Download The Role of Chemical Reactivity Data in Process Safety

The Role of Chemical Reactivity Data in Process Safety Management
By Michelle Murphy
Abstract: Chemical reactivity is addressed throughout the requirements of OSHA’s PSM
Standard. It is specifically required in the process safety information element. In addition, it is
necessary input to process hazard analyses, operating procedure development, emergency relief
system design, and mechanical integrity. As the understanding of the impact of chemical
reactivity hazards on the operation of a chemical process continues to develop, it is important to
have a method for developing this data. Equally important is a method for extracting meaningful
reactivity information from the data and incorporating it into process safety.
This paper will present a process for evaluating chemical reactivity hazards using an
Accelerating Rate Calorimeter (ARC®). It will then explain how to extract information from this
data to help define process safety elements such as safe upper and lower limits, emergency relief
system design, etc.
Introduction
OSHA’s Process Safety Management (PSM) Rule 1 was promulgated in February 1992.
Understanding, interpreting and implementing the requirements have been—and continue to
be—an ongoing process. Although many areas of PSM have seen a lot of attention, such as
mechanical integrity and management of change, the impact of chemical reactivity on process
safety has not gotten widespread attention in the chemical industry. Many companies have been
focused on the other areas of PSM, such as updating P&IDs and operating procedures (OPs),
conducting PHAs, gathering mechanical integrity information, instituting management of change
(MOC) programs, defining or re-defining the design basis for pressure relief systems, leaving
chemical reactivity largely overlooked.
Back in October 2002, the US Chemical Safety Board (CSB) published a report entitled,
Improving Reactive Hazard Management 2. Their first conclusion was “Reactive incidents are a
significant chemical safety problem”. They went on to conclude that the OSHA PSM Standard
and the EPA risk management plan regulation 3 have significant gaps in coverage of reactive
hazards. These conclusions resulted in recommendations to OSHA, EPA, NIST, CCPS, ACC,
SOCMA and other industry and trade groups to help facilities handling highly hazardous
chemicals achieve more comprehensive control of reactive hazards.
In 2010, Jordan Barab, Deputy Assistant Secretary of Labor for Occupational Safety and Health,
at a speech presented to the NPRA 4, berated the refining industry (and the chemical industry by
inclusion) for their poor safety record.
Recent incidents, including the 2007 explosion at T2 Laboratories, highlight the need for
comprehensive understanding of chemical reactivity hazards. This paper will present a method
for measuring chemical reactivity, evaluating reactivity hazards, and using them to define
process parameters such as upper and lower limits.
Chemical Reactivity Hazards
Chemical reactivity is addressed throughout OSHA’s PSM Standard. It is specifically required as
part of the process safety information element. Additionally it is used as input to process hazard
analyses, pre-startup safety reviews, operating procedure development, pressure relief and flare
system design, mechanical integrity programs, and management of change. Finally it is utilized
in developing training and contractor programs, conducting incident investigations and planning
for emergencies.
To develop reactive chemical processes, much research is conducted with respect to the desired
chemistry. Research and development is necessarily focused on understanding the operating
parameters associated with product development. Often, because the unknown is hard to
investigate, the evaluation of undesired chemistry goes incomplete.
“A chemical reactivity hazard…is a situation with the potential for an uncontrolled chemical
reaction...” as defined by the CCPS book, Essential Practices for Managing Chemical Reactivity
Hazards 5. It is a situation resulting in an increase in temperature, pressure or gas evolution. The
causes of reactive chemical incidents are numerous, including:
•
•
•
•
Unintentional interaction between incompatible materials
Self-reactivity
Accumulation of reactants
Loss of cooling
Many predictive tools exist, to characterize potential reactivity hazards. These work well as
initial screening tools. When the materials of interest are new or unknown, characterization of
chemical reactivity hazards is best done by experiment.
