JOVE TRANSCRIPTS

VLE Prac
Vapor-liquid equilibrium is paramount in engineering applications such as distillation,
environmental modeling, and general process design. Understanding the interactions of
components in a mixture is very important in designing, operating and analyzing such
separators. The activity coefficient is an excellent tool for relating molecular interactions to
mixture composition. Finding the molecular interaction parameters allows future prediction
of the activity coefficients for a mixture using a model.

Vapor-liquid equilibrium is a critical factor in common processes in the chemical industry,

such as distillation. Distillation is the process of separating liquids by their boiling point. A

liquid mixture is fed into a distillation unit or column, then boiled. Vapor-liquid equilibrium

data is useful for determining how liquid mixtures will separate. Because the liquids have

different boiling points, one liquid will boil into a vapor and rise in the column, while the other

will stay as a liquid and drain through the unit. The process is very important in a variety of

industries.

In this experiment, the activity coefficients of mixtures of various compositions of methanol,

isopropanol, and deionized water will be obtained using a vapor-liquid equilibrium apparatus

and gas chromatograph. Additionally, the binary interaction parameters of the system will be

determined using Wilson's equation and the activity coefficients.

Principles

Vapor-liquid equilibrium is a state in which a pure component or mixture exists in liquid and

vapor phases, with mechanical and thermal equilibrium and no net mass transfer between

the two phases. Vapor and liquid are separated by gravity and heat (Figure 1). The liquid

mixture is inserted into the system, which is put into a vacuum state with a vacuum pump.

The vapor is condensed and returned to mix with the liquid, which is then passed back to the

boiling chamber. Differences in the boiling point results in some separation of the mixture.
The boiling point of water is higher than that of the added components, so the volatile

components begin to evaporate.

Figure 1: A depiction of the apparatus

An activity coefficient is defined as the ratio of a component's fugacity in an actual mixture

to the fugacity of an ideal solution of the same composition. Fugacity is a property used to

show differences between chemical potentials at standard states. Vapor phase fugacities

can be expressed in terms of a fugacity coefficient [φ: fiV = φi yi fi0V ], with yi = mol fraction of

i in the vapor phase, and fi0V = the vapor standard state fugacity (the fugacity of pure

component vapor at T and P). For low pressures, as in this experiment, φi = 1 and fi0V = P.

Liquid phase fugacities can be expressed in terms of an activity coefficient γi: fiL = γi xi fi0L ,

with xi = mol fraction of i in the liquid phase, and fi0L = the liquid standard state fugacity.

At the saturation pressure (Pis) of this T, the pure component liquid fugacity would be Pis,

because the pure vapor and liquid are in equilibrium. Since liquid fugacities are only weak

functions of pressure, we can approximate the pure component liquid fugacity at T and P
(fi0L) as Pis, as long as the difference between Pis and P is not large. This approximation is

usually called "neglecting the Poynting correction". If experimenters use a VLE apparatus to

measure the compositions of the vapor and liquid which are in equilibrium, experimenters

can directly calculate the activity coefficients provided to also measure P and T. T must be

measured to determine PiS for all i.

The heart of the VLE device, used in this experiment to determine compositions of mixtures,

is a Cottrell pump which "spits" boiling liquid into a well-insulated, equilibrium chamber. Two

magnetically operated sampling valves allow for withdrawal of liquid and condensed vapor

samples. A large reservoir helps to dampen pressure pulses in the system as the on-off

control valve switches, and from fluctuations caused by the Cottrell pump. A slow leak can

be used to create a balance between the rate of withdrawal of air and the rate of input of air

to maintain a constant pressure, if necessary.

A comparable way to solve for vapor-liquid equilibrium is to use a variety of models. Raoult's

law, Dalton's law, and Henry's law are all theoretical models that can find the vapor-liquid

equilibrium concentration data. All three models are related to the proportionality of partial

pressures, total pressure, and mole fractions of substances. Wilson's equation has been

proven to be accurate for miscible liquids, while not being overly complex. Additionally,

Wilson's model incorporates activity coefficients to account for deviation from ideal values.

Procedure

1. Priming the system

1. Vent the VLE system using the vent/control 3-way valve mounted on the frame of the
apparatus, and (if necessary) by draining liquid out of the system into a waste flask.
2. Remove the sample tubes and replace with clean tubes (if necessary). The liquid will not
completely drain out of the system.
 3. For the first run of experiments, refill through the input valve with a mixture of roughly (vol
%) 50% methanol, 30% isopropanol and 20% water. For the second week, refill with roughly
25% methanol, 45% isopropanol and 30% water. For the third week, refill with whatever
liquid you need to repeat. The total liquid capacity is approximately 130 cm3.
4. Fill with liquid to just below the spot where the Cottrell pump intersects with the vacuum
jacket. Too little liquid will cause the system to require very high boiling rates to get enough
liquid to "spit" (when liquid pops while boiling intensely).
5. Use a beaker and pour the liquid into the addition port at the top of the equilibrium chamber.
Close the port.
6. Check the barometric pressure with the mercury manometer on the wall. Adjust the "zero" on
the digital pressure gauge to match (if necessary).
7. Switch the three-way vent valve to "control" and start up the vacuum pump and pressure
controller.
8. Open the throttle valve on the pressure controller several turns and observe the pressure
rapidly drop. Watch the pressure on the digital pressure gauge.
9. Set the control pressure set-point on the pressure controller to obtain ~ 700 mm Hg. Listen for
clicking of the control valve. Once the control point is reached, the noise from the vacuum
pump will be audibly different.
10. At this point, with the throttle valve opened several turns, every time the control valve opens,
too much air is dumped to the vacuum pump and the pressure dips below 700 mm before
slowly recovering. Close the throttle valve completely, then open it about ½ turn.
11. Wait for the control valve to begin clicking again, then close the throttle valve in small
increments until the pressure fluctuates only ~0.5 mmHg or less when the valve is open.
Make minor adjustments to the control point or the leak valve as necessary to maintain very
near 700 mmHg.
12. Once the mixture is within ±10 mmHg of 700 mmHg, turn on the heater power, heating
mantle power, condenser water and magnetic stirrer. Try 25-30 % heater power and 1.5-2
turns mantle power to start. The apparatus will require 20 min or less to approach equilibrium.
Keep adjusting the pressure during this time.

2. Running the experiment

1. Upon boiling, the Cottrell pump will begin to spit liquid and liquid can be seen dripping back
into the boiling chamber. Condensed vapor will require longer to appear. When equilibrium is
reached, experimenters should see steady drips of condensed vapor (2 - 3 drops/s) and
returned liquid (2 - 3 drops/s). The temperature should be stable to ± 0.03 ºC and the pressure
should be stable at 700.0 ± 0.5 mmHg. When these conditions have been established for at
least 2 min (or so), equilibrium is attained.
2. Open the magnetic valves (marked "1" and "2" on the controller) 4 or 5 times each for long
enough to collect about 0.5 cm3 of liquid in each sample tube, and close the tubes. If a valve
does not respond to its button, try flipping the power switch for the controller off then on.
This first sample will be used to wash the tubes and delivery system and will be discarded.
Washing replaces any remaining chemical on the sides of the tubes with the same chemical
that is being sampled, so it will not affect the composition of the test.
3. Momentarily turn off the power to the heater, wait 30 s for the boiling to subside, then vent
the system with the vent/control 3-way valve. Remove the sample tubes, swirl a few times,
then dump them into the waste pot.
4. Refit the sample tubes on the system, turn the vent valve back to "control", turn the power
back on to the heater, and wait for equilibrium to be reestablished. This will only take a few
min if the apparatus does not cool. A slight difference in temperature may be observed when
equilibrium is re-established. This can be due to a slight disturbance of the overall
composition due to sampling.
 5. Once equilibrium is re-established, take two new samples. Have two labeled vials with new
septa ready.
6. After taking ~0.5 cm3 samples in each tube again, turn off the heater, vent the system, remove
the sample tubes and pour them into the vials. Cap the vials and replace the sample tubes with
clean tubes if necessary.
7. While analyzing the samples, prepare a new sample. Drain ~15 cm3 of liquid into a beaker or
flask. Add ~20 cm3 of pure methanol or 50/50 methanol/isopropanol through the sample port.
This will give a new overall composition.
8. Be sure the sample tubes are completely empty, then close the system off, switch the vent
valve to "control", and turn the heater back on. If working quickly, equilibrium will be re-
established rapidly. Note that there should be a temperature difference from the previous
sample.
9. Repeat the equilibration and sampling procedure as before, remembering to take one sample
to wash, and then take the final sample. Continue the experiments by adding component(s).
Twelve data points are sufficient to determine the activity coefficients and (roughly) the
binary interaction coefficients.

