'WEAR

ELSEVIER Wear199 (1996) 169-180

Computational methods applied to oxidational wear |

T.FJ. Quinn
Wear C'ons~dtant..~8River Ban~ Road. Kendal. CumbriaLA9 5.15, UE
Received6July 1995;a~¢epted2 November1995

Abstract

Theories of oxidational weartend to involvethe numberof contacts (Nj betweensliding triba-elememsas wellas the contactlempenture
(Tv), i.e. the temperatureat the real areasof contact tglwecn these tribe.elements,This paperdescribeshow these importantparametencan
t~ found for a given tribe-system,namely high.chromiumfcrride steel pins sliding against rotating austcnitic stfinl:ss steel disks, by the
applicationof a numerical toothedto the analysisof the theo~ical and experimentaldivisionsof heat (DTFiand DE) belw¢¢ntribe.elements
undergoingofidational wear. The paper then shows how on~ can use these values of N and TFas inputs to a similar uun~rical methodin
order to deduce credible and eonsis2¢mvalues for the tribolugiea!oxidation constantsrequired for the th~mtical oxidational wear rale.s
( WRrrH) tObe equalto the experimenlaiiymeasuredoxidalionalwearrates ( WaSXT,r ) of the given ~bo-system.

Keyword~: Oxidaiionalwca~,High-chromiumfcrriticsteelpinslAus'~enilicslainlesssteeldisks;Divisionorheat;OxRlaUonalwearr~eg Nmae4"Jr.~lIln~hod;

Tdboiogicaloxidationcondones

I, Introduction 2. The bade dements o f oxidafiomd wear

The complicated and intexdisciplinary nature of tribologi- For conventional static oxidation tests, the mass uptake of
cal research, especially research into oxidational wear, makes oxygen per un it area ( dr, ) usually depends parabo!ieally upon
it a very suitable candidate for the application of computa- thu time (t) of oxidation, that is:
tional methods. In the early work involving compatalional
(dm):=(k~)*(t)=l(Fo)*(R.)*(THe)] 2 (I)
method~ and oxidatioaal wear, line-command programs in
FORTRAN and BASIC computer languages were used for where (kp) is the parabolic oxidation rate constant, (/,~) is
numeric'ally solving the non-linear algebraic eqnationswhich the fraction of oxide that is oxygen, (RH) is the density of
arose from the theoretical heat flow and the theeretical oxi- the oxide and (Txp) is the thickness ofthe oxide. The physics
dational wear analyses. In the more recent work in this field, behind this assumption lies in the fact that, aflcr the first few
electronicspreadsheet programming (working in a Windows moments of exposure to an oxidizingatmosphere, the metal
environment) has provided a more user-friendlyapproach and oxygen ions have to pass through an oxide film that
than either FORTRAN or BASIC. Although normally con- becomes thicker with time. 'rh¢ ~.,ina| oxtdatloaal wear
sidered to be mainly of use for business and accounting appli- theory [!,21 assumes a similar decrease in oxygen intake
cations, there is a growing awareness of the relevance of with increasing time at each of the N asperity contacts
electronic spreadsheets to science and engineering. In this between two sliding surfaces until, after a characteristic num-
paper, we show how electronic sp~eadsheets facilitate the ber of passes, n critical (i.e. a mezhanioally unstable) thicb
application of theories of heat flow und oxidational wear to a ness is roached. Under these conditions, parabolic oxidation
tribe-system of particular relevan,'e to the valve seat wear of kin©tics clearly apply. There have been proposals which
the exhaust valve system of the typical diesel engine, namely assume n linear dependence of (arm) upon dine [3], but that
high-chromium ferritic steel pins sliding against austonitic sort of dependence can only apply provided the whole of the
stainless steel. In order to do this, we must fi,'st review some critical thickness forms during the first few moments of con-
of the basics of oxidational wear. tact. This is considered to be an unlikely mechaniam.
The oxidation rate constant (kp) dependsexponentially on
Based on a p a w presentedat a conferenceon the "Theoreticaland the temperature of oxidation (T,,), namely:
Compu~lonat MethodsInl~rlctlon, Lubrlcellonand Wear" oi,gaatzodby
the T-ibologyGroupofhheImlituteof Physicsin Londonon 11May 1994. ( ~ ) = tap)* exp [ - Lgp/(R*ro) ] (2)

0043.1648196/515.00 ~ 1996ElsevierScienceS.A.AlldglusKeserved
$5D10043-1648.(95 ) 06856-2
[70 Z F.J, Oninn /Wear 199 (1996) 169-180

