Sco2 Supercritical Carbon Dioxide Engineering Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

In many sectors including the pharmaceutical, food, chemical and natural-product industries huge amounts of organic solvents are used for extraction purposes. In search for more environmentally friendly alternatives in the last couple of decades a lot of research has been done into using supercritical fluids (SCFs) as an alternative to traditional solvent extraction methods. SCFs can be divided into low-critical temperature SCFs and high-critical temperature SCFs which all have different selectivity characteristics and solvent power. The solvent power of high-critical temperature SFCs is much higher and therefore they are more capable of solving high molecular weight materials than low-critical temperature SCFs. The biggest problem with high-critical temperature SCFs is their high operating temperatures between 500K and 700K which demands specially fabricated equipment and even more important the possibility of degrading the extract material with these excessive temperatures is almost unavoidable. For purposes in the pharmaceutical, natural-product and food industries usually the low-critical temperature SCFs are more eligible and especially their lower (and density dependant) solvent power makes them more suitable for extraction of specific fractions of a solute, like a designated lipid, protein, alkaloid or oil by slightly modifying the thermodynamic properties of the SCF. Another advantage is that their low solvability makes it much easier to separate the SCF from the extract material later on in the process.

Carbon dioxide technology

One of the mainly used low-critical SCFs in pharmaceutical, natural-product and food industries is supercritical carbon dioxide (SCO2) because of its non-toxicity, inflammability and non-polluting properties. Also most CO2 is produced from waste streams, which makes it abundantly available and very cheap. When compressed into its supercritical state ( SCO2 is used for separation and fractionation of valuable components using it as a solvent for the extract material or the carrier material. By altering the pressure and/or temperature of the SCO2 or by adding an external agent the conditions can be manipulated in such a way that specific components will solve.


FeyeCon Carbon Dioxide Technology, further referred to as FeyeCon, develops and produces innovative products and processes based on CO2 technology. They are active in the pharmaceutical, food, textile, renewable energy and process industries.

FeyeCon is a company that is specialized in the design and construction of innovative products and processes based on CO2 technology. FeyeCon aroused as a hobby of a student at Delft University of Technology in 1992. Since 2003 FeyeCon is located in Weesp and has grown to a company that has 35 to 40 employees. Today FeyeCon is active in the food industry, pharmaceutical industry, process industry and renewable energy industry. FeyeCon develops designs and produces new, improved and more cost-effective equipment and processes for these industries. FeyeCon's Pharmaceutical Technology Department designs and produces processes, installations and equipment in which


For an existing spray drying test-setup developed by FeyeCon and which is used for forming micro particles from solutions by spraying these in SCO2 a CO2 recirculation module has to be designed, developed and implemented. The main problem is that the current equipment runs a process in an open cycle which results in an excessive use of CO2. The main issue with this is that during a process often multiple CO2 bottles are needed and switching them during a process is error prone and not user friendly at all.


The spray dryer shall be upgraded with a CO2 recirculation module. After the spray drying process the CO2 has to be purified from any left particles and solvents so that the left over CO2 is virtually as pure as it came out of the bottle. This CO2 must be brought back to its initial conditions and then reintroduced into the cycle. The current arrangement should be maintained wherever possible. Also the additions and changes that have to be made to the equipment have to fit inside the existing structure.


When this project is finished the recirculation module will be designed according to all different kinds of rules and regulations concerning high pressure equipment. The components have been developed and the whole module is implemented in the equipment. The whole setup has been tested and will be optimized in collaboration with the appropriate people at FeyeCon.

Project delineation and preconditions

Field of work

This project focuses mainly on the mechanical part of the processes and equipment. The chemical, pharmaceutical and biological aspects behind the processes will not be treated. Only when necessary more research will be done concerning these aspects.


Since space in the workshop area is scarce the extensions have to fit into de designated space.


The equipment has to remain in its current form as much as possible. The recirculation unit is an addition to the existing installation so in the future it can be restored to the old situation when preferred or necessary. Preferably and where possible existing or present components have to be reused in order to avoid the unnecessary spending of money. When new components are necessary first there will be looked at the possibility of fabricating these components within FeyeCon. When self-fabrication is not an option components have to be ordered from specialized manufacturers. Regarding safety components compromises will not be an option. When safety is at risk concessions will never be made and only the best is good enough.


In this phase of the project the efficiency is not the main focus. Since it concerns a test-setup it is first of all important that it works properly. In a later stadium efficiency will become of more importance.


Although sustainability is of paramount importance within FeyeCon, this will not be a primary requirement for this project.


To get to the presented goal the assignment is divided into the following seven phases.

The research phase

The analysis phase

The design phase

The realization phase

The test phase

The knowledge transfer phase

The evaluation phase

During all these phases consults will be committed with the designated coach within FeyeCon.

During the research phase a literature review is performed to get more comfortable with the subject. The thermophysical properties of CO2 will be analyzed with the help of its equation of state diagrams. The properties of SCFs and its appliances, the future possibilities of SCFs and the application of SCFs within FeyeCon will also be studied. Besides that the spray drying test-setup will be scrutinized to fully understand its applications and also some experimental runs will be executed to really get a feeling for the equipment.

In the analysis phase of the project first a function analysis will be made for all components in the old situation. After that a closer look will be given to various kinds of possible solutions for the recirculation module. These … solutions will be analyzed; there can be various kinds of methods to accomplish the same goal. Which method is the best and why? Eventually a final method will be chosen.