3. Shutting down the system

1. Turn the heaters off. When the apparatus begins to cool, shut off the stirrer and condenser
water.
2. Return the system to atmospheric pressure; set the controller >1020 mbar, close the throttle
valve, set the three-way valve to vent and open the valve on the tank.
3. Once atmospheric pressure has been reached, shut off the pump. Drain the liquid from the
reservoir until it reaches the level of the valve, but leave the rest of the liquid in the reservoir.
Close the 3-way valve.

4. Analysis

1. Using nonlinear regression and a standard sum of squared residuals objective function, use
the activity coefficients computed from the raw data to regress the 6 constants for the ternary
Wilson equation (below), for this system. Assess the quality of the fit by graphical methods
and computing the average percent relative deviations (APRD), which are average fit errors x
100.

2. Converge on the true optimal values from several different directions in response parameter
space by using a factorial method for the initial guesses. Compute the precision of the GC
measurements by sufficiently replicating one GC sample to determine relative precisions
according to the t-statistic, and use the precisions to determine whether to accept / reject a
particular GC measurement by appropriate hypothesis test.
3. Compare the relative precisions of the GC measurement to the APRDs, and discuss. Also
report the absolute precisions of the pressure and temperature gauges - determine these once
per day.
Results
The activity coefficients of the data do not show significant deviations from a mean value

for each component (Table 1). This is as expected because for intermediate component

compositions there are not large variations. However, components near 1 have γ's near 1.

Low composition components have high γ's. Components highest in concentration in a

mixture which will have a reduced deviation, therefore it will be closer to ideal (γ = 1).

Components with lower concentrations in a mixture will have higher deviations, so their γ's

will be greater than 1.

Table 1: Results of each sampling of the experimental data.

The data were fit to Wilson model parameters and the coefficients were calculated (Table

2). A simple reduction in the sum of squared residuals between experimental and Wilson

equation (1) activity coefficients was used. This was achieved using Excel's solver function.

The parity plot shown relates the Wilson's Equation model activity coefficients to the

experimentally found activity coefficients. The experimental activity coefficients were

calculated and graphically compared to the calculated model coefficients.

Table 2: Results of fitting the data to the Wilson model parameters.

(1)

The parameter values found were the best fit (Table 3). Ideally the correlation is along the

y=x line; however, a significant correlation resembling the ideal scenario was found (Figure

2). The activity coefficients of the data did not show significant deviations from a mean

value for each component, as expected. A reduction in the sum of squared residuals

between experimental and Wilson equation activity coefficients was used with Excel's solver

function. The parity plot relates the Wilson's Equation model activity coefficients to the

experimentally found activity coefficients.

Table 3: Model parameters with water (a), MeOH (b), and IPA (c). The experimental

values are compared to expected values.

Figure 2: Depiction of the correlation between the experimental activity coefficients

and the model activity coefficients.

Applications and Summary

This experiment demonstrated the equilibration of methanol - isopropanol - water vapor-

liquid mixtures at a constant P = 700 mm Hg and how to measure temperature and

composition and calculate activity coefficients. The activity coefficients of the data did not

significantly deviate from a mean value for each component, as expected. A reduction in the

sum of squared residuals between experimental and Wilson equation activity coefficients

was used with Excel's solver function. The parity plot relates the Wilson's Equation model

activity coefficients to the experimentally found activity coefficients.

In the petroleum industry, distillation is the primary process for separation of petroleum

products. Many oil refineries use distillation for crude oil1. Light hydrocarbons are separated

from heavier particles, separating based on boiling points1. Heavy materials like gas oils

collect in the lower plates, while light materials like propane and butane rise up 1.

Hydrocarbons, such as gasoline, jet, and diesel fuels, are separated1. This process is often

repeated many times to fully separate and refine the products1. Refineries run these

processes at steady state, constantly creating new products at maximum capacity, so

efficiency is key1. Chemical engineers working on these processes focus on optimizing the

efficiency of the production1.

Tray distillation columns are also used to separate a variety of chemical products. Ethanol is

one such product. Through closely related processes, a variety of products such as fuel-

grade ethanol, beer, and liquor can all be distilled2. Specific amounts of alcohol can be

separated from water in order to create a specific proof2. This process is limited to reducing

the percentage of water in the product, but cannot completely eliminate it2. In order to

remove water completely, azeotropic distillation is required, which uses extractor chemicals

to separate water from ethanol2.

FINNED TUBE HX PRAC

Heat exchangers transfer heat from one fluid to another fluid. Multiple classes of heat exchangers

exist to fill different needs. Some of the most common types are shell and tube exchangers and plate

exchangers1. Shell and tube heat exchangers use a system of tubes through which fluid flows1. One

set of tubes contains the liquid to be cooled or heated, while the second set contains the liquid that

will either absorb heat or transmit it1. Plate heat exchangers use a similar concept, in which plates

are closely joined together with a small gap between each for liquid to flow1. The fluid flowing

between the plates alternates between hot and cold so that heat will move into or out of the

necessary streams1. These exchangers have large surface areas, so they are usually more efficient1.

The goal for this experiment is to test the heat transfer efficiency of a finned-tube heat exchanger

(Figure 1) and compare it to the theoretical efficiency of a heat exchanger without fins. The

experimental data will be measured for three different flow rates of monoethylene glycol (MEG).

Two different water flow rates for each MEG flow rate will be used. Using the Wilson plot method

the heat transfer coefficients will be determined from the experimental data. Additionally, the

Reynold's number and the amount of heat transferred will be compared for flow with and without

the fins to evaluate heat transfer efficiency.

Figure 1: Finned-tube Heat Exchanger. 1) MEG outlet temperature 2) water inlet temperature 3)

MEG inlet temperature 4) water outlet temperature 5) water meter 6) MEG accumulation sight

glass/cylinder.
Principles

Heat exchangers transfer heat between two or more fluids. The exchangers use fluid species which

flow in a separate space from an opposing stream that is providing heat. Fins can be added to the

flow area to facilitate more heat transfer, as they increase the surface area available for

transference. The added fins decrease the area through which the species flows and provide more

surfaces on which boundary layers can form, resulting in flow that is less turbulent. The less

turbulent a flow, the larger boundary layer it will have. A boundary layer inhibits heat transfer, so

with less turbulent flow less heat is transferred. When the boundary layer is laminar, there is very

little mixing.

The relationship between the area through which heat can flow and the heat transfer coefficient is

used in calculating the total heat transferred. This relationship is calculated through Equation 1:
 (1)

where Q is heat transferred (Btu/hr), U is overall heat transfer coefficient, A is area through which

heat is transferred (ft2), ΔTLM is the logarithmic mean temperature difference.

The overall heat transfer coefficient equation is:

(2)

where Ab is the surface area of bare inner pipe, Af is the surface area of the fins, ALM is the

logarithmic mean area difference, A is the surface area of the pipe (o = outside, i = inside), Δx

thickness of the pipe, k is thermal conductivity of the pipe, h = Individual heat transfer coefficient.

(o=outside, i=inside)

The Wilson plot method uses experimental data to find UoAo from typical energy balance on the

MEG flow and plot its reciprocal to 1/Re0.8 of the inside pipe. By fitting a straight line and finding the

y-intercept, which is related to the heat transfer coefficient and is described in the first two terms on

the right of the equation above. A typical longitudinal rectangular profile fin efficiency equation is

used as the second equation to solve for the heat transfer coefficient and fin efficiency by

minimizing the sum of squares of an objective function. This method is applied to MEG flow

conditions with varying water flow rates.

