FCC CORROSION-CONCLUSION BASIC CORROSION CONTROL METHODS SOLVE VARIED PROBLEMS

Russell C. Strong, Veronica K. Majestic
Nalco Chemical Co.
Sugar Land, Tex.

S. Mark Wilhelm
Cortest Laboratories Inc.
Cypress, Tex.

Every fluid catalytic cracking (FCC) corrosion-control problem is different, but several basic principles should be understood and applied as needed.

These principles were used effectively to solve corrosion problems in three U.S. refineries.

CORROSION-CONTROL METHODS

The corrosion-control method chosen for any system will depend on both the chemical nature of the corrosion and the physical characteristics of the system. A good water wash system is the key to any FCC corrosion-control program.

Ideally, an effective water wash will at least move vapor-phase corrosive materials into the liquid phase, where they are more easily treated.

If the water wash does not provide complete protection, polysulfides and filming inhibitors can be added.

WATER WASH

An effective water wash dilutes and scrubs corrosive species-such as hydrogen sulfide, ammonia, chlorides, and cyanides-from the FCC offgas. The water wash rate, location, water quality, and injection equipment are critical to an effective program.

It may be necessary to inject wash water at several points, because of the limitations imposed by liquid route preference and flow regimes.

Corrosive gases are more soluble in water at high process pressures. Therefore, the best location for a water wash is downstream of each stage of gas compression, ahead of the coolers (Fig. 1).

If there are multiple banks of compressor aftercoolers, a water wash must be added to all exchanger banks.

A water wash in the main fractionator overhead vapor line can also be useful as a first step in removing volatile acid gases. Although the solubility of these gases may be low at the fractionator pressure, the ability to add a high volume of water for the relatively low cost of recycling the water from the reflux drum boot makes this an attractive option.

The conventionally recommended water wash rate into the gas compression stages is 1.0-1.5 gpm/1,000 bbl of oil feed to the FCC unit (FCCU). This rate, which translates to about 2 gpm/MMscf of gas, should be verified through flow regime calculations.

The objective is to create a stable mist for maximum surface-to-volume ratio of the water droplets and, hence, maximum contact efficiency. The wash water rate should be reliable and easily measured.

The water should be injected through a spray nozzle that can aid the flow kinetics required to achieve maximum surface-to-volume ratio. If the vapor and liquid rates are not in the correct range, the liquid will return to a stratified or wave flow.

The feed system should be designed for continuous water flow. Recommended wash water sources include boiler feed water (before alkalinity adjustment), steam condensate, and stripped process condensates.

Recycling from high to low pressure, also known as cascading water washes, is not recommended because it may create a "champagne effect," concentrating the corrosive species in the water and thereby aggravating corrosion.

A water wash alone does not always protect completely against hydrogen activity in FCCU fractionation trains. It is often necessary to supplement the water wash with chemical additives, such as ammonium or sodium polysulfide and filming inhibitors.

POLYSULFIDES

Polysulfide is injected into a water wash to react with hydrogen cyanide, producing thiocyanate. Because thiocyanates do not react with the protective sulfide film on metal surfaces, as cyanide does, hydrogen penetration is reduced.

Either sodium or ammonium polysulfide can be used, but ammonium polysulfide is most common. Concentrated solutions are red, while dilute solutions are yellow.

A number of considerations affect the use of polysulfides, including their susceptibility to oxidation, their instability at low pH, and their thermal instability.

Exposure to air oxidizes polysulfide to thiosulfate and sulfate, neither of which forms the desired thiocyanate. A change from red or yellow to colorless suggests oxidation of the polysulfide.

Based on their potential for oxidation, care should be taken to prevent polysulfides from being contaminated with oxygen. The polysulfide solution should be blanketed with hydrocarbon both to prevent odor problems and to reduce oxygen contamination.

Concentrated ammonium polysulfide is soluble in alkaline sour water. If the pH drops below 8, the polysulfide will decompose into ammonia, hydrogen sulfide, and sulfur. The formation of sulfur fouls the equipment.