where (Ap) is the parabolic Arrhenius constant, Qp is the nxidational wear rates for similar systems. Although these
parabolic oxidational activation energy and R is the universal Iribological Ap values were several ordcra of magnitude
gas constanL The original theory, of the oxtdational wear of greater than the static Arrhenius constants for iron [ 12], this
metallic materials assumes that, at any given instant, almost was not entirely unexpected, since the Arrhenius constant
all of the oxidation occurs at the N asperiliea that comprise part of the parahetic rate coustaar (k~) obtained from static
the real area of contact (A,) between opposing trib~ele- oxidation experiments must depend upon several f~etars,
manta. It also assumes that this oxidation occurs at a contact such as the conditions of the specimens, the method of heating
temperature (T~) that is normally higher than the general the specimens and other operational variables, for instance,
surface tem~rature (Tps or Tt~s) of either trlbo-element. The the partial oxygen pressure of the ambient atmosphere.
general theory of oxidalional wear [4], on the other hand, More recently, there have been some doubts cast on the
also includes the oxide growth that occurs at the general assumption that the static activation energies are the same as
surface temperature (Tps or Tns) on those parts of the real the tribologicai activation energies [ 13]. Furthermore, the
area of contact that are momentarily "out-of-contact". static oxidation data are not always available for any given
General acceptance of the oxidational wear theory has been tribe-system. By comparing the results of several pairs of
hampered by the formidable difficulties encountered in experiments in which high-chromium fenitie steel pins were
directly measuring the contacttemperatum (T~) and thenum- slid against auslenilie s!ainless steel disks run at similar
bar of contacts (N). Physical methods of analysis have been speeds and loads, each member of every sdeoted pair having
used to give some information about these parameters. For the same wear rate coefficient (WRc), it was possible [ 13]
example, X-ray.diffraction has been used [5] to identify the to deduce tribological values for Ap and Qp that were consis-
type of oxides present in the wear debris produced during tent with the theoretical wear rates ( assumingthat the original
oxidational wear, thereby providing information regarding wear theory was applicable).
the range of likely values for the contact temperature (T~). Unfortunately, although the Arrhenius constants (Ap)
Teansmission electron microscopy of replicas of worn sur- tended to be relatively constant at around 10 -4 kg 2 m '=s - l,
faces ~5 ] and the direct viewing of such surfaces by scanning the Qp values varied widely and tended to be much too low
electron microscopy [6] have been used to show that inter- compare d with values to be expected from the static oxidation
facial contact ozcurs on only one or two dominant plateaux, data for a ferrous material [ 12]. In an attempt to overcome
the surfaces of which lie about 4 x I 0- ~ re above the rest of the inconsistencies involved in using the original oxidafional
the apparent area of contact. wear theory, the author has recently published a paper [ 14]
Furthermore, Auger electron spectroscopy (AES) has in which account was also taken of the oxidation that occurs
been used [7] to show that these plateaux consist mainly of at the general surface temperature (Tvs) when the real amos
oxides of the substrata material. Although optical micro- of contact =wemomentarily "e- t of contact". The theoretical
graphs and infra-red scanning methots have bee.", u~cd [g] expression for the wear rate was called the "general theory
to cou::t the number of "hot spots" in the sliding interface of oxidational wear", to distinguish it from the expression
between transparent tribe-elements, the response time of based on the original theory..'!'h~ paper showed that the gen-
these techniques tends to be too slow for one to be sure that eral theory of oxidational wear [4] can provide a consistent
the detected "hot spots" are the only points of contact. Also, and realistic set of va[ues of the trihological activation ener-
the practical relevance of a transparent tribe-element, such as gies fern setof experiments involving high-chromiumferr itic
sapphire, is somewhat suspect! steel pins sliding against austeaitic stainless steel disks at
Because of these difficulties, some oxidadonal wear 2 m s- i under various loads.
researchers have looked elsewhere for an alternative (indi- The aim of the present paper is to describe a computational
rect) approach [9-11 ]. By measuring the experimentaldivi- method for a typical tribe-system in which numerical solu-
~inn of beat (DE) at the sliding interface, they were provided tions are perhaps the only way in which we can get the
with another parameter, the theory of which (D-m) also parameters we need to model that system, namely a system
depends upon T~, Nand the thicknesses ( Top and TaD) of the exhibiting oxidational wear, In order to accomplish this aim,
oxide plateaux. Using a numerical method, Sullivan et at. the paper must give a sufficiently detailed description of
[ I 1 ] found the number (N) that will satisfy the equation the oxidational wear model upon which the theories of
(DTn-D~) = 0 for oxidational wear experiments in which oxidationa] wear have been constructed.
low-alloy steel pins were slid against low-alloy steel disks.
T~ was then known, since it depended upon DrH, which itself
depended upon N. Inserting the theoretical values of Tr and 3. The oxidatloRal wear model
N, and the measured value of oxide plateau thickness (Top),
into the original expression for the oxidational wear rates, FLg. 1 is a schematic diagram of a surface of the pin in a
and assuming that the values for the static oxidational acti- pin-on-disk tribe-system that is undergoing nxidational wear.
vation energies (Qp) are also relevant to tribological oxida- At any given time, there are N contact areas, each of radius
tion, Sullivan et al. [ I l ] deduced tribelogical Arrhenius (A), on the surface ofthedominant plateau on the pin. These
constants (Ap) which, in turn, could be used far predicting N contact areas are matched by N areas on a similar dominant
 T.F.J.Quinn/Wear199~1996)169-1~ 171

( 1/g) contacts occur at a plateau on the pin surface before

it becomes unstable and delaminates at a Sickness Tap.
The theories of oxidational wear tend to concentrate on
deducing an expression for the/(-factor in terms of the appro-
priate operational vte'iables and material constants. The orig-
inal theory of oxidational wear [ I,:2] assumes that the oxide
N =onions nt any giveninstant thickness formed at the general surface temperature (T~j) is
coe~enfraled col :'he d0mnt0r,!
ploleOu of conlocr insignifieantcompared with the thickness formed at the con-
tact temperature (TF), whilst the general theory [41 allows
After I f K cOntacls hove occutree over whole ereo of file for the "out-of-contact" oxidation at the general surface tem-
5urfOce Of"Ihe OldfedulfttlJ oxiee filJTIdslon-eir/ofe$
poratme (Tes) over the whole of the apparent area of contact.
Fig, I, The oxld~oinal wear ¢or,la~ model,

