Document 3QNdo0n3e3vvzvm89G4VdMoY6

S c h u l e r , Wil k e h s o n , Ha lv o eso n & Williams R O N D A L. E L L IS D A V ID A. F U G E T T S T E V E N W. H A L V O R S O N * DREW LOVELL JASON MARK R IC H A R D D. S C H U L E R * J A M E S D. W IL K E R S O N . J R .* L O U I S L. W I L L I A M S * * B O A R D C E R T IF IE D C IV IL T R IA L L A W Y E R P R O F E S S IO N A L A S S O C IA T IO N ATTO RNEYS AT LAW B A R R IS T E R S B U IL D IN G 1615 F O R U M P LA C E We s t Palm B e a c h , F l o r id a 33401-2382 TELEPHONE 4 0 7 -6 8 9 -8 1 8 0 FACSIMILE 4 0 7 - 6 8 4 - 9 6 8 3 ADM INISTRATOR: CAROLYN SHAFER INVESTIGATORS: P A U L A D. FAVO CCI G R A C E E. S M I T H HANK VAN W IERINGEN. JR. April 8, 1994 Gerson Smoger, Esq. Smoger & Associates 1333 N. California Blvd., Suite 540 Walnut Creek, CA 94596-4521 Re: Document Request Moyer v Dow STATEMENT FOR PHOTOCOPY CHARGES Charge from PIP Printing for copies= (See copy of invoices attached) $2,086.82** TOTAL DUE **This represents 1/2 of copying charges since we had two copies made. PLEASE MAKE YOUR CHECK PAYABLE TO SCHULER, WILKERSON, HALVORSON & WILLIAMS, P.A. THANK YOU PRI1 -NT^ ING 1649 FORUM "PLAGE \ Vi ^VALDVfANNCCE PAYMENT RECORD '\ \ <n c VI v nJhank You for your order. WEST PALMBEACH, FLORIDA 33401 fOljR INVOICE (407) 689-7274 REG. CUSTOMER NO OTHER WILL BE SENT. It will be ready: M T W Th F Date. n - ~ f// ^ Time Y ^ MC V V ^ F X_ v fMyio l x J ^ a Zip. Phone Ext. Attn:. COD. TUB FILE xt~ NEW CUSTOMER MAILER PHONE BOOK NEWSPAPER/MAGAZINE RADIO/T.V. DROP IN RECOMMENDATION PERSONAL CALLS P.O.# ____ .____ INDICATE ON YOUR CHECK WHICH INVOICE NUMBERS YOU ARE PAYING Thank You! PHONE WHEN READY FILE ORIGINALS i dPRESSPERSON. V. O Time ____________ Date _____________ CUSTOMER CONTACT: 1. Date ~v. V D Phone Mail D^WItt. imments: QUANTITY v--' ePR ESS UP F IN IS H 'p x ^r \ . & X =~ !P RT N i , r ~"'yr- Y v? 9 ~ > n * CUTS jf N** " .1'^ ^ f tv >. " Mi ? t Vi 1^:0 - . > . ------ CUT FINISHED SIZE COLI-ATE SHEETS # /SET STAPLE - POS. TOTAL# SHEETS # PASS % SIZE COLOR STOCK WT. TYPE -Ax^yw -ia* Uw </ O .jj L r \lv i. 1 1 1f' 5 - 3 Y5 y-s >, y.J, UNIT PRICE AMOUNT ... $i JLjA*<*4*1,V\ `n 3 0. ~Yl ( ixi XlAPkl C i* . FOLD TYPE IN, " DRILL #H O L E S POS. SHEETS ' ER PD. POS. TOTAL PADS & PPASTE-UP MISC. > 2. Date Spoke With: Comments: Phone Mail 1649 FORUM PLACE WEST PALM BEACH, FLORIDA 33401 (407) 689-7274 X V - /v - -.psv Pt . IA 'On GV IA A - t)\ i ; \v ' r ~ *+ & / - (? 3 -3 i t ____ L. ", .. ___ **- + $ . UQ: + * ' +* . +s 'B TOTAL AMOUNT DUE CASH .^-ANNOUNCEMENTS BUSINESS CARDS METAL PLATES OUTSIDE BINDERY RUBBER STAMPS SECRETARIAL TYPESETTING VELOX $ COPYING L3 u/tj m c= 136V/I <40 C k # ________ RetYd by -------------- Date Amt. Paid In Advance Notes: PRINTING $ >X Authorized By Receiiv\feedcTBByyTand Date V + BINDER^ + O/SERVICES + * COPYING 4 = SUB TOTAL |s $ $ CODE yo Date Ck# . Balance Paid Rec'd by _ / Cash PAY THIS AMOUNT INCLUDES ALL CHARGES 3 ss?BALANCE DUE ss' RESALE = NET SUB. + TAX $ I r$ 3 Lm ;i?|$ SHIP/ HAND oW? gsjTOTA BY CUSTOM ER NAME 65"INVOICE NUMBER KARL KRMEL MOYER V DOW O^ CCR -1- OC/78.-.95' D O W 9f!77''lr> INTEOC L'CTion TA3L OF CONTENTS Page 2 RESULTS AND DISCUSSION ------------------------------------- 9 (a) The Mechanise of TCX Formation ---------------- -- 10 (b) The Distribution of TCX in the P r o c e s s ----- -------- (c) Reduction of TCX by Chemical M e a n s -------- 26 30 (d) Reduction of TCX by Physical M e a n s --------------- -- 38 GENERAL CONCLUSIONS --------------------------------------- 50 SAFETY & E C O L O G Y ----------------------------------------- 50 EXPERIMENTAL----------------------------------------------- 51 ACKNOWLEDGEMENTS ----------------------------------------- 56 R E F E R E N C E S ------------------------------- Us /o f'tfrS it -f J- OiN./^-MOq CA1MJ. - 2 - OC/73-35 ^ INTRODUCTION A process for preparing high purity 2,4-D was started in May 1977 ar 943 Building. The high purity molcar. acid is transferred by pipe line from 948 Building to 489 Building where it is formulated into esters and water soluble amine salts. Starting in lata October 1977, precipitates were observed regularly in diluted amines formulations that were found to be made up of a number of impurities associated with the process but that were never detected during laboratory or pilot plant development work. In addition, these impurities were not found in the old 4S9 Building 2,4-D process. R, McLachlan^ analyzed the precipitates and found 1,3,6,8-tetrachloroxanthone (I: TCX) , 1,1'3,3'6,618,8'-actachloro-9,9'-spirobixanthene (II: OCSX), and marginally soluble salts of 2,4-D. Cl Cl (ID V A screening program in the plant showed TCX at levels of 200-500 ppm in the crude reaction mass, the recycle solution, and in the final product. Levels of 1000-2000 ppm were|found in the solvent and as much as 104 in the still tars. McLuchian^ analyzed a sample or summarized in Table 1. still tars and his results are Vi- CC/73-3S Tabl3 1: An Analysis of Cne Tar Samole S true tur a Approximate Psrrantace Comment DC? ' 2,4-D TCX (Z) OCSa (11) Dichlcrophenyldichlorovinyl ether Dichlcrophenyltrichiorovinyl ether 60% 25% 0.5% 0.01% 1.5% 1.1% 1.2% ~ 1 0.8% ) Main Component From a heat exchanger leak (Dowtherm J) raw material Product Rxn product Plus: at least eleven other minor components structurally similar to the above. See Ref. 1 for details. v i 7 % 9 -4 OC/78-8S A pro act was started to study the formation or TCX and OCSX in some de tar to learn mere about their formation and fate in the process O fcaciu a it is net obvious how they ara made. Several months after start r.g this s t u d y , another significant impurity was found in the ''n ^ erode c and, to a lesser extant, in the tars* It was identified -sj as C I: "3-5") 2,2',4,4',5',7,7',9-octachlorospiro-(benzofuro(3,2-b) -i benzo yran-11,9 *-xanthene) . 8 ~*r* n o now OCR Imccr -*r.t Side Reactions CiCH22 2 2:ia -4* MaOH -5- OC/78-' 35" H0CH2C02Ma + NaCl GIvcoiic Acid . ONa l i Cl + Cl2CHC02Na The reaction is performed in two stages by first reacting 2.2 moles of DC? with 1.2 moles of caustic (added as 50% NaOH) to form a solution of NaDCP in D C P . High purity 2,4-dichlorophenol (2,4-DCP) made by the chlorination of phenol with sulfuryl chloride in the presence of FeCl^/Diphenyl sulfide catalyst is used with chloroacetic acid (MCAA) made by the oxidation of vinylidine chloride to make 2,4-D as shown above* The typical raw materials analyses for each is shown in Table 2. Table 2: Tvoical Raw Material Analyses for the 2,4-D Process v DCP MCAA 2,4-DC? 2,6-DC? Other Chlorophenols 98,5% ^0.3% MJ.5% MCAA Dichloroacetic acid Chloromaleic acid h 2o 99. Q| 0.0.05% <1.0% .^ ^ -6- OC/7S-35 o Q\'LL$7. A\0 A rapid initial screening of the toxicity of purified TCX and tars gave the following results. Tar (contains 3-55 t + OCSX) Oral LD-0 (single dose) >5g/Kg Skin/Eye slight transient initiation. Chloracne response after 10 applications of neat tars. v Purified TCX >4g/Kg not irritating, j Chloracne response after 10 applications of V* solution in CHCl^. to O '1 AAO OC?. -7 OC/78-Q5 l I io !^ z\ - -13 "T^O^cISS bOty 2G774 `(O?. - 3- OC/78-as These results are encouraging from a plant hygiene standpoint, but mere cats is necessary regarding long term effects. When the im purities were first discovered in the plant* the still was used very occasionally The still was originally added to the process as a -av of purging contaminants out of the stream to maintain high quality recycle solvent. At this point, tar impurities made up 60-75* of the precipitates formed in diluted amines formulations. About February 1978, the capacity of the still was increased and the average levels of TCX levels dropped to '''SO ppm in the crude reaction mass and product and to 200-500 ppm in the recy cled From the period of March-May 1978, the ability to produce quality 2,4-D continued to improve and the amount of ' tar im purities found in the dilution test precipitates dropped . The balance were sparingly soluble metal salts of 2,4-D and seme new impurities called Complex 1 and Complex 2, two o f which are shown below (IV & V) , In Jul 1973, the still capacity was increased and gained at understanding the parameters affecting formulation quuli c by S. Siegel (OCP), S. Schell (Production 439), and J. King (Formu anion - 9001 3ldg) tnd their co-w ^ l !\bi s Iiuv yi.ir<iclv improved S\r'*' OC/78-3;. the c m aistency with which 948 Bldg. 2,4-D car. be formulated. Based u p m car-i-icn coefficients and data on the rate of formation of TCX to he hiscossed later in this record, K. E. First4 (Process Engineer ca.-.sr oed a computer program which predicted that by distilling of recycle the steady state level of TCX in the product would he "''10 p?m and in the recycled perc the TCX level would be M O f ?= * The pur-pose of this report is to present data on the formation and fate ci the impurities in the 2,4-D process. In addition, the results of experiments aimed at reducing or eliminating the impurities by chemical and physical methods will be presented. RESULTS AND DISCUSSION This strudy wa3 divided into the following sections and the results of ^ach are discussed separately. (A) The mechanism of TCX formation. (B) The distribution of TCX in the process. (C) Reduction of TCX by chemical means. (D) Reduction of TCX by physical means. Although there are many impurities formed in this process, it was decided to focus the research on TCX for several reasons: (1) The rate and mode of formation of TCX was the most predicable of the major multicyclic impurities; (2) it was a major impurity; (3) a pure sarcle was readily obtained (by D. Humbert, Anal. Lab) ; (4) the analysis is not complicated; and; (5) structurally, it is the siniplest of the rulticyclic impurities. i * 13 \ D ' 10 MOq jrb L L K g en -10 OC/73-85 (A) The Mechanism'of TCX Forma sicn A brief search of the literature including Chemical Abstracts shewed that: zhe xar.nitr.s ring system is formed frem a number of reactions.^ A ccrrr.cn syncretic reaction is shewn fcelcw in which the -X is included + 1. MeOH/MeONa _____ C*`. > 2. cone to s h e w the orientation or reaction. Xanthone is formed along with other products from the pyrolysis of o-chlorobensoic acid, salicylic acid, aspirin, o-phenoxybenzoic acid, and salts of the carboxylic acids. Most of these cycli2ations take place in the presence of acidic catalysts such as P25' Alc^3 > Acetic anhydride, or sulfuric acid. Xanthone is prepared in reasonable yield^ by the pyrolysis of nenol salicylate with or without a catalyst. The ease of formation of ring systems that are structurally similar to xanthones is most clearly shown by the formation of fluorescein and fluorescein dyes. Fluorescein7 is made by reaction of phthalic anhydride and resorcinol as shewn on the following page. v 14 Q- VL A -11- OC/7 3- 35' DOW 207745 g Phnol=hthalain is also made by this process using phenol instead of resorcinol. v 15 v1 -12- " C C / 7 3 - 35 T iQ J OH ^ C 0 2H OH 01'LLVtZ MOq V These background data give some important clues as to reaction mechanisms as will be whown later in the report. v 16 t# c c r OC/78-35' in c::er to understand the pathway of TCX formation# a series of ampoule tests were performed to determine which chemicals and/or ; combinations give rise to TCa . Table 3 summarizes the results c : study. A series of ampoules ( 25 ml capacity) were charged with < 3-5 c=s of mixtures in molar ratios as shown inTable 3 and were * ; heated in an oil bath to 72 hours at lSO^C. The sources of the chemicals were as follows: 2/4-D: High Purity Rhone-Progil Acid DC? : Doubly distilled 948 Bldg material Pare :^ NaOH Glyoxal Glyoxylic acid Reagent materials Tetrachloroethane (50:50 1,1,2,2- & 1,1,1,2-isomers) V i 17 0-H ocn -14- OC/73-35 v Table 3: The Formation of TCX from Synthetic Mixtures Heated at 160 for 72 hrs. 18 p.o 'O C R 1 OC/73-35 Several compounds other than those found in the process were tested. The terrachlcroethane mixture is a known contaminant in tne 941 Bldg, chlcroacetvlcniorida that is used to make MCAA. 2,5-Dicaiorosalicylic ^ acid is a proposed intermediate from the reactions suggested below: o ro C12C*CC1, U P 2. -> Metal -si. -a 4 -'he decomposition or perc to yield phosgene is known but the second reaction is only speculated. The glyoxylic acid is postulated to come from bis 2,4-D as shown belcw: which is a well known acetal hydrolysis. The glyoxal was include^ to shew the general nature of the reaction whose mechanism is suggested later in this report. 19 moo m uM OC/73-85 Several tencative conclusions c m be made from the data in Table 3. (a) TZ'.l car. be made by several routes aN'iCK ar.d/cr MaDC? is necessary for the formatier. of TCX. obis-2,4-0 appears to be a key intermediate. c?erc is a source of TCX. (b) Significant amounts of TCX are probably made only in the `reactor crude storage tank in the plant since strong base is required. Run 14 (Table 3) simulates the Na 2,4-0 storage tank (V-501: 948 Bldg) and Run 1 (Table 3) simulates the product storage tank (V-602: 948 Blcg) in which no TCX was observed. In order to study the formation of TCX as a function of reaction parameters, a series of 2,4-D reactions were performed in the lab oratory. These reactions were run to evaluate the effect on the amount and rate cf TCX formation. The data are summarized in Table 4. In these experiments, 1.1 mol of DC? was treated wth 0.6 mol of caustic and the mixture was heated to 130C. The majority of the water that was formed was boiled off. To this mixture, 0.5 mol of MaCH and 0.5 mol of MCAA were con-added during 1 hr at 130 and the crude product: was heated an additional 1 hr. Water was continuously distilled out during the con add step. At this point the desired number of ampoules were charged with 5-10 gms of crude reaction mixture and ware heated in oil baths at 130, 145, and/or 160" for 24, 4a, and 72 hours. i 20 <0A 1 , C -17- OC/78-55 Tibi 2 ^ ^ i 3 J*i >i ^ c : ; : ; A-'fact;r. the For-.ation o TC.' ir. the Ge --i - ily ro vi5 P : 3 c : : : f e r P r a s a y -.s 2, 4-D 21 CM4 -1 s - OC/73-35 The caoa in Table 4 show a number of interesting points. It; mus be e.".'.:s izad that: in view of the fact that these data are based o r. parts-per-millicn chemistry, the precision must be subject to soma error. The data are reproducible to about 20%. O o hC ct cn CO I 22 9 -19- OC/7 3- 33 mere facile route to TCX from pero whose effect is masked at the higher tamparaturas by a primary route (compare Table 4, Runs 1 & 2 and R-ins 3 * 4 ) . (C) If fee: of Atmosphere in the Reactor Nitrogen blanketing appears to lower the TCX level to *v2/3 that ob served in air. It is not certain that this result is significant and since the magnitude of the drop was low, further work was not warranted. Since C ^2 is known to be present^-0 in the vapor space throughout the 2,4-D process, a run was made to determine its effect on TCX formation. Carbon dionide comes into the process from the caustic and from pero decomposition. The reaction was run as described earlier with 20 mol% of NaHCOj added before the boil down step and then blanketing the reaction mass with C O 2 during the pest reaction heating step. Since this is a high ;enperature, nearly anhydrous process# the following reaction was thought possible which could eventually lead to TCX (See Table 3# Runs 15 and 15). ONa ONa Comparing Runs 1 and 10 (Table 4) show that NaHCO^ and/or COj do v not measurably affect the formation of TCX. I 23 -20 * OC/73-35 (C) Changing MaDC?/JaMCAA Ratio o o w y jy n r , C c .r .r ir ::.: Runs 1 and 7 (Table 4) shew than lowering the NaDCP/XaMCAA ra:::s signifinanely reduces the TCX levels as was also predicted from clue date in Table 3 . The effaco was studied further and is discussed later in this report. (E) ..-Source or DC? A run was made using the old lower purity DCP from the 349 Bldg process and assure that 943 Bldg, high purity, DCP was not a source of TCX. Pun 6 shews that TCX formation is not associated with the source of DCP. (F.) Effect of Ferric Ion Runs 11-14, Table 4, show that Fe+^ clearly increases the rata of TCX formation which suggest that alkylation step(s) are involved. D. Humbert _nd colleagues7 performed a material balance of Fe in the fallowing scheme. NaOH--- DCP MCAA-- \/ V/ ... : -- u , V Recvcle Phenate 1 't Product v WW/ They found that most of the iron is entering the process in the caustic and that the reactor normally contains 20 pom of Fe+^ . A recurring problem for the past several months was how to explain the fact that 'i : the p-ar.c c o served TCX formation rates that were 5 times greater than ' the lab. The iron results explain at least part of this discrepancy and suggest that efforts aimed at removing iron from the process are : desirable. i \ T. Based upon the data presented up to this point, the mechanism shewn in Figure 1 for TCX formation is suggested. This mechanism is consistent with the observations of the effect of caustic, the fact that TCX can be mads from NaDC? and glyoxal or glyoxylic acid and the levels of bis-2,4-0. If this mechanism is correct, it is easy to see how perc can influence TCX formation when one considers the possible hydrolysis products using either NaOH or NaoCF. Some of these possibilities are shown in Figure 2 and many are known precursors to TCX. n addition to TCX, OCSX and "8-5" are known to be formed in measurable quantities during the reaction to 2,4-0. It was initially assumed that TCX is the probable precursor to OCSX and "8-5". However, a. series of ampcule experiments in which TCX was mixed with various ratios of D C ? , 2,4-0, NaOH, and pare showed none formed in detectable amounts. In addition, a reaction was run in the presence of 500 ppm of added TCX and gave normal levels of the impurities when the 500 pom of added TCX is subtracted out. It is apparent that OCSX, "8-5", and TCX are probably made by independent routes and proposed mechanisms are Shewn in Figures 3 and 4. v i \ 25 uC:< i"i~ OC/73-3 OC/73-35 . c --- "7 . Ct c, ~e of ths Possible Products cf the Hydrolysis of Psrchiorcechvlene with haCH or MaiCP ) 2: OH v -- ^ OH C C 1 2=C\ Cl H,0 ----7 ?, CHC12-C OH DCAA -a 'I --.2 ilp.cp- ClJj-OC-C^ known impurity ^z X Cl 2 CH / Cl OH ">V - rc\7 cr oh or CCl2CCl2 HO HO ^0=0^ Cl^ \ : l cl 6oci 2 Maoc?. \ cc/ / Cl 00C1- CKC1--C ^ 2 \H 1 7. 