THE AMERICAN INSTITUTE O F M I S I S G AsD IIETALLURGICAL ENGINEERS No. 1620-G. ISBCED ~ I T H I I I S I N G AND IIETALLURGY, DECEMBER, 1926. [ ~ B J E C T T O REVIBION

DISCC'SSION OF THIS PAPER IS INVITED. I t should preferably be presented in person a t the New York Meeting, February. 1927, when an abstract of the paper will be read. If this isirnpos- sible, discussion in writing may be sent to the Editor, Arnericsn Institute of Mining anp MetalIurgical Engiheers. 29 West 39th Street. New York, N. Y., for presentation by the Secretary or other representa- tive of its author. Unless special arrangement is made, the discussion of this paper will close March 1, 1927. Any discussion offered thereafter should preferably be in the form of a new paper.

Corrosion in an oil Refinery BY H. F. PERKINS,* PORT ARTHUR, TEXAS

(New York Meeting, February, 1927)

CORROSIOS as an economic problem is growing rapidly in importance not only because it entails a replacement of corroded parts, but because i t interrupts operation and causes hazards of damage and injury through failures of corroded equipment. Nearly all industries are interested in this subject. This paper, although concerned chiefly with corrosion in a complete refinery on the sea coast, discusses many of the various forms of chemical decomposition, most of which are found universally.

The petroleum refining industry, in the writer's opinion, is seriously afflicted by the presence of sulfur in nearly all crude oils. The complete removal of sulfur from the crude as it enters the plant would cause such a relief from corrosion that this problem for research would seem t o demand active attention. Next t o sulfur the most troublesome corrosive agent is the salt. magnesium chloride, found in the water associated with crude oil, the last traces of which are so difficult to remove. This is much less serious, however, than the sulfur. Refinery corrosion can be tabulated as follows:

Causes of Corrosion and Places of Attack i n a ReJi,nery CORROSIVE ELEMENTB MATERIAL ATTACKED

Air : Pipe lines hloisture Tanks COX Electric fixtures Sulfur gases Still towers Chlorine Condenser boxes Other acid-forming gases Ladders and stairs

Smoke stacks Steel buildings All exposed metal

Water: A1k:tli Fire lines and hydrants Acids - Cooling water lines Air Condensers and coolers Other dissolved gases Boiler tubes Steam Steani pipes and valves

Turbines Oil :

Sulfur compounds Tanks, pipes, valves, stills, towers MgC12-Hydrochloric acid Condensers

Air contamination

* Gulf Refining Co.

Copyright, 1926, by the American Institute of Mining and Metallurgical Engineers, Znc.

2 CORROSION I N AN OIL REFINERY

In addition, there is the corrosion of equipment in more or less localized operations such as that caused by caustic, sulfuric acid and other chemi- cals that may be used for refining. Corrosion caused by the commonly used chemicals is omitted in this paper, as the chemical industry has made special studies of them; their chemical composition is definitely known and is under control.

The distinct corrosion problem in oil refining is the variable nature and variable concentration of the active chemical elements, and lack of knowledge as to just what the elements sometimes are.

The mechanism of corrosion is being better understood every day and .in view of the large amount of recent literature on this subject, only a brief statement of the main controlling factors is given. Character- istics that influence corrosion are important to the understanding of this discussion.

Corrosion may be classified by causes as follows: 1. That due to so-called natural water in which oxygen plays the

most important part. 2. That due to dilute and very dilute acids where oxygen is still

important but the hydrogen ion concentration is most important. 3. That due to acid where hydrogen is evolved and oxygen plays but

little or no part. In general, corrosion is influenced by hydrogen ion concentration,

percentage of chemical dissociation, electrical conductivity, temperature, velocity of flow a t corroding surface, nature of corrosion products, electro-chemical potential of metals and solutions and overvoltage of hydrogen.

Sulfur

If the sulfur that occurs naturally in the crude is followed through the refinery, the first noticeable effects are found from gases such as HzS absorbed in the oil coming off during the light cuts. Probably a dozen forms of sulfur, including free sulfur, are found in most crude oils. These compounds have different degrees of corrosive activity, individual boiling points and dissociation temperatures, so that sulfur gases in some degree may be expected all through distillation down to coke.

