1 Introduction

Distillation columns are vital components for the Chemical Process Industry (CPI). As a result, they must be designed correctly and in the most cost-effective way possible. The interior of the distillation columns is arguably the most important part of the column as they are difficult to access after start-up. An internal breakdown in the column can impact the entire process. The internal components of the columns can either be plates or packing, which each have their characteristics that make one option more appropriate than the other for the desired separation application (Pilling & Holden, 2009).

2 Plate vs Packing

Various factors need to be considered when deciding on the internals of the distillation column such as the flowrates of the vapour and liquid phases, the reason for separation, the desired product characteristics and the physical properties of the vapour and liquid phases. Most of this data can be acquired from process simulation software such as Aspen Plus. This information, together with physical property models, can tell the user the reflux ratio, the optimum number of stages, the column’s economics and the purity of the product.

The following considerations need to be taken into account when deciding on the column internals: conditions and compositions of the streams entering and exiting the column, the chemical reactivity between the components being separated and the materials of construction, the temperature ranges, the tendency to foul, the corrosivity and the risk of contamination. These all affect the design choices for materials used in the column internals and are design constraints for the column internals (Pilling & Summers, 2012).

The final decision between a plate or packed column for a particular separation can only be made with complete certainty by costing each design. However, this will not always be beneficial, useful or necessary. The decision can be made by using previous experience as well as considering the pros and cons as outlined below (Sinnott, 2005, 589):

  1. Plate columns can be designed to accommodate a broader range of vapour and liquid flow rates, whereas packed columns tend to be appropriate for exceptionally low liquid flow rates. Plates are typically used with liquid flow rates more than 30 m3 m-2 h-1.
  2. Cooling in plate columns is much more easily designed by installing coils on the plates.
  3. The addition of side streams is much easier to install in plate columns.
  4. Packed columns are more appropriate for accommodating foaming systems.
  5. Maintenance of plate columns is much more accessible by installing maintenance holes on the plates to maintain fouling. However, with smaller diameter columns, it can be more cost-effective to use packing and replace it once it becomes fouled.
  6. When considering corrosive substances, a packed column tends to be always more cost-effective than a column with trays.
  7. The liquid holdup in a packed column is significantly less than in a plate column. This is an important safety consideration when the amount of toxic or flammable liquid needs to be minimised.
  8. Packing should be considered in vacuum columns as the pressure drop per equilibrium stage can be smaller for columns that contain packing than with plates.
  9. Packing is more appropriate for small diameter columns, approximately less than 0.6 m where plates can be more of a hindrance to install.

When considering the choice between plates and packing, two types of packing options exist: random packing and structured packing. Packed columns tend to be used for distillation, gas absorption and liquid-liquid flow. The flow can be counter-current; however, in some gas-absorption columns, co-current flow is used. The types of packing that exist include Raschig rings, Pall rings, Berl saddle ceramic rings, Intalox saddle ceramic rings, metal hypac and super Intalox ceramic rings. Three different types of plates exist: sieve plate (perforated plate), bubble-cup plates and valve plates (floating cap plates). When considering the type of plate to be used, factors considered include the cost, capacity, operating range, efficiency and pressure drop (Pilling & Holden, 2009).

3 Plate

Distillations plates, also called trays, allow for an amount of liquid holdup. This is to allow the vapour flow to encounter the liquid and allows for vapour-liquid mass transfer which is necessary for separation to occur.

The most popular type of plate contactor in distillation (and absorption) columns are cross-flow plates as depicted in the image below where the liquid flows in downcomers in the intervals between plates.

Figure 1: Example of a cross-section of a distillation column that depicts cross-flow (Hebert & Sandford, 2016)

Other types of trays exist which do not have downcomers and are termed non-crossflow plates. Instead, the liquid flows through large holes in the plates, and these types of plates are typically called shower plates. Furthermore, other kinds of non-crossflow plates exist that are designed for special cases, for example, in instances where a low-pressure drop is necessary.

There are three main types of cross-flow plates used in vapour-liquid mass transfer.

  1. Sieve Plates
  2. Bubble cap plates
  3. Valve plates which have three different subcategories: moving valve plates, fixed valve plates and enhanced fixed valve plates.