2. H 20 ni-CO-Na H2^ CHC12-C-H 3 OH' Oiis:,c*.c^^0H HO ci H"2Oq 00C1. 3 Naf9,^. (cij'fao) 2c*c Cl 4 OH or ^ 4 N'acc? ?? HC'CH HC-COjNa lCl .I H-C-CH 'bocl. 27 J. -t 35 NoiOEI 2 ring closures as shown in the TCa mechanism 28 D~  7* w* CC/73-35 OC/78-35 \ D O rcro. <1 -^100 290 ppm of this mi:cture in Uiw . A proposed mechanism for. the formation of Complex Acid I is shown in Figure 5. It must be emphasized that the mechanisms shown in Figures 3,4, and 5 are speculations and are only one of a number of possibilities. Much more work is necessary to prove their accuracy. 30 Figure 5: Proposed Mechanism for the Formation of the Majo Impurity in Complex I mixture (Compound IV) ; BiS h 7o > 2 DC? -t- 0 2 COMPLEX ACID I OC/78-35 S V Final Product; Storage Tank: Tests run at 120-175' on C O1 1 wet and dry 2,4-D '* skewed that there is an initial slow build up cf about 20 pen of TCX during the first 43 hrs at >160 that stabilits cut. This suggests a small amount of an unstable unknown species that decomposes rc ,,1 in acidic media. Molten 2,4-D can be treated for at least 100 hrs under tha above conditions without c r. PC affecting product performance. Since TCX is formed in the reactor and was found throughout 948 Bldg process, a laboratory study of the distribution was requested. In this experiment a reaction was carried out and the crude reaction mass was spiked with TCX. The purification process was then simu lated and the fate of the TCX was determined. Figure 6 summarizes the results of the extraction study performed at 90*/atm. j 32 These c i u show an excellent TCI: recovery in which, using clean perc, ail tne detectable TCh is removed from the Na 2,4-D solution. The faco the plant observes seme TCX in the product suggests that scrr.e entrained perc is carried overhead so that the impurities R e a m e d with it end in the product. The extraction of the perc/DC? mixture with caustic to recover the NaDC? also extracts about 51 of the TCX which is recycled to the reactor. This scheme was followed in a second experiment that was performed at 125"/ 35 psig which mere closely fellows the plant conditions. Under these con ditions the aqueous solutions can be more concentrated in Na 2,4-D, NaCl, and DCP which could affect the distribution Of TCX. A high pressure, mechanically stirred glass reactor was constructed and was charged with partially neutralized material from V-301 (the DC? neutralizer). The molten material was extracted six times with perc at 125aC and the molten 2,4-D was isolated. The TCX and "3-5" levels were monitored throughout and the analytical results Tor the products are summarized in Table 5 for the two runs. The data shew that TC)C is effectively removed at the more drastic condi tions. Table 5: The Analysis of the Product from the High Temperature Extraction of Crude Na 2,4-D with Clean Perchioroethylene Run 1 2 t c :c (ppm) ^ 4 +. 4 * ^ final 292 13 178 2 "8-5 ' (ppm) init final 54 N.D. 52 25 Metals (pom) Na Fe Ca Mo 84 5 40 13 91 3 50 16 It is interesting to note that "8-5" is not efficiently removed with perc extraction. Hence most of what is made goes out with the produce. OCSX was not detected in these samples. In summa v, TCX is forced in the reactor and most of it remains in the esc: relating perchloroethylene system, A certain amount of TCX sp il s into the product and the magnitude is directly related te the 1 vai in the perc. TCX is easily separated by distillation a .<*,* v- - - oK rfct rc still so by increasing the rata of distillation one would ac ieve a lower steady state concentration of TCX in the perc. O o rc c: e: Assuming a rate of TCX formation of 0.5-1.0 #/hr in the plant (basea upon analyses of plant samples) K. . First 4 has taken distribution coefficient data and has modeled the process in terms of TCX content in various streams as a function of % of perc distilled in the perc still. To date, there is not enough in-plant data to verify this model. (C) Reduction of TCX by Chemical Means Two accroaches for chemical reduction of TCX were studied. These were methods for inhibiting its formation during the reaction step and methods for reducing it from process streams. The following approaches were studied and each will be discussed in detail. 1. Changing NaDCP/NaMCAA ratio. 2. Post reaction neutralisation. 3. Effect or DCAA. 4. Bleaching Na 2,4-D solutions. (1 ) chanoine NaDCP/NaMCAA ratio Excess caustic or NaDCP was shown earlier to have a significant qualitative effect on the increased production of TCX. The 2,4-D reaction as developed by H. B r u s t ^ used a ratio of NaDCP/NaMCAA m 1.2. As described earlier in the discussion, at this ratio the product contains theoretically ^6.3 mole 5 excess alkalinity as JaDCP as shown on the following page. % 34 i>'31 - Ji " OC/73-3 5 , ~) Sta Mat'l Product 0.5 me1 DCP 0.5 mol DCP -- 0.6 mol NaOCP ---- > 0.1 ml NaDC? ' 2 %. > 0 .S ml NaMCAA 0.5 mol Ma 2.4 # 0.5 mol NaCl 13 14 Dhin^ra and Fear evaluated the effect of this ratio on the yield and kinetics of this reaction and concluded: (a) lower ratios (below 1.2 lead to lower yields based on MCAA and, (b) low HjO in solution gives higher 2,4-D yields. The effect of changing the NaDCP/NaMCAA was restudied in order to determine if the lower yield from MCAA could be justified by reduced TCX formation due to less excess caustic. The runs were carried out by using a constant amount of NaMCAA and varying the amounts of NaOCP which was dona by adding differing amounts of caustic in the initial boil down step. Tables 6 ana 7 and Figure 7 summarizes the results of this study. These data show that lowering the ratio from 1 . 2 to 1 . 1 results in a'two fold reduction of TCX along with a 1% loss in MCAA yield. This is reason able and was tested in the plant. After three weeks of operation, they did not note a measurable reduction for reasons that are not understood at this time. Any further reduction is not practical since the yield rapidly approaches that of the old plant thereby losing much of the advantage of the 948 Bldg process. It is inte resting to note that at ratios below 1 . 0 there is still a measurable amount of NaOCP which explains why formation of TCX is still observed. All of the data to date indicate that formation of TCX cannot be limited to much less than 60 ppm (in the standard 72 hr/160C ampoule test) . 35 II o c / / TABLE 6 EFFECT OF NaDCP/NaMCAA RATIO ON TCX FORMATION YIELD ON MCfiA RIJN REF DCP I OC 417-5-143 0.50 2 OC 417-5-144 0.55 3 oc 417-5-146 0.60 4 oc 640-1-1 0.61 Moles NaDCP 0.60 NaMCAA 0.50 0.55 0.50 0.50 0.50 0.49 0.50 Mol Ratio NaDCP NaMCAA 1.2 ppra HOG 951 1 . 1 1514 1 . 0 3608 0.98 3606 MCAA YIELD 96.4% 94.2 66.8 86.2 CRUDE ANALYSIS Na 2, 4-0 45.4 DCP NaDCP 31 .2 8.9 45.5 38.0 6.0 41.3 40.3 2.2 43.0 41 .7 2.5 CO V VP moq no L L M L 'M L W t M U u OCR -33- OC/78-6 5 TABLE 7 MaDCP THE RATE OF TCX FORMATION AS A FUNCTION OF TIME, TEMPERATURE & HaMCAA RATIO RUN REF TEMP 1 417-5-143 130 145 160 2 417-5-144 130 145 160 18 HRS 37 29 18 0 27 42 TCX/OCSX 42 HRS 66 HRS 34 06 ' 42 *19 39 74 19 70 118/107 66 33 64 93 3 417-5-146 130 <10 <10 15 145 15 20 41 160 29 47 55 4 640-1 130 17 145 23 160 36 2B 22 31 34 61 65 114 HRS 53 218 240 114 llrs 52 93 142 90 llrs 30 59 67 90 llrs IB 52 77 CO 'O ^ \ OC/73-35 c i f UC H -35- (2) Post Raction Neutralizacin OC/73- 85 . fermacion might be greatly reduced. This concept was tested using 32* KC1, and dry and wet 2,4-D as the neutralizing acid. The first experiment used enough dry 2,4-D to completely neutralize the residual NaDC?. The neutralized crude showed 1.4% NaDCP and a rate of TCX formation as shown below that is consistent with the data shown in Tables 6 & 7. ppm TCX Temp 160* 24 hr 24 48 hr 51 72 hr 62 The use of wet, molten 2,4-D ( 25% HjO or 32% aqueous HCl has some serious physical handling problems. 16 As was discussed in an earlier report , water added -to 2,4-D raises the freezing point of the mass creating a material similar to cottage cheese in consistency. In addition, the added salts (either Na 2,4-D or N a d ? also raised the freezing point of the reactor batch from a range c: 115-120c to 118-125'C. i In summary, whereas post reaction neutralization reduces TCX bv a factor of four, it offers several disadvantages: (a) it would add at least 30 min - 1 hr to the reaction cycle in*a plant already reactor limited near capacity operation. (b) Addition of acid could cause corrosion problems in the Incolfcy 800 decanter . (c) The possibility of precipitates forming in'the reactor would put e: ces3 ive stresses on the agitator. Their formation could not be avotd|<Q DOW CC/73-35 in the laboratory even by very slowing adding the aqueous acid rc * C7 (3) DCAA Content of MCAA The effect of DCAA was discussed earlier (4) Blaachir.c Na 2,4-D Solutions The major chemical difference between 943 Bldg and 489 Bldg techniques for processing crude 2,4-D is that 489 Bldg uses a bleach step in which residual DC? is oxidized out of the Na 2,4-D solution with 8-12" NaOCl at pH 10.5/100C. Several experiments were performed to determine if bleach would effect the impurity levels and/or improve the ability of the plant to consistently produce 2,4-D that passes the formulation dilution test. The results of these experiments are shewn in Table 8 . The results of the first experiment appeared very encouraging since the dilution test solids, the TCX and the "a-S" were all significantly reduced. The second series of runs showed a consistent improvement in dilution test solids when near or out of spec, but not much effect 4 on the impurities. A capital estimate performed by K. E. First showed the cost of implementing a bleach step in 2,4-D to be $50Q*M which immediately eliminated any further interest in this project in view of the marginal benefits. It is assumed, posthumously, that the main effect of the bleach was to reduce residual DCP which iff a knewn contributor to poor dilution test results. 40 D-31 0 Lr -37- OC/73-85 Table 8 : The Treatment of Na 2,4-0 Solutions with Bleach Na 2,4-0 Solution SCUC Z2 pH Mol Ratio NaOCl Na ,4-0 V-501 (4-7-78) 948 Bldg. starting material 10.5 0 .1 ml solids DMA-4 F-40 0.13 0.02 -* dO TCX CCSX "8- 58 N.D. 127 25 N.D. N.D V-301 (6-5-78) 943 Blcg starting material 10.5 0.1 10.5 0.2 1 0 .5 0.05 10.5 0.1 5 0.1 0.01 TR TR -- -- " TR TR TR -- -- 80 N.D. 85 n 85 n 104 tf 94 m 72 n 31 10 26 23 16 47 o TO CT -i <1 V i 41 0 '** "33 - OC/73-35 (D) Removal of TCX bv Physical Means A numberof attempts have bean made to see if the levels of TCX can be reduced by physical means that include: O 0 ho c: -J <1 po 1. Carbon treatment o f `the recycling pere and the molten 2,4-0. 2. Removing pere from the reaction step. 3. Improved washing of the molten 2,4-D 4. Recrystallization of Na 2,4-D and recrystallization of 2,4-D acid. The details of each alternative are discussed below. In all cases, these experiments were short term, range finding efforts and not comprehensive. Further work is justified only if the levels of im purities presently found in the product prove unacceptable in the future from a toxicity, environmental, or performance standpoint. (12 Removal of TCX with Activated Carbon The removal of TCX with activated carbon from perchloroethylene from v-402 ("clean1' perc storage tank) was evaluated by a standard isotherm method. A 100 ml portion of perc was treated with ground Pittsburg SGI, carbon at 70*0/72 hrs. The data, summarized in Table 9, were evaluated by a known technique that is summarized below. NA plot was made of X/M vs. C on log-log paper as shown in Figure 5. By extracoling C to incoming TCX concentration, the corresponding X/M value gives the amount of impurity absorbed per unit weight of carbon when that carbon is in equilibrium with the incoming concen tration an'd represents the ultimate capacity of the carbon. The 42 9 3? in \,U^i OC/73-3 *- '* a w C ^ .a -- \ '"< i 1 -- w- ^^ -- volume of liquid to be completely freed by TCX/g m of c i r : c r , i s c ilculated from the fcilcvinc equation: A T (.>c= Co Co V theoretical volumn to be treated * capacity/gm at incoming concentration V = vclume of liquid used in test Co incoming concentration For this experiment: VCo " lil X 100 * 166 ml/g of Carbon or 19.9 gal of perc/# C or 265 # perc/# C Carbon loading at 241 ppm of TCX in feed * 166 ml/g C X 1.6 g ml x .00024 64 mg/gc v j 4-1 0-4 H U K M oo oc?. OC/78-85 Table 9: The Absorp tion_ Iso t h e m for TCX Removal from Ferchloroe thy iene with Pittsburg SGL Activated Carbon 1------ -- -- t 1 ! Samcle iI * . Impurity (ppm) t c x QCSX "3-9 " (M) (C) Re sid WTC/I00ml ppm Pere TCX (X) ppm TCX Removed ( STG. Mat '1 1 1 -2 3 4 5 241 N.D. 19 159 It 108 it 57 n 17 40 ii 29 it 17 0 0.25 0.50 1.00 1.50 2.00 241 159 108 57 40 29 0* 82 133 184 201 212 (X/M) ppm TCX removed gm Carbon 328 266 184 134 106 I 44 JM ' nriWni ' n n n r\ OCR -41- OC/78-85 oc* 'U 00/78-35 This i s a low loading level and at the present rates a t TCX pro- ''N \ \ O O ^ ducctcn in the plant, about 15-20 # of carbon per hoar would be r e c u r r e d to remove the impurity. The data shown in Table 9 indicate he that carbon aces not remove "3-5" from perc. The shape of the isotherm curve indicates that more than one species is being ab- s o r b e d onto the carbon. No effort was made to determine the identity or that compound although it was probably perc. A run was made to remove impurities from wet molten 2,4-D. Since molten 2,-D is so difficult to handle at atmospheric pressure, pressurised system was built using a capillary feeder for controlling the continuous flow of molten 2,4-D at 100V 2 5 psig onto a vertical 2' x 0.5" column that was maintained: liquid full. The flow rate was 10 ml/min down the column for a mass flux of M ..0 gpm/ft. 2 The result-: are summarized in Tabl 10. Table 10: The Treatment of Molten 2,4-D with Activated Carbon no le TamD rci Pressure ps ig Approx feed rate ml/hr Approx. Prod, cut v v j xuiue (ml) Product Analysis (ppm) TCX OCSX 100 Ul 3 >-40* solid: (ml) ed 1 1 2 1 10 -2 33 2 1 10 33 n o 33 1 1 0 33 1 1 0 33 300 100 600 375 600 375 600 376 600 37S srmulaticn contains 2" Versene & 0% PG 4000 42 14 28 32 30 N.D. -- N.D. N.D. N.D. N.D. 32 -- 29 38 42 v 36 ------i--- .005 .04 .005 *se data are fairly crude in that the column was shorter than is desirable optimum column work (2 ' vs the recommended 5-6'^) and the flow through s column was a little faster than the more desirable 0.5 gpm/ft recommencec. 46 V-H3 gcr - 43- OC/73-35 DOW The ci:a show that TCX loadinc is rather lew. Breakthrough occurred attar 900 ml molten. D (^935g pure 2,4-D; was treated. From the a ~alysas, ^he 59.3 gm of Pittsburg SGL 8 x 30 granular carbon charred to the column absorbed 281 me of TCX for a loading of ^0 .2 mc/g cf carbon which is quite lew. Again, "8-3" was not absorbed by the carbon. Based uten these preliminary experiments, carbon absorption of these impurities is not an attractive puri fication technique. (2) Remove Perchloroethylene From the Reactor Since perc is known to give rise to TCX, several methods were examined to eliminate its recycle to the reactor. At present, since perc is soluble to 1.2% in the recycle NaDCP solution, there are about 350* returning to the reactor in each batch- Since a continuous percphenate phase separation is performed in V-401 just before recycle to the reactor, a brief study was undertaken to evaluate the phase separation to determine the time required for complete layer separation and the solubility of pare in NaDCP solution. A solution of 62% NaDCP and 2% NaOH in water was slurried with an excess of perc and vigorously stirred for 30 min at 75. The stirring was stepped and the p'nenate layer was analyzed for I perc as a function of time. The results are summarized in Table 11. These data show that the solubility of pare in NaDCP is 1.3t0.1% at 759C and that layer separation is complete within 15 minutes. The residence time in V-401 is ^4 0 minutes so that with proper operation, no layer separation problems should result. The question was also asked if the percvwas entrained in the NaDCP solution as an emulsion or was it in solution? A sample was centrifuged at 60 (minimum temperature) for 30 miri at ^4000 rpm and 1.2" perc was found in the phenata. It is concluded that the present equipment gives optimum layer separation and the 1 .2% of perc is soluble. 