Ammonia Destruction in the Reaction Furnace

Ammonia destruction in the reaction furnace

A study based on reaction fundamentals aids a comparison of the advantages of one-zone and two-zone SRU reaction furnace design in ammonia destruction

Simon Weiland, Nathan Hatcher, Clayton Jones and Ralph Weiland Optimized Gas Treating Inc.

mmonia typically evolves from nitro- destruction chemistry, and concludes with a gen-containing components in processing case study illustrating how design and operat- Acrude oil. Processing and handling sour ing parameters affect ammonia destruction. gas and sour water containing ammonia in an environmentally acceptable manner has always One-zone reaction furnace been a challenge in refineries and is the focal The one-zone furnace, also referred to as a point for many designs and patents. There are straight-through design, operates with all the two main objectives that these designs and amine acid gas (AAG) and sour water acid gas patents attempt to achieve: purifying ammo- (SWAG) feed being mixed together, then fed to nia-bearing sour water to prepare the water for the furnace via a single burner (see Figure 1). further processing or discharge; and disposing of This design is generally applied when the feed the ammonia once it has been removed from the gas contains no ammonia, or small concentra- sour water. Options for disposing of ammonia tions of ammonia, but with the addition of include destroying it, or selling it as a product if preheat and incrementally longer residence time it meets purity standards. This article focuses on it can also be used successfully when ammonia is a common method for disposing of the ammo- present in higher concentrations. nia: destroying it in the reaction furnace of a A benefit of the one-zone furnace is that all the sulphur recovery unit (SRU). When directed to acid gas passes through the burner flame, an SRU, nearly complete destruction of ammo- thereby providing the maximum available heat nia is needed to prevent plugging the SRU with for the destruction of contaminants such as ammonium salts. hydrocarbons and ammonia. Insufficient The effectiveness of ammonia destruction is destruction of hydrocarbons in the reaction strongly influenced by the configuration and furnace can have disastrous effects on the operation of the reaction furnace. The two converter beds downstream, especially if they hardware configurations considered here are contain aromatics such as BTEX. Another bene- one-zone and two-zone designs. The two-zone design that will be considered has one SWS acid gas Amine acid gas burner at the front of the first reaction chamber. These are two of the most common reac- Air To waste heat boiler tion furnace configurations used in SRUs. In this article, a Burner brief introduction to each reac- Reaction furnace tion furnace design is followed by a discussion of ammonia Figure 1 General layout of a one-zone reaction furnace

www.digitalrefining.com/article/1001297 PTQ Q4 2016 1 nia, comes from upstream acid SWS acid gas Amine acid gas gas removal units. It is split and directed to both the front-end and back-end zones. The rela- tive fractional split between the zones is a primary unit control Air To waste heat boiler parameter. The Claus process Zone 1 Zone 2 only requires that one-third of Burner the total H S be oxidised to SO , Reaction furnace 2 2 which leaves the remaining Figure 2 General layout of a two-zone reaction furnace4 two-thirds to act as a heat sink in the flame. Bypassing clean fit is a more straightforward control scheme than AAG around the front zone burner flame in a two-zone furnace because the one-zone increases the flame temperature by reducing the furnace controls do not have to address splitting amount of non-combusting gas in the flame that gas between zones. As with any design, however, is acting as a heat sink. there are also some disadvantages. The percentage of AAG being bypassed to the Depending on feed gas composition, one-zone second chamber depends on the desired flame

furnaces may experience problems with flame temperature as well as the concentration of H2S stability. Flame stability can be improved in a in the gas. The temperature of the front zone one-zone furnace by increasing the flame should be kept high enough to ensure adequate temperature by preheating the feed streams, ammonia destruction, generally in the range including the combustion air. With very low 2300–2700°F (1260-1480°C). The maximum

quality, lean H2S feeds, natural gas spiking is temperature is governed by the refractory sometimes employed to maintain flame stability. lining’s upper temperature limit. Generally, no

