8/6/2019 P45799

1/14

PILOT PLANT TESTING RESULTS INSUCCESSFUL LIQUID-LIQUID EXTRACTIONSCALE-UPAuthors:

Donald J. GlatzLori MasonKoch M odular Process Systems, LLC.

Copyright 2006Prepared for Presentation a tAIChE National Meeting

San Francisco, CANovember 2006

AIChE shall not be held responsible for statements or opinionscontained in papers or printed in its publications

KARR@and SCHEIBEL@are registered trademarks of Koch-Glitsch L.P.

8/6/2019 P45799

2/14

INTRODUCTIONPilot plant testing is a powerful tool that process engineers use to generate the datarequired for accurate design of liquid-liquid extraction (LLE ) column s. A well-designedand conducted test program followed by empirical scale-up to the production column isthe most accurate method used for the design of LLE columns2. This paper will coverthree (3 ) case studied where pilot plant testing was used for successful design andimplem entation of LLE c olumns. Each of these cases involved new processes for whichthe author knew of no previous operation for the same application. How ever, actual feedmaterial from operating plants was ava ilable. It is always desirable to use actual feedmaterial and not a synthetic mixture when performing pilot plant tests for scale-up. Thiswill insure that any potential problems such as emulsion form ation or mass-transferinhibition due to trace contamination will be exposed .Glass shell columns were used for testing each of the applications presented in this paper.For liquid-liquid extraction, the use of a glass c olumn during testing is highlyadvantageous during process development. Visual observation of the phase behavior(droplet formation, dispersed p hase hold-up, coalescence, interface behavior, etc.) allow sthe e ngineer or scientist conducting the test to quickly assess the colum n performanceand system atically tweak the process conditions to obtain optimal performance.

APPLICATION #1Extraction of Organic Impurities from an EM uent Sodium Hydro xide StreamIn a refinery operation, an effluent caustic solution was generated that wa s contam inatedwith som e unreacted monome rs and polymer precursors. The presence of these organicsin the caustic solution made it environmentally hazardous as a waste stream and also notuseful for plant neutralization or pH a djustments. In addition, the organic contaminantspolymerized upon heating causing fouling of equipment handling this stream. Thus,removal of the organic contaminants prior to heat treatment such as steam stripping wasessen tial in order to generate useful caustic so lution for reuse.The existing operating plant that generated the waste caustic solution also included apurification train for toluene. It was determined that if toluene could be used as theextraction solvent for this application, then the toluene effluent (extract phase) from theLLE step could be brought back into the toluene train for purification, without affectingthe final toluene quality.Laboratory shake tests w ere performed with the waste caustic stream and the planttoluene. The distribution coefficient for the most difficult organic impurity wasdetermined to be approximately 1.5 in the low concentration range tested. The Kremserequation was then used to calculate that greater than 99% removal of all the impuritiescan be achieved with 5-6 theoretical stages at a S/F ratio of 1 O.

KARR@and SCHEIB EL@ re registered trademarks of Koch-Glitsch L.P.

8/6/2019 P45799

3/14

With this as the basis, the next step in process developm ent was the selection of anextraction column and pilot plant testing to generate the data for scale-up. During theshake tests it was observed that this system had a high tendency to emulsify. Thus, theKARR@Colum n was selected as the optimal column type for this application4. The set-up for the pilot plant test program is shown in Figure 1. The set-up included a 1diameter KARR@Column with up to 10 plate stack height. The plates and spacers were316SS.Data generated in the pilot KARR@Colum n indicated the following:0 The colum n gave acceptable results for both aqueous and organic phase continuous.How ever, the tendency for emulsion was ob served to be less for the aqueous phasecontinuous.

Acceptable extraction efficiency was achieved at a solvent to feed ratio of 1.5 (as isweight basis) with 8 plate stack height.Capacities of 70 0 and 87 0 GPH /ft2 were tested succe ssfully. How ever, it wasdetermined that 700 GPH /ft2 would be used for scale-up.

Using the data generated in the 1 diameter, pilot KARR@Colum n; the standard scale-upprocedure2 was employed to design the production column for a waste caustic feed rate of4740 kg/hr. The resulting KARR@Colum n was 32 diameter x 32 plate stack height asshown in Figure 2. This column was fabricated and installed in 1999 and has beenoperating successfully since start-up. This has allowed the end use r to sell the effluentcaustic solution to a local plant to be used for neutralization purposes.

APPLICATION #2Arom atic Purification by Fractional Liquid-Liquid ExtractionIn fractional ex traction, a feed stream consisting of tw o or m ore solutes is introduced intothe m iddle of the extraction column and two imm iscible solvents are introduced neareither end. The heavy solvent will preferentially extract one of the solute(s) and the lightsolvent will preferentially extract the other solute(^)^. High recovery and purity can beachieved for both solute stream s by selecting the proper solvent system, solvent ratios,and number of stages above/below the feed location. A typical fractional extractioncolumn is shown in Figure 3Th e application here involved a light tar feed stream from a refinery operation thatincluded aromatic compo unds as well as many organic impurities. The goal was torecover a minimum of 90% of the aromatic compounds with less than 1YOmpurities.Preliminary laboratory work indicated that a mixed, polar solvent was very effective forremoving the aromatic compoun ds from the feed mixture. Howe ver, a significant amoun tof impurities would also co-extract with the aromatics. Since most of these im puritieswere non-polar, it was assumed that a straight-chain hydrocarbon that was no n-misciblewith the polar solvent would be effective for extracting these im purities.

