Columns, packed

Most packed column supports are prepared from diatomaceous earth, which is composed of skeletons of diatoms. The diatomite is basically an amorphous 

Liqmd stationary phase selection is extremely important for the satisfactory chromatographic separation. The stationary phase should be chosen for good chemical and thermal stabihty in addition to suitable selectivity. 

One of the most important parameters to characterise the stationary phase is the McReynolds constant which was first proposed by Rohrschneider [24-26] and then further developed by McReynolds [27], as shown in equation (33)  

Where RI is retention index (see equation 22), ARI represents retention index difference (for example, the retention index for benzene was 649 on a non-polar 20% squalane column, and 1169 on a highly polar SP-2340 stationary phase column under identical GC conditions, then ARI=520). 

From a large number of experiments (226 stationary phases were studied, and 68 compounds on 25 columns were analysed), McReynolds selected 10 most valuable compounds (the most valuable five are benzene, n-butanol, 2-pentanone, nitropropane, and pyridine) as probes to characterise columns. The polarity of the column as measured with benzene is termed X and is equal to ARI/100 for benzene. Similarly, y, z, u and s are the 1/100 terms for the other four probe compounds. The coefficient a, b, c, d and e for x, y, z, u and s terms are constants, which are defined for these five probe compounds. For benzene, a=100 and b, c, d and e=0. For n-butanol, b=100, and a, c,d and e=0, and so on for the other three probe compounds. Many GC manufacturers present the values of McReynolds constants for various stationary phases in their catalogues. Table 1 list McReynolds constants for some commonly used stationary phases. 

Packed columns are gaining ground on trayed columns. Lieberman states that based on his design and operating experience, a properly designed packed tower can have 20-40% more capacity than a trayed tower with an equal number of fractionation stages. 

Final design of packed columns should be performed by experts, but the layman is often required to provide preliminary designs for studies. This packed column section provides the information necessary for such estimates. 

Total Tower Height. This is an assumed maximum. 

Note Number Section Notes Height Allowance for Studies Reference 

Frank reports The liquid fall between the distributor and the top of the packing should be no greater than 12 in.  

Packed columns are used for distillation, gas absorption, and liquid-liquid extraction only distillation and absorption will be considered in this section. Stripping (desorption) is the reverse of absorption and the same design methods will apply. 

The gas liquid contact in a packed bed column is continuous, not stage-wise, as in a plate column. The liquid flows down the column over the packing surface and the gas or vapour, counter-currently, up the column. In some gas-absorption columns co-current flow is used. The performance of a packed column is very dependent on the maintenance of good liquid and gas distribution throughout the packed bed, and this is an important consideration in packed-column design. 

The design of packed columns using random packings is covered in books by Strigle (1994) and Billet (1995). 

The choice between a plate or packed column for a particular application can only be made with complete assurance by costing each design. However, this will not always be worthwhile, or necessary, and the choice can usually be made, on the basis of experience by considering main advantages and disadvantages of each type which are listed below  

Plate columns can be designed to handle a wider range of liquid and gas flow-rates than packed columns. 

The choice between a plate or packed column for a particular application can only be made with complete assurance by costing each design. However, this will not always 

CHAPTER 11 SEPARATION COLUMNS (DISTILLATION, ABSORPTION, AND EXTRACTION) 

Packed columns are as important as tray columns in the process industiy. Due to novel developments of packing elements the industrial use of packed columns is steadily increasing. In packed columns there exists a genuine countercurrent flow of gas and liquid as is shown in Fig. 5.4-14. An intimate contact between gas and liquid phases is established by packings that represent a solid structure with high porosity and large internal surface. The liquid proceeds downward in form of thin films or rivulets. Decisive for a good performance are a low pressure drop of the gas and a Uquid flow that is uniform over the cross section of the column. 

Very important for all types of packings is a uniform liquid distribution at the top of the bed and a limitation of bed height to 6 or 8 m. Beneath each bed, the liquid has to be collected, mixed, and redistributed. These measures intend to suppress the so-called maldistribution of liquid because it strongly affects mass transfer rates. The design of liquid distributors, liquid collectors, support grids, etc., should provide a large open area not to hinder the countercurrent flow of gas and liquid. 

