Training / Design & Simulation

Process Changes: Mitigate A Manager’s Mayhem

Readers recommend how to restore performance after ill-advised process changes.

This Month's Puzzler:

We manufacture 97%-pure ethylnylcyclopropane (ECP, also known as cyclopropyl acetate — CAS No. 6746-94-7). The batch ECP process involves these primary steps: reaction with an alkyl lithium compound; blowing in NH4Cl with high purity N2 to form the ECP; pumping the product solution through a bank of filters; quenching the alkyl lithium with cold heptane and 2-propanol; washing with chilled deionized (DI) water, followed by filtering; decanting in a bed filled with random packing to remove the salts, separating the organic and aqueous phases; and, finally, distilling the organic phase to separate the ECP (distillate) from the 2-propanol. The wastewater, which still contains a trace organic phase, undergoes batch vacuum distillation to remove the trace. The wastewater then passes through our new trickle-bed air bio-filter before discharge to the city sewer water plant.

Unfortunately, the new manager made some changes to reduce costs: 1) using air instead of N2; 2) replacing deionized water with plant (well) water; 3) bypassing the chilled water cooler for the quench; and 4) cutting back the 2-propanol to almost zero. He was following the advice of an experienced operator. The first campaign didn’t go well.

We had organic carry-over from the biofilter into the city sewer; the stream also contains n-hexyl lithium. In addition, our product quality suffered; we only achieved about 88% purity after distillation. What did we do wrong? And can we get rid of the 2-propanol without substituting something worse?

Adopt A Different Process

The synthetic method implied by the use of alkyl lithium is dimetallation of 5-chloro-1-pentyne, followed by cyclization:

Step 1: Cl-CH2-CH2-CH2-C=C-H + R-Li → Cl-CH2-CH2-CH2-C=C-Li + RH, where R is the alkyl group.

Step 2 is a second deprotonation: Cl-CH2-CH2-CH2-C=C-Li + RLi → Cl-CH2-CH2-CHLi-C=C-Li + RH

Step 3: Heating completes the reactions in Steps 1 and 2, and cyclization forms the cyclopropane moiety by eliminating LiCl:

Cl-CH2-CH2-CHLi-C=C-Li → LiCl + Li

Step 4: Quench the acetylenic anion with NH4Cl. Blowing in NH4Cl as powder with nitrogen gas seems odd and undesirable because DI water or well water was added later. This step will be quite exothermic, so slow addition of ammonium chloride dissolved in water would be quite satisfactory. Nitrogen or air should not be used because either of these would tend to carry away the fairly volatile cyclopropyl acetylene (CPA) product. The problem description notes n-hexyl lithium was “carried over,” even though the reaction was quenched with NH4Cl, isopropanol and water. (Certainly any hexyl lithium would have been totally hydrolyzed to n-hexane and lithium product, but thanks for the hint.)

Selecting appropriate raw materials can simplify the purification of CPA. A better lithium alkyl than n-hexyl would be n-butyl because butane formed by protonation could be flashed off. And better than n-butyl lithium in hexanes is n-butyl lithium in cyclohexane. This allows the CPA to be distilled more cleanly from the higher boiling solvent — no hexanes present. CPA boils at 52°C, butane at -0.5°C, n-hexane at 69°C and cyclohexane at 81°C.

This batch preparation yields an aqueous solution of lithium chloride and ammonia under an organic layer. If the NH4Cl was added as a 25% solution in water (nearly saturated), the amount of water added thereby will be enough to dissolve all the lithium chloride produced (it’s about twice as soluble). The requirement for washing, filtering and decanting depends on the side reactions that occur but typically a simple filtration will clear up the aqueous layer enough to flash off any CPA and cyclohexane present without presenting a foaming problem. (The high ionic content makes it unlikely that a significant amount of organics will be dissolved in the water, so eliminating the flash-off step might reduce costs.) The organic layer will be a bit more complicated: estimated composition will be 5–8% CPA and a little butane, but mostly cyclohexane. Whatever yield of CPA is lost to side reactions will probably generate higher-molecular-weight tars but these probably will be soluble in the cyclohexane.

Disposal of the organic residue after distillation will be fairly inexpensive as it will be useful as fuel (high BTU content). The water layer may be a challenge and an opportunity: As a fairly concentrated lithium chloride solution, it might have salvage value. Its high concentration of salts may make its purification by biological approaches less satisfactory. Separating the output into just four products: 1) CPA, 2) aqueous layer, 3) organic layer and 4) a little precipitate (from the aqueous filtration) minimizes waste generation and handling.

I would tell the new manager:
1. Use butyl lithium in cyclohexane, not n-hexyl lithium in hexane.
2. For the quench, dissolve NH4Cl in water, about 25%, and drip it in slowly with cooling; don’t use nitrogen or air.
3. DI water or potable well water is suitable for dissolving NH4Cl.
4. Use cooling, especially during the quench — a chiller or ice could work. Cooling the reactants at the beginning of the reaction will reduce the side reactions.
5. Isopropanol is neither necessary nor desirable for quenching, as it tends to unite the aqueous and organic layers. Plain water would quench the reaction but lithium hydroxide is not very soluble; so the addition of NH4Cl provides acidic chloride — but then NH3 is a product and the aqueous layer will be alkaline and smell of ammonia.

