy ing of the method, which, as the article explains, provides
or the mass of the reaction is not controlled properly, the valuable information for process design and operation.
reactor temperature and/or pressure may increase rapidly,
resulting in an accident. Ask and answer these questions
The U.S. Chemical Safety and Hazard Investiga- The questions that must be answered to properly
tion Board (CSB) studied 167 serious incidents in the characterize reactive chemistry (Table 1) deal with the
U.S. between January 1980 and June 2001 that involved intended reaction (Questions 1–5), as well as any side (or
uncontrolled chemical reactivity (1). Forty-eight of those other) reactions, such as an exothermic decomposition that
incidents resulted in a total of 108 fatalities — an average may occur at an increased temperature (Questions 6–8).
of five fatalities per year. The CSB concluded that reac- Table 1 also identifies how the experimental data are used
tive incidents are a significant chemical safety problem to answer these questions. References 2 and 3 provide
often caused by inadequate recognition and evaluation of additional information on the characterization of reactive
reactive hazards. Thus, to improve reactive hazard man- chemicals.
agement, the CSB recommended that factors such as the The most critical information about the reaction is the
rate and quantity of heat and gas generated be defined and rate of heat or energy generation (Table 1, Question 3),
examined within every process hazard analysis (PHA). which relates to the primary means of controlling the reac-
The first step toward preventing accidents related to tion. For example, heat-removal equipment must be sized
chemical reactions is to characterize the reactive chemistry, properly to remove the heat as fast as it can be generated to
which involves answering numerous questions about the avoid an out-of-control temperature increase and a run-
reaction. This article identifies the relevant questions, and away reaction.

26 www.aiche.org/cep July 2013 CEP Copyright © 2013 American Institute of Chemical Engineers (AIChE)
The calorimeter in a single run, in industrial practice, additional runs are
The data needed to answer the first five questions in necessary to verify the data, as well as to evaluate stoi-
Table 1 can be obtained from a single experimental run in chiometric and autocatalytic effects. Additional runs may
a specially designed adiabatic calorimeter. This experiment be conducted to identify and characterize side reactions or
also provides insight into the overall reaction order and the reactions at increased temperatures.
Arrhenius rate parameters. While the data can be acquired The experimental calorimeter must have several essential

Table 1. Answer the questions in the left column to characterize chemical reactivity.
Calorimeter experiments provide the data (right column) needed to answer these questions.
Question Calorimeter Method Used to Answer Question

1. At what temperature is the self-heating rate detected and Determine TO using the method shown in Figure 4.
sustained?

2. What is the maximum temperature due to adiabatic self- Determine the maximum temperature from a plot of temperature
heating of the reactants? vs. time data (Figure 3).

3. What is the maximum self-heating rate? At what time does this Determine the maximum self-heating rate and the time at which
occur? it occurs from the numerical derivative of temperature vs. time
data. Use Eq. 8 and the derived kinetic parameters to confirm
that the model predicts similar results.

4a. What is the maximum pressure during the reaction? Determine the maximum pressure from a plot of pressure vs.
time data.
Return the sealed calorimeter to its initial temperature — if the
4b. Is this pressure due to the vapor pressure of the liquid, or to
pressure is the same as the starting pressure, then it is due to
the generation of gaseous, noncondensable reaction products?
the vapor pressure of the liquids; a pressure higher than the
starting pressure means gaseous, noncondensable reaction
products were produced.

5. What is the maximum pressure rate? At what time does this Determine the maximum pressure rate and the time at which it
occur? occurs from the numerical derivative of pressure vs. time data.

6. Do any other side reactions occur, particularly at tempera- Operate the calorimeter at a temperature well beyond the
tures higher than the normal reaction temperature? If so, answer primary reaction temperature. If a side reaction is detected by
Questions 1–5 for each additional reaction. the calorimeter (by new exotherms), collect complete calorimeter
data and follow the full characterization method for each side
reaction.

