This Month’s Puzzler
A batch process that consists of a series of feed tanks for a fluidized rotary dryer relies on pH probes to control product quality. There are two steps: pH adjustment, then chemical addition. The pH is adjusted from 1.0 to about 6.5. If the pH is too low, the acid bound to the solid will corrode the dryer and the customer’s equipment. Also, low or high pH will cause a problem with another ingredient added to the customer’s product. Unfortunately, we can’t depend on the probes. They suffer from slow response, especially with negative drops, for example, from 1.0 to 1.1. Also, because of poor residence time in the process, some chemicals added during the pH adjustment have caused problems such as silicon gelling of the probes (in acid solution). The quality assurance department is convinced that we should be able to meet our customer’s needs if we can achieve an accuracy of 0.25 pH, i.e., both probes agreeing that closely. The pH probes last about a week with some product runs and only a few hours with others. Can you suggest improvements in this process or how it’s controlled?
FOCUS ON MIXING
Meter NaOH in by using powder feeders sized accordingly. Add bigger agitators and possibly larger motors to incorporate the powders completely and reduce reaction time.
Errol Williams, production engineer
AVEKA Inc., Woodbury, Minn.
USE ZETA POTENTIAL
I suggest using zeta potential measurements in place of pH for a few reasons. In measuring pH, what is really being measured is a mV signal, which is then converted to a pH value. It is this mV value which is important. This can be better measured via a zeta potential measurement. Zeta potential will vary as pH varies.
Once the shape of this relationship is known for a given material, then the process can be better controlled. (This will require development in a laboratory to define the relationship between zeta potential and pH.) For aqueous-based systems the rule of thumb is that a system will be kinetically stable so long as it is outside the electro-neutral region of ±30mV. If a system is sensitive to pH variation, then it is best controlled using zeta potential and adjusting the solution to a specific mV range.
Michael Sakillaris, technology fellow
Rohm & Haas, Marlborough, Mass.
I suggest you look at increasing the solution temperature or circulating the product in the pH control vessel to increase the residence time. Also look at the possibility of using a density measuring instrument to control you pH.
Lukas Koster, engineer
Ardeer Engineering, Modderfontein, South Africa
pH EQUALS SLOW RESPONSE
First you have to accept the fact that pH and slow response go hand-in-hand
with each other. With pH control you have to think outside the “normal
process control strategy” box. Here are some ideas: 1) when the process is too tough on the pH probes, don’t use them — measure something else instead; 2) don’t put the probes into the process fluid except when you need to measure the pH; 3) different pH process solutions require different pH probes; and 4) change the way the acidic slurry is neutralized.
An instrument is not necessary for neutralization. Try dead reckoning. For example, measure the amount acidic slurry added to tank one and then calculate the amount of dry NaOH needed for that amount based on a titration curve. Don’t put the pH probe into the acidic slurry tank in the first place. Use the tank level to determine the weight of slurry added to the tank. Titrate the slurry to only 5.5 pH in the first tank. Then the other two tanks only need to go from a pH of 5.5 to 6.5. It is best to use the titration calculation on those tanks as well. Then you have the remaining two tanks to adjust the final pH.
We have found a lot of success with NaOH-addition-based titrations in waste treatment applications but lots of problems with on-line dynamic pH adjustment control. You will need separate titration curves and calculations for each product neutralized.
With the titration curves you may be able to get away with only using two tanks. This way you can get longer residence time for the NaOH solids mixing by alternating using tanks one and two for the acidic neutralization. Then tank three is used for the final adjustment from tanks one and two.
Continual exposure to a harsh environment is often unnecessary and expensive. Consider using a retractable housing for the pH probes. This way the probes are only in contact with the process fluid during the time when you need to measure the pH. When retracted then the probe tips should be flushed with water to get the corrosive material off. While the solids are mixing and the pH is changing, don’t measure the pH. Wait until the acid is neutralized and then insert the probe to check the final result.
