SESSION 3
Topic : Agitated Vessel Reactor design
A. DATA GENERATION DURING PROCESS DEVELOPMENT IN
R&D LAB/PILOT PLANT :
Any Reactor design must consider the following:
• Reaction kinetics – Rate of reaction & reaction equilibrium in case of
a reversible reaction
• Heat of reaction
• Mass transfer efficiency
• Heat transfer efficiency
• Impurities generation
• Material of construction
• Reaction Safety – Ensuring run-away reaction can not take place in
the Reactor
Process Design Engineers have to work with R&D during Process
Development to generate enough data on above considerations to be able
to design the Reactor. Perfect generation of this data is very time
consuming & also costly. For large Petrochemical/Chemical Plants which
have very large capacities with Project Cost running into billions of dollars,
perfect data has to be generated, without which Capex & Opex cost may be
quite a bit higher than optimal. These are commodity Products where
margins are low and any cost inefficiency would make the Project unviable.
In such Projects, time & cost has to be spent to generate all the data
needed.
On the other hand, Specialty Chemicals Projects have normally very small
capacities & time to market has to be short because of which time
available for Process Development is short. In such cases, perfect data
generation becomes impossible & typically simple ways are adopted to
generate enough data for Reactor design.
I will discuss each of the above considerations separately on how the data
needed for Reactor design is generated:
• Reaction kinetics – In reaction kinetics, simplest case is of zero
order reaction where rate of reaction does not depend on
concentration of any reactant and the reaction rate is constant
throughout the reaction period at the reaction temperature, pressure
& pH. For most of the reactions this is not the case. Mostly, the rate of
reaction keeps coming down as the reaction progresses as the rate of
reaction is first order or second order with respect to one or more of
the reactants. First order with respect to a reactant means that
reaction rate is directly proportional to the concentration of that
reactant at any time. Second order reaction with respect to a
reactant means that reaction rate is directly proportional to square of
the concentration of that reactant.
Therefore, for a reaction A + B C + D, rate of reaction R at any
time during the reaction will be = k (Ca.Cb2) if the reaction is first
order with respect to reactant A & second order with respect to
reactant B, where k is the rate constant for the reaction and C a & Cb
are concentrations of reactants A & B at any time during the reaction.
This is for an irreversible reaction. If the reaction is reversible and the
reverse reaction rate is of second order with respect to the Product C
& first order with respect to the Product D, then the overall rate of
reaction R will be = kf(Ca.Cb2)- kr([Link]), where kf & kr are rate
constants for the forward reaction & the reverse reaction
respectively.
Normally, such details are gone into only for large Projects. For small
Specialty Chemicals Projects the simple approach used is to
determine the reaction time ‘T’ needed to complete the reaction at
the operating conditions. This is straight away used in the Plant, as a
simple substitute for reaction kinetics for Reactor design. Generation
of this data for different kinds of reactions is given below:
1. Reaction where all reactants & products are in liquid phase,
catalyst may be in liquid or solid phase - and heat of
reaction is not high. These reactions may be carried out in a
Reactor by putting all the reactants & catalyst together and
then heating the Reactor to the reaction temperature. After
this, the Reactor is maintained at the reaction temperature
till the desired level of conversion of reactants happen. This
time is the time ‘T’. Sometimes, temperature needs to be
raised above the reaction temperature to complete the
reaction. In this case we must have from lab both the
temperatures t1 & t2 along with times ‘T1’ & ‘T2’ at the two
temperatures in place of ‘T’. This assumes that very little
reaction happens during time when temperature is being
raised.
2. Reaction where all the reactants, & the Products are in liquid
phase, catalyst may be in solid or liquid phase - and the heat
of reaction is very high. In such a case, to manage heat
supply/removal (based on whether the reaction is
endothermic or exothermic), one or more of the reactants &
catalyst are added first into the Reactor & one of the
Reactants (say ‘X’) is added continuously at a rate to control
heat evolved/consumed so that the Reactor heat transfer
Area is sufficient to take care of the large heat of reaction.