47 p -H DOW 2 (>7778 OC/73-85 T^bis 11: The Solubility and Separation Rats of Perchloroethylene and 62*i NaOC? in Water at 75 3C TIMS* 5 rain 15 " 30 " 1 hr 2" 46" 24 " I Perc in Phenate Laver 1.9 1.3 1.4 1.5 1.4 1.4 1.2 1.3 *Tirae zero is when the vigorous stirring of the two layers is stooped. v * 48 CCi< OC/73-35 Several attempts were made to strip perc from the reaction mass and from the NaDC? solution. The plant tried to strip perc from the : ----- r. mass during the normal boil dcvr. star. Whereas normal -> opera > rAao t rc Where W 0 this < was s appro, the 1.3$ of perc from the 60% NaDC?-2% NaOH solution. The first involved a batch distillation from a standard solution from v-403. A total of 65.9 gms were distilled out and 73.8% of the perc was removed. The reaction was then carried out as usual and samples ware heated for 24, 48, and 72 hrs and the results are summarized below. These data suggest that distillation of 75% of the perc does not reduce TCX formation (See Table 4, Runs 4&5) . ppm TCX Temo 6 24 hrs 2l 48 hr 72 hrs 71 ' 145 The second approach taken for stripping out perc was using a falling film still. The results of two experiments are shown in Table 12. The results show that, as before, 70-80% of the perc is easily removed but that the last 20% is likely to be quite difficult. No more work is planned in this area until it can be better proven that removing perc offers any real advantage"in reducing rates of TCX formation. (3) Imccoved Washing of 2,4-0 Acid ^ Several experiments were run to see if improved washing would affect impurity levels in the 2,4-D product. A sample of 2,4-D product was 49taken from V-*602 (948 Bldg) and was treated in the following ways: (a) A 250 gm sample of molten acid was washed 3 times with 200 ml of water per wash at 1GQC. 'p-'fk V-M-*\ -46TABLE 12 OC/73 ""32 Preliminary Data for Falling Film Distillation of Parc From Na.DC? Soluhi Still: 1" x 11" Tube DOW Pressure Temp (col'm) " `(feed) Feed Pare (ave.) Overhead Tamp Feed analysis % Perc % DC? W t . Charged (Feed) Product: wt % Parc % DC? Overhead: Wt % Per: Run 1 120 75* 10 ml/min 85*-- > 90" 1.0 Sl.S 725 gms 652 gms 0.3% 53.9% 46.7 gms 1.5% Run 2 ATM. 75" 10 ml/min 90-- ^ 95 1.1 50.1 738 gms '645 gms 0 .2% 61.6% 75.0 gms 3.3% W# -si CO V 50 -47- OC/73-3 E . T8Ui>ZWvOQ (b) Another sample was reacidified to pH 0.5 with ccr.c HC1 and rewasned 2 times with water at lOQ^C. A synthetic V-501 mixture was made up usir.c Rhone Przcii 2.4- D to determine if any Impurities are made in the washing step. The synthetic mixture was acidified and washed and the 2.4- D was recovered far analysis. This experiment was repeated in the presence of 2750 ppm of added TCX. The results of these experiments are shown in Table 13. Based upon these experiments, it is shewn that the impurities are not made or reduced by improved washing. If anything, they are slightly increased in the product due to the greater solubility of 2,4-D in the hot water or brine. (d) Recrystallization of Na 2,4-D and 2,4-D Acid An attempt was made to determine if the impurities could be removed by recrystallization of Na 2,4-D from water and 2,4-D acid from organic solvents. Na 2,4-D was recrystallized by taking 500 gms of ' material from v-501 and adding enough water (350 ml) to form a homo geneous solution at lOO^C. The solution was cooled and the precipi tated :."a 2,4-D wag filtered and washed with 5% brine. The Na 2,4-D was redissolved in hot water and the 2,4-D was isolated and analyzed. The results are shown in Table 14. In twe separate experiments, 2,4-D from V-602 was recrystallized frem perchlcroethylene and ethylbenzene. A weight ratio of 3 parts solvent to 1 part 2,4-D was heated to boiling, the water contained in the molten 2,4-D was boiled out as an azatrope (the organic distillate was returned) and the solution was cooled. The 2,4-D was recovered by filera cron and the solvent removed by heating in a vacuum oven at '\,6Q"C- The results are summarized in Table 14. 51 p-H* 48- OC/73-35 Table 13: The Effect of Improved Washing of Molten 2,4-D on Levels of Imourities C? 3o Material Imsuriti es 1 | TCX o c s x "8-5" V-602 Starting Mat'l Wash three times Reacidify, wash two times Synthetic V-5Q1 Na 2,4-D: Naci DCP 21.5% 7.0% 0 .2% After Acidif, & Wash Repeat the synthetic V-501 spiked with 27 50 ppm TCX After Acidif & wash 56 N.D. 58 73 N.D. 60 96 34 68 N.D. N.D. N.D. 3150 N.D. 10 Mote: Rhone Progil 2,4-D shows no detectable TCX, OCSX & "8-5" V I 52 li*" O C / y 8-3 5 , O O * -Zj Tarie 14: <1 cc The Effect; of Recrystalli acion of Na 2,4-D and 2,4-D Acid on Impurity Levels Treatment Recrystallize Na 2,4-D Before* Recrvstallite 2,4-D From Pero Recrystallize 2,4-D From Ethylbenzene After Before After Before After Impurity Level (ppm) TCX OCSX "8-5" 53 -- > - 127 P-40** Dilution Test o .ia 81 53 52 N.D. 1 N.D. 52 N.D. 54 46 N.D. 46 0.3 + ** *1* N.D. N.D . N.D. Corme Predue Highly lored 91% Re verv o 2,4-D 87% Re verv o 2,4-D The impurities analysis was performed on a sample of 2,4-D acid isolated from Na 2, 4-D without any extra treatment. **2 0 :1 dilution in 1000 ppm hard wate anc no P-4000. formulation contained 25 Versene ***The recrystallized product behaved like the high purity Rhone Progil 2,4-D which is difficult to formulate as was mentioned earlier ig this report. i 53 v -w oc?. -50- O C / 7 3-3 5 I'R A i n r i M o a GENERAL CONCLUSIONS TCX and other non-acidic impurities are formed chiefly in the reaction step of the 2,4-D by several routes. A number of attempts to chemically and physically remove these species have net with limited success. SAFETY & ECOLOGY 2 ,4-Dichlorophenol, 50% NaOH, and chloroacetic acid are highly tcfciic and corrosive raw materials. When handling, the protective clothing included lab coat, rubber gloves and goggles, and when possible^ all operations were performed in a fume hood. A number of operations were carried out at elevated pressure which required the use of a face shield and secondary shielding in the hood. All waste samples^ and solutions were sent to the burner for disposal. -S' O C R 51- OC/78-3S M tr i7 The ?recarasion and Workuo of 2,4-D The following is a general description of the procedure used to prapara and isolate 2,4-0 when simulating 948 Bldg. A 1 liter round bottom flask equipped with a bottom drain, two dropping funnels, a mechanical stirrer, a thermometer, and a distillation head was charged with I79g (1.1 moles) of 2,4-DC? and 48 g (0.6 moles ) of 503 NaOH. The flask was heated with a heating mantel attached to * an 1^?. Thermcwatch controller. The reactor contents were heated while stirring and a DCP-watar a 2etrope was distilled out. The distillation was continued until enough water was removed so that a temperature of 130*C could be achieved. Normally 9-10 ml of HjO and 2-3 ml of DC? were removed. The DC? was returned to the pot. Then 47.3g (0.5 moles) of melted MCAA and 40.Og (0.5 moles) of 50% caustic were con-added from the two cropping funnels during 50-60 minutes at 130C. The rates of addition were carefully controlled so that neither added reactant was signi ficantly in excess of the other. Water and DC? continuously distilled out during the con-add and the DC? was returned to the reactor. Aftar the addition was complete, the reaction was heated an additional 60 minutes. About 43-47 gms of HjO was recovered in the con-add step. At the end of the post reaction samples of the viscous crude reaction melt were taken into ampoules,'if desired. v The wc rk up procedure for isolating the 2,4-D is as follows: (The amount s used assumes no samples were taken after reaction) , T^ie reacti on mass was diluted with about 500 ml of water, heated to boiling to er.sora complete dissolution and the pH was adjusted to 5.2SQ.2 with abou t 12 ml of cone HCl. The solution was then extracted with six 150 mi portions or perc at a temperature of >90oC to remove the DC? Occus icnaiiy SO-iQO mi of additional H2O was necessary to keep a1,,15 o5f the sc luds dissolved. The extracted Na 2,4-D solution was heated -b 9 DOW 26778G oca 52- CC/78-3 5 tc boiling and any traces of pore were distilled off. The pH of the scij.-L.zr. was than lowered to 0.5-0.7 with about 55 ml of ccr.c HCI added rapidly and the molten 2,4-Q layer was separated and drained into a beaher. The resulting brine was discarded. The 2,4-D was raslurried in 250 ml of hot distilled HjO in the pot and washed in this manner two times. The final pH of the aqueous layer was 2.7-2!.9. The 2,4-D was recovered and dried overnight at. ambient temoera fares. The analyses of product and intermediate streams were performed by t personnel in the 948 Bldg quality control laboratory. The analyses for TCX, OCSX, and "8-5" were performed as described earlier. The carbon Clean up of MoLten 2,4-D A glass pressure apparatus was assembled in which molten 2,4-D was pumped onto a carbon column (downflow). The apparatus is shown j schematically in Figure 10. The flow of 2,4-D was controlled by tontrolling the pressure drop across a capillary tube. To handle 56 --f CG N,A. prv 50# / Q -- OJ < ' r \ C i v. V J I Figure 9 : A Schematic Drawing of the apparatus used for Extracting DC? frcm Na 2,4-D with Perchloroethylane. 57 n . Schematic Drawing of the apparatus used to treat Molten 2,4-D with activated carbon 88 W 3 M OO o p .^uA ay Fiscxs*- r ~ ... : i_______ TC. `lessee* C /- tI v_ 'Tm m C. rc. ic.i%ASfc. i i " 22- O C / 7 3 - 3 S ' r mol tan 2,4-D reliably and effectively( required at least 10 psig/ vlOO-LDS'C to avoid flashing and freezing problems. The 2,4-D and a slight excess of water were placed in the feed tank and heated to 110"C. When the entire contents were melted the valve on the bottom of the column was closed, the column was filled with molten acid to 1 " above the top of the carbon, and the system stood for 60 min. The pressure drop across the capillary was adjusted to 7 psi (^10 ml/min flow) and the valve on the bottom of the column was adjusted so as to maintain the liquid level above the carbon bed. The results are shown in Table 10. UdLL9Z v 59 V DOW"" ACXNC'.'TLEDG NTS The un r.cr wish tie thank D. Hurtberc, R. Maciac hian, T. Evans, G. Jeve tc, P. Schicemann, and their coll eegues for their anaIvor cal support:, The assistance of K . First, and his process mod el ir*g efforts are also acknowledged. cz V 60 references /'~ \ \ A / 1 . ?.. McLacr.lan, D. Hu,mbsrr, G. Kallcs, AL 7800101, to be 2 . EST K--2372-- (18) February 9, 1973 . 3 . u Humbert et. al., unpublished resu its. 4. On published results to date. 5. -Seilstain 1/ 17, 354 . 6 . Organic Synthesis Coll, Vol 1, Pg. 552. 7. Bailstein 19 (3) 222, Merck Index 9, 4040. 8 . Morrison & Bovd Organic chemistry, Allan and Bacon, Boston (1953) Pg. 684. 9. . (a) S. Maclean, AL-75-20021. (b) V. Stevens, Communication, Inorganic Chem. TS4D. 10. D. Humbert, unpublished results. 11. X. L. Krmel, LR 78-74, 8-15-78. 12. K. F. Brust, OC 0730013-2, 2-26-73. 13. Y. R. Dhingra OCR Lab Book OC 327-5, p. 22-67, 1972. 14. . L. Fern, PS 73-10, 4-16-73. 15. F. G. Aerstin, private discussion. 16. X. L. Krmel, OC 78-17 LR, 39-78. 17. S. Siegel, Results to be published. 13. Absorption Handbooks Issued by Activated Carbon Div., Calgon Corp., Pittsburg, PA. v LABORATORY NOTEBOOK REFERENCES: R. F. Arnold OC 417, Pg. 109-150, OC 640 Pg. 1-73 K. L. Krmel OC 559 , Pg. 95-131. 61 j '.tfV ,h-'- V . -- r~- THE DOW CH EM ICAL CO M PAN Y 16ZC31M0(T#-. YS%Y: ' : ? x ; v - v - - MIDLAND DIVISION April 25, 1967 K. E. Coulter Midland Division Research & Development 566 Buildins CHLORACNE RESEARCH PROGRAM .15J9 o O'. History of Chloracrte Incidences at Dow: Historically, Dow Chemical has been involved in chloracne ,O incidences ever since Dow began the production of chlororh .nols. At first, the commercial production of chlorophenols was conducted at 206 Building. During the period 1934-36 there---- --- was a severe outbreak- among the employees which resulted in an unsuccessful lawsuit. The chloracne incidents were traced back to poor working conditions and the manufacture of ya 5*17-5 Sr Dowicide P. The manufacture of Dowicide P has since been opT ^ o w .3 +) terminated. New and improved working facilities for the I * upow. 6 ^ J Dowicide group were constructed at 265 Building in 1940. the late 1930's, Wes Stoesser in 20A lab got a serious chloracne attack from chlorinated diphenylene oxide. Drastic treatment:? were used to cure this Incident and no further work *W>!C ''Van cone on this series of compounds. It Is suspected that many of the other incidences -of chloracne are caused by 'chlorinated diphenylene oxides or analogues thereof. % " 'In the next ten years, another unfortunate situation occurred '..in the chloracne situation. Some of the Dow customers complained about dermatitis and/or chloracne from the use of Dow's Ddv/3 1+ . Dowicide 3 * I understand financial adjustments were made and' m ' Sul**S'Sbg production of Dowicide 3 terminated. A purified material n o M LHionoj Qf reiated structure is-now sold as Dowicide 31 and 32. PHBML ' During'the period 1940-65, the product 6x (diphenyl oxide V '" t i l / j <, chlorinated to the hexa level`) was manufactured at 206- Building wss-#:j+`.- and a {. 2.east one severe case of chloracne occurred because of this product. The production of 6X has been terminated. In the research.lab at 172 Building, there were some cases of chloracne from research exposures. In one case, a severe case resulted from the recycling of residues from the manufacture of 2,4,5-trichlorophenol using glycol as a solvent. In another case, several mild cases occurred hydrolyzing polychlorobenzenes using aqueous caustic at high temperature. The Chemical?FhysIcs Lab had several incidences of chloracne from the recycling of ,, caustic insolubles in the alcoholic caustic hydrolysis of T.:t - - tetrachl.orobenzene. 62,V r- 0; j T n 1965, the 2,ty,5-trichlorophenol hydrolysis step at 199 building was modified for' economy and safety reasons from the use of 100^ caustic to 2Jffo caustic. The rabbit test for chloracnigens in the caustic insolubles obtained from a pilot run at this time indicated , that the test response for chloracnigens showed no difference between caustic insolubles obtained by either procedure. haring the latter part of 1963# the production department, in order to increase capacity, raised the reaction temperature and increased ' the throughput. This meant that more caustic insolubles were produced, and more Dempster loadings had to be made. This meant ! that employees had more exposures to the caustic insolubles, and ' the caustic insolubles due to the higher temperature had larger concentrations of chloracnigens. Thus th higher concentration [ of chloracnigens and the more frequent exposure caused many mild incidences of chloracne in 199 Building and two severe cases (LTI's) In 1906, a new process (Boehringer) was put Into operation using a batch reactor at low temperature, and so far has operated satis factorily. * After they started up their Dowicide plant, the Canadians began to experience chloracne incidences in their employees. In Midland, more than half of the Dowicide employees have chloracne of varying intensity and it is impossible to say when or where these incidences occurred. Research Program in Progress for Chloracne Reduction: 'Because of the prevailing existence of chloracne in the Dowicide plants and a sincere desire to reduce or eliminate this, research has been initiated in 1966 on this-problem. Progress has been slow due.to the complexity of the problem. The problem involves isolation and determination of the identity of the chloracnigens. After being properly identified, work can progress on its reduction or elimination in the process. In commenting on chloracne.*, we must keep in mind that chloracne is a cosmetic evidence of the attack and serious liver damage, is an invisible effect of the attack. Rabbit ear tests are a positive sensitive method of determining the chloracne activity of chlorophenol Impurities. Present Knowledge of Chloracnigens: Compound Activity 2,5,7,8 Tetrachloro - Very positive Cl 1 Cl1 --3 - Not active Cl 1 Cls-7 - Possibly some cpds. active C1Q - Not active / Very active Activity I Very active 0uw1 2 'j i yv i : - Unknovm activity J CI1-2 - Unknovm activity CI3-6 - Some very active CI4-0 - Some activity? CI3 + - Some activity Materials possessing unexplained chloracne activity: Some Midland batches are mildly -- . -- reactive Many Canadian batches are mildl -- reactive ~ . NaOH Dowicide G .* ...?. ? All sludges are active 64 o - 'u ..-Compound >2033 Activity ...All decompositions are active OH - f Some have activity toW1232913 Cla-s Midland Research Program on Chloracne; -: The following chloracne activity fractions are being examined for isolation, identification," and minimization of chloracne activity. Unfortunately, due to shortage of technical help, the program is proceeding quite., slowly. 