PTQ Q1 2026 Issue

Identify corrosive regimes in hydrofluoric alkylation units: Part 2

How thermodynamic modelling can identify high-risk conditions and guide practical actions to reduce entrainment, adjust heat duty, and upgrade materials

Ezequiel Vicent OLI

H ydrofluoric (HF) acid alkylation units are workhorses of modern refineries. They convert isobutane and light olefins into high-octane alkylate, a gasoline blend stock valued for its clean burning and high research octane number (RON). The process employs HF acid as both catalyst and solvent, and consequently carries a unique set of safety and corrosion challenges. Part 1 of this article, published in PTQ Q4 2025 , and the following Part 2 demonstrate how a new thermodynamic model developed within a HF alkylation joint industry pro- ject (JIP) marks a significant advance in understanding and managing corrosion in HF alkylation units, demonstrating the ability to identify high-risk conditions, such as high entrainment and hot spots, and guide practical actions, such as reducing entrainment. The first case study looked at preheat and depropaniser feed. The second will consider the objective and approach. Case study 2: Objective and approach A particular unit in the US experienced corrosion in the depropaniser, the main fractionator for this HF unit. Corrosion was observed around the flash zone and in the top trays of the column. The refinery participated in the original JIP and built a full column model, which included the hydrocarbon phase saturation, the HF stripper, and overhead accumulator. This column contained 30 trays and, at the base entrainment level, it was found through simula- tion that a separate acid phase was forming below the flash zone (tray 17) down to tray 14. Therefore, for the present study, OLI’s team focused on determining the entrainment level that leads to acid phase formation on specific trays (14 and 17-20), calculating tran- sition temperatures and estimating the heat duty required to dissolve the acid phase on each tray. Unlike the previous study (see Part 1 in PTQ Q4 2025 ), which considered a preheat train, this analysis extended inside the fractiona- tion column. The goal was to determine whether entrained acid reached the trays and, if so, whether additional heat could dissolve the acid phase, at least containing it within the zone above the flash zone. The plant provided operating data for 2023. Initial entrainment, based on the acid boot test, was ~50 lb/h, but model inversion (using the thermodynamic model to

Calculated transition temperatures

Entrainment

Tray

14 TT

17 TT

18 TT

19 TT

20 TT

lbs/h 160 200 500

ºF

227 229 229

228 228 228

210 210 210

202 203 202

Table 1

increase entrainment until observable corrosion was local- ised) indicated that actual entrainment should be ~160 lb/h. To assess sensitivity, the study considered entrainment lev- els of 160 lb/h, 200 lb/h, and 500 lb/h. For each level, the model performed single variable tran- sition temperature (TT) surveys and contour plot (T-P) sur- veys for each tray. The transition temperatures were then compared with tray temperatures from the process simula- tion to determine whether the acid phase would persist on each tray. If the operating point was below TT, additional duty (heat) would be needed to dissolve the acid; if above, corrosion risk would be low. Transition temperature vs entrainment Table 1 summarises the calculated TT for the acid liquid phase as a function of entrainment for trays 14-20. Each marker corresponds to a tray; the lower group of points (around 205°F) belongs to the lower trays (19-20), while the upper group (≈228°F) corresponds to the flash zone and feed tray (tray 17). A tabulated version of the data is included below the plot. The results reveal several important trends. u Weak dependence on entrainment within the stud- ied range: Transition temperature increases slightly (2°F) as entrainment rises from 160 to 500 lb/h, but the differ- ences are small compared with measurement uncertainty. This result implies that once entrainment exceeds a certain threshold, adding more acid has little impact on TT because the temperature and pressure inside the column differ sig- nificantly from the fluid entering the column. In addition, vapour-liquid equilibrium (VLE) changes as the composi- tion of each tray is different from the next. v Substantial variation between trays: Trays 17 and 18 (around the column flash zone) require temperatures

PTQ Q1 2026