NASA Technical Memorandum 83465 __,

N?2JA Review of Liquid LubricantThermN/Oxidafive Degradation

|NASA-TH-83q65) A I_VIEW CF LIQUID N83-_ c. lq3LUBRICANT THERNAL/OXIDATIV_ EEGEADA_ICN

{NASA) 60 p HC ItOV,/MF AO! CSCL IIitUIIclas

G3/27 q2C99

William R. Jones, Jr.Lewb Research CenterCleveland, Ohio

/t

August 1983

1983026872

https://ntrs.nasa.gov/search.jsp?R=19830026872 2020-04-19T03:36:32+00:00Z

A REVII!WOF I.[qtlTI} LUBRICANTTIIERMAI.I(}X1DA'rlVEI}EGRAI)AIION

William R. aon¢,s, Jr.Nat.ional Aeronautics and Space Administration

Lewis Research CentErCleveland, Ohio 44135

_' AI_STRA(,1

I'ne rundall_enl;al processesoccurringduring the thermaland oxidatiw_.(t

, _ degradationof Iwdrocarbonsare reviewed. Particularemphasisis given to_, variousclassesof liquidlubricantssuch as mineraloils, esters,polyphenyl*.()

.-:, ethers,C.-etllers,and fluorinatedpolyethers. Experimentaltechniquesfor de-I

,,._ L-rminingthermaland oxidativestabilitiesof lubricantsare discussed. Therole ot inhibitorsand catalysisis also covered.

INTRODUCTION

Liquid lubrica._tsare being subject_:dto ever-increasingthermalstresses.l'herole el the lul)ricantsthermaland oxidativestabilitiesis becomingan im-portantfactor in tileirsurvivabilityat high temperaturesin oxidizingenvi-ronments. Maximumfluid temperatureshave been estimatedto be in excess ot?60° C for many applications(refs.I to 8}.

State-of-the-artfluids (esters,hydrocarbons,silicones,fluorinatedpolyethers,C-ethers,and polyphenylethers)have one or more deticiencieswhiclllimitor preventtheir performanceabove 260° C. Some of these defi-cienciesare relatedto physicalpropertiessuch as pour point or volatility,but tor many the chemicalstabilityat these hlgh temperaturesis the weakiink.

Ine objectiveof this paper is to reviewthe fundamentalsof thermalandoxidativebreakdownprocessesoccurringin liquid lubricants. In addition,the.probablemechanismsof lubricantbreakdownare reviewedfor severalchemi-cal classes(hydrocarbons,esters,C-ethers,polyphenylethers,and fluorinatedpolyethers).

, THERMALSTABILITY

It is unfortunatethat the literatureis often not explicitconcerningtl)eterm "thermalstability". It is sometimesused interchangeaolywithth_rmal.-oxidativestability. However,the proper definitionis reservedforproc_ssesoccurringin the absenceof oxygen.

Mechanism

In the case of hydrocarbonsand most other fluid classes,thermaldecom-posiLionor pyrolysisproceedsthrougha free-radicalcilainreactionprocessyieldingmany products. Free radicalsare organicf_agmentscontainingan un-paired electron. These radicalsare producedby hemolysis(breaking)of C_-C

- 1983026872-TSA03

[bonds. These radlcals can be generated by radiation, mechanical processes,and thermal energy (ref. g),

The production of free radica]s from a hydrocarbon by a thermal process(rats. 10 and 11) Is illustrated in equation (1)

H-Ctl2-CH3 R. + R-Ctt2.-Ctl2 + R-_It-CIt3 (1)

I

These radicals are highly reactive and start reaction chains by abstractinghydrogen atoms (H) from the parent hydrocarbon. The chalnllko reaction arisesfrom a simple mathematical principle - the sum of an oven plus an odd numberis always an odd number° Whena radical (odd numberof electrons) attacks anonradlcal (even numberof electrons) one of the resulting species must havean odd number of electrons. Therefore, it Is a radical itself and also cap-able of attackinga nonradical, Attackby this secondradical(possiblydif-ferentthan the initialradical)will producea third radical• Abstractionreactionsare illustratedin equation(2)

R" + R-CH2-CH3----_ RH + R-CH2-CH2 (2)

_RH + R-CH-CH3

This chain sequence will continue until the radicals are destroyed or allof the reactants are consumed. Hence, a single radical can bring aboutchanges in thousands of molecules• Radicals may be destroyed by recombination

(eq. (3)) I

R" + R" * R-R (3)

or through disproportionatlon reactions (transfer of an atom from one radicalto another) as in equation (4).

R, + H-CH2-CH2 �RH+ CH2 = CH2 (4)

Radica]sthemselvesmay also fragmentproducingnew radicalsand unsaturatedspecies (eq. (5)).

R-CH2-CH2 + R, + CH2 = CH2 (5)#

These reactions occur in industrial "cracking".

Aitllough higher molecular weight products may be produced in this seriesof reactions, the most general change In properties of the lubricant is an in-crease in the vapor pressureof the system. This is broughtabout by cleavageof large moleculesinto smaller,more volatile,gaseousfragments. This gase-

ous evolutioncan be utilizedto quantitatethe thermalstabilityof organiccompounds.

2 Ol" i;;{i_Oi'__UALIIY

1983026872-TSA04

ORIGINAL J_, ,,_ _OF POOR QtI_LI'I"Y

Arr_nius Rate Law

Tile rate of thermaldecomposition.sually varieswith temperatureac_or{l-in9 to the emplrlcalArrhenlusrate law (ref, 12) as shown in tilefollowin_equation:

k _ Ae-EIRT (6)!

where k is the rateconstant,A is the frequencyfactor'or preoxponentlalfactor,E is the activationenergy,R the gas constant,and T the absolute

' temperature. Accordingto this equation,a straightline shouldbe obtainedwhen log k is plottedas a functionof the reciprocalof the absolutetem-perature. It has been shown (ref, 12) that for most organiccompounds 1o9dp/dt versus I/T is also a straightllne, Then, by analogy,

dd-_t= A'e"E'IRT (7)

Using this eqt'atlon,one can define the rate constantfor thermaldecom-positionby measuringthe isothermalrate of vaporpressurerise at severaltemperatures. However,it is more convenientto have a singleparameterforthermaldecompositionrather than to tabulatevalues of A' and E' whichactuallydefine the rate constant.

Therefore,an arbitrarythermaldecompositiontemperature(TD) is de-fined as the temperatureat which the isothermalrate of vapor pressureriseis 1.85 Palsec (50 torr/hr). Then, the decompositionpointsfor a seriesoforganiccompoundsare the temperaturesat which all have identicalisothermalrates of vapor pressurerise. This is the techniqueused in the standardtestmethod (ASTM D2879) (ref. 13) for measuringthe initialdeco_npositiontempera-ture of liquids. This test uses a constantvolumedevice (the isoteniscope)which can also be used to measurevapor pressureas a functionof temperature.

Tenslmeter

An automateddevice (thetensimeter),based on the same principle,(ref. 14) also yields thermaldecompositiontemperaturesand vapor pressuredata. A schematicrepresentationof the tensimeterappearsin figure I. The

' samplecell is made of ordinaryborosilicate91ass j)ndhas a yo1_me of about5 ml (5xi0-6 m3). Three to four milliliters(3x]O-° to 4x10-° m_) of testfluid are placed in the samplece11. The sample is then degassedand refluxedunder a vacuum. The cell is placed in a temperature-programmedoven and heatedto an initialtemperatureabout 50" C below the suspecteddecompositiontemper-ature, After a 5-mlnutestabilizationperiod,the increasein vapor pressure,if any, is recordedas a verticalbar during a fixed time interval. Then theprogrammerautomaticallyraisesthe temperatureby a preset amount (usually5° C) and the same processis repeated. A typicalthermaldecompositioncurve

f for a synthetichydrocarbonis shown in figure 2. This is a plot of the 1o9-I'

arithmof the isothermalrate of vapor pressureincreaseas a functionot re-ciprocalabsolutetemperature. A straightline is drawn connectingthe tops

_ ot the recordedbars. The intersectionof this linewith the temperatureaxis

1983026872-TSA05

is the T[). In addition, the activation energy for decomposition (E') canL)ocalculatedfrom the slope of this line.

Useful Lives

From a knowledge of tile I"D and Ej, a useful life of a lubricant(intt=e absence of oxygen) can be calculated from the following equation (ref. _2).

t "-'-RIt_ lo9 antilog - 219 E' - (8)

where t ts time, MW Is molecular weight, and x is the percent decomposi-tion at temperature T. The useful life is typically defined as the time inhours required for 10 percent decomposition. A plot of useful ltves for sev-eral lubricants appears in figure 3 (from ref. 15).

GeneralIzatIone

Blake et al. (ret. 12) reportedon the thermaldecompositionof a varietyof differentchemicalstructures. From their data and that of others (refs.15 to 20) one can make the followinggeneralizations"

(I) The maximumthermalstabilityof a straightchain hydrocarbonor othercompoundscontainingsuch groups is about 350" C.

(2) Branchedchain hydrocarbonsare less stablethan straightChain hydro-carbonsdue to sterlceffectsand the fact that free radicalsofgreaterstabilityare formed.

(3) However,stericcrowdingaround largercentralatoms such as tin andsiliconactuallyincreasesthe thermalstabilitycomparedto the lesscrowdedtin and siliconcompounds.

(4) Aromaticbonds (C-Itand C-C) have higherdissociationenergiesdue toresonanceand thereforethese compoundsare much more stable thantheir allphaticanalogs. Maximumstabilitiesof this class approacll450" C.

(5) Esters of alcohols having 13-hydrogensdecomposethrough a low energytransitionstate with maximum TD'S near 280" C.

(6) Estersnot containingB-hydrogenshave the low energyreactionpathblockedand thereforeexhibitstabllitesapproachinghydrocarbons(320" to 340" C).

(7) Substitutionson an aromaticrlng decreaseits stability. Increasingthe number of substltuentscontinuallydecreasesstability. Increas-ing the chain lengthalso decreasesstabilityunt;l it approachesthatof aliphatlchydrocarbons. J

._ (8) Saturatedring compoundsare more stablethan their straightchainanalogs.(9) Completelyreplacinghydrogenwlth fluorinesometimesincreasesthe

therlnalstabilityof an organiccompound(some exceptionsare aroma-tics and esters).

_L

ORIGINAL P_;_ ;l'4 OF POOR QUALITY

1983026872-TSA06

Bond Dissociation Energy

These generalizations arise from the fact that thermal decomposition oc-curs at the weakest 11ok in the compound, Therefore,thermaldecompositionshouldhe a functionof tileweakestbond dissociationenergy (Eilv_)in thatcompound, labh.,I tabulates TD and EDIS values for _ varie_%f com-pounos, irle_nermaldecompositiontemperatureis then plottedas a function

, of the bond dissociationenergy in figure4. Here the generaltrend ol_ inn.creasing TD with increasing EDIS is apparent.

' Bond Length

Since bond length is inversely related to bond strength, it carl also becorrelated with thermal decomposition temperature, Figure 5 contains a plotof TD versusthe reciprocalof bond lengthfor a series of aromaticcom-pounds of the type (C6H5)nMwhere M is a Group V element, Here the correla-

i tlon is much better than In figure4 since the compoundsare structurallyvery( similarexcept for the centralatom.

DegradationProducts

In general,two types of thermaldegradationare observedwith organicfluids. One type is random degradationin which a great numberof decomposi-tion productsare produced. A saturatedstraightchain hydrocarbonis an ex-ample. A second type is calledchain depolymerization(or unzipping). Herethe compoundrevertsto its monomer. Polyolefinfluids are examples. Polymershave analogousdecompositionrouteswith polytetrafluoroethylenebehavingin anunzippingmode while polyethyleneexhibitsa random degradationor chain scis-sion (ref. 21). Both processesproceedby free radicalmechanisms.

Many fluids,however,exhibita combinationof these mechanisms. Twofactorsthat are importantfor chain depolymerizationto occur ere:

(I) The reactivityof the depropagatingradical(2) The availabilityof a reactivehydrogenatom which could allow chain

transfer.

For example,all polymershaving=-hydrogens(such as polyacrylates)yieldlittlemonomer. On the other hand, polymerssuch as polymethacrylatesgivehigh yields of monomer becauseof the blockingactionof the =-methylgroup.Polytetrafluoroethylenealso depolymerizesdue to the resistanceof the C-Fbonds to chain transfm.

CatalyticThermalDecomposition

Blake et el. (ref. 12) has reportedthat the thermalstabilityof mostchemicalclassesis not affectedby the presenceof metal surfaces. One ex-ceptionis esters,whichyield decreasesin TD from 35° to 60" C In thepresenceof steel (ref. 12) or iron powder (ref. 15). Kiaus has also shownthat a numberof hydrocarbonswere not affectedby the presenceof variouscatalystcoupons(ref. 20). However,recentwork (ref.22) with two esters(DES and TMPTH)has yielded interestingresultsin evaporationtests innitrogen. At temperaturesbelow 180 C, evaporationlossesare unaffectedby

5

1983026872-TSA07

metals, llowever,above IBD* C, the rat_ of evaporationis increasedand isdifferentfor differentmetals, This effect Is shown in figure6 (ref. 22).This suggeststhat therma]degradationis being catalyzedby the metalsyleldlng lower MW productswhich increasethe evaporationrate,

Effect of Additiveson ThermalStability

l'nereare many additivesthat are effectiveIn increasingthe oxidationstabilityof lubricantswhich will be discussedlater, llowever,additivesdonot normallyaffectthermalstabilityof the base fluid, Thls was confirmed

by evaporationtests using a formulatedester containing?henothlazine(PTZ),a pheny1-=-naphylamine(PANA),and tricresy|phosphate(TCP). No differenceswere notedcomparedto the nonadditlvefluid (ref. 22).

OXIDATIONSTABILITY

The oxidationof an organiccompoundwith molecularoxygen is usuallyre-ferred to as autoxidation. As was the case with thermaldecomposition,oxlda-

_ tion usuallyproceedsthrougha free radicalchain mechanism(ref.23). Hew-ever, with the additionalparticipant(oxygen)the reactionscan becomeexceed-Inglycomplex, The importanceof hydroperoxidesin the oxidationprocesswasshown by Crlegee (ref.24).