7. Can the energy generated by the desired chemistry heat the Assume the reaction energy is used solely to heat the reaction
reaction mass under adiabatic conditions to a temperature at liquid. The total energy generated by the reaction (in units of
which another reaction occurs? energy, e.g., J) is equal to the limiting reactant mass multiplied by
the heat of reaction. Divide this by the product of the total liquid
mass and the heat capacity of the liquid (in units of energy/mass-
temperature, e.g., J/kg-K) to obtain the temperature increase.
Add this to the initial temperature to obtain the final temperature.
If the final temperature is near or above the onset temperature for
any side reactions, then the side reactions may be possible.

8a. Can the energy generated by any chemical reaction (desired If the duty of the heat-removal equipment exceeds the maximum
or undesired) exceed the capability of the vessel or process to rate of heat generation by the reaction, then the answer to this
remove heat? question is no.
8b. If yes, at what temperature does this occur? Set the reaction heat-generation rate equal to the duty of the
heat-removal equipment (both in units of energy/time). Use the
model developed from the calorimeter data to calculate the tem-
perature at which this occurs. Any operating temperature above
this will exceed the heat-removal capability of the reactor.

Source: Adapted from (2).

Copyright © 2013 American Institute of Chemical Engineers (AIChE) CEP July 2013 www.aiche.org/cep 27
Safety

features. Its sample size must be small, ideally 10–15 mL, The behavior of a calorimeter
so that the procedure can be performed safely in a labora- will approach that of a large reactor
tory. The small laboratory reactor must be representative of
the behavior of a large commercial reactor (which could be if the calorimeter is adiabatic.
20,000 gal or larger in size).
The thermal responses of chemical reactors vary con- also absorb more heat than a thin-walled container, so its
siderably with size due to differences in the rate of heat self-heating rate will be slower. Thus, a thin-walled sample
lost through the reactor wall. Small reactors have a high container with a low heat capacity, but strong enough to
surface-area-to-volume ratio, and the heat loss through the provide reliable containment, should be used.
reactor wall dramatically affects the dynamic behavior of Many reactions produce gaseous products and may
the reactor; large reactors, on the other hand, have a low involve chemicals (reactants or products) that have high
surface-area-to-volume ratio, and the heat loss through their vapor pressures. A high-pressure reaction vessel is nec-
walls has a smaller effect. Thus, an exothermic reaction in a essary for such reactions. Unfortunately, this requires a
small beaker may have a self-heating rate of several degrees thick-walled containment vessel, and the heat absorbed
Celsius per minute, while the same reaction in a large indus- by the thick walls will delay the thermal response of the
trial reactor may have a self-heating rate 10 times larger system. This problem can be solved by placing the reac-
(or more). The fact that a large reactor will have a faster tion vessel inside a containment vessel, and then match-
thermal response rate than a smaller reactor is not intuitive. ing the pressure in the containment vessel to the pressure
Consider a small test volume in a large reactor as inside the reactor. Additional equipment is needed for this
shown in Figure 1. The test volume is completely sur- setup: another pressure sensor for the reactor, and a control
rounded by reaction fluid at the same temperature. Such system that can rapidly adjust the pressure in the outer
a system serves as an adiabatic test element. Thus, the containment vessel. The reactor in this system can have a
behavior of the calorimeter will approach that of a large thin wall, since it needs to withstand only the small pres-
reactor if the calorimeter is adiabatic. sure differential between the inside of the reactor and the
Adiabatic behavior can be achieved in a calorimeter containment vessel.
as shown in Figure 2. The entire calorimeter is encased in
an insulated box. The sample of reaction fluid to be tested Acquiring the data
resides in the sample cell, which is surrounded by heating To begin the calorimetry experiment, the sample to be
coils. Temperature sensors measure the temperature of the tested is placed in the sample cell. The controller first heats
sample cell and its surroundings, and a controller keeps the sample using the external heating coils at a constant
the temperature outside the sample cell the same as that rate, usually 1–2°C/min. When the controller detects
inside the sample cell. The control system must respond exothermic behavior (i.e., a temperature increase above
very quickly to changes in reaction temperature, since reac- the ramp rate), it switches to adiabatic mode. In this mode,
tion self-heating rates may be as high as several hundred the controller adjusts the temperature of the cell jacket to
degrees Celsius per minute. match that of the sample cell.
Another consideration in the design of the experimental A temperature vs. time plot (Figure 3) of the calorimetry
setup is the heat capacity of the sample container. A thick- data shows (A) the prescribed temperature rise (1°C/min),
walled container can tolerate high pressures, but it will (B) the onset of an exothermic reaction, (C) the establish-

