Document 2RxQemzOZjMZ929dr3GLkxmg

/ 753698 1 THSRWL 8a s n j T r of p i a j c i m c r u m s L` Mondeleorn Vestinghouse E lectric Corporation - Besearch Laboratories Pittsburgh, Pennsylvania 15235 ABSTRACT The thermal s t a b i l ity o f d ie le c tric flu id s la e le c tr ic a l equlpeest le gw w rally governed ty reaction* with other material* ead with A purltias th a t cum ot he avoided H A eral o ils ere suseapt i t l e to a e te l catalysed oxidative degradation! Involving a peroxide Intermediate. Suitable lohib- degradation as a function of temperature are obtained to determine i t s useful lif e under various operating conditions *s These ra te s are generally based on a simple c rite rio n such as single physical, cbealeel or d ie lectric change. F in a lly , d etailed stu d ies a re oade to determine lte r s can he need to decraase the rate* of thl* the sig n ific a n t chemical reactions and th e ir rates process* Both oxidative aad pyrolytic degradation* th a t occur during degradation under various con arc considered fo r silicone d ie le c tric fluid*. Askarels are b a sic a lly sta b le therm ally where they d itio n s . Only mineral o i l s , of the fluids d is cusoed here, have been subjected to a l l these a n presently used, provided.tiny are properly purified aad other M aterials present do not react stages o f in v e stig a tio n . The present statu s of knowledge of th e ra a l degradation of th e other w ith then* Thermal rating* o f a o lfn r hexafluoride flu id s is r e la tiv e ly meager. depend upon m e t i o s s o f the gas aad i t s e le c tric discharge products with the so lid s l a a device) aad Fur the Boat p a rt, in te re s t In the aging of aa adsorber can ha used to m o v e reactive products* s d ie le c tric flu id i s concerned with degradation e ffe c ts of one or more o ther m aterials to which m n p p ucnoK I t Is exposed in an e le c tr ic a l device. Sample* of these m aterials a re w ater, oxygen, Metals, solid This paper discusses the th eraal s ta b ility In su la to rs, and dissolved Im purities ouch as of acne d ie le c tric flu id s with regard to th e ir ion lia b le or re a c tiv e compounds and s a lts of metals. p x aetieal ose* These flu id s a re Mineral o ils , Gome re se t d ire c tly with th e d ie le c tric , as oxygen s tu c c o e s , sta re la and th e 91s , su lfu r h e * fa r e v p le , and o th ers, such ns copper s a lts , aay fln e rid e , whose operating tenperstores are in catalyse degradation. in contrast to ryrolytle th e range of 10O#C nod higher* Continuous demutde degradation of n d ie le c tr ic f lu id , where only i t s t o Increase th e operating te mper atu res of e le c tri own molecules a re involved In the reaction, these c a l equipment has generated considerable e ffo rt on unavoidable reac ta n ts generally cause the th eraal th is su b je c t. A general view of i t s significance s ta b ility to decrease. Therefore, mast aging and rn a lflc a tio n s la flrart pr esented here, in the studies o f d ie le c tr ic flu id s do not include Introduction. Shis w ill serve as the bssla for pyrolysis as a single A cto r. dealing subsequently with the various individual dielectric*. A ch aracteristic of o degradation process of a d ie le c tric flu id Is th a t th i s le often due to Knowledge o f th e ra a l s ta b ility characteris only a r a ll amounts o f reac ta n t and ensuing prod* . tic s , specifically aging, of d ielectrics deter- u cts- The reactin g m ateria la .noted above are nlne how Individual ones nay be used In various generally pr esent in low concentration A the ap p licatio n s. Where one my not be used, end Menus flu id o r a vary sm all p a r t o f i t la exposed to by which a d ie le c tric la Improved- The e sse n tia l them aa in th e case of a reactin g m etal. Degra c rite rio n fo r a c c e p ta b ility on th e basis of th er- dation products a t concentrations o f p arts per s a l s ta b ility is th a t th e physical and d ie le c tric B illio n have been found t o cause substantial propert i es ere not impaired by chealc a l reactions Increases A th e d issip a tio n A c to r of d ie le c tric s. ,1 i during reasonably long periods of operation a t th e expected saxlsna tem perature. Aging could* e re tle c , in A c t , sometimes enforce s ccopraaise Example* of snoh lsp o M tla s are hydrolysable chlorine A chlorinated diphenyls^ and carbonyls* in transform er o i l . This A c t , th a t reactants and of electrical or o f physical properties In choosing e d ie le c tric , i t I s noted th at e le c tric products involved A the degradation of a dielectric fluid are A s n a il amounts, presents certs A discharge e ffe c ts should not be confused with problems