Mechanism

T_e liquid-phaseoxidationof organiccompoundsand liquid lubricantshasbeen studiedby many investigators(refs.25 to 56). In general,the basicmechanismis thoughtto proceedas follows(ref.57).

Ri

In2-,-2 In, Initiation (g)

In. + RH �InH+ R. (I0)

Here an initiator(In) producesfree radlcals(In.) at reactionrate(Ri) which abstractsa proton from the hydrocarbonand producesan alkyl freeradical (R,). This highlyreactivespeciesreactswith oxygen in the propa-gation step of the chain reaction:

R. + 0z . RO_ Propagation (11) ,

This propagationstep producesa peroxyradical(R02,)which, in turn, reacts

with the parent hydrocarbon(RH)to producea hydroperoxide(RO2H).

kp

RO2" + RH-,-RO2H + R. (12)

OF POOR _UALITY

1983026872-TSA08

ORIGIN/_L I_t_GE_,. OF POOR QUALllYF

Itlisr_acti(}nr_g_nerat_sa alkyl fret;radical(R.)whicI_propa_jat_sth(_chain.

lh_ chain roactloncan he termlnat_dby radicalcouplingof two peroxy

radicals(ROR.).

2kt

P,ROp, --_ RO4R nonradical products (i3)l

or l}y(:rossterminationof an _lkyl radical(R.) arlda p(_rox.yr_(llcal(R()p.).4

R. + RO2. , RO2R (14)

_-_ filethird possibilityis the terminationreactionof two alkyl radicals(R.).

_R,.,,R-R (15}

Rate Equations

For normaloxygen concentrationswhere [ROT] >> R., reactions(14) and(15)can be neglected. Therefore,the rate of oxidationof hydrocarbonR-H canbe expressedas:

i (R1

The rate of oxidationIs then directlyproportionalto the hydrocarbonconcentrationand the square root of the rate of chain initiation. This rela-

i tionshipis illustratedIn figure 7, which shows the rate of oxidationof ethyllinoleate(initiatedby benzoylperoxide)as a functionof its concentration.For other types of chain termination,differentrate of consumptionequationswill be obtained. Some of these relationshipsappear in reference58.

Oxidlzability

' _he ratio kp/(2kt)I/2which appearsin equation (16) is referredto asthe oxidizability. For a specificrate of initiation,the autoxldationof ahydrocarbon(R-H) is then determinedby the valuesof kp and kt. The rate

' of c_ain propagation(kp)can be estimatedif the bond ehergyfor the weak._stC--Hbond is Rnown (ref. 59). The terminationrate constant kt can also beestimated. Thus, in theory,the oxidizabilityshouldbe predictablefor anypure hydrocarbon. Table II (ref. 60) containsa llstof oxidizabllltiesforvarioushydrocarbons. Thls llst illustratesthe wlde range of oxidationratesfor differenthydrocarbonstructures.

However,this simple relationship(eq. (16))breaksdown at highconverslons(>20percent). Complicationsare caused by the accumulationandreactionof secondaryproductssuch as aldehydesand ketones.

7i

J

1983026872-TSA09

ORIGINAL P_L_ _Cllain ]nltiatlm} QF POOF!QuI_LI'I_Y

AS tllerltlonsd_arller,chain Inltlatloncan h_ effectedby the dellberataaddition of an initiator. T.yplcal initiators are az_ c,ompoundsand peroxides,This circumventstll_.lOll!land ._olnet'Imef_IrreiJroduc11ile.inductionperiods(timebefore, oxyg(,,n al)sorptlon begins).

Inltlatiorlcan also take p1,1cnhy direct r_actlonwith oxygen:4

Ittl + 0 2 • II. "_'tli)_. (1'/) '

Ilowever,this reactionis thermodynamicallyand kinotlca'ilyunfavorable.|nitiationin tileabsenceof added initiatorsis probablydue to peroxidicimpuritlos,

llowever,a common sourceof free r'adicalsis the tI1ermaldecomposition1ofthe alkyl I1ydroperoxides,producingan alkoxyand a hydruxyradical:

RO211.RO. +.OH (18)

_, producingan alkoxyand a hydroxyradical, These radicalscan then react withthe parenthydrocarbon(RH).

RO. + RH + ROll+ R. (Ig)

•OH + RH . HOll+ R. (20)

producingwater, alcohols,and alkyl radicals• These reactionsare referredto as chain branching•

Chain Propagation

As indicated,the reactionof an alkyl radical (R.)and 02 (in itselfa biradical)proceedsvery rapidly. The rate controllingstep is then hydrogenabstractionby the alkylperoxyradicalin equation(12).

The alkylperoxyradicals(RO?,)are more stablethan the alkyl radicals(R.). They are quite persistentand selective. They preferentiallyabstract e

only the most labilehydrogenatom (weakestC-H bond). A group of H-bondenergiesappearsin Table Ill (ref. bT). It is obViOUsthat phenols_tIMols,aromaticaJaines,and phosphineshave the most labilehydrogenatoms. It isalso clear that the relativeattackfur primary,secondary,and tertiarybondsshouldbe in the order: tertiary(gl)kcal) > secondary(94 kcal) • primary(tO3 kcalj. Indeed,tillshas beee shown to he the case for Z-metI_ylpentanewhere the order is 3OO.JO.1(ref. 61)_ for tertiary,secondary,and primary

, bonds, respectively.

The propagationrate is also dependenton the type of hydroger,abstractingalkylperox¥radical. In order to correlaterates with C-IIbond energies,rateconstantsfor a seriesof hydrocarbons(R-H)should be comparedfor the same

iH

,, _. •

1983026872-TSA10

OF POOR QUALITY

alky]pP-roxyradical. 'lahl_IV (re#.57) show_ rate con._tant_l:ora s_rlos()f

hy{Irocarb_e_againsttheir own peroxy radicals(kp) and aBainstt_butylper_xyradlcaIs (k_). Obviously.reactlvltlesare vary _tructurallyd_pendent, l:or()xamplo.th(,_ bonloylperoxyradicali._40 000 times more reactivethan thet,-hutylporoxyradicaltowardbenzahlehyde. Comparingonly C,-libond enQrgle'_;(Table Ill)would leadone to concludethat aldehydn;_and alkylaromatlccoB|,..l)OUml._wn,hl oxidizeat similarrates, floweret,aldel)yde_oxidizeat apprec'l.-,al)lygr(._aterrat_ than the alkylaromatlcsdue to the strongel_ctronwlthdraw_

' ing effect of the carhonylgroup.

Chain Termination

As prevlouslymentioned,the norma'lterminationreactionIs

?.RO2,_ RO4R (13)

which is a tetroxide.

The decompositionof tetroxidesis also hlghlydependenton the nature ofthe R-group. For example,secondaryand primaryalkylperoxyradlcalsundergodisproportlonationto an alcoholand a ketone.

RO4R* R2CO+ RECHOH+ 02 (21)

Tert-alkylperoxyradicalsyield a differentmechanismresultingin the forma-tion of dialkylperoxidesand 02. The dialkylperoxidesundergofurtherde-composition, Therefore,primaryand secondaryalkylperoxyradicalsyield muchhigher terminationrates than tert-alkylperoxyradlcals.

The oxidationrate of a hydrocarbonis determinedby both the propagationrate (kp)and the terminationrate (kt)as indicatedin equation(16), Thisillustrateswhy a reactivehydrocarbonsuch as toluene(C6H5-CH_-H),which hasa C-H dissociationenergyof only 85 kcal/mole,has a rather low autoxidationrate. This is due to the high terminationrate of its primaryalkylperoxyradicals.

Inhibitionof Autoxldation

Autoxidationscan be inhibitedby the additionof scavengerswhich breakthe chai_1reactionby formingstablefree radicals:

RO2. + XH* RO2H + X. (22)

An exampleis in the use of subrtltutedphenols(?.6-di-t-butyl-4-methylpilenol)that would yield stablephenoxylradicals(ArO.)which are far lessreactivethan the RO2. radicalor the R. normallyformedby hydrogenabstractionfromthe parent hydrocarbon.

g

1983026872-TSAII

A secand type uf lnhihltur cause_ the de_truc_lOrl (*f Ilydroperuxides (R(121t),Ilere the inhibitor iX) reacts with the hydroperoxlde (R021i) yielding nonradl_alproducts.

R021t+ X ,_ ROll+ X() (2-3)

An example,of this type of Inhibitor is phenothtazto(; (PI'Z).

Metal coaters (sUCil as trlcre_ylpllosphate) can SOllietim_s b_ considered asa tlrlrd type Of Inhibiter. Ilere theIt action Is to prevent catalytic effectsby coating metal surfaces,

The eff(:cttveness of a chain breakln 9 lntlibttor is dependent on two fac-tors (1) the reactivity or stability of the radical iX) formed from the inhib-itor iX) and (?,) the reaction rate of the peroxy radicals (RO_.) with, the in-hibitor iX). The reactivityof X Is determinedby the ratio of kp/k_4 of thefollowing:

kpRO2" + RH"" ROyH+ R. (12)

k24RH + X" _ XH + R" (24)

The lowerthis ratio,the greaterthe inhibitionof the additive., If X. are

completelyinactive,kplk24 Is tery small and X is an excellentinhibitor. IfX, is still capableof chaln propagation,the chain reactionproceedsbut at alower rate.

: The Inhlbitorseffectivenessis also a functionof Its rate of reaction

with peroxy radicals(k22

k22

RO2" + XH _ RO2H+ X" (22)

The higher this rate (k22)the lower the concentrationof Inhlbltorrequiredtodecreasethe rate of oxidationby a certainfactor. The effectivenessmay Dedefinedas k22/ks where ks Is the reactionrate for some inhibitorin reac-tlon 22 taken as a standard. FableV (ref. 25) shows the effectivenessforseveralphenols.

Metal CatalysisIn Oxidation

: Metal catalyzedoxldatlonsare either homolytlc(oneelectronprocesses)or heterolytic(two electronprocesses). Homolytlccatalysisusuallyinvolvessolubletransltonmeta] salts (homogeneous)such as napthanatesof Co, Mn, Fe,and Cu, or the metal oxides (heterogeneous).Hemolyticcatalysisrequiresre-

_. _ cyling of the metal speciesbetweenseveraloxidationstatesby one electron

i_ 10 ORIGINAL P_ _5OF POOR QUALITY

- iL-

__C ....

1983026872-TSA12

ORIGINAl, I'I._.._L',_.q.IOF POOR QUALITY

ch;ln!|_;. In additi(_t_, t'v'_e Y'i_(lici_,|sa£_ pr(_dl_c_I_ddt_ring thl*,_ Illec,IlafFL,_,m.AlloXalllph__ t)t"l.hi_l;yp(:of'cat,rlly£i£illilhlsI:rat_din tlm foll_wln_l_qtlat.i(nl_.

+ (;nilIX 3 ------* -I,, '_ C()IIx_

+ X -----,H

! ' _ COIIIx3X _ collx?_

+ _ X + + fIXII

l_IthougI|this is an oxidativesubstituition'itdoes illustratethe homo.-lyric process(i.e.,the formationof free radicalsand one electrontransfer,

In constrast,heterolyticcatalysisinvolvesreactionsof compoundsco.-ordi_latedto transitionmetals, The metal complexacts as a Lewis acid andundergoestwo-equivalentchanges. Free radicalsare not involved, An exampleof this mechanismis shown here for anotheroxidativesubstitution.

_ pbIVx3 + X"

+ pbIVx4 _H

PI_IVX3_'-_ pbIVx3 + H+ (Z6)H

F-

_ pblVx3 "--'---_ _ X + pbIIx_

Catalysisof liquidphase autoxidationsoften proceedby homolytlcdecom-positionof alkyl hydroperoxides.The rapid decompositionof these hydroper-oxides in solutlorlscontainingiron,manganese,cobalt,and coppercompoundsis well known. The two basic reatlonsare"

t

! RO2H + M(n'1)+* RO, + Mn+ + IIO- (;_l)

RO2H + Mn+* ROz, + M(n-l)++ II+ (28)

L Therefore,metal complexescatalyzeautoxldationsby generatingchain initia.-Ling radicals(eqs. (27) and (28)). In additionalkyl radicals(R.)may beproduced directly from the parent hydrocarbon.

F-, 11

1983026872-TSA13

ONIGINAL pA_I_|_OF roOSt (_tll_,t,li r

R|I + Mn œ+ 11+ + M(n+])+ (29)

Mn++ Op. M(n+l) + Or. (3o)

t

IlumogenoousalldIlet_rogeneuus Cal;a'ly,si_;

s

I tie lun(hml(,_nt,_lchomh:al processes occurrill,9 during oxldat!on i_r'e L:,e ,_m)tc'Wlletiter t,be process takes place h} tl=o ceordl,ation sphere of a soluble metalcomplex (t}omogoneous)or at an adsorabate-motal interface. Th(; samemechaflis_.tic pathways aro available to botll kinds of catalysis, The local chomtcalstructure of the at'ttve site is more important than the macroscopic features ola system, such as Its phystcal state.

Chain granchln9

From the introductorymaterialIt might appearthat, at leasti_orpu_ecomponent systems, predictions of reaction rates and product dtstril}utl_nsshouldbe relativelyeasy to make. At sufficientlylow temperature,_(<1_0"C)and low conversions(<10percent)thls Is probablytrue for many sluplevdro,-carbon systems.

Above 100" C there is an increasingtendencyfor the p_ _ oxl¢_tlonproducts(alkylhydroperoxides)(R02H) to decomposehomolytl_allyintoalk_xyand hydroxyradlcalsas previouslyshown:

RO2H �RO'+ .OH (18)

Tt_esereactivespeciescan furtherreact_Ith the parent hydrocarbonRH toproducemore alky'l radicals. _I

RO. + RH . ROH + R. (19)

• OH+ RH+ HOH+ R" (ZO)

These branclHngreactionscan greatlyacceleratethe oxidation(auto- icatalysis)and compllcetethe reactionpicture. AnothercomplicationIs that

" the increasedreactionratemay drive the systemout of the kineticregion 1(i.e,,into a regionwhere oxygendiffusionbecomesa limitation). Therefore,thismust be taken into accountin high-temperatureoxidation.