Don’t use the same probe type for the acidic solution tank as for the neutral. Work with your supplier to determine what is best for each tank application.
Go back to the drawing board! Change the process. Consider only filling the first tank one-third; then add all the NaOH normally needed for a full tank of pH 5.5. This will over-neutralize the slurry to an alkaline mixture. But then top off the rest of the tank with the acidic slurry until the desired pH is achieved. This provides two benefits. First, you get a longer mixing time for the solid NaOH and, second, it can be easier to control the pH with the slurry addition rather than through the solids addition. This approach should be tested in the lab first and make sure the materials of construction are appropriate.
Also, consider using NaOH solution for the second and third tanks to improve the adjustment response time.
James Loar, applications engineer
Ciba, Basel, Switzerland
LOOK AT OTHER VENDORS
There are several pH probes that are designed for working in a harsh environment. The only caveat is that they are only accurate at one pH calibration point. If all you need to do is control at pH 6.5 and not worry too much about the absolute accuracy at lower and higher readings, it should work.
John Pocreva, engineer
Numeric-data-solutions, Lagrangeville, N.Y.
INSTALL A DRAFT TUBE
As neutralization reactions depend mainly on agitation, modify the reactor (see Figure 1: http://www.chemicalprocessing.com/Media/MediaManager/Figure_1_Draft_Tube_Model.pdf) and procedure as follows:
1) prepare an NaOH solution (30%) and add it instead of solid, as liquids are easier to mix;
2) keep the longitudinal baffles as per original design;
3) install a feed well to direct NaOH solution flow to the slurry;
4) construct a draft tube to direct flow to avoid short-circuiting and maximize mixing efficiency — the draft tube eliminates the effect of the longitudinal baffles as flow streams travel along the tank walls instead of towards them;
5) relocate the NaOH solution feeding point to be at the middle of the new draft tube;
6) install a riser to avoid short-circuiting and to keep mixture in the reactor longer;
7) make a new opening and install an overflow line at the top side of the tanks — this would maximize residence time, mixing time and minimize short-circuiting;
8) flip (or change the type of) agitator blades — this is necessary because of the draft tube; and
9) relocate pH meter away from the dead zone (of poor mixing), e.g., install it at the overflow.
Zeid Matouk, senior process engineer
Jordan Abyad Fertilisers and Chemicals Co, Amman, Jordan
Try modeling the chemical reaction; measure (weigh) accurately quantities of material going in and start with an accurate pH measurement of the slurry at the beginning of the process. This should be followed by occasional measured confirmation of the pH at intervals to compare with predicted pH, as the calculated amounts of -OH is added to the slurry. This way probes will last longer, especially if they are retrieved and rinsed during the non-measuring periods.
Steven Vidakovic, associate research fellow
Pfizer Inc., New London, Conn.
TRY VOTING LOGIC
If you are certain that the pH is 1.5 or less, maybe it would be best to keep the probe out of the low pH environment until partial neutralization has occurred. Often it is possible to control the pH by knowledge of the recipe, weight or operating profile. Add base with mass meters and a high performance density or conductivity measurement to verify the strength of base. This will allow tracking the amount of base, without exposing the probes to the less acidic environment.
Another quality assurance problem is precision of pH probes. With two probes it’s hard to know which is correct. A better practice is to have three probes and always take the middle value to be correct. A failed probe should alarm but probe failure, high or low, will allow the process to respond to a probe near the correct value.
A description of the mixture titration curve would have helped. If the pH
target is in a zone with a steep rate of change, dilute the concentration of base by one to three orders of magnitude to complete the neutralization. It will increase water volume but, if the material balance is known, using two different strength acids will improve the repeatability and avoid overshooting the endpoint.
If it is difficult to control the sequence of chemical addition, then automation seems warranted. It may be that your operators need training. Establishing the operating intent and methods of achieving product goals could improve quality.
