Jun-2017

SO2 breakthrough under the microscope

Plant simulation reveals that any breakthrough of sulphur dioxide will cause havoc in the tail gas treating unit.

NATHAN HATCHER, CLAYTON JONES and SIMON WEILAND
Optimized Gas Treating

Article Summary

The effects on a tail gas treating unit (TGTU) of an SO2 breakthrough can be outright devastating. Effects can include dissolved quench column piping, column internals failure, quench water loop plugging with elemental sulphur, amine system corrosion, lost selective treating performance, high sulphur emissions, and increased solvent make-up and reclaiming needs. The interaction of NH3 slipped from the sulphur recovery unit (SRU) reaction furnace with varying levels of SO2 ingress affects quench column pH and amine system performance. These effects are examined in the context of their chemistry together with techniques for prevention and mitigation.

A surprising finding from this study is that a major SO2 breakthrough can result from relatively minor swings in Claus combustion air if analysers are poorly maintained and there are poor reliability practices. It is shown how the first SO2 breakthrough can be a self-fulfilling prophesy; quench water pH neutralisation intervention targets are reviewed; and the survival of SO2 beyond the quench system is investigated.

Figure 1 shows the pathway and relevant chemical reactions that occur when SO2 enters an aqueous solution. It is this chemistry that is responsible for the damage mechanisms described above.

Keller2 provides a thorough summary of causes and characteristics of SO2 breakthroughs that happen in operation. In general, the types of SO2 breakthrough can be loosely categorised into:
•  Mild SO2 breakthrough with low or high H2S
•  High SO2 with low H2S breakthrough
•  SO2 only breakthrough – low and high levels without H2S.

This work uses a ‘virtual’ microscope in the form of the ProTreat mass transfer rate based process simulator to examine what happens at the molecular and ionic chemistry levels to accompany the macro observations that are more familiar to those who have had the misfortune of having had to deal with an SO2 breakthrough.

Base case: normal operations prior to breakthrough
The sample is prepared for the microscope by setting some basic assumptions regarding how the plant was operating prior to the SO2 breakthrough, that is, when everything was ‘normal’. Figure 2 shows the overall flowsheet for the SRU/TGTU train that was simulated, together with a few key plant details for the base case. The operating data and column details associated with this simulation are from an actual operating plant. At the time of data collection, the Claus unit was operating towards the upper end of low level oxygen enrichment. Consequently, there was plenty of excess hydrogen generated within the Claus unit itself from hydrogen cracking in the thermal section. The first column in Table 1 summarises the critical, key operating metrics for this unit and the more important parameters that were studied for the virtual SO2 breakthrough.
 
SO2 breakthrough case studies
One of the more common scenarios that can lead to an SO2 breakthrough event results from either faulty or poorly maintained analysers that are either ignored or placed in manual. The air demand analyser (ADA in Figure 2) and Claus combustion trim air feedback control are perhaps the most important loops critical to keeping operational and in cascade control. The reasons will become quite evident after examining the case studies in this work. Within the TGTU itself, the hydrogen analyser (H2 control in Figure 2) is important to watch as excess hydrogen must be present to hydrogenate SO2 and elemental sulphur to H2S. Furthermore, the quench water pH analyser can indicate when an SO2 breakthrough is in progress so that mitigating actions can be taken.

Case study 1: Claus combustion air
To simulate the process of an SO2 breakthrough in our virtual microscope, the Claus unit air demand feedback control loop was placed in manual. Hydrogen make-up was assumed to be in manual and held closed while perturbing the Claus combustion air (steam 29) upwards. Table 1 outlines key process parameters that change during these perturbations. Percent excess Claus air refers to the amount of excess air relative to the amount required for a 2:1 ratio of H2S to SO2 in the Claus tail gas. Cautionary parameters are shown in blue. Damage can be expected for the parameters that are highlighted in red.

As excess air is increased in the Claus unit, the first sign of impending doom is the dramatic increase in the outlet temperature from the TGTU hydrogenation reactor. If it was left going for long enough, most seasoned operators would notice this telltale. At roughly 5% excess air, the temperature across the catalyst bed has increased by 90°F (32°C), which is one of the reasons why this part of the plant is often refractory lined. At 10% excess air, the bed outlet temperature is a whopping 745°F (396°C). However, many of these events can occur quickly over a period of just seconds to minutes and be gone before the thermal mass of catalyst is fully heated up, masking the problem.

As excess air increases, the extra SO2 being fed to the TGTU hydrogenation reactor consumes more and more of the excess hydrogen. At 10% excess air, the virtual hydrogen analyser shows 2.1%. Although this is a shade on the low side, most in the industry would consider there still to be plenty of hydrogen to take care of the hydrogenation in the Co/Mo catalyst.

Between 10 and 12% excess air, there is a subtle change that occurs within the TGTU front end. Hydrogen is consumed rapidly and SO2 begins to emerge from the hydrogenation reactor. At 12% excess air, there is 0.2 ppmv SO2 in the feed to the quench tower. The SO2 breakthrough has begun. Nevertheless, there appear to be no further malicious indicators within the quench tower or amine system. However, as shown below, the unit is on the precipice of a very serious upset.

With only a small additional perturbation from 12 to 12.3% excess air, the train derails. Table 1 shows the following sequence of events:
• SO2 increases from 0.2 to 74 ppmv in the quench column feed.
• Excess hydrogen drops nearly to zero.
• The cold quench water pH plummets from 7.13 to 4.35. On the hot side, the column bottoms pH has dropped from 6.72 to 3.83.
• SO2 has entered the amine system, leading to further complications:
    ν    A significant portion of the MDEA has been neutralised (0.13 mol/mol loading x 2 mols of MDEA per mol of dissolved SO2 = 0.26 or 26% neutralised), resulting in lost scrubbing capacity for H2S.
    ν    H2S removal is no longer acceptable. The normal overhead concentration of 30–40 ppmv H2S is now over 3500 ppmv.
    ν    With H2S present in the circulating amine, as long as the solution is kept basic, the sulphite can be expected to convert to thiosulphate over time as noted previously in Table 1. However, the kinetics of this conversion are not immediate by any means and are not well understood. In our opinion, it is not inconceivable that sulphite will survive and enter the regenerator. Under the extreme case of no conversion to thiosulphate or elemental sulphur, ProTreat shows that some SO2 will actually be present in the vapour inside the regenerator, with somewhere between 1-7 ppmv in the regenerator reboiler vapour return. Herein lies proof by simulation of a damage mechanism that was correctly postulated by Keller.¹ The regenerator reflux water could be almost as aggressive as the quench tower water if the breakthrough remains unchecked.
    ν    Note: the model does not presently include provision for the aqua-Claus reaction to elemental sulphur, so some of the destructive nature of the SO2 in solution could in reality be transferred to problems with elemental sulphur forming in the amine system.
So what exactly happened?