EXPERIMENTALMETHODS

Reactionkineticsof liquid-phaseoxidationprocessesmust be studiedinthe kineticregion, That Is, the reactionmust not be oxygendiffusionllml-t_. Most standardoxldatlon-corroslontests developedin the past have beenbulk tests, Air or oxygen is passed over a staticfluid,bubbledthroughthefluid or over a constantlyagitatedfluid, An exampleof a macro-oxldatlon

12

1983026872-T,SA'I4

cell ls shown 1, ,lgure _. It ts not within the scope of this paper tod_scribethese macrotests,many of which are diffusionlimited, ReferenceP.6disc_4_edvarioustypes of tllesereactors. Other standardtests are alsodI_cus,_edin tileliterature(refs,62 and 63).

Oxygen Diffusion

()xygentransportInto the llqcdd-phaseinvolvesthree processes.

I_ [)iffusionof oxygen It,the gas pl,ase to the liquid-gasInterface,Dissolutionof oxygen into tileliquidat the 9as-liqui_interface,(3) Diffusionof the dissolvedoxygen into the liquidphase.

Process (I) is very rapid and not a limitingfactor. Process (2) is relatedto the partialpressureof oxygen a1_dHenry'slaw constant. Process(3) de-ponds on the interfaclalsurfacearea, rate of agitation,and the oxygencon-centratlongradient, Basically,to determineif oxygendiffusionis a problem,the amountof fluid or oxygen can be variedand agitationrates can be changed.If the rate of oxidationis altered,then the reactionis not takingplace Inthe k'neticregion.

" KineticCurves

The most common parametermeasured in most oxidationstudiesIs the ab-sorptionof oxygen. If this parameteris plottedas a functionof time, fourdifferentkineticcurves are observed(ref. 25) (fig. 9). Figure9(a) showsan autocatalyticeffectsometimesobservedwith mineraloils of low aromaticcontent. Figure9(b) is autoretardingand an exampleis highly aromaticmin-eral oils. Figure 9(c) is linearand is neitherautocatalyticnor autoretard-ing (a polybuteneis an example). And finally,figure9(d) shows mixed behav-ior (example- n-hexyldecylbenzene).

Microtests

Becauseof diffusionproblems,many investigatorshave designednew exper-imentsinvolvingsmall quantitiesof fluid. A numberof thin film tests havebeen developed(refs.64 to 66). One of the most successfulis that developedby Klaus and his coworkers(refs.30 and 67).

This test is illustratedin figure10. Basically,a very small quantityof lubricant(40 to 100 _l) is injectedonto the surfaceof a catalystafterthe entire apparatushas been equilibratedat test temperature. A constantflow of air is maintainedthroughthe air entry tube. Volatileoxidationproductscan be trappedfor analysis, At the conclusion,which could be a fewminutesor a few hours,d_pendingon test temperature,the apparatusis removedFrom the oven or bath and quenchedto room temperature.

Then, the degradedlubricantremainingo,ithe catalystsurfaceis dis-solved in arJappropriatesolvent. This solutioncan then be analyzedby avarietyof techniques,some of which will be highlightedlater. From thi_analysis,a rate of oxidationcan be determined.

ii

' 13

..... 1983026872-TSBO1

This apparatus ha._ a number of advantages. It, is _-imp]e, requires littletest _ample, has good reproducibility, and test durat%: _,., are not long, Sur-face area is also constantanddoe notvaryduringthe test. Anlnflnltvariety of catalysts can be _tudled and _ach can he recycled, ttowever, thereare some drawbacks,

TIle thin film te_t is a batch pro(:es_ ,'rod is subject to evaporation aswell as oxidation, This usually can be tak(;n Into account by running Identicaltests in nitrogen, llewever,some highly stablematerialsrequirehigher tem_ il)eraturesfor reasonablereactlo,rates to take place. Sometimesthe originalcharge disappearswell beforetest concluslon. Of course,this can he a11e-viated by runningtests above atmosphericpressureas has been done by others(ref. 25).

Anotherproblemcaused by the thin films is rapid lossof additivesfromformulatedfluids, Even with thin film tests,diffusionlimitationscan occur,especlallyat high temperatures,where localoxygenconsumptionIs great. Inaddition,these tests are usuallyrun to high conversion(up to 50 percent).At high conversionthe chain branchingreactionswhich take place can greatlycompllcatethe reactionpicture. Here many productsare formed and chemicalanalyslsand kinetictreatmentof the data becomedifficultif not impossible.Neverthelessthis techniquehas been successfulin reproducinghigh-temperatureoxidationdegradationfrom bearingtests (ref.56), and gas turbineenginetests (ref.68), and it has contributedto the fundamentalunderstandingof thedegradationof esters and hydrocarbons(ref.69).

StirredFlow Reactor

The thin film test just describedis an exampleof a batch process. Asinglechargeof material is used. In a continuousprocess,a constantflowof materialto be oxidizedis fed to the reactor. It has been shownthat theinstantaneousreactionrate can be determinedby balancingthe rate of reactionwith this flow (ref.70). This essentiallymaintainssteady-stateconditions.The stirredflow reactoris an example (ref.34). _

The materialto be oxidizedis fed into the reactorof constantvolume inwhich there is thoroughagitationproducinga homogeneousmixture. This mix-ture is then removedat the same flow rate as the inputof oxidizablematerial.Eventually,steady-stateconditionswill prevailand all reactants,intermedi-ates, and final productshave corstantconcentrations.The rate of consumptionor formationof any product,intermediate,or reactantcan then be calculatedfrom the followlngequation.

(X)_- (X)o (31)

where d(X)Idt is the rate of consumptionof formationof substance X, (X)zis the concentrationof X in the effluentat steadystate at residencetimet, and (X)o the concentrationof X in the enteringfluid.

This eqnationis valid for any species,regardlessof the complexityofthe reaction. This is obviouslya plus for kineticstudies. An empiricalrate

14

1983026872-TSB02

l,lwcan tii=mb_ determinedby findinga relationshiptilerde._cribesthe rat_.el l_rmation i_I a sp_,cie_ as a function of its concentration in the rc_actor.

An example _}i: ,_ stirred Ilow micr_r(_acter is shown i_l figure Ii, It con.-,_,ts ,_t tw() pyrex sph_;rt_,sA and d, The inside, spl_ere (t_) is perforated, l'hehydr()carbon(RII)enter,_at the t()pand is mixed with a streamof oxygen. Thi._mixturepa._se_thrnughthe perforatedsphere into sphereA. Tl}eoreticaliy,id(,.almixing,3_('the reactants,oxygen,and products(,ccurs.The outlet is onIHIJ:uppO.rrigllt.

l'llistechniquealso has some drawbacks, Since a continuousflow of reac-tants is necessary,a much greaterquantityof testmaterial is requiredthanin the thin film Lest. In additionpreactantpuritymust be stringentlycon-trolled, Nor is it clear how a catalyticsurfacecould be incorporated.Al-l:houghsolublemetal catalystscould be incorporatedin the reactant.

This type of apparatushas been successfullyused in kineticand mecilanis-t}c studiesof n-hexadecane(refs.34 and 35) and pentaerythrityltetraheptano-ate (PETH)(rei'.36). In addition,the mechanismsof two common antioxidants,4,4-dioctyldiphenylamine and 4,5-methylenebis(2,6-di-tert-butylphenol)

.. (ref. 71) have been studied.

MicrooxidationCorrosionApparatus

Anothermicroapparatusfor batch operationsis shown in figure 12. It isa modifiedversionof the type reportedby Snyder and Dolle (ref.46). Itconsistsof a pyrex decompositiontube and rod assembly. Metal catalystsmaybe positionedon the rod assembly. In a typicalexperiment,fluid is intro-duced into the decompositiontube. The system is evacuatedand backfilledwith

-_ oxygen to a known pressure. The tube is then placedin a preheatedfurnaceforthe specifiedtest time. At test conclusion,the systemis cooled and connec-ted to a vacuumsystem. The liquidnitrogennoncondensiblesare collectedquantitatively,measured,and analyzedby variouschemicalmeans. The liquid

i nitrogencondenslblesas well as the fluid residueare also analyzed. The rate

ot degradationis calculatedfrom the amountof liquidnitrogencondensiblesand is usuallyreportedas milligramsof condensibleproductper gram of ori-ginal fluid per hour of test.

This type of device has been used in a varietyof studieson linear_ef. 52) and branchedfluorinatedpolyethers(ref, 53). A similardevice,only using flowingoxygen)was used to study a branchedfluorinatedpolyetherand a perfluoropolyethertriazine(ref.49).

Again this apparatussuffersfrom the same disadvantagesas other batchreactors. Conditionsmust be chosenSo as to eliminateoxygendepletionordiffusionproblems.

_. ElectroricGas Sensor

As previouslymentioned,both thermaland oxidativeprocessesare accem-:_I partiedby the evolutionof gaseoushydrocarbonproducts. Ravner and WohltiJeni

i 15I"

] 983026872-TSBO

(r_-_t./27 l_aw_Lakefladvanta(l_uf tlli:_fact t_Jdeve,h)p a ratne,r silllplu,tt_(-ll-nique lur mm_itt}rin!jI,h_.,_)xidativebreakdown of |ubricants. l'heyus{;da soli(lstate m_t,al _}xidesomi(:_mductorgas se.nsor_)ri!ii_lally d{;vel(Jpedl;od(_Levt(._x--

i)l_3_Jv_] _:Otl(liLitJns. [h_ ._;yst_m is dia!iranml(:-'d in figure 13, (_urr_,lat.iorl tit tn,_(let.PCired !taS evuluLion w'ltll classical indJcai:_)rs (}f lubrJ(;_rll; oxidatiun (sucIiat; vi_i(:o_il, y all(I acld nl]lllber CIl'_n(j_s) i_; i IlusLrated in figilr(_ 14.

ANAL.YTI CAL TECIINI QUIS

Ilurim}, or at; LesL c()ncluston, a (:heroical analysis of I;he resulLing prod.-ucts is necessary. A variety of common chemical analytical techniques areused for these analyses. The following sections are not meant to be anexhaustive survey of these techniques. The intent is just to highlight a fewof tlleniorecommonly used procedures,

llighPressure Liquid Chromatography (fIPLC)

lhe deve]opment of HPLC has afforded a method for the separation of com-plex mixtures of organic compounds (such as a mixture of oxidized lubricantproducts) (ref. 73). This method separates materials due to an equilibriumdistribution of the materials between a stationary pi_aseand a mobile phasewllichpercolates through the stationary phase. HPLC is not used as extensive]yas gas chromatography (GC). However, it does have the advantage that whileonly _bout 20 percent of organic material is volatile enough for normal GC ana-lysis, a greater percentage can be solubilized for separation by HPLC. An ex-ample of this separation technique in the size exclusion mode is described inthe following paragraph.

b Size Exclusion Chromatography

Size exclusion chromatography (SEC) is the HPLC mode that separates com-ponents of a mixture according to molecular size. G-MIL-99 is a generic gasturbine engine lubricant developed at NASA for fundamental tribologica] stud-ies. It is very similar to proprietary blends meeting the MII.-L-23699speci- ",fication (ref. 74). G-MIL-99 is a single component basestock (trimethyolpro-pane triheptanoate) (TMPTH). This formulation also contains an antiwear addi-tive (2.5 percent tricresyl phosphate (TCP), two antioxidants, 1 percent di-uctyldiphenylamine (DOI)PA)and 1 percent phenyl=_-napthylamine (PANA), and acorrosion i,hibitor, 0.02 percent benzotriazole (BTZ). Figure 15(a) is a SECtrace of the unused fluid using an ultraviolet detector at 254 rim. As can beseen, three of the four additives have been partially separated and are iden-tified. Figure 15(b) contains the same sample but uslng a refractive index(RI) detector. Here, in addition to the additives, the primary ester peak isseen.

Tli(;next series of spectra (fig. 16) silowstireprogressive oxidation of asimilar tormulated ester in a gas turbine engine test (ref. 68). The spectrashow the progressive loss of additives and the formation of higher molecularweight products. Tilisdata is typical of the decomposition seen in iligii-,temperature tests with ester base fluids (refs. 29 and 69).

ORIGINAL P .......

16 OF POOR QUALITY :

,i !

1983026872-TSB04

Uitravio]et-VlslblnSpectroscopy

once an oxidizedmixturehas been separated,it is usefulto chemicallyidentify,at leastqualitatively,the variouscomponents. Many lubricantsandlubricantadditivesabsorb light in the _avel_ngthregion of _00 to 800 nm.AromAticcompounds,in particular,adsorbstronglyin the ultravioletregion(_00 to 400 nm). These energyadsorptionsare duo to electronictransitionsfrom ground to higherenergy states. A dlode-arrayspectrophotometercan be_isedLo help identifyvariouscomponentsas they olute from an iiPLCcolumn.l'hisdevice is capablee_ measuringan entire ultraviolet-visiblespectrum(UV-VI$)In I _econd. A seriesof UV-VISspectrafrom the HMW regionof fig-ure 16 (65 hr sample)are shown in figure 17.

InfraredSpectroscopy

Anothercommon identificationtechniqueis infraredspectroscopy(IR).llerethe spectralregionof interestis from 4800 to 400 cm"_. Adsorptionsin this region are due to variousmolecularvibrationssuch as stretchingandbendingmotions. IR spectrafor two componentsfrom a size exclusionchroma-togramof an oxidizedester appear in figure18. These spectraidentifythe

:: undegradedester basestockas well as the anti-wearadditive(TOP).

NuclearMagneticResonance

A third complimentaryspectra_techniqueis nuclearmagneticresonance(NMR). This techniquemay providemore structuralinformationthan eithertheIR or UV techniques. While IR and UV techniquesare often empiricalin nature,NMR requiesa more theoreticalbasis for proper interpretation.Simply stated,many atomic nucleipossessmagneticmoments. Normally,individualnuclearmo-ments are randomlyorientedso that there is no net magneticmoment. However,if that substanceis placedbetweenthe poles of a powerfulmagnetjthere willbe a small net orientationwith the externalfield. One can then measuretheenergy requiredto promotenucleito high energy states,out of alignmentwiththe field. An exampleof the use of NMR in oxidationstudiesis illustratedin figure19.

