Page 266 - Gas Purification 5E

Mechanical Design and Operation of Ahnolamine Plants 251

solution through a bed of activated carbon as described by Fife (1932). Not all activated car-
bons are suitable for amine filtration. Bourke and Mazzoni (1989) recommend the use of a
hard, fine mesh, steam-activated carbon having a broad range of pore sizes. They also rec-
ommend that the activated carbon have a low phosphorus content since leachable phosphates
can cause foaming. According to Pauley (1991), activated carbon with an iodine number
ranging from 900 to 1,100 is effective in removing amine degradation products, while car-
bons with a molasses number greater than 200 are more efficient in removing liquid hydro-
carbons. (Both the iodine and molasses numbers are measures of the pore area available for
adsorption.) Gustafson (1970) recommends a simple field test to judge the effectiveness of a
given activated carbon in reducing foaming. Two quart jars are partially filled with amine
solution, 1 wt% activated carbon is added to one jar, and both jars are then shaken vigorous-
ly. Comparison of the foam volume and duration of foaming indicates the degree of foam
reduction that can be achieved with that activated carbon.
While it is generally agreed that activated carbon filtration should be considered for all
amine units, it is highly recommended for secondary and tertiary amines (DEA, MDEA, and
DIPA), which cannot be thermally reclaimed online like DGA and MEA. The preferred loca-
tion for the activated carbon filter is on the lean side of the circulating system downstream of
the lean amine cooler. The activated carbon is more effective on the lean side and, if H2S is
present, safety concerns are not an issue when the carbon is replaced. However, there are
some circumstances when it may be advisable to locate the carbon filter on the rich side. See
the previous discussion on mechanical amine filtration for details. In filtration systems where
the carbon filter is located on the rich side there is a tendency for gas pockets to form inside
the carbon filter (Leister, 1996). These gas pockets can reduce the effectiveness of the car-
bon filter and, in rich amine filtration systems where H2S is present, care should be taken to
continuously vent the carbon filter to a downstream location to prevent flash gases from
being a safety hazard during filter changeouts as well as from accumulating and sealing off a
portion of the filter (Bacon, 1987).
Table 3-8 summarizes activated carbon filter design data. As indicated in this table, a
variety of activated carbons (8 x 30 mesh, 5 x 7 mesh, 4 x 10 mesh, etc.) derived from coal
or wood have been used for amine filtration. Activated carbon filters using the finer carbon
grades (8 x 30 mesh) are the most common type and are usually designed for a flux of 2 to 4
gpm/ft2 with a 10 foot bed height (15 to 20 minutes contact time). Filters using the coarser
carbon grades (5 x 7 mesh and 4 x 10 mesh) are designed for fluxes as high as 10 gpm/ft2
and usually have a minimum bed depth of 5 feet. All of these carbon filters operate in a
downflow mode and, in most cases, a 10 to 20% slipstream of lean amine is filtered. Small
amine systems may use an activated carbon cartridge element.
Per Table 3-8, the recommended size of the slipstream filtered by activated carbon varies
from 1% to 100% (full flow filtration). Early activated amine filters built in the 1970s were
often designed for a 1 to 2% amine solution slipstream. Experience demonstrated that these
filters were undersized, and most activated carbon filters are now designed for a 10 to 20%
slipstream (Bourke and Mazzoni, 1989). However, in some applications, e.g., Claus plant tail
gas treaters, activated carbon filters may be designed for full flow filtration. Typical activat-
ed carbon bed life is about 6 months to a year.
The activated carbon is replaced when the pressure drop across the bed exceeds design
values, when foaming or turbidity tests indicate that the carbon is spent, or when there is no
color change in the amine solution flowing through the filter. Recommended test procedures
are available from the activated carbon vendors. Ballard (1986A, B) suggests that better acti-
vated carbon filtration is obtained when the amine temperature is in the range of 120 to