In the design phase the chosen method will be further elaborated. All calculations concerning the components will be performed and the results will be discussed with the mentor.

After the design phase the realization phase will start. After the design is complete and the final results have been approved the components can be ordered or fabricated. When all components are present it is time to build up the module. This will happen in conjunction with the people at FeyeCon who are specialized in building this kind of equipment. When this is completed everything will be checked and double checked and when approved the test phase can begin.

When the recirculation module is integrated it is time to test the configuration. It will be tested in phases which will be determined when the unit is finished. In each phase a small part of the unit will be pressurized to a low pressure to determine everything holds and nothing leaks. In general the tests will indicate if the components work as calculated and desired. If there are any complications adjustments have to be made before commissioning will start.

When all the previous phases are completed it is time to conclude the documentation. All the acquired documentation will be transferred to FeyeCon. Also the answer to the research question will be formulated.

After realization of the recirculation module is finalized the whole project will be evaluated with FeyeCon. If necessary some small adjustments can be made to report and/or the spray dryer when appropriate. During this phase the final presentation will be prepared for the thesis defense which will take place at The Hague University.


Physical gegevens CO2


What is supercritical CO2

Why CO2




What do we do

Situation now

Open cycle

Where is it used for

How does it work

Fles kookpunt (20 graden 57.291 bar)

Eerst -10 graden

Solutions pumps

Max flow

Components and it's working (also fluctuations of pump BV1)




Temperature entropy diagram

How does the cycle go


Isothermal, iso….. etc,



Appendix…. shows a Process and Instrumentation Diagram (P&ID) for the current test-setup. In this chapter a function analysis will be made for the various components in this setup. In order to understand the whole process it is best to keep it simple so not every little component will be treated. Only components that have an essential or valuable contribution to the process will be discussed. These will be marked red in the P&ID. The analysis will be done on the basis of the temperature-entropy diagram from Appendix….

ST1: CO2 bottle

This is the CO2 source. The CO2 inside the bottle is a saturated liquid, that means it that it contains that much energy that it is about to vaporize. In the Ts-diagram this state of the CO2 corresponds to number 1.

HE1: Cooling heat exchanger

The CO2 is cooled to a temperature lower than the saturation temperature. The main reason for this is to ensure the CO2 that flows to the pump is a pure liquid since the pump is only capable of displacing liquids. In the Ts-diagram this process corresponds to step 1-2. The temperature at state 2 is dependant from the conditions inside the V1: Crystallizer vessel since the HE1 is indirectly connected to the V1 through the H1:heater.

P1: CO2 pump

The liquid CO2 is pressurized by a membrane pump to the desired operating pressure. The reason for the use of a membrane pump is to ensure no lubricants will infiltrate the CO2 flow. In the Ts-diagram this corresponds to step 2-3. In this diagram this process is isentropic. In reality this will not be the case because there will always be some exchange of heat with the surroundings and there will be energy losses due to friction inside the pump. For the rest of the process this is not very interesting and since there is only a pressure sensor after the pump it is not possible to measure either.

HE2: Heat exchanger

This heat exchanger heats up the CO2 to the desired conditions. At this point the CO2 is in a supercritical state. In the TS-diagram this step corresponds to step 3-4.

H1: Heater

This heater heats up an ethylene glycol-water mixture as heat source for HE2.

V1: Crystallizer vessel

The SCO2 flows into the crystallizer vessel where it is mixed with the solute and solution. The crystallizer is heated by a water jacket (HE3: water jacket) to keep the conditions inside the crystallizer vessel constant. In the Ts-diagram number 4 corresponds to the conditions inside the crystallizer vessel.

ST2: Solution buffer vessel

This vessel holds the solution that has to be sprayed into the SCO2 flow.

P2: Solution pump

Eisen en wensen

Not perfect (efficiencies etc.) because it's a lot of guess work, circumstances are not ideal. Lots of uncertainties. Therefore safetymargins taken into account

Materials PEEK, PTFE, (fluoroloy o-rings, back pressure regulator)

Pressure drop RE (NU PR?)





There are a lot of possible solutions for the recirculation of the CO2 flow. The approach will be to keep the existing setup intact as much as possible. This chapter will explain the sensible possible solutions; ones that make no sense at all will not be treated. For all the considered scenarios the starting conditions of the CO2 will be the same. These conditions are and the CO2 is boiling. The starting condition is marked by number 1 in Appendix … From there on for all possibilities the red line will be followed until the red dot which is marked by the red arrow and represents the conditions inside crystallizer vessel used for all scenarios and which are , .

An assumption is made that the compression of CO2 is an isentropic process. That means the process will be adiabatic and reversible. Another assumption that is made is that the expansion of CO2 through a throttle valve, or in this case the back pressure regulator, will be an isenthalpic process. That means there will be no change in enthalpy during this process. Making these assumptions will help simplify the explanation of the scenarios and since the considered scenarios will be used to select the best possible solution in terms of separating- and energy efficiency it won't really matter that the reality will be a little bit different. Images that support the scenarios are found in Appendix…. For all scenarios step 1 to 4 will be exactly the same as in the current situation described in chapter 3.

Scenario 1: Heating to T=90ËšC before expansion, separation at atmospheric pressure, recirculation at P=57 bara.