To calculate the heat transfer coefficient, the Reynolds Number is used, which is given by the

following equation:

(3)
where G is the mass velocity of fluid flow, D is the diameter of pipe where the fluid flows (Deq, the

equivalent diameter will replace D for calculations with fins), and µ is the fluid's viscosity. Fin

efficiency equation for a longitudinal rectangular profile fin is:

(4)

where m is √(2h/kt), h is the heat transfer coefficient, k is thermal conductivity of the pipe, t is

thickness of fin, and b is the height of fin.

Procedure

1. Start and Flow Rate Determination

1. Open the charge valve located below the steam generator.
2. Start the unit, and allow 15 min for steam to begin forming.
3. Calculate the flow rate of water
1. Start a stopwatch and monitor the gauge displaying the volume of water.
2. Stop the watch after 30 s and record the total volume of water displayed on the
gauge.
3. Divide the volume of water by the time to determine the volumetric flow rate.
4. Record the MEG flow rate from the flow meter.
5. Observe the temperature from the thermocouples, and record the values.

2. Varying the flow rate and shut down

1. To collect data for 6 different runs, set the flow rate of water to either a high or low flow
rate and run it with a high, medium, or low flow rate of MEG.
1. For reference, the previous flow rates have been used: 0.0439, 0.0881, and 0.1323
gal/sec for the low, medium, and high flow rates of MEG, respectively.
2. As previous, record the volumetric flow rates and temperature difference on the
thermocouple for each run.
3. When finished, shut down the instrument.
1. Close the valves to stop the flow of steam, monoethylene glycol and water.
2. Turn off the main switch.

3. Calculations
1. Use Equation 1 to calculate the total heat transferred, Q, with the temperature difference
read from the thermocouples (devices used to measure temperature) and the known
physical dimensions of the heat exchanger (found in the user manual for the unit being
operated). The temperature differences can be taken from the temperature readings of each
run.
2. Calculate the heat transferred for each unique trial run, and use the Wilson plot method to
find the heat transfer coefficients for the three MEG flow rates.
 3. Compare the calculated heat transferred and Reynolds number to theoretical values of the
heat exchanger without fins.
Results

The finned tube heat exchanger did not reach turbulent flow (Figure 2). The fins provide additional

surfaces on which boundary layers form, as known through laminar and turbulent flow theory. If the

fluid is not at a sufficient velocity, the fluid will not reach turbulence. The boundary layers between

the fins overlap in the laminar region, so the fluid will remain laminar.

Figure 2: Reynolds numbers for each setting.

The amount of heat transferred, Q, in the tubes with and without fins at different flow rates of MEG

was compared (Figure 3). The results show that a finned-tube transfers more heat than a tube

without fins at the same operating conditions. In this experiment, the fins clearly improved heat

transfer. This is because heat transfer is more effective when there is a greater surface area

available. The finned-tube heat exchanger transferred more heat (Figure 3), despite the lower

Reynolds number (Figure 2).

Figure 3: Heat transferred between exchangers with and without fins at each flow rate.
Applications and Summary

Heat exchangers are used in a variety of industries, including agriculture, chemical production, and

HVAC. The goal for this experiment was to test the heat transfer efficiency of a finned-tube heat

exchanger and compare it to the theoretical efficiency of a heat exchanger without fins.

Experimental data was measured for three different flow rates of monoethylene glycol (MEG) and

two unique water flow rates for each MEG flow rate used. The Reynold's number was determined

for flow with and without the fins and was used to calculate the heat transfer coefficient, surface

area, and fin efficiency for each unique trial run. This data was used to evaluate if turbulent flow is

possible without the fins and under which set of trial conditions the most heat transfer occurs. The

finned tubes did not reach turbulent flow. The results showed that a fin tube will transfer more heat

than a tube without fins at the same operating conditions because the flow of MEG through the heat

exchanger will not reach turbulence.

In the agriculture industry, heat exchangers are used in the processing of sugar and ethanol2. Both of

these products are processed into a juice, which must be heated to be further processed2. Heat

exchangers are used in heating the juices for clarification2. Once the juices have been processed into

even syrups, further heating with exchangers is necessary to continue processing and form

molasses2. Molasses is cooled using heat exchangers, after which it can be stored for later

processing2.

Heating, ventilation, and air conditioning systems, together known as HVAC, all make use of heat

exchangers3. Household air conditioning and heating units make use of heat exchangers3. In larger

settings, chemical plants, hospitals, and transportation centers all make use of similar heat

exchanger HVAC, on a much larger scale3. In the chemical industry, heat exchangers are used for

heating and cooling a large variety of processes4. Fermentation, distillation, and fragmentation all

make use of heat exchangers4. Even more processes like rectification and purification require heat

exchangers4.
ABSORBER PRAC

Gas absorbers are used to remove contaminants from gas streams. Multiple designs are used to

accomplish this objective1. A packed bed column uses gas and liquid streams running counter to

each other in a column packed with loose packing materials, such as ceramics, metals, and plastics,

or structured packing1. The packed bed uses surface area created by the packing to create a

maximum amount of efficient contact between the two phases1. The systems are low maintenance

and can handle corrosive materials with high mass transfer rates1. Spray columns are another type

of absorber, which uses constant direct contact between the two phases, with gas moving up and

liquid being sprayed down into the gas flow1. This system only has one stage and poor mass transfer

rates, but is very effective for solutes with high liquid solubility1.

The goal of this experiment is to determine how variables including gas flow rate, water flow rate,

and carbon dioxide concentration affect the overall mass transfer coefficient in a gas absorber.

Understanding how these parameters affect CO2 removal enables contaminant removal to be

optimized. The experiment uses a randomly packed water counter-flow gas absorption column.

Eight runs with two different gas flow rates, liquid flow rates, and CO2 concentrations were used.

During each run, the partial pressures were taken from the bottom, middle, and top of the column

unit, and the equilibrium partial pressures were calculated. These pressures were then used to find

the mass transfer coefficient, and the mass transfer coefficients were compared to theoretical

values.
Principles

A gas absorption unit (Figure 1) uses contact with a liquid to remove a substance from a gas mixture.

Mass is transferred from the gas mixture to the liquid via absorption.
Figure 1: Typical gas absorption column.

The overall mass transfer coefficient is the rate at which the concentration of one species moves

from one fluid to the other (Equation 1).

(1)
In equation 1, Gs is the gas molar flow rate per cross-sectional area of the column, pAg is the partial

pressure of CO2, p*A is the pressure in equilibrium with pAg, a is the interfacial area/volume or

“effective area” (a function of column packing), z is the height of the packing, and KG is the overall

mass transfer coefficient in mols/(pressure x interfacial area x time). Mass transfer depends on the

mass transfer coefficients in each phase and the amount of interfacial area available in the absorber.

Henry's Law or Raoult's Law is applied to approximate the partial pressures. They are two laws that

describe the partial pressure of a component in a mixture, and are used together in order to fully

describe the behavior of the mixture at the limits of the vapor-liquid equilibrium relationship. The

objective of a gas absorption column is to control the effluent partial pressure of contaminant. A

liquid solvent flows counter-current to the gas stream to remove the contaminant through

convective mass transfer. The overall mass transfer of a water counter-flow packed column is

measured in this study to determine the effects of water flow, gas flow, and CO2 gas concentration.

The coefficients will then be compared to theoretical values.

Procedure

The experiment uses a randomly packed water counter-flow gas absorption column. The column is

packed with 34 cm of 13 mm berl saddles with 465 m2/m3 surface (effective) area. The pressure

entering the system is about 1.42 bar with a temperature of about 26 °C, and valves at the entrance

and exit of the column allow gas to escape. An "Oxy Baby" Infra-red spectrometer, directly

connected to the unit at various locations, measures gas composition, and tanks of pure gas are used

for calibration.