plateau on the opposing surface (that is, on the surface of the 4, Analysis ofthe experimental and theoretical division
rotating disk). Dominant plateaux arise from the thermal o[ heat (DE and DTH)
expansion due to the high contact temperatures (To) gener-
ated at the real area of contact. This intermilqent thermal 4. I. The experimentaldivision ofheat (DE)
expansion is the basis of a whole subject of tribolngical
endeavour known as thermoelastic instabilities [ 15]. The The seven experiments chosen for discussion in thispapcr
plateau on the pin lies a distance (Tat,) above the general have been reported in a previous publication [ 17]. The pin-
surface of the pin, whereas the plateau on the disk lies a on-disk wear machine upon which the selected experiments
thstance (T.D) above the general surhce Icwl of the disk. were carried out consisted of a cylindrical pin surrouededby
Over a sufficiently large number of passes, it is reasonable to a hollow cylindrical insulator split in two halves along its
assume that the whole, surface a~a of the dominant plateau main axis. The insulator and pin were clamped, together with
on the pin surface oxidizes uniformly until it reaches a critical the three thermocouples, in u cylindrical copper calorimeter
thickness (Tn~). at which it delnminates and breaks up to (as shown in Fig. 2). This figure also shows the. linear rate
become oxidized wear debris. of heat flow (H1) entering the pin at the sIiding interface, the
The average Lime for a given plateau to reach the critical linear rate of heat flow (//2) leaving the exposed portion of
thickness (THp) is related to the "K-factor" of the Archard the pin and the linear heat flow (//3) entering ~ insulated
wear law [16], which relates the theoretical wear rate portion of the pin. The objective of this sub-section is to
(WRTH) tO thO real area of contact (A~), namoly: provide the basic equations Y~rdetermining the experimental
division of heat (DE) at the pin/disk interface, where De is
WRT. = g * ( A r ) =K*(W/Pp) (3) given by
where g¢ is the norraa] applied load and Pt, is the hardness of OF= (H1)/(HT) (4)
the substxate material of the pin. Because the wear ralz is
where
measured in units of volume removed per unit sliding dis-
tanea, it has the dimensions of area. Hence, K is a dimension- H~ = U*(FF) (5)
less factor of proportionality. Archard [ 16] inte~vreted Kas
being the probability of producing a wear particle at each In these cqumions (HT) is the total frictional hem dissipated
at the interface across which ~ shear frictional force (Yt,) is
microconlacL Following Arehard0 we may assume that
acting and U is the linear sliding speed.
The therraal analysis of this system is divided into calcu-
- ROTATING DISK lating the heat flow in the insulated portion of the pin (of

IL, length L3) and heat flow into the exposed portion of the pin
(of length L1). Rowson and Quinn [ 18] deduced the follow-
ing differential equation for the insulated portion of the pin:
(d2T/d~) = (M;)*T (6)
where
(M2) =2*(Kt)/[ (Ksr,)*(RT)2. In (R~,IR.r) ] (7)
The~ authors assurned that the axial h eat reaching an element
( ~ ) at a distance x from where the insulation begins, is
Fig.2. Healflowsand~ennocouplepodlions~levantIotheanM,sisaflhe diminished only by the radial heat losses from this clement
ex~filae~ilaldivisionof heat (De}. (see Fig, 3 for a perspective view of the insulated portion of
L72 Z E£ Quinn/Wear 199(1996)169--180

The boundary conditions are T(x) = Tps at x = 0 and TAat

x = Lv One also has the axial heat flows of H i atx = 0 and It2

~=o~ 1~ T:TA
at x = L v All the terms in Eqs. (14) and (15) have been
defined except H. This is the heat transfer coefficient for a
steel cylindrical surface losing heat by convection in an air
flow moving at approximately the same speed as the rotating
',+dx ~ T=Tx"Pdx disk surface. It is assumed that the temperature of the air is
not very different from To, the temperature of the copper
calorimeter which holds the pin. Rowson and Quilm [18]
solve the above differential equation and obtain the following
~L-~ T=TB expression for HI:
Ht = (gse)*f'ff)*(Rr)*(TA- To) * sinh (GD/Z

Rg, 3, Perslmctiv*'dew ofth~ insulatedportionof the pin,

+(H2)*eosh(G~) (l~)
where
th~ pin). (Kt) and (Ksp) are the respective thermal conduc-
tivitias of the insulator and pin materials. (G3) = (L.)/[Z*(RT) ] (1"7)
Solving this equation for the temperature T(x) at any dis- Clearly, the ¢~perimental division of heat (D~) is given
tance x between x ~-0 and x =/~. given that T(x) = Tt, at x = 0 by combining Eqs. (4), ('~), (15) and (16). Another peram-
and T(x) = Tn at x=L~, Rowson and Qainn [ 18] obtained vter can also b~ deduced from this analysis, namely the gen-
the following expression for T(x): eral surface temperature of the pin (Tps). This is the
r(x) = { [ (TA- T o ) - ( a , ) ] * exp (M*x) l temperature that is consistent with assuming that all of the
heat (H0 catering the pin at the pin/disk interface is dissi-
+ {(Gt)* exp ( - M ' x ) } + {To} (8) pated across the whole of die nominal area ofcontactbctwcen
where the two tribe-elements. The expression for (Tps) is given
below:
(Gt) = {(TA- To)* exp [M*(/~) ]
Tps = {(T A - To)* cosh (G~) } + [Z'H2* sinh (G3) /
- ( T s - T c ) } I { 2 * sinh [M*(L3)] ) (9)
[Ksp*(~)*(Rx) ] }+ Tc ( I g)
Since If(z) = - (Kse)*('ff)*(RT)2*[dZ(x)ldx], where
(,rr) = 3.141593, the authors were able to deduce H3 at x = 0,
namely: 4.2. Thetheoreticaldivi&ionofheat(DT,~)
Hs = (Ksp)*('ff)*(RT)Z*M*(G2) (10) Arehard [ 19] proposed that the theoretical division of heat
where (DTu) should be given by the following relation:
(G2) = { ( r ^ - re)* ~osh [M*(L3) ] ~ . = rD/(TD+Tp) (19)
-(r.-Tc))l{sinh [M*(L3)] } (ll) where (Tp) and (To) are the ~,ctitious temperatures one
would obtain assuming all the friulional heat (fiT) goes into
If ,~ is the heat flow rate per unit temperature gradient along
the pin anddisk, respectively. Arehard [ 19] didnot, however,
the thermoeonple wire measuring TA,then (H2 - fi~) is given
allow for the effect of any oxide films present in the real are as
by: ofeantaet between the # n and the disk. Quinn [ 10] modified
(1"12-1"13) = C * ( T ^ - T c ) / ( R A - R T ) (12) the expression for (Tp) to take into account an oxide film of
thiekness (THp). The modified equation is given below:
Hence
Tp= [ (HT)/[(Tr)*(A)*(Ksp)*N]}
Hz=H~+{C*(TA-Tc)I(Ra-RT)} (13)
+ [ (fiT)*(Tlip) / [ ('ff)*(A2)*(Kop)*N] } (20)
Having obtained an expression for Hz, the heat flow rate
leaving the exposed portion of the pin, P.owsvn and Quinn Sullivan et at. [12] later modified the Archard [19]
[ 18] could then deduce H1, the rate at which heat flows from expression for (TD) to allow t'or a thiekaass (TH~) of oxide
the real area of contact into the pin, by solving the following film on the real areas of instantaneous contact on the disk.
differential equation f~r the exposed portion The modified equation is:
(dZTId.F") = T/[ (Zg) *(RT) 2] (14) (To) = { - (0.102t'C) +0.8005}*(K~ +K2} (21)
where where
Z~= (gsp) / [2*~*(R,)] (15) KI=(HT)I{(~)*(A)*(N)*(KsD)} (22)
 ZF.Z Q~inn/Wear199(1996)169.-I80 173