The CO2 is heated to 90ËšC (4-5) before expansion to atmospheric pressure (5-6), which is 1.013 bara, where the separation process will occur. To establish this 710W of heat has to be transferred to the CO2. At atmospheric pressure the temperature of the CO2 will be 5ËšC. When separation has finished the CO2 will be compressed to 57 bara (6-7). The temperature will rise to 355ËšC during this compression step. Then the CO2 has to be cooled to 20 ËšC (7-8) and condensed (8-1) so the CO2 will be for 100% in its liquid phase. For this cooling step a cooling power of 2350W is needed.

Scenario 2: Heating to 90ËšC before expansion, separation and recirculation at 57 bara

The CO2 is heated to 90ËšC (4-5) (710W) and then expands through the back pressure regulator to 57 bara and 60ËšC (5-6) where the separation process will happen. Then the CO2 is cooled to 20ËšC (6-7) and condensed to its liquid phase (7-1) for which 760W of cooling power is needed.

Scenario 3: Heating to 90ËšC before expansion, separation and recirculation at 57 bara

The CO2 is heated to 90ËšC (4-5) (710W) and then expands through the back pressure regulator to 57 bara and 60ËšC. (5-6) Then the CO2 is cooled to 20ËšC (6-7) where separation will occur. At these conditions separating the methanol form the CO2 is more efficient as Appendix shows. After separation the CO2 is condensed to its liquid phase (7-1). The total cooling power needed is also 760W. (smaller difference, but much less volume)

Scenario 4: Expansion to 57 bara separation and recirculation also at 57 bara

The CO2 is directly expanded to 57 bara and 20 ËšC (4-5). At this stage the CO2 will be in transition from liquid (55%) to vapor (45%). The difference in densities for liquid CO2 and liquid methanol are too small for efficiently separating these substances. Therefore the CO2 is heated (5-6) to ensure all the CO2 evaporates. For this 380W of energy in the form of heat has to be transferred to the CO2. Now the separation can take place. After separation the CO2 has to be condensed back to its liquid phase (6-1) for which 630 W of cooling power is needed. Heat tubing before vessel

Scenario 5: Heating to 57 ËšC before expansion, separation and recirculation at 57 bara

The CO2 is heated to 57ËšC and then expanded to 57 bara and 20 ËšC (380 W). At this stage the CO2 will be in a 100% vapor phase. Now the separation can take place. After separation the CO2 has to be condensed back to its liquid phase for which 630 W of cooling power is needed.

Scenario 6

Scenario 6 is similar to scenario 4. Only there is no need for an extra vessel where the separation process will happen. The CO2 is expanded to 57 bara and 20 ËšC (4-5) where the CO2 will be in transition from liquid to vapor. The CO2 flows into a vessel where it will be heated (5-6), evaporated and separated from the methanol simultaneously. For this 380W of energy in the form of heat has to be transferred to the CO2. After this step the CO2 has to be condensed back to its liquid phase (6-1) for which 630 W of cooling power is needed.

Extra vessel


+ no changes to existing part of setup

- extra compression step

- lots of cooling power needed


+ no extra compression step

- cooling needed

As indicated before the reason the CO2 is heated before expansion to avoid crossing the saturation line in the temperature-entropy diagram for CO2 when expanding. If this saturation line is crossed during expansion the CO2 changes to a transitional state which means the CO2 flow becomes a mixture of vapor and liquid. When expanded further it will cross the critical point (, ) where solid CO2 or dry ice will be formed. This is an unwanted situation since solid CO2 can easily clog the tubing behind the backpressure regulator.


Not changing BMM arrangement

There are various ways to close the cycle

Possible solutions naar 1 atm? 3weg klep?

Pro's each

Counter pro's each

Pressure entropy diagrams

Choice of concept


Overall design

High pressure

Materials (pressure, temp. and co2)


Changes to be made


Back pressure regulator

How it works

Materials (pressure, temp. and co2)

Start conditions

End conditions


Solvent is (m)ethanol in this case

Smaller than 16,6 micrometer will not settle.

Also separation in condenser

Tests will tell more

Cyclone or gravity settler

Gravity settler

Perry's chemical engineers handbook


For high pressure purposes almost any kind of vessel can be fabricated but that comes with a price. To keep the price as low as possible it is better to choose a vessel off the shelf. The two most common vapor-liquid separators are the horizontal and vertical vapor-liquid separators.

(The basic working principles of both these separators)

In theory both these separators will work fine for the setup, but for various reasons the vertical one will be chosen. Tabel shows some advantages and disadvantages of the two different separators that have had impact on the choice of design.


Horizontal separator

Vertical separator



Higher separation efficiency

Liquid surface area independent of liquid level

Capable of handling large liquid volumes

Requires small footprint

Larger vapor residence time

More efficient for high vapor/liquid ratio

Diameter and length are much less dependent from each other



Requires larger footprint

Less suitable for high liquid/vapor ratio

Diameter and length are dependent from each other

One of the main reasons I choose to work with a vertical vapor-liquid separator is that all the motion occurs in the vertical plane. In a vertical separator the velocity of the vapor stream is only affected by the diameter of the vessel so it is unaffected by the liquid level. In a horizontal vessel the vapor velocity rises with a rising liquid level. According to IPS-E-PR-880 it is also preferred to use a vertical separators when (some) evaporation has to occur inside the vessel. Other reasons are:

The much smaller footprint is something that is very important to meet the imposed requirements.

The fact that the diameter of a horizontal separator is inextricably dependent of its length makes it more difficult to find a suitable vessel while that is much less the case with a vertical one.

