Page 59 - Adsorbents - fundamentals and applications

44 SORBENT SELECTION: CRITERIA

the numerical method are discussed in detail elsewhere (Rege and Yang, 1997;
Hutson et al., 1999). The merit of the sorbent pairs was judged by subjecting
them to similar PSA cycles and by studying the performance parameters, such as
product purity, recovery, and throughput. The cycle parameters were so adjusted
that two of these performance parameters were nearly constant and the third was
compared to determine the better sorbent.
+
The first group of sorbents considered includes LiX (Al/Si = 1.0, 100% Li −
exchange) and NaX (13X) zeolites. The Langmuir equation parameters for the
N 2 and O 2 isotherms at 298 K for this group, as well as LiAgX, are summarized
in Table 3.2. The results of the PSA simulation runs with corresponding process
parameters previously given by Rege and Yang (1997) are shown in Table 3.3.
The performance of the adsorbents for air separation to give about 95% pure
O 2 product (with Ar impurity included in the O 2 product) was determined by
keeping the product throughput fixed at approximately 0.03 kg O 2 /h/kg sorbent
at constant product purity. A wide range of pressure ratios (P H /P L ) from 2 to 10
were considered. It has been shown that LiX is a superior sorbent compared with
NaX in the entire range of pressure ratios (Rege and Yang, 1997). The sorbent
parameter S was calculated for each of these cases and was plotted against the O 2
product recovery as shown in Figure 3.3. The figure clearly shows an increasing
product recovery with the value of the sorbent parameter. More importantly, the
value of S for LiX sorbent is much greater than that for NaX, which shows that
it is a valid parameter for comparing sorbents.
The second group of sorbents consisted of LiLSX and LiAgX sorbents. The N 2
and O 2 isotherms for these sorbents, as well as a comparison of their performance
as sorbents for air separation, appear in Hutson et al. (1999). The parameters of
the PSA cycle, as well as the corresponding PSA sorbent selection parameters, are
shown in Table 3.3. Two different cycles were used with different adsorption and
desorption pressures. In this case the product purity and recovery in each run were
kept roughly the same for both sorbents by manipulating the PSA parameters,
and the product throughputs were compared. As was shown by Hutson et al.
(1999) the LiAgX sorbent (with 1 Ag -ion per unit cell) showed a 12% higher
+
product throughput compared with LiLSX sorbent. This is the result of a slightly
higher N 2 loading on LiAgX compared with that on LiX. The product throughput
for the two runs was plotted against the parameter S andisshowninFigure 3.4.

Table 3.2. Langmuir parameters for N 2 and O 2 isotherms at 298 K on the adsorbents
used
Adsorbent N 2 O 2
−1
−1
q s (mmol/g) b(atm ) q s (mmol/g) b(atm )
LiX (Si/Al = 1) 2.653 0.946 2.544 0.086
NaX 0.982 0.901 0.276 0.624
LiAgX 2.635 1.170 1.863 0.131
From Rege and Yang, 2001.