A pinpointing causes of degradation. It' theraal effects, since both any occur during a functional aging te s t of a d ie le c tric . way be d if f ic u lt to determine th e nature of these e n te rta in , or even whether the f A id i t s e l f mas degraded. - The source o f th e deleterious effects investigations o f th e ra a l aging of a dielec any be decomposition or reactio n products of tr ic flu id usually follow th is sequence. an Impurity, where the d ie le c tric i t s e l f Is not F i r s t , conditions, esp ecially temperature lev els, involved < Likewise, problems a re ancountered A th n t cause s lg n lfic s n t degradation are established. attem pting to elu cid ate e ffe c ts which may be due This in to n a tio n i s often su ffic ie n t to decide on to very sm all amounts o f c a ta ly s t. the use of a specific d ie le c tr ic . There are n rasnber o f A6TX t e s tr v b le h , with sooe sa lific a tio n s However, once a d eleterio u s reactant and/or vfatpe necessary, any ierve t h i s purpose. As product le known, tbsn o ften advantage can be in te re st In a d ie le c tric grows, lta xttas of taken of I t s s n a il concentration to Improve the 11 NPC00026362 753699 thermal e ta b illty of the flu id , nil* la tehieved by us* o f on additive or suitable p u rification, or both . A 1 1 amount of a aubstance th a t specif ic a lly re se ts v itb a harmful reactant can be added t o r emove I t or change I t to a re la tiv e ly Innocuous fe rn , there la some leeway in choosing an ad d itiv e aa I t need not have as good d ie le c tric properties oa the fluid I ts e lf because i t la used in lov concentration. An additive t i n t rcncte with a reactan t la known as an in h ib ito r, and one reacting v itb tbe products la called a scavenger, the purification process of a d ielectric fluid lc often designed to remove m aterial th a t are capa ble of degrading easily, v e il as those that impart poor d ie le c tric pr opertie s to I t In th e ir original f a n . She fhetors contributing t o th e thermal stab ility of n dielectric fluid in an electrical device can be conveniently s u n a rlse d under two main categ o ries. One deals v itb degradation e ffe c ts and th e other with means to upgrade th e them e 1 s ta b ility . Degradation effects are related to t a . She reaction of the flu id v itb e n teric Is Area vbich i t cannot be iso la te d , such as oxygen and m etals, sod b . C atalytic effects due to certain impuri ties* increased thermal s ta b ility i s achieved by: a Purification. b . in h ib ito rs , and c . Scavengers iKDivsroyL piEUBgngcs Mineral o il S pecially p u rified and fractionated mineral o ils are uaed extensively in transform ers, cables, cap acito rs, and c ir c u it breakers o f high power r a tin g s . Their tssxlxiai operating temperature fo r continuous service here they do so t have nevere re a c tions with th e other components of the device has usually been rated a t ID5* . These o ils are mixtures Vhose major co nstituents e re a lip h a tic hydrocarbons, arcmatlce and naphthenes. Much a tte n tio n has been devoted to th e ir thermal s ta b ility to assure prolonged function of the equipmen t in which they are used. There are a number of co n flictin g conclusions shoot Influences on t h e i r th e ra a l s ta b i l i t y . This occurred probably because of differences in composition, Im purities, fractio n atio n and p u rificatio n o f th e o ils studied, and use of various c rite ria for theraal s ta b ility . However, a t th is p o in t, acme o f the e ffe c ts o f degradation o f these o ils In e le c tric a l equipment a s v a il aa a number o f tbe facto rs th a t a ffe e t th is process are fa irly wall understood. This is based on studies o f transform er o i l , which w ill be emphasized here. Thermal s ta b ility considerations o f u mineral e l l Invariably deal with i t s reaction with oxygen and th e ca ta ly sis of th is reaction with copper and iro n . Here, moisture is p articu larly harmful, and g reet car* is token to exclude i t . Exposure to copper and Iron cannot be avoided In most e le c tric s ) devices. The f in a l products of th is reaction are o d d , sludges (insoluble acids and aoept>}f5 mater and carbon dioxide. Their deleterious effects Include lowering tbe in su latin g q u a lity of the o il, deterio ratio n of paper or preasboard b arriers and corr osion of conductors and casin g Sludge can a lso obntnict flow of the o il th a t i s necessary to d issip a te b e a t.4 There are ASTM te s ts fo r acid* and fo r sludge" formation in tran sfo rm er.o il. Besides th e ir use In quality control, these te sts have been the experimental basis fo r various thcxnml s ta b ility studies