Jim Gaidis, technical director

Arcal Chemicals, Inc., Woodbine, Md.

Go Back To Back To Original Design

I am amazed that the new manager changed or was allowed to change the process based on the advice of an “experienced” operator and he did not consult the company’s technical personnel about the changes. Why didn’t anyone ask if the operator was a qualified chemist or chemical engineer? Changes made without a science and technology basis should have been a no-no, at least in my book.
If the company wants to get back to producing the desired product, then I suggest the following:

1. Go back to the original process as designed and get the plant producing quality product as was being done before the “new manager” made the process changes.

2. Get the process technologists involved to review the operator’s suggestions and understand the basis for the modifications.

3. Rationalize the changes suggested by the operator. Cost reduction changes should be based on science, chemistry and chemical engineering rather than on the operator’s experience.

4. Technical staff should test out the operator’s suggested changes in a pilot plant or even in the plant (if there is no pilot plant) under very strict supervision. This will let the company understand how the changes will work out in an operating plant.

5. The operator’s suggestion indicates there are opportunities to lower the cost of manufacture. Time and effort are needed to pursue all avenues to achieve necessary cost reductions.

6. The company needs to review its operating practices and should not allow changes to be made in the process without complete understanding of the changes and their impact.

Girish Malhotra, PE, president

EPCOT International, Pepper Pike, Ohio

Change One Variable At A Time

This situation is what management of change was meant to avoid. In addition, it seems like bad experimentation to change several variables at once.

N2 is meant to protect the alkyl lithium from oxidation. Using air, especially plant air with its relatively high concentration of water — typically, 0.02 lb of water per lb of dry air at 20°F dew point — means that the catalyst and the 1-butyne, 3-chloro are exposed not only to oxygen but also to water. Depending on the availability of metal catalysts, this could pose more than a contamination or yield problem — it could be a safety problem.

DI water is slightly acidic, pH 4.5–5. Hard water, especially well water, typically is slightly basic. The EPA recommends water should have a pH of between 6.5 and 8.5; rainwater has a pH of about 5.3. Acid helps dissolve salts but also catalyzes the conversion of alkynes to a ketone, a side reaction, instead of allowing them to be carried away to distillation. So, there would be slight benefit to raising the pH. However, DI water has consistent properties, well water does not. Water, by itself, is a necessary part of the quenching process: R-Li+ reacts with water to form an alkyl compound and LiOH.

The 2-isopropanol is also part of the quench solution. Without it, the alkyl lithium could become a hazard. In addition, trying to do distillation without the liquid in the trays will be difficult. I wonder if this is causing the impurity problem?

As for chilling, this probably is required to improve the density difference between the organic and aqueous phases. Without it, some organics, especially polar ones, will be drawn to the water. Take samples near the phase break to determine the driving force between the two phases. I think you’ll find a problem.

The trickle bed is a fast, highly efficient means of removing organic compounds in trace amounts from water. The residence time reported is 20–90 seconds. However, these beds typically are designed for a certain maximum organic loading that you exceeded by “dumping” organics during quenching and decanting. You should look at the economics again; consider increasing the capacity of your wastewater system if your changes warrant it.

Dirk Willard, contract process engineer

Jedson Engineering, Cincinnati, Ohio

December’s Puzzler

We use N2 at 31 psig to drive 2,200 gallons of a viscous polymer at 235°F through an electrically heat traced and insulated 2-in., sch-10, 347-ft line from a reactor, which, once emptied, is ready for the next batch (Figure 1). This takes about 210 minutes. Initially, we can put about 25 gpm through the pipe. By the time we’re done, the polymer viscosity has doubled from 380 cP and the flow has dropped to 13.8 gpm. The polymer specific gravity is about 1.05 at 235°F. The reactor is rated for 40 psig. Originally, we used lobe pumps; we switched to the pressurized system because of monomer emissions, contamination of pump barrier fluid, a heel in the reactor, and shearing by the pump. Even with N2, the product continues to cross-link, resulting in the increased viscosity and the slowed-down transfer. Measurements show that the apparent viscosity increases to 1,200 cP for a 3-in. line and 3,600 cP for a 4-in. line. How can we speed up the process? The production manager wants the transfer only to take 30 minutes but believes that boosting the polymer flow rate will substantially increase the N2 demand. What do you think?

Send us your comments, suggestions or solutions for this question by November 13, 2015. We’ll include as many of them as possible in the December 2016 issue and all on Send visuals — a sketch is fine. E-mail us at or mail to Process Puzzler, Chemical Processing, 1501 E. Woodfield Rd., Suite 400N, Schaumburg, IL 60173. 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.

Free Subscriptions

Chemical Processing Digital Edition

Access the entire print issue on-line and be notified each month via e-mail when your new issue is ready for you. Subscribe Today. E-Newsletters

Stay ahead of the curve with free e-newsletters and alerts. They are delivered to your inbox on a regular basis and you may unsubscribe at any time. Subscribe Today.