Corrosion does not seem to be serious in the absence of liquid water. As far as the distillation apparatus is concerned there would be little to worry about if water could be kept out of the vapor lines and condenser coils. Unfortunately, water is present from three sources: steam used in distillation, natural water in the crude and leaks in condenser coils. Water is usually separated from the condensed oil a t the receiving house, so corrosion beyond this point is not severe.

II. F. PERKINS 3

Any place in the vapor systems a t a temperature low enough to coli- dense water vapor will be subject to strong corrosion. As high tempera- tures accelerate the action, the most destructive corrosion would be expected a t the point where water vapor first condenses. This is con- firmed by experience.

Furthermore, in most cases of refinery corrosion the presence of water is necessary for attack to take place. This is because the active agent is usually in the gaseous form and inactive until absorbed by water when an active acid is produced.

Fractions of oil distilled over and bottoms, even coke, will contain some sulfur or its compounds. Some of the free-sulfur gases will go to the gasoline recovery plant where, with a small amount of water vapor present, they attack the compressor, valves and cooling coils. Where the absorption method is used, the absorbing oil becomes saturated with sulfur, and if the pressure is reduced on this oil the slight cooling effect is sufficient to precipitate free sulfur which will clog the pipes and valves. Sulfur in the gas hinders to some extent the complete recovery of gasoline present.

Where the gas from stills is of small value, the sulfur gases are exhausted to the atmosphere and add to its corrosive effect.

Fuel oil is apt to contain considerable sulfur which is converted to gas during combustion and is delivered to the atmosphere. The oils that are treated to remove sulfur, have to be redistilled. Sometimes sulfur gases come off in large quantities causing serious corrosion. The crude or fraction of crude that is cracked in pressure stills gives off sulfur gases which in their dry state a t the high temperature used are very active on iron

Sulfur gases arising from oil in storage combine with the condensed moisture from atmosphere on the inside of tanks, principally on the roof. Compounds are formed with iron that cause an increase in temperature. I t has been suggested that the temperature under special conditions might reach the ignition point and be responsible for some mysterious tank fires. A sprayed coating of some immune metal might prevent this condition.

Hydrochloric Acid

Hydrochloric acid is formed, probably from the decomposition of magnesium chloride found in the water associated with crude oil, when enough water is present. This acid usually is disposed of early in the refining process but where sulfur gases are also present, the combination is very active, as insoluble compounds formed by the sulfur acids will be acted on by the hydrochloric acid. However: the crude is freed from this corrosive agent usually by the time the gas oil is removed and gives no more trouble except to add activity to the atmosphere.

4 CORROSION I N AN OIL, REFINERY

Water

Water is the third principal source of corrosion in a refinery which, in this respect, is probably worse off than many other kinds of plants. A refinery needs large quantities of water for making steam, condensing oil vapor, cooling oil, flushing, testing, sewage and fire purposes. The amount, "depending considerably on temperature, may be as high as 1000 gal. per bbl. of crude oil. The disadvantages are high temperature in summer, salt content, and natural contamination from the atmosphere and its gases and from manufacturing wastes.

The corrosion from this source is most active in the pumping equipment. In one plant observed it has recently become noticeable in the fire mains but other water lines have not as yet given trouble. Brass- tube gasoline condensers give some trouble. Sprayed cooling coils are of course severely attacked. There has been no unusual trouble due to water with steam boilers. The steam-condenser corrosion is about average. The refrigerating plants, using circulated brine, have given no trouble in 10 years' operation. The closed system which excludes air is used.

Cooling water from deep wells is found to be very corrosive, and this can hardly be attributed to the 1 per cent. salt content. On examination this water is found to be saturated with oxygen which is more likely to blame.

Atmosphere

Atmospheric corrosion is no more severe than that found in any industrial district near the ocean. The source being high humidity result- ing in abnormal precipitation of moisture, and the contamination of air by combustion products, sulfur, carbon dioxide and HC1 gases, which form active acids with the water.

Corrosion of steel frames, especially at bdttoms of windows, in concrete buildings is serious as it causes warping which prevents the window from operating. This can easily be prevented when the building is constructed by coating the steel with the wax tailings compound mentioned later.