}]erea proton NMR spectrumof an as-receivedfluorinatedpolyetherisshown in figure 19(a). The presenceof a doublet in Lhe NMR spectrumat 5.9ppm Is indicativeof the presenceof a C-H bond. This is confirmedby thespectrumof a fully characterizedfluid (C3FTOCF(C_)H)known to containa C-Hbond (fig. i9(b)). The fluorinatedpolyetheris supposedto be devoidof suchbonds. Since the bond energyof a C-H bond is much less than thatof a corre.-spendingC-F bond, thls presenceof C-H in such a fluid providesan unwantedwear link. The consequencesof this are discussedIn the sectionon fluorina-ted polyethers. A spectrumof the same fluid after thermalpretreatmentInoxygen at 343" C appearsin figure 19(c). Here the C-H impuritieshave beendegradedand the doublethas disappearedand a much more stablefluid results.

' ..... 1983026872-TSBO[

ElectronParamagnet%cResonance

Somewhat analogousto NMR I_ a tecllnlqueknown as electronp_ramagnetlcresonance(EPR) (ref, 75). ilerewe are concernedwlth electronsrul;herthannuclei. 'Thespin in a magnetlcfield may be orientedin two ways and each i)aveslightlydifferentenergies. If radiationis Bnpplled.a transitionmay occur,Radiationin ti|emicrowaveregiop is used. In practice,the wavelengthot theradiationis kept constantwhile the intensityof the magneticfield is varied,"_hismetho_ can detect unpairedelectrons(free radicals)In cnncentrationsaslow as _0_I molar (ref. 76).

An exampleof a stable (long-lived)free radicalelectrochemicallygoner~ated C-etherlubricantis shown in figure 20. Actually,two free radicalsarepresent. These speciesmay be reactionproductsof initiallygeneratedradicalions and may be responsiblefor the rapid productionof high molecularweightmaterial (sludge)often generatedby C-ethersin high-temperaturebearingtests(refs.54 and 56).

OTHER EXPERIMENTALTECHNIQUES

Chemiluminescence

"_ Chemiluminescence(CL) is simply lightemittedfrom a chemicalreaction,

It was first observedin biologicalsourcessuch as the firefly(ref. ll). Ifa reactionis to produce light severalcriteriamust be met. (1) sufficlentenergy for excitation,(2) a speciescapableof formingan excitedstate, (3)an emitterto give off excitationenergy,(4) a rapid chemicalreaction,and(5) a reaction-coordinatesystemfavoringproductionof an excitedstate overdirectground state formation.

It has been observedfor many years that CL is associatedwith many oxi-dation processes(ref. 78). Referringback to the sequenceof reactionsin theautoxidationof hydrocarbons,it is believedthat equation(13) is responsiblefor the observedCL:

2RO_'-m'ROH+ R2C = 0 + 02 (13)

This disproportionationreactionproducesa small amountof excitedstates(R2C=0)*. For most hydrocarbons,only one photon is emittedfor each 108 to

1010 terminationreaction. This inefficiencyis caused,in part, by quenching &'. reactionsbetweenthe excitedketonewith oxygenand other species.w

(R2C=O)*+ Q �R2C=O+ Q (3_)

. (R2C=O) * R2C=O + hv (J31

;I 18 ORIGINAL PAGI _OF POOR QUALI'W

_C) Ol.I/_.l.lL.Jl/._ l_-_l...sv'.

The intensity of the emitted radiation (by) Is dtrectly proportional to

square of the concentration of peroxy radtca]s [RO_*] 2. This dependencethe

ts the basis for determining the concentration of peroxy radtcals and thus thekinetics of hydrocarbon oxidation. The advantages of CL are (1) It ts ex-treme]y sensitive, (2) ttts nonlnvasive, (3) it provides a continuous monitorof peroxy radicals, and (4) only small samples are required. A numberof In-vestigators have used CL tn oxidation studies (refs. 7g to 81). An example ofa CL apparatus appears tn figure 21. A thorough review of CL appears In ref-erence 82 and a recent review concerning Its applications to fuels and lubri-cants appears tn reference 83.

Shock Tube Studies

Another method that has been used for kinetic studies of hydrocarbon py-rolysis arid oxidation is the shock tube technique (ref. 84). Its use in gasphase hydrocarbon oxidation is reported In reference 85. In this technique,gas samples are heated during a very short tim_ interval (0.5 to 5 usec) andthen quenchedby expansion. The reaction products can then by analyzed outsidethe shock tube for use in kinettc studies. Since this analysis can be carriedout at leisure, a variety of ana]ytical methods can be used. Therefore, eachexperiment can yle|d data on the concentration_ of many compounds. A varietyof rate constants for various hydrocarbon reactions appear in reference 8b.

CLASSESOF LUBRICANTS

Hydrocarbons

Most conventionalautomotivelubricantsare mineraloils which are com-plex mixturesof hydrocarbons. Althoughsome fundamentalstudies(refs.20,33, and 69) have examinedthesematerials,others (refs.34, 35, 38, 43, 44,45, and 86) have opted Lo study a pure component.

Jensen and coworkers(refs.34 and 35) have performeda seriesof elegantexperimentswith pure n-hexadecane. Autoxidationwith molecularoxygenwascarriedout at 120° to 180° C using a stirredflow reactor. Inltlalstudiesconcentratedon identifyingthe primaryoxidationproductsat 120°, 160°, and180" C. These primaryproductsincludedmono-, di-, and trihydroperoxides,hydroperoxyketones,cyclic peroxldes_and some trlfunctionalproducts. Neg11-gible amountsof estersor acids were formed in these low conversionstudies.

The hydroperoxldesare formedby both Interhydrogenand intrahydrogenab-stractionreactions. Furthermore,it appearedthat a high percentageof hydro-gen abstractionswere takingplace by hydroxyradicals (°OH). It was felt thatthe Intramolecularabstractionsshouldplay a role in the autoxldattonof alln-alkaneslargerthan n-butane.

In a continuationof thls work (ref. 35), secondaryor c]eavageproductsof the autoxldatlonof n-hexadecanewere determined. Productsat low conver-sion (2 percentor less) Includedmethy]ketones,a]kanolcacids,hydrogenper-

i

19

1983026872-TSB07

; oxide, aldt_hydL_..%ethersDcarbondioxide,and carbonOlOnOxidc_,At IMglmrCOil-..v_r_ions(<16 percent)_.ste.rsand y-lar.toflf_sare fornled.

The be.sicmechanismProl)OS_dby tilea_4thor5to accountfor l:heobs(_rv(.,dproductsis suumlarizedas follows:

Chain initiationoccurs via homolyslsof hydroperoxidosformingalkox.y(RO,7 and hydroxy(,OH7 radicalswhich subsequentlyabstractprotonsproducin!ialkyl radicals(R,). TI)ealkylradicals(R,) react witiDoxygenproduclnyperoxy radicals(If02,)whlch_ in turn, abstracthydrogensintra.-orintermolec.ularly producingmono-, di_, and trih)droperoxides.Chain terminationoccurs via_in?olecu)arreactionsof chain carryingperoxyradicals(R02,7. Methyletones (CI|3COR)ann aIKanolcacids (RC0211)are formedby cleavagereactionsof a,y_hydroparoxyketones(IIOOR_O).Exc_essacid and other cleavageproductsare derivedfrolnreactionsof alkoxylradicals(RO,).

Esters

Most esters in use teday as lubricantsare gas turbineengine lubricants.PresentformulationsmeetingmilitaryspecificationsMIL-L-23699or MIL-L-7808are eitherbased on trimethyolpropaneor pentrerythritol.These alcohols,which containno @-hydrogens,are estrifieowith variousacids to producethefinal ester products. In practicethese lubricantscontaina mixtureof com-pounds.

However,as in the case of hydrocarbons,utilizinga pure compoundforoxidationstudiesis advantageous. Hamiltonet al. (ref. 36) have studiedthemechanismof the autoxidationof pentaerythrltyltetraheptanoateat 180° to220" C. This was an analogousstudy to their work with n-hexadecanepreviouslydiscussed.

The proposedreactionscheme for PETH is very similarto that reportedforn-hexadecaneand is reproducedin figure22. Again it consistsof initiationby hydroperoxidedecompositionformingreactivefree radicals(I), formationof monohydroperoxidesby intermolecularhydrogenabstraction(37, and dihydro-peroxidesand hydroperoxyl<-e1_o'nesby intramolecularabstractions(4 and 4"7.Chain terminationproceedsvia perox_ical recombination(6). Secondaryproductsof acids and methylketonesare formedfrom hydroxyketones(77. Inaddition,high MW carboxyand acetylsubstitutedtetraesterswill be formed.

A kineticanalysiscomparingrate constantsfor the two systems(PE'rHand ;;'n-C|6)showeda remarkablesimilarityfor most reactionsteps. This wouldindMcatethat indeedboth compoundsappearto be reactingby a similarmech.-anism.

Continuingresearchat PennsylvaniaState Universityover the last fewyears has also led to a fundamentalunderstandingof the thermal-oxidativereactinnstakingplace in ester based systems(refs.27 to 32).

Bulk oxidationtests to (500° F) 260° C on di-2-ethylhexylsebacate(DES) and trimethyolpropanetriheptanoate(TMPTII)were reportedby Czarnecki(ref. 28). At 260" C, testswith theseesters were autoretardlng(seefig. 97.It was felt that this was due to the in situ formationof inhibitingmaterials.

ORledNALp_ 15;20 OF POOR _IALIIY

)

1983028872-TSB08

mm

=floweret,there were obviousoxygen diffusionlimitationssince an approximate3D C rise in test temperatureonly doubledthe reactionrate.

Reactionproductsfor both tests were analyzedby gas chromatographyan{lmass _pectroscopy. For DESa llalfacid ester and a high boiliflgfraction(>510" C) were reported.

For TMPTH, a diesterand smallerfragmentswere reportedas well as ahlgh boilingfraction, No definitivereactionschemewas theorizedbut somesort of a condensationreactionwas proposedfor the formationof the hlghboiling (i.e,,hlgh molecularweight)fractions.

All (ref.O7) reportedfurtheron the chemicaldegradationof estersagalnusing a bulk oxidationtechnique, Some diesters,TMPTH and PETH were studied,Oxidlze_productswere separatedby gel permeationchromatography. A chromat-egram for di-2-ethylhexylsebacateis shown in figure 23. Typicallyall theesters producedthree fractionsof low, intermediate,and high MW. The indl-vidua] fractionswere isolatedand furtheranalyzedby IR, UV, and NMR tech-niques. It was deducedthat, initially,a low molecularweight productIsformedwhich eventuallypolymerizesto a high MW sludge (fraction3). Spectro-scopic analysisindicatedthat the polymerizedfraction(up to MW of 50 000)containedcarbonylgroupsconjugatedwith one or i_re double bonds {-C=C-C.O).

Cvitkovic(ref. 30) describedthe developmentof the microoxidationtest

previouslydiscussed. Ester oxidationwas representedby a simplephenomeno-i logicalmodel. It was concludedthat this techniqueadequatelysimulatessev-

ere degradationenvironmentssuch as those presentin high-temperaturegasturbineengines.

_- Productanalysisreinforcedthe theory that the first step in ester oxi-dation is the formationof low MW unstablespecieswhich subsequentlypolymer-

'. ize into high MW sludge. UV absorptionwas found to increasewith increasing

M_Jof oxidationproductsand with the extentof oxidation. In additionGPCanalysiscombinedwith atomicabsorption(AA)of samplesfrom 4-ball tests

indicate that the iron surfaceparticipatesin the reactionsgeneratinghlgh

MW material.

Furtherwork was reportedby LockwoodeL at. (refs.3_ and 88), with

estersusing the same apparatus. Here a kineticmodel was devisedwhich pre-dicted inhibitoroxidationrates,evaporationrates,and lubricantstable lifefrom i/4° to 247° C. Post stable lifetests wlth Fe catalystsindicatedrapid

_, basestockoxidationwith the productionof hlgh MW products and organic ironproducts.

_-_ A model for ester oxidationis presentedin reference88. It is similar

|to that proposedby Hamiltonet al (ref.3b) exceptthat the model is carried

., further to indicate how a final condensation polymer containing carbonyl groupsconjugated with double bonds could be formed by dehydration of ¢-hydroxyketones.

r_

_" ORIGINAL PAGE i

OFPOOROUALrr i

1983026872-TSB09

OF POOR ,Qtll_LrlY[)11 I,) -112() (II II " II

R-£1tCII?CR_ RCII,.,CHg..-R ( ,_4)

| i_ addi t Jon i_ iiiP,,_,harlJsmwhereby _n i ntf:rmerli at( _,I)rOdllct oJ' kp,,torff._ (;undo.N-sati_)n(:nuhlre.a(:twith an iron surfaceto l}rOdu(:P-_olubloiron product_wa._prt!;_(!nt(_(l,

0tt ()

I"eOIt + R-_It Ctl2CR'--" --_ (J_]

C

c/ \co# -----,'olym rlcproduct

X,. Soluble iron productswere almosttotallysuppressedby the pr_._enceof a known

metal coater - TCP.

A furtherdescriptionof the Penn State microoxidationtest appearsinreference69. Here oxidationtests of esters and a superreflnedmlneralo11are reported. Oxygen diffusionlimitationswere negligiblefor small samples(40 pl or less) at 245" C or less. Reactionkineticsindicatedthat esters 'and the mineraloil followeda first-orderreactionrate from 0 time up to 80 _

percentconversion. !I

As indicatedin previouswork, primaryoxidationproductsfur the esterwere ot lowerMW than was the basestock, For a mineraloil, these productswere in the same MW region as the basestock. Secondaryand tertiaryreactionsinvolvethe primaryreactionproductsin a condensationpolymerization.Theprimaryproductsdo not accumulate, Tnereforejit was concludedthat primaryoxidationis the rate determiningstep and the polymerizationreactionis muchmore rapid.