Ammonium polysulfide is thermally unstable and very difficult to handle and pump in cold weather. Care must be taken when the solution temperature falls below 38 F. because the polysulfide precipitates at this temperature and at temperatures above 250 F., causing precipitation of excess sulfur.

Some refiners make ammonium polysulfide in situ by injecting air into the wash water, forming polysulfide in a process similar to a Claus reaction.

This technique raises concerns about the contribution of oxygen from the air to corrosion activity. In addition, the oxygen could "vapor pool," creating a potential for explosion when the equipment is opened.

Polysulfide reacts with cyanide in an equimolar ratio. Unfortunately, cyanides are difficult to measure in process waters, making the dosage of polysulfide difficult to regulate.

The best standard for determining the dosage is the appearance of the process water from the water knockouts. It should have a distinct transparent, pale-yellow color. Excess polysulfide can increase corrosion at elevated temperatures and, because of its instability, can foul process equipment.

CORROSION INHIBITORS

Most FCC corrosion inhibitors are filmers, which provide a thin barrier of organic material on the inside surface of the equipment at risk. Dosages are only a few parts per million, based on the mass flow rate of the treated process.

If effective, this thin barrier will prevent an aqueous phase from reaching the surface, thereby preventing the corrosion that can cause hydrogen permeation or carbonate cracking.

Some inhibitors are passivators, which interact chemically with the surface to anodically or cathodically inhibit the electrochemistry of corrosion. Passivators may not form a complete organic barrier, but eliminating the cathodic or anodic site interrupts the corrosion.

To date there are no commercially successful vapor-phase inhibitors.

The effectiveness of several inhibitors was evaluated by Cortest Laboratories Inc., an independent testing firm. The experimental procedure used an electrochemical cell divided into two compartments by a thin (0.03 in.) steel membrane.

One side of the membrane was exposed to a sour FCC environment. Hydrogen atoms resulting from the corrosion reaction diffused through the membrane and were electrochemically oxidized in the second compartment.

The amount of hydrogen permeating the membrane indicates the rate of corrosion and the efficiency of the permeation step.

The experimental data provide a measure of the rate of proton discharge (corrosion rate) and the ratio of atoms absorbed to atoms generated (permeation efficiency).

An effective inhibitor can function by reducing the corrosion rate, by lowering the permeation efficiency, or by doing both.

In a typical corrosion inhibitor evaluation, a range of concentrations is examined over a range of temperatures. The corrosion rates and permeation efficiencies in inhibitor solutions are compared with those in "blanks."

The tests are run in solutions containing ammonium, cyanide and sulfide ions, and trace impurities peculiar to the location of the sour process stream.

Tests were conducted on several proprietary FCC treatment formulations: two oil-soluble inhibitors, a water-soluble inhibitor, and a water-soluble passivator.

The four formulations are described as follows:

Inhibitor A: Oil-soluble diamide inhibitor
Passivator B: Water-soluble passivating inhibitor
Inhibitor C: Oil-soluble imidazoline inhibitor
Inhibitor D: Water-soluble quaternary amine inhibitor.

Most of the compounds tested lowered the corrosion rate substantially (Fig. 2). But not all produced interfacial conditions that reduced hydrogen absorption or permeation (Fig. 3).

This difference has important ramifications in selecting efficient treatment chemicals. Mixtures can be designed containing ingredients specific to both corrosion rate and hydrogen absorption.

Once the physical limitations of the system are established so that the chemical can be applied with maximum contact effectiveness, a product can be selected.

Factors that should be considered when choosing an inhibitor are:

The nature of the problem (general corrosion or hydrogen permeation)
The product's demonstrated ability to inhibit activity at high pH
The product's oil or water solubility, depending on the primary phase in which the problem occurs
The product's solubility in light ends, if oil based
The product's emulsification or foaming tendencies.

CASE STUDIES

The following case studies demonstrate the importance of various factors in developing a corrosion-control program for the FCCU. Factors of primary concern in Case Studies 1, 2, and 3, respectively, were chemistry and phase partitioning, flow regime, and process monitoring and maintenance.