Both the still and the furnace suffer corrosion from combustion. Iron oxidizes rapidly a t high temperatures which it sometimes reaches on the still bottom due to heat insulating deposits inside of the still. Refinery fuel is not always ideal, sometimes it is high in sulfur and sometimes it contains caustic. There has been no noticeable effect on the furnace due to sulfur but the alkalies and caustic act strongly on the brick work where high temperatures are maintained, being most noticeable in thesteam- boiler furnace. Cements containing chromium seem to withstand this action and are being- tried as a remedy.

, . H. F. PERKINS'

Corrosion in Tankers

Tanks carrying light oils are said to corrdde niost rapidly Although light oils, particularly gasoline, have a very low sulfur content and it is doubtful that this sulfur plays any part in corrosion, the information so far obtained gives some indication tha t the light-sulphur compounds may be responsible for the excess corrosion. A bad feature of oil-tank corro- sion is the1 absorption of oil by the scale formed, this oil is difficult t o remove unless the scale is removed also, and may have been the cause of tank explosions after they were steamed out and men started t o work . in them.

Tankers of course, are also subject to salt-water corrosion, when'water is used for ballast, but this effect does not yet appear serious.

The most acute phases in oil refining are: 1, Where the steam used in distilling acid-treated oils is condensed; 2, where water vapor is condensed when distilling certain high-sulfur crudes; 3, where excessive hydrochloric -

acid is formed; 4, where there are large amounts of sulfur and HCl gases with water vapor condensing; 5, in pressure-still equipment; 6, where cooling water is sprayed over pipes and tank roofs; and 7, in cast-iron fire lines.

MEANS OF COMBATING CORROSION

There are three ways of combating corrosion; the corrosive elements may be neutralized, a type of metal or material having high resistance to any special form of corrosion may be used, or a protective coating, includ- - ing inhibitors which form protective coatings, may be used. Each of these methods is useful in its particular sphere, but the problem being one of economy, the principal difficulty after a careful study has been made in a particular case, is to determine what means to use to reduce the cost. Cost, of course, includes all cost of replacement and interruption to operations, not just the cost of material.

Neutralization

Neutralization. of acids by ammonia has been suggested and is being used to some extent in various plants. Caustic soda, where it is a plant by-product, can be used instead of ammonia and is less expensive. his method can be used in t h first four cases, but in the firstcase i t has not yet'.been.determined what would be most economical as the amount of caustic required. is large. Tbe!use of caustic is economical in the other cases though its real advantages have not yet been fully determined.

There are othei cases. of -extremely active vapors from chemically treated oils similar to the first case mentioned above. Where it is possi-

/

ble to scrub these vapors before they condense, an arrangement for this purpose has been used. The vapors are first passed through a cast-iron box, where they are thoroughly scrubbed with sufficient water to absorb all the corrosive gases, then the water is led off through a barometrically sealed tower or large trap lined with acid-proof brick. The condensed oil and uncondensed vapors are drawn off from the surface of water in the tower and pass through the regular condenser without corrosive action. As the water carries off considerable oil i t is necessary to con- struct an efficient skimming trap to recover the oil.

Metals That Resist Corrosion

Metals, such as aluminum alloys and chrome iron, may be used for pressure-still towers though some prefer to use heavy iron. The princi- pal corrosive here probably is HzS, and the corrosion takes place when the'iron is above 600" F., increasing rapidly with temperature and depending on the amount of HpS present.

Practically all refinery equipment has been built of cast-iron, wrought- iron or steel in the past; i t is the lowest-priced metal and readily worked, therefore the progress of fighting corrosion will be based on this standard. Any metal giving greater resistance t o corrosion will cost more and usually many times more than iron. There are places such as still bottoms, espe- cially in coke stills, and tubes for pressure stills where the increased serv- iceability may more than warrant the increased cost of material.

The use of special metal of relatively high cost for handling sulfuric acid has been accepted, and special metals or alloys for other places will be adopted when their value becomes a necessity.

Protectice Coatings

Paints are not satisfactory protective coatings, but a compound made of high-melting-point wax tailings with chromium added is very promising. This compound has three disadvantages, however: As it always remains greasy it cannot be applied to parts that will be touched or walked on, as it melts a t around 130" F. it cannot be used on surfaces of higher temperature, and it is dissolved by oil. On the other hand, if applied in )..is-in. layer it will stand the electrical test, will last a long time under flowing water and is not readily attacked by acids or alkali, and not only adheres well to metal but penetrates to the small cavities, intro- ducing chromium which is an excellent rust inhibitor. This compound is being used extensively to coat underground pipe and is being tried on water-sprayed cooling coils and tank roofs.