CatalystStudies

Catalyststudiesindicatedthat the type of metal affectsboth the primaryrate as well as thu condensationstep. For the primaryreactionthe followingorder o_ decreasingcatalysiswas observed:

Fe > AI > Pb > Cu

The ordur tot decreasingcatalysisof the condensationreactionwas a littledifter'(.,nt:

Fe > AI > Cu > Pb

2_

1983026872-TS B10

r

CF POOl1 QUALITY

Q.alitatlveanalyalsindicatedthat tour characterlstlcchemical9roup_wore identifiedin all oxidationproducts(primaryand secondary)from e_tersand the mineralell, These were hydroxyl(_OH),carbonyl (C.O),alkenes(C.C),and conju@ateddienes (C.C-.C.O).

More recent work (ref, _P,)with the microreactorindicatesthat oxygen isdefinitelyrequiredto inducethe po|ymerizatlonreactionwith esters. Thiswork also establlshedthat wllil_low carl)onsteel and stainles,__teelcatalyzeoxidation,Pb, Zn, Sn, and CU act as Inhibitors, AA analysisindicatesthatthe concentratlonof dissolvedmetal in a luhricantdoes not follow the orderof catalytlcactivity. Low concentrationsof iron are activepromoterswhilehigh concentrationsof lead act as Inhibltors.

Polyphenyl Ethers

The polyphenyl ether fluids have been studied (refs, 89 and gO) as possi-ble high-temperature lubricants for many years, They are thermally stable toapproxlmately450" C and oxJdativelystableto approximately275" C. Most oftheir problemshave been relatedto poor low-temperatureproperties_nd poorboundarylubricatingability.

Most of the mechanismsof oxidationdegradationof these fluidswere per-formed in the early 60's (refs.55 and 91). At 288° C aridabove these fluidsbegin to consumeoxygen (&fiefan inductionperiod}. These were bulk oxidationtests using a Dorntetype apparatus. As oxidationproceeds,the fluid turnsblack with an accompanyinglinear increasein viscositywith increasingoxygenconsumption. In addition,EPR spectraindicatedthe presenceof a stablefreeradicalin the oxidizedproduct.

In contrastto the previouslydiscussedautoxldationmechanismsof alipha-tic hydrocarbons,the initiationstep appearsto be an sttackby molecularoxy-gen at the phenyl-oxygen-phenylcarbonsratherthan C-ilpositions. Symmetricaland unsymmetricalcleavageariddehydrogenationreactionsoccur to yield phenoxyand substitutedphenoxyradicals.

Propagationoccurswhen these radicalsattackthe parentmoleculeat the

C-O-C carbonsprodu_:ingmore phenoxyradicalsand with c]eavageto form ahighermolecularweightproduct (containingsolaeC-C bridges). An overallreactionwould be

5C_4H1803+ 40Z * _0C61140g+ 51t2() (36}

Becauseof tlleinherentstabilityof these fluids,conventionalantioxi-dents (such as ..,inesaridphenols}are rlotuseful. This is relatedto thevolatilityor thermali_stabilityof these additives, lluwever,some additivesare effectiveabove 260 C. These includearomatictin compoundssuch as tet-raphenyltin (ref.9Z} and certainother organometalliccompounds(ref.gl).Cuprous and cupricoxides (refs.gZ and 93) were also effective,

A continuationof this work (ref. 95) showedthat solublealkalior alka-line earth metal pheno_idesare also effectiveinhibltorsto 310" C. Oxides,hydroxides,and carbonatesof alkalimetals and barium are very effectivein-

23 ;'t

i

1983026872-TSB 11

llibit._)rs (bJ approximatcqy _7(|" C). A propound m_)cbani:;mis tile d_i:ruci:i_rl_I p_._r_x:/radlcals by ;_ _i_pne_xld_ l:ormati_n (r_f, g4),

|11 1,_;1., ox'idid:i_n/=:orrosion t:=_;t_, (r(d', gfi) _hownd that th(._pr_m:n ot,_ varitd:,y qt ili_)Lal(:_)Ul)Or)_reduced l:h(_()xJd_d;Joflof a pol.ypllnnyl etilP.r (IIIPO!_.ur'=_dby Vl_-;f:osity ln(_r'(_,_()), I.ead alld c_)pl)f_f'S(.'(_lll(_.dto be tit(._ mo!;I;fH;fc_¢:L1ve,1'11i,_1_ In ,=_jr¢.,(_m=_H;with th(._ ftndh)9_ of Kla==set el, (Pet, _,_) (li_cus_d Inl.h_ (:_L¢)r _(!cLiori,

(:,,(;.th_.rS

An_)thor c'ta,_ of t l_=l(|s, (C,,.eth()rs) or thlo,..¢.,.tho,rs, are ._tru_turally slm ,_i l_r to the I)olyph(myl ether; (Per. 97). l'bO basic col_lpo_tt'io_l of this cla_sis .*;hewnin figure ?.4. These fluids h_ve,certain advantag(:s ow)r _ho poly.-.phenyl ethers such as lower pour points and bettor boundary lubricating I_(_r'-tormancu (ref.g8). However,their Inh{)rentthermalstabilityit somuwha_lower (_,pproxlmately390" C) and oxidationstability(bulktests) Is about260" C. However,at with the polyphenylethers,these fluidsdo producelargequaetitiesof a high molecularweightsludgeunder l_hricatingconditions(ref. 99).

Macr(}oxidation.-corrosiontests of formt_latedC-ethersindicateddeposit

w formationwas a problem(ref. 54). T_,oC..etl_erformulationswere studiedbyJones and Morales (ref.56) at 353 C using tl)oPenn State a macro.-reuctorde-sign. Degradationproductswere analyzedby gel peimeationchromatography.¢_-50steel and Ag were used as catalysts.

In general,both cata]yzedC-etherdecompositionto about the same extent.Ft=isis it=contrastto the data reportedin reference54, wt=eresilverproduced2 to 5 times the amountof high molecularweightproductsas M-50 under similartest conditions. In additionthere was littledifferencebetweenair and ni=trogen atmospheretests based on high MW productformation.

Severaldiscretepeaks of higherMW were seen on many of the chromate-grams. Peaks appearedat about 450, 600, 700, and 80D as well as a broad peakcenteredat 240U (see fig, 25). It would appearthat couplingreactionsre-sultingin a seriesof higherMW oligomersbasicallydifferingin MW by theadditionof anotherphenyl-sulfurspecies. A processsimilarto that describedfor the polyphenylethers could be operatinghere.

PerfIuoroalkylethers

Perflooroalkylethersare a class of fluidswhich exhibitsexcel'_ntther=realand uxldatlvestability(refs.100 and 101). Combinedwith good viscositycharacteristics(ref. 43), good elastohydrodynamicfilm formingcapabilities(ref. 102), good boundarylubricatingability(refs.46 and 4g), and nonflam--

i abilityproperties(ref.51) make these fluidspromisingcandidatesfor high--t¢:mperaturelubricantapplications.

l_asically,ti_ereare two types of perfluoroalkylethers,an unbranchedanda branchedclass. ]'begeneralstructuresof these classesare simwn in figure26. Tn_ most importantrepresentativesof the branchedmaterials(fig. 2.6(a))

are based on th_ polymerizationof hexafluoropropyleneoxide (HFPO).

1983026872-TSB12

OIdGINPJL.Pt_lil _L_OF POOR (_II/_LITY

Brancl)ed l)erflueroalkylether_

Initial tllermal and oxldatlve stability tests on tile I}rafl_lled cla_swereperformed on very pure, dgly claracterlze(l Jllaterials (ref. 100), This earlydata Indicated tliat tlle_e, inat_:rlal_ slmuhl h(_' thor,lal ly stab _ t_ 4it) C andthat oxygen would not accelerate tilts degradation,

IlowQVer. test;s p_rformed en commercial saml,He,_hover ytulded this 1deal-.tzed _tahtllty, Thermal stah111ty et el}out 390 (. (ref, 49) and oxldattonlcorrosion stabt'llty to _60= C (rof. 46) In tim i}rosencc, of ferrous and tttantumalloys was obtained, 1'his was confirmed by I)aclorek et el, (ref, 47) tn bulkoxtdat ton/corrosion OXl)orimonts,

In tim early work by G,mprocht (ref. 103) he showed that pure (t.e.. cool,_plotel.y, fluorinated)Itl;I)O tlutds tt|(eorotlcally yt!.l(I.mall)lY CF3CFr,CI:;._.C,F3C.OF.and COl.?..Under oxldlzln9conditions(1)acloveKet el,) (tel'.4/) Old nol;el)-serve any CF3CF-CF2. Thls may be due to |:heexidatlonreaction

0

CF3CF-CF2 _ CF3COI"+ COF;) (37)

or more 11kelydue to the unzippingreaction

0

CF-CF._0CF-CF O..--'C%COI'* CoF2 (3.)CF3 CF3

'; The major productsisolatedat test conclusionwere 5iF4, COZ, and BF3. Theseare obviouslydue to reactionef the plimaryproductswith the glass surfaceofthe oxldation-corroslonapparatussuch asl

]

2COF2 + Si02-=-SiF4 + 2C0,_ (39)

Based on the rate of productionof these volatileproducts,it was deter-mined that the commercialHFPO fluidswere oxidatIvelystableto about 343" Cin the absenceof metal. The limiteddegradationat 343 C was thoughtto bedue to oxidativeinstabilityof a weak linkor end group. 5inca at test con-clusionthe residualfluid was essentiallyunchanged(i,e.,same MW, viscosity,and IR spectra),it was deducedthat a certainnumber of cllainshad to behydrogenterminated. Indeed,pretreatmentat 343 C with oxygenylelded amore stablefluid,presumablydue to burningoff tim weak ilydrogenterminate(lchairs. ProtonNMR (fig. 19) eventuallyconfirmedthis theory. The as-receivedfluid shows the preserlceof CF3CFIIO- as a characteristicdoubletat 5.9 ppm. After thermal-oxldativeexposureat 343= C timse peaks completelydisappear(fig.19(c)).

In the preserlceof metals aridalloysat 316° C sucllas Ti (4 AI, 4 Mn),Ti, and Al, the degradationrate is increased(fig,_l) and is higher for thenonpretreatedfluid. This would indicatethat the metals are acceleratingtile

25 ii

1983026872-TSB13

P ORIGINAL P_:;/_!!:

OF POOl1 _I)t'_LI i',:

rh.'_.ir#i(aLi()(of wi_ak llnk_, llow_)vP.r,SlICCO.Q5;'_e i-ll_aSllrelll[}rlt5with the. TI (4 AI,4 Mu) alloy at 21_II"(;._Ilawa c(mtinual upt,_Ic.,_f_x.yg_lwlth l;ilIie,(lbvinu!_lyafl_I;II_rme(:hanlsmI_ ()pP-rativellere, Analy:_is(iftilere_irlualtluid ._II_wL_dadra_tic dncr_as(.,I11MH wirlcllw_i,'IdIrlrli(:atr_tll_Ita chain s_:is_iiOll,}roc_i i_iIik_.,ly,._ucIIafid(.,IHCl;ndl)n.l(iw

ICI:3

A II

Frayllient A can furtt_er degrade and II is just a lower MW t(!'lomer of the; IIFPUt, laid.

_' Therel'ore, it appears that metals and metal alloys promoLe d_y_'adation in: HI-'POfluids by a chain scisslon process, lhe III:PUfluids that have been ther-

mally pretreated at _43° C in oxygen are stable to oxygen at tillstemperature.Nor is the pretreated fluid degraded by M.-50steel or Ti (4 AI, 4 Mn) alluysat 316° C. llowever,degradation does occur at 343" C with these alloys.

_: Unbranched Perfluoroalkylethers

A new class of perfluoroalkylethers based on the photo._oxidationot_luoro-olefins (ref. 104) has been developed, This class of materials, whos__.general chemical structure appears in fi!lure26(b)_ has an unbranched struc_ture (rlopendant CF3 groups). These fluids have better viscosity-temperaturei_ropertiestIianthe branched (III=PO)class (ref. nO), However_ the linear tluidclass has exhibited lower thermal-oxidative staoility compared to the HFPO flu-ids (ref. 50). Tillsis surprising since tilechemical bending in both classesis very similar. In _act, it has been shown (ret. 103) that tertiary carbon-fluorine bonds are less stable than those involving primary or secondary carbonatoms. This would lead one to conclude that the HFPO fluids (which contain

tertiary carbon atoms) sliouIdbe less stable than the linear fluids.

Recent work (ref. 5_J has confirmed tilatthe unbranched fluids are inher-.e tly u_stable at 316° C in oxidizing atmospheres. This instability is rlotdueto hydrogen chain termination or residual peroxide linkages. In the pres'_in'ceof M-SIJsteel or Ti (4 AI, 4 Mn) alloys at 288° arld316° C in oxidizing atmo-splmres, tlreunbranched fluids exhibit much greater degradation than in uncat..llyzedtests (fly. 2B). llowever,these alloys do not promote degradation at316° C in inert (nitrogen) atmospheres.

Irl addition, catalyzed tests indicated that pure metals (TI, AI) did notpromote as much degradation as a11oys [Ii (4 Al, 4 Me)] (fig, _.9). Again, a._,

f

1983026872-TSB14

i ORIGINAL Pt_Gi: _;OF POOR QUALI'I'Y

with the flFP0 fluids, the metals end alloys promoted degeadatton via A chainsclsslonprocesswl|Ichis depictedin flg_Ire30°

Ini_Ibltors

Two differentdegaradatieninhibitorshave been studiedwlth perfluere-alkyletherfluids,a perfluorophenylphosphlne(P-3) (refs.48 and 52) and apho_phatriazine(C_PN3)(ref. 5_).° Both of these Inhibitershave beenshown to be auite effectlveat 2_8_ C in oxygenwith the unbranchedfluids(fig. 31), but are only marginallyeffectiveat 316" C, Similarresponsesoccurredwith the branchedfluids (ref. 53).

MODELINGAND SIMULATION

In the last 10 years, modeling (ref.105) has been increasinglyused inthe study of complexreactions(ref. 106), includingoxidation(ref. i07),Modeling involvesa systemizationof informationfor a pArticulArcomplexreac-tion, This systemizationshould lead to a compilationof all possiblemechan-isms. This resultsin a chaoticset of chemicalreactions.