o f mineral o il. The discussion of the degradation process, here, I s based p rin cip ally on s review and two specific studies of the degradation of mineral e l l d ie le c tr ic s . The review Is the proceedings of a eysposiua os in su latin g o ils th a t was held Id 1958 a t the In s titu tio n o f E le c tric a l Engineers under the chairmanship of Goesling." One of the studies wes made by Melchlore and M ills,1**0 who. mode carefu l analysis o f interm ediate products of the degradation and a ls o controlled th e compo s itio n o f tbe o i l . The o ther la by Teal, S tuart ft Boaal3 who emphasized d e ta ils o f o i l compos itio n in developing th e ir conclusions. Tbe degradation .of a m ineral o i l , according to caygen uptake, generally pr oceed k lo e tlc a lly as follows ." #9 f i r s t th e re la aa induction period, n*xt the r a te becomes n u to cataly tlc, then I t i s constant, and f in a lly th e ra te of oxygen uptake decreases. The d eleterio u s e ffe c ts In th e o i l become evident In th e la s t two stages. The in h ib itio n stage la diminished by tb e presence soluble copper o r iro n .6 9 a meaningful mechanism of th is process must deal e ffe c tiv ely with th e auto co taly tic phenomenon. The co une o f the degradation o f the o i l has been shown to be associated w ith the production end concentration o f s hydroperoxide Interm ediate I t Is the basin o f the following scries of free ra d ic a l reactions th a t have been proposed. Tbe o i l Is designated as RH, end B*, ROg* and BO* are free radicals. BH + 02-k R* 3Dg* F- 02 -w H02* ROg* * BH-* BOOH ' R* 2-- BOOH (1 ) (2) (3) (t) These reactions would sc count f o r the in h ib itio n atoge. The n u to cataly tlc stage th a t follows Is explained by a number o f re a c tio n s, where the hydroperoxide decomposes to regenerate free radicals sod thereby propagates a free rad ical chain. One such reactio n could be s unlaolecular decomposition o f hydroperoxide J 3 BOOH-- R* OH* . ( 5) However the evidence p oints to a c a to ly tle o lly Induced frtS? ra d lc o l chain, with copper or Iron NPC00026363 753700 0 i .i, ,rii .Si&i c n ta ly st. I t fcac been proposed^7 ^ th a t soluble copper ions react as follows with hydroperoxide. POOH-* Cu*S*+ Cu*1 * H* + HO*2 BOOH * Cu+3VCu*2 * OH" HO* (6) (7) Also a se rie s o f reactions was suggested where hydroperoed.de re a e ts v lth copper oxide, oe the orface of copper, to fora soluble ccorplexes which deccBpose to fre e ra d ic a ls . Soluble copper te a been shewn to be e ffectiv e a t concentration of l e s t than 200 parte per b illio n causing auto-. a to ly s is and an increase in d issip atio n factor* t t e inhlbltlco period depresses v lth lncreesins concentration o f copper. fid s mechanism ood ossociatad effects probably also apply for cataly sis by other a e tala such aa iron* file autoc a ta ly tic reaction i s kept in check, from becaalng explosive, by te m ln a tln g reactions here free ra d ic a ls combine to fin non-reactive products. Nelehlore ana Mills'* e ls ia tb o t catalysis vlth copper la enhanced by nitrogen and su lfu r larpur l t i e s th a t occur in the oil* They suggested th a t these Impurities pranote solubillcstloa of copper. Howeve r , J e z l, S tu a rt and Boss13, and others m in ta lu th a t th e aulfiir ecopounds n et as inhib ito rs to the oxidation. She f in a l products, i n l y acids and sludge, a re fam ed during th e l a s t two stages of the oxidation, when th e r a te of osqrgen uptake is constant and then decreases. Acids are derived * fre n th e a lip h a tic hyd rocarbons in the o i l , and th e sludge, larg ely insoluble acids and carboxylic s a lt , from th e aromatic*.o#9/*3 R>e mechanic f o r fin a a tio n o f these a a te ria lo must be very complicated, and i t involves decomposition of hydroperoxide, whose concentration decreases during t h i s period, and various reactions of. the oxygen ated fre e r a d ic a ls .*3** 7ra* th e above i t can be seen th a t deterlom tlo o o f a m ineral o i l can be inhibited by deacti vating hydroper oxide, th e fiv e ra d ic a ls, or the m etal c ataly st* tach o f these modes of in hibition w ill be discussed, as well as purification as means to increase the thermal s ta b ility of e nlit eral o il. Mineral o ils for e le c tric a l applications ere gene r a lly p u rifie d by successive extractions with s u lfu ric a d d ) but hydrogenation end solvent ex tractio n a re a ls o used* Various studies hove shown th a t tb s su lfu ric acid process can be opplied to a lim ited ex ten t.^ 9b f i r s t few extractions achieve increased s ta b ility .