Packed columns can be operated within certain limits of gas and liquid loads only. A typical operation region of packed columns is shown in Fig. 5.4-17. The mechanisms that set limitations to gas and hquid flow rates are flooding and poor surface wetting. Flooding is a strong limitation that caimot be surpassed. Nonsufficient surface wetting is a soft limitation that can be crossed at the expenses of poorer mass transfer efficiency, ft is important to note that the operation range of the liquid is 

The operation point has to be chosen so that a sufficient safety margin to the operation limits exists. In practice, an operating pressure loss of approx. 3 mbar/m is often recommended for colunm dimensioning. 

Correlation of flooding of packed columns according to Mersmann (1965, 

Packed columns can be used for both preparative and analytical separations. The preparative columns have typically an ID of 1-2 cm and sample volumes up to 1 ml can be injected. The amount of stationary phase in partition chromatography (GLC) is 20-30% (w/w particle). 

Analytical packed columns have an ID of 2-4 mm, and the amount of stationary phase in GLC is 10% (w/w). 

The column body material is glass or metal (e.g., stainless steel). 

Glass is rather inert and allows inspection of the packing efficiency, but requires skill to connect the column to the injector and detector. Metal columns are easier to handle, but can have some catalytic activity. 

Packed columns are often used for gas/liquid contacting when a high volumetric mass transfer rate has to be combined with countercurrent gas and liquid flows. They are particularly suitable for chemically enhanced absorption. This is the phenomenon that the absorbed component is completely converted by the rapid chemical reaction within the diffusion layer, close to the gasAiquid interface (see section 5.4.2.1), In such processes there is no need for a large liquid volume, the reaction rate per unit volume is then directly proportional to the interfacial area. 

The design of packed columns is too specialistic a subject to be treated here. 

For a first design method b is the more reliable. For an accurate optimization one needs the correlations one can find in the literature. A useful review paper on the subject was presented by Carra and Morbidelli (1987). 

Sahay and Sharma [8] carried out measurements in a packed column filled with 1 packing elements. They arrived at the following equation (Equation A7.11) for kody in which the coefficients depend on the geometry and the material of packing elements  

Onda et al. [9] developed the following correlation for in packed columns. The correlation is considered suitable for most packing elements, with the exception of Pall rings. A requirement for using this correlation is that the mass transfer area (fly) is estimated using some other correlation  

For packing elements with a saddle or a ring geometry d = 0.95-7.6 cm), a modification [ 10] of the traditional Sherwood and Holloway correlation is suggested  

The most frequently used correlation for the mass transfer area was developed by Onda 

Packing Element Material Nominal Size (m) Estimated a (1/m) Porosity (—) 

When the packing is also a catalyst, both countercurrent and cocurrent flows are applied. In the latter case both upflow and downflow operations are encountered. With upflow operation the contacting between gas and liquid is superior, but the pressure drop is higher and there are restrictions on flow rates and packing diameter because of flooding. The downflow cocurrent column packed with catalyst may operate in two distinct flow regimes the trickle flow regime when the gas phase is continuous and the liquid phase is dispersed, or 

Locus of Reoctioii Packed Plate Empty Stirred Vessel Miscellaneous 

Fluid phase only CounterciiiTent flow CounterciiiTent Coiintercunent Absorber or Venturi 

Trickle bed reactors have grown rapidly in importance because of their application in hydrodesulfurization of naphtha, kerosene, gasoil, and heavier petroleum fractions in hydrocracking of heavy gasoil and atmospheric residues in hydrotreating of lube oils and in hydrogenation processes. In trickle bed operation the flow rates are much lower than those in absorbers. To avoid effectiveness factors in the reaction that are too low, the catalyst size must be much smaller than that of the packing used in absorbers, which also means that the overall void fraction is much smaller. 

The fixed bed is preferred to a slurry-type operation when the gas flow rate is relatively low because it leads to a gas and liquid flow pattern that better approximates plug flow. Only for high gas flow rates would an operation with suspended catalyst be preferred — when the catalyst size permits it — to avoid the pulsed flow regime that might be encountered in fixed bed operation. 