Gary Holleran, senior engineer
BASF Corp., Beaumont, Texas
IMPROVE THE DESIGN
I believe the process design needs a review. Addition of solid caustic in three different tanks suggests too many variables and process will be out of control as is being experienced.
My solution assumes that chemicals are miscible liquids and will be added only after the desired pH has been achieved.
Using the flow diagram and not having much information of the liquid concentration of the slurry, I would make the following changes to have better process control:
1. Use liquid 50% caustic instead of solid caustic. If solid caustic is not completely dissolved due to uneven mixing, variable pH readings will result; this is due to incomplete and delayed dissolution of solid caustic.
2. Use only one tank with a pump-around system and sufficient residence time for proper mixing; this should give a stable reading. pH probe would be in-line and controlling flow of liquid caustic into the tank. pH probe will have to be strategically located.
3. Check the agitator for proper mixing and design.
4. If addition of caustic as liquid overloads the dryer, re-evaluate the dryer design. Also check the liquid mass balance of the slurry against the dryer design. Necessary changes might have to be incorporated.
Girish Malhotra, president
EPCOT International, Pepper Pike, Ohio
SPLIT THE FLOW
I suggest reconfiguring the tanks to operate in parallel instead of in series. Coordinate your operators so that when one tank is charging material, which has met the pH requirement, the other two feed tanks can be prepared for their batches. I see several advantages in this approach: 1) the material volume in each tank is certain, e.g., the quantity of solid NaOH charged to the tank can be calculated to reach the pH value quickly; 2) there is enough residence time for the probe to attain a reliable pH value; and 3) dividing the dryer demand between three tanks offers greater flexibility for probe maintenance.
Bin Yu, senior process engineer
Rohm and Haas (China) Holding Company Ltd., Shanghai, China
RELY ON CONDUCTIVITY PROBES
I had similar problems of slow pH-probe response with one of our chemical scrubber systems. We monitored pH to identify when the scrubber media was “spent” and required replacement with new media. The pH was erratic, requiring continual recalibration of the transmitters and cleaning and replacement of the probes.
We replaced the PH probes with conductivity probes. By correlating conductivity with pH, we were able to make our process considerably more stable and reliable. These probes have been operating for more than three years without replacement or any major issues.
Darwin De Los Santos, plant engineer
IMTT-Bayonne, Bayonne, N.J.
CHANGE THE PROCESS
It seems to me that a change of a pH 1.0 to about 6.5 is quite wild. Assuming this is SiO2-to-SiO4 neutralization, if the remaining SiO2 is highly acidic, it might explain probe fouling. It seems that the stuff is clumping up and, as has been suggested, the reaction may be slow. Have you considered homogenization, dewatering or filter press? If water addition is not an option (because of dryer load or economics), have you considered depolymerization/polymerization?
Michael Waugh, consultant
We use a carbon-steel spiral-tube heat exchanger as a condenser in our spray tower. Operating conditions normally are 100 psig at about 150°F. Treated chilled water at 45°F runs in the tube side. The shell side handles metal chlorides, e.g., FeCl2, VCl4 and TiCl4. These chlorides condense to form liquids and evaporate to form gases. Severe corrosion from chlorine and erosion was the reason why the spiral exchanger was chosen over a shell-and-tube one. The chlorine is usually dry but, occasionally, spikes of water upstream of the exchangers cause corrosion. One of our engineers is pushing the idea of cladding Type 304L stainless steel with a high-nickel alloy. Coupon tests suggest this will improve service life. One concern is how bending the tube will affect cladding. How can we assure our nervous operations manager that this cladding will work? Do you have any other ideas? Keep in mind replacing the heat exchangers with another style will require a major change in piping around the units.
Send us your comments, suggestions or solutions for this question by December 12, 2008. We’ll include as many of them as possible in the January 2009 issue and all on CP.com. Send visuals — a sketch is fine. E-mail us at [email protected] 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.