PH26631A

PH26631A - Method for treating caustic-refined glyceride oils for removal of soaps and phospholipids

Info

Publication number: PH26631A
Application number: PH35248A
Authority: PH
Inventors: William Alan Welsch; James Marlow Bogdanor
Original assignee: Grace W R & Co
Priority date: 1986-05-14
Filing date: 1987-05-13
Publication date: 1992-08-19
Also published as: CN1029318C; EP0247411A1; EP0247411B1; PT84208A; MX170388B; ZA873334B; PT84865B; JPS6327600A; CA1298853C; AU600485B2; PT84865A; AR242428A1; ATE59060T1; CN87101626A; IN171401B; DE3766651D1; ES2019324B3; PT84208B; AU7294687A; GR3001427T3

Description

Lo Ca TARTR OLN TONNE . . - : : . a a Livy nyse oo oo | LAU Mg : hen i} : SN EEE UREA oo | 4PECLIICATION | .

To ALL HAOM IT HAT COHCERIL ! 9 66 3 1 : . v0 ’ ) gE IT KHOHH THAT WE, WILLIAM ALAN WELSH AND

JAMES MAKLON BOGDANOE, both citizens “of. the

United States of toericn, «of 125680 Hall Shop - bo . Road, Fulton, Harylond- 20759, United States of . . oo America and 5457 Tiited Stone, Columbia, -

Maryland 21048, Inited States of American . ’ respectively, ‘have lnvented a certain new and * oo - uaeiul Lprovement In “HETHOD FOR TREATING nL !

CAUSTIC-REFIHED GLYCKRIDE OILS FOR REMOVAL oF

SOAPS AND PHOSPHOLIPID3", <f which the ) following 1s & specification. | : \ oo — “

Co ——

BAD ORIGINAL 9 :

i ’ . .

This invention relates to a method for | : refining glyesride oils by contacting the “oils with an adsorbent capable of removing Certain” : | impurities. More apecifically, it has ‘bean found that amorphous silicas are . gulte : effective in adsorbing both soaps and oo phospholipids from caustic treated or caustic veined glyceride cils, to produce oil products with asubotantially lowered concentrations of . oo 19 these Impurities. For purposss of this gpacitication. the term "impurities refers be goaps ant phospholipids. The phospholipids are oo BE agaasclated with metal tong aud together they Lo wilt he referved to az “trace contaminunts.” 16 The torm “glyceride olls” as used herein 1s totandead to encompass both vegetable and animal : of ls. The term i18 priwarily intended to ; i dese i be the so-called adlble oils, Le, olls . !

Jevived from fruits ov ceeds of plants and used . : 20 chitefiy in feodstaffs. but it is understood oo a. that oils wheso end use is ag non-edibles are . te be included ac well. The invention is - applicable to oils which have been wybjected to : . o, { . .

BAD ORIGINAL 9 . : .

caustic dreatment, which ie the refining etep : in uhich soape are forwed in the oil. , : Crude glycaride oils, particulusly vegetable ©3ls, are refined by a multi-stage. procags, the 2irst slop of which typleally is ‘ “dsgumming” or “desllming” by breatment wlth oo water or with a cheminal such 3s phouphoric : acid, eltric acid or acetic anhydride. This : . treatment removes some but not all gums and » | 10 certain other contaminants. some of the So ’ ) Fhosphorug content af the oll ia removed wlth ’ : the SU, Jy 1 eye in] e m ) ' 5 Either crude or degummed oil nay ba . breatad in a chemlval or o ; . brocess. The additi Hustle, refining s ag oe : : Lion of & ~ SY .

Castle soda Tor exam Woalkali solution, : degummeq 011 o “Xample, {o 4 crud : : Cause pe de or o

PULTE isnt nn . tut anil Erp , fatty acids to Form eeape. Chie sbeg iw Eke refining process will be referred to herein as } : "eausbic treatment” and olle treated in this manna will be referred to as "cauatle treated oils. Scaps geimrated during caustic treatment

Anas na } ! on Ari “are an impurity which must be removed from the oil because Lhey have a detrimental affect oni . : r . . : . oo BAD ORIGINAL 9 the fluvor and stability of the finished oil.

Moreover, the presence of scape is harmful to the catalvets used in the o¢il hydrogenation } } process. to

Current industrial practice is to first . remove goapg be centrifugal separation ) (referved to as "primary centrifugation”). In : i } this specification, oily which have been

Co subjJucted to caustic treatment and. primary 12 centrifugation will bs referred to ae "caustic refined” oil. Conventionally, the caustic ’ - refined oil, whick still has significant acap conten, 13 subjected te a water wash, which digsolves the goapa from the oll phase lnto the 16 aqueons phags. the tuo phases ary separated by : } centrifugation, although complete separation of i i the phases 1s not possible, even under the Lest : | cf conditions. The light phase discharge is oo ol water-vashed oil which now has reduced Bop content. The heavy phase 1s a dilute soapy water solution, Frequently, the water wash and ’ senteifugation steps must be repeated in order : } oo | BAD ORIGINAL 9 bo ree biins Gh sea conteut of Lhe ¢il below ‘ hans Ta ean, The dater-wached oil they must :

Lo deved te romove reaidual wo istnre to below about, 0 1 yeighl neraoant. The dried oil is 6 Tsu wither trancforred Go Lhe Liaaching }

Dra or do shipped or stored ax once-refined wal '

A wlandricaut part of the wasts diachavrge fiow ihe caauiic reXiniag of vegetable oil } 14 vedas dren the quater pach proosas used to oo Deis BOLD, In tact, a primary reason for aN voflaers’ dew of Lhe physical refinlug pro2ess _ | i to weeded Lhe Wages locum Froduit ton oo avon tated with remeval or Seals gensirated In ’ IO We wunitie ruling proseos: sinee no canatic - Co boo used CoLn physial solisiug, Mo S00RE are doanrated, In addition, ia Lhe caustic

Voda Prac, bibl aii Se lout io thie pater ah Prove. In the caugtlic vefianlag PEOCESS ‘ “ul ver abloh Lhis tavoutlis on valataen, morauver, the So datat Soap, Yosh neds beg Garton Lerore : . dluposal;, typically with ai fnorgants acid such

Lo oo oo : roo-

BAD ORIGINAL 9 ug auld forie ald an 4 PEGE GE te: rmed acidulat ion. Guilin bo add ae frodguentld veil, oo IL can be even hab aad bez ee TAGE LF separato . anit operations mabe up ili ovap rosa al : ‘ H proces, wach of which resulbe Tie woes dagroaa of oll loge. The opera wend A hspeesald oof aanps : i and aguecns soapshock ls one af the ww : conglderable problums snus raked with the

CC emaatie refloing of glyoer ide olla, : . 19 Ti addition bo removal of soaps created fun the crustic refining process, phosphorae- . } containing Trae cotaminands wai he removed row Lies Sil. The presepow of theo trace . . conbamlnaubn «aa dural GE colors, wdors tnd

Clare Ver bbe Finlobesd S31 prod at theaa componads ae chrogpbiod Lpdds, vi th whieh are atic fated tonic Cogms of Lhe wealails ole ba,

MAZES bE, Fare mach soppy. For purposssot . : thi Loveoblon, valeraesod tu Lhe pomornl or of adaurpl ion of phospho tiplda jgintonded algo Lo . refer Ls vomoral oo or adeurpblan of rhe apaosiated fer bal fiona. Adeovrpblon of . : rr ,