Packing in (mm) Percent voids (0 Specific surface (a ) nf/nf Dumped weight kg/m (Ib/ft ) Packing factor (F) 

Rate = Aiky(yA — yAi) for the vapor phase, and Rate = A kx(XAi — xa) for the liquid phase. 

Ai is the interfacial area between the liquid and vapor phases, xa and are the interfacial mole fractions, xa and y A are the bulk-phase mole fractions, and ky are the individual mass transfer coefficients for the liquid and vapor phases, respectively. 

Since the interfacial area, Ai, is difficult to determine accurately, these rate equations may be rewritten as  

is the liquid mole fraction that would be in equilibrium with and y is the vapor mole fraction that would be in equilibrium with Xa  

In order to obtain a good rate of transfer per unit volume of the tower, a packing is selected which will promote a high interfacial area between the two phases and a high degree of turbulence in the fluids. Usually increased area and turbulence are achieved at the expense of increased capital cost and/or pressure drop, and a balance must be made between these factors when arriving at an economic design. 

The construction of packed towers is relatively straightforward. The shell of the column may be constructed from metal, ceramics, glass, or plastics material, or from metal with a 

Gas is distributed directly into packed bed - no hydrostatic head - gas and liquid flows through separate openings in plate 

At the top of the packed bed a liquid distributor of suitable design provides for the uniform irrigation of the packing which is necessary for satisfactory operation. Four 

Uniform liquid flow is essential if the best use is to be made of the packing and, if the tower is high, re-distributing plates are necessary. These plates are needed at intervals 

An alternative approach to the HETP method for calculating the column packing height, also discussed in this chapter, is based on the height of a transfer unit (HTU) and the number of transfer units (NTU). In this approach the packing height is calculated by a theoretical analysis of mass transfer phenomena across the liquid and vapor phases. 

Another parameter in the design of a packed column is the column diameter, which is determined on the basis of hydraulic considerations, such as pressure drop and flooding conditions. The hydraulic behavior of the column depends on the packing type and operating conditions. Correlations based on experimental data are available for calculating column packing hydraulics, and are discussed in this chapter. 

The use of packing instead of trays in multistage separation columns is common for column diameters 3-4 ft or smaller. More recently, packing has been used for larger columns because of its low pressure drops, favorable efficiencies, and high vapor capacity. Packed columns are also the preferred choice where corrosion is a potential problem. The packing material used in these situations is ceramic or polymeric. Another characteristic of packed columns is their low liquid holdup, which could reduce the amounts of off-specification products at startup and shutdown. 

Pink diatomaceous materials derived from crushed firebrick, for example, Chromosorb P. The material is used for high performance stationary phases due to its high surface area (4.0m g ) and support for high stationary phase loadings, up to 35% (w/w). It is particularly suitable for alkanes but must be deactivated by silanisation for polar compounds. 

White diatomaceous materials prepared from calcined diatomite, for example, Chromosorb W, a fragile packing of lower surface area (1.0 m g ) than Chromosorb P, suitable for polar compounds, Chromosorb G is the hardest Chromosorb and is twice as dense as Chromosorb W. Maximum stationary phase loading is approximately 5% (w/w), equivalent to 12% loading on Chromosorb W because of its higher density. 

Chromosorb T is a polytetrafluoroethylene (PTFE) support material. It is inert and hydrophobic and therefore suitable for analysis of small polar molecules such as water, organic acids, phenols, amines and acidic gases (HF, HCl, SO2, NO c). Chromosorb T can be coated with stationary phases such as polyethylene glycol, Apiezon or a fluorocarbon oil (Kel-F, Fluoropak-80) but can be difficult to pack. 

HMDS is useful for on-column silanisation since the by-product of the reaction is nitrogen, which will not harm the metallic surfaces of the 

---

Packed columns are widely used in gas absorption, but particulates are also removed in the process (see Fig. 11.2a). The main disadvan-... 