A typical vertical separator has three ports. A vapor-liquid inlet half way and two outlets, a vapour outlet on top of the vessel and a liquid outlet on the bottom of the as can be seen in Figure….

The dimensions of the vessel are based on some specific design rules that have to be met as shown in the same figure. Especially the mixture inlet nozzle on the side makes it quite hard to find a high pressure vessel with these exact specifications. The solution to this problem is to take a standard open ended sample cylinder and make two ports out of one end as shown in Figure…. The vapor-liquid mixture comes in from the top and goes down through the tube to be released in the vessel at the required level. Than the evaporation of the remaining liquid CO2 and the settlement of the liquid methanol drops can occur and the CO2 vapor stream goes out on the top as well.


When the CO2 - methanol mixture expands through the back pressure regulator to the desired pressure and temperature the CO2 can be in the liquid phase, but in most cases will be boiling and thus be in a transitional phase. The following steps is to separate the methanol from the CO2 to improve the purity of the CO2 stream that goes back into the cycle. This separation will happen in the vapor-liquid separator. While it is possible to separate the two from each other in liquid form due to their difference in densities using gravity separation, it is much easier to separate a liquid from a vapor. As can be seen in Eq…(Sounders-Brown) the rate of gravity separation depends for a great deal on the difference in densities of the components. The easiest way to do this is to divide the two different operations between different stages in the cycle. The first step then will be to evaporate the CO2 after which the mixture will flow into a vapor-liquid separator. Out of cost and space reduction consideration I have decided to do both transitions in one cylinder. That means that inside the separator the present CO2 in liquid form has to be evaporated into a vaporous state while the separation of the two components happens. That means that the mixture has to be slightly heated to the saturated vapor state of the CO2. This can be done in various ways. For some of the considered options I will explain the reasons why I did or did not choose them.

A heating element inside the cylinder

Plunging the cylinder in a hot water bath

A heating mantle

Heating the cylinder with heating tape

One of the biggest problems with a heating element inside the separator is that the size of the element is limited to the size of the openings in the cylinder. Because the openings are relatively small and the ends of the cylinder are spherical it will be very hard to put an element inside that is big enough to transfer enough energy to the mixture in the time it will be inside the separator. Another disadvantage is that anything inside the cylinder will have an influence on the flow inside of the separator.

Plunging the cylinder in a bath of hot water is a good option because it is very easy to maintain a constant temperature inside the vessel. A disadvantage is that it will take a lot of space because you need a heating element that heats the water and you also need a pump or a rotor to keep the water in motion so a constant temperature can be maintained. Plunging the cylinder in a water bath also makes it harder to remove the cylinder for cleaning purposes.

A heating mantle is also quite a good option. Heating mantles come in various forms and sizes. There are electrically heated mantles as well as mantles that are heated by a hot liquid that runs through them. A heating mantle has to fit exactly around the cylinder for maximum energy transfer which makes it quite hard to find one that have the exact specifications you need. They can be ordered to specification but that comes with a price.

Heating tape is by far the cheapest option. It doesn't need a lot of space and can be easily wrapped around any surface. They come in various sizes and assuming it is properly insulated the power density per square millimeter, is also quite high. A downside of all heaters and especially electrical heaters is that they react very slowly.

For this purpose the choice fell on the heating tape. 1,5 x power.

Gravity separation

The separator utilizes gravity as the primary mechanism for separating the liquid droplets from the gaseous CO2. The forces acting on a droplet are gravity, buoyancy and drag. The separator has to be designed in such a way that the gravity force succumbs the drag and buoyancy force. In that way the liquid droplets will disengage. The magnitudes of the forces are defined as follows from Eq…. to ….


Gravity Force ()

Buoyancy Force ()

Drag Force ()

Vapor density ()

Liquid density ()

Droplet volume ()

Terminal velocity of droplet ()

Drag coefficient (-)

Droplet surface area ()

Gravitational acceleration ()

When the net gravity force equals the drag force the droplets will settle at a constant terminal velocity.


Droplet diameter (m)

Eq…. makes sizing a separator vessel very difficult since there are quite some uncertainties. Usually the droplet diameter is unknown and since the drag coefficient is dependent on the droplet size this is also an unknown variable. In 1934 M. Souders and G.G. Brown (Souders & Brown, 1934) developed a method to simplify the sizing of a separator vessel. They describe an empirical constant for separator sizing which can be described as in Eq…:


Empirical constant for separator sizing

In their journal they recommend to use certain factors for designing vapor-liquid or liquid-liquid separator setups and they also supply correction factors for inter alia pressure increments or for the use of mist eliminating devices. The resulting factor can be used to calculate a maximum allowable vapor velocity inside a separator vessel.


Maximum allowable vapor velocity inside vessel ()

According to the Gas Processors Suppliers Association Table , the value for a small vertical vapor-liquid separator lies somewhere between 0.036 m/s and 0.073 m/s. Since the value of depends on; the properties of the media, the size of the drops, the vapor velocity, the design of the separator and the required degree of separation, it is best to use the lower boundary for rough calculation to ensure better results later on.