and is particularly useful for setting out all the data a~Jd all fl',e
equations given above and almost instantly seeing the effect
K~=(H.r)*(T.D)/{(~)*(A2)*(Kov)*(N)] (23) the choice of N has on all the many parts of F.xI. (28), namely
and Eqs. (19)-(26), For details of the eleclronic spreadsheet
program used to deduce the number of contacts (H), and
c= u*a/ [2*(xg) } (24)
b e t a tlg contact temperatore (To), interested readea'scan
where ~ntact the author at Ihe address given above.
(XK) = (XsD)/{ (Rov)*(e) } (25)
In these equations, (Xa) is the diffusivity, (RAy) is the den-
sity and (c) is the specific heat ~apa~ityof the disk material.
The radius (A) of each individual contact is found from: 6. Deterrdning the tribolegienl olddatien constants
using the original oxldallonal wear theory
A = ~/1W/[,V*(-~)*(Pt,) 1} (26)
where (Pc) is the hardness of the pin material.
Before going on to determine (N) from the heat flow, we The original oxidational wear theory is based on an inter-
should briefly mention the relationship between the theoret- pretation oftbe K-factor in the Archard wear law, Eq, (3).
ical division of heat (DT~I) and the contact temperature (Y~). namely, iLis assumed that a total number (1 IK) of passes are
In fact, Arehard [ 19] has shown that the excess temperature needed for the whole of the dominant oxidized plateau to
(T~- Tt,s) can be obtained from the relation: build up to a critical thickness T.t, on the pin (or Tun on the
disk) at which it I~,_comesunstable and breaks away from its
(T~- Tps) =DTtt*Tp
substrata as a delaminalcd flake. In the usual aibo-system.
Hence the flake will be trapped between the sliding surfaces and
broken up into wear debris particles which ate then thrown
Tv= (DTA*Tt,)+Tt, s (27)
out of the contact area, The total time (r) for (It'K) passes
This equation has been the basis of many papers dealing with is also the time during which the plateau is being subjected
the temperatures of robbing surfaces, and will be used in the to the oxidizing temperature (TeL Thus:
current paper,
r= (ILK) {(2*A) I U} (29)

$. Determining the number of contacts from the heat where (2*A) is the diametex of an asperity contact. From
flow Eq~ (!)-(3) an~a (59), we get the original expression for
the oxidational wear rate (We-m), namely"
Let us define (Fn), a function of N, by the following
equation: WR~= {2*.4.*W*Ap*exp -Qpt[R*(TF+273)])}/
(F~) = (Din) - (De) (28) {p p*U* ( Fo*Ru *T.p ) 2} (30)
where DE, the experimental division of heat, is given by Eq.
(4) and D.ra, the theoretical division of heat, is given by Eqs, This equation depends upon the number of contacts (N)
(19)-(26), All the parameters of these equations are known through (A) and (To), measu~d in degrees Celsius, as given
except 17 (the number of contacts) and THt,and T.v (the in Eqs. (26) and (2?), The remainder of the pammeten are
thicknesses of the oxide plateaux), Now, these thicknesses known, c~ccpt fortbe appropriate values to use for the An'ba-
can be measured using the scanning electron tuieroscope ntus constant (At,) and the oxidation activation energy (Qt,).
[17], leaving 17 as the only unknown. It is analytically As mentioned in the Introduction, there are some dmthts
impossible to solve Ecl. (28) for N, so recourse must be about using the static oxidation values for these constants,
made to a numeric[ tuethod, The choice of method is up to so we will use the experiments them~lves to provide the
the individual. However, the method of "halving the inter- tribologicalA# and Q#,
val" is recommended because there are no problems of non- The author [13,14] has used the. following method for
convergence to the s:flulion, namely (F~)=0, with this deducing At, and Qt, from Eq. (30) applied to experiments
meff,od, All one has to do is to find two values of N, namely with high.chromium ferritic steels sliding on austenitic stain.
N~ and N2, which give Fm and Fm above and below zero to less steels:
the required degre~ of accuracy. I, Replace WnTUwith the expedmantal values for the wear
Tbe choice of the initial values for 17. namely N~ and N:, rates for two experiments eat'fled out at the same speed
is probably the most important step in the analysis. It is and only stightly different loads and which have the same
essential to ~e what effects one's choice for Nhas on F~ and wear rate coefficient (Wnc), i.e. the same specific wear
the component parts of Eq, (28), The electronic spreadsheet ~eate.
174 ZF3. Qu&n/WearI99 (1996)169-180