KARR@and SCHE IBEL@ re registered trademarks of Koch-Glitsch L.P.

8/6/2019 P45799

4/14

Preliminary tests in a 3 diameter, pilot RDC colum n indicated that many extractionstages were required as w ell as a high ratios of the non-polar solvent to the polar solventand to the organic feed. Also, it was determined that the feed location mu st be near thetop of the colum n, with most o f the stages below the feed location (used for extracting theimpurities). Due to the large number o f theoretical stages required for this application,the RDC column was not a good choice. Instead, the SCH EIBEL@Column was selectedas the best design because this column design h as a very high stage efficiency4.The set-up for the pilot plant test program is show n in Figure 4. The set-up included a 3diameter SCHEIBEL@Colum n with up to 99 agitated stages. Wh ereas standard LLEtesting can generally be comp leted in 2 - 3 days, testing for fractional L LE can takesignificantly longer, which was the case for this application. During a 2-week testprogram , the following variables were evaluated in the pilot plant equipm ent:

The total number of stages was varied between 65 - 99 stages. The stages above theorganic feed location were varied between 12 - 34 stages, and the stages below thefeed location were varied between 42 - 65 stages.The overall column capacity (specific through put) was varied between 5 15 - 625Many variations of the ratios for polar solvent to feed and non -polar solvent to feedwere tested. The range for the polar solvent to feed was between 0.5: 1 and 1.4: 1 andthe range for the non-polar solvent to feed was 1.1 1 and 3 :1.The agitation speed range w as varied between 400 - 850 RPM.

0

G P H / ~ ~ ~ .

0

At the com pletion of the test program, it was determine d that the optimal operatingconditions were as follows:Capacity 600-62 5 GPH/ft2Total Num ber of Stages 65 agitated stagesOrganic Feed Location Stage 13 from the top of columnPolar Solvent : Feed Ratio 0.7 : 1.0Non-polar Solvent : Feed Ratio 1.8 : 1.0

For these con ditions the aromatic recovery was 9 2% with

8/6/2019 P45799

5/14

APPLICATION #3Neutralization / Washing P rocessIn a specialty chemical plant, a reaction process produced a nitrated organic productcontam inated with un-reacted nitric acid. Purifica tion required removal of the nitric acidand salts to very low levels (4 ppm). Based upon experience with similar systems, itwas understood that a single column could be used to neutralize the nitric acid withammonia solution and also wash out residual salts generated by this reaction with freshwater in a single column. Based upon solubility data, the extract stream (water + salts)would contain a significant amount of the nitrated organic product m aking it a hazardouswaste stream. Thus, a second extraction column was proposed using toluene to extractthe nitrated organic from the aqueous steam. After steam stripping the resulting aqueoussteam to remove the toluene, the w ater could be safely disposed through the plantseffluent treatment system. The proposed flow scheme for the two-column extractionprocess is shown in Figure 6.Due to the ha zardous nature of the nitrated organic feed, this material could not be easilyshipped to a pilot plant facility for testing. Thus , for this applica tion, the pilot plantequipment was shipped to and set-up in the operating plant. The set-up is shown on theattached Figure 7. Due to the fact that there was a high expe ctation for emulsionformation in the neutralizatio dwa sh colum n, a 1 diameter x 10 plate stack heightKARR@Column was used for the test program. Since the organic phase was thecontinuous phase, the plate stack consisted of PTFE plates and spacers to insure theproper wetting characteristics2. T he same colum n was used for testing the tolueneextraction column, however 3 16SS plates and spacers were used because the aqueou sphase was the continuous phase for this step.For the Neutralizatiod Wa sh Colum n, data generated by the test work indicated:0

0

The optimal design capacity was a specific throughput of 900 GPH/ft2.The volume ratio for the phases was:Organic Feed :Amm onia Water : Fresh Water = 1.0 : 0.087 : 0.26The full 10 plate stack height was required to give acceptable results. Also, it wasdetermined that the best location for the organic feed was in the middle, with 5 platestack height above and below.

Using the data generated in the 1 diameter, pilot KARR@Colum n; the standard scale-upprocedure w as em ployed to design the production column for a nitrated organic feed rateof 3200 GPH. The resulting KARR@Column was 30 diameter x 40 plate stack height.For the Toluene Extraction Column, data generated by the test work indicated:0 The optimal design capacity was a specific throughput of 110 0 GPH/ft2

The volume phase ratio was:Toluene :Aqueous Feed = 0.27 : 1.0

KARR@and SCHEIBEL@are registered trademarks of Koch-Glitsch L.P.