In a one-zone furnace, there is no acid gas more than 60% of the total H2S in the SRU feed bypass available to increase the flame tempera- should be bypassed to ensure the atmosphere in ture. Lower flame temperatures can leave the front zone does not operate in the oxidising one-zone furnaces more susceptible to flame-out region and cause NOx levels to increase. If the in the event of feed composition changes and atmosphere in the furnace operates in this can also lead to multiple operating and reliabil- oxidising region, substantial amounts of NOx ity issues caused by incomplete destruction of and SOx can be produced which, under the right contaminants. By placing preheaters in the feed conditions, can form hot aqua regia, an gas and air streams, this operating instability extremely corrosive material capable of dissolv- can be mitigated; however, it comes at the cost ing even gold and platinum. The controls for the of adding capex for the equipment as well as amount of acid gas bypass are generally based opex for maintaining and operating that on the flame temperature in the front zone as

equipment. well as the concentration of H2S in the AAG feed. One of the benefits of a well-designed two-zone Two-zone reaction furnace furnace is improved ammonia destruction over An alternative furnace design is the two-zone the one-zone furnace. To a large extent, greater one-burner furnace (see Figure 2), also referred ammonia destruction results from higher flame to as a front side split. This design is generally temperatures. This temperature control advan- applied when the ammonia content in the acid tage in the front zone comes from the ability to gas exceeds the nominal 2 mol% limit, but it can bypass some of the amine acid gas around the also be used in cases with less ammonia. The front burner, relieving some of the cold acid gas general layout directs all the ammonia-bearing load on the flame. When the heat sink of cold SWAG to the front zone where it passes through uncombusted acid gas is removed from the flame the burner. The front zone operates at a higher by bypassing it to the back zone, the flame temperature to encourage adequate ammonia temperature can be greatly increased. High and contaminant destruction. The AAG, which is temperatures are known to benefit ammonia

rich in H2S but contains relatively little ammo- destruction. The benefit derives from improved

2 PTQ Q4 2016 www.digitalrefining.com/article/1001297 kinetics for the complex and somewhat coun- of ammonia to free nitrogen and water. The ter-intuitive ammonia destruction mechanism second (Reaction 2) is thermal cracking to free described in the next section. nitrogen and hydrogen. Both pathways are well One of the negative aspects of the two-zone documented and both occur in the reaction design is that the fraction of AAG that is being furnace. However, there is another pathway that bypassed to the second zone does not pass has been reported by Alberta Sulphur Research through the burner flame at all. Thus, any Limited6 (ASRL), which usually dominates in an contaminants that may be present are not operating plant. subjected to the high temperatures of the flame. The influence of hydrocarbons and other This can pose a problem if the contaminant contaminants in sulphur plant feeds will be set levels in the clean AAG become excessive. aside so the main focus can be placed on the

Another negative is that the portion of the acid reactions involving H2S and ammonia. It is well

gas that is bypassed has a much shorter resi- established that H2S is present in significant dence time, making it harder to achieve quantities in the front zone where ammonia

equilibrium of the Claus reaction in the furnace. destruction takes place and where H2S oxidises

Finally, since the two-zone design has an addi- to SO2. The kinetics of H2S oxidation have been tional control point for splitting the clean acid shown by ASRL3, 6 to be much faster than the gas flow between the zones, the control system reaction between ammonia and oxygen. Most, if will be inherently more complex. There are also not all of the oxygen present in the flame envi-

safety concerns associated with bypassing acid ronment is consumed by H2S long before gas around the front zone. If too much is ammonia even has a chance to react. It being the bypassed, the temperature of the front zone can case that oxygen is rapidly consumed predomi-

exceed the refractory thermal limits and compro- nantly by H2S, one must ask the question as to mise the integrity of the reaction furnace lining, what oxidises ammonia? ASRL has shown that a and therefore the furnace wall itself. pathway for this oxidative reaction begins with