Say time taken for complete addition of the reactant ‘X’ is
‘T1’. After complete reactant ‘X’ addition, reaction would still
not have gone to completion. Reactor is then maintained at
the reaction temperature, or at a higher temperature, for
some more time (called the cooking time) to complete the
reaction. Let this be ‘T2’. In this case ‘T’ = ‘T1’ + ‘T2’
3. Gas-Liquid reaction- Reaction where one or more reactants
are in liquid phase, catalyst may be in solid or liquid phase &
one or more reactants are in gas phase. Typical example of
this is Hydrogenation of Organics using Ni, Pt or Pd catalyst.
In these reactions, Liquid reactants & catalyst are first
added and then Hydrogen is added continuously under
Reactor pressure control. In these reactions also like in
point 2 above, there is time of complete Hydrogen addition &
then after that cooking time to complete the reaction.
Therefore, as in point 2 above ‘T’ = ‘T1’ + ‘T2’
• Heat of Reaction – Heat of Reaction can be calculated by the
following three methods:
o By doing the reaction in a Reaction Calorimeter (RC1e) and
calculating from data generated by that. This is the most
accurate way but many companies may not have this
Equipment. There are companies having this facility & do this
work for other companies on commercial basis
o By putting temperature instruments in the inlet & outlet of
heating/cooling utility of a Pilot Reactor & putting flowmeter in
the utility line. Flow multiplied temperature gain/loss of utility
will give the total heat of reaction for the Batch. This total heat
of reaction generated in the batch divided by Product formed in
the Batch in Kg gives the heat of reaction Kcal/Kg of Product.
Temperature gain/loss will vary as the reaction proceeds. We
must break down the time into a number of time intervals (say
10) and get temperature gain/loss during each of these time
intervals. Cumulative heat of reaction over the ten intervals is
the total heat of reaction. This does not take into account the
heat loss/gain by the lab Reactor to the atmosphere and
therefore to that extent will be inaccurate
o From heat of formations of the reactants & the products. We
have found this reasonably accurate in many cases. If we use
this method, we will have to make some judgement from
lab/Pilot Reactor whether the calculated figure generally
appears correct. We must take 30% or more margin of safety
(according to judgement) in using heat of reaction calculated
theoretically from heat of formations.
• Mass transfer efficiency – Mass transfer efficiency directly impacts
the Reaction time. The Mass transfer efficiency depends on agitation.
It is desirable to take the reaction to a Pilot Reactor where agitator
type is similar to what we are conceptualizing for the Plant Reactor.
From dimensions of the Pilot Agitator & the Reactor, using software
such as Visimix we can scale up the Agitator to Plant level and
thereby have same mass transfer efficiency as the Pilot Plant. If we
do not have Visimix or some other reliable Software for agitation
scale up, we can do it by providing Agitator in the Plant with same
Power per unit Volume as in the Lab/pilot plant and agitator
diameter/vessel diameter in Plant same as in Lab/Pilot Plant Reactor.
In a situation where it is not possible to do piloting, we should do the
reaction in the lab Reactor with varying agitator RPM from very low to
very high to understand whether the reaction time has much
dependency on Agitator RPM. This should give us some
understanding of whether we need intense agitation or mild agitation
works as well. We should get Plant Reactor Agitator designed by a
professional Agitators designer. We should give him input on
perceived requirement of agitation intensity.
• Heat transfer efficiency – Providing arrangements for
removing/supplying heat of reaction in a Plant Reactor normally is not
a major concern as we have a number of ways to remove/supply heat
of reaction like:
o Jacket
o Internal Coils
o Agitation creates improved heat transfer coefficients
o Circulating reaction mass through an external Heat exchanger
Therefore, we do not need any inputs from lab on heat transfer.