BAD ORIGINAL J

. x phcsphorns on varivus adsorbouba 9for crampla, bicaching warth? hoo been practised but on ly : With regpact to sila undergoing physleal ) | refining (Ln ubiloh fie woaps sre generated} or in canst refiniog subsequent bo water wail - shops (Lon which hha weaps are removed). Ble adueri-lbicn process has osccomplished the removal

Sof bitoh wuers suid phospholipids alt en early

Cubase or Gass t je to lining whoa large

C10 quant ition of ooape are presont, : : . © The praseonb Lavand ien provides a & Lap leo ploeloal adouscpliou process by which coup snd phospholipids can be reo ol from

Cova ba Lmao cr Gaus ble rel ined co vegetable : 0) vive do om Gang. vol operable. Tiddosunique i } Ld Soap besten ly wllainated Lies nosed Lo sublet causlic Leeabaed or cadable roebiaed. wil

Lo a waber washing prossas be sede Lo Lemony es fwaps it alu aliwloabos whe pew) for wo . or Goparnbe adoospi dun process Loo redone the phosphor lipid contaeet of Lhe oad, Lhe process described hevelns ubiilass sworphoues silica : i

BAD Origa, | 9 adgorbents having an average pore diameter of oo

Co greatur than 6UOA which «¢an remove all or

Co CE substantially all sours from the oil aud which } raduce the phospholipid content on the coil to 6 at least below 1H parts per million, preferably below 5§ paris per million, most preferably . : cubatantially to ABLD. } deep Lok of soaks and chospholipids - {together with asgociated contaminants) onto ' 13 aworphous silica in the warmer described herein

Gffer:; proeal advantage tu caust ia refining ‘by eliwinat ing Lh. overall unit operations : voguired when zenventlonal water-washing, cenbrifugation and dryviug are employed to 1% remove soaps rom the oils. In saddztion, the nei method eliminates the need for wastewater treatment and disposal from those operations. : "Over and above Lhe woul savings realized from: thio tremendous simplification of ble wil : 2a processing, the everall values of Lhe product is increased since a sispificant by-product of } conventional cavistic refining ia dilute

Bs © BAD ORIGINAL 9

: aqueous soapstock, which is of very loi value oo and reguires substantial treatment before : disposal ig permitted Ly anvironmental authority. ; oo & It 1s also intended that use of the method : of this invention may reduce or potentially : eliminate Glas need for bleaching earth {reatnenl. Iu this inveniiom only ane ) adsorption step 1g used for removal of both oo , 10 soaps and phospholipids. Additional treatment dith bleaching warth Lo remove these ’ impurities typically will not be required.

Redu.tion or elimination of an additional Co bleashing garth step will result in substantial 15 . oll conservation as this step typically rasulle : in significant oil leas. Moreover, since spent hlaaching earth has a tendency Go undergo ' oo . spontaneous combuation, reduction or elimination of this step will yield an i

BY ocoupationally and envivonmental ly safer

Process. : oo

BAD ORIGINAL 9

Fal additivnal vhijuct of the invention is to } : simplifs the recovery costs and progessing now agcoecialaed wilh preparation of the aqueous soaps tach Tor ase in the aulwal Deed industry. .

Eb The spent silica adosrbent can be used in animal foeds gither as ls or after acidulalion to convert Lhe scaps Into free fatty acide.

The need in the conventional caustis refining ‘ oo

Co Proce for drying or concentra ing dhe Jdilutc in : 10 soapstock 1s eliminated by this Invention. la the accompanying drawings, Flgure 1 is : a graphic represents of adsorption isotherms for ithe capacity of amorphous silica for ceombinad phospholipids and Soaps. Tha ’ 16 Juotherms are based on the results of Example ’ IT as shown in Table V.

Figure 2 ig a graphle representation of

Nn aloorplion igotherws for Lhe capacity of amor phous silica for phospholipids, for treated : : 20 oil with £ 3d parts per million residual soap.

The lsolherms are based on the resulte of :

BAD ORIGINAL 9 a ) ’ " © Example Il as ohoun io Table ©.

Tie piecont daveabion io based on the | }

Jigoerers Lihat aero sllicas are «oo . Co particularly weil suited far reno ing both ) & B00 PE ard phosping Lip dis from causiiac ‘rafiued plyceride ile. The AR wi Lhe Las ent lon

For Lhe peposal of fd bsie WELLE ded Comp Lied . sont eobars a ouerab bz teeolesd or doal vo ore fYonesd 7 .

E cog Yraarids all wih Leek robe SALE and : 1 chosphel tpidas wlth morpho arlicw, Af Plowing - the Gounps avd preocspboel iid bo be adeorbaed cnto othe sa rppos co Llica, and wep waking the - oo absorvenl rested 01 Crew the adoorbunt.

By thea Le wv thila dovaatlon mode vind . phospholivlde are pomeved Frone olla Loa single advorption shep. Povsover, 16 has Laan Found ' Lhat iporasaing levels of soap in the oil to be oo

Epontad wobuad ly baore coed bb capacity of Lhe

Amey tia wad bie a no allsuerh ph ooophe oad, Troal’ i 20 1a, Lhe pagers of goals ab Levels belo fhe gave dpa adsorbent capac gk of the od Jioa makes :

Sd possible Lo anbatantisdly roduee phosphorao 84D ORIGINAL 9 cont onl al bower Gili asage than dd rogulved in ile abaeress of soaps. .

The Oils | )

The Provess described herein can be used b for the removal of phospholipids from auy . cauwila refined glycoride all, for ex anple.

Obdis wi soybean, peanut, rapasaad, Gorn, auniiower, pslm, coconut, olive, cottonseed

Po ebe. The caustls refining process involves the 113 pov ralisation of Lhe free fatty acid contant : , | of crude or degummed oil by treatment with oo bape, such wes sodium hydrostde or sodiam carbonate, whlch typically sre used in ,Aduecus } uo dui lon, The neutralized reo fatly acld 16 presant as the alkali or alkalloe cart salt ia doliaed a8 soap. The seap content of caustic

Lraavesd old will vary dependiog ou the free

Paley content of the unrefined oil, Values divelesed ag bypiloal la Lhe industry vary from

Sleels 30 ppm goap for caustic treated oil.

CiEricicon, B31. Haodbook of Soy Oil Process Lng wad Utidisation, Chapter 7, “Refining,” po 81 ! : BAD ORIGINAL 9

{ . : C oi 0 . oo ¢ © . . co : DEY Combo Ga BY ppm oss. for caus : : ST LL Cali PA hay Someries g LE oo

Cohan beswion, olor Gowloe Piccenedng 0 and : : ) ce RLY WON od LT Pe La axed ad cog cd, . ) Gos pinball Geir. Usb homie Soe oy May U0 . CL

Cel rally rian sila mmen pave goap Co : Covad es ES EAS Lr dara. Ceaventlonsat oo 0 0

Dauparut ion ared Wold osmahing Drocoisss pomovve 0 0 : Coiba dn of Ther goes content gone meted by kha So

Ce hd Cites Phe Lreallienl Gh) The prosogs dase loosed ol oo ‘ ‘ ’ bored wa sil yaduce wpe 4 Irevala acceptable . To . . Coo be nduantey, than Lo Tess chan aboab ig in : : pin pans fhapand, Peco ban lout 5 pom, most Le Co . preforebiy abou eos pa, W10hed tl bBhe use Hl : .

