JPT logo
SPE logo

Technology Update

Oxidation of Mercaptan, Dimethyl Sulfide in Aqueous Waste

Tert-butyl mercaptan (TBM) and dimethyl sulfide (DMS) are odorous compounds added to natural gas to warn of a gas leakage so that an alarm can be activated. In a storage project for gas produced from platforms offshore Ireland, TBM and DMS were expected to be present in a low-volume aqueous waste stream. Project developers designed an economical process to treat these compounds so that the stream could be accepted at a water treatment facility. Several permutations of plant design and disposal routes were evaluated, and an advanced oxidation process trial was found to yield a straightforward process for mercaptan treatment.

Background

The business model of seasonal gas storage is to inject natural gas into an underground reservoir when gas is relatively cheap (summertime) and withdrawing the gas when the price is higher (wintertime). In this project, the injected gas is odorized with TBM and DMS to 4.7 ppmv and 1.3 ppmv, respectively. The proposed offshore storage reservoir is tied to the shore terminal by a 34-mile pipeline.

To inhibit hydrate formation, monoethylene glycol (MEG) and methanol (MeOH) will be injected into the withdrawn gas. The gas exiting the reservoir is expected to be water saturated and water will condense because of the loss of pressure as the gas travels up the wellbores and through the pipeline. At the shore terminal, the aqueous MEG/MeOH will be separated from the gas stream for recovery. The aqueous waste stream must be treated before disposal. Because the injected gas will contain TBM/DMS, the withdrawn gas and waste stream will contain these compounds, too. The waste stream will also contain trace amounts of native petroleum species (alkanes, aromatics, ammonia, etc).

The waste stream is expected to have relatively low-volume flow rates on the order of 1 m3/d to 10 m3/d. This flow rate is seasonal and can change daily, even within the season, in line with fluctuating customer nominations. There is also considerable uncertainty at the upper flow limit, which makes it difficult to size the plant throughput.

Initial Treatment Options

The waste treatment options and disposal routes considered by the process developers were

  • Basic site treatment and sea outfall
  • Site treatment and ground infiltration
  • Site treatment and surface discharge
  • Collection and off-site treatment by a third party.

The site treatment methods were

  • Biological aerobic
  • Biological anaerobic
  • Aerobic/anaerobic combinations
  • Membrane or oxidative treatment

The sea outfall disposal route was discounted because the low flow rate made it uneconomic.

Site treatment was discounted because the a priori determination of waste flow rates was uncertain and could easily lead to overcapacity or undercapacity design. Furthermore, the combinations of biological treatment were discounted because of concerns about maintaining a viable biomass during the injection season, and maintaining a system to contain mercaptan odors that could raise erroneous gas leak warnings. Despite high confidence that biological methods could treat the waste, they were discounted on practical grounds.

Membrane and oxidative treatments of the waste on site would be expensive because of the MEG and MeOH loading and were discounted.

Complex combinations of the above processes were not considered feasible for a seasonally operated plant with low throughput.

The best option was to transfer the aqueous waste to a third-party treatment plant that cannot accept a waste with mercaptan. Waste handling experience from other industries locally was that waste apparently free of mercaptan odor at the time of leaving the waste producing site could release odors in the tanker during transport, or release a trace mercaptan upon opening the tanker lid at the receiving treatment facility. The project, therefore, required a selective destruction method for the mercaptan.

Mercaptan Destruction Process

Many of the options for selective mercaptan destruction either contained the same inherent weaknesses as the site treatment methods or transferred the waste into a different form with its own subsequent handling issues. None could guarantee that low parts-per-trillion mercaptan concentrations could be attained. For example, sparging the liquid waste and treating the off gas with a two-stage carbon adsorption was considered. The questions raised were (a) the process cannot guarantee total mercaptan removal, (b) the gas carries the methanol, increasing the load on the carbon, and (c) the carbon waste is still a waste stream and a handling concern, and remains an operational overhead, but in a different category. For example, chemical quenching with hypochlorite was discounted because of its potential to generate halomethanes.

Advanced oxidation processes have been researched (Hwang et al. 1994). The indications from published trials indicate that mercaptan can be destroyed by oxidants: ozone (O3) and/or hydrogen peroxide (H2O2) in combination with ultraviolet (UV). Other research contains reaction profiles for glycols (Turan-Ertas and Gurol 2002; Takahashi and Katsuki 1990) and alcohols (phenol) (Mokrini et al. 1997). A comparison of the destruction profiles indicates destruction in the order of one to several hours. Given that MEG and MeOH were expected to be present in concentrations of 100 mg/l to 1000 mg/l and TBM/DMS concentration at ~1 mg/l range (in the liquid phase), there was concern that MEG/MeOH would be significant competitors for the oxidant. A trial was required to verify the destruction of TBM/DMS in the presence of MEG and MeOH. Because of the expected competitive behavior, a comprehensive pilot trial was developed.