But there is a major concern of forming a hot spot in an
exothermic reaction where heat of reaction is high and we are
adding one of the reactants continuously through a dip pipe. Hot
spot depends on flow rate of the reactant through the dip pipe. In
the lab, the flow rate is very low. In the Plant flow rate through the
dip pipe, if we provide only a single dip pipe, will be several
hundred or several thousand times the flow rate in the lab.
Therefore, a single dip pipe in the Plant Reactor can lead to a
massive hot spot. In cases of high heat of reaction, the reactant
continuously added should be through a sparger or multiple dip
pipes to avoid formation of a severe hot spot. Also, high flow
agitator & placement of dip point(s) in highest flow region reduces
temperature of hot spot. A severe hot spot raises the local
temperature at the point of addition. This can give rise to very high
level of impurities formation or even an unexpected runaway
reaction. Hot spot formation possibility & it’s implications must be
very carefully studied.
• Impurities generation – Impurities generation must be studied
in the lab at various temperatures. As described above, in the Plant
there can be much higher temperature in case of a hot spot
formation for an exothermic reaction. Similarly, there can be
significantly higher skin temperature at the surface of heat transfer
in case of an endothermic reaction because of much lower surface
area to volume ratio in case of the Plant Reactor compared to the
lab Reactor. Such higher temperatures could lead to higher level of
impurities formation than found in lab. Therefore, we must lower
hot spot formation by appropriate sparging & agitation. In our
Colleges we mostly understand that sparger use is for better mass
transfer. My experience is that it is more for avoiding hot spots.
• Material of Construction (MOC) – At early stage of Process
Development ideal MOC of Reactor as per literature must be
determined. MOC must be suitable for all reactants, reaction
intermediates & products. Ideally all reactions in lab during
Process Development should be done In a Reactor of selected
MOC. This ensures that we get reasonably long term performance
of the MOC as Process Development normally takes many
months (my experience for specialty chemicals is average 6
months). Reactor thickness & surface appearance must be
checked for metallic Reactors after Process Development is over
to look for corrosion/pitting that may have happened. For any
MOCs to be considered or even the selected MOC, corrosion
coupons with weld joint may be tied to the Reactor Agitator for
several batches to calculate corrosion rate & make observations
on change in roughness/pitting/impact on weld joint compared to
the original coupon.
MOC selection should consider i) corrosion performance of the
MOC ii) does it colour the reaction mass iii) does it catalyse any
side reaction that leads to impurities formation.
Sometimes a major MOC blunder happens during Process
Development in Glass Reactor. Some reactions get catalysed by
natural light when we do reaction in a Glass Flask in the Lab. As
we did not provide any light for the reaction, we do not report this
as a Photo reaction in the Process Development Report. When
this reaction is taken to a Glass lined Reactor in the Plant, the
reaction will not proceed at all because there is no atmospheric
light available which was there in the lab Reactor.
Reactor MOC selection is not limited to the Reactor Shell, it
applies to all parts which come in contact with reaction
materials, like Agitator, internal coils, Agitator Shaft mechanical
seal internals, Gaskets and any ‘O’ rings. Non-metallic MOC
selection for Gaskets & Elastomers for ‘O’ rings should also be
done in early stage of Process Development & tried thoroughly.
• Reaction Safety – Before the Process is taken to Pilot Plant or
Commercial Plant DSC (Differential Scanning Calorimetry) must
be done on the Reaction mass to check for any runaway reaction
possibility. Runaway reaction refers to any reaction that may run
uncontrolled thermally or pressure caused by gases formed in the
uncontrolled reaction.
B. PROCESS DESIGN OF REACTOR
Agitated Vessel Reactors are mostly used as Batch Reactors. I will
therefore discuss below design method for Batch Reactor only.
Process Design of Reactor consists of the following:
• Reactor Vessel sizing
• Sizing of heat transfer provision to be made
• Agitator specification
Design for different types of Reactors is given below:
• METALLIC REACTOR
o Reactor Vessel Sizing
This is done based Plant Annual Capacity & Batch Cycle
Time (BCT) for the Reactor.