Next this set of reactionsis systematizedinto a network. In particular,to a sequencenetworkwhich will describethe possiblepathwaysof given atomsvia certainselectedspecies, An exampleof a sequencenetworkfor the oxida-tion of ethylbenzene(RH2) is shown in figure 32. Experimentaldata then pro-rides feedbackfor the initialmechanismcompilation. Reactionsfound not tooccur may be eliminatedwhile othersmay be added. Certainexperiments(invitro)may be performedfor additionalinformation. For example,the impor-tance of a specificsubsetof reactionsmay be studiedby excludingother com-peting reactions.

It is obviouslyimportantto determinethe relationshipbetweenth_reactionmechanismand kinetics. Any mechanismcan be expresssedby thefollowingdifferentialequationsystem

= (41)

where _ is the column vectorof the first derivativeof the speciesconcen-trationswith time, A Is the stoichiometricmatrixdeterminedby the mecha-nism,_ is the rate-constantmatrix of the elementaryprocessesin the mecha-nism, and c is the column vectorof the time dependenceof the speciescon-centration,-

There are three analysispossibilities:(I) Only the importantreactionsare considered(zerosare insertedin

the A matrix for eliminatedreactions)and the reducedsystemsolved.T2) The concentrationsof all speciesare :neasuredand the equations

solved for k.

(3) By _rbitrarilyvaryingthe elementsof k, differentsolutionsmay beobtainedand then comparedwlth experimentaldata. Thls enablesone to deter-mine the importanceof the variousreactionsin the overallmechanism, Thlsapproachis called computersimulation,

27

1983026872-TSC01

p

A de:tailed explanation for a modeling process for the liquid phas=._ oxida-tion of etllylt}enzene appear_ in reference 107. A general introduction to sim-ulation ?,_pears In reference !O_.

GLOSSARY

Free Radical A specl_:s having an unpaired electron bnt no chargt_

llonJnlysis Brookingof a diamagneticmolecule'Intotwo parama|jn(;ti(:

species

Autoxidation (Jxi(latlonof an organiccompoundwith molecu'laroxygen

in the absenceof a flame

Catalyst Any substancethat increasesthe rate of a chemical

reaction

Initiator A substancethat startsa free radicalchain process

Oxidation Electronremovalfrom a chemicalspecies

bisproportionation Transferof an atom from one radicalto anotherforming

a saturatedand an unsaturatedmolecule

Lewis Acid Speciescapableof acceptingan electronpair

SYMBOL;

A preexponentialfactor

AA atomicabsorptionspectroscopy

Are. phenoxyl radical ORIGINAL PA_t_ _i_'

CL chemiluminescence 0;: POOR QUALrrv+

DES di-Z-ethyihexy]sebacate

E activationenergy

ELLIS bond dissociationenergy

ESR electronspin resonancespectroscopy

IlO2, hydroperoxyradicalIn initiator

IR Intraredspectroscopy

k rate constant

kp rate ot chain propagation

kt rate of chain termination_W molecularweight

NMl4 nuclearmagneticresonancespectroscopy

PETH pentaerythrityltetraheptanoate

28

+

• " -- -- ........... i9830-2-6872 TSC02

PTZ phenothtazine

R gas constant

R-N hydrocarbon

R, rate of chain initiation

RO, alkyl free radical

RO_ alkoxyradical

R02_ peroxy radical

RO2H hydroperoxlde

RO4R tetroxideSEC size exclusionchromatography

T absolutetemperature

TD thermaldecompositiontemperatureTCP tricresylphosphate

TMPTH trimethyolpropanetriheptanoate

UV ultravioletspectroscopy

X inhibitor

REFERENCES

I. Bisson,E. E.; and Anderson,W. J.: AdvancedBearingTechnology. NASASP-38, 1964,

D( : 2. Loomis,W. R.; Townsend,D. P.; and Johnson,R. L.: Lubricantsfor

InertedLubricationSystemsin Enginesfor AdvancedAircraft. NASA TN

D-5420,Sept. 1969.

3. Parker,R. J.; Bamberger,E. N.; and Zaretsky,E. V." BearingTorque_'=_ and FatigueLife Studieswith SeveralLubricantsfor Use in the Range

500° to 700° F. NASATN D-3948,May 1967._ 4. Zaretsky,E, V.; and Ludwig,L. P." Advancementsin Bearings,Seals,

_ and Lubricants. AircraftPropulsion,NASA SP-2Sg,1971, pp. 421-463.

L 5. Sliney,H. E.: Bearings,Lubricants,and Seals for the Space Shuttle,Space TransportationSystemTechnologySymposium,Vol. Ill - Structures

and Materials,NASA TM X-52876,1970, pp, Y89-296.6. Bucknell,R. L." Influenceof Fuels and Lubricantson TurbineEngine

Designand Performance,Vol. II - Fuel and LubricantAnalyses.

PWA..FR-5673,AFAPL-TR-13-52~VDL-2,Pratt and WhitneyAircraft,1973,(AD-769309.)

7. Lansdown,A, R.: Liquid Lubricants- Functionsand Requirements._J InterdisciplinaryApproachto LiquidLubricantTechnology,P. M. Ku,I ed, NASA SP-318,1973, pp. 1-55.

29

" ............ 1983026872-TSC02

ORIGINAL PAGI; {_;OF POOR f_UALrI'Y

B. Russell,T. E.; and Mattes,R. E.: A Study of tliyhl'emp(-_ratureFu(_l__md Lul)ricantson SupersonicAircraft/EngineSystem Performance. SAEPaper 740473,Apr. 1974.

9. Kochi, J. K., ed.: Free Radicals. Vol. I, Wiley, 1973.

I0. l_acK,M. II.: Pyrolysisot llydrocarbons.The Mechanismsot Pyruiy_is,()xidation,and Burningr)tOrganicMaterials,L. A. Wall, ed., NIISSP 357, 1972, pp. 17-31.

11. Layokun,S. K.; and S1ater,D. li.: Mechanismand Kineticsof PropanePyrolysis. Incl.Eng. Chem. Process.Des. Dev., vol. 18, no. 2, 1979,pp. P32-236.

12. Blake, EdwardS.; et al." ThermalStaoilityas a Functionof ChemicalStructure. J. Chem. Eng. Data, vol. 6, no. I, Jan. 1961, pp. 87-98.

13. Vapor Pressure-TemperatureRelatlonshipand InitialDecompositionTemperatureof Liquidsby Isoteniscope. Am. Soc. Test. Mater. Stand.D-2879-75,Part No. 24, 1980.

; 14. Fowler,L.; and Trump,W. N." A RecordingTensimeterand Decomposition

I TemperatureDetector. Analysis Instrumentation,Vol. 7. B. Connelly,I..Fowler,and R. Krueger,eds., ISA, 1969,pp. 141-144.

15. Jones,W. R., Jr.; and Hady, W. F.: BoundaryLubricationand Ther,nalStabilityStudieswith Five LiquidLubricantsin Nitrogento 400° C.NASA TN D-6251,Mar. 1971.

16. Beerbower,A.: EnvironmentalCapabilityof Liquid Lubricants. Inter-disciplinaryApproachto Liquid LubricantTechnology,P. M. Ku, ed.,NASA SP-318,1973, pp. 365-431.

17_ Jones,W. R., Jr.: Friction,Wear, and ThermalStabilityStudiesofSome Organotinand OrganosiliconCompounds. NASA TN D-7175,Mar. 1973.

18. Richardson,G. A.; and Blake,E. S..' PartiallyFluorinatedPoIyaromaticEthers: Synthesis,ThermalStability,and Other PhysicalProperties.USAF AerospaceFluidsand LubricantsConf.,P. M. Ku, ed., SouthwestResearch Inst.,1963,pp. 130-139.

19. Gunderson,R. C.; and Hart, A. W.: SyntheticLubricants. ReinholdPubl. Corp., 1962.

2U. Klaus, E. E.; and Perez, J. M.: ThermalStabilityCharacteristicsotSome MineralOil and HydrocarbonHydraulicFluidsand Lubricants. ASLETrans.,vol. 10, 1967, pp. 38-47.

21. Van Krevelen,D. W.: Propertiesof Polymers,Their EstimationandCorrelationwith ChemicalStructure. Seconded., ElsevierScientificPubl. Co. (Amsterdam),1976.

30

1983026872-TSC04

22. Ugwuzor, D.I.K.A.: The Effects of Metals on Htgh Temperature geBrada-tlon of Ester-Type Lubricants. M.S. Thesis, Penn. State Univ., Nov.1982.

23. Backstrom, H.L.J.: The Chain-Reaction Theory of Ne9atlve Catalysts,J, Am, Chem. Soc., vol. 49, no. 6, June 1927, pp. 1460-1472.

24. Crtegee, R.; Ptlz, H.; and FlYBare, H.: Oleftn Peroxides, Chem. Ber.,vol, 72B, 1939, pp. 1799-1804.

25, Emanuel, N. M.; Dentsov, E. T.; and Matzus, Z. K. {(B. J, Hazzard,Transl.)}: Ltquld Phase Oxidation of Hydrocarbons. Plenum, 1967.

26. Dukek, W. G.: Fuels and Lubricants for the Next Generation Aircraft -the Supersonic Transport. J. Inst. Pet. London, vol. 50, no, 491, 1964,pp. 273-296.

27. Booser, E. R.: Liquid-Phase Oxidation of Pure Hydrocarbons. Ph,D.Thests, The Pennsylvania State University, 1968.

28. Czarneckl, J. R.: HlBh Temperature De9radatlon of SomeOr9anlc Esters.M.S. Thesis, The Pennsylvania State University, Sept. 1971.

29. All, A. R.: Characterization of Liquid Phase Oxidation Products. M.S.Thesis, Pennsylvania State University, 1975.

30. Cvttkovlc, E.: Reaction Rate Studies on Ester Oxidation. M.S. Thesis,Pennsylvania State University, May 1976.

31. Lahijant, J.: Effects of Metals on Oxidation Rates and Products fromEsters. M.S. Thesis, Pennsylvania State University, 1977,

32, Lockwood,F. E.: Ester Oxidation Under Simulated Boundary LubricationConditions. Ph.D. Thesis, Pennsylvania State University, 1978.

33. Korcek, S.; and Jensen, R. K.: Relation Between Base Otl Compositionand Oxidation Stability at Increased Temperatures. ASLETrans.,vol. 19, no. 2, 1976, pp. 83-94.

34. Jensen, R. K.; et al.: Liquid-Phase Autoxldatlon of Organtc Compoundsat Elevated Temperatures. 1. The Sttrred Flow Reactor Technique andAnalysis of Primary Products from n-Hexadecane Autoxtdatlon at120"-180" C. J. Am. Chem. Soc., vo]. 101, no, 25, Dec. 1979,pp. 7574-7584.

35. Jensen, R. K.; et al.: Liquid-Phase Autoxidation of 09rantc Compoundsat Elevated Temperatures. 2. Klnet;cs and Mechanism of the Formation ofClevage Products tn n-HexadecaneAutoxldation. J. Am. Chem. Soc.,vo]. 103, no. 7, Apr, 1981, pp. 1742-1749.

36. Hamilton, E. J.; et al.: Ktnetlcs and Mechanismof the Autoxtdatton ofPentaerythrity] Tetraheptanoate at 180-220 ° C. Int. J. Chem, Kinet.,Vol. 12, no. 9, 1980, pp. 577-603.

31I

1983026872-TSC05

_]7.Maho_ley,I. R.; et el.: l_ffectsef Structur[:on the Thermexid_ll:IveStahilltyet Synthe.ti("Ester Lubricants: Theory and PredictiveMethodDevelepment. Presentedat the Div. ef Petrel.them,,Amer. Chem. Sec.(las Vegas,Nev.),M_r. 2(_.Apr.2, 1982.

3_. Boss, li.D,; and llazlett,R. N.: Oxidationof llydrocarbonstn theLiquidPhase - n-l)odecanein a BorosilicateGlass (]namberat 20()° C.Can. J. Chem. vel. 47, 1969, pp, 4175_41t12.

39. Mill, T.; et al.: Gasand l.iquid-PhaseOxidationel_ n.lIutane.J. Am.Chem. Soc., vol. 94, no. ]9, Sept. 1972,pp. 6802_6811.

40. Van Sickle,D. E.; et al.: IntramolecularPropagationin the Oxidationof n-Alkanes,Autoxidationof n-Pentaneand n-Octane. J. Org. Chem.,vol. 38, no. 26, 1973, p. 4435.

41. Brown, O. M.; and Fish, A.' The Extensionto Long-ChainAlkanesand toHigh Temperaturesof the HydroperoxideChain Mechanismof Autoxidation.Prec. R. Soc. LondonSer. A, vol. 308, no. 1495, Jan. 1969, pp. 547-568.

42. Sniegoski,P. J.: Selectivityof the OxidativeAttackon a Model EsterLubricant. ASLE Trans.,vol. 20, no. 4, 1977, pp. 282-286.

43. Betts,J.: Kineticsof HydrocarbonAutoxidationin the LiquidPhase.Quart. Rev. Chem. Soc., vol. 25, no. 2, 1971, pp. 265-288.

44. George,P.; Rideal,E. K.; and Robertson,A.: Oxidationof LiquidHydrocarbons. Prec. R. Soc. London,voI. 185, no. 1002, Mar. 1946,pR. 288-351.

45. Morton,F.; and Bell, R. T. T.: Low-Temperature,Liquid-PhaseOxidationof Hydrocarbons. J. Inst. Pet. London,vol. 44, 1958, pp. 260-272.

46. Snyder,C. E., Jr.; and Dolle,R. E., Jr.: Developmentof Po|yper--- fluroalkylethersas High TemperatureLubricantsand HydraulicFluids.

ASLE Trans.,vol. 19, 1976, pp. 171-180.

47. Paciorek,K. J. L.; et al.: ThermalOxidativeStudiesof Poly(HexafluoropropeneOxide) Fluids. J. Appl. Polym.Sci., vol. 24, no. 6,1979, pp. 1397-1411.