- However, when they a re continued beyond a c e rta in p o in t, where the o rc a a tlc content- of th e o i l begins to be sig n if ic a n tly diminished, file o i l becomes lees etable t o oxidation. I t was concluded th a t the eruoatles in the o il a c t, a t le a st in p ert, es an Inhibitor to the oxidation. An in h ib ito r tfast extends tbe in h ibition period u n til i t I s consisted i s known ec an a n ti oxidant. When i t Is exhausted, the o i l m y be replenished v lth more a f t e r the o i l i s suitably refined again. Ihe compound DHPC, 2, 6- d i- te r tb u ty l-jara-ereso l, end occasionally some of Its anslogues, hove been uied e ffectiv ely for thle purpose. I t s presence bet been shown to double tbe l i f e o f a t r a n s f o r m e r .^ There a re e number an ti-o x id an ts, of tbe nine caxpound type, that are more e ffectiv e fo r alnezal o ils than DBFC* But they cannot be used In e le c trie s l applications because they cause large increases la the dlsslpetiaa factor. There ore d ifferen t opinions about the reac tions of on antl-oxldsnt th at Inhibit oxidation. One* melota Ins th a t i t reacts v lth the hydroper oxide to yield re la tiv e ly in e rt product*. Another claims th a t the antlaxldsnt suppresses the propagation sequence by reacting p referen tially v lth the chain carrying free rad icals .*7 These active rad icals are cooverted to stab le aolecules, vb lle the anti-oxidant ra d ic a l, which ensues, is relatively in ert. 1 Deactivation of a metal catalyst to Inhibit oxidation has been achieved by use of certain chelating agents* Metal ions form complexes v lth these compounds in which they lose th e ir ionic character, and hence th e ir a b ility t o propagate free ra d ic a l chains. Also, a metal sutlhee may be passivated by adsorbing tb e chelating agent* This would prevent 'reaction of tbe metal oxide, copper oodLde fo r example, v lth hydr operoxide to form a complex th a t generates free ra d ic a ls. Sinee a metal deactivator resets v lth limited supply of c a ta ly s t, which is not replenished during the oxidation, I t s e ffe c t should be re la tiv e ly perma nent. in contrast, the effect of an anti-oxidant I t not,permanent as shown shove. Melehlore end M ills11 found b e tte r in hibition of oxidation by s m etal d eectlv star than by an an ti-o x id an t.' However, p r e s e n tly d i- te r t- b u ty l p sn -c ra a b l Is used much more extensively than metal desctivm tors. I t would seem lo gical to add both kinds of . in h ib ito r to an o il, ooe th a t reacts v lth hydro peroxide and one th a t deactivates e m etal c a ta ly st, to get o rela tiv e ly high degree of thermal stab il i t y . However, th is may re s u lt in too high a dissipation factor since these additives are fa irly polar and only a s a il amount can be to lerated In the o il. Also,decreasing th e ir concentrations to s a tis fy th e d issip atio n facto r requirement may re su lt in both being altogether ineffective. The following i s e reeeipe of s su ffic ie n tly sta b le transform er o i l , far operation e t about 100*C th a t i s based on exper ie nce and th e th e rm ! s ta b ility considerations noted above* A rcm tlcs - 30% Baphthenlcs - y% A liphatic - 65% Sulfur ( f ) and nitrogen containing compounds <1D p m eneb Additive - *15$ in h ib ito rs a re generally not added to the O il fo r power transform ers, where speciel care is .taken to avoid causes of oxidation. Oxygen la carefully excluded, and the o i l i s aa lo ts Iced ms dry * possible. N PC00026364 753701 S ilico n e Fluid There a r e number o f e ilie o n e flu id which a re r t t t d fo r re la tiv e ly high tra p e ra tu re a p p li c a tio n s , g e n e ra lly fetr 150*C and s tiro s even above 20G*C, d i e l e c t r i c . They a r c o f th e ethyl and phenyl s ilic o n e s , which a re polymers with tbeae general fornula. _ -- 2 R1 R2 - 81 - 0 - - 8 1 - 0 - 8 1 - 0 - _ * 1 --JD 1 R2 where th e H 's a re methyl or phenyl group. We C-H (p rim a ry ), Bi-C , and 81*0 bonds in th e s e s tru c tu r e s re r e la tiv e ly s ta b le th erm ally and r e s i s t a n t t o o x id a tio n , coopsred to th e C-H (seco n d ary and t e r t i a r y ) and C-C bonds in a In e ra l o i l . The s ilic o n e o ils a ls o bave sco re ry d e sira b le d ie le c tr ic and p h y sic a l proper" t i e s , such as low lo s s e s and low v is c o s ity -te m p e ra tu r e c o e f f i c i e n t s . However t h e i r h ig h c o s t bas re s tric te d th e ir use to s o s ll sized d ev ice s. The follow ing a re sow o f the c h a ra c te ris tic s of the degradation of silico n e flu id s th a t are p e rtin e n t to th e ir use in e le c tric a l dev ices. P y ro ly sis occurs above 200*C, and onuses v o l a t i l isa tio n and a decrease in molecular w eight. Exposure to oxygen i s d e le