1. Operating the Gas Absorber

1. Turn on the master switch and close the adjusting valve used to control the amount of water
in the column
2. Open the air flow valve completely and the adjusting valve for column pressure.
 3. Set the air flow rate to the desired level (use a minimum of 20 L/min and increase as
needed), and set the column pressure to ~ 1.4 bar and 25°C using the adjusting valve for
pressure.
4. Start the carbon dioxide flow rate at ~ 4 L/min.
5. Set the water flow at ~ 75 L/h, and adjust the water level to maintain a constant height.
Tweak if necessary while running to ensure constant height.
6. Sample the CO2 partial pressure at the base, center and head of the column using the
pressure taps and the infra-red spectrometer.
7. Perform eight different runs, using two different gas flow rates, liquid flow rates, and
CO2 concentrations. This will enable determination of the most important variables.
8. Allow the system to achieve steady-state when any flow rate is altered. This typically takes
30 - 45 min.
Results

Partial pressures were taken from each trial run. Mass transfer coefficients were calculated from

these and compared to predicted values (Figure 2). The predicted values arise from the calculated

operating line for the absorber (see reference 2 for an in-depth discussion of the operating line).

Solid lines represent the values calculated using the operating line, while triangles represent the

experimental mass transfer coefficient values. Confidence intervals for the model values and the

mean mass transfer coefficient were plotted with dashed lines. These values were compared to

determine how the experimental parameters (liquid flow rate, gas flow rate, and CO2 partial

pressure) affected the overall mass transfer coefficient. Under these operating conditions, only

liquid flow rate had a statistically significant effect on mass transfer when compared to the

confidence interval. The results showed that gas flow rate and feed composition had little to no

effect on the mass transfer coefficient.

Figure 2: Model of the predicted and actual values of the mass transfer coefficient.

Theoretical KG values for a high (30 L/min) and low (20 L/min) were calculated from mass transfer

coefficient correlations and are shown as blue and green lines, respectively, in Figure 3. The

experimental KG values at a variety of liquid flow rates were plotted against the theoretical values

and showed similar trends, verifying the dependence of KG on liquid flow rate. The theoretical values

showed some variation from the experimental values, attributable to minor experimental error.

Figure 3: A graphical depiction of experimental value compared to theoretical values.

Applications and Summary

The goal of this experiment was to use factors of gas flow rate, water flow rate, and carbon dioxide

concentration to determine the overall mass transfer coefficient in a gas absorber. The experiment
used a randomly packed GUNT CE 400 water counter-flow gas absorption column. Eight runs with

two different gas flow rates, liquid flow rates, and CO2 concentrations were performed. Partial

pressures were taken from the bottom, middle, and top of the column unit, and these pressures

were then used to find the mass transfer coefficient.

Under these operating conditions, only the liquid flow rate had a significant statistical effect on mass

transfer when compared to the confidence interval for the given conditions. The process is liquid-

phase mass transfer controlled. Gas-related factors such as CO2 concentration and gas flow rate will

have little to no significance.

Gas absorption is an important mechanism for safety in the production of chlorine3. During normal

operation, gas absorbers treat any consistently occurring leaks. The start-up of a chlorine operation

must be treated until it produces a gas-free product. In the event of a breakdown in the process,

absorbers must be used to treat the gas that has been produced. Additionally, when new leaks form,

the main emergency response unit is the standby gas absorbers. Treatment units are vitally

important in these operating conditions, as they help create a safe environment when dealing with a

dangerous product3.

When refining natural gas, absorption towers are used to remove natural gas liquids from the gas

phase4. An absorbing oil with an affinity to natural gas liquids removes the liquid from the gas phase,

purifying the product. The oil with natural gas liquids is then further purified to recover the liquids,

such as butane, pentanes and other molecules. The oil can then be used again for treatment.
Absorption is also used to remove the major impurities CO2 and H2S from wellhead natural gas,

converting it to pipeline gas. The process uses aqueous amines or glycols as solvents at low

temperatures (typically <40 °C)5

TRAY DRYING PRAC

Dryers are utilized in numerous industrial processes. The function of a dryer is to use heat transfer

processes to dry solids. A variety of dryer types exist. Adiabatic dryers use convection and direct

contact with gases to dry solids, whereas non-adiabatic dryers use methods other than heated gas

contact to dry1, including conduction, radiation, and radio frequency drying1. Dryers can be used for

batch processes or they be in continuous use1.

In this experiment, the effects of temperature and air velocity on the drying rate of sand will be

determined using a tray dryer. Three different power settings (1000 W, 1500 W and 2500 W) for two

different air flow rates will be tested, providing a total of six data sets. From this data, the heat and

mass transfer coefficients can be calculated.

Principles

Tray dryers are one type of batch dryer, which also include fluidized-bed dryers, freeze dryers and

vacuum dryers. Tray dryers use convective heat transfer to flow heated air over solids to dry them.

They are used by a variety of industries including for the production of pharmaceuticals and other

chemicals1. Continuous dryers on the other hand are common to large volume product industries,

such as the food industry1.

To begin the process in a typical tray dryer, the tray is filled evenly with a wet solid, such as sand,

and loaded into the apparatus. The dryer’s adjustable fan and heater allow for continuous variations

in air flow rate from the fan through the drying channel, and heat duty variations in 500 watt

increments. As the dryer operates, water evaporates from the sand into the air. The drying rate is

then calculated by weighing the initial solid/water mixture and subtracting the weight of the final

dry solid and at various timed intervals.

Heat transfer is driven by the temperature difference between the sand and the surrounding air. A

simplified Newton’s Law of Heating (Equation 1) can be used to model the heat transfer between

the heated air and sand-air interface to obtain an experimental heat transfer coefficient. Other heat

duty terms are negligible compared to the term in Equation 1,

Equation 1

where q is the heat transferred, ṁ is the water evaporated in an allotted amount of time or rate of

evaporation, ∆Hvap is the enthalpy of vaporization, hy is the heat transfer coefficient, Tair is the air

temperature, and Ts is the sand’s surface temperature.

In order to obtain an experimental mass transfer coefficient, the transfer of water from sand to air

will be modeled as mass transfer flowing across a true phase boundary. The drying rate equation

(Equation 2) is this model.

Equation 2

where ky is the mass transfer coefficient, C is the concentration of water, and A is the surface area of

the boundary. Concentrations of water in the sand (Cs) and air (C∞) will be obtained by using a mass

balance and psychrometric charts, respectively. These are used to solve for the drying rate.

Theoretical values can be compared to the experimental data by calculating heat and mass transfer

coefficients. The theoretical heat (Equation 3) and mass (Equation 4) transfer coefficients are

obtained from the properties of the substances involved from correlations.

Equation 3

Equation 4
where Re is the Reynolds number, Pr is the Prandtl number, Sc is the Schmidt number, DAB is the

diffusivity of water in air, L is the length, and k is thermal conductivity.

Procedure

The experiment will consist of four runs, each testing a different combination of one of two fan and

heat settings.

1. Tray Dryer Operation

1. Prepare the slurry by mixing 500 g of sand and 150 mL of water and load into the
experimental tray for the unit. Spread the mixture evenly in the tray.
2. With the main unit off, place the tray in the drying chamber.
3. Turn on the main unit, then turn on the blower and heater.
4. Set the air velocity and temperature for each run. The three air velocities should range from
0.8 ft/s to 2.0 ft/s (one high, medium, and low) with a constant temperature around 195 ºF.
The three temperatures should range from 130 - 200 ºF with a constant air velocity of 1.8
ft/s.
5. Take measurements every 5 min over the entire run, which should last 45 min. The data
collected should include inlet air temperature, sand temperature, sand weight, outlet air
temperature, outlet air flow rate, dry bulb temperature, and wet bulb temperature. Use the
digital thermometers for temperature readings, air flow settings for the air flow, a digital
scale for sand weight and a sling psychrometer for the wet and dry bulb temperatures.
6. Repeat the process for each set of settings, totaling four unique runs.
Results

From the data collected, the following information can be obtained. Use psychrometric charts to

determine the absolute humidity, which gives the concentration of water present in the air. The heat

transfer coefficients can be calculated using the measured temperatures and Equation 1. And finally,

the change in mass of the wet sand can be used to calculate the concentration of water in the sand.

The moisture content of sand decreased linearly over time. As expected, the evaporation rate was

found to increase with larger flow rate and heat duty. According to their equations, both heat and

mass transfer coefficients are directly proportional to the evaporation rate at the sand-air interface.