48. Snyder,C. E. Jr.; et al.: Synthesisand Developmentof Improvedlligh-TemperatureAdditivesfor Perfluoroalkylethe_Lubricantsand Hydraulic

_ Fluid. Lubr. Eng., vol. 35, no. 8, Aug. 1979, pp. 451-456.

---- 49. Jor_es,W. R. Jr.; and Snyder,C. E., Jr.: BoundaryLubrication,Thermaland OxidativeStabilityof a FluorinatedPolyetherand a Perfluoro-polyetherTriazine. ASLE Trans.,re). 23, no. 3, July 1980, pp. 253-261.

50. Snyder,C. E. Jr.; Gschwender,L. J.; and Tamborski,C.: Linear Poly-perfluoroalkylether- Based Wide-Liquid-RangeHigh-TemperatureFluids

_!_-_ and Lubricants. Lubr. Eng., vol. 37, no. 6, June 1981, pp. 344-349.

] 983026872-1-SC06

51. Snyder, C, E, Jr.; Gschwender, L, J.; and Cambe11, g. B,: Developmentand MechanlcalEvaluatlonof NonflammableAerospace(-54° C to 135" C)HydraulicFluids. l.ubr.Eng., vol. 38, no. I, Jan. 1982, pp. 41-61,

52. Jones,W. R. Jr.; et el.: ThermalOxidativeDegradationReactionsofLinearPerfluoroalkylethers.NASA TM_82834,Apr. 1982,

: 53. Jones,W. R. Jr.; et el.: Metals and InhibltorsEffectson ThermalOxidativeDegredatlonReactionsof Perfluoroalkylethers.Presentedatthe Sixth Winter Symposiumon FluorineChemistry(DaytonaBeach,Fla.),Feb. 6-11, 1983.

54. Clark,F. S.; and Miller,D. R._ Formulationand Evaluationof C-EtherFluids as LubricantsUseful to 260° C. (MRC-SL-IO07,MonsantoResearchCorp.;NASA Contr_cLNAS3-19746.)NASA CR-159794,Dec. 1980.

55. Wilson,G. R.; Smith, J. 0.; end Stemniski,Jr.: Mechanismof Oxidationof PolyphenylEthers. ML TDR-64-98,Wrlght-PattersonAFB, Dec. 1963(Availableas AD-457120).

B_. Jones,W. R. Jr.; and Morales,W.: Thermaland OxidativeDegradationStudiesof FormulatedC-Ethersby Gel-PermeationChromatography.NASATP-1994,Mar. 1982.

57. Sheldon,R. A.; and Kochi,J. K.: Metal-CatalyzedOxidationsof OrganicCompounds. AcademicPress, 1981.

58. Emanuel,N. M.: 8O's in the Field of LiquidPhase Oxidationof OrganicCompounds. OxidationCommunications,vol. 2, No. 3-4, 1982, pp. 221-238.

59. Korcek,S.; et al.: AbsoluteRate ConstantsFor HydrocarbonAutoxida-tion XXI. ActivationEnergiesfor Propagationand the CorrelationofPropagationRate Constantswith Carbon-HydrogenBond Strengths,Can. J.Chem.,vo]. 50, 1972, pp. 2285-2297.

60. Howard,J. A.: AbsoluteRate Constantsfor Reactionsof Oxyl Radicals.Advancesin Free-RadicalChemistry,G. H. Williams,ed., AcademicPress,1972, pp. 49-165.

61. Bennett,J. E.; Brown, D. M.; and Mile, B.: ElectronSpin Resonanceofthe Reactionsof AlkylperoxyRadicals. I. AbsoluteRate Constantsforthe TerminationReactionsof AlkylperoxyRadicals. Trans.FaradaySoc.,vol. 66, no. 2, 1970, po. 386-396.

62. Dornte, R. W.: Oxldat!on of White Oils. Ind. Eng. Chem., vol. 28, Jan.1936, pp. 26-30.

63. Bolland, J. L._ Ktnetlc Studies in the Chemistry of Rubber and RelatedMaterials. I. The ThermalOxidationof Ethyl Linoleate. Prec.R. Soc.London,vo]. 186A, no. 1005, July 1946, pp. 218-236.

64. Diamond,H.; Kennedy,H. C.; Larsen,R. G.: OxidationCharacteristicsof LubricatingOils at High Temperature. Ind. Eng. Chem.,vol. 44,1952, pp. 1834-1843.

33

i

1983026872-TSC07

!

P

6_. Brook, J. II. T.= A Clr_u]atory (011) Oxidation Test. J. Inst. I)(_t.London, vol. 4_, 1962, pp. 7_12.

66. liepplewhite,H. L.; and Ob_rrlght,E. A.: Thin Film ()xldatlonTest:otLubricantsfor Gas ]urbln_En_Ine_.Prec. USAF AerospaceFluids and LubricantsConf., P. M. Ku, od.,SouthwestResearchInst.,1963, pp. 62-69.

_7. Cvitkovic,E.; Klans,E. E.; and Lockwood,F.= A Thin-Filmlesl;ForMeasurementof the Oxidationand Evaporationof Ester-TypeLubricants.ASLE Trans.,vol. 22, no. 4, 1979,pp. 395-401.

68. Jones,W, R. ,Jr.;and Mora|es,W.= Analysisof a MIL-L-27502LubricantFrom a Gas-TurbineEngineTest by Size-ExcluslonChromatography.NASATP-2063,Jan. 1983_

69. Klaus, E. E.; et al.: StabilityStudiesof OrganicAcid EsterLubricants. PennsylvaniaState Univ. AnnualRep. =or NASA GrantNAG3-35,Mar. 1981.

70. Denbigh,K. G.: Velocityand Yield in Continuous-ReactionSystems.Trans. FaradaySo(:.,vol. 40, 1944, pp. 352-373.

71. Mahoney, L. R.; et al.: Time-TemperatureStudiesof High TemperatureDeteriorationPhenomenain LubricantSystems= SyntheticEster Lubri-cants. AFOSR TR-80-0065,Ford Motor Co., 1979. (AD-A080135.)

12. Ravner,H.; and WobltJen,H.: The Determinationof the OxidativeStabilityof Sewral DeuteratedLubricantsby an ElectronicGas Sensor.ASLE Preprint82-AM-5A-4,May 1982. _

73. Morales,W.: Use of High PressureLiquidChromatographyin the Study ofLiquid LubricantOxidation. NASA TM-83033,Oct. 1982.

74. LubricatingOil, AircrattTurbineEngine,SyntheticBase. MIL-L-23699C,Dept. Of Defense,June 1981.

75. Assenheim,H.M.: Introductionto ElectronSpln Resonance. Hilger andWatts, LTD (London),1966.

76. Gould, E. S.: InorganicReactionsand Structure. Halt, Rinehart,andWinston, Inc., 1962.

77. Harwy, E. N.: Bio|uminescence,AcademicPress, 1952.

i._ 78. Vassil'ev,R. F.: Chemiluminescencein Liquid-PhaseReactions. Pro-9ress in ReactionKinetics,voi. 4, G. Porter,ed., PergamonPress,1967, pp. 305-352.

79. S|awinski,J.: The Use of Chemiluminescencefor Investigationof theKineticsof the Oxidationfo Hydrocarbonsin the Liquid Phase. Int.Chem. Engr.,vol. 6, no. I, Jan. 1966,pp. 160-162.

34

1983026872-T8C08

b •

HO, Lloyd, R. A.: Law Level Chemiluminescence from Hydrocarbon AutoxldattonReactions. Part I. Apparatus for Studytn9 Thermal Decomposition Reac_lions and Observations on Benzoyl Peroxide tn De-oxy9enated Benzene,Trans. Faraday Soc., vol. 61, no. 514, 1966, pp. 2173-2181.

81. Lloyd, R. A.: Thermal Decomposition of Benzoyl Peroxide, CumeneHydro-peroxide, and UV Irradiated Solvents. Part II, Low Level Chemilumines-cence From Hydrocarbon Auto-oxidation Reactions. Trans. Faraday Soc.,vol. 61, no, 514, 1965, pp. 2182-2193.

B?. ttercules, D. M.: Physical Basis of Chemiluminescence. The CurrentStatus of Liquid Scintillation Countln9, E. O, Bransome, ed,, Grune andStratton, 1970, pp. 315-336,

83. Clark, D, B.; Weeks, S. J.; and Hsu, S. M.: Chemiluminescenceof Fuelsand Lubricants ° A Crtttcal Review. ASLEPreprlnt 82-AM-5A-1, May 1982.

84, Belford, R. L.; and Strehlow, R. A.: Shock Tube Technique tn ChemicalKinetics. Ann. Rev. Phys. Chem., vol. 20, 1969, pp. 247-272.

85. Khandelwol,S. C.; and Skinner,G. B.: Shock Tube Studiesof )i "' HydrocarbonOxidationShock Waves in Chemistry, A. Llftshitz,ed.,

Marcel Dekker,1981, pp. 1-57.

86. Larsen,R. G.; Thorpe,R. E.; and Armfield,F. A.: OxidationCharacter-isticsof Pure Hydrocarbons. Ind. Eng. Chem.,vol. 34, no. 2, Feb.1942, pp. 183-193.

)87. All, A.; et al.: The ChemicalDegradationof Ester Lubricants. ASLE

i Trans., vol, 22, no. 3, _uly 1979, pp. 267-276.

88. Lockwood,F.; and K1aus, E. E.: Ester Oxidation- The Effectof an IronSurface. ASLETrans., vol. 25, no. 2, 1982, pp. 236-244.

89. Mahoney, C. L.; et al.: Meta-Linked Po]yphenyi Ethers as High-Temperature Radiation-Resistant Lubricants. ASLETrans., vo]. 3, no. 1,

_" Apr. 1960, pp. 83-92.a--

90. Jones, W. R., Jr.; Hady, W. F.; and Swtkert, M. A.: Lubrication WithSomePo]yphenyl Ethers and Superrefined Mineral Ol]s in a 600" F(316" C) InertedVane Pump Loop. NASA TN D-5096,Mar. i969.

i

i91. Archer,W, L.; and Bozer, K, B.: OxidativeOegradatlonof the Poly-

i(_ pheny] Ethers. Ind. Eng. Chem. Prod. Res. Oeve]op., vo]. 5t no, 2, June

!1966, pp. 145-149.

92, Smith, J. 0.; et al.: Research on Htgh Temperature Additives fop Lubri-cants. WADC-TR-59-191, Part IV, Feb. 1962. (Available as AD 281831.)

i'_ 93. Ravner, H.; Russ, E, R.; and Ttmmons, C. 0.: Ant]oxidant Action ofMetals and Metal-OrganicSalts on F1uorestersand PolyphenylEthers.

_=_ J. Chem. Eng. Data, vol. B, no. 4, Oct. 1963, pp. 591-596.

i.p. 35

r

1983026872-1-SC09

i i

I

: 94. Ravner,li.;Monlz,W. B.I and Blachly,C. H.: Hlgh TempertureStabili_zation of Polyphenyl Ethers by Inor9antc Salts. ASLETrans., vol. 15,no. 1, 1972o pp. 45_53,

95, Ravner, H.; and Kaufman, S,: High-Temperature Stabt|lzatlon of Poly-phenyl Ethers by Soluble Metel-Organic Salts. ASLETrens., vol. 18,no. 1, 1975, pp, I-4.

g6. Stemntski, _. R.; et al.: hnttoxidants for High-Temperature Lubri-cants. ASL_ Trans., vol. 7, 1964, pp. 43-54.

97. McHugh, K. L.; and Stark, L. R.: Properties of a New Class of Poly-aromatics for Use as High-Temperature Lubricants and Functional Fluids.ASLE Trans., vol. g, no. 1, Jan, 1966, ppo 13-23.

98. Jones, W. R., Jr.: Boundary Lubrication of Formulated C-Ethers in AirL to 300" C. Lubr. Eng., VOlo 32, no. 10, 1976, pp. 530-538.I

99. Jones, W. R., Jr.: The Effect of OxygenConcentration on the Boundary-Lubricating Characteristics of a C-Ether and a Poiyphenyl Ether to300° C. Wear, voi. 73, 1981, pp. 123-136.

100. Gumprecht, W. H.: PR-143-A New Class of High Temperature Flutds. ASLETrans., vol. 9, no. 1, Jan. 1966, pp. 24-30.

101. Slanost, O.; et al.: Perfluoropolyethers - Their Physical Propertiesand Behavior at High and Low Temperatures. Wear, vol. 18, 1971,pp. 85-100.

lO2+onesw, e, Pressureviscosit,,e.so,emootsforse.ralLubricants to 1 Newtons per Square Meter (8x10_ psi) and149" C (300" F). ASLETrans., vol. 18, no. 4, 1975, pp. 249-262.

103. Gumprecht, W. H.: The Preparation and Thermal Behavior of Hexafluoro-propylene Epoxide Polymers. Presented at the Fourth InternationalSymposiumon Fluorine Chemistry (Estes Perk, Colorado), July 1967.

104. Sianesi, O.; et al.: Perfluoropolyethers by Photo-Oxidation of F1uoro-olefins. Chlm. Ind. (Milan), vo]. 55, no 2, 1973, pp. 208-_21.

105. Franks, R. G. E.: Modeling and Simulation and Chemtcal Engineering.John Wiley and Sons, Inc., 1972.

ZOb. Happel_ J.; and Sellers, P. H.: Multtple Remction Hechantsms inCatalysis. Ind. En9. Chem., Fundam., voZ. 21, no. 1, 1982, pp. 67-76.

1U7. Gal, O.; et a1.: SomeAspects of Hodeltn 9 Hydrocarbon Oxidation.Eighteenth Symposium(International) on Combustion, The CombustionInstitute, 1961, pp. 1321-1331.

108. Schmtdt, J. W.: Funda+entals of Digital Simulation Modeling. 1981Winter Slmuletton Conference Proceedings, Vol. 1, T. F. Oren, C. H.Oelfosse, and C. H. Shub, eds., IEEE, 1981, pp. 13-21.

i

36 i)

: !