Table : Typical Values of K for vertical separators at atmospheric pressure: From "Engineering Data Book, 12th edition" by Gas Processors Suppliers Association, 2004, Tulsa, Oklahoma

Table 2 assumes the vessel is equipped with a wire-mesh mist extractor. According to GRAVITY SEPARATOR FUNDAMENTALS AND DESIGN when no mist eliminating device is used, the value for has to be multiplied by 0.15 to provide a margin of safety and to allow for flow surges. This factor will be called for further calculations in Appendix….. Also the operating pressure of around 57 bara has an effect on the factor as can be seen in Error: Reference source not foundTabel , which will be called and lies somewhere between 0.75 and 0.80. Also here the value that will be used is 0.75 to ensure the best possible conditions for the methanol droplets to settle out.

Tabel : Effect of Pressure on Allowable K factor: From "Engineering Data Book, 12th edition" by Gas Processors Suppliers Association, 2004, Tulsa, Oklahoma

The resulting maximum vapor velocity and vessel diameter calculations from Appendix C are shown in Tabel .


Separator sizing


Density liquid methanol


Density CO2 vapor


Mass flow











Maximum vapor velocity


Minimum vessel internal diameter


Design maximum vapor velocity


Vessel internal diameter for critical droplet size



Reynolds number for flow inside vessel

Transition flow

As long as the vapor velocity, which is directed upwards, is smaller than the droplets terminal velocity ( the droplets should settle out. The methods used for determining settlement behavior are based on laws like Stokes Law and Newton's Law. These laws describe the behavior of a single, perfectly spherical droplet or particle in a continuous viscous fluid. Since the flow is in transition and the droplets will interact with each other and hit walls or other obstacles inside the separator vessel, the separation will be hindered. To compensate for this WU recommends In Drum separator design WU that the design maximum vapor velocity () for a vertical separating vessel should be between 75% and 90% of the terminal velocity to ensure the droplets with a diameter equal to or larger than the critical diameter will settle out.

The design maximum vapor velocity will be set to 0.75

The critical droplet diameter can be calculated if we rewrite Eq… as follows:

Now it is possible to select a vessel with the sufficiently large internal diameter. Appendix…. shows the chosen vessel from Swagelok. With the given internal diameter and the given mass flow we are now able to predict for different droplet diameters whether they will settle out or not with only gravity separation.

The drag coefficient is a function of the droplets Reynolds number (Eq…). Assuming that the shape of the droplets can be described as solid spheres the relationship between drag coefficient and the Reynolds number is showed in Figure .


Droplets Reynolds number (-)

Vapor velocity inside vessel ()

Droplet diameter ()

Density CO2 vapor ()

Absolute viscosity of the vapor (

Figure : Drag Coefficient and Reynolds Number for Spherical Particles: From "Engineering Data Book, 12th edition" by Gas Processors Suppliers Association, 2004, Tulsa, Oklahoma

According to the GPSA the curve in Figure can be described by splitting it into three sections. At low Reynolds numbers () there is a linear relationship between CD and Re corresponding to a laminar flow. Stokes law can be applied and Eq… can be rewritten as:

For the terminal settling velocity can be described by the Intermediate settling law:

For Newton's Law is applicable. The description of the terminal settling velocity will then look like:

The drag coefficient for a smooth sphere can be numerically estimated using the following (Gerhart & Gross, 1985) ;

It is almost impossible to predict the exact size of the methanol droplets that have to settle from the CO2 vapor. This makes it very hard to predict the exact events that will happen inside the separator. Sizing of separation equipment is still largely based on empirical methods. Using the internal diameter () of the chosen vessel it is possible to estimate the size of the methanol droplets that will settle out. As described before, droplets will settle out if their terminal velocity is larger than the vapor velocity inside the vessel. Knowing the vapor volumetric flow and the vessel diameter the vapor velocity can be determined. Filling in the vapor velocity in Eq… the theoretically used K-factor can be determined. In compliance with the necessary safety factors it is safe to say that droplets with a K-factor larger than the calculated K-factor, which is , will settle out. When looking at Figure or table 4 in Appendix… it shows this coincides with a droplet diameter of around 180 μm. Determining the terminal velocity of different droplet diameters using Eq… to Eq… and comparing them to the vapor velocity inside the separator vessel it is said that droplets with a diameter smaller than 16.6 μm will definitely not settle out. This diameter corresponds to a K-factor of 0.0024 which in its turn is exactly the K-factor without the proper safety factors. What happens with the droplets inside this range is a bit of a grey area. When these smaller droplets clash with other droplets they will merge together and form larger droplets which will settle out. If not these droplets will be carried further in the system with the CO2 vapor flow.

Figure : Theoretical K-factor for liquid methanol droplets in CO2 vapor as a function of the droplet diameter at T = 20 ËšC and P = 57 bara


The secondary principle that is applicable on this separator is the principle of distillation. Distillation is a process that uses the difference in the boiling points of two or more compounds for separation. The pressure inside the separator is around 57 bara. The CO2 that flows into the separator is boiling and has a temperature of around 293 K. At the same pressure the boiling point of methanol lies around 493 K. When the vessel is heated to a temperature between 293 K and 493 K the methanol remains a liquid and the liquid CO2 evaporates thereby separating itself from the methanol. The CO2 vapor flows to the condenser where it is cooled and condensed back into a liquid closing the distillation cycle. Since the molar fraction of methanol in CO2 is about the principal of distillation will be disregarded.