Cdk BLL wasn alg, ’ BE : " 4 : ; . : co J . - B B

RL an Co tora contact emis Lh : { yi rte | aed nd cic eaten wae a Lele ) ? Foon . Co EK : ’ . . . : Lo . Loon . : oy . oo make cig Ye Le on nd cand Thode svierpe Ln Lr ER rh A

Cerda nee as Low papain lea Cle Takes al f oo Ted

Co LZ Aland alte why Tau rn dn 1h. Yin ahaeid ceil, } : [ET ’ Coben i ie CA eed Lm rth Catia. oof to ) ’ oo . NC oo Co . ,

Co Cobble rch Ge dL aah any fod aio ig. . : hes d Be bell ated den oy bea Tega, 1y codes } } ’ : ) ' ! ’ ' : rt a ta . + + - . . . . . ‘ . LEE 1 ! . ) . . : . . . . : .

. ~ than olont H&E, pacar ding Fo general induasivy practice As on 1llusltration of Lhe

. “ ’ refining goals with regpect Ko trace contaniusabo, typical phosphorus levela in 6 soybean 11 at various sbages of chanical : veiling are howe in Table I. : oe f } : BAD ORIGINAL 9 : . be : 15 oo

+ vv 3 - ’ ! ce pI i cide

Lo i b) 2 Is , ls Le

SEY Ea RRS {

JAY 3 & - 8 - ] ; i 1

Sa o : 0 l~ | © Ly : £ coo i :

SRS 9 al= ~o . i

N - i" 0) i oc Q H nein x £0 a [ cco © oo "a oH a. vee . oo nw J 0

E|O man oN oe oC © HB : coco © vO . ~M : es = oi um O 8 0 Mme } & ~C-d w no wy ™ nL QO ~ < 5 Mose . ne ~~ ~~ ¢ gle] PL <1 He OE 0

Sl “re TC 0 +«sgC J

T . . . Ag © 3 5. 0 U a : ec ~uuC KL " neg PO . : © Nn Nn He oO ¢ > g 5194 Vorb Un eH bs BR : 519 LALA ogc ww E i

LE or WE ’

E MM eH CU — © ow dd » nnn no de 0UAHE C : i. — cjm Pr a cu

Hl clo —-—-—- © +E HL TO

S : gEecPU 0 oo 5 oo Mm Na aE @ NC TL «4% Or-mM : © cco © HO UD®E B <| ®© non © uw ¢ — Cob

Bl M ~~ a re £e¢ > 4g - ele 111 | ol PH EE & coc o Cem L Dvd © ~ ° nS © Cc ~ £ oem Or Mod . o CE kh Med | God 1% a MU GC

EOC i ~ £0 E>dT ; —- © — co OL i © Cave EP -f 3 © © ap cw ew Ek 3 Gog © TO

BOW ENED HB H

¢ A @ ¢AMmOETUE HE + [® [=] Nn — - QO Q “wm a ed wed sup ~ 0

HO £8 wa CH —~0 EQ h-

OH oO QU Do we» od :

CH OU K3 tc ou EL Ww : no Sg dT O00 EY - couL »™ LO OE HHO ” En ooh Qm-A AOKU \ u a Fe] T° a,

Oo: vow EW @ sSog HA3T 1 ! ry) Om HO ® EE 0 oQu A oOMA — o~ ‘

In addition to phospholipid removal, the E ’ Process of this Invention alse removes from

Edible alle domo forme of the wetale calolum, .-

Magnes iom, fran and supper, which sre belleved :

Lo be chemically associated with bhospholiplds, and yhlch are removed ig conjunction with the phospholipids. These metal long themes ves have a deletsrious effaot on the refined oil oC Froductys, Calcium ung magnesium ions can } 18 result in the formation of pracipltates, | Lo particularly With frac fatty asides, rouulting in ‘undesired 80aps le tle finished ail. The Co presence of Jeon ard Copper ons proute

OX1dal iv instablityty, Moreover, each Of these ms tal long is associated with catalyst

Poluoning where there fined oll iu catalytically - . hydrogenatag. Typlanl tencentrations of these etal in Soren Gil oat, various stages of ‘ chemical refining Len Shown in Table I.

Throughout Lhe desuriph ion of this invention, urilegs otherwise ladicated, relerence to the removal ef phospholipid. i& meant to COCOUEIRE 6 . the renosval of associated metal ions ag vell, Co ’ . 1 - °, } ’ . . | Le {

BAD ORIGINAL 9

. . 26631

The Adsorption Process - The smorphous oo sliicas described below exhibit very high aapaoitly for adsorption of Boaps and phogphol lipids. The capacity of the silica for phosplivlipids is improved with increasing soap levels In the sterbiog oil, provided that sub ficisnt silica 1s used to obtain adsorbent peaked oil with soap lavels of approximately 33 pew or less. IE 1s whens the residual soap 19 levels (in the adsorbent-treated oil) fall bolouw about 38 ppm that the lucraaced capacity ol Lhe silica for phospholipid adsorption is aeu. Lois pelieved that the total available adv pbloen capacity of awcerphous silicals about ’ 15 NE A Th wi. % ona dry Laois. ‘ } | The silica uvage should be adjusted so that

Lhe total soap and phoopholipid content «ff the caus lic treated or cangatic refined oll does } nods creacd about BO to TH wt. kh of the wesilllca : } 2a added on a dry basis The maxioum adsorption capacity observed in a particular application

La sxpoeted Lo be a functlcn of the specific

ER Bn BAD ORIGINAL 9 .

properbise of the silica used, the all Lype and lage of relinement, apd processing conditione ‘euch ug temperature, degree of mixing and sllica-oil contact Line. Caleulations for a

Co b specific application are well within, the : knowledge of no peveon of ovdinary skill ase . . oo quldsd by thie specification. co The adsarrt fon step Lha=ll le acconpliehed - "by convenbicnal methods in which the amorphous oo 143 tlllen and the oll are contnebed, preferably in a mmannsr vhich facilitates the adgorvphtlon. -

The adsorption shap may b= by any conven lent batch or coontlonuousg praoasss. In any case, . agitation or wihar mixing will euhance the adsorption «fflaiency of the silica. :

The adsorphlon can ba copdacted at any : convenient btemporature at which the oil ts « :

Tiquid. The caustic refined oll and amorphous

Co gllica are contacled as degoribed above for. a 2a period enfficient bi achieve bhe decired levels of soap und phospholipid fn the treated : - 19. — f .

BAD ORIGINAL I t hk: ji | 5 ol ee Bat it BE § «iol Fe § a

Re Rew ped ~ in i!

CRE Bre Bod y, Bs a . For es

TET FL dehs } Rob

ACT . piu rK ] Ras oil. The specific contact time will vary Ek £iN ! i of ? cM . fag _ somewhat with the selected process, i.e., batch Eo 4 2 1. ily k Co ERE i 3 or continucus. In addition, the adsorbent 8h yo ¥ : pero] 0 i usage, that is, the relative quantity of For : ; 3} adsorbent. brought into contact with the oil, £1 will affect the amount, of BODE and & phospholipids removed. The adsorption usage is ¥ . . EF quantified as the weight percent of amorphous Fi

Fol gllica (on a dry weight basis after ignition at od %. 1 19 1750°F or 955°C), calculated on the basis of £1 the weight of the oil processed. The preferred gol . Ear ’ . . = yd adsorbent usage is at least about 0.01 to about Er :

BN 1.0 wt. %, dry basis, most preferably at least iE about 0.1 to about 0.15 wt. %, dry basis. RE vo

L 18 : CAB geen in the Examples, significant gif oo par # by gyi ‘, Fart ge Sag Te TURRET ag ed en Lo ! EA A

Sl : HFedudtion” in soap and phosphdlipid content is RE" x

CL achievved by the method of this invention. The go ry] soap content and the phosphorus content of the xl sd

HN) : NR E734 treated oil willl depend primarily on the oil tr a BA]

ELL

2 itself, as well as on the silica, usage, od l 5 [I process, eto. For example, by reference to Lol

Table I, it will be appreciated that the mE !