In the case of a plate column the performance of a real plate is related to the performance of a theoretical one by the plate efficiency. In the case of a packed column the height equivalent to a theoretical plate HETP) gives a measure of the contacting efficiency of the packing. 

Used in virtually all organic chemistry analytical laboratories, gas chromatography has a powerful separation capacity. Using distillation as an analogy, the number of theoretical plates would vary from 100 for packed columns to 10 for 100-meter capillary columns as shown in Figure 2.1. 

The column is swept continuously by a carrier gas such as helium, hydrogen, nitrogen or argon. The sample is injected into the head of the column where it is vaporized and picked up by the carrier gas. In packed columns, the injected volume is on the order of a microliter, whereas in a capillary column a flow divider (split) is installed at the head of the column and only a tiny fraction of the volume injected, about one per cent, is carried into the column. The different components migrate through the length of the column by a continuous succession of equilibria between the stationary and mobile phases. The components are held up by their attraction for the stationary phase and their vaporization temperatures. 

The most common mobile phases for GC are He, Ar, and N2, which have the advantage of being chemically inert toward both the sample and the stationary phase. The choice of which carrier gas to use is often determined by the instrument s detector. With packed columns the mobile-phase velocity is usually within the range of 25-150 mF/min, whereas flow rates for capillary columns are 1-25 mF/min. Actual flow rates are determined with a flow meter placed at the column outlet. 

Packed Columns A packed column is constructed from glass, stainless steel, copper or aluminum and is typically 2-6 m in length, with an internal diameter of 2-4 mm. The column is filled with a particulate solid support, with particle diameters ranging from 37-44 pm to 250-354 pm. 

To minimize the multiple path and mass transfer contributions to plate height (equations 12.23 and 12.26), the packing material should be of as small a diameter as is practical and loaded with a thin film of stationary phase (equation 12.25). Compared with capillary columns, which are discussed in the next section, packed columns can handle larger amounts of sample. Samples of 0.1-10 )J,L are routinely analyzed with a packed column. Column efficiencies are typically several hundred to 2000 plates/m, providing columns with 3000-10,000 theoretical plates. Assuming Wiax/Wiin is approximately 50, a packed column with 10,000 theoretical plates has a peak capacity (equation 12.18) of... 

This method uses a short, packed column that generally produces a poor resolution of chromatographic peaks. The liquid-liquid extraction used to extract the trihalomethanes is nonselective. Besides the trihalomethanes, a wide range of nonpolar and polar organic constituents, such as benzene and... 

Time, Cost, and Equipment Analysis time can vary from several minutes for samples containing only a few constituents to more than an hour for more complex samples. Preliminary sample preparation may substantially increase the analysis time. Instrumentation for gas chromatography ranges in price from inexpensive (a few thousand dollars) to expensive (more than 50,000). The more expensive models are equipped for capillary columns and include a variety of injection options and more sophisticated detectors, such as a mass spectrometer. Packed columns typically cost 50- 200, and the cost of a capillary column is typically 200- 1000. 

Kovat s retention index (p. 575) liquid-solid adsorption chromatography (p. 590) longitudinal diffusion (p. 560) loop injector (p. 584) mass spectrum (p. 571) mass transfer (p. 561) micellar electrokinetic capillary chromatography (p. 606) micelle (p. 606) mobile phase (p. 546) normal-phase chromatography (p. 580) on-column injection (p. 568) open tubular column (p. 564) packed column (p. 564) peak capacity (p. 554)... 

Graham, R. C. Robertson, J. K. Analysis of Trihalomethanes in Soft Drinks, /. Chem. Educ. 1988, 65, 735-737. Trihalomethanes are extracted from soft drinks using a liquid-liquid extraction with pentane. Samples are analyzed using a packed column containing 20% OV-101 on 80/100 mesh Gaschrom Q equipped with an electron capture detector. 

An internal standard of 1-butanol is used to determine the concentrations of one or more of the following impurities commonly found in whiskey acetaldehyde, methanol, ethyl acetate, 1-propanol, 2-methyl-1-propanol, acetic acid, 2-methyl-1-butanol and 3-methyl-1-butanol. A packed column using 5% Garbowax 20m on 80/120 Garbopak B and an EID detector were used. 