Two-zone furnaces also require a relatively H2S oxidation, then proceeds to ammonia oxida- large nozzle on the side of the reaction furnace tion as follows: vessel to allow the bypassed acid gas to enter the

back zone. This nozzle has accompanying H2S oxidation (very fast reactions) mechanical challenges associated with construc- tion and maintenance of the refractory lining,

which are avoided in the one-zone furnace ! ! ! ! design. 2 𝐻𝐻 𝑆𝑆 + 𝑂𝑂 → 𝑆𝑆 + 2 𝐻𝐻 𝑂𝑂 𝟑𝟑 ! ! ! ! ! 𝐻𝐻 𝑆𝑆 + ! 𝑂𝑂 → 𝑆𝑆𝑆𝑆 + 𝐻𝐻 𝑂𝑂 𝟒𝟒 Ammonia destruction pathways ! ! ! Adequate destruction of ammonia is extremely 2 𝐻𝐻 𝑆𝑆 → 𝑆𝑆 + 2 𝐻𝐻 𝟓𝟓 ! important for reliable SRU operation because 2 𝐻𝐻 !𝑆𝑆 + ! 𝑂𝑂! → 𝐻𝐻! + 𝑆𝑆! + 𝐻𝐻!𝑂𝑂 𝟔𝟔 insufficient destruction can lead to precipitation NH3 oxidation (fast reaction) of ammonium salts of sulphide, bisulphide, sulphate, sulphite, and bicarbonate. This typi-

cally occurs downstream in the sulphur ! ! ! ! ! condensers where temperatures are lower. It can 2 𝑁𝑁𝑁𝑁 This+ mechanism 𝑆𝑆𝑆𝑆 → 𝑁𝑁 + 2 𝐻𝐻 proposes𝑂𝑂 + 𝐻𝐻 𝑆𝑆 𝟕𝟕that because the

also cause catalyst deactivation. There are two reaction rate of H2S with oxygen is so much ammonia destruction pathways that are faster than between ammonia and oxygen, the

commonly assumed to dominate in the reaction H2S consumes oxygen first, producing SO2

furnace: according to Reaction (4). The SO2 then oxidises ammonia according to Reaction (7). In the ProTreat simulator, the furnace model is based ! 2 𝑁𝑁𝑁𝑁 ! + ! 𝑂𝑂! −→ 𝑁𝑁! + 3 𝐻𝐻 !𝑂𝑂 (1) on reaction kinetics. It uses this mechanism together with data from ASRL’s recent work.

! ! ! This kinetic furnace model also predicts the 2 𝑁𝑁𝑁𝑁The −first→ 𝑁𝑁 pathway+ 3 𝐻𝐻 (Reaction (2) 1) is the oxidation extent of destruction of aliphatic and aromatic

2 PTQ Q4 2016 www.digitalrefining.com/article/1001297 www.digitalrefining.com/article/1001297 PTQ Q4 2016 3 Figure 3 Flowsheet for straight-through reaction furnace with preheat of SWAG and AAG

hydrocarbons on the basis of fundamental reac- combined AAG feed stream, consisting of bulk tion rate data measured by ASRL.5 amine acid gas and tail gas unit (TGU) recycle Because the ProTreat kinetic model is based gas streams, and a typical SWAG composition. on the correct fundamentals, it is a predictive In a one-zone design, all of the AAG is combined tool that can be used to anticipate accurately the with the SWAG prior to entering the reaction effect of process changes. This is a capability furnace. However, in a two-zone design, all of that an empirically fitted model simply does not the SWAG is combined with only a portion of have because such models are the AAG. In actual practice, the based on empiricism, not phys- Acid gas feed compositions portion of AAG that is combined ic-chemical fundamentals, so with the SWAG is determined extrapolation can be quite Component Amine acid Sour water acid through a feed-back control gas (AAG), gas (SWAG), uncertain. mol% mol% loop based on the effluent H O 1.4 25 2 temperature of the first zone. H S 92.1 33.5 Case studies 2 The combined gas is then fed CO2 5 0

Using the ProTreat Sulphur NH3 0 41.5 through the burner and passes CH 0.5 0 Plant Model allows one to look 4 through the front reaction H2 0.0012 0

at a comparative case study N2 0.9988 0 chamber. The front zone efflu- involving both the one-zone and Total flow rate, ent combines with the bypassed lbmol/h 462 200 the two-zone designs in the AAG and enters the back zone context of ammonia destruction. of the reaction furnace. In the The case studies use a typical Table 1 back zone, the hot gas from the