Pilot Trial Setup

A pilot plant was sourced from a commercial vendor. The process schematic diagram is shown in Fig. 1. The plant is able to dose O3, H2O2, or (O3+H2O2) in a side stream pumped from a small 200 L agitated tank. The side stream also passes through an UV lamp. The O3 generator requires an oxygen bottle and can produce up to 8 g/hr ozone. The holding tank was agitated and the pump flow rate was ~1 m3/h to 1.2 m3/h.

Trial batches were made up to 50 L, using preprepared vials containing aliquots of chemicals to yield the desired concentrations. The trial was split into three phases. Phase 1 was designed to prove the concept of chemical oxygen demand (COD) destruction of MEG/MeOH using the combinations of oxidative reagents, and establish the ability of the oxidation to treat the waste sufficiently to allow on-site ground infiltration. Phase 2 was designed to determine the destruction behavior of odorant in competition with MEG/MeOH and other petroleum species. This was expected to be an iterative process with some degree of trial and error to define the optimum process. Phase 3 was to verify the destruction of the odorant using the selected process scheme with confirmation by certified laboratories.

Trial Results

Phase 1 was designed to confirm the basic function of the process and establish background rates of destruction for MEG/MeOH (Hwang et al. 1994). Methanol at the target concentrations was largely unaffected by the oxidation processes. Batches were run for up to two hours and achieved destruction of ~50%. An anomaly was discovered: the H2O2 apparently increased COD. Information searches revealed* a known interference effect of the standard COD reagent with H2O2. As COD was not the prime trial target, this phase of testing was discontinued. The testing moved to Phase 2, the trial of the oxidation of mercaptan.

Phase 2 dosed mercaptan and then mercaptan and petroleum species into MEG/MeOH solutions. The trials were carried out in a remote area to avoid public concern over a possible odor release and a false alarm. Vials were opened under liquid surface to prevent gas escape, and equipment was rinsed with hypochlorite to destroy the mercaptan odor. The trial equipment was also placed under a fume hood with an extract fan and filter containing KOH/KI-impregnated activated carbon.

Phase 2 proceeded successfully. In the first test, the mercaptan odor was not detected by operators after 30 minutes. Repeatable results were obtained for varying solution strengths; increasing the background COD from 28 mg/l to 2400 mg/l did not affect the mercaptan destruction rate (Fig 2). Attempts to identify a rate of reaction were not possible, because the reaction proceeded faster than one residence time. The slow COD destruction rate at higher COD concentrations meant that there was no point extending the process time to achieve a significant COD reduction. This did not infer that the MEG/MeOH was not destroyed, but that COD remained, likely from intermediate products of the reaction (Turan-Ertas and Gurol 2002).

Phase 3 proceeded immediately. Two batches were processed for 120 minutes to ensure repeatability and to collect redundant samples. The analysis confirmation approach was threefold:

  • Liquid samples were taken at time intervals and analyzed for mercaptan to confirm its destruction.
  • The reaction was stopped after 120 minutes and the batch transferred to a barrel for headspace analysis using graphite adsorption tubes.
  • Liquid samples were taken at time intervals. They were stored and subsequently sampled by a volunteer odor panel, which confirmed the destruction of the mercaptan within 60 minutes and, hence, the process.

Conclusion

The advanced oxidation process using ozone/UV rapidly destroyed mercaptan in an aqueous waste containing methanol, monoethylene glycol, and petroleum species. The oxidation process did not destroy COD at a sufficient rate at the high target concentrations.

The authors acknowledge the assistance of Henk van der Puil Treatment Systems.

References

  • Hwang, Y., Matsuo, T., Hanaki, K., and Suzuki, N. 1994. Removal of Odorous Compounds in Wastewater by Using Activated Carbon, Ozonation, and Aerated Biofilter. Water Research 28 (11): 2309–2319. http://dx.doi.org/10.1016/0043-1354(94)90046-9
  • Mokrini, A., Ousse, D., and Esplugas, S. 1997. Oxidation of Aromatic Compounds With UV Radiation/Ozone/Hydrogen Peroxide. Water Science and Technology 35 (4): 95–102.
  • Takahashi, N. and Katsuki, O. 1990. Decomposition of Ethylene Glycol by the Combined Use of Ozone Oxidation and Electrolytic Methods. Ozone: Science and Engineering 12 (2): 115–131. http://dx.doi.org/10.1080/01919519008552214
  • Turan-Ertas, T. and Gurol, M.D. 2002. Oxidation of Diethylene Glycol With Ozone and Modified Fenton Process. Chemosphere 47 (3): 293–301.

* http://www.h2o2.com/technical-library/analytical-methods/default.aspx?pid=
75&name=Analytical-Interferences-Caused-by-Residual-Peroxide