Absorption Column Design

Absorption column design is a cornerstone of chemical engineering, essential for processes like removing pollutants from flue gases or recovering valuable chemicals from process streams. Whether you're working with packed or tray columns, mastering this design ensures efficient mass transfer and cost-effective operation.

Fundamentals of Gas Absorption and Column Types

Gas absorption is a unit operation where one or more components (solutes) are transferred from a gas phase into a liquid phase (solvent). The heart of this process is the absorber column, where intimate contact between the gas and liquid is facilitated. You'll primarily encounter two column internals: packed columns, which are filled with inert materials to maximize surface area, and tray columns, which use a series of plates or stages to create distinct gas-liquid contact points. The choice between packed and tray designs hinges on factors like pressure drop, fouling tendency, and the required separation efficiency. For both types, the core design methodology revolves around material balances and mass transfer principles to size the column correctly.

The design process begins by defining the feed conditions: the inlet gas flow rate $G$ and composition $y_1$, and the desired outlet gas composition $y_2$. You also specify the inlet liquid solvent flow rate $L$ and its composition $x_2$. The goal is to determine the column height needed to achieve the separation. This requires building a model that relates the changing compositions of the gas and liquid phases as they move counter-currently through the column. The equilibrium relationship between the solute in the gas and liquid phases, often expressed as $y^{\ast }=f(x)$, is a critical piece of data, as it defines the theoretical limit of absorption.

Constructing the Operating Line from Material Balances

For a continuous, steady-state absorber, a material balance on the solute provides the relationship between gas and liquid compositions at any point in the column. This balance yields the operating line, which represents the actual, working compositions of the two phases. For a column where the gas and liquid flow rates ($G$ and $L$) can be assumed constant (a valid assumption for dilute systems), the operating line is straight. Its equation is derived from a balance around the top of the column:

$$y=\frac{L}{G}(x-x_2)+y_2$$

Here, $y$ and $x$ are the mole fractions of solute in the gas and liquid, respectively. The slope of this line is $L/G$, the liquid-to-gas ratio. Graphically, you plot this line on a $y$ vs. $x$ diagram. The operating line must always lie above the equilibrium curve for absorption to be thermodynamically feasible. The vertical distance between the operating line and the equilibrium curve at any point, $(y-y^{\ast })$, is the driving force for mass transfer. In concentrated systems, where solute transfer significantly changes total flow rates, $G$ and $L$ are not constant. Here, you must use solute-free flow rates ($G^{\prime }$ and $L^{\prime }$) and mole ratios ($Y$ and $X$) to construct a curved operating line: $Y=\frac{L^{\prime }}{G^{\prime }}(X-X_2)+Y_2$.

Worked Example (Dilute System): An air stream containing 5.0 mol% SO₂ ($y_1=0.05$) is fed to a packed column at 100 kmol/h. The target is 90% recovery, so the outlet gas composition is $y_2=0.005$. Pure water ($x_2=0$) is used at a rate of 300 kmol/h. The equilibrium is given by $y^{\ast }=25x$.

1. Calculate slope: $L/G=300/100=3$.
2. Operating line equation: $y=3(x-0)+0.005$ or $y=3x+0.005$.
3. The endpoint at the column bottom (where gas enters) is $(x_1,y_1)$. Using the material balance: $G(y_1-y_2)=L(x_1-x_2)$, we find $x_1=(100(0.05-0.005))/300=0.015$. So the line runs from point $(0,0.005)$ at the top to $(0.015,0.05)$ at the bottom.

Determining the Minimum Liquid Flow Rate

The minimum liquid rate $L_{min}$ is the smallest solvent flow rate that can theoretically achieve the desired separation. It corresponds to an operating line that just touches the equilibrium curve at the point where the driving force becomes zero—a pinch point. At $L_{min}$, the required column height becomes infinite because the driving force for mass transfer vanishes at the pinch. You can find $L_{min}$ graphically by drawing the steepest line from the point $(x_2,y_2)$ that touches the equilibrium curve. For a dilute system with a straight equilibrium line $y^{\ast }=mx$, it is calculated directly:

$${\left(\frac{L}{G}\right)}_{min}=\frac{y_1-y_2}{\frac{y_1}{m}-x_2}$$

In practice, you always operate at a liquid rate greater than the minimum. A typical rule of thumb is $L=1.2$ to $2.0$ times $L_{min}$. This provides an economic balance between column height (capital cost) and liquid pumping costs (operating cost). Choosing this multiplier is one of your key design decisions.

Applying to the Worked Example: With equilibrium $y^{\ast }=25x$, $m=25$, and $x_2=0$. $$(L/G{)}_{min}=\frac{0.05-0.005}{(0.05/25)-0}=\frac{0.045}{0.002}=22.5$$ Thus, $L_{min}=22.5\times G=2250$ kmol/h. Our actual $L=300$ kmol/h gives $L/L_{min}\approx 1.33$, which is within the practical range. This example illustrates that the equilibrium constant dramatically affects the minimum rate; a high $m$ value (low solubility) demands a very high liquid flow.