2. Apply Eq. (30) tu (say) Expts [ and 3, and obtain: Table 1(c) is concerned with the determination of the acti-
Qp=G*[In (S/V)] (31) vation energies for tfibo[ogica] oxidation (Qe) and the
accompanying A~henius constants (An) for this particular
wbero tribe-system. Rows 92-94 contain the spreacL~h©etsoftware
G= ( Tp,*T~z*R) / ( L , - T~) (32) varaions of G, S and V for each adjacent pair of experiments,
as given by suitably modifying Eqs. (32)-(34). Rows 95
SffiWRo*A3*U~*(Tap3 2 (33) and 96 contain ~nditional statements related to whether or
and not the loads used in any adjacent pair of experiments arc less
or greater than WT. If hath loads are less, then row 95 gives
V~ WRc3*A,*U~*(T,p~) 2 (34) the tribological activation energy for mild oxidational wear
(QPM) for adjacent pairs of experiments, according to Eq.
3, Having deduced Qe from Eq. (31), obtain At, from the (31 ). Row 96 will then show the type of oxidational wear,
following equation: namely "mild". If both loads are greater than WT, row 95
gives the type of oxidational wear, namely "severe". Row
Ap=M/Q (35) 96 will then give the tribologi,'al activation energy for severe
where oxidatioaal wear (QPs) for adjacent pairs of experiments,
also according to Eq. (31). If the loads are below and above
m-m WRcl*Ul*PPl*(Tnpl*Fol*Rm) 2 (36) Wr, then there is no solution, and "no soln" appears in both
and of rows 95 and 96. Similar conditional statements are con-
tainco in rows 99-102, which relate to the Arrhenius con-
• Qf2*A** exp { -Qp/[R*(T~:1 +273)]] (37)
stants for tribological oxidation according to Eqs. (35)-
(3'7,). Obviously, ArMI relates to the constant for mild oxi-
4. Repeal the above steps for Expts 3 and 5, Expts 5 and 7, dation al wear of the first experiment of hh.epair, whereas Apm
Expts 7 and 9, Expts 9 and 1 I, and Expts I 1 and 13, to relates to the second experiment. Similar definitions obtain
obtain Qp and Ap for each pair, for Aps~ and At,s2 for severe oxidatiooal wear.
Table 1(a) is a part of the oleclronic spreadsheet created From Tabte l(e) we can deduce that the QrM values for
to enable the tribological oxidation constants to be calculated Expts (1) and (3), (3) and (5) and (5) and (7) give an
from the seven wear experiments using high-chromium far- average value for QvMof about 0.226 kJ me1- ] and an aver-
fitic s~el pins sliding against austenitie stainless steals first age value for Qes of 5.795 kJ mol- ~. InseTting these values
reported by Allen et at. [ 17] in 1986. Rows 1-6 of the spread- into Eq. (35), we obtain an average Apst of about l a x
sbee t are reserved for title and/or constants that relate to every 10-4kgZm-'ts -~ and an A~ of about 4.2x10-4kg 2
experiment. Rows 7-14 contain stone of the experimental m-4 s- t. All of these values are well I~low the static oxi-
input data for each experiment, Row 15 contains the software dation of iron values obtainable from the results reported by
formula version of Eq. (4) for the total heat (HT). ROWS17- Caplan and Cohen [ 12]:
38 contain a mixture of (i) more input data (such as the wear • For T< 450 "C--Qt, = 208 kJ mot- ~; Ar = 10~ kg~ m-'*
rate coefficients) and (it) the values of quantities that have $-i;
been transported from other parts of the current spreadsheet • For 450 < T< 600 °C--Qp ~ 96 kJ real- ~; Aa = l0 ~ kgz
or from uther sprcadsbeets. Row 25, which is described as m-4 s-l;
"wear rate transition (WT)", gives the load at which the • For T> 800 ~'C--Qe =-7-10 IO real- ~; A~ = 10~ kg2 m -4
graph of "wear rate vs load" changes slope by about a factor $-i.
of two. Following Lira and Ashby [20], the type of mild These differences between the values for static oxidation of
wear at loads less than WT wilt be called "mild oxidational iron and the tdbological oxidation values for this systam we.re
wear". At loads greater than W.r,the mild wear will he called originally thought to be due to the high proportions of chro-
"severe oxidational wear". It must heemphasized that severe mium in the pin and disk materials [13]. It was suggested,
oxidational wear is n sub-classification of mild wear. It is not however, that bettor agreement [mtwecn the two setsofvalues
scver~ wear, as described by At'chard ~ d Hirst [ 21 ]. Severe might be obtained if one assumes that the "out-of-contact"
wear is characterizco by metallic wear debris, low electrical oxidation [4], which occurs at the general surface tempera-
contact r~sistanee and wear rates an order of magnitude ture (Tps) oftbe pin during oxidational wear, makes a sig-
greater than mild wear. nificant contribution to the total thickness of the oxid~
Table l(b) contains the sub-routines required for calcu- plateaux. In ~,&er words, it was suggested that the general
lating (i) the heat transfer coefficient (H), (it) the experi- theory of oxidational wear might apply to so-calla0 "room
mental division of heat (DE) and the general surface temperature" experiments, as well as to expedtoeats carried
temperature of the pin (Tps) and (iii) the effect of(Tvs) and out at externally induced elevated tamp, futures. The conse-
(Tvs) on the hardness of the substrates and the thermal Con- quences of applying the general theory of oxidational wear
ductivity of the oxides on the pin and disk, respectively. to this tribe-system are described in the next section,
 T,FJ, Quinn/ WearI99 (1996)769-180 175

Ta61c I

I
2
3 Xg= 8.8 >~10- ~, Ksr = 26.8, Ks,~= 15
4 R 8.314,K~-0.1045,R^=0~0795
5 RTffiO.OO3175,C=O.OO~OIIT,M=29,031721
6
7 F.xper~me~tno,