HE

—- 20 - {

BAD ORIGINAL 9 i ;

Sr se EE —

LA

EN

EN , ’ B, ri - ; id ‘

GS

Sr ga

Bin x il initial soap content will vary significantly ' J 3 a : depending whether the oil is treated by this : gf 1 adsorption method following caustic treatment iE

Sy - or following primary centrrifugation. i 2 5 Similarly, the phosphorus content will be } : ! somewhat reduced following degumming, caustic a treatment and/or primary centrifuge. However, ‘r phosphorus levels of less than 15 oom, ! preferably less than 5.8 ppn, and most : : 112 preferably less than 1.¢ ppm, and scap levels : of less than 58 ppm, preferably less than about 19 ppm and most preferably substantially sero ” prm, can be achieved by this adsorption method, i Following adsorption, the soap and i Dee iE PRA eB Ap Ge he Lo ei Co : ws phospholipid enriched silica is removed from the adsorbent-treated oil by any convenient : t means, for exauple, by filtration or centrifugation. The oll may be subjected to additional finishing processes, such as steam § : i refining, bleaching and/or deodorising. With ; [ low phosphorus and soap levels,it may be ¢ . A feasible to use heat bleaching instead of a bi - 21 - ] 1 \ (1

0 ase

Bo 4 fi yr 7 25 i

SO

3 Le : 2] fg ;

SE

A ; ! Eo ) i” JO . :

HO § - . i -.

FE ) 66 3 1

NO ;

AE i » ; 3 I .

BR § bleaching earth etep, which is associated with ;

E>

R slgnificant oll losses. Even where blesching i r earth operation are to be employed, } i

J = i . simultaneous or sequential treatment with

BF i amorphous silica and bleaching earth provides x an extremely efficient overall process. By 1 first neing the method of this invention to ~ l

Il decrease the goap and phospholipid content, and : then treating with bleaching earth, the ! no 109 effectiveness of Lhe latter step ies increased. i i } co iB% stan woo Therefore,. either the quantity of bleaching ak earth required can be significantly reduced, or . the bleaching earth will operate more effectively per unit weight. The spent silica i may be used in animal feed, either as is, or ] following acidulation to reconvert the soaps into fatty acide. Alternatively, it may be i feasible to elube Lhe adsorbed impurities from : the spent silica in order to re-cycle the 2» silica for further oil treatment. tl

Bi “ { ’ 1 - 22 - i : i i

Eli i i oN

ARTI ;

CE

.-. : % ; : i ‘ [he Adsorbent : The tern “amorphous silica" as used herein : is intended to embrace Bilica Bels, . Precipitated sllicas, dialytic gilicas and ; 5 fumed silicas in their various prepared or 1 activated forms, Both silica gels and } Precipitated 8ilicas are Prepared by the 5 . ! destabilization vf aqueous silicate solutions by acid Neutralization, In the Preparation of 1a silica Bel, a silica hydrogel igs formed which # then typically is Washed to log salt cantent,., Cd] ve ofr The we hE be milled, or it may be “ dried, Ultimately to the point Where its structure no longer changes as gq result of 1 shrinkage. The drizd, stable silica is termed a Xerogel., Ip the Preparation of Precipitated i sllicas, the destabilization is carried out in the presence of Polymerization inhibitors, such y a8 inorganic salts, which cause Precipitation 2 of hydrated silica, The brecipitate typically i is filtered, Washed and dried. Fop Preparation of gels or Precipitates useful in this i 1 : i tl

J k 4 i a 3 i 5

T. B 3 E invention, it is preferred to initially dry the = 1 | gel or precipitate to the desired water : p 1 3 content. Alternatively, they can be dried and : [

Hi. ; i 3 then water can be added to reach the desired =r if 5 water content before use, Dialytic silica is 3 1 prepared by precipitation of silica from =a 5 . fh 3 ’ soluble gllicate solution containing B a electrolyte salts (s.@., NalOy, NagS0,, KNO) © i { 1 while electrodialyzing, as described in pending i. . ¥ 19 U.S. patent application Serial No. 533,206 i»

I (Winyall), "Particulate Dialytic Sllica,” filed = er owl ea September 8051988% non ‘United States Patent N : ¥ No. 45@B6@7. Fumed &ilicas (or pyrogenic ] silicas) are prepared from sllicon tetrachloride by high-temperature hydrolysis or other canvenient methods. The specific } a. manufacturing process used to prepare the : amorphous silica is not expected to affect itg utility in this method. ;

In the preferred embodiment of this invention, the silica adsorbent, will have the ; highest possible surface aren in rores which

I i - 24 - Il ig

Cd i . di

¥ bo i

A

* po are large cnough to permit access to the soap | ] 3: E i . and phospholipid molecules, while being capable ' § § of maintaining good structural integrity upon g : contact with the oil. The requirement of ;

F 5 structural integrity is particularly important where the silica adsorbents are used. in J

H

3 : continous flow systems, which are susceptible to disruption and plugging. Amorphous sllicas i suitable for use in this process have surface i § 19 areas of ap to about 1200 square meters per

By : . “ he ! gram, preferably between 18@ and 1200 square | IB ; : meters per gram. It is preferred, as well, for 2 iq. 88 much ss possible of the surface area to be iB

RZ ‘ §

Co § i ms SOE, Porgs,vith diameters greater than. E

T 15 60a 3 ; The method of this invention utilizes § ; 1) amorphous silicas with substantial porosity = ) bo ‘ contained in pores having diameters greater * » than about 6A, a3 defined herein, after : 2 appropriate activation. Activvation typically ; is accomplished by heating to temperatures of i about 453 to TORCF (239 to 370°C) in vaccum. B . : if ¥ - 25 -

A

¥ i x . [ iB of 1

IE

2 's : Ri !

Ci J » ] 3 b« One convention which describes silicas is I 3 be average pore diameter ("AFD"), typically R

Ch 3 3 2 defined as that pore diameter at which 50% of I 3 3 the surface area or pore volume is contained | I

I : ? - 5 in pores with diameters greater than the stated ¥

SA E yr APD and 520% is contained 1n pores with

A 5 diameters less than the stated APD. Thus, in I

EI 1 ik . amorphous silicas suitable for use in the } 3 ik method of this invention, at lesst 58% of the y ov IR x 10 pore volume willl be in pores of at least BOA } . § 3 diameter. Silicas with a higher proportion of E ¥ oo pores with diaweture greater than 6A will be 2 yi k!

Ea preferred, ae these will contain & greater I ovo ast number © potétitial “adsorption elites. The ; a x

I 15 practical wpper APD limit is about, 52P0A. a

Silicas which have measured intraparticle 3

APDs within the stated range will be euitsble x for use in this process. Alternatively, the 5 required porosity may be schieved by the 28 creation of an artificial pore network of interparticle voids in the 60 to S5@00A ranges. i

For example, non-porous silicas (1.e., fumed jl ] 1 - 26 - r n 1 i i} : ‘

EE == ~ ea ———————— tt I ) .. i a.

BC dl. : Eo po + . ir . [1] : pt 1.