NO. 1: TEXACO, ANACORTES

Texaco U.S.A.'s Puget Sound refinery in Anacortes, Wash., experienced a sudden onset of severe general corrosion in the absorber-bottoms piping (Fig. 1).

Corrosion rates were at times in excess of 400 mils/year, as measured by electrical resistance probes, and the piping was becoming dangerously thin.

The company formed a Quality Action Team (QAT) to investigate and solve the problem. The team included members from various operating departments and management, as well as the local Nalco representative and representatives of the Nalco and Texaco research groups.

Historical and current data collected, collated, and discussed by the QAT indicated that there had been no similar corrosion in the history of the unit and no changes to the feedstock or unit operations.

Visual inspection of the piping showed severe pitting, similar to what might be seen with aggressive dew-point corrosion. However, very little corrosion-product deposit was found in the piping or in the pits.

Although the corrosion was extremely aggressive, no organic acid was found. Thus, despite the high temperature of the absorber bottoms, the QAT concluded that the corrosion was aqueous rather than organic.

The QAT identified the root cause of the problem as ineffective coalescer operation in the wet lean oil.

In the presence of high concentrations of H2S and CO2, the water flowing downward to the absorber bottoms was creating a severe general corrosion problem. Only a high volume of water introduced into the tower could reach the bottom piping where the corrosion was occurring.

The solution was to:

Improve maintenance of the coalescer.
Improve monitoring of coalescer performance.
Apply an oil-soluble inhibitor effective at inhibiting general corrosion under the conditions found in the FCC.

Items 1 and 2 were addressed by the operations and research members of the QAT, and Item 3 was addressed by applying Inhibitor A (Figs. 2 and 3). The problem is now under control.

NO. 2: CHEVRON, PHILADELPHIA

Chevron U.S.A. Inc.'s Philadelphia refinery was experiencing blistering in the light-ends fractionation equipment (Fig. 1).

The most evident indication of this hydrogen charging problem was concentration of ferrocyanides in the tower overheads. The problem was severe enough to require re.placement of the debutanizer and depropanizer condenser shells.

The existing corrosion-control program consisted of a water wash in the interstage coolers and treatment with filming inhibitors and ammonium polysulfide.

With these measures in place, the cyanides in the compressor section should have reacted and been scrubbed and should never have reached the downstream fractionation unit.

A study of the piping and process flows indicated that the polysulfide wash might be ineffective for two reasons:

The flow regime was a wave pattern, which prevented good water-vapor contact in the piping leading to the coolers.
The water route preference forced the water predominantly through one set of coolers, thereby preventing contact with the vapors in a parallel set of coolers.

This study resulted in two recommendations for process changes.

First, because of the diameter of the piping, the water rate was increased to above normally accepted standards. This higher rate encourages an annular distribution of water and improves the efficiency of water and polysulfide contact with the vapors.

Since the water rate was increased several months ago, cyanide levels have dropped 40-50%, as shown in Fig. 4. During this time, the polysulfide rate was also varied, with less effect.

Second, the point of polysulfide injection will be changed to overcome the limitations imposed by liquid route preference. The piping needed to improve distribution is in design. Once it is installed, cyanide levels are expected to drop another 4050%.

NO. 3: MARATHON, GARYVILLE

Because of a wet feed, Marathon Oil Co.'s refinery in Garyville, La., operates an unsaturated gas plant absorber column with a total-draw tray to remove water from the tower. For years, this tower required no corrosion-control program.

The absorber column environment is well-monitored, with hydrogen probes on the compressor knockout pots feeding the tower, on the absorber intercooler coalescer, and at other locations (Fig. 5). All the probes typically showed little, if any, activity.

In late 1988, the hydrogen probe in the absorber intercooler coalescer began to show an increase in hydrogen permeation rate. This change suggested that the potential for hydrogen-charging corrosion was increasing.

