PTQ Q2 2023 Issue

at the TGU are known to contribute to poisoning. Although not ideal for an SRU, this method of disposal is sometimes taken as the one with least consequences. SRUs that pro- cess BTEX-containing acid gas can pass those species on to the TGU, especially in lean acid gas situations. The effect of BTEX is thought to be reduced at temperatures below 240°C, as discussed elsewhere. 3 Deactivation is connected with catalyst pore and active site distribution. 4 Catalytic activity is conveniently defined in terms of the observed external rate constant k obs , which is equal to the product of the catalyst active site-based intrinsic rate constant k intr , the effectiveness factor η , and the active site surface number density, σ (number of sites per area of surface), and an ageing factor, AF:

1.0

0.8

for h > 2 T

0.6

0.4

η 1/h T

0.2

0.0

h or Φs/2.46 T

Figure 2 Effectiveness factor vs Thiele modulus

sulphate). Activity testing reported on abused catalyst (due to exposure to extreme temperatures) in spent conditions expresses about 50% of fresh activity, with carbon typically 0.1 or 0.2%. Catalyst deactivation and poisoning is a vast and fasci- nating subject in its own right, and it deserves much more than a cursory treatment. Part 3 will take a deeper plunge into the subject so, for now, the discussion of mechanisms and causes of deactivation is deferred. Suffice it to men - tion that hydrothermal ageing, sooting, chemisorption of poisons (especially of oxygen) and, by sulphation, coking and sintering are all causes of deactivation. Ultimately, the loss of specific surface area directly affects catalyst perfor - mance as the number of active Al-OH surface sites falls. Quantification of poisoning Generally, fresh alumina catalyst has a specific surface area of 300-350 m 2 /g; with initial ageing, surface area declines to 240-260 m 2 /g and then stays relatively stable, declining only slowly over several years until ‘spent’, at approximately 120 m 2 /g. Hydrothermal ageing tends to occur uniformly throughout the catalyst bed, approaching about 50% of fresh activity when spent. Poisoning is treated as activity loss related to any of sev- eral contaminants in the feed. Certain streams that wind up

k obs = k intr η σ AF

Poisoning corresponds to a loss of active sites, i.e., σ = σ 0 (1 - α ), where α is the fraction of sites poisoned. The effect on activity is a combination of site number density, poison selectivity, mass transfer resistance, and loss of surface area. Deactivation directly affects: • Selectivity: how quickly the poison interacts with the cat- alyst active sites; selective poisoning preferentially affects sites near the pore mouth and slowly progresses along the pore, vs non-selective, which progresses more or less uni- formly along the entire length of the pore. • η : effectiveness factor, i.e., reaction rate with mass trans- fer resistance/intrinsic reaction rates without mass transfer resistance. • h T : Thiele modulus, i.e., the ratio of kinetic rate to mass transfer (diffusion) rate. The activity response to poisoning depends on the com- bination of selectivity and Thiele modulus. The approximate order is as follows: • Half-order for non-selective, large h T • First-order for small h T (<2) • Reciprocal function (1/(1+ σ h T )) (selective with large h T ). Catalysts used in TGU service have enhanced pore struc- tures with macro-, meso- and micropores. These facilitate good transport of reactants into the interior surface and active Co/Mo sites with moderate diffusional resistance. Classic particle geometry estimation of characteristic pore radius, pore length, and tortuosity gives a rather low value for Thiele modulus and an overly conservative estimate of effectiveness factor. Comparison between experimentally determined whole and crushed catalyst activity coefficients is the best way to determine effectiveness factors because finely crushing the catalyst eliminates pore diffusion. The effect of poisoning on overall activity depends on the product of effectiveness factor and the fraction of sites poi- soned (see Figure 2 ). Selective poisoning can have a dra- matic effect when the Thiele modulus is large (h T ≫ 1). A linear relationship represents both selective and non-selec- tive poisoning when h T ≤1 (see Figure 3 ). This was selected for the model. Poisoning is seen to occur at the inlet section of the reactor bed, caused by the presence of contaminants and SO 2 , with the poison moving through the bed. The effect of poisoning on TGU reactor performance

0.8

= (1-σ)

0.5

0.6

= (1-σ)

0.4

0.2

= (1/1-σ h ) T

0.2

0.4

0.6

0.8

σ, fraction sites poisoned

selective, h =10

non sel, h >5 T

h <1, non sel T

h <1, selective T

Figure 3 Relative activity vs fraction sites poisoned

PTQ Q2 2023