A protective-coating method, which may often prove advantageous where high-priced metals are prohibited, is the applying of a metal coat- ing by spraying on the melted meta1,desired. A coating of any thickness

H. F. PERKINS 7

may be applied but usually 0.002 in. is sufficient. Four metals are being thus tried in a tanker carrying light oils, aluminum, copper, brass and zinc. The copper was badly corroded and the brass showed some attack I

on the first trip. No water was carried on this trip. Lead coating would probably resist the acid atmosphere better than

galvanizing which has a poor protective value. Aluminum paint has shown much promise as a protection against atmospheric corrosion. I

The best metal to use in any case is a matter of close study; the chance of being right by haphazard guess is very rare. Just because a metal does well in one place is not even an indication that it will do well in another place.

Monel metal is popular and it is wise to use it under natural water- corrosion conditions and in connection with steel or iron. It is electro- negative to iron and when used in comparatively small surfaces, such as bolts, will give consistent life to apparatus. Sweater-pans are another example. It would seem advisable to line the iron pans with thin sheet monel metal and spray aluminum or nickel on the outside.

After a resisting, acceptable metal is selected, good judgment must be used to see that there will be no chance for long range electro-chemical action between it and some other metal after installation. This is par- ticularly true where parts are under water as in condensers. The heads of steel bolts in water boxes were found to dissolve rapidly. On testing this head was found to be electro-positive to all other iron. Testing different grades of cast-iron, iron plate and steel as well as iron nuts and bolts, we find there is always a potential difference which is not always consistent. When a metal is positive in an electrolyte and an electric circuit exists the metal goes into solution. Monel metal bolts greatly outlast the life of the cast-iron pipe which is corroded from the inside by the acid conditions cited.

We are making a special study of fire-line failure, a brief outline of which will be made.

The fire lines began to give trouble from failures about 3 years ago. .

They carry a normal pressure of 80 lb. and have been tested frequently a t 120 lb. which is the fire pressure. These lines are class C thickness and, like all cast-iron pipe, do not change in thickness or appearance as they corrode, so that no matter how badly corroded they are there is no indication of it without close examination. (Fig. 1.)

The corrosion process seems to remove all the iron and about half of the carbon, leaving a structure made up of a relatively large amount of silica, considerable carbon and the balance iron oxide. This structure cuts as eadly as graphite and is porous. The water oozes through until

8 CORROSION IN AN OIL REFINERY

there is sufficient flow to wash out the structure. A hole of 8 to 10 sq. in. is the result. Although this is the more usuaI type of failure in some cases extensive cracks joined a number of corroded spots.

PIG. 1.-TEX-IXCH CAST-IRON FIRE M A I N IV SERVICIG 12 YR.. Partly corroded spots dirl not develop until found by hnmmering. Other small

dents show the firmness of the remaining surfacc.

FIG. 2.-VARIATION I N CORRODED DEPTH.

The inside surface is corrodcd evenly; the outside surfacc is not corroded a t A, but is corrodcd nearly through a t B, B. The upper edge of large hole shows the extcnt of a large corrodcd spot.

The inside of pipe is always corroded evenly one-eighth to one-quarter of the thickness, whereas the outside is corroded in spots and much of the

outside surface is not touchcd by corrosion. It is these outside spots that finally cxtend through the pipe wall. (Fig. 2.)

Cast-iron pipes are supposed to have been coated with coal-tar pitch varnish a t high temperature, and it is evident that the flowing water inside gets through the varnish and causes corrosion. On the outside, where the pipe is in contact with damp or wet soil, there seems to be no corrosion where the varnish was originally intact, but the outside coating is damaged on nearly all pipe before the pipe is laid and i t is a t these points that corrosion takes place unrestrictedly. Under these conditions 12 years seems to be the life of cast-iron pipe. We are coating new pipe outside and trying to use bctter watcr inside ant1 in thc next 15 or 20 years we expect to have some more data.

Examination of cast-iron pipe conveying cooling water a t 35 lb. pressure and in service 7 ycars, so far has shown only a trace of inside corrosion, and i t may be supposed that the pressure on the water might influence its penetration of t,hc varnish coating.