Theoretical values of heat and mass transfer coefficients were found to have a strong positive

correlation with a R2 of 99%. The experimental values only showed a weak correlation after testing.
The relationships between air flow and evaporation rate and between temperature and evaporation

rate both increased linearly (Figure 1, Figure 2). Increased air flow (Figure 1) and increased

temperature (Figure 2) both increased the evaporation rate. These graphs show that when air flow

or temperature increase and the other variable is held constant, the evaporation rate will increase at

an equivalent rate and follow a positive linear trend. The air flow variation test was a measure of

convective heat transfer, while the temperature variation test was a measure of conductive heat

transfer. The sum of the two tests shows that both convective and conductive heat transfer follow a

linear relationship with evaporation rate.

Figure 1: Depiction of the relationship between air velocity and evaporation rate, which increased

linearly.
Figure 2: Depiction of the relationship between temperature and evaporation rate, which

increased linearly.

There are many sources of error in the measurements with the greatest sources for error being the

relative humidity and temperature of the air-sand interface. Also, the air velocity effect on the

weight of the tray was deemed unimportant but it is a source of error. Some of this error may have

also reduced the correlation of the heat and mass transfer coefficients. These coefficients were

calculated theoretically and proven to be correlated. However, the experimental data did not show a

significant trend, despite being theoretically similar.

Applications and Summary

A tray dryer was used to measure the drying rate of sand with respect to convective and conductive

heat transfer. Using the dryer at three different power levels and two different flow rates, six

experimental data sets were found. Measurements were taken by weighing the sand/water mixture

at five minute intervals.

This experiment made use of Newton’s Law of Heating, drying rate modeling, and heat and mass

transfer modeling. Heat and mass transfer coefficients were determined with the use of a boundary

layer model. Theoretically, the heat and mass transfer coefficients show a very strong positive linear
correlation. Even though the experimental results showed a positive trend as well, the data was too

inaccurate to display any significant correlation between the two.

Tray-drying can be used in a variety of fields. One such field is pharmaceuticals. In pharmaceuticals,

tray dryers are used to dry many different base materials, including sticky, granular, and crystalline

materials2. Many plastics used in pharmaceuticals can be dried in tray dryers2. Additionally,

precipitates, pastes, and other wet masses can be dried with a tray dryer, along with crude drugs,

chemicals, powders, and tablet granules. Even some equipment is dried in the dryers2. Tray dryers

offer many advantages to this industry, since they are used for batches, which can vary in size and be

handled without losses2. The dryers are also readily adjusted to accompany other materials in an

efficient manner2. In some cases, tray dryers in a vacuum are used to dry heat sensitive products like

vitamins2.

Tray dryers are also used in food processing3. Food can be spread out thinly and evenly onto the

trays for drying3. Depending on the type of food, drying can be performed by heating with air

moving across the trays, conduction from heated trays or shelves, or radiation form other heated

surfaces3. Air can be used with the additional benefit of removing moist vapors, though this can be a

problem for some foods3.

PBR PRAC

The goal of this experiment is to determine the magnitude of maldistribution in typical packed bed

reactors in both single phase and two-phase (gas-liquid) flow and evaluate the effects of this

maldistribution on pressure drop. The concepts of residence time distribution and dispersion are

introduced through the use of tracers, and these concepts are related to physical maldistribution.

Channeling in a single-phase flow can occur along walls or by preferential flow through a larger

portion of the bed cross-section. Channeling in two-phase flow can result from even more complex

causes, and simple two-phase flow theories seldom predict pressure drops in packed beds. A design

goal is always to minimize the extent of channeling by finding the optimal bed and particle diameters

for the design flow rates and by packing a bed in a way to minimize settling. It is always important to

quantify how much maldistribution might occur and to over-design the unit to account for its

occurrence.

The permeameter apparatus measures pressure drop, ΔP, and the concentration of tracer (dye)

exiting horizontal packed beds of armored glass for either water, air, or two-phase flow (Figures 1

and 2). Water enters through a control valve and can be routed through manual valves to any of five

beds (48" long, 3" I.D.) with different size glass bead dumped (random) packings. The pressure drop

is measured using a pressure transmitter. The water flow is measured by a differential pressure (DP,

orifice) transmitter and the air flow by a dry test meter (similar to a home gas meter). The dye

sample is injected upstream by an automated sampling valve. The exit concentration of the dye from

a bed is measured using a UV-Vis spectrometer. Residence time distributions are calculated from the

tests and compared to the predictions of theories on dispersion in packed beds. Two-phase flow will

be studied in bed 5, which contains the largest particles.

Figure 1: Process and instrumentation diagram of the apparatus.
Figure 2. 3-D rendering of the apparatus. Bed #1 is at the top, bed #5 at the bottom. The water

control valve is on the left (red bonnet). The DP transmitter is at the top center (blue).
Principles

Gas-liquid countercurrent vertical packed beds (packed columns) are frequently used in separation

processes such as distillation, absorption, and stripping.1 Cocurrent horizontal packed beds are often

used as reactors or adsorbers with a solid catalyst or adsorbent. In both cases (as separators or

reactors), the packing increases the surface area of vapor-liquid contact.1 Packing can exist in two

forms: dumped packing, consisting of random or simple geometric shapes of materials such as clays,

metals or ceramic oxides, or structured packing from common metal and plastics, consisting of
highly defined interconnected geometric networks (usually corrugated metals or plastics) that can

reduce pressure drop compared to most dumped packings.1 However, whether horizontal or

vertical, maldistribution (channeling) can degrade the performance of the separator, reactor or

adsorber; sometimes, various types of flow distributors can be employed to mitigate the

effects.2 The single phase packed bed ΔP's can be compared to the predictions of the Ergun

equation.3

Tracers are dyes that are injected instantaneously into the upstream flow, and whose composition

as a function of time is measured in the flow downstream of a bed.4 The measurable tracer

molecules are assumed to be characteristic of all the molecules making up the liquid flow. The

volume of the injected tracer must be small relative to the system volume. If perfect plug-flow (no

axial mixing) occurs in a packed bed, then the tracer injected at time zero would exit the bed at a

later time as a spike. For any real bed, the tracer will disperse exiting the reactor at lower

concentrations over a longer time period. If the flow is not maldistributed, the spreading will be

described by the Gaussian (Normal) distribution, with the peak of the curve observed at the average

residence time. The more the tracer spreads out in time, the worse the maldistribution, and typically

the poorer the separation or reaction process.

The residence time distribution (RTD) describes the distribution of times that molecules can spend in

the bed. If M is the total mass of tracer injected into the system, Q the volumetric flowrate, and C(t)

is the effluent concentration, then the mass balance on the tracer is:

(1)
The left-hand side of Equation (1) represents tracer mass in, and the right-hand side represents mass

out. E(t) is the bed-exiting residence time distribution (RTD), a probability distribution. Using

Equation 1 for the integral, E(t) it can be calculated as:

(2)

E(t)dt is the fraction (probability) of molecules in the exit stream of residence time

between t and t+dt. The terms E-curve and RTD are synonymous. For packed beds, the residence

time is related to the void volume (product of total reactor volume V and porosity ) divided by the

volumetric flow rate, Q. The mean residence time, , can be defined and related to E(t)dt, the

probability a given molecule entering the bed at t = 0 will exit at t:

(3)

As seen from Equation 3, E(t) has units of inverse time. Sometimes the dimensionless E-curve is

plotted instead of the E-curve. This dimensionless E-curve, E(t/ ), is obtained by multiplying the E-

curve by . Its average is 1. Another good way to quantify the deviation from plug flow (the

"dispersion") is to compute the variance of the E-curve (σ2) divided by its mean squared:

(4)

This quantity should be invariant with respect to flow rate for a packed bed, if maldistribution is not

present. The range of values due to molecular diffusion only should be:

(5)
for Rep < 40, where Rep is the particle Reynolds number, dp the average particle diameter and L bed

length. Greater values of experimental σ2 than predicted by Equation 5 and deviations from the

Gaussian distribution indicate flow maldistribution, as does an 'early' peak in the E(t)-curve, or a long

tail on the main peak.