This experiment provides an alternative approach to measuring the partition coefficient (Henry s law constant) for volatile organic compounds in water. A OV-101 packed column and flame ionization detector are used. 

Method f2.i describes the analysis of the trihalomethanes CHCI3, CHBr3, CHChBr, and CHClBr2 in drinking water using a packed column with a nonpolar stationary phase. Predict the order in which these four trihalomethanes will elute. 

Bentone-34 has commonly been used in packed columns (138—139). The retention indices of many benzene homologues on squalane have been determined (140). Gas chromatography of C —aromatic compounds using a Ucon B550X-coated capillary column is discussed in Reference 141. A variety of other separation media have also been used, including phthaUc acids (142), Hquid crystals (143), and Werner complexes (144). Gel permeation chromatography of alkylbenzenes and the separation of the Cg aromatics treated with zeofltes ate described in References 145—148. 

Fig. 2. Packing materials for packed columns, (a)—(f) Typical packing elements generally used for random packing (g) example of stmctured packing.

To use all of these equations, the heights of the transfer units or the mass transfer coefficients and must be known. Transfer data for packed columns are often measured and reported direcdy in terms of and and correlated in this form against and... 

Experimental Mass Transfer Coefficients. Hundreds of papers have been pubHshed reporting mass transfer coefficients in packed columns. For some simple systems which have been studied quite extensively, mass transfer data may be obtained directiy from the Hterature (6). The situation with respect to the prediction of mass transfer coefficients for new systems is stiU poor. Despite the wealth of experimental and theoretical studies, no comprehensive theory has been developed, and most generalizations are based on empirical or semiempitical equations. 

Other correlations based partially on theoretical considerations but made to fit existing data also exist (71—75). A number of researchers have also attempted to separate from a by measuring the latter, sometimes in terms of the wetted area (76—78). Finally, a number of correlations for the mass transfer coefficient itself exist. These ate based on a mote fundamental theory of mass transfer in packed columns (79—82). Although certain predictions were verified by experimental evidence, these models often cannot serve as design basis because the equations contain the interfacial area as an independent variable. 

Pig. 22. Schematic representation of typical pressure drop as a function of superficial gas velocity, expressed in terms of G = /9q tiQ, in packed columns. O, Dry packing , low Hquid flow rate I, higher Hquid flow rate. The points do not correspond to actual experimental data, but represent examples. 

Nonisothermal Gas Absorption. The computation of nonisothermal gas absorption processes is difficult because of all the interactions involved as described for packed columns. A computer is normally required for the enormous number of plate calculations necessary to estabUsh the correct concentration and temperature profiles through the tower. Suitable algorithms have been developed (46,105) and nonisothermal gas absorption in plate columns has been studied experimentally and the measured profiles compared to the calculated results (47,106). Figure 27 shows a typical Hquid temperature profile observed in an adiabatic bubble plate absorber (107). The close agreement between the calculated and observed profiles was obtained without adjusting parameters. The plate efficiencies required for the calculations were measured independendy on a single exact copy of the bubble cap plates installed in the five-tray absorber. 

This type of equation has been found useful in correlating drop diameters in packed columns where the packing si2e exceeds the drop diameter (65). 

In the case of a packed column, the terms on the right-hand side should each be divided by the voidage, ie, the volume fraction not occupied by the soHd packing (71). In unpacked columns at low values of the sHp velocity approximates the terminal velocity of an isolated drop, but the sHp velocity decreases with holdup and may also be affected by column internals such as agitators, baffle plates, etc. The sHp velocity can generally be represented by (73) ... 

Fig. 12. Unagitated column extractors (a) spray column (b) packed column and (c) perforated-plate column.

Pulsed Columns. The efficiency of sieve-plate or packed columns is increased by the appHcation of sinusoidal pulsation to the contents of the column. The weU-distributed turbulence promotes dispersion and mass transfer while tending to reduce axial dispersion in comparison with the unpulsed column. This leads to a substantial reduction in HETS or HTU values. 

---