For the liquid drops to settle out of the gas stream the vapor velocity must be smaller than the terminal velocity of the droplets. In paragraph 5.2.5 Gravity separation some additional safety factors are included to ensure the best possible conditions for separation. When the maximum allowable vapor velocity is known, the minimum vessel diameter () is calculated using Eg…


Minimum inner vessel diameter ()

Vapor volumetric flow ()

Design vapor velocity (

There are various methods to determine the length of a separating vessel. In this case the method will be used stated in Gerunda. For common vertical separators there is a minimum liquid level required to prevent the forming of a vortex on the bottom of the separator. In this case that will not be necessary due to the fact that the bottom of the separator will be closed during the process. This way there will never be a vortex. On the other hand there is a strict limit to the maximum liquid level inside a separating vessel to prevent the inlet nozzle from flooding. When the methanol settles on the bottom of the separator it will stay there until the process has finished. That means that the liquid level inside the separator will slightly rise during the process. According to Gerunda for optimal separating conditions the maximum allowed liquid level inside can be derived from Eq….


Distance between inlet nozzle and maximum liquid level ()

The volume flow rate ()

Vapor velocity ()

To ensure that the liquid level will not exceed this limit it is important to transfer enough heat energy into the vessel that for every second the same amount of liquid CO2 evaporates as flows inside. When the amount of energy transferred to the CO2 is too small part of the CO2 will stay a liquid which results in a rising total liquid level inside the vessel. When too much energy is transferred into the vessel the temperature of the CO2 vapor will rise what is also not preferable. In the worst case scenario the CO2 that flows into the separator is for 100% in the liquid phase. The solution will not be injected in the crystallizer vessel until the desired conditions are reached. That means that the temperature inside the vessel is already at the design temperature and that the CO2 that flows out of the separator is in a vaporous state. The maximum liquid level than depends on:

the maximum volume of the ISCO pump

the mass flow of CO2

the internal volume of separator vessel

The efficiency of separation

The maximum total liquid height can be calculated using Eq…(and Figure). This height will be reached when the CO2 flowing inside the vessel is for 100% in its liquid phase and 100% of the methanol is collected inside the separator.


Maximum liquid level ()

Maximum volume ISCO pump ()

Density methanol at 20ËšC and 1 atm. ()

Density at 20ËšC and 57 bara ()

Mass flow of CO2

Density of CO2 at 20ËšC and 57 bara ()

Figuurtje!!! Geen low level geen draaikolk

The ASHRAE is a little bit less conservative and recommends to use Eg….


Distance between inlet nozzle and maximum liquid level ()

The internal diameter of the vessel ()

In this case both equations result in a value for H of 16 mm.



What must happen

Materials (pressure, temp. and co2)

Evaporation or distillation + calculations





Calculations (Chap 14 and 17 Perry)







MAY 1997

Cost aspects



The condenser is a large heat exchanger that must ensure the CO2 vapor stream is condensed back to a liquid state. Therefore inside the condenser the latent heat has to be given up by the CO2 vapor and has to transfer to the coolant.

The condenser is a self-fabricated heat exchanger that consists of a large stainless steel pipe which holds the CO2 flow and a copper spiral on the outside which holds a flow of coolant, in this case demineralized water. The coolant will be cooled by an available unused cooler, probably a 2 kW ……. cooler. The coolant will be pumped through the copper tubing using a circulation pump with a maximum flow of …. The whole composition will be covert in a thick layer of insulation to make sure the efficiency will be as high as possible for the given circumstances. One of the main requirements was to use available parts where possible. The heat exchanger, the cooler and the pump are components that are already used for variable applications within FeyeCon. For this condenser 2 kW of cooling power is hugely oversized so it will not be necessary to x . Calculating the efficiency of the condenser is quite hard since there are no indications on the performance of this part.

What must happen

Materials (pressure, temp. and co2)

Condensation + rest precipitates




Whit PI one giant TI


End conditions

Safety valves

All vessels, tubing systems and other equipment in high-pressure systems have a design pressure. This is an upper limit pressure based on design codes that the weakest component can safely handle during normal operation. The design pressure is highly dependent from the operating temperature. Figuur shows the influence of a raise in temperature to the shear- and tensile strength of some commonly used metals.

Material strength.jpg

Figuur : Influence of temperature on shear- and tensile strength. Reprinted from Onkenhout (

If for any reason the pressure inside a vessel, tube system or other piece of equipment exceeds the design pressure this overpressure has to be relieved by a pressure relief system. A safety relief valve or pressure relief valve works solely with the process gas or fluid as source of power to prevent it from malfunctioning during a period of power failure. A relief valve is designed to open at a set pressure, when this set pressure is reached a valve seat that is kept in place by a spring loaded piston is forced to slightly open by the process gas or liquid. This results in a small flow through the valve. The pressure force on the piston increases and when in overcomes the spring force the valve pops fully open. Usually the difference between the set pressure and the overpressure at which the valve goes fully open is about 10%. When the valve is fully opened a rated flow capacity flows through the valve until the pressure drops to a safe level and then the valve closes again.


Figuur : Schematic cross section of a safety relief valve. Reprinted from The Michigan Chemical Process Dynamics and Control Open Text Book, by D. Katzman, J. Moreno, J. Noelanders, and M. Winston-Galant, 2007, Ann Arbor: The University of Michigan.

Since pressure relief systems are used to protect process equipment and also people from hazardous situations it is very important that the valves are adequately sized and verified. The calculations are made using the ISO!!! standard API RP 520 (American Petroleum Institute, January 2000) procedures.