4 PTQ Q4 2016 www.digitalrefining.com/article/1001297 Figure 4 Flowsheet for straight-through reaction furnace without preheat of acid gas feeds

front zone (not a burner) provides the heat dence time required to destroy ammonia required by the reactions. All cases use the feed depends on both temperature and ammonia compositions given in Table 1. concentration in the feed gas. When using the There are two case studies: two-zone design, the front zone often has a • Case study 1: ammonia destruction in a longer residence time than the back to ensure Figure 3 Flowsheet for straight-through reaction furnace with preheat of SWAG and AAG one-zone reaction furnace design is assessed by fairly complete destruction of ammonia. varying the amount of preheat provided to the A one-zone design is modelled in ProTreat AAG and SWAG feeds. Figures 3 and 4 both show using a single reaction furnace block consisting of one-zone designs. Figure 3 shows SWAG and burner plus reaction chamber where the AAG, AAG feed preheaters. Figure 4 shows no SWAG, and combustion air are all mixed together preheating. before entering the furnace itself. The two-zone • Case study 2: ammonia destruction in a design is modelled using two reaction furnace two-zone reaction furnace design is assessed by blocks in series. The front zone (Zone 1) contains varying the percentage of clean acid gas that is a burner and all the SWAG, part of the AAG, and bypassed to the back zone of the two-zone the combustion air are mixed together as Zone-1 furnace. Figure 5 shows the two-zone design. feed. The back zone (Zone 2) does not contain a Downstream of the furnace(s) the flowsheets burner. It is fed with effluent from Zone 1 are all identical and are typical of a three-bed combined with the bypassed portion of the AAG. Claus unit. Reaction furnace parameters are set to typical values: the residence time in the reaction Results and discussion: one-zone furnace chamber is arbitrarily set to 1.1 seconds. The The ammonia concentration for the combined length-to-diameter (L/D) ratio is 3.5 for the SWAG and AAG stream is about 12.5 mol%. This one-zone furnace and for each zone in the is well above the typical maximum ammonia two-zone reaction furnace. concentration for a one-zone design and the The typical residence time for a reaction results show why this is the case. The informal furnace is nominally one second; however, it industry standard for the maximum acceptable should be pointed out that regardless of whether ammonia concentration in any reaction furnace the design is one-zone or two-zone, the resi- effluent is 100–150 ppmv (wet basis). This isto

4 PTQ Q4 2016 www.digitalrefining.com/article/1001297 www.digitalrefining.com/article/1001297 PTQ Q4 2016 5 Figure 5 Flowsheet for the two-zone (front-side split) reaction furnace

prevent the precipitation of ammonium salts in furnace calculation, the ammonia level in the the downstream sulphur condensers. After effluent would be only about 0.5 ppmv- ammo running the simulation in Figure 4 using the nia. The cause of the huge discrepancy is the ProTreat kinetic model, it was found for the equilibrium model’s failure to account for the one-zone furnace without preheat that the actual kinetic rates for each of the important ammonia concentration in the calculated efflu- reactions. An equilibrium model assumes all ent was approximately 343 ppmv. This is two to reactions come to equilibrium – nothing could three times the advised limit. If this same simu- be further from the truth, especially for ammo- lation were run using a simple equilibrium based nia destruction. When preheating was used on both the AAG and SWAG feed 345 streams (but not on the combus- 340 tion air stream) to the one-zone 335 furnace (see Figure 3), to provide a combined acid gas stream 330 temperature of around 380°F 325 (193°C), the ammonia level in 320 the effluent dropped from 343 315 ppmv (no preheat) to 309 ppmv.

Ammonia, ppmv 310 (Figure 6 summarises the results 305 from various levels of preheat.) Preheating of just the combus- 300 0 150 200 250 300 350 400 tion air alone to 450°F (232°C) Preheat temperature, ºF was also studied. (There is nothing to corrode the steel with hot air and there are no hydrocar- Figure 6 Effect of gas preheat temperature on ammonia from one-zone bons to crack, so air preheat can furnace be used to provide higher