1. Dichlorophenol still residue. 2. Pentachlorophenol process samples. a." Dowicide 6 and 7 "active" batches b. Dowicide G sludge. c. Dowicide G scrubber sludge. The method of research is t o `first concentrate the sample (remove .chlorophenols); then fractionate by chromatography; test fraction on rabbits, then further fractionate by chromatography, then test fraction on rabbits; etc., then determine structure by micro analysis; then determine method of analysis in original sample; and .then investigate process changes which will minimize the chloracnigens in the process. Th Benzene Research Lab_and the Biochem Research Lab are collabora^ ting in this project. The Benzene Research Lab does the chemical research and the Biochem Research Lab does th testing on the rabbits on their charge. Alex Widiger Benzene Research Laboratory 4.74 Euilding de cc; W. H. Haberstroh, 474 Bldg R. C. Sauers, 47A Blcg. SC . TL. Bl j eMndAeArM," 1721 y O Blcg.13 "I * M e ! c ! Staehling, 253 Bldg. L. Silversiein, 1701 Bldg. it 65 I R & D REPORT -< I DOW C H E M IC A L U.S.A. R E S T R I C T E D : for u s e within T h e D o w C h e m i c a l C o m p a n y DEP A RTM ENT DA TE ISSUED October 2 7. 197S P R O B L E M NO. ORGANIC CHEMICALS RESEARCH 19 io o o o a o TI T L E A STUDY OF THE FORMATION AND REMOVAL OF IMPURITIES IN THE SEMI- 57 HYDROUS PROCESS FOR 2,4-D___________________________ OC UJ 203 A U T H O R IS) Z) Z K . L. Krumel & R. F, Arnold A U T H O R ( Si S I G N A T U R E S) uO C i - . kcL-- C^ SIG N A TU R E This report is: ti INTERIM FINAL and mainly: PAGES IN FUL REPOR' X NEW REVIEW D ESC RIP TIV E SUMMARY WITH C O N C L U S IO N S : ( I n c l u d e in this s p a c e r e fe r e n c e s to data b o o k s , and to earlier relatec reports, patents and publications*) Shortly after the startup of the new 2,4-D_ process-- in--9-4.8_Building, a new and unexpected class of nonacidic impurities were isolated, in which two of the major components were tetrachloroxanthone ancl octachlorospirobixanthene . These impuritie's"weT5~Ccnising -problemstQ oin the subsequent formulation of 2,4-D as amine salts. A project was started to learn the source of these impurities a n o ^ ^ methods for controlling them. It was found that the impurities are formed mainly in the 2,4-D reaction step by several different^-^ routes and that the rate of formation is increased by increased caustic ratios, perchloroethylene, heat and iron. -yj'^ A number of techniques were evaluated for removing the impurities including carbon treatment, recrystallization, bleaching, varying reaction conditions, and improved washing. None of the treatments were totally successful. In the plant, the impurities tend to concentrate in the recirculating perchloroethylene system. It has been found that by increasing the capacity of the perchloroethylene distillation column from 0.7 to 15 gpm, the steady state concentration of impurities is reduced to where they do not adversely affect general product quality or the ability of the plant to operate at optimum rates. i pjd DISTRIBUTION: DEPARTMENT FILES R & D ADMINISTRATION c e n t r a l REPORT INDEX '566 3ldg. -- Midland) 5 COPIES Distribution list is c o n t in u e d on attached page. O C R -1- OC/78-85 DOW 2 6 7 7 3 5 TABLE OF CONTENTS INTRODUCTION -------------------------------------------- Page 2 RESULTS AND DISCUSSION ---------------------------------(a) The Mechanism of TCX Formation--------------------(b) The Distribution of TCX in the Pro c e s s ------------(c) Reduction of TCX by Chemical M e a n s ----------------(d) Reduction of TCX by Physical M e a n s ----------------- 9 10 26 30 38 GENERAL CONCLUSIONS ------------------------------------- 50 SAFETY & ECO L O G Y ---------------------------------------- 50 EXPERIMENTAL -------------------------------------------- 51 ACKNOWLEDGEMENTS---------------------------------------- 56 REFERENCES 57 og.R -2- O C / 7 8 - 8 5 INTRODUCTION A new process for preparing high purity 2,4-D was started in May 1977 oQ at 948 Building. The high purity molten acid is transferred by pipe ro line from 948 Building to 489 Building where it is formulated into esters and water soluble amine salts. Starting in late October 1977/ precipitates were observed regularly in diluted amines formulations -0i3 that were found to be made up of a number of impurities associated with the process but that were never detected during laboratory or pilot plant development work. In addition, these impurities were not found in the old 489 Building aqueous 2,4-D process. R. McLachlan^ analyzed the precipitates and found 1 ,3,6 ,8 -tetrachloroxanthone (I: TCX), 1,1'3,3'6 ,6 18 ,8 '-octachloro-9,9'-spirobixanthene (II: OCSX) and marginally soluble salts of 2,4-D. Cl Lo Cl Cl Cl IQl Cl c r Cl 0 (I) C.J& Clv/^s/ CUyJ O| Cl Cl (II) A screening program in the plant showed TCX at levels of 200-500 ppm in the crude reaction mass, the recycle sodium dichlorophenate (piaDCP) solution, and in the final product. Levels of 1000-2000 ppm were found in the perchloroethylene (Perc) solvent and as much as 10% in the perc still tars. 67 McLachlan analyzed a sample of perc still tars and his results are r summarized in Table 1. p-f DW 267737 OCR -3- OC/78-85 1 Table 1: An Analysis of One Pere Tar Sample Structure Approximate Percentage Comment Perc Diethyl benzene DCP 2,4-D TCX (I) OCSX (II) Dichlorophenyldichlorovinyl ether Dichlorophenyltrichlorovinyl ether 60% 25% 0.5% 0.01% 1.5% 1.1% 1.2% 0.8% Main Component From a heat exchanger leak (Dowtherm J) raw material Product Rxn product of NaDCP & pere & its impurities Plus; at least eleven other minor components structurally similar to the above. See Ref. 1 for details. / k it W 2677,38 OCR -4- OC/78-85 A project was started to study the formation of TCX and OCSX in some detail to learn more about their formation and fate in the process because it is not obvious how they are made. Several months after starting this study, another significant impurity was found in the product and, to a lesser extent, in the perc tars. It was identified as (III: "8-5") 2,2',4,41,5',7,71,9-octachlorospiro-(benzofuro(3,2-b) benzopyran-11,91-xanthene). Cl Cl A description of the 2,4-D process in 948 Building is as follows: Main Reactions OCR -5- Important Side Reactions ClCH2C02Na + NaOH -----> H0CH2C02Na + NaCl Glycolic Acid O C / 7 8 - 8? DOW 267739 The reaction is performed in two stages by first reacting 2.2 moles of DCP with 1.2 moles of caustic (added as 50% NaOH) to form a solution of NaDCP in DCP. High purity 2,4-dichlorophenol (2/4-DCP) made by the chlorination of phenol with sulfuryl chloride in the presence of FeCl^/Diphenyl sulfide catalyst is used with chloroacetic acid (MCAA) made by the oxidation of vinylidine chloride to make 2,4-D as shown above. The typical raw materials analyses for each is shown in Table 2. Table 2: Typical Raw Material Analyses for the 2,4-D Process DCP MCAA 2,4-DCP 2/6-DCP Other Chlorophenols 98.5% ^0.8% ^0.5% MCAA Dichloroacetic acid Chloromaleic acid h 2o 99/.0% 0.3% vO. 05% <1.0% OCR -6- OC/78-85 The solution is then boiled down to remove most of the water. Then, 1.0 moles of NaOH (added as 50% NaOH) and 1.0 moles of MCAA are added to the solution at 130 and the water formed is continuously q boiled off. The crude product, whose approximate analysis is shown below, is transferred to a storage tank that feeds the continuous purification equipment. 267740 Approximate 2,4-D Crude Product Analysis Na 2,4-D Na DCP DCP NaCl h 2o 45% 10% 30% 12% 3% The 2,4-D is purified by partial acidification to selectively neutra lize the DCP, extraction with Perc to remove the DCP, and then final acidification to liberate molten 2,4-D which is decanted from the brine and washed with water. A schematic flow sheet of the process is shown below. A rapid initial screening of the toxicity of purified TCX and perc tars gave the following results. Pere Tar (contains 3-5% TCX + OCSX) Oral LDS0 (single dose) >5g/Kg Skin/Eye slight transient initiation. Chloracne response after 10 applications of neat tars. Purified TCX >4g/Kg not irritating. Chlor acne response?after 10 applications of 1 % solution in CHCl^. OCR -7- O C / 7 8 - 35 267741 i/1 \ u o 0 0 1 ~~-J" 0\y ,267742 OCR -8- O C / 7 8 - 85 These results are encouraging from a plant hygiene standpoint, but more data is necessary regarding long term effects. When the im purities were first discovered in the plant, the perchloroethylene still was used very occasionally and at a rate of <1 gpm vs a total perc flow of 40 gpm. The still was originally added to the process as a way of purging contaminants out of the stream to maintain high quality recycle solvent. At this point, perc tar impurities made up 60-75% of the precipitates formed in diluted amines formulations. About February 1978, the capacity of the perc still was increased to 3 gpm and the average levels of TCX levels dropped to ^50 ppm in the crude reaction mass and product and to 200-500 ppm in the recy cled perc. From the period of March-May 1978, the ability to produce quality 2,4-D continued to improve and the amount of perc tar im purities found in the dilution test precipitates dropped to 1 0 % of 3 the total . The balance were sparingly soluble metal salts of 2,4-D and some new impurities called Complex 1 and Complex 2, two of which are shown below (IV & V). co2H IV In July 1978, the perc still capacity was increased to 15 gpm and 70efforts gained at understanding the parameters affecting formulation quality by S. Siegel (OCR), S. Schell (Production 489), and J. King r (Formulation - 9001 Bldg) and their co-workers have greatly improved X? ' OCR -9- OC/78-85 ,GW9Z Moa the consistency with which 948 Bldg. 2,4-D can be formulated. Based upon partition coefficients and data on the rate of formation of TCX to be discussed later in this report, K. E. First 4 (Process Engineering) .generated a computer program which predicted that by distilling 15 gpm of recycle perc, the steady state level of TCX in the product would be VLO ppm and in the recycled perc the TCX level would be M 0 ppm. The purpose of this report is to present data on the formation and fate of the impurities in the 2,4-D process. In addition, the results of experiments aimed at reducing or eliminating the impurities by chemical and physical methods will be presented. RESULTS AND DISCUSSION This study was divided into the following sections and the results of each are discussed separately. (A) The mechanism of TCX formation. (B) The distribution of TCX in the process. (C) Reduction of TCX by chemical means. (D) Reduction of TCX by physical means. Although there are many impurities formed in this process, it was decided to focus the research on TCX for several reasons: (1) The rate and mode of formation of TCX was the most predicable of the major multicyclic impurities; (2) it was a major impurity; (3) a pure sample was readily obtained (by D. Humbert, Anal. Lab); (4) the analysis is not complicated; and; (5) structurally, it is the simplest of the multicyclic impurities. i no b 0 L L 93 Moq OCR -10- O C / 7 8 - 85 (A) The Mechanism of TCX Formation A brief search of the literature including Chemical Abstracts showed 5 that the xanthone ring system is formed from a number of reactions. A common synthetic reaction is shown below in which the -X is included O OH 0 to show the orientation of reaction. Xanthone is formed along with other products from the pyrolysis of o-chlorobenzoic acid, salicylic acid, aspirin, o-phenoxybenzoic acid, and salts of the carboxylic acids. Most of these cyclizations take place in the presence of acidic catalysts such as P25/ Acetic anhydride, or sulfuric acid. Xanthone is prepared in reasonable yield** by the pyrolysis o phenol salicylate with or without a catalyst. The ease of formation of ring systems that are structurally similar to xanthones is most clearly shown by the formation of fluorescein and fluorescein dyes. Fluorescein is made by reaction of phthalic anhydride and resorcinol as shown on the following page. / 71 OCR -11- OC/78-85 DOW 2 6 7 7 4 5 g Phenolphthalein is also made by this process using phenol instead of resorcinol. I <y OCR -12- O C / 7 8 - 85 \ DOW 267746 colorless red These background data give some important clues as to reaction mechanisms as will be whown later in the report. 72 OCR -13- OC/78-85' DOW 267747 In order to understand the pathway of TCX formation, a series of ampoule tests were performed to determine which chemicals and/or combinations give rise to TCX. Table 3 summarizes the results of this study. A series of ampoules ( 25 ml capacity) were charged with 3-6 gms of mixtures in molar ratios as shown inTable 3 and were heated in an oil bath fo 72 hours at 160C. The sources of the chemicals were as follows: 2,4-D: High Purity Rhone-Progil Acid DCP : Doubly distilled 948 Bldg material Perc : j NaOH : / Glyoxal Glyoxylic acid Reagent materia Tetrachloroethane (50:50 1,1,2,2- & 1,1,1,2-isomers) OCR -14- OC/7 8-8 5 Run # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 2,4-D 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Molar Ratio Bis DCP NaOH Pere 2,4-D Other 1.0 1.0 2.0 1.0 1.0 1.0 1.2 1.16 1.4 1.16 1.2 1.16 0.06 1.0 0.26 0.25 1.2 1.16 1.15 1.0 4.7 1.5 1.0 1.0 1.0 0.16 1.0 1.0 0.26 1.0 0.26 O O t\o cn -vl U CO TCE mixture*: 0.11 NaCl:1.5, H20: 26.0 3,5-dichlorosalicylic acid: 0.16 tl 11 Glyoxal: 0.26 Glyoxylic Acid: 0.55 rpm TCX N.D. N.D. N.D. N.D. N.D. N.D. 8( 282 244 24f 21E 12: 189S -ND- 10 N.D. 104 213 *TCE Mixture: 50:50 Mixture of 1,1,1,2- & 1,1,2,2-Tetrachloroethane Table 3: The Formation of TCX from Synthetic Mixtures Heated at 160 for 72 hrs. i 73 .p- DOW 267749 OCR -15- OC/78-85 Several compounds other than those found in the process were tested. The tetrachloroethane mixture is a known contaminant in tne 941 Bldg, chloroacetylchloride that is used to make MCAA. 2,5-Dichlorosalicylic acid is a proposed intermediate from the reactions suggested below: c i 2c = c c i2 -- /'2 > Metal 0 Cl-^^-Cl OH(Na) Cl OH 0 k OH The decomposition of perc to yield phosgene is known 9 but the second reaction is only speculated. The glyoxylic acid is postulated to come from bis 2,4-D as shown below: which is a well known acetal hydrolysis. The glyoxal was included to show the general nature of the reaction whose mechanism is suggested later in this report. OCR -16- OC/78-85 D \y 267750 Several tentative conclusions can be made from the data in Table 3. (a) TCX can be made by several routes aNaOH and/or NaDCP is necessary for the formation of TCX. bis-2,4-D appears to be a key intermediate. flPerc is a source of TCX. (b) Significant amounts of TCX are probably made only in the reactor crude storage tank in the plant since strong base is required. Run 14 (Table 3) simulates the Na 2,4-D storage tank (V-501: 948 Bldg) and Run 1 (Table 3) simulates the product storage tank (V-602: 948 Bldg) in which no TCX was observed. In order to study the formation of TCX as a function of reaction parameters, a series of 2,4-D reactions were performed in the lab oratory. These reactions were run to evaluate the effect of different raw materials, perc, air/^* and iron on the amount and rate of TCX formation. The data are summarized in Table 4. In these experiments, 1.1 mol of DCP was treated wth 0.6 mol of caustic and the mixture was heated to 130C. The majority of the water- that was formed was boiled off. To this mixture, 0.5 mol of NaOH and 0.5 mol of MCAA were con-added during 1 hr at 130 and the crude product was heated an additional 1 hr. Water was continuously distilled out during the con add step. At this point the desired number of ampoules were charged with 5-10 gms of crude reaction mixture and were heated in oil baths at 130, 145, and/or 160 for 24, 48, and 72 hours. i 74 P--7T OCR -17- O C / 78-85 O O V /267751 Tabic 4: The P a r a m e t e r s A f f e c t i n g the F o r m a t i o n of TCX in the S e m i - H y d r o u s Process for P r e p a r i n g 2 , 4-D l~y As. H sat kw Scy ah iN ta M ftAT- s c , . Tar 3r e a r s i i A * - ^ > j u TCX rA T i * D C? m caa 41 2NiwHOitf *C *r Ew rttc. Ix u d* co ItkMl. ;* s <*- a 7i-* C tfA yM C N T l X CC`'.-lie 3I 1! i ^ i DduJ 'y 33x 3 io w ' 1 <?/8& a s 0.4 t . z . a / / a 130 r / .J . rt J 3 JA A . _ w r 23 8 110 ,/ to w U S ' je o J / 3 0 13 21 r/t/ / # -- /*' 6 u;/fr JA n /*7 13X 230 i xy X b tJ B O 'V Hi sue. / * A), uj. /JO /O C rfD T * A3> re- 20 0*4 it SI 111 IS tot / o y ec.iil-S-lsS 2 J fioCZtMl'0. 1. / i a,m *s y CCV 2)0 risutr0.2 3/ . - cc.ii7-s-u: 2/ f S u e . t&escjiTo . z /.X Z.irhS'-iii -2)0 AO2U f j e / r cyehss-ai ^CuJ Wtu,4 ts*4**r r 7) / 3k 0y6 W^O-l*J X mb u Jtlie fiASttf /wio. 3/ - y OC-Ulo-IA 'Diu) wetta 0.*i / A C<Cy.vfo-<4 D o J `it8Seat 0-t IX. Ait* __ /hi- / / % fi/i. A,t. ' AiM, Ae- -- - A)4. i -- Ae- _ 30 Sl/H s __ M3\ /60 Ms xi MD Sx IX 0 -- J? V It /j I -- MS .loo i__ -- M>4i ax S'? S3 90 n % lit * - ,/ o Me. w 91 -- Sje>x' /f ici/cM.r cTutLUaiCj-ti-astiOos/OJ; cieneJceSeS.Sa - 'xx 6' / o M.e. T&- SJl - o r * *At*e.i. lutes Mser/'ti- 1 - n /&3 r?D . . . -- . .......... - SYlta 'HitA ulirii eut* A 7 ten ST7S' - /o - /X, ti Si /Jo - tbo C co j i /Jo* 740 S 1 /X, 21 + f-fiFne. no /es ti.er/*'i. - 0/ 211 4H + iM/riAL #irot. imisxecs Mirao hol% Mo-HiOj i P.