In some cases, the nature and magnitude of the maldistribution can be observed visually. This is

especially true in two-phase flow. Two simple models exist for two-phase flow, the homogenous

model and the stratified model.3,5 For homogeneous flow, the basic assumptions are that the actual

gas velocity, UG, actual liquid velocity, UL and averaged velocity of the liquid-gas mixture, Utp are

equal:

UL = UG = Utp (6)

Then the two-phase density is given by G/Utp (G is mass velocity), and the averaged two-phase

viscosity, µtp, is given by:

μtp-1 = μL-1 (1 - X) + μG-1 X (7)

where X is the quality (weight fraction of vapor in a vapor-liquid mixture) and µL, μG are the

viscosities of the respective liquid and gas phases.

For stratified flow, once the pressure drop, total porosity, and both volumetric flow rates are known,

the gas volume fraction in active (i.e., not stagnant) flow, α, can be computed by setting the Ergun

equations (or similar equations for ΔP) equal for both phases. One can then predict ΔP/L. No matter

the type of flow, both phases must have equal pressure drops because they are in parallel. The mass

balance relates the two-phase velocity to the actual phase velocities:

Utp = UL (1 - α) + UG (α) = G [ (1 - X)/ρL + X/ρG] (8)

The effect of the gas flow on the liquid is both to reduce its effective cross-sectional area and to

provide a nearly zero-shear interface. The effect of the liquid flow on the gas is also to reduce its

effective cross-sectional area. Therefore, actual two-phase flow pressure drops typically exceed ΔP's

calculated simply on the basis of measuring or calculating α and applying a packed bed ΔP equation

(using α instead of ε).

Procedure

1. Starting up the apparatus

The apparatus is primarily operated through the distributed control system interface. A Perm P&ID

schematic appears and opening/closing automated valves is point and click.

1. To establish water flow to either bed #4 or #5, open the inlet and exit valves to the bed
being tested and the water supply solenoid.
2. Use the flow controller to start water flowing through the bed, raising it gradually. Good
starting points are 400 mL/min for bed #4 and 500 mL/min for bed #5. Monitor the
differential pressure across the beds. Vary the flow to cover the entire possible range of the
DP transmitter.
3. Power up the spectrometer equipment and establish communication with the control
console. Spectrometer procedures are detailed in the operating manual (SpectraSuite). The
calibration of the spectrometer for the fluorescent dye standards will be provided.
4. Perform one tracer test each on beds #4 and 5 using 50 ppm dye in DI water as the tracer, at
a single average flow rate for each bed.
5. Insert the spectrometer probe into the probe sample point (Fig. 1). At the PERM interface,
change injection valve status from "Running" to "Charging."
6. Inject the tracer using the syringe provided into the sample valve. Change the status to
"Running".
7. Clean out the injection chamber of the sample valve by changing its status back to
"Charging", detaching and loading the syringe with water, then injecting at least 100 mL of
water into the valve. When the injected sample has completely exited the bed
(spectrometer absorbance returns to base line), change the valve state back to "Running"
and let the water flow through the valve for 10 - 15 min at a high flow rate before using it
again.

2. Conducting Two-Phase Flow Pressure Drop Experiments

Be sure that the water valves to the beds are closed, the inlet and exit valves to bed #5 are open, the

drain valve is open, and that the manual valve for the air to the beds is closed.
1. Slowly open the air regulator to establish an air flow (< 5 psig at first). Open the manual
valve for the air to the beds.
2. Set water flow controller at desired setpoint (700 mL/min) and open manual valve. Route
water/air flow to gas-liquid separator (see valving in Fig. 1).
3. Confirm that water is exiting to drain. Close the valve to the drain for a period of time to
build up a liquid head in the gas-liquid separator. This will result in better separation of the
air and water.
4. Adjust air flow (typically < 2 SCFM) as desired using the pressure regulator and the dry test
meter on the gas exit line. Close the drain valve for short periods of time to get a correct gas
flow reading on the wet test meter.
5. Conduct two-phase flow pressure drop (use DP transmitter) experiments using bed #5, at
multiple air rates. Try to cover the range of the DP transmitter. Disconnect the dry test
meter if you see water exiting from the gas exit line.
Results

Obtain the RTDs (E-curves, using Equations 1-2) after subtracting an appropriate baseline (if

necessary) from the spectrometer signals. An example of baseline correction for Bed #3 (not used

here) is in Figure 3. Using Equations 1-3, calculate the average porosity, tracer mass, mean residence

time, variance and variance divided by mean squared from the RTDs. Compare calculated tracer

mass with injected mass - if they aren't within expected precision, examine how the baseline was

determined in the spectrometer measurements (and perhaps determine differently). Examine how

the variance compares to the prediction from dispersion theory (Equations 4-5); deviations denote

excessive channeling.
Figure 3. Bed #3 dimensionless RTD E-curve (390 mL/min, 50 ppm tracer injection) with and

without baseline correction. The calculated from Equations 2 and 3 was 3.6 min. The baseline

correction was made by subtracting two average baseline values, one before and one after the

maximum. The one before was subtracted from all values prior to the maximum, the other after

was subtraced from all values after the maximum.

Once the porosities of the beds (Equation 3) have been found, the Ergun equation can be used to

predict the ΔP's for the water flow experiments. The average particle diameter must first be

calculated. Because particle drag is related to area for flow, the surface-area (d2) weighting is usually

the best way to obtain the average diameter for a range of particles. The average diameter can be

computed as follows, obtaining the particle diameters from the information in the Materials List

(ωi is the wt fraction of particles of diameter di):

(9)

The calculated porosities can be used to pinpoint the cause of any discrepancies between predicted

(by the Ergun equation) and measured ΔP's. For example, the minimum porosity for close-packed
spheres is 0.36. It is unlikely that any real ε of an entire bed is less than 0.3. Predicted ΔP's >> actual

ΔP's suggest channeling (short-circuiting) along the walls or in the upper portion of the bed when

settling occurs. Such phenomena would result in a low ε calculated from the E-curve, leading to high

predicted ΔP's. This is the case in Figure 4 for both beds #3 and 4. Note that the more expected ε =

0.36 reproduced the Ergun equation results except at a very high flow rate where a high percentage

of the flow was through low voidage regions. This channeling can actually be observed in the

experiment.

Figure 4. Experimental ΔP's compared to the predictions of the Ergun equation, both at ε = 0.36

and the ε values determined from the E-curves.

Predicted ΔP's << actual ΔP's suggest channeling only through the lower half of the bed, or partial

bed blockage. For these beds, this is unlikely.

For the two-phase flows, compute predicted ΔP's by both homogeneous flow and stratified flow

theories using Equations 6-9. For stratified flow, one must solve the Ergun equation and Equation 9

simultaneously to obtain α, setting Ergun ΔP/L (liquid) = Ergun ΔP/L (gas). Then compare computed

vs. actual ΔP's and see which theory applies best, or if in fact either theory applies. Other flow

regimes (e.g., slug, mist or inhomogeneous bubbly flows) are possible, as are major flow distortions

due to channeling, which is often more prevalent in two-phase flows.

For two-phase flows through bed #5, the ΔP's calculated using homogeneous flow theory prove to

be better than those using stratified flow theory (Table 1), although as seen neither theory applies

exactly. The high actual ΔP's suggest severe channeling in a horizontal bed during two-phase flow -

the liquid is confined to a small portion of the cross-sectional area. Indeed, the gas volume fractions

estimated by visual inspection looked to be at least 0.90. The liquid was also confined to the non-

wall region of lower voidage, which increases ΔP. The results reflect the limitations of the simpler

rheological models for two-phase flow, and why far more sophisticated microrheological models are

finding more use today.

Table 1: Gas volume fractions α and pressure drops in two-phase flow, bed #5.