Calculating Transfer Units and Column Height

Since driving force changes along the column, we integrate its effect using the concepts of Number of Transfer Units (NTU) and Height of a Transfer Unit (HTU). The NTU, specifically $N_{OG}$ for gas-phase control, is a measure of the difficulty of separation. It is calculated by integrating the change in gas composition divided by the driving force:

$$N_{OG}={\int }_{y_2}^{y_1}\frac{dy}{y-y^{\ast }}$$

For a dilute system with straight operating and equilibrium lines, this integral has an analytical solution:

$$N_{OG}=\frac{y_1-y_2}{\Delta y_{lm}}$$ where $\Delta y_{lm}$ is the log-mean driving force: $$\Delta y_{lm}=\frac{(y_1-y_1^{\ast })-(y_2-y_2^{\ast })}{\ln \left(\frac{y_1-y_1^{\ast }}{y_2-y_2^{\ast }}\right)}$$

The HTU, or $H_{OG}$, is a measure of packing efficiency—the height of column required to achieve one transfer unit. It is obtained from vendor correlations or experiments and is a function of packing type, flow rates, and fluid properties. The overall packed height $Z$ is then simply:

$$Z=H_{OG}\times N_{OG}$$

For tray columns, you use the Number of Theoretical Stages (NTS) and stage efficiency instead of NTU and HTU. The McCabe-Thiele graphical stepping method between the operating and equilibrium lines is the standard approach to find NTS for both dilute and concentrated systems.

Approaches for Concentrated Systems

In concentrated systems, the assumptions of constant flow rates and constant $H_{OG}$ often break down. You must shift to using solute-free coordinates ($Y$, $X$, $G^{\prime }$, $L^{\prime }$). The operating line becomes straight in $Y-X$ coordinates, but the equilibrium curve may not be. The $N_{OG}$ calculation now requires numerical integration:

$$N_{OG}={\int }_{Y_2}^{Y_1}\frac{dY}{Y-Y^{\ast }}$$

Furthermore, $H_{OG}$ can vary significantly along the column due to changes in mass transfer coefficients with concentration and flow rate. A common design strategy is to divide the column into several segments, assuming constant conditions within each, and sum the heights: $Z=\sum (H_{OG}\cdot N_{OG}{)}_{segment}$. This approach, while more computationally intensive, yields a more accurate design for processes like scrubbing concentrated acid gases.

Common Pitfalls

1. Using Dilute Method Assumptions for Concentrated Cases: Applying the simple log-mean driving force formula to a system where the solute removal exceeds ~10% of the feed can lead to significant underestimation of required height. Correction: Always check if total flow rates change by more than 10%. If they do, switch to solute-free coordinates and perform numerical integration for $N_{OG}$.

1. Ignoring the Minimum Liquid Rate in Economic Optimization: Selecting a liquid rate based solely on intuition can lead to either an excessively tall column (if $L$ is too close to $L_{min}$) or prohibitive pumping costs (if $L$ is too high). Correction: Systematically calculate $L_{min}$ and perform a short cost analysis around the typical $1.2-2.0\times L_{min}$ range to find the optimum.

1. Confusing HTU with Physical Packing Height: $H_{OG}$ is not a physical property of the packing; it's a performance parameter that depends on operating conditions. Assuming a single $H_{OG}$ value is valid across different gas and liquid rates is a mistake. Correction: Use reliable, experimentally derived correlations for mass transfer coefficients to calculate $H_{OG}$ specifically for your design's flow rates and fluid properties.

1. Misplacing the Operating Line on the $y-x$ Diagram: Drawing the operating line below the equilibrium curve makes the process thermodynamically impossible (it would represent desorption, not absorption). Correction: Double-check your material balance and endpoints. Remember, for absorption, the operating line must always be above the equilibrium curve.

Summary

• The operating line, derived from a material balance, defines the actual gas and liquid compositions in the column and must lie above the equilibrium curve for absorption.
• The minimum liquid rate $L_{min}$ is a theoretical limit; practical design uses a multiplier of 1.2 to 2.0 times this value to balance capital and operating costs.
• Column height is determined by the product of Number of Transfer Units (NTU), which quantifies separation difficulty, and Height of a Transfer Unit (HTU), which measures packing efficiency: $Z=H_{OG}\times N_{OG}$.
• Dilute systems allow for simplified methods with constant flow rates and the log-mean driving force, while concentrated systems require the use of solute-free coordinates and often numerical integration.
• Always validate assumptions about flow constancy and ensure the correct graphical construction to avoid thermodynamically infeasible designs.
• For tray columns, the McCabe-Thiele method is used to determine the number of theoretical stages, which is analogous to the NTU approach for packed columns.