-sez ui-l , hj.iiJi- "fisr tZiUTZ.i'J tf* J o n r 2tl 700 /OOZ^ 2W W A ss ot I-X. y' IO C .erO -0 7 j |c2 o ,J 3s .* J 9 8(' xW < j 0/ M i X -------------- So-73c c - D o J Cu>c*f r ? _ .J 0/ ! / * i 1i !! !i i 1 ! ; A,e. u "k 300 / J O /o A e. - 3-00 /JO /bo ' 9 m 43 t i Ss~s~ 7c& // 77J 37 / l i f - X-iiFrte. /AOFes Meer,-s . 3br 93ih A,al - A S - 1*0 6c T / 17J J it i 1 1 OCR -18- OC/78-85 92 MOQ The data in Table 4 show a number of interesting points. It must be emphasized that in view of the fact that these data are based on parts-per-million chemistry, the precision must be subject to some error. The data are reproducible to about 20%. (A) Effect of Dichloroacetic Acid in MCAA Runs 1,4,8, and 9 show that DCAA affects TCX formation presumably through the formation and decomposition of bis-2,4-D. Run 9 was especially interesting because not only was TCX made but bis 2,4-D was observed as shown below: to ppm TCX % Bis Hrs at 160C 24 12 0.4 48 41 0.3 72 56 0.2 These data indicate there is another route to bis-2,4-D besides reaction of NaDCP with dichloroacetic acid since the monochloroacetic acid used in this experiment showed no detectable DCAA (< 100 ppm) by liquid chromatography. (B) Effect of Perchloroethylene Whereas Perc showed a significant increase in TCX levels in the am poule experiments, its effect in the reactions is uncertain. In most of the runs, the reactions contaminated with perc showed more TCX at 130C and similar amounts at 160C. K. E. First, Process Engineering, 4 ran a computer analysis of, tlhe data which gave the best fit when calculated as two consecutive first order reactions with an Ea = ^30 Kcal/mol. In the presence of perc the Ea dropped to M.5 Kcal/mol. These data suggest that there is another 75 P- OCR -19- O C / 7 8 - 85, D W o67753 more facile route to TCX from perc whose effect is masked at the higher temperatures by a primary route (compare Table 4, Runs 1 & 2 and Runs 3 & 4). (C) Effect of Atmosphere in the Reactor Nitrogen blanketing appears to lower the TCX level to ^2/3 that ob served in air. It is not certain that this result is significant and since the magnitude of the drop was low, further work was not warranted. Since C02 is known to be present1^* in the vapor space throughout the 2,4-D process, a run was made to determine its effect on TCX formation. Carbon dioxide comes into the process from the caustic and from perc decomposition. The reaction was run as described earlier with 20 mol% of NaHCO^ added before the boil down step and then blanketing the reaction mass with C02 during the post reaction heating step. Since this is a high temperature, nearly anhydrous process, the following reaction was thought possible which could eventually lead to TCX (See Table 3, Runs 15 and 16) . ONa + CO 2 NaOH ONa Cl Comparing Runs 1 and 10 (Table 4) show that NaHCO^ and/or C02 do not measurably affect the formation of TCX. VZLL92 A\Oq OCR -20- OC/78-85 (D) Changing NaDCP/NaMCAA Ratio Comparing Runs 1 and 7 (Table 4) show that lowering the NaDCP/NaMCAA ratios significantly reduces the TCX levels as was also predicted from the data in Table 3. The effect was studied further and is discussed later in this report. (E) Source of DCP A run was made using the old lower purity DCP from the 349 Bldg process and assure that 948 Bldg, high purity, DCP was not a source of TCX. Run 6 shows that TCX formation is not associated with the source of DCP. (F) Effect of Ferric Ion Runs 11-14, Table 4, show that Fe+3 clearly increases the rate of TCX formation which suggest that alkylation step(s) are involved. D. Humbert 7 and colleagues performed a material balance of Fe in the following scheme. / 76 9 'P OCR -21- O C / 7 8 - 85 DOW 267755 They found that most of the iron is entering the process in the caustic and that the reactor normally contains 20 ppm of Fe+3 . A recurring problem for the past several months was how to explain the fact that the plant observed TCX formation rates that were 5 times greater than the lab. The iron results explain at least part of this discrepancy and suggest that efforts aimed at removing iron from the process are desirable. Based upon the data presented up to this point, the mechanism shown in Figure 1 for TCX formation is suggested. This mechanism is consistent with the observations of the effect of caustic, the fact that TCX can be made from NaDCP and glyoxal or glyoxylic acid and the levels of bis-2,4-D. If this mechanism is correct, it is easy to see how perc can influence TCX formation when one considers the possible hydrolysis products using either NaOH or NaDCP. Some of these possibilities are shown in Figure 2 and many are known precursors to TCX. In addition to TCX, OCSX and "8-5" are known to be formed in measurable quantities during the reaction to 2,4-D. It was initially assumed that TCX is the probable precursor to OCSX and "8-5". However, a series of ampoule experiments in which TCX was mixed with various ratios of DCP, 2,4-D, NaOH, and perc showed none formed in detectable amounts. In addition, a reaction was run in the presence of 500 ppm of added TCX and gave normal levels of the impurities when the 500 ppm of added TCX is subtracted out. It is apparent that OCSX, "8-5", and TCX are probably made by independent routes and proposed mechanisms are shown in Figures 3 and 4. I Oovv 2 6 7 7 5 C OCR -22- Figure 1: Proposed Mechanism of TCX Formation OC/78-85 O H 3 Cl Cl C-CO~Na I2 H TCX 77 i p-%4 OCR -23- OC/78-85 0 qnM2 6 7 7 5 Figure 2: Some of the Possible Products of the Hydrolysis of Perchloroethylene with NaOH or NaDCP OH CC l 2=C\ Cl 0 ,4h 2o CHCl DCAA OH NaDCP. C1 (h-ocUc^ ^ XC1 2 OH 3/^ Cl OH c\ cr oh or c c i 2= c c i2 HO .HO s-c=c/ Cl^ 'Cl Cl 6oci 2 NaDCP^ V _ n/ ------- > / - c Cl 00C1. known impurity 0 CHCl~-C ~ 2 \H 1. Tautom ^ 2. H 20 O h S-C02Na h 2o CHC12-C-H 3 OH- V B * HO-" \ l nh 2,,uo \ [ o HCII-C02Na 00C1. 3 NaDCP> (Cl2<j)0)2c=c Cl 4 OH or 4 NaDCP ?? HC-CH O Cl \ H-ICI-CIH ^\ OOC1. OCR -24- OC/78-85 Figure 3: Proposed Mechanism for the Formation of OCSX By a Route That is Independent of TCX DOW 267758 NaOH 2 ring closures as shown in the TCX mechanism Cl Cl Cl Cl } 78 P" OCR -25- OC/78-35 . 0 OV/ 267759 Figure 4: Proposed Mechanism for the Formation of "8-5" By a Route that is Independent of TCX OCR -26- OC/78-85 s DOW 267760 The formation of OCSX uses an intermediate in the TCX mechanism as its starting material. The apparent fact that TCX is not an intermediate suggests that the more activated dihydroxy ketone is necessary for alkylation. The mechanism for "8-5" formation involves a major component in the Complex I and II mixture as a starting material. Recent analytical results^ have shown ^100-200 ppm of this mixture in the process. A proposed mechanism for the formation of Complex Acid I is shown in Figure 5. It must be emphasized that the mechanisms shown in Figures 3,4, and 5 are speculations and are only one of a number of possibilities. Much more work is necessary to prove their accuracy. (B) The Distribution of TCX in the Process As mentioned earlier, TCX appears to be formed only in the reactor section of the process. Each of the process hold tanks were tested using either synthetic mixtures and/or actual plant material under normal operating conditions and the results are shown below: V-201 and V-202: Reactor & crude reactor hold tank. Tables 3&4 summarize these results. V-403: Phenate Recycle Storage Tank. 60% NaDCP was heated for up to 7 days in presence and absence of air in presence and absence of steel at 90c and showed no detectable TCX. V-501: Na 2,4-D Storage Tank Mixtures of Na 2,4-D, NaCl & I^O were heated at 145-160 for 3 days in presence and absence of traces of perc and DCP: no detectable TCX was formed. 79 P DOV/ 267761 OCR -27- OC/ 7 8-8 5 Figure 5: Proposed Mechanism for the Formation of the Major Impurity in Complex I mixture (Compound IV) ; Bis + H20 > 2 DCP + 2e 6$ COMPLEX ACID I OCR -28- OC/78-85 V-602 Final Product Storage Tank: Tests run at 120-175 on wet and dry 2,4-D'*'''' showed that there is an initial slow build up of about 20 ppm of TCX during the first 48 hrs at >160 that stabilizes out. This suggests a small amount of an unstable unknown species that decomposes in acidic media. Molten 2,4-D can be treated for at least 100 hrs under the above conditions without affecting product performance. $oo co OT -J CD CO Since TCX is formed in the reactor and was found throughout 948 Bldg process, a laboratory study of the distribution was requested. In this experiment a reaction was carried out and the crude reaction mass was spiked with TCX. The purification process was then simu lated and the fate of the TCX was determined. Figure 6 summarizes the results of the extraction study performed at 90/atm. Figure 6: The Distribution of TCX in a Laboratory Simulation of the 948 Bldg 2,4-D Process REACTION (TCX not detected) MASS Pere 30% NaOH -----0.248g TCX added (1000 ppm) \' N/ . -DC:p EXTRilCTION Na 2,4-D Solution (TCX not detected) _\O n p:rc SOLIJTION I (0.240g TCX Recovered) 0.016g TCX Recovered) NaDCP SOLUTION f Distribution Ka I ^ Pere SOLUTION 80 (0.214g TC> Recovered) OCR -29- OC/ 7 8- 85 .. DOW 267763 These data show an excellent TCX recovery in which, using clean perc, all the detectable TCX is removed from the Na 2,4-D solution. The fact that the plant observes some TCX in the product suggests that some entrained perc is carried overhead so that the impurities [carried with it end in the product. The extraction of the perc/DCP mixture with caustic to recover the NaDCP also extracts about 5% of the TCX which is recycled to the reactor. This scheme was followed in a second experiment that was performed at 125/ 35 psig which more closely follows the plant conditions. Under these con ditions the aqueous solutions can be more concentrated in Na 2,4-D, NaCl, and DCP which could affect the distribution of TCX. r A high pressure, mechanically stirred glass reactor was constructed and was charged with partially neutralized material from V-301 (the DCP neutralizer). The molten material was extracted six times with perc at 125C and the molten 2,4-D was isolated. The TCX and "8-5" levels were monitored throughout and the analytical results for the products are summarized in Table 5 for the two runs. The data show that TCX is effectively removed at the more drastic condi tions. Table 5: The Analysis of the Product from the High Temperature Extraction of Crude Na 2,4-D with Clean Perchloroethylene TCX (ppm) Run init final 1 292 2 178 13 2 "8-5 " (ppm) init final 54 N.D. 52 25 Metals (ppm) Na Fe Ca Mg 84 5 40 13 91 3 50 16 It is interesting to note that "8-5" is not efficiently removed with perc extraction. Hence most of what is made goes out with the product. OCSX was not detected in these samples. OCR -30- OC/78-85 In summary, TCX is formed in the reactor and most of it remains in the recirculating perchloroethylene system. A certain amount of TCX spills into the product and the magnitude is directly related to the level in the perc. TCX is easily separated by distillation in the perc still so by increasing the rate of distillation one would achieve a lower steady state concentration of TCX in the perc. o ?o 267764 Assuming a rate of TCX formation of 0.5-1.0 #/hr in the plant (based upon analyses of plant samples) K. E. First4 has taken distribution coefficient data and has modeled the process in terms of TCX content in various streams as a function of % of perc distilled in the perc still. To date, there is not enough in-plant data to verify this model. (C) Reduction of TCX by Chemical Means Two approaches for chemical reduction of TCX were studied. These were methods for inhibiting its formation during the reaction step and methods for reducing it from process streams. The following approaches were studied and each will be discussed in detail. 1. Changing NaDCP/NaMCAA ratio. 2. Post reaction neutralization. 3. Effect of DCAA. 4. Bleaching Na 2,4-D solutions. (1) Changing NaDCP/NaMCAA ratio Excess caustic or NaDCP was shown earlier to have a significant qualitative effect on the increased production of TCX. The 2,4-D reaction as developed by H. Brust 12 used a ratio of NaDCP/NaMCAA = 1.2. As described earlier in the discussion, at this ratio the| product contains theoretically <\/6.3 mole % excess alkalinity as NaDCP as shown on the following page. 81 p-fi OCR -31- OC/78-85 DOW 267765' Stq Mat11 0.5 mol DCP 0.6 mol NaDCP ---- > 0.5 ml NaMCAA Product 0.5 mol DCP 0.1 ml NaDCP 0.5 mol Na 2,4-D 0.5 mol NaCl Dhingra13 and Fear14 evaluated the effect of this ratio on the yield and kinetics of this reaction and concluded: (a) lower ratios (below 1.2 lead to lower yields based on MCAA and, (b) low HjO in solution gives higher 2,4-D yields. The effect of changing the NaDCP/NaMCAA was restudied in order to determine if the lower yield from MCAA could be justified by reduced TCX formation due to less excess caustic. The runs were carried out by using a constant amount of NaMCAA and varying the amounts of NaDCP which was done by adding differing amounts of caustic in the initial boil down step. Tables 6 and 7 and Figure 7 summarizes the results of this study. These data show that lowering the ratio from 1.2 to 1.1 results in a two fold reduction of TCX along with a 1% loss in MCAA yield. This is re^ able and was tested in the plant. After three weeks of operation, they did not note a measurable reduction for reasons that are not understood at this time. Any further reduction is not practical since the yield rapidly approaches that of the old plant thereby losing much of the advantage of the 948 Bldg process. It is inte resting to note that at ratios below 1.0 there is still a measurable amount of NaDCP which explains why formation of TCX is still observed. All of the data to date indicate that formation of TCX cannot be limited to much less than 60 ppm (in the standard 72 hr/160C ampoule test). OCR -32- OC/78-85 TABLE 6 EFFECT OF NaDCP/NaMCAA RATIO ON TCX FORMATION & YIELD ON MCAA RUN REF DCP 1 OC 417-5-143 0.50 2 OC 417-5-144 0.55 3 OC 417-5-146 0.60 4 OC 640-1-1 0.61 Moles NaDCP 0.60 NaMCAA 0.50 0.55 0.50 0.50 0.50 0.49 0.50 Mol Ratio NaDCP NaMCAA 1.2 PPm HOG 951 1.1 1514 1.0 3808 0.98 3606 MCAA YIELD 96.4% 94.2 86.8 86.2 CRUDE ANALYSIS Na 2,4 -D DCP NaDCP 45.4 31.2 8.9 45.5 38.0 6.0 41.3 40.3 2.2 43.0 41.7 2.5 NaCl 1 0 .0 7.4 7.4 7.9 <3 00 fo 9393 MOq ^ OCR -33- , 91193 thOQ OC/78-85 TABLE 7 NaDCP THE RATE OF TCX FORMATION AS A FUNCTION OF TIME, TEMPERATURE & NaMCAA RATIO RUN REF TEMP 1 417-5-143 130 145 160 2 417-5-144 130 145 160 18 HRS 37 29 18 0 27 42 TCX/OCSX 42 HRS 66 HRS 34 86 42 19 39 74 19 70 118/107 66 33 64 93 3 417-5-146 130 <10 <10 15 145 15 20 41 160 29 47 55 4 640-1 130 17 145 23 160 36 28 22 31 34 61 65 114 HRS 53 218 240 114 Hrs 52 93 142 90 Hrs 30 59 67 90 Hrs 18 52 77 t * OCR -36- OC/78-85 in the laboratory even by very slowing adding the aqueous acid. O o 267770 (d) Control of the acid addition would be difficult and, in addition, the plant would have to run a slightly smaller batch size in order to provide room for the acid. Further work on this concept is not warranted at this time. (3) DCAA Content of MCAA The effect of DCAA was discussed earlier. (4) Bleaching Na 2,4-D Solutions The major chemical difference between 948 Bldg and 489 Bldg techniques for processing crude 2,4-D is that 489 Bldg uses a bleach step in which residual DCP is oxidized out of the Na 2,4-D solution with 8-12% NaOCl at pH 10.5/100C. Several experiments were performed to determine if bleach would effect the impurity levels and/or improve the ability of the plant to consistently produce 2,4-D that passes the formulation dilution test. The results of these experiments are shown in Table 8 . The results of the first experiment appeared very encouraging since the dilution test solids, the TCX and the "8-5" were all significantly reduced. The second series of runs showed a consistent improvement in dilution test solids when near or out of spec, but not much effect 4 on the impurities. A capital estimate performed by K. E. First showed the cost of implementing a bleach step in 2,4-D to be $500 M which immediately eliminated any further interest in this project in view of the marginal benefits. It is assumed, posthumouslythat the main effect of the bleach was to reduce residual DCP which ;is a known contributor to poor dilution test results. 