Experimental Hazard Evaluation
Experimental hazard evaluation can be conducted using many instruments, including:
•
•
•
Thermogravimetric analyzers (TGs)
Calorimeters, such as differential scanning calorimeters DSC) and the Accelerating Rate
Calorimeter (ARC®)
Low phi-instruments, such as the Vent Sizing Package (VSP) and the Automatic Pressure
Tracking Adiabatic Calorimeter (APTAC™)
TGs provide mass vs. temperature data which can help define physical transitions and
decomposition reactions that generate gas. DSCs measure heat flow as a function of temperature.
This data can be used to develop onset temperatures and heats of reactions. ARCs add another
layer of critical information by measuring pressure vs. temperature and time. These instruments
provide heats of reaction, rates of reaction, and rates of gas generation. The high thermal inertia
associated with ARC instruments can slow down energetic reaction so that they can be
measured. The VSP and APTAC produce similar information while operating in a low thermal
inertia environment. These can produce direct scale-up results for less energetic reactions. This
paper will focus on evaluating data from the ARC, APTAC and VSP as the measurement of
pressure is important in defining reactivity hazards.
ARC tests measure the temperature and pressure response of the system with respect to time and
temperature. The nature of the design of an ARC test allows for measuring energetic systems at
elevated temperatures and pressures to uncover potential hazards. The high thermal inertia of the
system slows down reactions to allow for quantification of temperature, pressure and rates. Data
from an ARC Test can provide invaluable guidance in managing process safety. The data
generated includes:
•
•
•
•
•
•
•
•
To – Reaction Onset Temperature
∆Hd – Heat of decomposition
Pmax – Peak pressure achieved at decomposition
Tmax – Peak temperature achieved at decomposition
dT/dt – Rate of temperature rise
dP/dt – Rate of pressure rise
∆Hrxn: Measured heat of reaction for normal chemistry steps
Vapor Pressure
The process for evaluating chemical reactivity should begin with tests of each raw material,
followed by a test of the reaction mixture. Additionally, known potential variables can also be
evaluated in the safety of the laboratory.
Evaluation of each component can define potential runaway reactions, help define scenarios for
pressure relief design, and establish safe operating limits and the consequences of deviation.
Evaluating the reaction mixtures provides information on potential higher temperature runaway
reactions/decompositions, gas generating side reactions, adiabatic temperature rises, onset
temperatures of undesired reactions and appropriate normal operating limits.
In an adiabatic environment, reaction heat from one reaction can often increase the temperature
of the system above the onset temperature of an undesired reaction (often times a decomposition
reaction with high gas generation). For a full scale reactor, if the cooling system fails or is not
sized properly, this can result in an uncontrolled runaway reaction, increased gas generation and
accompanying pressure. Consider, for example, a process where the onset temperature of the
desired reaction is 35°C and the heat of reaction is such that the adiabatic temperature rise is
109°C. This system is also known to have a decomposition reaction beginning around 117°C. If
the system suffers a loss of cooling, the temperature can rise to 144°C (35+117°C) which will
start the undesired decomposition reaction and most likely overpressure the reactor. In this case,
it is critical to know what these values are to define operating limits and specify necessary
controls to protect against overpressure.
Similarly, understanding the onset temperature of a reaction can define a lower operating
temperature to prevent reactant accumulation and overwhelming of the cooling system.
Generic Example
Figures 1 and 2 show ARC test plots from an example thermal hazard evaluation. These plots are
from a test of one of the individual reactants in a reaction system. Here, called Reactant A. In
this process, the reactor operates at 95°C. For Reactant A, Figure 1 shows an energetic
decomposition reaction measured to begin at 159°C (432 K). Figure 2 shows that pure Reactant
A, in a runaway condition, can generate pressure up to 1500 psi with a residual pressure over 600
psi. The high residual pressure indicates generation of a large amount of non-condensable gas.
