THIS MONTH'S PUZZLER
We use a pyrophoric catalyst in a polymerization reactor. The ZrCl2 catalyst reacts violently with air and water. According to the catalyst manufacturer, the reactor, monomer feed tank and other tanks must be purged to 10 ppm-volume (ppmv) oxygen and 10 ppm-mass water. Monomer comes by railcar; the data sheet says the monomer meets a 10-ppmv oxygen standard; there's no measurable water present. The monomer goes through a molecular-sieve dryer, then to the feed tank after hydrotesting, air-drying and a nitrogen purge. However, we still can't get the monomer oxygen down to 100 ppmv in the feed tank. After 30 additional hours of purging, we can't get below 200 ppmv. The water in the tank is about 50 ppm. What can we do? What could be the source of additional oxygen and water?
A SIMPLE SWEEP WON'T SUFFICE
I have the following questions: 1) Did you guys check the O2 level yourselves or did you trust the certificate of acceptance (COA)? 2) Did you check after your mole-sieve beds to make sure they are not adding water? (They can add water if they are saturated.) 3) Are you sure the O2 meter is working correctly? 4) How are you purging?
At my plant, we assume that the N2 supply has essentially no O2. Our O2 meter is on N2 purge unless we're using it. We zero our meters with N2.
We have found that a simple N2 sweep is not very effective. What we do is pressure up the tank with N2, the air and N2 mix, then we bleed it down to flare. We cycle pressure up to about 15 psig several times, then we vent to flare.
Why don't you use the pyrophoric material to spend the air and water? After we clean our tanks we add a pyrophoric in a solvent until samples show that it is not being consumed. We never air-dry a tank.
M. Goodeaux, senior process engineer
[Company name withheld on request], Beaumont, Texas
GO WITH PDP
Instead of N2 purging, use a pressurization and depressurization (PDP) process. The PDP process is very effective compared to simple purging because N2 is introduced at multiple nozzles located at various elevations in the reactor. Depressurize from nozzles near the bottom portion of the reactor. If you still don't get the required dryness, use the reactor jacket to drive water into the vapor state. In some polymer reactor systems such as that for high density polyethylene a small quantity of dilute tri-ethyl aluminum (TEAL) in a solution of n-hexane is injected into the reactor to reduce moisture and the O2 level after a hydro test. This is followed by drying and the N2 PDP process.
Check the sampling and testing procedures. If the samples are analyzed on-site (or in the field) the instruments must be thoroughly purged: > 5 minutes. Sample bottles must be dry and purged before use.
C.C.S. Reddy, lead process design engineer
Singapore Refining Company, Singapore
THE NITROGEN FLOW MAY BE TOO LOW
Check all inlets and outlets to make sure the tank is isolated except for the N2 purge and its vent. The N2 could be fed with a properly sized regulator controlling the flow and the pressure in the tank could be held at an inch or two of water head depending on the tank properties.
Richard H. Smith, engineer
Texas Commission on Environmental Quality, Austin, Texas
PURGING DIFFERS FROM DRYING
Assuming a 5,000-gallon tank, a target of 8 ppmv O2 and a starting O2 content of 21%, it will take about 4.5 hours with an N2 flow of 1,500 ft3/hr (at 60°F and 14.7 psia).
The feed tank is probably not a pressure vessel or I would recommend a vacuum cycle purge: vacuum, purge, vacuum, purge, etc., until the cycle is complete. For the same tank, with a 100-torr vacuum, it would take about five cycles and 103 ft3 of N2 or 66 ft3 of CO2. Some sources recommend jet-mixing to improve purging.
The trouble with these textbook calculations is that they assume 100% penetration into every corner of tanks and process. It's only 20-20 hindsight but, next time, perhaps a pneumatic pressure test would have been a better choice. Maybe inefficiency is the issue. Allow a few more days. What about drying?
Water can come from two sources: 1) equipment, and 2) the atmosphere. Assuming the mole-sieve dryer is functional and the piping was well drained after hydrotesting, it seems unlikely that water came from there. However, some moisture will be trapped in low spots and gaskets and other soft materials.
Moist air is a strong possibility. At 75°F and 14.7 psia, air at 100% relative humidity has a moisture content of about 18,500 ppm-m. The slightest contamination while taking a sample could lead to hours of fruitless running around on a rainy day.
Purging and drying are different processes. A steady bleed into a tank might work well for O2 purging but be useless for drying. Drying requires that the drains be located in spots where water will flow by gravity. Driving water from a tank should be at a vigorous flow with heating, if possible. I would do the drying first with a high flow and heat and then do the purging. This idea is supported by the pool evaporation equation: http://www.engineeringtoolbox.com/evaporation-water-surface-d_690.html. Maybe the mole-sieve column could be used at the last stage of drying to remove trace water.
Another option might be to purge initially with a heavy gas. For this to be effective, the gas should flow top to bottom allowing gravity to work against the O2. A light purge gas should flow upward from the bottom. Based on price for 99.999% purity (N2 at about $140/unit, CO2 at $157/unit, and argon at $184/unit), C O2 would be the most cost effective choice. A heavy purge gas might be less likely to mingle with the gas in the tank. Purging with heavy gases is quite common in laboratories doing sensitive work.
Dirk Willard, consultant
TRY COPPER CATALYTIC OXIDATION
Use copper fixed on a support, e.g., aluminum oxides, in the form of a bed of granules, with the bed having a high length/diameter ratio. Thereafter, the mole.sieve adsoption, preferable under pressure, also relatively long, should be able to remove water to the necessary concentration.
Walter Schicketanz, Dr.-Ing., Consultant
Readings from two radar level transmitters on our LPG storage tank markedly disagree. The radar elements are installed in stillwells connected to a large pressurized tank currently operating at about 800 kPa (g) and 24°C. The gas feeds burners via underground piping. Right now, demand is so low there's no need to use the vaporizer, evaporation suffices. What could cause these elements to differ so radically? Should we be concerned? Is there anything that can be done to line up the level transmitters?
Send us your comments, suggestions or solutions for this question by September 10, 2010. We'll include as many of them as possible in the October 2010 issue and all on CP.com. Send visuals — a sketch is fine. E-mail us at ProcessPuzzler@putman.net or mail to Process Puzzler, Chemical Processing, 555 W. Pierce Road, Suite 301, Itasca, IL 60143. Fax: (630) 467-1120. Please include your name, title, location and company affiliation in the response.
And, of course, if you have a process problem you'd like to pose to our readers, send it along and we'll be pleased to consider it for publication.