84 V OCR -37- O C / 7 8 - 85 Table 8 : The Treatment of Na 2,4-D Solutions with Bleach Na 2,4-D Solution Source PH Mol Ratio NaOCl Na 2,4-D V-501 (4-7-78) -948 Bldg. starting material 10.5 0 .1 ml solids DMA-4 F-40 0.18 0.02 Impurities (ppm) TCX OCSX "8-1 58 N.D. 127 25 N.D. N.D. V-501 (6-5-78) 948 Bldg starting material 10.5 0 .1 10.5 0.2 10.5 0.05 10.5 0 .1 5 0 .1 o % ro 05 -a 0.01 TR TR -- -- ------ TR TR TR -- -- ---- 80 85 85 104 94 72 N.D. It fl II II II 31 10 26 23 16 ,7 / OCR -38- OC/7 8-8 5 DOW 267772 (D) Removal of TCX by Physical Means To date, a stage has been reached in 948 Bldg, where the 2,4-D plant can be run to consistently produce in-spec material and has demonstrated an ability to run at near capacity of 1 0 0 - 1 2 0 M#/day. With the perc still running at ^15 GPM, the 2,4-D product contains ^20 ppm of TCX and ^20 ppm of "8-5". A numberof attempts have been made to see if the levels of TCX can be reduced by physical means that include: 1. Carbon treatment of`the recycling perc and the molten 2,4-D. 2. Removing perc from the reaction step. 3. Improved washing of the molten 2,4-D. 4. Recrystallization of Na 2,4-D and recrystallization of 2,4-D acid. The details of each alternative are discussed below. In all cases, these experiments were short term, range finding efforts and not comprehensive. Further work is justified only if the levels of im purities presently found in the product prove unacceptable in the future from a toxicity, environmental, or performance standpoint. (1) Removal of TCX with Activated Carbon The removal of TCX with activated carbon from perchloroethylene from V-402 ("clean" perc storage tank) was evaluated by a standard isotherm method. A 100 ml portion of perc was treated with ground Pittsburg SGL carbon at 70C/72 hrs. The data, summarized in Table 9, were evaluated by a known technique that is summarized below. A plot was made of X/M vs. C on log-log paper as shown in Figure 5. By extrapoling C to incoming TCX concentration, the corresponding X/M value gives the amount of impurity absorbed per unit weight of carbon when that carbon is in equilibrium with the incoming concen tration and represents the ultimate capacity of the carbon. The 85 tol V OCR -39- O C / 7 8 - 85 o % ro cn -a -a -3 theoretical volume of liquid to be completely freed by TCX/gram of carbon is calculated from the following equation: V.Co Co Co V VCq = theoretical volumn to be treated = capacity/gm at incoming concentration Co V = volume of liquid used in test Co = incoming concentration For this experiment: VCo = X ^0 = 1^6 ml/g of Carbon or 19.9 gal of perc/# C or 265 # perc/# C Carbon loading at 241 ppm of TCX in feed = 166 ml/g C x 1.6 g ml x .00024 64 mg/gc / $5 A DOW 267774 OCR -40- OC/78-85 Table 9: The Absorption Isotherm for TCX Removal from Perchloroethylene with Pittsburg SGL Activated Carbon Sample STG. Mat '1 1 .2 3 4 5 Impurity (ppm) TCX OCSX "8-9" (M) (C) Res id WTC/100ml ppm Pere TCX (X) ppm TCX Removed (X/M) ppm TCX removed gm Carbon 241 N.D. 159 II 108 11 57 II 40 II 29 11 19 17 17 0 0.25 0.50 1.00 1.50 2.00 241 159 108 57 40 29 0. 82 133 184 201 212 _ __ 328 266 184 134 106 i 86 OCR O -41- OC/78-85 OCR -42- OC/78-85 This is a low loading level and at the present rates of TCX pro duction in the plant, about 15-20 # of carbon per hour would be required to remove the impurity. The data shown in Table 9 indicate that carbon does not remove "8-5" from perc. The shape of the isotherm curve indicates that more than one species is being ab sorbed onto the carbon. No effort was made to determine the identity of that compound although it was probably perc. O O ro <x> <E vl -a CD A run was made to remove impurities from wet molten 2,4-D. Since molten 2,-D is so difficult to handle at atmospheric pressure, pressurized system was built using a capillary feeder for controlling the continuous flow of molten 2,4-D at 100/25 psig onto a vertical 2' x 0.5" column that was maintained liquid full. The flow rate 2 was 10 ml/min down the column for a mass flux of *\1.0 gpm/ft. The results are summarized in Tabl 10. Table 10: The Treatment of Molten 2,4-D with Activated Carbon Sample # Temp (C) Pressure psig Approx feed rate ml/hr Approx. Prod, cut volume (ml) Product Analysis (ppm) TCX OCSX "8-5" F-40* solids (ml) Feed 1 2 3 4 5 112 110 2 110 110 110 110 33 2 33 33 33 33 300 100 600 375 600 375 600 376 600 375 42 N.D. 32 ---- -- 14 N.D. 29 28 N.D. 38 32 N.D. 42 30 N.D. 36 .005 .04 .005 *Formulation contains 2% Versene & 0% PG 4000 These data are fairly crude in that the column was shorter than is desirable >r optimum column work (2' vs the recommended 5-6' 18 ) and the flow through the column was a little faster than the more desirable 0.5 gpm/ft 2 recommended. 87 OCR -43- OC/78-85 267777 The data show that TCX loading is rather low. Breakthrough occurred ^ after 900 ml molten D (^935g pure 2,4-D) was treated. From the O analyses, the 59.8 gm of Pittsburg SGL 8 x 30 granular carbon charged to the column absorbed 281 mg of TCX for a loading of *^0.5 mg/g of carbon which is quite low. Again, "8-5" was not absorbed by the carbon. Based upon these preliminary experiments, carbon absorption of these impurities is not an attractive puri fication technique. (2) Remove Perchloroethylene From the Reactor Since perc is known to give rise to TCX, several methods were examined to eliminate its recycle to the reactor. At present, since perc is soluble to 1.2% in the recycle NaDCP solution, there are about 350# returning to the reactor in each batch. Since a continuous percphenate phase separation is performed in V-401 just before recycle to the reactor, a brief study was undertaken to evaluate the phase separation to determine the time required for complete layer separation and the solubility of perc in NaDCP solution. A solution of 62% NaDCP and 2% NaOH in water was slurried with an excess of perc and vigorously stirred for 30 min at 75. The stirring was stopped and the phenate layer was analyzed for % perc as a function of time. The results are summarized in Table 11. These data show that the solubility of perc in NaDCP is 1.30.1% at 75C and that layer separation is complete within 15 minutes. The residence time in V-401 is 'MO minutes so that with proper operation, no layer separation problems should result. The question was also asked if the perc was entrained in the NaDCP solution as an emulsion or was it in solution? A sample was centrifuged at 60 (minimum temperature) for 30 m|n at ^4000 rpm and 1.2% perc was found in the enate. It is concluded that the present equipment gives optimum layer separation and the 1.2% of perc is soluble. \0 $ 1 * cv>- DOW 267778 OCR -44- O C / 7 8 - 85 Table 11: The Solubility and Separation Rate of Perchloroethylene and 62% NaDCP in Water at 75C TIME* 5 min 15 " 30 " 1 hr 2" '4 " 6" 24 " % Perc in Phenate Layer 1.9 1.3 1.4 1.5 1.4 1.4 1.2 1.3 *Time zero is when the vigorous stirring of the two layers is stopped. 88 Q.iol OCR -45- OC/78-85 DOW 26777!) Several attempts were made to strip perc from the reaction mass and from the NaDCP solution. The plant tried to strip perc from the reaction mass during the normal boil down step. Whereas normal operation involves returning the distilled organic layer to the reactor, in this experiment the recovered organic layer was discarded. Whereas normal boil down returns 350# of DCP to the reactor, in this experiment even after 4000# of DCP was distilled off, there was still detectable perc in the distillate. As a result, this approach was judged impractical. Two attempts were made to distill the 1.3% of perc from the 60% NaDCP-2% NaOH solution. The first involved a batch distillation from a standard solution from V-403. A total of 65.9 gms were distilled out and 73.8% of the perc was removed. The reaction was then carried out as usual and samples were heated for 24, 48, and 72 hrs and the results are summarized below. These data suggest that distillation of 75% of the perc does not reduce TCX formation (See Table 4, Runs 4&5). ppm TCX Temp 160 24 hrs 21 48 hr 72 hrs 75 145 The second approach taken for stripping out perc was using a falling film still. The results of two experiments are shown in Table 12. The results show that, as before, 70-80% of the perc is easily removed but that the last 20% is likely to be quite difficult. No more work is planned in this area until it can be better proven that removing perc offers any real advantage in reducing rates of TCX formation. (3) Improved Washing of 2,4-D Acid ^ Several experiments were run to see if improved washing would affect impurity levels in the 2,4-D product. A sample of 2,4-D product was taken from V-602 (948 Bldg) and was treated in the following ways: (a) A 250 gm sample of molten acid was washed 3 times with 200 ml of water per wash at 100C. OCR -46- O C / 7 8 - 85 TABLE 12 Preliminary Data for Falling Film Distillation of Perc From NaDCP Solution Still: 1" x 12" Tube DOW 267780 Oin Pressure Temp (col'm) " (feed) Feed Rate (ave.) Overhead Temp Feed analysis % Perc % DCP Wt. Charged (Feed) Product: Wt % Perc % DCP Overhead: Wt % Perc Run 1 120 10 ml/min 85-- > 90 1.0 51.5 725 gms 652 gms 0.3% 53.9% 46.7 gms 1.5% Run 2 ATM. 75 10 ml/min 90-- > 95 1.1 50.1 738 gms 645 gms 0.2% 61.6% 75.0 gms 3.3% 89 0 -(O'! OCR -47- OC/78-85 \8LL$ZhhOQ (b) Another sample was reacidified to pH 0.5 with cone HCl and rewashed 2 times with water at 100C. (c) A synthetic V-501 mixture was made up using Rhone Progil 2.4- D to determine if any impurities are made in the washing step. The synthetic mixture was acidified and washed and the 2.4- D was recovered for analysis. This experiment was repeated in the presence of 2750 ppm of added TCX. The results of these experiments are shown in Table 13. Based upon these experiments, it is shown that the impurities are not made or reduced by improved washing. If anything, they are slightly increased in the product due to the greater solubility of 2,4-D in the hot water or brine. (d) Recrystallization of Na 2,4-D and 2,4-D Acid An attempt was made to determine if the impurities could be removed by recrystallization of Na 2,4-D from water and 2,4-D acid from organic solvents. Na 2,4-D was recrystallized by taking 500 gms of material from V-501 and adding enough water (350 ml) to form a homo geneous solution at 100C. The solution was cooled and the precipi tated Na 2,4-D was filtered and washed with 5% brine. The Na 2,4-D was redissolved in hot water and the 2,4-D was isolated and analyzed. The results are shown in Table 14. In two separate experiments, 2,4-D from V-602 was recrystallized from perchloroethylene and ethylbenzene. A weight ratio of 3 parts solvent to 1 part 2,4-D was heated to boiling, the water contained in the molten 2,4-D was boiled out as an azetrope (the organic distillate was returned) and the solution was cooled. The 2,4-D was recovered by filtration and the solvent removed by heating in a vacuum oven at ^60C. The results are summarized in Table 14. DOW 267782 OCR -48- OC/78-8 5 Table 13: The Effect of Improved Washing of Molten 2,4-D on Levels of Impurities Material V-602 Starting Mat'l Wash three times Reacidify, wash two times Synthetic V-501 Na 2,4-D: NaCl DCP 21.5% 7.0% 0.2% After Acidif. & Wash Repeat the synthetic V-501 spiked with 2750 ppm TCX Impurities TCX OCSX "8-5" 56 N.D. 73 N.D. 96 34 58 60 68 N.D. N.D. N.D. After Acidif & wash 3150 N.D. 10 Note: Rhone Progil 2,4-D shows no detectable TCX, OCSX & "8-5" 90 D -w OCR -49- OC/78-85, POW 2G7783 Table 14: The Effect of Recrystallization of Na 2,4-D and 2,4-D Acid on Impurity Levels Treatment Recrystallize Na 2,4-D Before* Recrystallize 2,4-D From Perc Recrystallize 2,4-D From Ethylbenzene After Before After Before After Impurity Level (ppm) F-40** Dilution TCX OCSX "8-5" Test 58 -- 127 0.18 81 53 52 N.D. 54 0.3+ 46 *** 1 N.D. 52 N.D. N.D. 46 * N.D. N.D. N.D. Commen Product Highly lored 91% Ree very of 2,4-D 87% Ree ver{ f 2, 4-D *The impurities analysis was performed on a sample of 2,4-D acid isolated from Na 2,4-D without any extra treatment. **20:1 dilution in 1000 ppm hard water, formulation contained 2% Versene and no P-4000.* ***The recrystallized product behaved like the high purity Rhone Progil 2,4-D which is difficult to formulate as was mentioned earlier in this report. OCR -50- OC/78-85 DOW 267784 Based upon these results, recrystallization of Na 2,4-D is not a good option since the TCX tends to concentrate in the product and the product was highly colored. It was also observed that the iron level in the product increased from 22 ppm to 91 ppm which could explain the off-color. Recrystallization from an organic solvent clearly improves the product quality although it would be difficult and expensive to implement in the plant. No further work is planned in this area unless it is determined that extremely low levels of impurities are necessary from a toxicity or an environmental standpoint. GENERAL CONCLUSIONS TCX and other non-acidic impurities are formed chiefly in the reaction step of the 2,4-D by several routes. Additional caustic, perchloroethylene, elevated temperatures and iron all promote their formation. A number of attempts to chemically and physically remove these species have met with limited success. It was found that th impurities are concentrated the recirculating perchloroethylene system and that by increasing the capacity of the clean up distillation column from 0.75 to 15 gpm, the levels of impurities in the process do not cause serious operating or quality problems. SAFETY & ECOLOGY 2 ,4-Dichlorophenol, 50% NaOH, and chloroacetic acid are highly toxic and corrosive raw materials. When handling, the protective clothing included lab coat, rubber gloves and goggles, and when possibly, all operations were performed in a fume hood. A number of operations were carried out at elevated pressure which required the use of a face shield and secondary shielding in the hood. All waste samples and solutions were sent to the burner for disposal. 91 D -3 OCR -51- OC/78-85 EXPERIMENTAL o ro The Preparation and Workup of 2f4-D ctd The following is a general description of the procedure used to ^ prepare and isolate 2,4-D when simulating 948 Bldg. A 1 liter CH round bottom flask equipped with a bottom drain, two dropping funnels, a mechanical stirrer, a thermometer, and a distillation head was charged with 179g (1.1 moles) of 2,4-DCP and 48 g (0.6 moles ) of 50% N2aOH. The flask was heated with a heating mantel attached to an I R Thermowatch controller. The reactor contents were heated while stirring and a DCP-water azetrope was distilled out. The distillation was continued until enough water was removed so that a temperature of 130C could be achieved. Normally 9-10 ml of I^O and 2-3 ml of DCP were removed. The DCP was returned to the pot. Then 47.3g (0.5 moles) of melted MCAA and 40.Og (0.5 moles) of 50% caustic were con-added from the two dropping funnels durir;g 50-60 minutes at 130C. The rates of addition were carefully controlled so that neither added reactant was signi ficantly in excess of the other. Water and DCP continuously distilled out during the con-add and the DCP was returned to the reactor. After the addition was complete, the reaction was heated an additional 60 minutes. About 43-47 gms of i^O was recovered in the con-add step. At the end of the post reaction samples of the viscous crude reaction melt were taken into ampoules,'if desired. The work up procedure for isolating the 2,4-D is as follows: (The amounts used assumes no samples were taken after reaction). Tbjk reaction mass was diluted with about 500 ml of water, heated to boiling to ensure complete dissolution and the pH was adjusted to 5.20.2 with about 12 ml of cone HC1. The solution was then extracted with six 150 ml portions of perc at a temperature of >90C to remove the DCP. Occasionally 50-100 ml of additional H 2O was necessary to keep all of the SOlids dissolved. The extracted Na 2,4-D solution was heated OCR -52- OC/78-8 5 DOW 267786 to boiling and any traces of perc were distilled off. The pH of the solution was then lowered to 0.5-0.7 with about 55 ml of cone HCl added rapidly and the molten 2,4-D layer was separated and drained into a beaker. The resulting brine was discarded. The 2,4-D was reslurried in 250 ml of hot distilled H20 in the pot and washed in this manner two times. The final pH of the aqueous layer was 2.7-2.9. The 2,4-D was recovered and dried overnight at. ambient temperatures. The analyses of product and intermediate streams were performed by personnel in the 948 Bldg quality control laboratory. The analyses for TCX, OCSX, and "8-5" were performed as described earlier^. The Extraction of DCP and Impurities from Na 2,4-D Solution Under Pressure A glass apparatus was designed and fabricated from heavy wall glass pipe as shown schematically in Figure 9 . The main pot was equipped with a mechanical stirrer and was loaded with 700g of material from V-301 (948 Bldg). The reactor was sealed and heated to 125C. The perchloroethylene was added in 649 gm increments from the pressurized shot tank. The perc layer was drained into the bottom receiver where it was cooled before draining into a bottle. All analyses on the perc and Na 2,4-D layers were performed in the 948 Bldg Q. C. Laboratory. The Carbon Clean up of Molten 2,4-D A glass pressure apparatus was assembled in which molten 2,4-D yas pumped onto a carbon column (downflow). The apparatus is shown^ schematically in Figure 10. The flow of 2,4-D was controlled by controlling the pressure drop across a capillary tube. To handle / OCR -53- OC/78-85 PRV 50# pOV* 2 6 7 7 8 7 Figure 9: h A Schematic Drawing of the apparatus used for Extracting DCP from Na 2,4-D with Perchloroethylene. OCR -54- OC/78- Figure 10: A Schematic Drawing of the apparatus used to treat Molten 2,4-D with activated carbon r\ OOVV 2 6 7 7 8 8 c L OCR -55 OC/78-85 molten 2,4-D reliably and effectively, required at least 10 psig/ ,vl00-105oC to avoid flashing and freezing problems. The 2,4-D and a slight excess of water were placed in the feed tank and heated to 110C. When the entire contents were melted the valve on the bottom of the column was closed, the column was filled with molten acid to 1" above the top of the carbon, and the system stood for 60 min. The pressure drop across the capillary was adjusted to 7 psi (vlO ml/min flow) and the valve on the bottom of the column was adjusted so as to maintain the liquid level above the carbon bed. The results are shown in Table 10. &2LL3Z th Q / A OCR -56- OC/78-85 ACKNOWLEDGEMENTS The authors wish to thank D. Humbert, R. MacLachlan, T. Evans, G. Jewett, P. Schloemann, and their colleagues for their analytical support. The assistance of K. First and his process modeling efforts are also acknowledged. D o 267790 94 D -ut OCR -57- OC/78-85 pow 267791 REFERENCES R. McLachlan, D. Humbert, G. Kallos, AL 7800101, to be issued. ^1 2. P. Keller, et. al., HET K-2372-(18) February 9, 1978. ro -vl D. Humbert et. al., unpublished results. S 4. Unpublished results to date. h-4 5. Beilstein I, 17, 354. 