There was no apparent cause for the increased activity, and none of the other probes showed a comparable response. All process flow rates and compositions had consistently remained within normal ranges.

When the problem was first observed, the dosage of oil-soluble filming inhibitor to the main fractionator was increased, without effect. Next, the inhibitor was injected directly into the feed at 6 ppm.

This inhibitor had been effective at reducing hydrogen permeation in absorber towers at this location and many others, so it was expected to work in this situation. After 3 weeks it was concluded that the inhibitor was ineffective against this problem, despite its track record.

Next, a water-soluble inhibitor that had inhibited corrosion in sour water strippers - a similar environment - was tried.

The side effects of the water-soluble inhibitor prevented its use for more than a few hours. (Many water-soluble inhibitors are surfactants that can seriously aggravate foaming and emulsification, and that was true here.)

Marathon and Nalco engineers agreed that the use of a water-soluble inhibitor was necessary because of the location of the active probe and the flow kinetics in the unit.

Cortest Laboratories' research data showed that one water-soluble product was very effective at inhibiting hydrogen permeation. Further testing showed no negative side effects of foaming, emulsification, or salting.

This product, Passivator B from Figs. 2 and 3, was applied as indicated in Fig. 5, and monitored closely to ensure that bench studies had modeled the system accurately.

Results showed an immediate reduction in hydrogen probe activity (Fig. 6).

For the 37 days prior to treatment with the new chemical, the average increase in pressure was 0.6 psig/day. For the first 1 00 days of treatment, chemical dosage was varied between 9 and 12 ppm to find the optimum level, and the activity was cut by almost an order of magnitude to 0.08 psig/day.

Dosage was fixed at 12 ppm, and the activity over the next 177 days was cut in half again to 0.04 psig/day.

As this case shows, any chemical approach to solving a corrosion problem must take all the discussed principles into account. The resulting solution should be proven to be effective in similar laboratory or field environments before testing in situ.

Editors note: At the request of the refineries mentioned in this article, any questions about the case histories should be directed to the authors.

BIBLIOGRAPHY

Neumaier, B.W., and Schilmoller, C.M., "Deterrence of Hydrogen Blistering at a Fluid Catalytic Cracking Unit," API Publication, Vol. 35, 1955, pp. 92-108.

Gutzeit, Joerg, "Corrosion of Steel by Sulfides and Cyanides in Refinery Condensate Water," Materials Protection, Vol. 7, No. 12, December 1968, pp. 19-23.

Skei, T., Wachter, A., Bonner, W.A., and Burnham, H.D., "Hydrogen Blistering of Steel in Hydrogen Sulfide Solutions," Corrosion, May 1953, pp. 163-72.

Merrick, R.D., "Refinery Experiences with Cracking in Wet H2S Environments," Materials Performance, January 1988, pp. 30-36.

Kmetz, J.H., and Truax, D.J., "Carbonate Stress Corrosion Cracking of Carbon Steel in Refinery FCC Main Fractionator Overhead Systems," NACE Paper 206, Corrosion/90, Apr. 23-27, 1990, Las Vegas.

Baker, O., "Designing for simultaneous flow of oil and gas," OGJ, July 26, 1954, pp. 185-95.

Oranje, L., "Condensate behavior in gas pipelines is predictable," OGJ, July 2, 1973, pp. 39-44. Fortuin, J.M.H., Hamersma, P.J., Hart, J., Smit, H.J., and Been, W.P., "Calculations predict condensate movement at 'T' junctions, " OGJ, Jan. 21, 199 1, pp.3740.

Fortuin, J.M.H., Hamersma, P.J., Hart, J., Smit, H.J., and Been, W.P., "Experiments verify predictions of condensate movements," OGJ, Jan. 28, pp. 91-93.

Kane, R.D., Wilhelm, S.M., and Oldfield, J.W., "Review of Hydrogen Induced Cracking of Steels in Wet H2S Refinery Service," International Conference on Interaction of Steels with Hydrogen, Mar. 2830, 1989, Paris.