te r io u s sto v e 150*C, and re s u lts In Increases in v lse o slty and, fin a lly , g e l l a t l c n . Copper was found to i n h i b i t th e o x id atio n o f a silic o n e flu id e t 200*C, w hile o th er t a i s , such a s Ir a i s t e e l , t i n and t i n e , had no s l^ a lflo a n t e ffe c t/6 This Is in osrked c o n tra st to th e oxidation behavior of m ineral o il s . A cids, which can o fte n be avoided, Induce g e L lstio n . A stu d y was made by B e a ls, Hlckam and L o e ffle r !9 ,a o 0f some o f th e k in e tic s o f o x id a tio n o f s i l i c o n e s , which Included d im eth y l s i li c o n e (DC 200, Dow C orning) Boss o x id a tio n o f t h i s Liquid was d ete c te d a t 150*C; and i t g e lle d s t 250*C, when th e r a t e o f oxygen a b so rp tio n was r e l a t i v e l y h ig h . A mechanism we proposed f o r th e o x id a tio n o f s i l i c o n e s , which la e s s e n tia l ly th e same as th a t given shore fo r m ineral o ils , w ith e hydroperoxide Interm ediate in th e propagation of a free ra d ic a l c h a in . O e lls tio n Is p rob ably induced by th e a c id s th a t a re g en erated , slnee acid s advance polymer iz a tio n and cro ss-lin k in g o f s s ilic o n e . Use Of i n h i b it o r s to su p p ress o x id a tio n e f f e c ts In silic o n e flu id d ie le c tric s is severely r e s tr ic t ed by th e rag n itu d e o f th e tam perettares, above 150*C, a t which th ey have to be a f f e c ti v e . I t i s i mpor t a n t t h a t th e in h i b it o r does n o t in tro d u c e s q u a n tity o f io n s In to th e d i e l e c t r i c medium th a t would s u b s ta n tia lly in crease i t s d is s ip a tio n f a c t o r . While a number o f a n ti-o x id a n ta have been found which c o n trib u te l i t t l e to th e d is s ip a tio n fh e to r a t te m p e ra tu re s o f 100- 125*C in m in e ra l o i l s , t h i s la le s s l i k e l y t o be achieved a t 2002J0*C fo r s ilic o n e flu id s because th e d is s ip a tio n fa c to r increases rap id ly w ith tem prature. The h ig h e s t tem perature r a t in g o f a s i li c o n e f l u id can be o b tain ed by i s o la t in g i t t o avoid o x id atio n , sad then the therm al s ta b ility is based on p y ro ly tic r e a c t io n s . S e h ie f e r21 determ ined apparent i n i t i a l deconposition tem peratures fo r three silico n e d ie le c tric flu id s, in nitrogen. These s r e based on measurements of vapor p re s su re as a function of Increasing tem perature, s t a rate o f 2*C p e r m in u te, SDd a re th e p o in t where th e lo g vapor pressure versus l/(absolute tem perature) p lo t changed s lo p e . They were found to be 315*C to r dim eth y l s i li c o n e , 320*C f o r p b enylm etbyl-dlm ethyl s ilic o n e and 372*C f o r phenylmethyL s i l i c o n e . T heir magnitude is re la te d only q u a lita tiv e ly to the degree of therm al s ta b ility , as they probably r e f le c t f a i r ly high ra te s o f decom position due to the high ra te s o f tem perature r is e used fo r the vapor p re s s u re measurem ents. In f a c t , th e maximum operating tem peratures of these silic o n e flu id s e r e n o t more th a n 250*C, which I s c o n sid e ra b ly below th o s e v a lu e o . A mechanism f o r th e p y r o ly tic decom position o f a s ilic o n e f l u id was su g g ested by L ew is. 3 x t was based on w eight lo s s d a ta fo r a a e r ie s o f d im eth y l s ilic o n e liq u id s (DC 200, Dow C orning) F i r s t , th e polymer chain was assumed to c le a v e s t a s i t e where an im p u rity s tom was lo c a te d . One o f th e fragm ents would th e n r e a c t t o g iv e a product w ith a cyclic s tru c tu r e This kind o f p ro d u c t has been found to be f eu wed in th e prolysls of a silico n e. AsV arela A number o f d ie le c tr ic flu id s which a re widely used in c a p a c ito rs sod tra n sfo rm e rs a r e known a s - a s k s r e l s . T h is I s a g e n e ric name t h a t p e r ta in s t o th e i r n o n -flm m sb lllty . Tbeae d ie le c tr ic flu id s sre gen erally c e rta in cbloro diphenyls and chloro b en z e n e s. The moat common ch lo ro d ip h e n y ls a r e . # m ixtures of isom ers, o f tric h lo ro diphenyl (Aroclor 12l2), o f te t r a c h l o r o d ip h en y l (A ro clo r 12A8) sod p e n t c h lo ro d ip h e n y l (A roclor 125A ). T here a r e two c h lo ro benzenes In t h i s group, th e 1 , 2, Atr ic h lo ro and th e 1 , 2, 3 A -te tra c h lo ro d e riv a tiv e a . They a re used in so lu tio n w ith a ch lo rin ated diphenyl, g en erally to lower i t s v i s c o s i t y . The th erm al s t a b i l i t y o f a s k a r a la la r e l a ti v e ly high w ith regard to p y ro ly sis and oxidation, in fa c t, th e ir