Insulation
Heating
Coils

Sample Cell Controller

Sample
Thermocouple
External
Thermocouple

p Figure 1. A small volume in a large chemical reactor behaves adiabati- p Figure 2. A temperature-tracking control system allows a calorimeter to
cally since the surrounding fluid is at the same temperature. achieve near-adiabatic behavior.

28 www.aiche.org/cep July 2013 CEP Copyright © 2013 American Institute of Chemical Engineers (AIChE)
 200 hand, temperature can be measured easily and rapidly. The
D
End reaction conversion, x, can be related to the total tempera-
160 Test
ture change during the reaction by the approximation:
Temperature, °C

120
A B and C T −TO
x= ( 3)
80 T F −TO

40 1°C/min where TO is the reaction onset temperature (K) and TF is

the final temperature (K).
0 This approximation is based on several assumptions:
0 10 20 30 40 50 60
• the reaction can be characterized by an onset tempera-
Time, min
ture and a final temperature
p Figure 3. A typical temperature-vs.-time plot can be generated using • a procedure is available to determine TO and TF
data obtained from calorimeter experiments. The plot shows (A) the • the overall heat capacity of the entire reaction sample
prescribed temperature rise (1°C/min), (B) the onset of an exothermic
reaction, (C) the establishment of adiabatic conditions, and (D) the remains constant during the reaction — a condition
maximum temp­erature reached when the reaction is complete approached if one of the reactants is in excess or the reac-
tion system includes a solvent
ment of adiabatic conditions, and (D) the maximum temp­ • there are no heat losses
erature reached when the reaction is complete. • the liquid reaction mass is heated only by the reaction.
The pressure data obtained by the calorimeter is also While industrial reactions generally do not meet these
important, particularly for relief-system design. Pressure assumed conditions, the calorimeter is typically operated
data can be used to determine the vapor pressure of the at conditions that minimize the effects of the assumptions.
reaction liquid; in this case, the reaction vessel is evacu- Once the data are obtained from the calorimeter, the results
ated before the reactants are added, and the pressure sensor can be combined into a more general model that is not
measures the vapor pressure of the reaction liquid. To limited by these assumptions.
determine whether the reaction system produces non­ The onset temperature is used only to relate the con-
condensable gaseous byproducts, cool down the vessel to centration of the reactants to the temperature of the reactor;
the starting temperature, and watch the pressure. A higher it should not be interpreted as a “safe” temperature below
pressure indicates the production of noncondensable gas- which the reaction will not occur. Data from different calo-
eous byproducts. rimeters will yield different onset temperatures depending
Finally, a comparison of the time and temperature at on the sensitivity of each device.
which the maximum rates of temperature and pressure The following dimensionless variables can be defined:
increase occur can be used to classify the reaction type as
T F −TO
either gaseous or hybrid (2). B= ( 4)
TO
Calorimeter modeling Ea
Consider a reaction occurring in an adiabatic batch Γ= (5)
R g TO
reactor. Assume that the overall reaction can be represented
by nth-order kinetics: τ = k (TO )C 0n −1t (6)
dC
= –k (T )C n (1) where B is the dimensionless adiabatic temperature, Γ is
dt
the dimensionless activation energy, and τ is the dimen-
where C is the reactant concentration (mol/m3), T is the sionless time.
absolute temperature (K), and k(T) is the Arrhenius reaction- These dimensionless variables are combined with the
rate constant (s–1 for a first-order reaction) given by: reaction conversion (x) defined by Eq. 3 to obtain the fol-
lowing dimensionless form of Eq. 1 (2, 4):
⎛ –E ⎞
k (T ) = A exp ⎜⎜ a ⎟⎟ ( 2) dx n ⎛ ΓBx ⎞
= (1− x ) exp ⎜ ⎟ (7)
⎝ R gT ⎠ dτ ⎝ 1+ Bx ⎠
where A is the pre-exponential factor (s–1) and Rg is the The calorimeter can now be modeled using just three
ideal gas constant (0.00831 kJ/mol-K). parameters: n, B, and Γ. Once these parameters are known,
Measuring the concentration as the reaction proceeds Eq. 7 can be integrated with respect to dimensionless time.
is difficult, particularly if the reaction is fast. On the other However, to convert the result to temperature and real time,

