Page 121 - Gas Purification 5E

Alkanolamines for Hydrogen Surjide and Carbon Dioxide Removal 11 1

Tray Versus Packed Columns

Although bubble-cap trays and raschig ring packings were once commonly used in amine
plant absorbers and strippers, modem plants are generally designed to use more effective
trays (e.g.. sieve or valve types) and improved packing shapes (e.g., Pall rings or high-per-
formance proprietary designs). Very high-performance structured packing is seldom used for
large commercial gas treating plants because of its high cost and sensitivity to plugging by
small particles suspended in the solution.
The choice benreen trays and packing is somewhat arbitrary because either can usually be
designed to do an adequate job, and the overall economics are seldom decisively in favor of
one or the other. At this time, sieve tray columns are probably the most popular for both
absorbers and strippers in conventional, hge commercial amine plants; while packed columns
are often used for revamps to increase capacity or efficiency and for special applications.
Tray columns are particularly applicable for high pressure columns, where pressure drop
is not an important consideration and gas purity specifications can readily be attained with
about 20 trays. Packing is often specified for C02 removal columns, where a high degree of
C02 removal is desired and the low efficiency of trays may result in objectionably tal!
columns. Packing is also preferred for columns where pressure drop and possible foam for-
mation are important considerations. Packing should not be used in absorbers treating unsat-
urated gases that can readily polymerize (propadiene, butadiene, butylene, etc.) as gum for-
mation can lead to plugging of the packing. Also, packing should not be used in treating
gases containing H2S which are contaminated with oxygen because of the potential for plug-
ging with elemental sulfur. General factors affecting the choice between tray and packed
towers are discussed in Chapter 1.
Table 2-12 gives a comparison of trays and packing based on an analysis by Glitsch. Inc.
and presented by Gangriwala (1987). In the table, valve trays on 24-in. spacing are assigned
capacity and efficiency indexes of 100, and various other column internals are assigned com-
parative values. The data show, for example, that a column can be modified to give 132% of
the original capacity at the same efficiency by converting from valve trays at Win. spacing
to #3 Cascade Mini Rings. Alternatively, the efficiency of a trayed column can be increased
at essentially the same (or slightly higher) capacity by converting to #2 or #2.5 Cascade Mini
Rings. Conventional random packings (Ballast Rings) do not show any significant
capacity/efficiency advantages over trays: however, they maj7 offer other advantages such as
a low pressure drop. Although the table covers only Glitsch products, it should be noted that
other vendor products in the same categories show very similar performance characteristics.
The effect of contactor design on the selectivity of amine solutions for absorbing H,S in the
presence of C02 has been studied by Darton et al. (1987). They conclude that selectivity, as
represented by the ratio of overall mass transfer coefficients for H2S and CO?, is about the
same (100) for trays and ?-in. Pall rings in a MDEA contactor operating at atmospheric pres-
sure. Their work indicates that even higher selectivities should be possible with alternative con-
tactor designs, such as cyclones, centrifuges, and cocurrent gadliquid flow tubes.
Vickery et al. (1988) compared valve trays with 2-in. steel Pall rings for treating high-
pressure natural gas with 50% MDEA solution. The results of this analysis, based on GAS-
PLANT software, are given in Table 2-13, These results, which indicate a much higher
selectivity for Pall.rings than for valve trays, do not appear to agree with those of Darton et
al. (1987): however, the two studies were based on widely different operating pressures, and
other parameters may also have differed substantially.