6. Organic Synthesis Coll, Vol 1, Pg. 552. 7. Beilstein 19 (3) 222, Merck Index 9, 4040. 8 . Morrison & Boyd Organic Chemistry, Allan and Bacon, Boston (1963) Pg. 684. 9. (a) S. MacLean, AL-75-20021. (b) V. Stevens, Communication, Inorganic Chem. TS&D. 10. D. Humbert, unpublished results. 11. K. L. Krmel, LR 78-74, 8-15-78. 12. H. F. Brust, OC 0730013-2, 2-26-73. 13. Y. R. Dhingra OCR Lab Book OC 327-5, p. 22-67, 1972. 14. D. L. Fern, PE 73-10, 4-16-73. 15. F. G. Aerstin, private discussion. 16. K. L. Krmel, OC 78-17 LR, 39-78. 17 . S. Siegel, Results to be published. 18 . Absorption Handbook: Issued by Activated Carbon Div., Calgon Corp., Pittsburg, PA. LABORATORY NOTEBOOK REFERENCES: R. F. Arnold OC 417, Pg. 109-150, OC 640 Pg. 1-73. K. L. Krmel OC 559, Pg. 95-131. DOW 267792 DISTRIBUTION 1710 Building *A. E . Young A. J . Vogel P. W. Owen S. Siegel *T. W. McGee *S. D . McGregor *S. W. Tobey 948 Building R. Young E. Hoy W. Fothergill *A. Lueck 834 Building *D. S. Morehouse *D. T. Buzzelli D. Woodward W. Haberstroh *E. Parsey *J. Maddox 574 Building *E. V . Luoma *W. Westover D. Humbert *D. MacLachlan *G. Jewett *C. Mendoza B. Dersnah, 489 2020 Building *M. E. Pruitt 1803 Building L. L. Smith S. J. Gorzinski D . Kociba *B. Schwetz 9001 Building* *R. D. Moss J . King *K. Seymour *C. S. Williams Bahia, Brazil T . L . Kennedy F. G. Aerstin *P. Schloemann R. Manfrom 566 Building *R. R. Bumb *M. J. Mintz 633 Building *J. Bleiweiss K. E. First j W . Ward. 9008 i 95 STATE OF FLORIDA COUNTY OF ORANGE ) ) SS: ) IN THE CIRCUIT COURT FOR ORANGE COUNTY, FLORIDA KIMBERLY MOYER, as Personal ) Representative of the Estate of) ROBERT W. MOYER, II, Deceased, ) Plaintiffs, vs. ) )Case No Cl 89-8657 )Div: 32 Thompson ) ) DOW CHEMICAL COMPANY, et a l ., ) Defendants. ) ) FORM LASER BOND A PENGAD/INDY 1-800-631-6989 The deposition upon oral examination of KARL KRUMEL, Ph.D., a deponent produced and sworn to before me, Aprille Rigsbee Lucas, RPR, a Notary Public at large in and for the State of Indiana, taken on behalf of the Plaintiff at DowElanco, 9001 Purdue Road, Quad III, Indianapolis, Marion County, Indiana, on March 10, 1993, at 9:30 a.m., taken pursuant to the Florida Rules of Civil Procedure in the above-captioned matter. ASSOCIATED REPORTING, INC. TWO MARKET SQUARE CENTER 251 EAST OHIO STREET SUITE 940 INDIANAPOLIS, INDIANA 46204 | 96 9 -p -2' A P PEARANCES FOR THE PLAINTIFF: Richard D. Schuler, Esq. Schuler, Wilkerson, Halvorson & Williams, P.A. 1615 Forum Place West Palm Beach, FL 33401 FOR THE DEFENDANT DOW CHEMICAL: Alan F. Wagner, Esq. Carlton, Fields, Ward, Emmanuel, Smith & Cutler, P.A. First Florida Bank Building P.O. Drawer 190 Tallahassee, FL 32302 FOR THE DEFENDANT HELENA CHEMICAL: H. Roger Lutz, Esq. Lutz, Webb, Bobo & Baitty, P.A. Two North Tamiami Trail Sarasota, FL 34236 INDEX DIRECT EXAMINATION, Questions by Mr. Schuler PAGE 4 FORM LASER BOND A PENGAD/INDV 1-800-631-6989 J 97 EXHIBITS Plaintiff's Exhibit 1 - R&D Report dated 10-27-78 Plaintiff's Exhibit 2 - R&D Report dated 10-27-78 Plaintiff's Exhibit 3 - "Analytical Method" dated 12-16 -79 Plaintiff's Exhibit 4 - "Analytical Method" dated 2-18- 80 Plaintiff's Exhibit 5 - Affidavit of Lawrence G. Silverstein Pag 38 45 75 77 109 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 i 98 4 KARL KRUMEL, P h .D ., having been first duly sworn to tell the truth, the whole truth and nothing but the truth, relating to said matter, was examined and testified as follows: DIRECT EXAMINATION, QUESTIONS BY MR. SCHULER: Q Would you state your name, please. A My name is Karl Krumel. Q What is your professional address? A My professional address is 1710 Building, the Dow Chemical Company, Midland, Michigan, 48674. Q Are you employed? A Yes, I am. Q By whom are you employed? A By the Dow Chemical Company. Q What is your position with Dow Chemical Company? A My title is associate scientist. Q As associate scientist, just in general, now, what are your duties and responsibilities? A Well, the associate scientist is a generic ^itle for people in research who have reached a certain level in the company. It doesn't in any way q q FORM LASER BOND A PENGAD/INDV 1-000-631*6989 5 define the job. Tell me a little bit about the structure, to give me some understanding of how the research department is organized. And if you would, could you also tell me, and you don't have to be completely thorough about this, but just tell me in general what the structure is and the various positions that are within the research department. MR. WAGNER: I object as to breadth, compound. You mean today? MR. SCHULER: Right. A Are you asking me about the structure of research within the total Dow or within the structure within my department, which is a small part of the overall? Q Let's start with your department. A The department I'm in is agricultural chemicals process research. This department, the name has changed several times over the years that I've been in it, but that's the current name. Our structure, we have a technical director. Wj have two research managers, and these research managers s then control groups of on the order of 30 * 6 technical professionals each. I happen to be in one of those groups as a senior scientist in the department. Q So there are roughly 62 or 65 people in the department? A In the current department, yes. Q Now, to expand that to the, to give me some idea of how the department of agricultural chemicals process research fits into the overall scheme of things, what other departments are there, other than that, if you know? A Oh, goodness, there are a number of departments located in Midland that support other areas of Dow R&D, you know, of the Dow Chemical Company, including pharmaceuticals, analytical, polymers, resins. Q You're the only department that has to do with agricultural chemicals in -- A In the Midland, area, yes. MR. WAGNER: You meant research, of course? j MR. SCHULER: Right. Q Are there other research departments that deal 101 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 7 with agricultural chemicals in other areas of the country? A Well, we have small groups that will support the production of ag chemicals in other locations other than Midland, yes. Q That are right within the factories, themselves, you mean? A They're within the plant area, themselves, yes. Q Who is the technical director presently? A The gentleman's name is Jim Love, L-O-V-E. Q How long has he been technical director at the agricultural chemicals process research department? A He's been there since 1988. Q Do you know who his predecessor was? A His predecessor was, yes, the gentleman's name is Alex Vogel, V-O-G-E-L. Q How long was he there, if you know? A I can't give you the exact time frame. It's on the order of four years or so. Q Would that position of technical director op the agricultural chemicals process research department have been in existence back in 1975, let's say? 102 FORM LASER BOND A PENGAD/INDY 1-000-631-6989 8 A In '75, the organization was pretty much the same. The name of the department at that point was different. Q What was the name at that time? A Organic chemicals research. Q And who was the head of the department back in 1975, if you recall? A 1975, the technical director was a gentleman named Don Morehouse. Q Were there also residents or research managers back in 1975? A Yes. Q Do you recall who the research managers were in 1975? A Well, I certainly recall my research manager at that point, which was Michael Mintz, M-I-N-T-Z . Q Is Mr. Morehouse still with the company? A No. He retired a number of years ago. Q Is he still in Midland or has he moved away? A He moved -- well, where he is today, I really don't know. He moved away to Florida, but 've lost track of him. Q How about Mr. Mintz, it is he still with the 103 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 9 company? A Yes, he still works for Dow. Q And what is his position now, the same? A Oh, no, no, no. He's off in another area. I can't tell you exactly what his job is today. Q What is your educational background, Dr. Krumel? A In 1960, I received a Bachelor of Arts degree at Grinnell College, Grinnell, Iowa; in 1962, a Master in chemistry; in 1962, a Master of Arts in chemistry from DePauw University, D-E-P-A-U-W, in Greencastle; and then in 1965, Ph.D. in organic chemistry from Michigan State University. MR. LUTZ: What year was that? THE DEPONENT: 1965. Q After '65, what did you do? A I joined Dow in June of '65, right out of college. Q So you have worked for Dow basically all of your working professional life? A Correct, since graduation from college. Q Trace for me, if you would, the positions that you've held at Dow, starting from when you jbegan work there after graduating in 1965. A From the standpoint of the departments in which I 104 9 ' 130 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 10 worked? Q Yes, and the positions that you held. A In 1965, I joined Dow as a research chemist, which was an entry-level position for a Ph.D. at that time. I joined the benzene research lab, which was part of the laboratory organization at that time known as organic chemicals production research. In 1969, I moved over to an area of polymer chemistry, cellulose ether chemistry. The department I moved into was -- Q I'm sorry, that last statement was? A Cellulose, C-E-L-L-U-L-O-S-E, cellulose ether chemistry. Q Cellulose ether? A Yes. And the name of that departmentchanged a couple of times during the time I was there, but in 1975, I then moved into what was then known as organic chemicals research, and I've been in that department, or whatever that department was renamed, since that time. j Q So I can save myself some time here, prior to 1975, would it be fair to say that you did not 105 D'v3' FORM LASER BOND A PENGAD/INDY 1-800-631-6989 11 have any experience, within Dow Chemical Company, with the agricultural chemicals that the company was producing? A No, because I hired into benzene research, and benzene research, at that time, one of their responsibilities was supporting the agricultural chemicals business at that time. The period of time of '69 to '75, I had no involvement at all in ag. Q Go back to '65, from '65 to '69, then, did you have any involvement with polychlorinated phenols, * I think they're called, those type of chemicals? A The polychlorinated phenols, yes. In 1966, I began doing production research on dichlorophenol. Q Any others? A And, of course, by association, I was aware of what was going on in the other chlorinated phenol areas of research at that time. Q To be more specific, did you produce any documentation from 1965 through 1969 with regard to any research on 2,4-D or 2,4,5-T or Silv^x? A No. Q Did you work onany projects thatinvolved 2,4-D, 106 FORM LASER BOND A PENGAD/INDV 1-800-631-6989 12 2.4.5- T or Silvex during that time period? A No. Q Did you produce any research during that time period, from '65 through '69, that had to do with any of the component parts that went either 2,4-D, 2.4.5- T or Silvex? MR. WAGNER: I object as vague. A I'm sorry, would you rephrase the first part of your question? Q Between '65 and '69, did you produce any research at Dow Chemical Company on any of the component chemicals that went into making 2,4-D or 2,4,5-T or Silvex? A Yes, I did. Q And which chemicals arethose, as yourecall? A 2,4-dichlorophenol. Q And that is involved withthe production of 2,4-D? A Yes. Q Did you have one or were theremany projects involving 2,4-dichlorophenal? A It was one major project. j Q And without getting too technicalon me, can you tell me basically, in laymen's terms, what the 107 p.tfS FORM LASER BOND A PENGAD/INDY 1-800-631-6989 13 project involved? A The project involved trying to improve the isomer purity of the raw material, of the 2,4-dichlorophenol. Q And did you produce a paper or papers regarding that? A Yes. Q Was the process that you were involved with ultimately implemented commercially? A No. Q And do you recall the reason why? A Subsequently some additional work was done that found something better. Q And who did that additional work, do you recall? A Let's see, one of the gentlemen's names was W. David Watson. Dave Watson was the primary researcher. He would have been the primary researcher in that area. Q Was there a fellow by the name of Brust or -- A Harry Brust, yes. Q Was he also working in that area in the process for 2,4-dichlorophenol, or attempting to improve the isomer from which 2,4-D is made? 108 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 14 A By the time I got to know -- well, I didn't actually get to know Harry until I came back into organic chemicals research in 1975. I'm aware of what he did in years before that, but only because I knew -- I got to know him when I came back into the department. And he was not involved in the dichlorophenol part, as best I can recall. Q Between '65 and '69, though, he was not directly involved in the research and development for 2,4-D as far as you know -- A I don't know what Harry Brust did in that period of time. Q Between '65 and '69, did you have occasion in your research to analyze or write about any potential contaminants of 2,4-dichlorophenol or 2,4-D, itself, such as any of the dioxins, TCDD or any of the dioxins, for example? MR. WAGNER: I object as to the breadth. A Are we specifically talking about polychlorinated dibenzo-para-dioxins when we use the term dioxin? Q Yes. j A Not that I recall. Q What about furans? I know that's a broad C9 9 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 15 statement, but did you do any papers or analysis with regard to the presence of -A Again, polychlorinated dibenzofurans? Q Yes. A Not that I recall. MR. WAGNER: He sounds like a scientist. THE DEPONENT: Well, I'm going to do that. Q Did you do any toxicological research between '65 and '69? A I have -- no. Q Let me ask you a very general question here as well, do you consider yourself, I know you're an organic chemistry researcher, do you consider yourself to have expertise in toxicology? A No, I have no training whatsoever in that area. Q Again, a general question, which will save time later on, as far as the, when I'm going to ask you some questions about the report you did, I think in 1978, although the date is hard to make out here, but with regard to any contaminants tjliat you ultimately may have dealt with in your research with 2,4-D or 2,4-dichlorophenol, you would not be 1 $ '& > FORM LASER BOND A PENGAD/INDY 1-000-631-6909 16 an expert on the toxicological impact of those contaminants ? A That's correct. Q That type of research would be carried on by others at Dow? A Yes. Q Between '65 and '69, did you do any research that was connected with any Dow involvement with any claims that either 2,4-D or 2,4,5-T or Silvex caused various injuries or diseases? MR. WAGNER: I object, over broad and vague. A Are you asking me did I consult with Dow attorneys? Q Yes, that would be part of it, I assume. A No, I never did. Q Did you ever testify in any formregarding 2,4-D or 2,4,5-T or Silvex from, excuse me, from '65 to '69, again, I'm talking about? A No. Q Did you ever relate any communicationsto a^y governmental agencies regarding 2,4-D or 2,4,5-T or Silvex or their component parts? i n 3*f FORM LASER BOND A PENGAD/INDY 1-800-631-6989 17 A No. Q Did you, and I'm going to open this up to the entire time all the way to the present, did you ever give a deposition in any any case involving allegations of personal injury as a result of exposure to 2,4-D, 2,4,5-T or Silvex? A This is myfirstdeposition. Q This is your first deposition, period, you've never given one before? MR. WAGNER: You're it. Q I have to get the answer from you, though. This is your first one? A Yes, sir. Q There has been testimony in this case that, I'm not sure as I look back at it, the exact date, but there was a meeting regarding contaminants found in 2,4,5-trichlorophenol of various manufacturers, I think either in '64 or 65. You wouldn't have attended that meeting, would you? A Well, no, obviously not, because I didn't join the company until June of '65, and I had no j involvement whatsoever in the area of halogenated phenols until after 1966. 