Bs Dy = ol ) 8 Bp gy ; ‘4 SI. HE 8 J: | J

A g . | : fs ft s hi, Jd

N RY" +. be B

EA silica) can be uscd ae aggregated particles. a be Silicas, with or without the required porosity, JB po 3

B . . BR “I may be used under conditions which create this : k 2 EB

A i artificial pore network. Thus the criterior i 3 ¥ i 5 for selecting suitable amorphous eilicas for ; 3 ? x

SE use In this process 1s the presence of an JE r } ‘ id tH pe. : ph i "effective average pore diameter” greater than 13 i I a ! =

Kr 64. This term includes both measured EE : 1 J: : . i intraparticle APD and interparticle AFD, jt

E 10 designating the pores created by aggregation or he , . =~ packing of silica particles. Ei . JAK ¢ . JR + py «F The APD value (in Angstroms) can be 1% fot Bh

ARNE . . Ip wa, Hs tee de IRABUTEd cu DY s..EEVETEL. methods -or Cas bo DOR 0 oo {1 ‘ i, ' : * J bi 4 ‘ Co . I '§ . approximated by the following equation, which i i

VL

" ‘ 5 |§ i aa 15 assumes model pores of cylindrical geometry: q i g - i (1) AFD (A) = 42,088 x PV (cc/am), y

I. sa (M%/gm) - + Where PV is pore volumme (measured in cubic ; 24 centimeters per gram) and SA is surface area (measured in square meters per gram). 1 i

E - 27 -

KX : i i

Rl i £ } a 3 i - ]

B I

1 Co] 4 q i i pa ye b 3 Pars: i : Ft ! 4 . ge I

Ln he

TE : “ . i i i A §i BY. , ; Hl fi Both nitrogen and mercury porosimetry may k -

TE be used to measure pore volume in Xerogels, : 3 precipitated eilicas and dislytic silicas. - Pore volume may be measured by the nitrogen

A I

Eo 5 Brunauer-Enmett-Teller ("B-E-T") method i po - described in Brinausr et al., J. Am. Chen. 3 i pj: 4 Soc., Vol 6@, bv. 3@9 (1838). This method

VEG i

OE

3 3 depends on the condensation of nitrogen -into !

E: | : 3 the pores of activated silica and is useful for v i “a 1 measuring pores with diameters up to about y 621A If the sample contains pores With ’ i k 3

NH | diamters greater than about B@PA, the pore size ¥ distribution, at least of the larger pores, ig 5 : I o determined by mercury porosimetry as described - \ 2 38 . TU ue AMER ed cde Y wr HN UE Re RITESE et AL Ind. Eng. Chem. Anal. Ed. H : ih J it 17,787 (1945). This wethod is baged on

T H

Ty determining the pressure required to force mercury into the pores of the sample, Mercury ] . 8 i porosimetry, which is useful from about 3p to - 28 about 12,000 A, may be used alone for measuring 3 pure volumes in gilicas having pores with diameters both above and below 6DA .

Alternatively, nltrogen porogimetry can be used ! } : ; ¥ 1 : - £8 - i : : p

-_— TT EE ———— ee " n wu BR 3 HI .

A BL a. 9 bi ! ) E-. pi. : i yO i

NE iH 4g = 1

A i.

TA. po N ‘IN p E x . p 7H 9 o 1” I a. IB px. lg i a

E: in conjunction with mercury porosimetry for 3

Co. Ei

E18 bi

Ro these silicas. For measurement of APDs below

OC pl

HE J. i: 4 : 601A, it may be desired to compare the results gE -

I. i: . wl obtained by both methods. The calculated PV J iE 3 5 volume 1s used in Equation (1). - il -

Be 4 . i . 3 ig aE ‘I. sk : . wl For determining pore volume of hydrogels, a } ‘I a in } # . different procedure, which assumes a direct =F '-

Ee i . relationship between pore volume and water i. i ;

SE content, ie used. A sample of the hydrogel is i.

SEER :

CE RE i wry a 19 weighed into a container and all water is ! k

Lr HA, bo ete 110 Ua oh vA 0 gh ge deat Wire ae eh dead Td Cte, Catenin mel he eg f

LA ie} CAR te tyanioved PRSH the” shnple by vacuum at’ low. iy, temperatures (l1.e., about room temperature). {I . oF - i The sample is then heated to about 4508 to TPF . (239° 379°C) to activate. After activation, 1 the sample is re-weoghed to determine the weight of the silica on a dry basis, and the a. pore volume is calculated by the equation: ] (2) PV (co/gm) = % TV I eee meee 1 80 - % TV : 1 21 where TV is total volatiles, determined by the i

BR i , g § ; ¥ I ¥ ; ; : 8 4 I ho - 29 - iF jo . ; : :

vA la = —— ’ J a ; fu i

ANE

, ingle a fi

CTL i" i#H Eee i 14 i 0 t

HL Ch

TE § 2 p: ' ve

Het and dry Wzight differential. An

FE 3p ,

AE alternative method of seleslatiog TV is to a. . measure weight lose on ignition at 175@°F i. (955°C), (see Equation (9) in Example II). The “E oF 5 FY value calculated in this manner is then used § in Equation (1). J : The surface area measurement in the APD €quation is measured bythe nitrogen B-E-T i fF surface area method, described in the Brunauer ] 19 et al., article, gupra. The surface area of ! gf . all types of appropriately activated amorphous

Le ’ J

Ts 5% ) . !

SE : : gllicas can be measured by this method. the re 3 !

TE : rR measured SA is used in Equation (1) with the ! Hori : 5 i gis fy Wh Ve eg cn ER pe mary me gp PERE roe EL w ey . } aR i measured FV to calenlate the APD of the silica, i i Hi f 15 The purity of the amorphous silica used in : . : this Invention is not believed to be critical j in terma of the adsorption of soaps andphospholipids. However, where the finished products are intended to be food grade oils 2 care should be taken to ensure that the silica used dees not contain leachable impurities - 3a - ix 1

WE : A 2 1 { 4 bi ! ;

EC

3 Pr ’ "I

TO ;

FN = ee i ke 42K an ‘ 3 fn }

Ep ! , Bhi LU. ‘I. :, HO 9 3 1 ba

I ER

A. : ’ > h

SE La - bl

J which could compromise the desired purity of Li"

EE the product(s). It is preferred, therefore, to a

A zs pr B hy

E23 use a substantially pure smorphous esllica, . 2

Rg. 3 although winor amounts, i.e., less than about Ed

Ta. b &

Eo: A a ! . ~ RB

EE 5 19%, of other inorganic constituents may be i. <4 |i ils . hy

Ee present. For exanple, sultable sillicas may Eo 1 vie

LE ’ i

CE . . : Jb

SE comprise iron as Feqgls, aluminum as AlsO4, kL } gi : titanium as Ti0g, calcium as Cal, sodium as ££ ; Ei -R , Na,0, zirconium as Zrl,, and/or trace elements. BG 3 2 RA

E Bg 1 LIE f 10 The examples which follow are given for B 4 rr JE & illustrative purpoges and are not meant to AE 5 ' Bi is

LR Ra 1% “t limit the invention described herein. The EE it : . : ! a yf . oor sail + following abbreviations have been used 1 ve 1h hb Be FE ASU Ee tr Se pe LC . . JT deer bn ped Cg i] EN f gd throughout in describing the invention: CI aE » FE bs of Be oy I of 15 A - Angstrom(s) 1 i A ro ' AFD - average pore diameter Eo > ) oe i oi: , B-E-T - Brunauer-Emmett-teller 2 ~ 5

C - gapaclity i!