Copyright © 2013 American Institute of Chemical Engineers (AIChE) CEP July 2013 www.aiche.org/cep 29
Safety

the onset (TO) and final (TF) temperatures, reaction order rate of change (dT/dt) vs. –1,000/T (Figure 4). The onset
(n), pre-exponential factor (A), and activation energy (Ea) temperature is calculated from the value on the x-axis cor-
must be known. Note that the onset and final temperatures responding to the point at which the dT/dt curve begins to
are also coupled to the dimensionless parameters B, Γ, and increase; similarly, the final temperature is calculated from
τ and cannot be independently specified. the point at which dT/dt is the same as the rate at TO. This
The maximum self-heating rate is found by differentiat- plot can also be used to determine the maximum experi-
ing Eq. 7, setting the result to zero, and solving for x. The mental rate of temperature change (i.e., highest slope).
resulting conversion at the maximum self-heating rate, xm, Once TO and TF are known, the dimensionless adiabatic
is given by: temperature rise (B) can be calculated from Eq. 4.
1 ⎡ The remaining parameters are determined by taking the
xm = – (Γ + 2n ) + Γ [Γ + 4n (1+ B )] ⎤⎦ (8) natural log of both sides of Eq. 7 and converting this into
2nB ⎣
dimensional variables:
⎡ dT ( ⎤
)
Estimate parameters from calorimetry data
ln ⎢ dt ⎥ = ln ⎡⎣k (TO )C 0n −1 ⎤⎦ + Γ ⎛⎜T −TO ⎞⎟ (9)
The calorimeter model (Eq. 7) can be used to estimate 1−n n
⎢⎣ (T F −TO ) (T F −T ) ⎥⎦ ⎝ T ⎠
the important process parameters related to the reactive
system. Since first-order reactions are most common (others are
The first parameters that must be determined are the possible, including fractional reaction orders), an initial
onset temperature (TO) and the final temperature (TF). This reaction order of n = 1 is assumed. Next, the left side of
is done by plotting the log of the calorimeter temperature Eq. 9 is plotted vs. (T – TO)/T. If the plot is a straight line,
then the assumed reaction order is correct. The intercept of
the straight line is:
100
ln ⎡⎣k (TO )C 0n −1 ⎤⎦ (10)
and the slope is Γ.
dT/dt, K/min

Final
Temperature For a first-order reaction (n = 1), Eq. 9 reduces to:
10
⎡ dT ⎤
ln ⎢ dt ⎥ = ln A − E a (11)
Onset Temperature ⎢⎣T F −T ⎥⎦ R gT
1
–3.4 –3.2 –3.0 –2.8 –2.6 –2.4 –2.2 –2.0
The left side of Eq. 11 is then plotted vs. –1,000/T to
–1,000/T, K–1
produce a straight line with a slope of ln(A) and a y-inter-
cept of –1,000Ea/RgT (Figure 5). Selecting which data to
p Figure 4. The onset temperature (TO) and the final temperature (TF) use to fit a straight line requires some judgment; in this
can be determined by plotting the log of the calorimeter temperature rate case, the data points fitting a straight line with the least
of change (dT/dt) vs. –1,000/T.
variation were selected.
4 Y = 9015.5X + 23.072
R2 = 0.9878 460

2 Data Chosen for Trendline 440 x Experimental Data

Data Excluded
420 Theoretical Model
⎛ dT / dt ⎞
⎟
⎝ TF − T ⎠

0
Temperature, K

400
–2
ln ⎜

380

–4 360

340
–6
–3.50 –3.25 –3.00 –2.75 –2.50 –2.25 320
20 21 22 23 24 25 26
–1,000/T, K–1
Time, min
p Figure 5. For a first-order reaction, a plot of the left side of Eq. 11
vs. –1,000/T is a straight line with a slope of ln(A) and a y-intercept of p Figure 6. The theoretical model agrees well with the experimental data
–1,000Ea/RgT. for the example reaction.