operation cap ab ility a t high tem peratures, above ljO 'C , ts lim ited by th e occurrence o f high d issip atio n factors re su ltin g from d isso lv ed im p u ritie s, and not to th e ooset o f therm al i n s ta b i li ty . O ften, however, th e therm al r a tin g o f a device lo which they a re p re s e n t has t o be r e l a t i v e l y low, o r an a s k a r a l can n o t be used, because of deleterious so lu b ility e ffe c ts, as th ey a c t as so lv e n ts and sw ellin g ag en ts fo r any organic re sin s sod components. This has to be considered regarding the s t a b il it y of d ie le c tric or stru c tu ra l m aterials th a t are exposed to an a sk a re l, and tba possible In tro d u ctio n ' o f an im purity from th e so lid to th e liq u id , where i t can be harm ful. Thermal d e g ra d a tio n o f m sk arcls b eg in s to become s i g n i f i c a n t a t a b o u t 200*C. Munch2* determ ined "> Monsanto tr a d e name. <4 NPC00026365 .91 753702 th e e ffect* of a i r , copper, Iron, le a d -tin older and luminum oo tb e change In d is s ip a tio n fa c to r o f A ro c lo r 12li2 due to h e a tin g a t 20C*C fo r up to 800 h o u r . D ia a lp a tio c f a c to r change* a re o fte n ore se n sitiv e to the presence of d eleterio u s degradation products, than re s u lts of chemical re a c tio n * used t o d e te c t them . He found, h e re , th a t none o f th e s e v a ria b le s in d iv id u a lly had a s u b s ta n ti a l e f f e c t on tb e th erm al s t a b i l i t y o f th is f l u id . The decrease in therm al s ta b ility due to s i r vos fu r th e r enhanced by th e presence o f c o p p e r. An ASTM aeth o d 2 5 1 a v a ila b le to d eterm in e th e s t a b i l i t y o f an ash o re 1 a t 200*C and 21f)*C, v lt h a i r p r e s e n t. I t se rv e s a u ln ly f a r quolity control. Aakarola fo r e le c tr ic a l ap p licatio n s a re re q u ire d to ba fr e e of an Im purity which e a s ily generates hydrogen chloride. 3 This a ffe c t can lead to in s ta b ility of an e le c tric a l d e v ic e . Bueh an Im p u rity may be a compound w ith su b stitu te d c h lo rin e, in an a lip h a tic hydrocarbon or In an a lip h a tic branch of an arom atic, or an ercam tle v lth added ch lo rin e. The p resen ce o f th e s e d e le te rio u s Im p u ritie s la detected by analyses fo r hydrolyzable chlorine (w ith a lc o h o lic sodium h y d ro x id e ).2^ These t e s t s are sensitive to parts per m illion, * . 1 , hydrolyzable chlorine. Sulfur H erafluoride The use o f s u lf u r h ex aflu o rid e a s a d ie l e c tr ic medlTB in h ig h v o lta g e equipm ent, which In clu d es high power c ir c u it break ers, tran sfo rm ers, x-ray apparatus and high frequency devices, has increas in g ly focused a t t e n t i o n on i t s therm 1 s t a b i l i t y c h a r a c t e r i s t i c s . The s a in concern la I t s re a c tiv ity , as w ell as th a t of it s products of d is c h a rg e , w ith th e s o lid components of an e le c tr ic a l d e v ic e . In th e gas phase, however, where th e re is no e ff e c t o f 0 s o lid , s u lfu r h e f l u o r i d e 1 q u ite s t a b le p y r o l y ti c e ll y , a t l e a s t to 500*C .2? , 28 very l i t t l e i n f o n * t ion is p resently av ailab le about the various possible th e n m l reaotlen* o f su lfu r b e flu o rid e with m etala, in s u la to rs , semiconductors aod other cportent* o f e le c tr ic a l equipment, to serve as a b aila fo r tem perature ra tin g s , in co n tra st, more has been published about re a c tio n s o f it s products of e le c tric a l discharges. The s ig n if ic a n t consequences o f th e n m l reactio n s o f su lfu r hexafluoride in e le c tric a l equipment a re lo sa o f th is gss and d e te rio ra tio n o f tb e s o l i d s . A m eaningful study o f t h i s would have to consider possible e ffe c ts or im purities th a t y be p re s e n t, sueh as m oisture aod s i r , on th e r e a c t i o n s . Some d a ta was o b tain ed by H ears2? on h a s tin g o f su lfv ir h e f l u o r l d e w ith v a rio u s m etals and in s u la tin g s m te ria lo a t 150*C and a,t 223*C fo r 2TO d a y s. Tbe m etals were s i li c o n s t e e l , mild s t e e l , copper and luminum. This l i s t i n g 1 in o rd e r o r t h e i r observed degree of re a c tiv ity , with silic o n s te e l being aost re a c tiv e . I t appeared th a t a l l the m aterials tested are q u ite compatible w ith su lfu r hexa flu o rid e a t 150*C, which should not be exceeded. A ctiv a te d a lu a ln a was t e s te d , In t h i s stu d y , as a scavenger for reactive gaseous products. I t did reduce the concentration of co rro siv e gss th a t was produced by decom position o f s u lf u r h e x a flu o rid e When th