6 PTQ Q4 2016 www.digitalrefining.com/article/1001297 temperatures.) This amount of air preheat produced 287 ppmv 260 ammonia. This was better 240 performance than with acid gas 220 preheat alone because the mass flow of the combustion air was 200 considerably larger than the 180 mass flow of acid gas, hence the 160 preheat effect was greater. Both 140

cases left ammonia in the Ammonia, ppmv 120 furnace effluent well above the 100 advised limit; however, the 80 results indicated some ability to 5 10 15 20 25 30 mitigate ammonia levels by Proportion bypassed, % preheating in a one-zone furnace design. With preheat alone, in some cases the ammonia levels Figure 7 Effect of acid gas bypass on ammonia in Zone 2 effluent may be sufficiently reduced to meet the advised limits when using a one-zone reaction 2650 furnace. Another simulation study was 2600 Figure 5 Flowsheet for the two-zone (front-side split) reaction furnace done for the one-zone design with the acid gas preheated to 380°F (193°C), by holding all 2550 variables fixed and varying the residence time in the furnace. 2500 To meet the 100-150 ppmv Temperature, °F Temperature, industry guideline with a 2450 one-zone furnace, more residence time would be required 2400 for this particular case. An 5 10 15 20 25 30 ammonia level of 95 ppmv could Proportion bypassed, % be achieved by increasing the residence time to 1.4 seconds. It is quite possible to destroy Figure 8 Effect of acid gas bypass on temperature in Zone 1 ammonia adequately in a one-zone furnace to acceptable levels and lower. approaches zero, ammonia from the back zone However, this may require the one-zone furnace approaches 275 ppmv. As already seen with the to be outfitted with adequate preheaters as well one-zone furnace, the outlet without preheat was as to have enough volume to provide the neces- approximately 343 ppmv ammonia. The differ- sary residence time. ence in this zero-bypass case results from the mixing characteristics that are assumed in the Results and discussion: two-zone furnace front and back zones of the two-zone furnace. In For the two-zone furnace, an acceptable ammo- the front zone, the burner mixing characteristics nia level was achieved with acid gas bypass of are set equivalent to a high intensity burner by around 30%. The ammonia concentration in the selecting this option in ProTreat. This selection effluent from Zone 2 was approximately 90 ppmv accounts for mixing imperfections in the cham- on a wet basis at this percentage bypass. In Figure ber of a commercial size unit but with burners 7, the concentration of ammonia in the effluent having good mixing characteristics. The back from the back zone is shown as a function of the zone is set to the perfectly mixed option because percentage bypass. As the bypass percentage there is no burner located in that zone.

6 PTQ Q4 2016 www.digitalrefining.com/article/1001297 www.digitalrefining.com/article/1001297 PTQ Q4 2016 7 the furnace, and proper mixing of combustion air with the acid 100000 Zone 1 gas entering the burner. It is Zone 2 important to ensure an almost 10000 completely homogeneous mixture of combustion air and acid gas to ensure complete and 1000 even combustion of the acid gas and complete ammonia destruction. 100 Conclusions Effluent ammonia, ppmv Effluent There are many considerations 10 pertinent to each reaction 0.8 0.9 1.0 1.1 1.2 1.3 1.4 furnace design. The one-zone Total residence time, seconds furnace is usually able to relieve the rest of the SRU adequately Figure 9 Effect of total residence time in the two-zone furnace on ammonia from hydrocarbon contaminants in effluent from each zone with 30% bypass by ensuring all the acid gas passes through the high temper- As the bypass percentage is increased, the atures of the burner. Compared to the two-zone concentration of ammonia from the back zone reaction furnace, the acid gas controls are decreases almost linearly. This is due in part to simpler. One-zone furnaces generally require the increasing temperatures in the front zone, as more equipment upstream of the furnace to seen in Figure 8. Additionally, the concentration ensure the acid gas is adequately preheated to

of SO2 is increasing in the front zone. At a ensure a stable flame. Conversely the two-zone bypass of around 30%, the temperature furnace has some further mechanical and main- approaches the upper end of the minimum tenance challenges associated with the additional acceptable temperature range (2300–2700°F, nozzle in the reaction furnace vessel. 1260-1480°C). As discussed earlier, higher The two-zone furnace will generally achieve temperature is achieved through bypassing the higher ammonia destruction compared to a

portion of H2S that does not need to be single zone. This results from the ability to