BCT for the Reactor is the summation of time taken for
each operation in the Batch. These typically are:
1. Raw Materials & Catalyst charging Hrs
(normally 1 Hr)
2. Heating/Cooling to Reaction temp Hrs
3. Reaction Time (incl cooking time) Hrs
4. Agitator stopping & settling for sampling Hrs
(normally 1-2 Hrs)
5. Sampling & receiving analysis Report Hrs
(normally 1-2 Hrs)
6. Cooling/Heating for Batch unloading Hrs
7. Batch unloading Hrs
Total BCT BCT Hrs
Let the plant Capacity be ‘P’ Kgs/year
Let us assume plant annual operation time 7200 Hrs/year
(300 days, 3 shift operation)
Therefore production/Hr = ‘P’/7200
Therefore production required per Batch = BCT x ‘P’/7200 Kg
Let the lowest density of the reaction mass over the entire
BCT be = D
Then, the Volume of Reaction Mass per batch will be
= BCT x ‘P’/7200/D Liters
Considering Reactor filling of 70%. Reactor Volume required
= BCT x ‘P’/7200/D/0.7 Liters
If there are internal coils in the Reactor then volume of
internal coils should be added to the Reactor Volume
calculated above.
Generally L/D = 1.5 is taken for the Cylindrical part of the
Reactor
• Heat transfer Area provision – Adequate heat transfer area
needs to be provided to ensure that temperature of the Reactor
can be stably maintained at the desired reaction temperature.
This depends on reaction temperature, temperature of
heating/cooling utility to be used for reaction temperature
control. In a batch reaction heat of reaction is
released/required depending on whether the reaction is
exothermic/endothermic. Let us say that the reaction is
exothermic. Heat will be released as per progress of the
reaction. Theoretically, heat required will be at the peak when
the rate of reaction is at the peak. For a batch reaction, peak
heat of reaction calculation is quite tricky. Some reactions are
not one Step reaction and the reaction happens through an
intermediate formation. Rate of intermediate formation may be
very different than the rate of intermediate conversion to the
final product. Also, heat of reaction of intermediate formation
may be very different than the heat of conversion of
intermediate to the final product. It could even be that
intermediate formation is endothermic while intermediate
conversion to the product is exothermic or the other way
round. Reaction Calorimetry to make the heat
release/consumption curve against time is ideal. We should
take the peak heat release/heat consumption rate to calculate
the heat transfer area required to maintain the temperature at
the desired level.
Calculation of heat transfer area for the peak reaction time &
also for Reactor heating & Reactor Cooling time should be
done & highest of the three should be used for design. Heat
transfer Area requirement calculations should be done with the
help of a Software such as Visimix. Generally, we have found
Overall Heat Transfer Coefficients in the range of 200-500
Kcal/[Link].K. Heat transfer area does not help too much in hot
spots prevention. For that Sparger & Agitator play major role as
discussed in previous pages.
• Agitator Scale up – Ideally, we should get Agitator scale up
done by an Agitation Design specialist based on successful
reaction batches in the Lab. For that we need to provide the
Vessel dimensions, Batch quantity & physical properties of all
reaction starting materials, catalyst & products. Any
observations in lab regarding change in viscosity of the reaction
mass during the reaction must also be provided. At the
reaction temperature & lowest temperature in the Reactor.
GLASS LINED REACTOR – For Glass Lined Reactors we need to
specify Volume of Reactor, Process Conditions, Process Materials &
Agitation objectives. Glass Lined Reactor Vendors have standard designs
& we need to choose from them with the help of the Vendor. My
experience with Glass Lined Reactors is that it is very restrictive in Heat
transfer & Agitation. Heat transfer coefficient is poor & Heat transfer Area
availability is a major constraint because it is available only as a Jacket.
For low heat of reaction there may not be a problem. But, when heat of
reaction is high you will have to slow down your reaction by slow feeding
of reactant being continuously fed. As per my experience one should be
going for a Glass Lined Reactor only when there is no other compatible
MOC available, particularly when heat of reaction is high.