30 www.aiche.org/cep July 2013 CEP Copyright © 2013 American Institute of Chemical Engineers (AIChE)
 Figure 6 shows how the theoretical model (Eq. 7) Find the temperature at which this occurs from Eq. 3:
matches the experimental data for a set of calorimeter data
for a reaction of methanol and acetic anhydride. T =TO + x (T F −TO ) = 298 + (0.861) ( 447 − 298) = 427 K (16)
The heat of reaction can also be estimated if the heat
capacity of the liquid reaction mixture, the mass of the The dimensionless maximum self-heating rate, from
limiting reactant, and the total reaction mass are known (2). Eq. 7, is:
Once the kinetic parameters have been estimated,
dx  ΓBx 
they can be used in any reactor model — steady-state or = (1− x ) exp 
dynamic; batch, semi-batch, continuous, or plug-flow — dτ  1+ Bx 
to determine the reactor conversion and temperature.  ( 29.4 ) ( 0.50 ) ( 0.861) 
= (1− 0.861) exp   = 967 (17 )
Putting the procedure into practice  1+ ( 0.50 ) ( 0.861) 
Consider the following example: A chemical reaction The dimensional self-heating rate can be calculated by
occurs in a batch reactor with temperature control provided working through the dimensionless parameters:
by internal cooling coils. The reactor is well-insulated so
dT ⎛ dx ⎞
adiabatic behavior can be assumed. The reaction has been = (T F −TO ) k (TO )C 0n −1 ⎜ ⎟ (18)
characterized by calorimetry experiments that generated dt ⎝d τ ⎠
the data in Table 2. and substituting n = 1:
Suppose that the flow of coolant to the cooling coils is dT
interrupted, and the reaction runs away adiabatically. The = ( 447 K – 298 K ) (1.24 × 10 –5 s –1 ) (967)
dt
calorimeter model (Eq. 7) can be used to determine the
=1.78K s = 107 K min (19)
maximum rate of temperature change (K/s) and the temp­
erature at which this will occur. The total heat-generation The heat load is defined by an energy balance around
rate at the maximum reaction rate can also be estimated. the reactor (Table 2):
The runaway behavior of this reactor will depend on
dQ dT
where in the reaction cycle the runaway occurs. If it occurs = mC P ( 20)
at the beginning of the cycle (i.e., as the reactor is being dt dt
charged or immediately thereafter), the maximum temp­ where dQ/dt is the heat load (J/s), m is the total reaction
erature reached and the rate of temperature increase will mass in the reactor (kg; in this example, 1,000 kg), and CP is
be higher than they would be if the runaway conditions the heat capacity of the liquid (2.31 J/kg-K). Substitute the
were to occur near the end of the reaction cycle, when the values for the example reaction system into Eq. 20 to get:
concentrations of the reactants are lower. For this example,
dQ
assume the worst case — i.e., that the runaway starts at the = (1,000 kg ) ( 2.31 J/kg-K ) (1.78 K/s) = 4,112 J/s ( 21)
beginning of the reaction cycle. dt
First, calculate the dimensionless parameters from Equation 21 gives the maximum heat load, which is the
Eqs. 4 and 5: amount of energy generated by the chemical reaction that
T F −TO 447 − 298 must be removed. Of course, the cooling system will need
B= = = 0.50 (12) to have a higher capacity than the calculated maximum
TO 298
heat load to handle upsets and excursions, and to maintain
Ea 72.8 kJ/mol adequate control.
Γ= = = 29.4 (13)
R gTO (0.00831 kJ/mol-K ) ( 298 K ) The theoretical model can also be used to estimate the
time required to reach the maximum self-heating rate by
Then use Eq. 2 to calculate k(TO):
Table 2. Calorimetry was used to characterize a
⎡ –72.8 kJ/mol ⎤ hypo­thetical reaction and determine these parameters.
k (TO ) = (7.27 ×10 s ) exp ⎢
7 –1
⎥
⎣ (0.00831 kJ/mol-K ) ( 298 K ) ⎦ Reaction Order (n) 1
Onset temperature (TO) 298 K
= 1.24 ×10 s –5 –1
(14)
Final temperature (TF) 447 K
Next, use Eq. 8 to determine the conversion at the Activation energy (Ea) 72.8 kJ/mol
maximum self-heating rate:
Pre-exponential factor (A) 7.27 × 107 s–1
x m = 0.861 (15) Heat capacity of reaction mass (CP) 2.31 J/g-K