3.1 Ideal Tray Design

Figure 2: Tray configuration with a sieve plate (Pilling & Holden, 2009).

The deck in a typical cross-flow plate contains an inlet region where the liquid is supplied as depicted above. The downcomer is the outlet. Plates need to have holes for vapour to flow through. The perforations make up approximately 5 % to 15 % of the tray area. The spacing between the trays is outlined below. The downcomer area consists of approximately 5 % to 30 % of the cross-sectional area of the column. This is dependant on the liquid load.

The weir regulates the amount of liquid build-up on the plate surface and any excess liquid froths over into the downcomer. The weir height usually is 50 mm (Pilling & Holden, 2009).

3.2 Operating Conditions

The operating conditions of plates are restricted to minimise weeping at low throughputs and flooding (also termed entrainment) at high throughputs. Weeping happens when the pressure drop on the vapour-side of the plate deck is too small to bear the liquid on top of the plate. As a result of this, the liquid trickles through the perforations. A dry-plate pressure drop less than 12 mm H2O increases the likelihood that weeping will occur. Generally, 20 % of weeping results in a decrease of 10 % efficiency.

On the other hand, flooding happens when very large vapour velocities transport liquid droplets to the tray above. This is termed jet flooding. Entrainment is more damaging to the ability of the tray to produce the desired separations as it causes back-mixing of the liquids. Generally, 10% of entrainment results in a decrease of 10 % efficiency.

The flooding of the inlets or backup may result in the flooding of the downcomers. The liquid head determines the height of the froth in the downcomer at the plate inlet or downcomer outlet, pressure drop across the plate as well as friction losses. A froth level rise in the downcomer is due to an increasing pressure drop across the plate and an increase in the flow rate of the liquid. When the amount of froth surpasses the downcomer height, the tray above will flood. This can all be prevented or minimised by widening the spacing between trays, reducing the pressure drop across the trays, enlarging the downcomer clearance or lowering the outlet weir height. Typically, extending the length of the downcomer is not effectual as the losses due to friction in the downcomer are insignificant (Pilling & Holden, 2009).

3.3 Sieve Plate (Perforated Plate)

This is the most basic version of a cross-flow plate. The vapour flows through the holes in the plate whilst the liquid is retained on the tray by the vapour flow. However, there is no adhesive material in the holes to prevent liquid from trickling out and therefore “weeping” can occur through the perforations. As a result, this can decrease the plate efficiency. The hole size can vary, but bigger perforations are more common. The advantages and disadvantages of sieve plates are listed below (Separation Technology, 2012).

Figure 3: Illustration of sieve tray perforation, top view and side view (Sinnott, 2005: 559).


  • Low entrainment.
  • Inexpensive and easy to install.
  • Low running costs.
  • Reduced fouling.


  • Not as adaptable to varying loads when compared to other types of tray designs.
  • Increased occurrences of weeping.

3.4 Bubble-Cap Plates

Bubble cap plates are pictured below and are the most traditional type of cross-flow tray type. Many versions exist, and they can be customised for specific design requirements. The most prominent feature of the bubble-cap plate are the risers which are small pipes where the vapour flows through that are covered by a cap. This feature ensures that a certain amount of liquid remains on the plate regardless of the vapour flow rates (Separation Technology, 2012). This aids in reducing weeping.


  • Most flexible to loading and can accommodate both high and low liquid and vapour flow rates.
  • Can provide excellent turndown.
  • Reduced weeping.


  • High entrainment.
  • High fouling tendency.
  • High-pressure drop.
  • Expensive

Figure 4: Depiction of a Bubble cap tray constructed of stainless steel (Separation Technology, 2012).

3.5 Valve Plates

These have become the most popular tray choice in mass transfer applications in the CPI Industries.

The valve tray consists of a sieve plate with large diameter perforations that are covered by movable flaps which rise as the vapour flowrate increases. The valve plate design has an improved efficiency when compared to sieve plates. This is because the valves are closed at lower flow rates, and weeping is minimised.


  • Good contact between the liquid and vapour phases for enhanced mass transfer.
  • Higher capacity.
  • More flexibility when compared to sieve trays.
  • Can accommodate higher loadings.
  • Lower pressure drop than bubble cap plates.

Two types of valve plates exist: Fixed valves and moving valves. Fixed valves remain open whereas moving valves are opened when vapour flows through the perforations and raises the valve cover.