Re” : RE

Ca - calcium _ ag 21 oC - cubic centimeter(s) ne

A. ; cm ~- centimeter EE

Cu - copper : . i ! "1 ! 1 li 4 - 31 - :

AH i i 3 : i i ] ki :

Sr — - eee eee ee ——————————e y/ E 3 RE i il WN 3 ge. bi hi . Ba ! JES © po HB : ey i. a A 17 Rj!

Fa g 3 gh {I

I - 3 “BR 2 1 ea = i EY CE

Cg E oc - degrees Centigrade Bi ur bf i fr: i 3 pe db - dry basis E 1 4 x 3 Rn ; | her

SI CF - degrees Fahrenheit 3 o a 2 ! i i: 7 Fe - Iron .

H ;

Wo pr 4 : 8 gn ~ gram(s) E 3 be BE ‘EE ICFP - Inductively Coupled Plasma Ee

Ef E

A m - meter i

Ld 3 1 | % - Mg ~ magnesium EE

Ei k:! . a NK: ‘ min ~- minutes BE

LAS \ Be 3 g on 10 : ml - milliliter(s) B

A }

I Ei

Eo 3 P - phosphorus i o PL - phospholipids B

Ie R | =

A E: ppm ~ parts per million (by weight) gd ge B

TRL BREET | oC TERE RE pF im Rove volume To ) f ix ht ky or . By 2, 15 % - percent a - IE

Ir RH - relative humidity BC > J ‘- - rpm - revolutions per minute p q . = 5 - A0aps '. . i

I Sa - surface area / . : 20 sec - seconds

TV - total volatiles

WL - weight 1 i — 3 jo s B 1 i 1 3 3 i ] i 3 2 3] ¥ 3 i

4 +i & 2 4 i, lt bi ’ Es

PT 8 ho) N.. ' ‘S = y A 3 Pp (Anorphons S51lica 011 Samples)

EE 4 ¥ p The properties of the amorphous silica used ig 1 in these examples are listed in Table II. - 5 TABLE 11 3 fb Silica surface Fore Av. Pore Total

CAE Sample Ares Yo lume? Dizsmeter Yolati lest

SF Hydrogel” 911 1.8 80 64.5

HE -—

SE - 2 3 i 1- B-E-T Surface Area (5A) measured as tg -. described above. . : a. z2- Pore Volume (PV) measured as described , Ho dw ag nibh, Ae nd Sp EL PST PIG 4g whet ea pedi : : : ee Tas A rt

Re ISR vedi Loe a HERET UB THE” hydrogel method. Te ol Co Co 3 3- Average Pore Diameter (APD) calculated as i

AE 15 described above. 3 4- Total volatiles, in weight percent (wt.%) on ignition at 175@0°F. (995°C). 5- The hydrogel was obtained from the

Davison Chemical division of W. R. Grace ad & Co. i ~- 33 - 4 4 i

- iH

LE pa aE 4 E The 0il1 Samples used in the following eo

A y examples were prepared by combining 0il A (see - Table III), a caustic refined soybean oil i. sampled after caustic treatment and primary

E | 5 centrifuge but before water wash, with elther x 0il Sample Eor 0il Sample E° degummed soybean ¥ oils prepared as described below and not E x ¥ subjected to caustic treatmment. 0il Sample E

CH was prepared in the same manner as 011 Sample E y 1 of Table III, for which analytical results are shown; Insufficient quantities of 0ill Sample : E° precluded separate analysis, but it is 1 "assumed that the identically degummed oils were ” : substantially identical. 0i1 Sample A of 15 contained large quantities of soaps (362 ppm) {i gli... dsternined by measuring alkalinity expressed ss oo q godium oleate (ppm) by A.0.C.S. Recommended ) fl ) Practice Cc 17-79. The acid degummed oils, 3 : having not been contacted with caustic, contained no soap, but contained significant levals of phosphorus, as indicated by the A values for 0il Samplle E, which contained 22.0 i ppm phosphorus, measured by inductively coupled 8 f - 34 -

-— ees ” ] ] J 3 8 3 | oo : s Eo! : . 3 po . } bf 4 ys ]

F Fore By ] i - : : KR 3 28 : J i 2 . - CE 1 : plasma ("ICP") emigsion spectroscopy. £ y I B, 3 § gi,

J. E - E 011 Sample A was mixed in varying A ht : R: ks t : proportions (as indicated in Table III) with E

J . vo 0il Sample e or E° to prepare 0il Samples B, C 3

By R-: ' E!

EE. 5 and D, which are relatively constant dor gi 1 h, iN: pr x phosphorus and associated metal ions but which 5. rl: . B 1 contain significantly different levels of 3

HE . i 3 soaps. 0il Sample B contained 75% (011 Sample A - ¥ and 206% 0il Sample E. 011 Sample C contained a i: : ia 12 5% 0111 Sample A and 50% 0il Sample E°. (il | BE ’ I : } Do -, , Sample D contained 25% 011 Sample a and 75% Oil g

Eon oh E sun ny av pita BEBRLE Ela opBach.01d. Sample wag analyzed as. Cl a a: iE

A described above for trace contaminants (P, Ca, ¥ fo i i x Mg, Fe and Cu) and for scaps. the results are : 1 r 3 15 shown in Table III. LL .

RN

. a

The acid degummed oils (0il Samples E and "

E’) were prepared by heating 508.0 am oil, i covered with foil and blanketed with nitrogen, gl in a 48°C water bath. Next, 500 ppm 85% phosphoric acid (8.25 gm) was added to the oil and stirred for twenty minutes while i ’ : I i Ji o g - 35 - !

I fi

Ek 2 9 5 JB 58 a. i i X: , »

Ek: pl J ii p. J

TA i

A. 5

E 1 maintaining the nitrogen blanket. Ten i x milliliters of de-ionized water was added and i 4 X I

A mixed for one hour. The sample was centrifuged k ; at 2300 rpm for thirty minutes. The top layer I

R a i } 1 o was the degummed o0ill used in the experiment 4

BE

(the bottom layer, comprising the gums, was 8 ‘ ™ 4 discarded). pe 3 i:

B , )

I

Bs 1

EE inl bed RE a an Hehe Re ae Te Wg, RETR EAL hs "3 F : il : d 1 o 3 CH 3 ? : & 3 ¥ 3 ] RG : { \ f ] i i 1 1 - 35 - ‘3 ' E fe oe TTT ‘ : - cep A b 4 33 iE ‘ 4 : io - I , :

LA RR $7 . , [ i ve SY i Busi fi 4 ht

CR | 1 3 9 F - 19 - | 5 “i TARLE IIT $e i | SS bi 1

Lh { iy 0il Trace Contaminants m! 3 : ; i Sample P Ca Mg Fe® Cu Soap, ppm §

F A 13.4 0.93 1.03 0.02 o¢.02- ° | 362.0 x 5 B 19.4 2.08 1.92 0.00 0.02 180.0

Cc 20.8 3.04 2.46 0.06 0.01 70.0 . D 23.7 3.84 3.01 0.07 0.02 30.0 J ;

E 22.9 4.27 3.17 0.11 0.03 0.0

E' * * * * * * ; i 10 1- Trace contaminant levels (P, ca, Mg, Fe, Cu) measured i" in parts per million by ICP emission Spectroscopy. Sh I ) 2~ Fe and Cu values reported were near the detection Cd ] limits of this analytical technique. Ly 1K 3- Soap measured by a.o0.c.s. Recommended Practice Cc Cl i 1. 15 17-79, Cl " oF *- Oil Sample E' was prepared from the same crude oil as Cl B ii 0il Sample E, and by identical acig degumming Steps. Ce = smaiNfUfficient quantities. of 0il Sample E' were’ avaijabié' Po :

Cr A CdeinAag o8 analysis, but it is assumed that the values are » comparable to those of 0il Sample E, iy B . . j i oH | B 1 EXAMPLE IT | kK 3 (Treatment Of 0il Samples With Silica) : i b 3 The Oil Samples Prepared in Example I were treated vith B® ; the amorphous silica described in Example I. For each test | 8 ] 25a 100.0 gm quantity of the 0i1 Sample (a, B, C, D, or E}) was i i 5 heated at 100°C, and the silica was added in the amount. Ve B - indicated in Table IV. The mixture was maintained at 100°C, gE while being stirred vigorously, for 0.5 hours. The silica B was separated from the oj) by filtration. The treated, 1 ] 30filtered 0i1 Samples were analyzed for Scap and trace ih ; contaminant levels by the methods described in Example I.