8 i 3 5 `1 9 11 13

(a) Upper part of eleclronic spreadsheet for dexlucmgthe tnbological oxid~lioaal C~ns~'~csfor the given n-il~.syslem
9 Speed{/..') 2 2 2 2 2 2 2
I0 Load (W) 12.5 31.25 3"1.5 50 75 81.25 87.5
11 Frictionforce {F~) 5,12 13.12 15.38 22 28.12 30.87 32.81
12 ThermocoupleA (T~.) 48 92 75 115 220 22"1
t3 Thermoc~pIe B (T~) 34.5 50 44 69 88 97 7~
14 ThermocoupleC (To) 32 39 39 51 "19 77 89
15 Tolalhe~ (HT) 10.24 26.24 3026 44 56.2,:1- 61.74 65.62
16
I'~ Temperalumof pi~.surface (Tn) 56,250543 114.79153 97.885649 145.30555 301.77541 308.89105 338.1686
[8 Temperatureo f disc surface 40 39 32 47 6`1 98 98
(TD~)
19 Wearcoefficient(W~c) ]JOE-]4 HOE-t4 1.10E-14 ]JOE-14 2.80E-14 2.8013-14 2.8E-t4
20 Healflow along p~n (H~) 0.9849~M 3,0406638 2,3078119 3.6057744 9.5827176 9.6370782 11A1903
21 Oxide lhickness,pin [Tt~) 3.25E-06 3.25E0-06 3,25E-06 3.25E-06 3,251~-06 3,25ff.-06 33E-06
22 Oxidelhlckne~s,disc(TuD) 3.50E-06 3.50E-06 3.50E-06 3.50E-05 350~-eJ6 3.5DE-06 3.5E-06
23 Pinhvzdness(Pv) 3.50E+(~ 3.50E+09 3,50E+C~ 3.50E+09 3.50E+09 3.50E+0P 3.5E+09
24 Then'nMconductivityofoxide 6.2070689 5.8189422 5.9310281 5.6166342 4.579239 4.5320623 4.337952
(K.~)
25 Wear r~le transidom ( WT) 65 65 65 65 65 65 65
26 Dischardness(Po) 3,92E+09 3,92E+C0 3.93E+09 3.9UE÷09 3.8EE-I.09 3.79E+09 3~8R+0~
27 Thermal cold of oxide (Ken) 6.31481 6.32t~ 6.367~ 6,2684 6.1358 5.93027 5.93027
28 Distmnce(L~) ~tween 0.0165 0,0i65 0.0171 0.016 0.0t6 0,016 0.016
29 fl~ennocouplcsA mid B
30 F~n~ardne.ss(Pe) 3.50E+09 3.50E+09 3.50E+09 3.50E+09 3.50E+09 3.50E+09 3.5E+09
31 Distance(L~) belween A 0.0079 0.00"/ 0.~% (},008 0,008 0.008 0.008
32 thctmucoupleand interface
33 Vi~tualpintcmpemtu~e(Tp) 2071.0638 2208,463~ 231C'AI71 2564.4878 2422.`1609 2494.5328 2518.083
34 Vi~aldisctemperamm(Tn) 220.39724 289.45634 ]87.40239 228.9]797 497.59935 461,3<1461 530.5083
35 Divis[ortofheal(theory}(Dr~) 0.096182 0.1158"/9 0.0"150~o4 0.0819494 0.1"/0389"1 0.1560913 0,174018
36 Contact~emperatuteITl~) 255.4495~ 370.70597 27t.227'91 355.46383 714.58895 698.26597 776.3592
37 Radiusof COnlacl (A) 6.78E-~ 6.58~-- 06 6.88E- O~ 6.81E-06 5.72~- 0'~ 5"9~ I [ 5'~ I ~
38 Numbe~ofcontacls(N) 24.720748 65,566458 72,005781 98.142~26 20L~4619 210.95311 250,8981

(b) Middle l~'U'tof cb~tmnic spzeadsheetshowing the sub-mutin~ n:quin:d for cMculaOngH. De, T~, PD,Pp,A'oPandKoo
52 CaIcuialion of heal tram let ~ecfficiem
53 Reynolds number (RB) 778.09624 `17809624 `178.0%24 778.096~4 778,09624 778.09524 778.0962
54 Nusseldt's number (R~) 13.680332 13.680332 13.680332 13.680332 13.680332 13.680332 13.6'~033
55 He~t~r~sfar coeffici~t (H) 58.538007 58,538003 58~538007 58.538007 58,538~7 58,538007 '~8'~3801
58 C ~ J c ~ Y ~
62 Z=~(lfst,/2*Rr"H) Z= 8.491,0534 8.4910534 fiA9L0.T,¢34 8.4910534 g.,¢910534 8A91U534 8.491053
63 F=L/[Z*Rr) F= 0.29303~ 0.2595526 0.356095 0.2967459 0.2967459 0,29,57459 0.296746
B= 133 42 31 46 132 130 163
D= 2.5 II 5 ]g 9 20 -]2
66 E f ~'~- Tc B= 16 53 36 64 141 150 I$1
67 Cmf¢osh(M*L~) Cm= 1.1169424 1.1169424 1.1257793 1,1098374 1.1098374 1.1098374 IA~'837
68 ,.,t'a~=finh (M * L~} Sm= 0..~97,¢5~5 0.4~'7.%,~5 ~.51`1087 0,4813929 0.4813929 0.4813929 0.481393
69 ~., H~= ~.','617,.173 238(;5~'~ ',"~q'Z9835 2,7143355 "1.5491695 7.49739"15 9.192134
70 /'/~ H~= 0.~55-~ 23i72',.kE L,i~"b¢4 2.8"/16884 7.89583"16 "1.8661933 9.563389
7t Cmfccsb(F} CH:= ;'~ 32433 I {.~33,¢~;.- i~:~i0"146 1,0443531 1.0443531 1.044353]. 1.044353
SHz= 0,~.~;"24S5 U.2";~:J801 (~:~;36686 0,~11202 0.3011202 0.3011~ 0,~0112
73 Hi Hz = 0.984903~ 3~L~r~38 2,3(]78119 3,6057"144 9.5827176 9.5370`182 11.41903
176 ZF.J.Qai~#/ Weart99 (t996)169-180

Table I (continued)

74 DefH,/Hr De= 0.096182 0.115879 0.0750264 0.0gl9494 0.1703897 0.1560913 0.17401S

75 Tt,s Tea= ~6,250543 114.79153 97.885649 145,30555 301,77541 305,S9105 338.1686
78 Effectof Tvsand T~so~ ha~xtncssesand lhermalcmductivityof oxidefilms
82 Hardnessofdisk Pa= 3.92[/.+09 3.92E+~9 3.93F.+09 3,90E+09 3.g6E+09 3.79Eq-09 3.$E+09
83 Hardnes$ofpia P~= 3.50E+09 3.50E+09 3.50E+09 3.SOE+09 3.50E+09 3.50E+09 3.5E+09
84 Thermalconductivi~ Kor= 6,20706S9 5.~t189422 59310281 5.6166M2 4.579239 4.5320623 4.337952
85 of pin & diskoxides Key= 631481 6,32144 636785 6.2684 6,1358 5,93027 5,93027