3.5.1 Moving Valves

The moving valve became a standard addition to the process industry after its invention. The main advantage it offered was that it allowed for movement that could accommodate fluctuating vapour flowrates. The maximum valve movement is restricted either by the legs (below left) or by the cage height in the design (below right)

The full advantage of the moving valve which provides a range of motion is that maximum movement of the valve at full height can be attained by allowing high vapour flow rates. Moving valves have a nominal turndown (ratio of maximum vapour flowrate to minimum vapour flowrate of 4:1). Moving valves produce an incredible turndown ratio. This is because at high vapour flowrates the valve operates at maximum height whereas at low vapour flowrates the valve partially closes due to low-pressure drop across each plate. This results in the weeping of the liquid and poor column efficiency. However, due to the moving valves, they partially close and minimise weeping.

Turndown ratios of 10:1 are feasible when the moving valves are designed well by installing valves in rows of alternating thickness. This allows for control over which valves open and closes first when vapour flowrates fluctuate (Hebert & Sandford, 2016).

The disadvantage of moving valves is that fouling can cause the extruding portion to stick to the tray deck. Also, unexpected operating conditions such as a sudden high vapour flowrate can cause the extruding portion to “pop” off, thereby resulting in parts of the deck having a sieve plate design. That defeats the purpose of the valve tray and causes all the disadvantages of sieve plates.

Figure 6: Moving valves. Open position with a simple extruding cover (left) and caged assembly design (right) lower (Hebert & Sandford, 2016)

3.5.2 Fixed Valves

The most basic fixed valves are formed from metal sheets that have been punched. They have the advantage of no moving components. As a result, they do not stick, pop, erode or corrode. Furthermore, the larger open perforation prevents fouling. The valves can be customised to have varying heights by varying the cut and geometry of the perforation.

Figure 7: Fixed valve consisting of an extruded perforation with a smaller diameter (left) and larger diameter (right) holes (Hebert & Sandford, 2016).

Moreover, the hydraulic capacity is inversely proportional to the size of the valve perforation. The smaller the opening, the larger the capacity. Additionally, the smaller the net rise, (as depicted in the figure below) the less entrainment is anticipated. Another advantage of fixed valves is they are resistant to process upsets due to mechanical reliability (Hebert & Sandford, 2016).

Figure 8: Net rise is the difference between the level position to the full height of the extruded cover (Hebert & Sandford, 2016).

3.5.3 Enhanced Fixed Valves

Thicker metals do not allow for as much stretching and extension and, as a result, net rise movements are restricted. These limitations can be tackled by a fixed valve plate that integrates large perforations in the deck with separate cover pieces that overlap the opening as illustrated below.

Figure 9: Enhanced Fixed valves with cover piece that overlaps the perforation for thicker metal sheets that form the deck (Hebert & Sandford, 2016).

The tapered cut minimises any stagnant liquid on the deck, and this design also minimises any weeping. Due to the large hole and fixed components, fouling is also suppressed. Despite the shape not being able to be formed by extrusion and having to be mechanically fit together, the design is still more mechanically robust than moving valves (Hebert & Sandford, 2016).

3.6 Deciding on the right plate type

The main points to consider when deciding on the right plate type between valve plates, sieve plates, and bubble-cap plates are pressure drop, efficiency, operating range, capacity and price. However, when trying to minimise the cost (and excluding bubble cap trays), the diagram below illustrates the level to which various plates should be considered when fouling plays a factor in the column design.

Figure 10: Levels of fouling resistance and relevant plate type that is recommended (Hebert & Sandford, 2016).

Pressure Drop
Pressure drop is a significant factor to consider, especially in vacuum columns. The pressure drop will depend on the specifics of the plate design. However, sieve plates produce the smallest pressure drop, with valve plates causing a slightly bigger pressure drop and bubble-cap plates producing the largest pressure drop (Sinnott, 2005: 560).
The efficiency used to calculate the plate efficiency is called the Murphree efficiency and is the same for all three types when considering the design flow range (Zuiderweg et al., 1960). Therefore, this point is insignificant when used as a determining factor between the different plate designs.
Operating Range
This is the most important point when considering the three types. The operating range refers to the span of liquid and vapour flow rates over which the tray can operate well and stably.