Considerable trouble has been espericncetl with brass tube condensers for steam stills. Whcre the gasoline is free from sulfur, corrosion takes

FIG. 3.-CORRODED ADMIRBLTY lfKT.111 TUBE FROM (;:\SOT.IXk: CONDENSEIZ.

place only on the out (water) side of the tubc and in the first pass which is the hottest. The excessive corrosion is due to the high temperature mainly but is assisted by deposition of mud on tubes and by conducting

10 CORROSION IN AN OIL REFINERY

water. Coating tubes with tin, completely insulating tubes, partly insulating from tube sheet and making complete contact are being tried. Tubes with high nickel content are also being tried.

Fig. 3 illustrates the condition where sulfur compounds work inside and water outside. This condenser was in service 955 months. The small hole on top corroded through from outside and the large hole corroded from inside, illustrate the problem. Condensing gasoline vapors and steam were inside of tube and cooling water outside.

The corrosion of steam condensers in one power plant is of an entirely different nature, but is similar to general experience with such appa- ratus. While the tube wall is reduced about 25 per cent. in thickness in 2 years, 95 per cent. of the failures are due t o splits and this is really a manufacturing defect. Some drawn-nickel tubes are about to be tried on one ship, but they will have to show about 5 times the life of Admiralty metal to pap for themselves.

It is pointed out that while the troublesome corrosive elements are present in only a very small percentage the total amounts are appreciable.

In .one day there is approximately 360 bbl. of salt water, producing 120 lb. HC1, 109 tons of sulfur, in crude oil to stills and 9 tons of sulfur delivered to the atmosphere by combustion. Many tons of sulfur, caustic and other refinery wastes are turned into sewers and find their way in more or less neutralized condition into cooling water.

The standard unit of corrosion depth per year should be adopted and used in all specific corrosion data. Grinding the surface to determine pitting would add greatly to the value of test data. ~ a b o r a t o r ~ testing of metals and materials under specific corrosion conditions, although not giving exact quantitative results as compared with the natural condition, will give a reliable indication on which a choice of material may be based. This is a profitable procedure in most any case of corrosion and will lead t o much saving.

A general guess can often be made as to the principal agents of corro- sion present but in specific cases of refinery corrosion such determination of the quantities,' concentration or other elements is impractical. Samples of the water vapor, including gases, may be taken as it condenses and used for laboratory tests, or small samples of material may be sus- pended in the condensing vapor for information on corrosion resistance.

Considering the variation and extent of refinery corrosion, it will be appreciated that a paper on this subject could not contain much detail or technical data, but i t may contain information tha t will be heartening

H. F. PERKINS 11

to some of its readers. The main thought for active consideration is the large and important part sulfur and its compounds play in raising refining expenses and retarding refinery operations.

I

I SUMMARY

Corrosion is an economic problem, its reduction must show good returns on the investment.

Sulfur is the chief refinery corrosive element; water, atmosphere, hydrochloric acid and caustic are the other important factors. Sulfur compounds form sulfur gases on heating, which combining with water give sulfur acid.

Hydrochloric acid gas coming from decomposition of magnesium chloride salt in crude; give's hydrochloric acid when combined with water. These acids may be neutralized by ammonia or caustic soda with econ- omy, when in small amounts. In large amounts it may be advisable to use acid-proof jet condensers or special costly metals.

Atmosphere is made excessively corrosive by combustion products, including sulfur, and by humidity. Protective coatings of metal, metallic paint or specially prepared wax tailings and chromium are advantageous.

Water contains organic matter, salt, atmospheric gases, weak sulfur acids, hydrochloric acid and caustic, and is high in temperature half the year. Special metals, protective coatings and neutralization by caustic may be employed.

Caustic in fuel oil causes rapid deterioration of furnace brick work. Fire cements containing chromium are under investigation as aprotection.

Sulfur compounds in light oils will corrode metals under certain conditions. Aluminum and tin are being tried as a protection.

Uncoated cast-iron, 36 in. thick, when kept wet will corrode through in about 12 years, so that it will not hold water a t 80 lb. pressure, though there is no change in thickness. There is the same effect on oil condenser cast-iron coils, which have a life of about 2 years where steam and acid gases are condensed.

Accuracy is not possible, and the problem is complicated because of changing operation, variation in crude oils and the wide variation in the composition of corrosive agents. Laboratory tests and standard methods of reporting are of great value.