(c) Lowerpart of electtcoiespreadsheetfor dedooingthe trihelegicalexLdaflonalcoastaotsforthegiventribe-system

¢}0 DeternlinationoftM triUologicaloxidationconstants(assumingthe originalmildoxidalionalwejrtheory)
92 G=(T~*Tvz*R)/{Tv,-Tv:~}= 6830.9259 8403,2384 9515,7386 5880.5228 2S4148,51 57713.867
93 S=Wsct*A2*Ut*T~t*T.pt= 1.53E-30 1.60E-30 1.58E-30 1.33E-30 3.5E-30 3.33E-30
94 VfWec2*At*Vz*Yvn*Tan= 1.55E-30 1.53E-30 I.~OE-30 4.03E-30 3.38E-30 .~.50E-30
93 Typeo['wear/activation 201.95853 372.429 104.64127 No solution Severe Severe
energy=
96 Typeofwear/acfivation Mild Mild Mild Nosoht~on 8707,1229 2864,2138
energyffi
9"/ Demityofoxides(R~)= 5220 5220 5220 5220 5220 5220
98 FractioRoxygenin o~ides= 0.2945 0,2945 0,2~45 0,2945 0.29¢5 0.2945
99 Anhet~iusconstant(Ansi)= 0,0001484 0,0001565 0.0001429 0,0001412 Severe Severe
100 Arthenius~:oastant(A~,t~)= 0.0001519 0.0001429 0.0~1412 S~vet~ Scvcrc Sevele
10l Atrheniusconstant(A~) = Mild Mild Mild Mild 0,000427"/ 0.0004133
102 Arrheniusconstant(Apse)= Mild Mild Mild 0.0004277 0.0004133 0.0004344
103 Expt l&3 Expt3&5 Expt5&7 Expt7&9 Exptg&ll Exp~11&13

7. Determining the tdbological oxidation constants D = (~)*W*(2*RT) z (43)

using the general theory of osidafienal wear
E=Ap2* exp I - O.p/[R*(Tps+273)] } (44)
= ~*J=e*o*,V*(W/(N*C~)*Pr) Ie s (45)
In this section, we will describe the calculations recently
carried out by the author [ 14] in which he applied thegeneral K = (Fo~*Rm*Txp) 2 (46)
theory of oxidational wear to the results of"room tempera- Wn.m~ = {W*2*R.r/JF*(M*L) °'s (47)
ture" oxidafional wear experiments with the high-chromium
ferritic steel/austeniti¢stainless steel tribe-system running at where
a speed of 2 in s- x. He created an nicotinic spr~adsMet with
rows 1-86 exactly the same as Table l(a) and (b). He J= U*Pe*Fo~*RH~*Fo~*Ru~*(Tl.lp) 2 (48)
replaced rows 87-103 of Table l(c), however, with rows L = (exp -QdtR*(Tes+273)])*(Ap,/N) (49)
87-146 now changed so that he could apply the general theory
to each set of results (i.e. each column) in order to deduce M= (Tr)*ap,* exp { - Qd[R*(Tv+273) ] } (50)
the appropriate tribological oxidation constants. Before dis-
Table 2 containsthe lower part (rows 87-146) of aspread-
cussing this spreadsheet, the expression for the wear rate
sheetused to deduce the valuesof the tribolog]ealoxidational
(Wt~.nt), as given by the general oxidatioaal wear theory, will
activation cncrg[cs (QpM and OPs) for mild and scvvrc oxi-
be quoted without derivation:
dational wear that will satisfy Eq. (38). In order to do this,
( 38 ) one has to know the appropriate values for (i) the Arrhaniu~
WR.m= WRmj + WnTm + WR.m3
constant (Apt) for tribological oxidation at the contact tem-
where perature (To) and (ii) the Arrhenius COnstant tap2 ) for tri-
bological oxidation atthe general sm'face temperature (Trs).
WR'mm= (a*B)/C (39) The author [ 14] showed that the Apt values obtained from
applying the original theory to this trilm.system were also
a=2*W*A
(4Oa) relevant to any application of the general theory. Hence, the
B=Apt* exp { -QpI[ R*(TF + 273) ] } (40b) addresses of the cells in Row 99 of Table 1(c) for ArM, and
in Row 100 of Table l(c) for Aps, were inserted into the
C = V*P p*( F~* Ra~*r.p ) 2 (4l) appropriate cells of Row 06 of Table 2. The assignment of
Wn~I2= D* E/ ( G*K) (42) values to ap2, however, was not so straightforward, Sine©the
Arrhenius constant should be independent of oxidation tem-
where perature, pmvidad the composition of the oxide(s) does not
 T.F.LQuimt/Weor 199 (1996)169-I~0 177

Table2
Using an eleotronic spreadst,,e¢.tto deduce the values of Q¢ to smis[y W ~ - Waexvr= 0 to wilhin0.001%

Exl~dmenl no.