The results, shown in Table IV, indicate that; | i fq 0s | 1 j 1. The amorphous silica adsorbent removed soaps and i trace contaminants (phospholipids and associated metal ions) 1 35 from the 0il Samples in a single operation. wl ’ B EB i

Candle ] pte Bi RE } rR i el 8 i A § 3 RR i .

CRE RU ER

\ } ) ’ . Lo : og $i i NT oy i I ro K: . 13 E ’ fl pi A i FF k - 3 i

Ky: gE - a. a. Soaps appeared to be preferentially

SO be adsorbed ag compared to trace contaminants. [Ip

RE x Many cases there were po soaps found in the i 8ilics treated oil, while there were

E S considerable trace contaminants remaining in : the oil. 3. The capacity of the silica adsorbent for phosphorous appeared to increase with increasing Soap levels in the 0il Samplles. 3 12 For example, ip Oil Sample a (362 ppm soap), a ie . i . . cet

Cs J . UT . ye : *

Eoglh wins Bllden- doadingsofuonly B.15 wt.% was Yequired

Srl en pl ut Ai 05 dhl eg dli nd TN ’ - y to reduce the phosphorus level to welll below y 1.@ ppm, while in 013 Samples C, D and E (7a, i 30 and © ppm Seap, respectively) silica i ; 15 loadings orf 9.5 Wt.% were required to reduce phosphorus levels to below 1.9 ppm. The bresence of soaps in the 0il therefore made it 4 Possible to reduce Phosphorus levels to below 1.9 ppm at a much lower silica Usage than wag { revwuired in the abuence of B0aps. ! ; ] The data obtained from Example II if i i ] if } i - 38 - 1 u

SE — es

EE i :

E § demonstrate that the capacity of amorphous i L- silica for phospholipid and soap removal fg actually increases with increasing soap content 1 of the starting oll until a maximum adsorbent I x 8 capacity is approached. The maximumm adsorbent

Be capacity of the silica hydrogel used under the i

A conditions of Example II was approximately §5 - Wt.% soaps plus phospholipids. i x 1 I

Bi The data in Table V were calculated from F : 19 Table IV inorder to obtain valnec for the ’ ig . adsorption capacity of the ammorphous gillica, | {

I» i i Calculations Were made as follows. The 8 oil come pra yp hE BUOTEROUS eilics for combined 3

A BOaps and phospholipids (Cg_pr). expressed ae a 8 i - 16 percent, can be defined ag: I 1 (3) Cop, = [ 8 (prm) + FL (ppm) 9 E —_— X18 FE 1 Silica (db, wt.%) I where the change in concentrations of soaps and iB

Zi phospholipids in the oil (from before to after i contact with the silica adsorbent) are defined fi

J '.

Ii

Ji

Ah - 39 - ; : | i

A l

- --..s TT EE 3 fv ARE - Bride

Le ik 1 i ick ‘ : ! ' DRNTES tid : p. fh 3 i : i; Hl a

R et VUE Ee a. Ep fT Age

TE. as: 3 ; (4) (ppm) = 5 (PW) ypj4iq1 - 5 (PPM) ginal fe E: . , - (5) PL (ppm) = F (ppmu) x 30 . : i 3 (6) FP (ppm) = FP (PPO) jq54441 - F (PPO) ping) I

Ea To af f > a (7) Silica (db, wt.%) = Silica (db, gm) yo ik x 190 oH - a I . ] " 0il (am) NEE ok : i 0 Co | { a where "Silica (db, gm)” is the weight in grams ok of the silica after ignition at 175@°F. if - od 19 (995°C). 4 ] ) } ] : : (8) Silica (db, wt.%) = Silica as is, gm) x 100 - TV 100 ] (3) TV = 100 x Silica (as is, gm) - Silica (db, gm) j -

Silica (as is, am) 1}

The capacity of the amorphous silica for il ] Fb 1 IE Fl 4 B h on 4 85 kB ETE EE 3 i IN § { . k . , 1 ] {

- 4 b phospholipids alone (Cpr) expressed as Aa

Eo percent, can be defined as:

A. * (189) Cpp, = FLotepm) 10-2 § | 5 Silica (db, Wb.%) : The calculated values for changes in phosphorus (P), phospholipide (PL) and soap : (3), combined phospholipid and soap (5-PL) remaining in the oll, capacity for combined - ) 1 soap and phospholipid (%Cg_pp) are given in $ Table V for each of the treated 011 Samples, 4 along with starting phosphorus and soap values. nth, mittee s 0 The ddta” from Table V were plotted in FIGURE © 1 © Co in the form of sdosrption isotherms, with the wt.% phospholipids and scapes adsorbed on the silica ( o-FL) plotted on the ordinate versus the amount of scap and phospholipid remaining in the adesorbent-treated oil (Remaining S-FL) plotted on the abecisea. The data were plotted 218 in this manner in order to correct for the phenomena typleally observed for adsorption of increasing capacity (up to some plateau value ; - 41 - ! .

J

#

A

) 3 FL . 3 ih v ' i t | :

A. Ch

AE as a result of saturation) with increasing - 3 Eo adgorbate remaining in the treated material. 3 5 This phenomenon is predicted from equilibrium i

Ek considerations.

B : 4 5 The data from Table V were also plotted in b ! 1 ° +

Ko . i “3 FIGURE 2 in the form of adsorption 13 1

H : isotherms,with the wt.% phospholipids adsorbed . : " on the silica ( PL) plotted in the ordinate i ¥ versus the amount of phosphorus remaining in : 10 the adsorbent-treated oil (FP) plotted on the ” § . fp ' ! k abscissa. FIG 2. shows data for adsorbent- od treated 0il Samples with < 32 ppm residual 4

Te

JAN . eo: a i soaps. :

Ce gC gS af i ede ee SL a. Cy Co v i ! , -

Ho The data from Table V and Figures 1 and 2 y 15 indicate the following: ; 1. The capacity of the silica for

GQ rhospholipid and soaps tends to increasing levels of soap in Lhe starting oil. : f 2. Increasing soap content on the silica ) 200 tends to increase the phospholipid capacity of Ny ! 1

I

: 43 -

A fl ‘ ] 1 I i

J

RE Eee a i: iN py: 3 ps 4d . “i Kg: - “al cL 4 Cy

EB the silica when the soap content in the treated : - + oil has been significantly reduced for example,

Wi 4 in this cage, about 30 ppm 5008p, aE bi demmonstrated in Table V and Figure 2, for

S o these O11 Samples and this adsorbent. : ); ! . i:

Jt : Ae dB ue eR Corn rt TET . i . : . H :

iH ig “i Bb i

A. ’ 1. »

Ly

BL - 23 - ‘ il 0 14 : . E

E | . ’

A + ON oO <r : o [ed i ~| oN —= AC tn wn c

He . . . . . . .