ΔDP ΔDP
α
Q water Q air Q air (stratified) (homogeneous)
(stratified)
psi psi

mL/min ft3/min mL/min

1100 1.62 45900 0.58 2.2 12

1100 1.26 35700 0.47 1.7 10

1100 1.11 31400 0.38 1.5 9.5

1100 0.930 26300 0.19 1.3 8.6

500 0.73 20700 0.58 0.66 3.4

500 0.50 14200 0.47 0.50 2.7

500 0.39 11000 0.38 0.40 2.7

500 0.16 4250 0.19 0.29 1.4

Applications and Summary

In this experiment the real flow behavior of horizontal packed beds, both in single and two-phase

flow, was contrasted to the simpler theoretical models for pressure drop and dispersion (flow

spreading in the axial direction, deviating from plug flow). The utility of tracer tests in probing for

maldistribution ("channeling") in such beds has been demonstrated, and it has even been shown

that certain metrics calculated from the tracer tests can give some idea of the cause of the

channeling. These calculations using the tracer tests, such as computing the E-curve, are normally

known as "residence time distribution" (RTD) theory.

Channeling in single-phase flow can occur along walls or any other low voidage region, for example,

if settling occurs in a horizontal bed. Channeling in two-phase flow can result from even more

complex causes, and as seen simple two-phase flow theories seldom predict pressure drops in

packed beds. Channeling increases downstream separation costs or can ruin the product. A goal of

design is always to minimize the extent of channeling by finding the optimal bed and particle

diameters for a given desired Q, and by packing a bed in a way to minimize settling.
The tracer method of testing is a simple way to quantify the RTD. However, the tracers are seldom

the same molecules as used in the process (although they can be close, if isotopes are used).

Therefore, tracer molecules may not behave in exactly the same ways as reactant or adsorbate

molecules in the fluid phase. In particular, it is important that the tracer not adsorb on the solid

particles, because then it cannot be fully characteristic of a fluid molecule.

The time each molecule of a reactant spends inside a chemical reactor is an important determinant

of the macroscopic conversion and selectivity to the desired product. The occurrence of "dead

zones" (regions of stagnant flow) often lead to poorer than expected selectivities even when the

conversions are not much affected. This is one reason why RTD theory is so important in reactor

design.4

Tracers are also used by environmental and petroleum engineers to help characterize subsurface

solid packing structure. In these applications, two wells are drilled some distance apart; a tracer is

injected into one and recovered at the other. Because the earth's subsurface is highly

heterogeneous, the effluent profiles (E-curves) are typically nonsymmetrical, indicating the presence

of preferential flow paths. This information helps characterize the structure of the subsurface strata,

which is important for modeling petroleum recovery and contaminant transport in groundwater.

In environmental engineering, the use of partitioning tracers can be used to locate and quantify

organic contaminants in subsurface strata. An inert tracer is injected to characterize the flowing

(aqueous) phase between two wells. A partitioning tracer is then injected, partitioning preferentially

into an organic contaminant phase if one exists. The tracer is light enough that it will eventually

diffuse out of the organic phase. This behavior manifests itself as a time delay as compared to the
inert tracer, and comparison of these two can be used to deduce the volume of stagnant organic

phase present.

Two-phase flows are also commonly found in power plants, in non-reactor, non-adsorber

applications. An example is boiling heat transfer, with the steam created in a boiler. They are also

found in all distillation columns, absorbers and strippers, although in vertical rather than horizontal

configuration.

Materials List
Name Company Catalog Number Comments

Equipment

25-40 mesh (50%)

Bed #3 – glass beads Grainger Packed in parallel
60-120 mesh (50%)

60-120 mesh (90%)

- glass

80-120 mesh (6%) -

Bed #4 – glass beads and
Grainger Mixed together
blast sand glass

120-200 mesh (4%)

- sand

Bed #5 – glass beads Grainger 5-10 mesh

Dry test meter Singer Model 803

Fiber-optic UV-Vis Ocean Includes Ocean Optics DT-1000

Model USB2000
spectrometer Optics light source

Test tubes VWR 10 mL For calibration

Reagents
 Yellow/green fluorescent Cole- Used to make up tracer
0298-17
dye Parmer solutions
Materials

Name Company Catalog Number Comments

Equipment

25-40 mesh (50%)

Packed in parallel
Bed #3 – glass beads Grainger
60-120 mesh (50%)
60-120 mesh (90%) - glass

Bed #4 – glass beads and blast sand Grainger 80-120 mesh (6%) - glass Mixed together

120-200 mesh (4%) - sand

Bed #5 – glass beads Grainger 5-10 mesh
Dry test meter Singer Model 803
Fiber-optic UV-Vis spectrometer Ocean Optics Model USB2000 Includes Ocean Optics DT-
Test tubes VWR 10 mL For calibration
Reagents
Yellow/green fluorescent dye Cole-Parmer 0298-17 Used to make up tracer so
LLE PRAC

Liquid-liquid extraction (LLE) is a separation technique used instead of distillation when either: (a)

the relative volatilities of the compounds to be separated are very similar; (b) one or more of the

mixture components are temperature sensitive even near ambient conditions; (c) the distillation

would require a very low pressure or a very high distillate/feed ratio.1The driving force for mass

transfer is the difference in solubility of one material (the solute) in two other immiscible or partially

miscible streams (the feed and the solvent). The feed and solvent streams are mixed and then

separated, allowing the solute to transfer from the feed to the solvent. Normally, this process is

repeated in successive stages using counter-current flow. The solute-rich solvent is called

the extract as it leaves, and the solute-depleted feed is the raffinate. When there is a reasonable

density difference between the feed and solvent streams, extraction can be accomplished using a

vertical column, although in other cases a series of mixing and settling tanks may be used.

In this experiment, the operational goal is to extract isopropanol (IPA, ~10 - 15 wt. %, the solute)

from a mixture of C8-to-C10 hydrocarbons using pure water as solvent. A York-Scheibel type (vertical

mixers and coalescers, one each per physical stage) extraction column is available. Like most

extractors, the overall efficiency (number theoretical stages/physical stages) of this column is quite

low, especially in comparison to many distillation columns. The low efficiencies arise from both slow

mass transfer (two liquid resistances instead of one as in distillation) and often also from

maldistribution of the phases. The effect of agitator speed on both the solute recovery in the extract

and the overall column efficiency will be evaluated.

Principles
Either the a) McCabe-Thiele method, or b) a process simulator (e.g., ASPEN Plus, HYSYS, ChemSep)

may be used to estimate the number of equilibrium (theoretical) stages. The McCabe-Thiele method

is employed on a solvent-free basis, meaning both the solubilities of the solvent in the extract and of

the diluent compound in the raffinate are neglected. A stagewise representation of counter-current

liquid-liquid extraction is shown in Figure 1, where F' is the molar flow rate of the feed

(approximately constant), S' is the molar flow rate of extract (approximately constant), Xf is the mole

fraction of solute in feed, Ys is mole fraction of solute in the solvent, Ye is mole fraction of solute in

solvent in extract stream, and Xr is mole fraction of solute in diluent in the raffinate stream.

Figure 1: Stagewise representation of the extraction process.

At steady state, a material balance on the solute between the feed end of the column and any stage,

n (dotted outline above) leads to the operating line:

(1)

In particular, the equation is satisfied at both ends of the column, so the points (Xf, Ye) and (Xr, Ys) lie

on the line.The equilibrium data in the Appendix can be used in conjunction with this equation

(either graphically or numerically) to step through the column.

The process simulators can do more rigorous stage-to-stage calculations, but still

assuming equilibrium stages. Either the NRTL or UNIQUAC methods (both sets of parameters in the
Appendix) can be used to model the equilibrium relationship. Note that the big advantage of the

simulators is that they DO tell you how much solvent winds up in the extract and how much diluent

winds up in the raffinate. They also can give the exit temperatures for an adiabatic column, or the

heat duty needed to keep the column isothermal.

A York-Scheibel apparatus is shown in Figure 2. Feed can be introduced at the bottom (11 stages) or

at the middle of the column (6 stages).

Figure 2: York-Scheibel liquid-liquid extraction apparatus.

The extraction unit consists of a 2" I.D. Pyrex column, with 11 extraction stages, each consisting of a

one-inch mixing section and a four-inch wire mesh packing (coalescing) section. The column is

mechanically agitated by paddlewheel-type (Rushton turbine) agitators. A variable speed motor,

with a control knob and digital readout on the control panel, controls the speed of the agitator.