Copyright © 2013 American Institute of Chemical Engineers (AIChE) CEP July 2013 www.aiche.org/cep 31
Safety

integrating Eq. 7 until a reaction conversion of 0.861 is can then be translated into real time (relative to the actual
obtained. This can be done with a spreadsheet using a suit- time at which the reactor reached the onset temperature).
ably small step size (e.g., Dτ = 0.005). The dimensionless This information is useful for devising and executing emer-
time (τ) for which the reaction conversion is equal to 0.861 gency response procedures.
In actual industrial practice, an exothermic
reaction of this type would most likely be done
Table 3a. The information determined via the calorimetry method
discussed in this article can be applied to design and operations.
in a semi-batch reactor or in a batch reactor with
a solvent. For the semi-batch reactor, the reac-
Process Design or Suggested Method to
Operating Issue Apply Calorimeter Data
tion heat-release rate is controlled by the rate at
which the reactant is fed into the reactor. For the
Reactor vessel size Divide the target production rate (mass/time) by
the product production rate in the calorimeter
batch reactor with a solvent, the boiling temper-
volume (mass/time-volume). Apply a suitable vapor ature of the solvent and an overhead condenser
volume. are used to control the reactor temperature.
Pressure rating of reactor Use the maximum pressure from the pressure vs.
time data. Apply a suitable safety factor. Controlling reactive hazards
Temperature rating of Use the maximum temperature from the temperature Tables 3a and 3b suggests how the informa-
reactor vs. time data. Apply a suitable safety factor. tion obtained from the calorimeter can be used
Operating temperature of The maximum production rate will occur at the in process design and operations. Such informa-
reactor temperature of the maximum adiabatic self-heating tion is particularly important for designing the
rate. Other temperatures may be selected based on heat-removal equipment and the control system
other factors, such as catalyst life. needed to maintain the proper reactor operating
Maximum heat-removal Calculate the reactor energy balance. Multiply the temperatures (2, 3). Of course, other design and
rate from reactor to maximum adiabatic self-heating rate at the reactor operation issues, such as process siting, operat-
control temperature operating temperature (temperature/time) by the
ing specifications of other process equipment,
heat capacity of the liquid mass (energy/mass-
temperature) and by the mass of the liquid. See the and startup and shutdown procedures, are also
example provided in this article. important and must be considered.
Maximum reactant feed Divide the reactor heat-removal capacity
An understanding of these calorimetric
rate for semi-batch or (energy/time) by the heat of reaction (energy/mass). methods — and their associated assumptions —
continuous reactor is important to ensure that the data are applied
Table 3b. The issues listed in this table are interdependent;
properly in an industrial environment. Before
only the calorimeter parameters of interest are listed. you can handle reactive chemicals safely, you
Process Design or
need to know what you are dealing with. CEP
Operating Issue Calorimeter Parameters of Value
Reactor temperature Onset temperature
control and sequencing Maximum temperature Literature Cited
Temperature at maximum adiabatic self-heating rate 1. U.S. Chemical Safety and Hazard Investiga-
Maximum adiabatic self-heating rate tion Board, “Improving Reactive Hazard Man-
Heat of reaction agement,” Report 2001-01-H, CSB, Washington,
DC (2002).
Calorimeter parameters for side reaction if present
2. Crowl, D. A., and J. F. Louvar, “Chemical
Alarm/shutdown Onset temperature Reactivity,” Chapter 8 in “Chemical Process
temperatures and Maximum temperature and pressure Safety: Fundamentals with Applications,”
pressures 3rd ed., Prentice Hall, Englewood Cliffs, NJ,
Temperature at maximum adiabatic self-heating rate
Emergency response pp. 381–418 (May 30, 2011).
Time to maximum temperature and pressure rate
procedures 3. Johnson, R. W., et al., “Essential Practices for
Calorimeter parameters for side reaction if present Managing Chemical Reactivity Hazards,” Center
Relief device sizing to Temperature and pressure values and rates (Ref. 2, for Chemical Process Safety, American Institute
protect the reactor from Chapter 10) of Chemical Engineers, New York,
overpressure Heat of reaction NY (Mar. 2003).