Some variability should be included in this operating range during times when the flow rate will drastically differ out of normal operating range. For example, start-up and shutdown. The turndown ratio refers to the ratio of maximum flow rate to the minimum flow rate.

Bubble cap plates have a positive liquid seal and work well at exceedingly small vapour flowrates. Sieve plates depend on the vapour flow to maintain the liquid on the tray and work poorly at a low vapour flow rate. However, with enhanced designing, sieve plates can operate well.

Valve plates provide an improved operating range than sieve plates and are cheaper than bubble caps.
When evaluating the capacity factor, which is defined as the diameter of a column necessary for a given flow rate, the capacity is rated as sieve > valve > bubble cap.


Bubble caps are the most expensive of the three. It should be noted that cost is also based on the material of construction used.

3.7 Perforated Area

The amount of area that is available for the perforation is dependent on the area that will already be taken up by the structural support and the calming zones. Calming zones are regions of the tray at the inlet and outlet sides of each plate. Standard dimensions exist for the width of each zone. For diameters less than 1.5 m, a zone of 75 mm is recommended, and for diameters greater than 1.5 m, a 100 mmm zone is necessary. The support ring for the plates is usually between 50 mm to 75 mm; it should be noted that the support ring should not block the downcomer area (Sinnott, 2005: 572).

3.8 Plate spacing

The height of the entire distillation column is dependent on the plate spacing. The typical spacing of trays usually ranges from 0.15 m to 1 m. This value is contingent on the operating conditions as well as the distillation column diameter. The smaller the diameter, the closer the spacing. Closer spacing is also used when the headroom of the distillation column needs to be compact. In distillation towers with a diameter over 1 m, plate spacings between 0.3 to 0.6 m are used with a spacing of 0.5 m used as an initial approximation in the design iteration process. Larger spacings are necessary when feed and side streams are located between plates (Sinnott, 2005: 573).

4 Packing

The main stipulations that need to be met for packing are that it should meet the following conditions:

  1. Allow for a large surface area for the interface between liquid and gas.
  2. The packing should be configured in a way that allows the least resistance to the gas flow, preferably by having an open arrangement.
  3. It should enhance steady liquid flow on the exterior of the packing.
  4. It should enhance a steady gas and vapour distribution across the cross-section of the distillation column.

Numerous types of packing exist in various shapes and sizes to meet various demands. However, they can be split into two main categories:

  1. Packing with a consistent layout, termed structured packing. Examples of structured packing include stacked rings, proprietary structured packing, as well as grid packing.
  2. Random packing in the form of proprietary shapes, saddles and rings. Random packing is arbitrarily unloaded into the distillation column and takes up a random order within the column.

    Figure 11: Different types of packing options available (a) Raschig rings (b) Pall rings (c) Berl saddle ceramic (d) Intalox saddle ceramic (e) Metal Hypac (f) Ceramic, super Intalox (Sinnott, 2005, 590)

Packing with a consistent layout, for example, grids have an advantage in that they provide a much larger surface area for substances to flow through. This allows higher flow rates of gases where a minimal pressure drop is important such as in cooling towers (Sinnott, 2005, 590).

4.1 Structured Packing

Structured packing usually consists of alternating layers of mesh, gauze or fine corrugated sheets as illustrated in the image below. These are typically made of a wide range of materials including ceramics, graphite, plastic or metal alloys. The material is arranged with a regular geometry to obtain a large surface area with a large void fraction. Structured packing usually results in smaller pressure drops and overall improved efficiency with shorter bed heights when compared to random packing. However, they are typically more costly to construct, and the time of construction is also lengthier (Pilling & Holden, 2009).

Figure 12: Structured packing (Pilling & Holden, 2009).

The main advantage of structured packing when compared to random packing is that the separation efficiency, typically expressed in the height of packing equivalent to a theoretical equilibrium plate (HETP) is usually less than 0.5 m and the low-pressure drop of 100 Pa/m.

Structured packing is more favoured than random packing for the following cases:

  1. In column revamps: in scenarios to enhance the capacity of the column and minimise the reflux ratio.
  2. High vacuum distillations.
  3. For separations that need numerous stages, for example, in the separation of isotopes due to the more favourable HETP.