112 V '* FORM LASER BOND A PENGAD/INDV 1-000-631-6989 18 Q So you wouldn't have attended any meetings with regard to the presence or absence of TCDD in any of the Dow chemicals from 1965 to 1969? A Correct. I'm sorry, would you ask that question again? I want to make sure -(The requested material was read by the reporter.) A The question should probably be, I attended no meetings, I was not involved in anything dealing with 2,4-D, 2,4,5-T or Silvex. Q Nothing whatever during that time period? A Correct. Q Let's skip forward to your moving intothe organic chemicals research in 1975. Were you assigned responsibilities when you went into that particular department back in 1975, any responsibilities with respect to 2,4-D or 2,4,5-T or Silvex or their component parts with respect to doing research? A Yes, I was. Q What responsibilities were youassignedinitially when you first went into the department? A When I first went to that department, I was 113 V 'M FORM LASER BONO A PENGAD/INDY 1-800-631-6989 19 assigned the responsibility of doing process research for the new 2,4-D plant that was scheduled for start up in approximately two years from the time that I joined the group. Q And when you say process research for the new 2,4-D plant, 1975, when you went back to organic chemistry research or the organic chemicals research department, 2,4-D was being made by Dow Chemical Company at that time? A Correct. Q It was being made pursuant to a process that they had been using for a number of years, correct? A That's also correct. Q Do you know the history of the process of 2,4-D, in other words, the process that was being used in 1975? When you went to the organic chemicals research department, do you know, was there a name for the process they were using at that time? A Well, it was just simply called the 2,4-D plant. Q And what building was it located in? A 489 building. | Q And there wasn't any specific name for the process, who put the process together or anything 114 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 20 like that? MR. WAGNER: I object, compound. A Not to my knowledge. It was just known as the 2,4-D plant. Q What was the purpose, if you recall, for attempting to put together a new process for the manufacture of 2,4-D? A We developed some new technology that was much more efficient, produced much higher yield, and produced isomer purity in the final product. Q When you say we developed? A Our department. Q I understand, but were there certain individuals that were responsible for developing that new process? A Well, the scientists. David Watson, that I mentioned before, was the primary investigator in improving the quality of our 2,4-dichlorophenol from an isomer specific standpoint. Harry Brust is the chemist who developed the chemistry that was eventually implemented in the 2,4-D process. Q Is Mr. Watson still with the company, by the way? A Yes. 115 FORM LASER BOND A PENGAD/INDV 1-800-631-6989 21 Q Is he in your department? A No. He's in another department. He's located at Midland. Q Do you know what department he's located in? A The department is called design thermoplastics research. Q What about Mr. Brust, is he still with the company? A No. Harry retired a number of years ago. Q Is he still in Midland, or did he move away somewhere? A I believe he's moved away, although I see him in town every once in a while. Q Did you know whether or not, with the process that was being used in 1975, for the manufacture of 2,4-D, whether there were any problems, from an organic chemistry standpoint, with contaminants being found in the 2,4-D being manufactured, I'll call it by the old process that was in existence, 1975? A No. j MR. WAGNER: I object, vague. Q You don't know one way or the other? FORM LASER BOND A PENGAD/INDV 1-800-631-6989 116 22 A No, I don't. Q Do you know who might have, if anyone, done research on the 2,4-D that was manufactured in 1975, to determine its purity or to determine whether there were any contaminants in the 2,4-D? MR. WAGNER: Calls for speculation. A Could you be more specific? Q Well, it's relatively specific. Let me see if I can make it more clear. Do you know who the individual was in the organic chemicals research department, if anyone, who might have done research on the 2,4-D being manufactured in 1975, to determine whether there were any contaminants in the 2,4-D? MR. WAGNER: Same objection. A No, I can't recall. Q Let me ask the question this way, maybe this will make it more clear, was there anyone that you knew back in 1975, in the organic chemicals research department, that was responsible for doing the basic research on 2,4-D back then? j A Well, Dr. Brust was responsible for the research on the new developing process. I did some 117 FORM LASER BOND A PENGAD/INDV 1-800-631-6989 23 research in the, as it relates to the old process, not specifically targeted at trace impurities. Q In other words, you weren't just looking for trace impurities; you were just doing research on it for other purposes? A For other purposes, yes. Q Did you produce tapes that outlined your research back in 1975? A Yes, I would have. Q Do you recall whether, in the course of doing your research, you found any impurities in the process for making 2,4-D in 1975? A For me to answer that question properly, I guess I'd like to know what impurities you're specifically asking about. Q TCDD, for example. A No. Q No, you did not find TCDD? A No, nor did we analyze -- was I responsible for determining TCDD in that time frame? Q You were not looking for that? * | A Correct. Q Do you know if there was anyone that was looking 118 V ' i-H FORM LASER BOND A PENGAD/INDV 1*800-631-6989 24 for TCDD, for example, in 2,4-D back in 1975? A Noi Q You don 't know, no, or -A No, I don 't know. Q How big was the department back then in 1975? MR. WAGNER: Organic chemistry research? MR. SCHULER: Yes. A More or less the same size. At this point I don't recall exactly how many people we had. Q What I'm getting at, do you have an idea of who may have been involved in that type of research, or you're just not sure who it was? A I just simply don't recall anyone doing any research at that time, and I would, if it were done, I'm sure. Q Okay, to get back to my earlier question, I asked you about whether in your research I think you found any TCDD. I know you weren't looking for these things, but in the course of your research on the 1975 process, did you find any of the other polychlorinated dibenzo-p-dioxins, did I pronounce it properly? A What you're specifically asking me about is did we 119 V' ^ FORM LASER BOND A PENGAD/INDY 1-800-631-6989 25 look for these compounds in the old 2,4-D process, in the 1975 time frame? Q Correct. A No. Q Let me change the question a little bit to, even though you weren't looking for them, did you find any of those? A No. Q Did you work with someone in doing your research on 2,4-D in 1975? A Within our department? Q Yes. A No. I was doing the work. Excuse me, are you asking about the old process work? Q The old process, that's right. A Yes, that was done by m e . Q You didn't have someone that worked with you in the department, it was just you? MR. WAGNER: Asked and answered. A As best I can recall at this point. Q In your work on the old process, can you ju^t tell me what your research did involve back in 1975? A At that time, again, we were just -- one of my 120 V' A/t FORM LASER BOND A PENGAD/INDV 1-800-631-6989 26 roles as a process chemist is to make sure we're using the most modern, up-to-date technology and chemistry that we can do. And my research was essentially aimed at doing just that, making sure the plant was operating, that the chemistry in the lab was operating most efficiently within the boundaries in which we could operate in the plant, maximizing yield, maximizing product quality, that sort of thing. Q Would you try different things in the laboratory to see if you could improve the quality to maximize yield and maximize purity? A Correct. Q And if they appeared to be, the things that you did were improvements and appeared to be things that could be incorporated into the process, then that suggestion would be made, is that how it worked? A It would certainly be considered. Q Who would make the ultimate decisions as to whether a particular change in the process ^ould be incorporated? I know there's probably a number of people it went through, but just give me a 121 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 27 general idea how it worked. A At that time the, well, the ultimate -- the person ultimately responsible for the plant would be the plant superintendent. Q Who was the -- A But the plant superintendent would work with people in the agricultural chemicals production department because we had some senior level experienced engineers that were located in that department, whose responsibility it was to maintain high standards of technology from the engineering perspective. Q That was the agricultural chemical production department, you say? A Yes. Q Who was the plant superintendent for the 489 2,4-D process back in 1975? A At the time I joined the group, the man's name was Bill Jones, William Jones. Q Is he still with the company, to your knowledge? A As far as I know he is. j Q Do you know where he is in the company, what department? 122 FORM LASER BOND A PENGAD/INDV 1-800-631-6909 28 A I haven't any idea at this point. Q In the plant, would records be kept about the process, in other words, the yields that were being obtained, the quality of the product, would that periodically be tested and records be kept on that? MR. WAGNER: 1975 time period? MR. SCHULER: Right. A During that time period, yes. Q Do you know what other records that would be kept in the plant, primarily? A Well, they would keep operational records of how the plant was running, product quality, the results of quality control testing, records on plant maintenance and et cetera, et cetera. Q Would the plant superintendent be specifically responsible for the quality control testing? A He's the person who is in charge of the people who do the work, yes. Q Some of these questions, I'm sure, seem very basic to you, but I need to know, do these plants^run continuously? A Yes. i23 0 ' m FORM LASER BOND A PENGAD/INDY 1-800-631-6989 29 Q 24 hours a day? A Yes. Q And besides operational records and records dealing with yield and purity and quality control testing, are there any other types of records kept in the plant back in 1975, I'm talking about? A I think I covered the ones that certainly I'm aware of . Q Was it ever your function to go and review the records in the plant to see how the plant was doing? A Oh, yes, I would, on a regular basis, review records on how the plant was operating and how the product quality was holding up. I didn't have any reason to review records on maintenance or scheduling or things like that. Q When you say product quality, and taking 2,4-D specifically in 1975, and again, as generically as you can, what would you look for to determine product quality? MR. WAGNER: I object to the fornyof the question. A We -- well, we had certain specific indications 124 0 -150 FORM LASER BOND A PENGAD/INDY 1-800-631-6909 30 that we -- that the product was required to meet by our customers, and we would maintain a track of that and try to maintain an awareness of how the product quality was related to the specs that we were -- specifications that we were expected to meet. And if problems or issues came up with regards to meeting specs, then we would attempt to solve the problem or issue. Q What type of specs are you talking about? A Assays, A-S-S-A-Y, the tracking of normal isomer, isomeric impurities. We had several final product quality control tests that we would monitor on a regular basis. Q When you talk about assays, what are you talking about, what type of assays? You mean concentration of the product? A Well, the quality of the technical product, the assay at that time was expected to be 95 percent pure 2,4-D, so that's what I mean by assay. Now, what I'm talking about is the technical 2,4-D assay, which is 2,4-dichlorophenoxyacetic aid, A-C-E-T-I-C, acid. And the spec at that time was 95 percent purity? 125 FORM LASER BOND A PENGAD/INDV 1-800-631-6989 31 A Yes. Q And when you say 95 percent purity, was that in what form, was that in liquid or powder? A Powder form. Q Was the powder put in suspension before the product was shipped, or was it shipped both ways, how did that work, if you know? MR. WAGNER: I object to the form of the question. A It was -- we did supply, some of our materials were supplied to the customer as the acid form, but then it was supplied in a number of other physical forms. Q Depending on thecustomers' requirements, more or less? A Right. Q Now, when you said they were tested for isomeric impurities, was there any test for the dioxins that we discussed earlier, for example, TCDD? A I simply don't recall how that was handled in those days. | Q What about any of the other polychlorinated dibenzo-p-dioxins, were there tests for any of the 126 0' 152- FORM LASER BOND A PENGAD/INDY 1-800-631-6989 32 others ? A Not that I specifically recall. Q Same question with regard to the furans, were there any tests for those? A Not that I specifically recall. Q What type of impurities, when you say isomeric impurities, do you recall any that was tested for? A We would, our quality control procedures at that time did not require that we analyze for other isomeric impurities. Q I thought you told me earlier that there was testing for isomeric impurities. Did I misunderstand you? A I'm trying to recall 18 years ago. If I said that, I guess I must have misspoken. At that time, the specs that we were required to meet in our quality control was 95 percent assay, a 95 percent assay. We had the ability to analyze for other isomeric impurities, but that was not an official specification. That may be where we got confused a little bit. jj Q Does that mean that it wasn't done or it was done periodically or what, testing for isomeric 127 FORM LASER BOND A PENGAD/INDY 1-800-631-6989 33 impurities ? MR. WAGNER: I object to the form of the question. A It was more likely to be done on an irregular basis, and most likely from a research standpoint. Q As opposed to a quality control standpoint? A As opposed to a quality control standpoint, yes. Q You mentioned initially you did some research on the old process for manufacturing 2,4-D, and at some point you were given responsibilities to do process research for what was potentially going to be a new process, correct? A Correct. Q Was there any other research, before we get to that, that you did with regard to the old process? I know I've focused on impurities, and you've told me basically what you did, but was there any other aspect of your research that we haven't covered from the old process? MR. WAGNER: This is 1975, before he started working on the new one? j A As it relates to the 2,4-D acid process in the 489 building? 128 P' M FORM LASER BOND A PENGAD/INDY 1-800-631-6989 34 Q Right. A As best I recall, I'vetold youessentially everything I did at this time. Q From a chronological standpoint, how long were you involved in the research in the old process before you turned your attention or your responsibilities were directed to doing research on the new process? A Only a few months. Q And when your attention was directed to the new process, what was your initial assignment with regard to that? A Essentially to provide chemistry support to the construction of the new plant, focusing on gaining a personal understanding of the way the new process was going to operate. Q Let me get some time parameters in here, too, so perhaps this can save me some effort later on. Did the new process -- was it implemented? A Yes. Q Can you tell me at what point in time the n^w process was implemented? A We started the new 2,4-D process in the spring of 129 p - ISb FORM LASER BOND A PENGAD/INDY 1-800-631-6989 35 '77, I think. I see you've got my report. I think it talks about May of ' l l . Q Okay, was it May of '77, is that when the process was began? A That's when we started the production of 2,4-D. Q And up until May of 1977, the old process was being used? A Yes. Q Was the new process, did the new process, was that put in the same building, 489? A No. It was a new structure. Q What was the new building? A 948 building -- 948-949 building. Q From your recollection, if you know, when the new process was started in the spring of '77, was the product, in May of '77, was the product sold right away or was there any delay in the sale of the product? MR. WAGNER: I object to the form of the question, vague. A I have no personal knowledge as to when we j actually began -- the old process continued to operate for approximately 18 months after, as best 130 p -'lS 'lo FORM LASER BOND A PENGAD/INDV 1-800-631-6989