converted to SO2 and which would have acted as bypass clean AAG around the burner into the a heat sink had it been allowed to enter the first second zone, which raises the flame temperature zone. Front zone temperature increases with in the first zone. However, to achieve the higher increasing bypass. Ammonia destruction kinet- level of ammonia destruction, a more compli- ics, of course, respond exponentially to higher cated control scheme is needed to control the temperature, providing greater extent of bypass, primary acid gas, and the flow rate of reaction. combustion air. This is not to say a single-zone Another interesting aspect of two-zone furnaces furnace is unable to achieve adequate ammonia is the effect of residence time on the ammonia destruction because with the addition of preheat- concentrations in the effluent from the two ers and increased residence time, a one-zone furnace elements. As shown in Figure 9 for a furnace is able to achieve the desired destruction two-zone furnace, ammonia concentrations from levels. This might be a trade-off against the each zone decrease as the total residence time in operating complexity and bypass cost of a the two combined zones increases. Over the range two-zone system. of practical residence times (0.8–1.4 seconds), A two-zone furnace increases the likelihood of ammonia destruction trends exponentially with subjecting the SRU to contaminants that are residence time. present in the otherwise clean (ammonia free) Three important factors play a large role in the amine acid gas being bypassed around the adequate destruction of ammonia in the reaction burner unit. Bypassing aliphatic and aromatic furnace, namely, residence time, temperature in hydrocarbons around the flame can have devas-

8 PTQ Q4 2016 www.digitalrefining.com/article/1001297 tating effects on downstream catalyst beds. – Sulphur Recovery, Laurence Reid Gas Conditioning Conference, Besides the temperatures in the furnace, Norman, Oklahoma, 2007. another key parameter in ammonia destruction 5 Clark P D, Dowling N I, Huang M, Chemistry of the Claus front- is the residence time in the reaction chamber. As end reaction furnace. hydrocarbon reactions and the formation and destruction of CS , ASRL Quarterly Bulletin, Vol XXXIII, No. 4, shown in Figure 9, increasing the total residence 2 1-49, Jan-Mar 1997. time, whether in a single- or two-zone furnace, 6 Clark P D, Dowling N I, Huang M, Mechanisms of ammonia has a significant effect on the ammonia concen- destruction in the Claus front-end furnace, ASRL Quarterly trations in the effluent gas – of course, the Bulletin, Vol XXXIV, No. 4, 1-50, Jan-Mar 1997. impact of residence time will only be seen in a furnace model that includes chemical reaction G Simon Weiland recently joined Optimized Gas Treating as a kinetics. By using the ProTreat SRU model, Software Development Engineer. He holds a BS in chemical highly realistic simulations that take into engineering from the University of Oklahoma. account deep-seated chemistry and reaction Clayton E Jones joined Optimized Gas Treating, Inc as a Software kinetics vs just equilibrium can be used to inves- Development Engineer in 2012. He holds a BS in chemical tigate different reaction furnace designs engineering from McNeese State University and a MS in chemical quantitatively. Using a fundamentals basis engineering from the University of New Mexico. Nathan (Nate) A Hatcher is Vice President, Technology greatly improves model reliability and realism, Development for OGT. He holds a BS in chemical engineering and it results in a model that is truly a virtual from the University of Kansas. plant capable of revealing new insights into Email: [email protected] complex processes. Ralph H Weiland is President of Optimized Gas Treating, Inc (OGT). He holds BASc, MASc and PhD degrees in chemical engineering ProTreat is a registered mark of Optimized Gas Treating, Inc. from the University of Toronto. Email: [email protected] References 1 Anonymous, Refineries face issues of increasing ammonia, Sulphur, Mar–Apr 1995. LINKS 2 Grancher P, Advances in Claus Technology; Part 1: studies in More articles from: Optimized Gas Treating reaction mechanics, Hydrocarbon Processing, Jul 1978. 3 Li D, Dowling N I, Marriott R A, Clark P D, Kinetics and More articles from the following categories: Mechanisms for Destruction of Ammonia in the Claus Furnace, Gas Processing and Treatment Alberta Sulphur Research Ltd. Chalk Talk, Calgary AB, Canada, Jan Sulphur Removal and Recovery 2016. Process Modelling and Simulation 4 Hatcher N H, Nasato E, Chow T, Huffmaster M, Fundamentals

8 PTQ Q4 2016 www.digitalrefining.com/article/1001297 www.digitalrefining.com/article/1001297 PTQ Q4 2016 9