8/6/2019 P45799

6/14

0 A plate stack height of only 5 was required to achieve the desired remov al of thenitrated o rganic product.Using the da ta generated in the 1 diameter, pilot KARR@Colum n, the standard scale-upprocedure was employ ed to design the production colum n for a waste water feed rate of1250 GPH . The resulting KARR@Column w as 16 diameter x 15 plate stack height.Both KARR@Colum ns were fabricated and installed in 1999. They have been operatingsuccessfully since start up one year later.

CONCLUSIONPilot plant testing follow ed by emp irical scale-up is the most reliable and effectivemethod for the design of liquid-liquid extraction column s. This report presented threecase studies where data w as generated in a glass-shell, pilot extraction colum n and thisdata was used to design the required production size columns . In each case, smooth start-up a nd operation was achieved in the plant after installation of these columns.

KARR@and SCHEIBEL@are registered trademarks of Koch-Glitsch L.P.

8/6/2019 P45799

7/14

LITERATURE CITED1. Cusack , R., Pilot Plants Confirm Process Validity, Chemical Engineering , June1998.2. Cusa ck, R., Karr, A., A F resh Look at Liqu id-Liquid Extraction, Extractor Designand Sp ecification, Chem ical Engineering, April 199 1.3. Glatz, D., Parker, W., Extraction Enhancement, One Step at a Time, ChemicalEngineering, Novem ber 20044. Cusa ck, R., Frem eaux, P., A Fresh Look at Liquid-Liquid Extraction, Inside theExtractor, Chemical Engineering, March 1991.5. Karr, A., KARR@Reciprocating Plate Extraction Column, Bulletin KC-3, Chem-Pro Corporation, Novemb er 1997.

KARR@and SCHEIBEL@ re registered trademarks of Koch-Glitsch L.P.

8/6/2019 P45799

8/14

H O R

H O S

UG A L >-

- M E S H

A S T E C A U S T IC

1" IA - K A R R C O L U M N-16SS P L A T ES O N2" P A C I N GH O R -LttH O S

I::IIIz10 Fr-

rl55 G A L 55 G A LD R U M D R U MA Q U E O U S A Q U E O UR A F F I N A T E R A F F IN A

55 G A LD R U MF R E S H T O L U E N ES O L V E N TP R O C E S S F L O W D I A G R A MP IL O T P L A N T SET U PE X T R A C T I O N O F O R G A N I C IM P U R I T IE SF R O M W A S T E C A U ST IC

I F I G U R E - 1 IIK O C H M O D U L A R P R O C E S S I

8/6/2019 P45799

9/14

E X T R A C T(O R G A N I C O U T )I N T E R F A C E

\W A S T E C A U S TIC

8/6/2019 P45799

10/14

H E A V Y S O L V E N T-"1L I G H T S O L V E N T

(IN)

T

L IG H T P H A S E O U T(6 - R I C H )

I - H E A VY P H A S E O U T(A - R I C H )T Y P IC A L F R A C T I O N A LE X T R A C T IO N P R O C E S S

I F I G U R E - 3 IIKOCH M O D U L A R P R O C E S S I

8/6/2019 P45799

11/14

I I - A

r

L

I II IIt iII II IrT A G E 99

S T A G E 77

S T A G E 65

S T A G E 53

S T A G E 42

1

55 G A LD R U M5 G A L 55 G A LD R U M D R U MO R G A N I C N O N - P O L A R S O L V E N T E X T R A C TF E E D ( A L K A N E )P R O C E S S F L O W D I A G R A MP IL O T P L A N T S E T U PA R O M A T IC PU R I F CAT10NB Y F R A C T I O N A L E X T R A C T IO N

I F I G U R E - 4 IIKOCH M O D U L A R P R O C E SS I

55 G A LD R U MR A F F I N A

8/6/2019 P45799

12/14

L IG H T P H A S E(OUT)

I

P O L A R S O L V E N T( I N )

0 -P O L A R S O L V E N T( I N )

O R G A NF E E D("1C

O V E R A L LH E I G H T73' 9''

S C H E IB E L C O L U M N - P R O D U C T IO N U N I TA R O M A T IC S R E C O V ER YA N D P U R IF IC A T IO NF I G U R E - 5

K O C H M O D U L A R P R O C E S S

8/6/2019 P45799

13/14

N I T R A T E DO R G A N I C-F E E D

-M M O N I AW A T E R

F R E S HW A T E R

E X T R A C T # 1( W A T E R t S A L T+P R0DUC T)

I

T

E X T R A C T #+ P R O D U C T- T O L U E N E

T O L U E N EE X T R A C T O R

W A S T E W A T E R( W I T H S A LT S)

N E U T R A L I Z A T I O N+ W A S H C O L U M N

O R G A N I C P R O D U C T

P U R IFICAT10N P R O C E S SF O R N IT R A T E D O R G A N I CF I G U R E - 6

IKOCH M O D U L A R P R O C ES S I

8/6/2019 P45799

14/14

I .

tI I c2WI