Relief valves are designed according to worst case scenarios. For properly selecting and sizing a pressure relief valve the set pressure, the required relief capacity and the physical properties of the media have to be determined. The set pressure is based on the limits of the vessel or system. The upstream relief pressure is the sum of the set pressure, the overpressure and the atmospheric pressure. In general the overpressure is 10% of the set pressure. As said before the pressure limit is dependant from the temperature of the media. Using the Temperature entropy diagram (Ts-diagram) for CO2 (Appendix A) and the thermophysical property tables of CO2 it is possible to determine the worst case scenario conditions and the corresponding properties of the CO2 for which the relief valves will be designed. The CO2 recirculation module consists out of two separate parts, with its own relief valve, divided by ball valve 2 (BV-2). (Appendix…) contains a description of the worst case scenarios for both separate parts as well as the necessary calculations to properly select a relief valve.

Safety Valve 4

Safety Valve 4 protects the separator vessel from over pressurization. Figuur shows the pressure-temperature rating of the chosen vessel. For the worst case scenario an assumption has to be made according the conditions of the CO2 before it flows through the backpressure regulator. Since the normal operating conditions inside the crystallizer vessel never exceed a pressure of 150 bara and temperature of 60 ËšC these conditions will be used to determine the worst case scenario. When CO2 expands over the backpressure regulator from 150 bara to 100 bara the temperature drops from 60 ËšC to 45 ËšC or 318.15 K. This process is isenthalpic. That means the CO2 is in its supercritical phase. As stated in API RP 520 Part I, there are no recognized procedures for certifying the capacity of pressure relief valves in two-phase service. Since supercritical fluids have both liquid and vapor characteristics it is suspected that supercritical fluids fall within this category. This makes it much more complex to size and select a pressure relief valve for this occasion. Within FeyeCon the theory of compressible flow is used to select a proper valve. This has always been sufficient in practice and for further calculations this theory will be used but the accuracy is unknown.

Figuur : Pressure-temperature rating of vessel 304L-HDF4-1000 according to manufacturers' literature

Safety valve 3

Safety valve 3 (SV-3) has to protect the condenser form over pressurization. Since the condenser is self-fabricated and consists of a stainless steel vessel, inside which the CO2 has to condensate back into its liquid phase, and a copper spiral which serves as a cooling spiral and holds the flow of coolant in this case demineralized water. Before we can determine a set pressure for SV-3 it is necessary to determine the maximum pressure the vessel can withstand. This will be done according to the guidelines given by AD-Merkblatt B 1 Berechnung von Druckbehältern, Zylinder- unde Kugelschalen unter innerem überdruck (Arbeitsgemeinschaft Druckbehälter, Vereinigung der Technischen überwachungsvereine, & Verband der Technischen überwachungs-Vereine, 2002). The vessel is basically a seamless stainless steel tube with:

, and


The length of the pipe

Outer diameter of the pipe

Smallest wall thickness of the pipe


As can be seen in Appendix… the maximum allowed pressure inside the vessel is 84 bara. This method uses different kinds of safety measures for uncertainties to prevent hazardous situations. This makes this method safe enough to adopt a set pressure of 80 bara for this safety valve

Sizing method

When sizing relief valves it is important to know the critical flow pressure. This is the back pressure for which the flow reaches sonic velocity. When the velocity of the CO2 inside the relief valve reaches sonic velocity it is going so fast that it cannot go any faster. This is known as the critical flow. According to (American Petroleum Institute, January 2000) the critical pressure ratio is described as follows from Eq…


Critical flow nozzle pressure

Upstream relief pressure

Isentropic exponent

If the back pressure is less than or equal to the critical flow pressure, then critical flow will occur, Eq…. If the back pressure is larger than the critical pressure than subcritical flow will occur, Eq…. Since it can be confusing talking about critical flow, subcritical flow and supercritical fluid from now on the terms choked and non-choked will be used.


Back pressure ()

Upstream relief pressure ()

The real back pressure consists of superimposed backpressure and built-up backpressure. According to Superimposed backpressure is defined as:

The static pressure that exists at the outlet of a pressure relief device at the time the device is required to operate. It is the result of pressure in the discharge system coming from other sources, and it may be either constant or variable. Built-up backpressure is the increase in backpressure because of relief of the pressure relief valve (American Petroleum Institute, January 2000, p.4).

Since the piping system at the back of the relief valve is directly connected to the outside world the superimposed back pressure will in general be 1 atm. The built-up backpressure will be negligibly small. When sizing a relief valve it is best to select a specific manufacturer or even a specific valve in advance, because for sizing the valve some standard manufacturer data is necessary to do so. Since the chosen manufacturer has limited the maximum back pressure at 15% of the upstream pressure, this is still smaller than the critical pressure and therefore there will always be choked flow, this will be used for sizing the relief valve.

For choked flow the orifice surface area can be obtained using Eq… According the ISO 4126-1 standard a slightly bigger surface area will be the chosen. Error: Reference source not found in Appendix… shows some basic specifications of the chosen valve in question.


Required flow through the safety device ()

Coefficient determined from an expression of the ratio of the specific heats ()

Certified derated coefficient of discharge, shall not be greater than 90% of the coefficient of discharge (from manufacturers literature) ()

Capacity correction factor due to back pressure ()

Upstream relief pressure ()

Relief absolute temperature of media ()

Molecular weight of media ()

According to (American Petroleum Institute, January 2000) coefficient C can be expressed as follows from Eq… and the Capacity correction factor can be obtained from Figuur . Since we use 15% of the set pressure as the maximum back pressure the capacity correction factor will be 1.00.