Structured packing is typically more expensive than random packing. However, this comes at the advantage of an improved efficiency when compared to random packing.

Geometry and materials

Due to the slim size of structured packing and the large amount of surface area, there is negligible corrosion allowance. As a result, the material of construction is a crucial decision choice. Wire gauze packing made from woven metal cloth is the best choice for separations that require a high efficiency with low liquid loadings. However, sheet metal packing that has both a high surface area as well as a superior quality liquid distributor is also a good alternative.

The surface area available in structured packing can span from 40 m2/m3 to 900 m2/m3. Grid packing typically lies within 40 m2/m3 to 90 m2/m3 and has a robust design with a large open area and has a high capacity. Grid packing is typically used when fouling and hydraulic load stresses are high. Structured packing greater than 500 m2/m3 is typically used in air separation and the manufacture of refined chemicals (Pilling & Holden, 2009).

Structured packing is typically perforated and textured, which assists the liquid distributors and can be designed to accommodate two liquid phases. This requires careful design as the liquid phases can segregate on the surface of the packing resulting in reduced efficiency.

When many stages are required, structured packing is a good alternative due to its high efficiency (HETP).  However, it is important to weigh the design choice between structured packing and plates because plates are preferable when the liquid flow rate is high, and the pressure drop is high. Furthermore, structured packing needs collectors and distributors, which can require a taller column than the equivalent column design that uses plates in the column internals.

When separating components where thermal degradation is a concern, structured packing provides two advantages. Firstly, the liquid holdup is low in columns with structured packing. This results in short residence times of the components in the column leading to a lower risk of degradation. Secondly, structured packing has a smaller pressure drop, and this means that the bottom of the column can be operated at lower pressures and temperatures, which also lowers the risk of thermal degradation. The execution and efficiency of structured packing is a result of the packing geometry and liquid distribution (Pilling & Holden, 2009).

4.2 Random Packing

4.2.1 Raschig Rings

Raschig Rings were the first kind of packing used in mass transfer processes, and they consist of a hollow cylindrical tube with a height to diameter ratio of 1:1. They display an exceptional strength to weight ratio and are much more resistant to fouling than other types of packing (The Pall Ring Company, 2020).

4.2.2 Pall rings

Pall rings are a progression from the Raschig rings and have the same cylindrical shape and ratio. However, they have two rows of punched out holes with webs from the centre which drastically enhances the execution of the packing with regards to the pressure drop, throughput and efficiency (The Pall Ring Company, 2020).

Pressure drop versus capacity

They are most appropriate when a low-pressure drop is required in high capacity scenarios and can be customised in a wide variety of sizes for optimum performance for separation. They create a large amount of randomness, and the high surface area to volume ratio enhances mass transfer. Pall rings provide a consistent ratio of free and blocked pathways regardless of the orientation.

Mechanical Strength

The internal structure with the web-like design results in a mechanically robust type of packing that is appropriate for use in packed beds that are deep.

Metal pall rings can be made of stainless steel alloys or other metals and provide the advantage that they are capable of handling higher temperature separations and are highly resistant to fouling and are also mechanically robust. They are typically used in separations with H2S, NH3 and SO2 as well as absorption and stripping, steam stripping, quench towers and direct contact cooling.

4.2.3 Berl Saddles

The ceramic Berl saddles are an improvement from both the Pall rings and Raschig rings with regards to an enhanced flow distribution. The shape of the saddles has both the external and internal surfaces exposed, and the distribution of vapour-liquid is the same on both sides. When comparing Berl saddles to the previous two types of packing, they produce a lower pressure on the surrounding column walls (Ultimo Engineers, sa).

Berl saddles typically consist of chemical grade porcelain which is chemically resistant to many substances apart from Hydrogen Fluoride and strong alkalis. It is most appropriate for sulphuric acid applications.

4.2.4 Proprietary shapes

The Hypac and Intalox rings are improvements from the Pall ring and Berl saddles. Furthermore, Intalox saddles, Hypac and super Intalox are all trademarked designs by the company Norton Chemical Process Products Ltd (Sinnott, 2005).

4.3 Overall Evaluation

Packing can be customised in various materials including carbon, plastic, ceramic and metals. However, rings with thinner walls make the packing more efficient. Thin packing is easiest to manufacture when the material of choice is metal or plastic rings. Furthermore, the composition of the packing is determined by the operating temperature and the properties of the fluid: fouling, corrosivity etc.