Side reaction calorimeter parameters 4. Balakotaiah, V., et al., “Runaway Limits for
Homogeneous and Catalytic Reactors,” Chem.
Source of reactor pressure: vapor pressure or Eng. Science, 50 (7), pp. 1149–1171 (Apr. 1995).
gaseous products.

32 www.aiche.org/cep July 2013 CEP Copyright © 2013 American Institute of Chemical Engineers (AIChE)
 Additional Reading Acknowledgments
Townsend, D. I., and J. C. Tou, “Thermal Hazard Evaluation by an Amy Theis of Fauske and Associates and Chad Mashuga of Dow Chemical
Accelerating Rate Calorimeter,” Thermochimica Acta, 37 (1), provided valuable input to this article.
pp. 1–30, (Apr. 1, 1980).

Daniel A. Crowl is the Herbert H. Dow Professor for Chemical Process Safety in fuel science from the Pennsylvania State Univ. and an MS and PhD in
at Michigan Technological Univ. (Houghton, MI; Phone: (906) 487-3221; chemical engineering from the Univ. of Illinois.
Email: crowl@mtu.edu). He has been involved with process safety educa-
tion and research since 1985. He is the author or editor of several books Jason M. Keith is a professor of chemical engineering, Earnest W. Deaven-
on process safety, has produced many modules and student certificate port, Jr. Chair, and the Director of the Dave C. Swalm School of Chemical
programs for the Safety and Chemical Engineering Education (SAChE) Engineering at Mississippi State Univ. (Email: Keith@che.msstate.edu;
program, and is a past editor of AIChE’s quarterly journal Process Safety Phone: (662) 325-2480). Prior to joining Mississippi State Univ., he was
Progress. He worked for two years as a process control engineer for St. an assistant and associate professor at Michigan Technological Univ.
Regis Paper Co., and then spent 16 years teaching at Wayne State Univ. He has received numerous teaching and research awards, most notably
in Detroit before joining Michigan Tech in 1993. His research is in flam- the Raymond W. Fahien Award from the Chemical Engineering Div. of the
mability and reactivity, and he consults with industry in these areas. He is American Society for Engineering Education (ASEE). He leads research
a member of the AIChE SAChE Committee, the 11a Committee on Loss Pre- programs in modeling of transport systems for pollution reduction and
vention, the Chem-E-Car Rules Committee, and the ASTM E27 Committee alternative energy applications. He has taught courses in process con-
on the Hazard Potential of Chemicals. He is a Fellow of AIChE, the Center trol, engineering mathematics, transport phenomena, separations, fuel
for Chemical Process Safety (CCPS), the ACS Div. of Chemical Health and cells, and hydrogen energy fundamentals during his academic career.
Safety, and the National Speleological Society. He has received numerous He received his BSChE from the Univ. of Akron and a PhD in chemical
awards, including most recently the Merit Award from the Mary Kay engineering from the Univ. of Notre Dame. He is a member of AIChE
O’Connor Process Safety Center at Texas A&M Univ. He received a BS and ASEE.

www.aiche.org/cep or Circle No.119

Copyright © 2013 American Institute of Chemical Engineers (AIChE) CEP July 2013 www.aiche.org/cep 33