e use o f s u l f u r h e ra f lu o rid e in e le c t r i c a l equipment e n ta ils co n siderable d isch arg es, the re a c tiv ity of the products of discharge with tbe o th er components has to be considered regard ing th e o v e r a l l s t a b i l i t y . The p ro d u cts t h a t a re of concern a re those th a t remain a f t e r th e d is charge, or a re a ls o formed th e n . These a re a c i d i c , 30 com prising o f flu o rin e and low er f l u orides of s u lfu r, as w ell as hydrogen flu o rid e and o j^ flu o rld e s which a re due to tb e presence of m o istu re , oxygen and o rg an ic compounds .31 ab so rb ents such a s a c tiv a te d alumina have been used, q u ite e ffe c tiv e ly to lower th e ir concentra tio n s .30,31 schumb, Trump end P r i e s t 2? found t h a t both m etals aod n onoetal* a r e a tta c k e d by th e p ro d u cts o f a low c u rre n t de d is c h a rg e in s u lfu r h e x a flu o rid e . The m etals th ey te s te d followed th is order of re a c tiv ity w ith these products: brass 7 copper > s te e ly sta in le ss s t e e l > aluminum This order o f r e a c tiv itie s should be compared w ith th a t fo r th e therm al reactions of m etals w ith su lfu r h e flu o rld e . The d iffe re n c e in r e la tiv e p o sitio n s o f copper may be o f some s i g n if i c a n c e . some d a ta w ere, obtained re c e n tly by Kanasashl31 and by ThkuaaJ2 an th e expoaure o f se v e ra l polymer film s , which a re need a s in s u la to rs , to th e products o f corona d isc h a rg e s i n s u lf u r h e x a flu o rid e . On th e b a s is of weight lo s s , th e follow ing order o f re sista n c e t o th e s e p ro d u c ts was found: p o ly t e tr s f l u o r e tb ylene > polyethylene 7 polystyrene7 polypropylene > polyethylene te re p h th a la te ^ vinyl chloride > cellulose trie c e ta te . Understanding o f the reactio n s of su lfu r h exafluoride and o f i t s discharge products has as a back up th e chem istry of flu o rin e and hydrogen flu o rid e . Reference to th e re a c tio n s of these m a te ria ls and methods used to handle the* is useful fo r e le c tric a l application of su lfu r hexafluroide to help exploit it s d esirable d ielectric properties. COWCUUDIHO RZXMOC P resently, the various d ie le c tric flu id s that a re a v a ila b le and th e methods used to p rep are and m aintain them a re adequate to s a tis f y moat th e r a a l requirem ents fo r e le c tr ic a l eqiupment. I t le the s o lid I n s u la tin g member, o f o rag n lc nature-, th a t Often lim its th e therm al ra tin g and o p e ra tio n a l l i f e of a d e v ic e . However, as a s h i f t occurs to mare th e rm a lly s ta b le .a n d more so lv e n t r e s i s t a n t so lid In s u la to rs , the promise o f increased th e r mal ra tin g s w ill eventually c a ll fo r sim ila r 5 a \ /J NPC00026366 753703 Improvements In th e flu id m edia. This vould be achieved by upgrading the th e r a a l s t a b i l i t y o f p resen t flu id d ie le c tric s and developing "v oneo. The use o f d i e l e c t r i c f l u id s a t h ig h e r tem per a tu r e s th an new, v i l l demand even a c re a tt e n ti o n t o t h e i r d e g ra d a tio n phenomena and e f f e c ts on d issip atio n fa c to r- I t is reasonable to expect t h a t v ery c a re f u l and new technique v i l l have to be used to exclude im p u rities th a t could induce decomposition or increase the d issip a tio n fa c to r. Mechanisms o f deg rad atio n and acens to in h ib it I t , and th e n a tu re o f d e le te r io u s Im p u ritie s may v e i l be q u ite n o v e l, e s p e c ia lly where new liq u id s and solid d ielectrics are used. m ra:am 1 . 1968 Book Of AfiTH S ta n d a rd s, P a r t 29, E le c tr ic a l Insulating M aterials. Philadelphia: .American S o c ie ty f o r T estin g and M i te r l a ls , 1968 2 . T. V. Daklb, "E le c tric a l In su latio n d e te rio ra tio n tr e a te d a s a chem ical ra te phenomenon," A2ZE TRANSACTIONS, V ol. 67, P a r t 1 , pp. 113118, 19*8 . 3 7 . P . Byrne, and K. V. G uardlpee, "A method to measure the s ta b ility o f chlo rin ated in s u la tin g l i q u i d s , MATERIALS RESEARCH AND STANDARDS, V ol. 2 , p p . 825-828, O c t. 1962. k . J . J . M elchior, and I . V. M ills, "Factors affecting sta b ility of elec trical insulating o i l s , " H D TRAN8 . OH ELECTRICAL IHSUIATION, V ol. EZ-2, p p . I 5O -I35, December 1967. 5 L . B o ls a e le t, and C. Hourst o f f , 'The o x id a tio n p ro d u c ts o f l u b r ic a t in g o i l s , " AHH. COMBUS T IB L E S u Q u i r e s , v o i . 19, p p . 69- 76, 193V. See r e f . 9 . 6 . 7* M. C la rk , In s u la tin g > te te rla la f o r Design and to g in e e rln g P r a c t i c e . New York, London: John Wiley and Sons, I n c ., 1962. 