Ceramic packing is preferred for corrosive fluids but is inappropriate for fluids with high alkalinity. Furthermore, plastic packing can react with some organic fluids and can be restricted by temperature rise and is therefore not typically used in distillation columns. Due to the robust nature of metal packing, it is preferred in situations where the distillation may be unstable instead of ceramic packing, which is much more easily broken.

When considering the cost-effectiveness of the materials, Raschig rings are less expensive (price per unit volume) when compared to Pall rings or saddles. However, they are less efficient. When evaluating the cost-effectiveness of Raschig rings, the column works out to be more expensive when taking into consideration the whole column. When designing new columns, Pall rings and Berl or Intalox saddles are typically used (Sinnott, 2005,590 – 592).

4.4 Packing Size

The rule of thumb is that the biggest size of random packing appropriate for the dimensions of the column should be used. However, an upper limit of 50 mm should be put in place. It should be noted that small-sized packing is considerably more expensive. Packing sizes greater than 50 mm do not show any cost-benefit when considering the reduced mass transfer efficiency. This is because of larger packing results in substandard liquid distribution (Sinnott, 2005, 592).

Recommended size ranges are:

  • For a column diameter smaller than 0.3 m diameter, use a packing size smaller than 25 mm.
  • For a column diameter between 0.3 m and 0.9 m, use a packing size of between 25 mm and 38 mm accordingly.
  • For a column diameter above 0.9 m, use a packing size between 50 mm and 75 mm.

4.5 Liquid Distribution

Optimal operation and separation are achieved by attaining adequate liquid distribution along the length and breadth of the column. Poor distribution is the cause of issues in packed columns. To obtain the highest possible efficiency, the column needs to have an evenly wetted surface area. The surface area of the packing needs to be covered entirely in an ideal situation. However, this is not practically possible when allowing for enough fouling resistance and an even liquid flow along the packing surface.

Suppose the actual liquid flowrate is less than the lower limit of the design value. In that case, the liquid levels in the distributor will be too small, and the flow to drip points along the length of the column will be non-uniform, resulting in poor separation efficiency. Similarly, if the actual liquid flowrate is larger than the upper limit of the design value, the liquid levels in the distributor will be too high, and overflows will occur. The separation efficiency will also be negatively impacted (Pilling & Holden, 2009).

5 Conclusion

Distillation column design is a complex and intricate process that requires a good comprehension of various concerns. Each design choice has an impact on the operational range, pressure drop, cost, capacity and efficiency of the overall processes. Observing the suggestions outlined here will allow for more knowledgeable decisions.

6 References

Sinnott, R.K. (2005), Coulson & Richardson’s Chemical engineering Design, Volume 6, 4th Edition, Elsevier, Massachusetts, USA.

Pilling, M. and Holden, B.S. (2009), “Choosing Trays and Packings for Distillation,” Back to Basics CEP, Sulzer Chemtech USA and The Dow Chemical Co.

The Pall Ring Company (2020) “Products for Mass Transfer Operations, Pall rings – Plastic”, The Pall Ring Company, https://www.pallrings.co.uk/products/pall-rings-plastic/, [2020, September 20]

The Pall Ring Company (2020) “Products for Mass Transfer Operations, Raschig Rings”, The Pall Ring Company, https://www.pallrings.co.uk/products/raschigrings/

, [2020, September 20].

Ultimo Engineers, (sa), “Ceramic Berl Saddles,” Ultimo Engineers, https://www.pallringsindia.com/ceramic-berl-saddles.html [2020, September 20]

Separation Technology (2012), “Distillation Column Tray Selection & Sizing,” Separation Process Industry, http://seperationtechnology.com/distillation-column-tray-selection-1/ [2020, September 20].

Pilling, M. and Summers, D.R. (2012), “Be Smart about Column Design,” Sulzer Chemtech USA, Reactions and Separations, American Institute of Chemical Engineers AiCHE

Hebert, S. and Sandford, N., (2016), “Consider Moving to Fixed Valves,” Koch-Glitsch, LP, Reactions and Separations, American Institute of Chemical Engineers AiCHE