OU NON ONO OO no Co Cc Oo o coc oo OOOO

J Cc oo << ~ mM oN mM ? ‘4 wl om — — 3 . f aN aN 0 — O~N

OT <r << WOO — C0 UN NOSSO Mn oN WN Ww OCT MO

Nn —— ON oN << ON ON ~~ ON ~~ — = Tot 4 a COCO OO O-~OO coc oco ooo oocCco

QO . . . . . . . . 0» . . . . . . . . . . . . » ‘ . Cow . — ~~ in ™ : aw oI oC wo mo 0 nO I~ oO ~~

RE ~—eoN N QO Ww ~ NNEC O eo oo aU COoOO0COCm ooo coco coo w -~oMOo 4 «es 0» . . oe 5» *« + = 0» * eo 0

E « wu oN ~~ MC OM NON ~~ On oh — r= OWN ~~ o Mm a CT MONG A= Mmm T MON oncom —- 00 WU > CT «se COCO NT — +s a f~M so 2M ss oC :

Ol 2] Ame ver He ee NHL NN re MN ;

Sd 2. a ! vm ' { 1-1 s ~ ~ aarA Or 000 ND LMR O << ONC 00 MC ~~ s L cle ON ON WON Oo <UL OO XD 0 Oh oO noo ¥ ‘3m no se COO CT « eo oN s 2 eM «eo 2 OO i 5 x UO] ©dr os 0 tN 1 ss MOH «oo MON + TON Rh . 3 ’ ’ § 1.0 i } y v [3

CF : . op de nfl. BL dabei emi hee Erp dee ee cr . i . © NOAM ON TMOM~ CHO NHN OHO rR . ! ~ so ox ~~ O sc I~ On + — \O OC « sfi~I~ « oPf~em / = Mmm «ONO OV eo oN © oe o> Mr «OQ NN eM I : el — rd OY es ee — oe oN Ne Nr» Nr 1

I ) i” i 0g 4 bi ] b [i] i oO ord i : . - 1} w :

Co oe www o «<n Us} un nN A . CO rw MY orm — MO — MD ~ tN \D co + « oo eo 1 « eo » i os i « oo ) « os a : = | COO OO | OOO {CoC loo oOo COCO : i al + ord ! o CCR Lom coo oR a ERR li 3 nol ub

En a.

Ef t -— Jd - 1

<i Ag. ? no . 5 . do - ‘ Ll i

I 3 : 1 : jr in

Ce Cre oe . N LH,

A - . pet bo le cB . - 24 - bE oo : Wo call - Cd

Clu — REE a Ca Nc OT < -O ~N Cn CM Lo

A 1 « eo sb 1 . =e i +. eo. 1 ss 0 I + es . i in v | HOMO I Ino | ~wo 1 nw | — Oo SoH 3 ra — ON Won ON oN = Ne ; 9 1 3 mn Nuno mI oe low INO © | © - 3 <3 1 s eo so se i « I +s «0» i o eo 0 1 » + 2 ; i wn | I~ On Mean no ~~ — wo " x nth wn ON oe ~~ mo oN — oN : ¥ O : . . g 2 . 4 DE ; = : ~ ll CNoO we Ino I Mee | NON Immo ‘ p ol | vec 1 <n | — oc 1 mo lwo i © 3 me i ~ mo moo ! 1. EH [| { mw; i : i — Co i

Ig 4 jen amm — . ;

AE] Tnoam oo sant AND NAHMNN Ono ‘ V Hl . HAD eC tO oO ~ « oP o t= t th en hf OND Oh sHN © «ev ae ee IN : Fi — — rd OY + oo — . oO oN oO. ’ i i i Hl \ ! <c or nN @ nn nn nN i

Ww fcoHmMmw 1o~=™m | = ND 1 Me — ow I

J ee see § eo so I so 0 {| +» os I a eo H i op coccoo coc coo coc joocc h i y ] 3 , p . : I ; —_ i ; . A ! 2 CREE Lama VOLO cocoa KERR . h i }

Co i. 1 k 1 - ‘ . i

H

{! : i

Claims

- st 3 i 3 3 if ir Sr i i i I; 1 : ;

. i x! . ji; : “ied ho : - | \ 1 pf it If 5 ] 0 WE CLAIM: 1 3 B 1. A process for the removal of soaps and i of {

1.

a. phospholipids together with assoclated metal V - - ions from a caustic-treated or caustic- refined EB i a. 5 glyeceride oil which contains soaps and ! = } _ phospholipids which comprises contacting said : if :

i.

i. oll with amorphous silica, allowing said soaps in and phospholipids to be adsorbed onto the A 3 ! bi amorphous silica, and separating the adsorbent- A 18 treated oil from the adsorbent. ! f, | : 1 :

4 2. The process of claim 1 in which the ] A F said glyceride oil is soybean oill. i i t : erie Mtg bens oe co ya Gaon s The process of claim 1 in whieh said cere i caustic-treated or caustic-refined glyceride I 4 15 oil comprises at least about &@ parts per ES million scaps and at least 15 parts per million R phosphorus, }

4. The process of claim 1 wherein the soap content of the glyceride oil 1s reduced to - 47 - 3 ee — a

- i we i.

Ta. i ki a : i ig i ol ol f x 2 “a Co Ci Co

PL . So | i

T. ; 3 i 2 below about 5@ parts per million, and the § CL phosphorus content Lo below about 15 parts per i x ; 4 million. ] 3 4 ! 2 gE 7 5. The process of claimm 4 wherein the : EB ! I i! o- 5 soap content of the glyceride oll is reduced to i iE below about 18 parts per million and the

(. . 3 in yp ! i phosphorus content to below about 5 parts oer I million. . , 08 & il 6. The process of claim 5 wherein the soap te ayrrho MR Rega Tp Tree ee ny Co x 19 content of the glyceride oil 1s reduced © to Sree "i aubstantially zero parts per million, and the ‘ 4 phosphorus content to below about 1 part per 7 million. ’ -t

T. The process of claim 1 in which at least 50% of the pore volume of said silica is contained in pores of at least 60 Angstrom in diameter.

8. The process of claim 1 in which sald + A amorphous silica has an artificial pore network

4 ie veil : ir * ; 4 * ot A Ty CG Col bg 3 i a. ¥ RN Po kt 2 a of interparticle voids having diameters of 3 EL about 68 to about LPP Angstroms. Ai, ain CE 9. The process of claim 1 in which said ig Li amorphous 6ilica 1s gelected from the @droup . : a5 5 ° consisting of silica gels, precipitated B x gilicas, dialytic cilicas, and fumed sillicas. i) ¥ | | “

- . 14. The process of claim 8 in which said 1 gilica gel is a hydrogel. th, aS Apel i CIMA mh ae dp re AT et itn So 1 11. The process of claim 1 if which gaya eo or ’ 1 10 amorphous eilica has a surface are of about 1 199 to 1200 square meters per gram. : 12. The process of claim 1 in which sald § : caus bic-treated or ocaunstic-refined oil is ¢ k ® contacted with amorphous silica at a silica usage of 0.01 vo 1 welght percent. / 13. The process of claim 1 in which the silica-treated glyceride oil 1s treated wlth bleaching earth. —- 49 — - Co po _ oo Ce Sau > al