The plots show that the reactor operates fairly close to the onset temperature of an undesired
decomposition of Reactant A. Many chemical reactivity hazards are evident, including:
1. Reactant A should not be heated without additional reactants
2. An excess of Reactant A should not be present in the reactor
3. Cooling needs to be sufficient to assure that the heat of reaction from the desired reaction
cannot cause the system to reach the onset temperature of the undesired decomposition
reaction
4. Safe operating limits should be defined to prevent the system from reaching the onset
temperature of the decomposition reaction
Since Reactant A is only one component in the reaction system, these hazards are focused on
avoiding the undesired situation of isolating Reactant A in the vessel and heating it up.
Figure 1: Example Temperature Rise Rate Curve
Figure 2: Example Pressure vs. Temperature Curve
T2 Laboratories Explosion
On December 19, 2007 T2 Laboratories suffered a massive explosion, killing 4 people and
injuring 32. Debris was found up to 1 mile away.
The CSB Investigation 6 reported two contributing factors: (1) no redundancy in the cooling
system and (2) the pressure relief system was not sized for a runaway reaction. This is clearly a
case where additional chemical reactivity information may have prevented this tragedy.
The system was tested in a 1-liter glass reactor up to 380°F. During the start-up phase there were
many near misses suggesting that runaway exothermic reactions were possible. The rupture disk
was sized for normal operations, not expected failures, at 400 psig. The vessel MAWP was 600
psig.
During their investigation, the CSB conducted reactivity testing of the normal process mixture
(molten metallic sodium, methylcyclopentadiene (MCPD), and diethylene glycol dimethyl ether
(diglyme)) and a combination of two of the reactants (sodium and diglyme).
Figure 3 shows a calorimetry trace from the normal process mixture test. This plot shows that the
desired reaction mixture, if allowed to increase in temperature, will result in a dangerously high
reaction rate of 2340 °F/min and a pressure generation rate of 32,000 psi/min. In the test, this
reaction burst the test vessel. Before failure, pressures of 1788 psig were measured at a
temperature of 1222°F.
A review of this data shows:
1. The initial reaction generates enough heat to set off a secondary reaction
2. The secondary reaction generates very high temperatures and pressures, well above the
MAWP
3. This dangerous secondary reaction initiates around 390°F
Had this information been available during the design phase, it would have identified the need
for multiple controls:
1. A reliable cooling system that could prevent the lower temperature reaction from
reaching the initiation temperature of the secondary reaction
2. A pressure relief device capable of venting the system fast enough to keep the pressure
below the MAWP plus accumulation
Additional analysis and observations can be found in the report.
Figure 3: Temperature and Pressure vs time data from CSB Report on T2 Laboratories Explosion
Conclusion
Chemical reactivity hazards continue to be a cause of deadly accidents in the chemical industry.
The use of experimental analysis methods can help uncover unknown hazards allowing for the
analysis of prevention and mitigation methods to prevent unwanted reactions and to mitigate
their effects should they occur. Each reaction system is unique and the effects of process and
chemical variables cannot always be predicted using theoretical analysis alone. As we develop
and improve our chemical processes, experimental hazard analysis is another tool to help keep us
safe.
1
Occupational Safety and Health Administration (OSHA), 1992. Process Safety Management of Highly Hazardous
Chemicals, 29 CFR 1910.119, OSHA.
2
The United States Chemical Safety and Hazard Investigation Board (CSB), 2002a. Improving Reactive Hazard
Management, Report Number 2001-01-H.
3
Environmental Protection Agency (EPA), 1994. Chemical Accident Prevention Provisions, 40 CFR 68.
4
Barab, Jordan, 2010, “Learning From Industry Mistakes,” National Petrochemical and Refiners Association
National Safety Conference.
5
CCPS, 2203. Essential Practices for Managing Chemical Reactivity Hazards, AIChE, p.1.
6
CSB, 2009. Investigation Report T2 laboratories, Inc. Runaway Reaction, Report Number 2008-3-I-FL.