7 . 1968 Book o f A8TM S ta n d a rd s. P a rt 29, E le c t r i c a l I n s u la tin g M ite r la ls , D66A, D190A-61T, D193*-62, and D2U 0- 65T . P h ila d e lp h ia : American S ociety f o r T esting and M tte ris la , 1968. 8. i b i d , D1313-5*. D190k-6lT and D2Wio-65T . 9 . "Symposium on I n s u la tin g o i l s , " JOURNAL OF THE INSTITUTE 07 PETROLEUM, V ol. k k , p p . 259366, S e p t. 1958. Note 7- Morton, sod R. T. B ell, ib id , pp. 260-272. 10. J . J . M elchior, and I . V. m i l s , "The ro le of copper during the oxidation of transform er o i l s , " JOURNAL 07 THE EIZCTIDCHMCAL SOCIETY, V o l. 112, pp. 390-395/ A p ril 1965. 11. J . W. M ills , and J . J . M elch lo re, " E ffe c t o f arom atic and s e le c te d a d d itiv e s on o x id a tio n s t a b i l i t y o f tra n sfo rm e r o i l s , " I and EC PRODUCT RESEARCH and DEVELOPMOTT, V ol. 6 , pp. h0-k2, m reb 1967. 12. G. R. D lm eler, I . V. M illo , and J . J . M elchlore, "The scope o f h y d ro g en atio n a r e f in in g to o l fo r the manufacture of tra n sfo rm e r-O ils," EIOHTH ELECTRICAL INSULATION CONFERENCE, pp. 183- 186, December 1968. 13* J . L. J e i l , A . P . S tu a r t, and E . S . R oss, `The e f f e c t o f com position on th e o x id a tio n s t a b i l i t y o f e l e c t r i c a l o i l s , " AIEE TRANS ACTIONS, V ol. 77, P o rt I I I , p p . 715-721, O ctober 1958. I k . H. M. Emanuel, E T . D enisov, and Z . K. Mnlxua, Liquid-Phase Oxidation of H ydrocarbons. T ra n sla te d by B- J . Hsxxard, New York": Plenum Press, 1967. 15 . E. A. Hornsby, R. Irv in g , and E. A. fh tte rs o n , "Operation of transform ers e t high tem pera t u r e s , " PROCEEDINGS 3EE, V ol. 112, p p . 586601, March 1965 16. C. J - Pede *een, "Mechanism o f a n t i o x id a n t a c tio n in p s o l l n e , " INDUSTRIAL ENGINEERING CHB-CSTOY, V ol. k8, p p . l 88l - l 88k, O ctober 1956- 17 C. D* Cook, "O xidation o f h in d ered p h e n o ls , 1 , O xidation o f an ox id atio n in h ib itio n by / 2 ,6 d l- te r t- b u ty l- k - m e tb y lp h e n o l," JOURNAL S ' * 07 ORGANIC CHEMISTRY, V ol. 18, p p . 261-266^ March, 1953- .' 18- D. C* A tk in s , C. N. Murphy, a o d 'c J5. S aunders, "Polymethyla1lo n en es...th erm al and oxidation s t a b i l i t i e s , " INDUSTRIAL AND ENGINEERING CHEMISTRY, V o l. 39, 1935s*1*01, Nov. 19*7- 19- L. C . 80a l a , sod V. N. Hlckam, "Therm al and o x id a tiv e d e g ra d a tio n o f s i l i c o n e s , " INDUS TRIAL AND ENGINEERING CHEMISTRY, V o l. 90, PP- 1583-158k, O ctober 1958. 2 0 . L. C. S e a ls , V. M> Hicham, and M. H. L o e f f le r , "E ffe c t o f oxygen p re s su re on th e therm s 1 d e g ra d a tio n o f a M e th y lp h e n y lsilie o n e ," JOURNAL OP APPLIED POLYMER 8CTENCI, V ol. I I , p p . 297- 301, 1959 2 1 . H. M. S e h le f e r , Dow C om ing C o rp ., M idland, Michigan, P riv a te coomiunlcatlon. 22. C V. L evis, 'The p y ro ly sis of dlm ethylp o ly a llo x a n e s ," JOURNAL Or POLYMER SCIENCE, v o l. 33, PP- 153-159, Dec. 1959- 23. Cl W. Lewis, T^je p y ro ly s is o f d lm e th y lp o ly s llo s n n e s . I I . JOURNAL OT POLYMER SCIENCE, V ol. 37, PP- *25-*i3i, June 1959- 2 k . R. H. Munch, Monsanto C o rp ., S t . L o u is, Mo. P riv ate cosnun len t ion. G NPC00026367 753704 t 2 5 # 1966 Book o f ASTH 8tandsrda. Part 29, 30. V. C. Bcbvsab, J . G. Truaap, and G. L* P r i e s t , E l e c t r i c a l In s u la tin g H te ria la , D193^"^ "E ffect of high voltage e le c tr ic a l discharges P h ilad elp h ia: American S o ciety fo r T esting on s u lf u r h e x a flu o rid e ," INDUSTRIAL AND and feterial*. niGurmnna chdostfj, v o i . 4i , pp. 131*8-1351, i Ju ly I9l9* \ i I b id , D1820- 65, MM1I-65T. i 31* M. JCanazashl, " D e te r io ra tio n o f SPg (polym er i 0 . C a a l l l l , G. 6 . Gordon and R . I - P lw p , film ) in su la tio n systems by corona d isc h a rg e ," 9 "Gaaeoua In s u la tio n fo r h ig h -v o lta g e 1968 ANNUAL REPOST , CONFERENCE ON ELECTRICAL } tr a n a f o r a e r s ," TRAN8ACTI0R A U S, V ol. 71, F a rt 3, p p . 31*8-356, J m * ry 1952* IRSUWTION AND DIEU3CTRIC PHENOMENA, p p. 9096, P u b lic a tio n 1705, H atlO M l Academy of S ciences, W ashington, D. C ., 1969. 28. B. M. Bochberi, "Sulfur b g f lu oride -___ .i gaaeoua e l e c t r i c i n s u l a t i o n ," EIZKTRICHESTW, V ol. 3 , PP* 15-19; 391*7. 32. T* TakuoB, "Corona d e te r io r a t io n o f organ ic in s u la tin g m a te ria ls in SP&," ELECTRICAL ! s}4 29. V. H. H ears, A llie d C heaieal C o rp ., S p e c ia lty Chernies la D iv is io n , B u ffa lo , Nev York, ENGINEERING IN JAPAN, V o l. 88, pp. 68- 75, October I968. j P r iv a te cosaura l e s t lo o . .1tI A { \ tI 7 NPC00026368 753705