_ .......1983026872"' ' '

TAOI.EI. - THERMALDECOMPOSITIONTEMPERATURESAND

BONDDISSOCIATIONI_NERG!ESFOR VARIOUSCOMPOUNDS

Compound Bonds EDXS,a TD,kJ/mole "C

Octacosane C_C 337 350

11-ethyl-11-methyl C 314 331

pentacosane -C-CC

p-quartaerpheny] _-_ 432 454

Polyphenylether(5P-4E) _-0-_ 423 443

E-bls(E-chlorophenoxy) _-CI 419 409benzene

p-bls(p-bromophenoxy) _-Br 335 387benzene

Fluorinatedpolyether CF3-CF3 406 390Syntheticparafflnic C-C 337 314

Alkylatedbenzene _-C-CH 335 340

aFromref.21. I

iI ,_

f_

1983026872-TSC 11

IAI_LEII, - OXII)IZABILITYOF VhRIt}U_ORfiANICCOMPOIJNDS

.... [From raf. 60.]

Subatrata kp/( ?.kt )1/_x1Q3(M_I/_aac,_11?}

2,3-1)hne.thyl_2-bl_te,e 3.

Cyc]ohexo.ne , 7.3

L,(}cLen_) .06

l!thylbenzerle ,P.I

'l'oIuem,_ .01

_.-Xyleno .OB

Bonal dehyde 290

Benzylalcohol .85

2,4.6-Trlmethylhoptane .09

I

1983026872-TSC 12

k

I'AI_LE III. ,-X_II BONI_ENERGIES ORIG|NtIL PAGe's'_[From ref, _l.] OF POOR QIIt_LITY

Comp_imd F_nergy,kcal/m_l

Clt3-,It 103

n-_C3117-tl 9g

_t-C3117-li 94

_-C4tlg-H 90

CIt2,,Ctl.--It 105

C6Hr,-H 103L)

CH2"CH-CHp-H 85

PhCH2-H 85RCO-H 86

CH3S-H 88

CH3PH-It 85PhO-H 88

PhNH-H 80

ROO-.H 90I

J

1983026872-TSC 13

TABLF IV, - R_,TECgN._TAN'ISPERLAI_IL£ tiYl]R_%l:NF(_ll

REACFION01: StlBSTRATi_SWITIt THEIR OWNPEROXY

RAI]ICALS(kp) ANI)WITit Cert_ BUTYLPEROXY

RAI)I_ALS(k_,)_t ,_o'e.[From rn,f. !_7.

k /k'Sl_b;;t.r_lt_! kl} kp I} P

(M-1_nc-l,) (M"I,_oc-I

I-0ctene O,5 O,Oil4 6,0I •(,y_.lohexene 1.5 ,80 1.9

Cyclopentol_e I,7 ,85 2,_)J

2,3-1)imethyl-R-butene I.14 ,14 1.0

Toluene .()8 ,01_ 6.7

Ethylberlzene ,65 .I0 6,5

Cumene .18 .22= .g

l'etralin 1.6 .5 3,?.

Benzylether 7.5 .3 ?.5,0

l}enzyIalcohol 2.4 .065 37.0

BeHzylacetate 2.3 .0075 307

lJenzylchlorlde 1.50 .008 190

P IJenzylbromide .6 ,006 100t_

" Beny] cyanide 1,56 ,01 156 _z:

E I_enzatdehyde 33 000 ,85 _ 10 000 }

_ " 1983026872-TSC 14

I

TABL_V, - _FFECTIVENESSOFINHIBITION

_VVARIOUSPHENOLS

[Fromref. _.]

Pllenol 0°02?

2.Methy_phenol .17

4-t-Butylphenol ,05

4-Methylphenol .095

_-t-gutylphenol ,26

2,5-Dimethylphenol .314

2_4-Di-t-butylphenol .465

2,4,6-Tri-t-butylphenol .47

2-Methyl-4-t-butylpheno| .515

2,6-Oi-t-butyl-4-methylphenol ,65

"" 2,4,6-Trlmethylphenol 1.00

2t4-Ol-t-butyl-6-methylphenol 1.05.i

.................. ........... 198"0"68""3z _c....................... ...................................................................":........................... TS[

I,

ORIGINAL PAGE ISOF POOR QUALITY

Electronicsmodule

Flguro1. - Recordingtenslmeter.

Slopo._ proportional,= toactivation= energy-,

r,. %%

_" Thermal'_ _ decomposition

<_ temperature

150 20O 4OO _[_Tenlperaturo,_C

Figure2. - Typlcotthermaldecompositioncurveforsynthetichydrocarbon,Heat..IngIntorvalopoC.

1983026872-TSD02

ORIGINAL PAtti ISOF POOR QUALrI'_

]lli 1_) Syntheticparoffhl

A Pnrofflnlc.r_slnC) lil_ol llorlvallv_

'_ 13 Advancedof,tar

,_ • TltonualilacoelpesltlontoiTiporature,1D

,o

25O 300 350Temperature,°C

Figure3, - Usefullifo(basedonlO-parcentdecomposition)offourlubricantsin nitro-genasfunctionoftemperature(rot. 15).

35t

_ 0____l___!_ J,- 300 ---_ 351) 4"_ 4POBonddissociationenergy(Edls),

kJImole

Figure4. - [hern_ofdecompositiontemperature(Td)asafunctionof

J

1983026872-TSD03

ORIGINALPAGE|_OF POORQUALITY

45_- / N

35_

_ 250

I OBiI

,4 .5 .6 ,7 .BRec[procalofbondlength,k

FigureP, - Thermaldecompos_tlontempera-lure (Td)asarunctlo;,ofthereciprocalbondlengthfor a seriesof9 'oupV aro-maticcompoundsofthetype_C6HpInM[ref. Z2).

: 1983o26872:-r,sD[

UHIGII@_L,P/_ |4OF pOOR QUALI'_'Y

,6

c,¢ '_

,2 ,4 .6 .8 l,OConcentration(molefractlonl of

ethyl Itnoleate, rll

Figure 7, - Oxidationrate as afunction of concentrationforethyl llnoleate(raf. 25).

1983026872-TSD05

ORIGINN,. pAr_E I_OF pOOR rJUN..ITY

jsd

Time

Figure 9. - Typesof kineticcurvesof tho absorptionof oxy_n(reto_h

Alrtluw

1) P':mHeated | / l lUI)I

l /l ".".--r o:}_,eIf--',J'f--- $alllpla..... Catalyst

.................. r \1 J i

L 2.,cm_

Figure ]0 - Mlcro-r,xldetlonapparatus(refs. 30 and66),

it

1983026872-TSD06

ORIGINALPP,GE !_OF POORQUALITY

Itydrocarbon,Ittl.,_J

1 f._ _Outlet02

Pyrex ,_.,._._ )spheres -_:- _._l_l

Figure1]. - Stirredflowmieroreactor_ret.341.

,-4mint.I),/

HolderpositlonerI I

/ Itofit over , I I ,iI sampleholder)

_L I I

Detailsoflpo_sltlonlng

- 450 mm Spacer sampleholder

, _.i -H°16er-.* *-- 3 millO.D,

I ,,- Metal

sltlor_ /t_e -/

. SpacerJ__Deters_f r_t_

___ ltl mm specimenarrallgement

I:lgure12, - Thermaloxidativedecompositiontube,

1983026872-TSD07

OF POURQLIALFIy

Figure13. - Oxidativedegradationapparatususln9 anelectronicgassensor(ref, 72L

ORIGINAL,PAGEISCF POORQUALITY

- Molac,larwalfiht

TCP+

I}ODPh

Ultravioletdetector(2P4nm)

response

1

PANA

(a)Ultravioletdetector.

Flqure1_ - Sizeexclusionchromato_ramofan unusedformu-latedester(G-mll-99)tref.?_.

EsterbasastockRefractive

Indexdetector

TCP4"

)ODPA

"---Nalecularweighttb)Refractiveindexdetector.

Figure15.- Concluded.

1983026872-TSD09

pORIGINALP_N|_I_OF pOOR_|IIL.I-I'TY

I.uhrlcalll I:J:,uru

s_lnplu /_. _._I*lhnln'ypeekI)unemJ 4 1_'!

------ NIl_l'3illlr 5 Jr I

...... ^fiordPhr () _ _

I .... I...........I.............. I.... J.................... I(_))_tdetuctertlOx).

_. ' • region

.... Primarypeak

i I

Molecularweight(b) UVdetector(1AUFS).I

Figure10, - Summeryofsizeexclusionchromatogramsfor unused_0-anddS-hrsamplue(ref.68),

9Ion

.z( .--_

o ................... i .............. j .............. l ................... .......... I .!

Figure17,- Ultruvlolet-vlelblespectrafromsizeexclusionseparationofoxidizedester(d5-hrsampleD,

1983026872-TSD10

iJ

ORIGINALPt_(_] 19OF POORQIJAI,I'I¥

I_ ldlB l_O I132 b98

1 II re,EsterI,a_stocl_

weight. _00 16_3 12_ 699Wevanumbers

(hiTricresylphosphate

Figure18. - Inlraredspectraof extdlzedesterseparatedbysizeexclusionchromatography,

8 7 O 5 4 _ 2 I 0ppm

j s _e_Fluorl.at_p_gether_esr_eI_.

I

I ...........

--J H_LL8 7 6 P 4 t 2 I 98765432 lO

ppln ppm

(b)C3FTOCF(CF3Itl, (c)Fluorinatedpolyether(afterthermalpre-treatmentin 02at3430C).

Figure19. - ProtenNMRspectre,referencedto tatrarnethy|sllaneIT,s) Iret. 53).

1983026872-TSD1

OF PUUIil,!l.iAIII'_'

1983026872-TSD12

i

ORIGINAL P,oitD_t_','i_7

OF POOR QU/ILIW

I ll,dOJ ,if r llfll' I_AIII I_1

ll_i_ilJlll _II _ IlQOllfJ_l §lOW l_J

Ill)l, _* ,, tIl@Oll III-OW 141

,llO(lliof_l _'I _ IIOQIlOI* FI*§T l_lJi*

IIUOllOll,_llll ,A n Ill]Ol_OOIlill I Ill,l)W I_fJ

IIO01Wi, ,_ ,, IIUO_O01I aI,OW(4'1IIC_I_OII it t ,. tlo_rtlo(,,oll i°A_t (n}

• ,y. Faoiv,o,_. ,comli-_-cu) OLOWI_1

lllll lll_" =ill "0"_' _lll_lll,,_lll -_1#1=_lll -I_1fl I p_ TII

K(ll,lllOl_ i 14001101i XO.lllll I _ IlO_l_Oi _Oell O,

Flguro22, - Reaetlor_schemefortheauto-xldotlonofPETII(rof. 3GI,

-- Hours Peak 11i oxidized hehjht.I', at533K orb, urllts

L-- 0 II.0

F --- 21 6.7_ ....... 47 d,0

k

:_ - Fractionl< ..'" ",. Fraction2

--., ...,, _" raetlon

lOgO0 _t300 2rig0IIIUli _00 200_w

_. L........ l 1 I l I I

" llllntJl_Figure_3. - Gelporeleatloilchroll_ato(Jral,Sobtainedfor oxidizeddl=2-

f ethylhexylsebacato,nocatalyst(ref. 871.

I,

+_"_:_:_....... :;_2z2;///il;;:.'_2 ........::':_'21 'I,'_Z],_ ]...................171_........i-_i.Z ........if__ T_IZ ' i

1983026872-TS D13

L)f++POLIR(jl_t+L,t+Y

OF POOR QUALITY

C_I7("(CF_rpOll)Rt

CF3(hi iloxofluorol_rol)Ylonooxhlo(III:PO)IJaso(II)rlli_chodfh_lds.

I_IOtCF'2CF_O)m(CI:20)nRf(I)lUnl)ratlchodporfltzoroalkyletherfluids.

Fkjuro,_, - ChondcalstructurosofI)orfluoroall_yletherfluids(let• CF3orC,_:5).

100--

[] PrefreatedIn OZat343°C

_ Notpretreated"_lO

.g,E

iiii!ii!ii

iii!i iiiii!ii iiiiiiiii

_ iii!ili

I

iiii;iil I _

_!',ii! I i!i!i!i!!!-1.UI 1114AI, 4 IVin) AI

Figur_Z/, - Effectofmetalsond_radaf[onofbranchedperfhmroalkylefher_6° C, 02, _4hr (r_f. 53L

1983026872-TSE01

(_i,.;&!r')!& F'n ','L",l_lOi; PO01_ _UI_LI'I'Y

M'50stool :='i',_

, '[1(AI, 41Vh/) :,;....

'it :" ,',:,,#ill r!,'_ IJl

.llj I Iiii

.... fl ,'l//'/J ,i L_, _/,,_f/// i ,,

I-.. -- ,_i.7.:. "///.,,_,

= V//. "=8 ..... : _._. "///

,:% _, ////_

///

.Ol C, N2 2_8oC, 02 g16oC.02

Figure28. - Ratesofcondensibleproducttormotlonfor unbranchedfluoroalkyletherIn thepresenceofmetalalloys(ref. 52).

1983026872-TSE02

OF POOR QUALITY

MFX•_APr'"(_F2COF '_C 3 0 C;_._AA, -,

DC"M$ CII_OHC_:50 CF2 0 CF3

C_:50 CF2 0 CF2(;1:20CF3+

C21:50CF2 OCF2 CF20 CF20 CF34

CF30 CF2Cl.2 CO2CH3+

CF30 CF2CF20 C,"2CF20 CF2CO2CiI3

Figure30. - Degradationof unbranchedperfluordalkylethersin thepresenceofTitdAI, 4Mn) at2880Cin 02(ref. 53).

100_ _ _ NoadditiveZ P-3- _ C2PN3

_ I _" _ ._

£ -

kl_r

• _ i':_l,_!&8oC,O2, M-SO 310oC,0b M-SO

Figure31. - Effectofdegradofioninhibltereonthe rateofcondensibleproductformationofanunbranchedperllu-oroalkvletherIrof. 52).

,_ i '_

1983026872-TSE03

!

w13

_i wl4 .o,' Fig,re 37. - Carbon-skeletonsequencenetworkof theoxida-

tionofethylbenzenein thetemperaturerangefrom400Cto130oC basedonthepossiblemechanism(ref. 1071,

1983026872-TSE04