Figuur : Back Pressure Correction Factor, Kb for Balanced-Bellows Pressure Relief Valve. Reprinted from Sizing, Selection, and Installation of Pressure-Relieving Devices in Refineries: Part I-Sizing and Selection (7th ed.) American Petroleum Institute. (January 2000). Washington, D.C.: The Institute.

Materials (pressure, temp. and co2)

Problems (dry ice)

Too big orifice

Too small orifice


Needle valves

How does one work

Why necessary


The tubing that will be used for the recirculation is seamless 316L stainless steel instrumentation tubing. This chapter will show that the pressure rating of the used tubing is more than sufficient to cope with the design pressure of the recirculation. Also the pressure drop inside the tubing will be treated in this chapter. Figure … in Appendix… shows a piece of the Pressure rating data from the manufacturer regarding the used tubing. Since the recirculation works with quite low pressure and temperature it is sufficient for all the piping to have the smallest diameter available. To show this to be true the calculations for minimal wall thickness of the piece of tubing between the separator vessel and the condenser are shown in Appendix …

Pressure rating

The pressure rating of the tubing can be derived from table… in Appendix... In this appendix also some calculations that support the manufacturers' specifications are made. All used fittings have

Pressure drop

When a liquid or vapor flows through a tube there will be a pressure drop along the pipe due to friction. This section will demonstrate that the pressure drop inside the recirculation tubing has such low values that it can be denied in the design phase.

The pressure drop inside a pipe or tube due to elevation change can be derived from Eq…


Pressure drop due to elevation change ()

Density of the media ()

Gravitational acceleration ()

Change in elevation ()

The pressure drop due to friction can be approximated using the Darcy-Weisbach equation, eq…


Pressure drop along the pipe or tube ()

Length of the pipe or tube ()

Diameter of the pipe or tube ()

Density of the media ()

Velocity through pipe ()

Coefficient of friction (-)

Darcy friction factor (-)

The Darcy-Weisbach equation is normally used to calculate the pressure drop due to friction of an incompressible liquid inside a pipe or tube. For compressible liquids or gases the pressure drop is not as easy to determine since changes in pressure result in changes of the density and also in changes of vapor velocity inside a tube or pipe. When it is known that the pressure drop will be very small the change in density will be relatively small and it is acceptable to assume the density is a constant along the tube or pipe. When this is the case the gas can be treated as an incompressible fluid and the Darcy-Weisbach equation stands according to the Chemical Engineering Portal (MyChemE, 2012).

The exact location of the to be integrated recirculation module is still unknown. Therefore, at this time it is hard to speculate on the exact lengths of the tubes, the number of bends and the elevation change. Therefore the pressure drop will be derived per meter of tube, per meter of elevation, per valve, or per bend. Known is that there will be one ball-valve integrated in this part of the recirculation module and that all the bends will be 90Ëš bends with a radius .

Pressure drop due to friction

For laminar flows when the Poiseuille law will be used to calculate λ;

For turbulent flows the value for can be read from a Moody-diagram (Moody, 1944) Figuur , or be determined using the Colebrook-White equation.


Figuur : Moody diagram. Moody, L. F. (1944). ASME Transactions. Friction factors for pipe flow, Vol. 66, p 671.


Absolute roughness of material ()

Internal diameter of the tube ()

This equation is developed using experimental results for turbulent flow and is very hard to solve since the Darcy friction factor () can be found on both sides of the equal sign. For that reason there are many people who developed an equation that approximates the Colebrook-White equation. One of which is the Swamee-Jain Equation developed in 1976 and which approximates the Colebrook-White Equation within a 2% error margin.

Tabel : Resulting pressure drop table

Resulting pressure drop

Pressure drop per meter of tube



Pressure drop per 90Ëš bend



Pressure drop through ball valve



Pressure drop per meter in elevation change



The resulting pressure drops can be seen in Error: Reference source not found. Since the available space for the recirculation unit is limited, the module will be compact. Which means there won't be any tubes longer than one meter and the elevation change will be no more than one meter as well.


ASHRAE, 1997, Handbook of Fundamentals, American Society of Heating, Refrigeration

and Air Conditioning Engineers, Atlanta, GA.

ASHRAE, 1998, Refrigeration Handbook, American Society of Heating, Refrigeration and

Air Conditioning Engineers, Atlanta, GA.

E.W. Lemmon, M.O. McLinden and D.G. Friend, "Thermophysical Properties of Fluid Systems" in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, Eds. P.J. Linstrom and W.G. Mallard, National Institute of Standards and Technology, Gaithersburg MD, 20899,, (retrieved December 4, 2012).

Gerunda, A., 1981, How to size liquid-vapor separators, Chemical Engineering, May, v. 88,

n. 9, pp. 81-84.



Gerhart, P.M., R. J. Gross, 1985, Fundamentals of Fluid Mechanics, Addison-Wesley

Publishing Co., Reading, MA.

GPSA Engineering Data Book

Gas Processors Suppliers Association | GPSA Press | 2004 Year | 12th Edition | ISBN: B00006KFVO | English | PDF | 67.5M | RS

Richtlijn 97/23/EG van het Europees Parlement en de Raad van 29 mei 1997 inzake de onderlinge aanpassing van de wetgevingen der lidstaten betreffende drukapparatuur 

Publicatieblad Nr. L 181 van 09/07/1997 blz. 0001 - 0055

( par. 7.2)