Rotameters on the feed and solvent inlets are used to measure those flow rates. Flow rates of the

extract and raffinate can be measured with a graduated cylinder and stopwatch.
The following equations relate the rotameter readings to volumetric flow rates (the flows can also

be checked with a graduated cylinder):

(2)

(3)

where Ff is the feed flow rate (~10 wt. % IPA) in mL/min,Rf is the feed rotameter reading,Fs is the

solvent flow rate in ml/min, andRs is the solvent rotameter reading.

Procedure

In this experiment, the properties of n-nonane are a good approximation to those of the

hydrocarbon mixture for equilibrium data purposes. The ternary system water/isopropanol/n-

nonane exhibits Type I equilibrium behavior (there is some composition range over which phase

splitting will not take place) at room temperature. The equilibrium data for this system can be found

in the Appendix.

1. Operating the York-Scheibel Column

1. Fill the extractor with hydrocarbon mixture /IPA feed (if necessary) and bleed air from the
feed line. Turn off the feed flow.
2. Start the mixer and keep the agitator speed constant.
3. Open the solvent, feed, extract, and raffinate ball valves; and start the flow of solvent
(water) into the column.
4. If no interface is present between the solvent entrance and the raffinate exit, let the
dispersed phase rise and form the upper interface.
5. When the upper interface forms, (re-)start the feed flow.
6. Control the interface level by adjusting the height of the inverted U on the extract line from
the bottom of the tower.The upper interface level adjustment (inverted U) is sensitive.
Movements of a fraction of an inch are often sufficient.
7. Periodically check the raffinate stream for steady state using gas chromatography. The gas
chromotagraph will separate and quantify the components in the sample.
8. Use a hydrometer to measure the specific gravity of the extract stream and determine the
composition. (This also can help confirm that steady state has been reached.) The extract
stream composition vs. specific gravity tables can be found in Perry's Handbook.3 This data
can be used to interpolate for weight percent of IPA.
9. Use a hydrometer to measure the specific gravity of the feed and raffinate for use in
subsequent calculations.
2. Shutdown Procedure
1. Once the experiments are complete, turn off the agitator and main power switch.
2. Close the feed and solvent ball valves, leaving the raffinate and extract ball valves open.
Results

Figures 3 and 4 show results when both the agitation and feed flow rates were varied over a wide

range. The overall efficiency and recovery increase before becoming asymptotic, which is fairly

typical of liquid-liquid extractors that are not at or near flooding. At near flooding conditions, the

overall efficiency and recovery are expected to sharply decrease. Note that, unlike distillation,

flooding can take place in liquid-liquid extraction at either high solvent or high feed rates (or

ratios).1 In this experiment, the lighter organic phase is also the dispersed (droplet) phase, so at high

feed rates it is expected that the droplets coalescence prior to flooding, leading to lower rates of

mass transfer and, therefore, lower recoveries and efficiencies. At high solvent rates the droplets

should remain small, so it is expected that the recovery and efficiency remain high until very near

the flooding point.

Figure 3: Percent recovery of IPA from hydrocarbon mixture into water, for a York-Scheibel

column, 11 stages, 16 - 18 mol% IPA in Isopar E (feed), S/F (molar) = 1.5.

Figure 4: Percent overall stage efficiency for IPA extraction using a York-Scheibel column, 11

stages, 16 - 18 mol% IPA in hydrocarbon mixture (feed), S/F (molar) = 1.5.

As seen from Figures 3 and 4, increasing the agitation rate increases both the recovery and overall

efficiency. This is because with greater power input the droplets of the dispersed phase are smaller -

the observed dependence is roughly inverse with respect to agitator speed.4 The "a" parameter

(interfacial area/total volume) that appears in mass transfer correlations and fundamental mass flux

equations can be written as follows for uniform-size spherical droplets:

a = 6 ε/d (4)

where ε is the volume fraction of the dispersed phase. While ε can increase with an increase in

either phase's superficial velocity, its changes are usually less marked than the change of diameter

with respect to speed. So (usually) the more speed, the more interfacial area, leading to faster mass

transfer.

The exception to the above discussion is at very high speeds, which were not reached in Figs. 3 and

4, where the two phases are so well-mixed that if the interfacial tension between them is low,

emulsification will take place. The formation of an emulsion negatively impacts recovery and

efficiency because the phases can no longer separate cleanly to move up or down to the next

physical stage.Emulsification is a problem in many liquid-liquid extractions and where it cannot be

limited a series of mixer and settler vessels in series is often preferred to column-type designs such

as sieve trays or York-Scheibel units.

Applications and Summary

Liquid-liquid extraction (LLE) is an alternative to distillation which relies upon solvent-feed

immiscibility (or slight miscibility) and favorable solute partition coefficients to attain high solute

recoveries in a solvent phase at as low a solvent/feed ratio as practical. Although the range of flows

(the "turndown") over which LLE will be effective is often limited, and while stage efficiencies are
low such that phase equilibrium is not attained, certain mixtures just cannot be separated using

other methods in a continuous countercurrent process. Mathematical analysis of the equilibrium

operation of such extractors follows a familiar McCabe-Thiele-type procedure (although reflux is

often lacking, so only one operating line). The non-equilibrium ("rate-based") analysis of LLEs is

complex and depends strongly on the relative velocity between the two phases (the "slip velocity"),

bubble size, and dispersed phase fraction, all of which can be observed but are difficult to predict.

To perfectly describe the hydraulics and mass transfer of a typical LLE is beyond the capability of

even the most sophisticated process simulators, at present. Therefore, design of industrial units still

relies on scale-up from pilot-plant-type units, such as that which was tested in this experiment.

Normally the engineer attempts to duplicate key descriptors such as the "a" parameter, solvent/feed

ratio, total agitator power input/volume, feed location and number of physical stages to keep the

stage efficiency and recovery constant during scale-up. Even so, scale-up is an inexact science, and

impurities, which alter the interfacial tension, can greatly impact the performance of real systems.

The more factors that are held constant, the more likely the scale-up will be successful.

There are many different LLE contactors: a series of mixers - settler vessels, structured packings

similar to those used in absorbers and distillation columns, sieve tray columns, rotating disk

contactors (similar to the York-Scheibel, but with baffles instead of mesh), Kuhni contactors (a

combination of rotating disk and sieve trays), and Podbielniak contactors ("Pods"), where the flow is

radial and centrifugal force is used to enhance liquid phase separation.5

A classic example of industrial LLE is the separation of acetic acid from water using ethyl ether or

ethyl acetate;6 it is preferred over distillation at lower acetic acid concentrations. Possibly the

biggest volume LLE process is that of propane deasphalting, which is used to refine lubricating oils in
refineries at near-supercritical conditions.1 However, most applications are found in the production

of specialty chemicals and in pharmaceutical industries, ranging from citric acid extraction from

fermentation broth to purification of antibiotics and protein purifications.1In these cases, a wide

variety of oxygenated organic solvents or two-phase aqueous systems (with one phase being mostly

water and the other aqueous dissolved salts and polymers) are utilized. For the latter, a typical

polymer system is poly(ethylene glycol)/dextran with NaCl and Na2SO4 as salts. Applications include

red blood cell separation and extraction of the phophofructokinase enzyme from S. cerevisiae.7

Appendix – Equilibrium Data8

Experimental Tie Lines in Mole Percent at 25 °C

Specific Model Parameters in Kelvin

UNIQUAC NRTL (a = 0.2)

I J AIJ AJI AIJ AJI

1 2 -186.05 104.6 814.26 -468.11

1 3 361.91 621.82 3151 1367.4

 2 3 -126.43 311.7 581.79 -25.91

R1 = 0.92 R2 = 2.7792 R3 = 6.523

Q1 = 1.4 Q2 = 2.508 Q3 = 5.476

Mean Deviation between Calculate and Experimental Concentrations in Mol. %

UNIQUAC (specific parameters) 1.4

NRTL (specific parameters) 0.54

UNIQUAC (default parameters) 1.68