I 3 5 7 9 II ]3

(a) First term in the ge~r~ theory of oxid~ional wear

89 OLD Qe,= 24675 26208 2/1490 30160 34880 35028.5 36132
90 OLD Q n = 24675 26208 24490 30160 34880 3S028.5 36132
93 OI.D Qp~- 24675 26208 24490 30160 M880 35028.5 36132
92 Wgexpr;Wnc*W= 1.388~13 3.44E-13 4,13E-13 5.50E-13 2.10E-12 2,28E-12 2.5E-12
93 W~a=(a*B~}/C= 5.24E~16 2.75E-15 1,88E-15 1.71E-15 3.00E-14 297E-14 3.9E-14
94 a=W*2*,l = 0.0001695 0.0004115 0.0(~5162 0.0006807 0.~8579 0.0009617 0.000986
95 B~=A~*exp[-Qv~l 5AOE-07 1,17E-06 6,37B--07 4.40E-07 6JIE-06 ~.40~-06 6.9E-06
[R* (T~+273))) =
96 Apt mApM/S(original theory) ~ 0,0001484 0.0001565 0.0001429 0.0001412 0.0004277 0.0004133 0.0004~I
97 C=U*Pp*(f~*RH~*T.p) ~ 174733,58 174733.58 174733.58 174733,58 174733,58 174733.58 174733.6
98 f,=fJ'a~tioaofoxygeninoxides{a'~Tt,) 0.2945 0.2945 0.2945 0.2945 0,294-5 0.2945 0,2945
99 R~,,=denf~tyofoxi~s(atT~)= 5220 5220 5229 5220 5220 5220 5220
I00 Ws-relts=(d*B2)/C= 5.24E~16 2.75E-15 1.88E-15 LTIE-15 3.00E-14 2.97E~14 3.9B-14
101 Bz=Ap~*exPI-Q~/ 5.40E-07 1.17E-06 6.37E-07 4.40E-07 6.11E-06 5AOE~06 6.9B-06
[R* (Ts+273)1} =
1{72 Wa.n4t~(a*B~)/C= 5.24E-16 2.75-15 1.88E-15 1,7LE-15 3.00E-14 2.97E-ld,. 3.9E-14
103 8~=Ap~*exp(-Qeal 5.40E-07 I,r7E-o~ 6.37E.-07 4.40E-07 6.11E-06 5,40E-06 6.9E-06
~R* (T~+273)]] =

104 (b) Secondterm in tt,~ g-,mczalthoryof oxidational wear

105 A ; fraction ofosygcn in oxides 0.2945 0.2945 0.2945 0.2945 0.2945 0.2945 0.2945
106 (at 7'~)
107 R~,=densilyofoxid:'s(~Te~)= 5220 5220 5220 5220 5220 5220 5220
108 Are=averagcA~/S,ori~nallh¢o~ 0+0001473 0,0001473 0.0001,673 0.0004222 0,0004222 0.0004222 0.0004222
109 W:~2~=D*~/(G','K) 1.2[E-13 2~g5B-13 3,59E~13 4,90E-13 1.63E-12 t.791]-I2 1,9E-12
110 D='n*W*(2*RT} ~= 0.00i5835 0.0039587 0.0047504 0,{)063338 0,0095008 0.0102925 0,011084
Ill E~=Ap~eeXp[-QPt/ t.79E-08 4.34]/-08 5.24E-08 7.231~-0,8 2.85E-07 3.03E~07 3.416-07
JR* (Tn+273)1}=
112 G=4*P~*U*N*2*~f(WI(lv'*'~*P~)) 9387822.6 24173801 27750984. 37410448 66789906 69918808 "/9128292
113 f=(F~'~m*TnP) 2= 0.000025 0.0~25 0,000025 0.000025 0.000025 0.000025 0.00~
114 Wsm==(D*U~2)/(G*K)" 1.21F,,-13 2.85E-13 3,59E-13 4.90E-13 1.621::.-12 [.79E-12 1,9E-12
115 E~=Ap:*exp(-QP21 1.79E-08 4.34E-08 5.24E-08 7.23E-08 2.85E-07 3.03E-0"/ 3.4E-07
[R* (T~s+273)] }
116 W~=(D*E3)I(G*K)= 1,21E-13 2.85E-13 3,59E-13 4,90.B-13 1,63E-12 t.79E-12 1,9E-12
117 E~=Ap:.exPl-OP,J 1.791:-08 ,6.34E-0S 5.24E-08 7,23E-0g 2,85E-07 3.03E-07 3.4E-07
JR* (T~s+273)]] =
118

119 (c) Third lerm in the ~eneral theory of oxide.tin,u1wear

]26
121 W e ~ - { W * 2 * R T / ] ] * ( M I * L , ) °'~= 1.59E-14 5.60E-14 5.20E-14 5.80E-14 4A2E-13 4.61E-13 5.5B-13
122 ./= (U*Pv*f~*Re~*.f**Rm*~v) = 174733.58 174733.58 174"/33.58 ]74733.58 174733.S8 174733.58 174733.6
123 L~[exp-Qm[ %25E-10 6.63E-10 7,27E-10 7.37B-10 1.37E-09 IA3E-09 1.4E-09
[R*(T~+ 273)I}*ApzlN
L24 M~=T*Avl*exp{-Qe./ I.?OE-06 3.67E-06 2.00E-06 i.38E-06 0.00~t92 0.000017 0.000022
[R* (FF+273) $}"
i2:5 Wg.rH~2-~W*2*R.r/J}*(M2..I,a':°5~ 1.59E-14 5.60E-14 5.20E-14 S.80E-14 4.428-13 4.61E~13 5,,.~--13
126 /.:=lexp-Qn/[R*(Tps+2'73)|l* 7,25E-10 6.63E-10 7.27B-10 7.37E- 10 1.37E-~69 1.43X10-09 1,4J~-'09
A~IN
127 M.~='n*.4r,~*expI-O~/ 1.70E-06 3.67E-~6 2.0013-06 1,38E-06 0.0000192 0,000017 0.0,{~022
(R* (7"i.+273)])
[28 Wa~'{W*2*E.r/J}*(M3*L.~) a'~= 1,59~-14 5,fi0E-|4 5.20E-14 5.80E-14 4,42E-13 4.6lE-13 5,5E-13
129 Iq=(exp-Qp~l[R*(Tps+2?3)]j* 7.25E-10 6,63E-10 7.27B-10 ?.37E-10 1.37E-'09 IA3B-09 IAE-09
ApzlN
130 M3=~*A#~*~p{-Qp~I 1.70E-06 3.67E-U6 2,99E-06 1,381]-06 0.,{X~0192 0.000017 0:~0022
tR*(7"F+273)I}
178 T.F.£ Q~inn/ Wear199 (I 996) 169-180

Table2tconlinaed)

]31 f~/Qp)-W~xvr-W~where -7.33E-17 -I.72E-17 -4,25E-16 -1.67E-16 1.26E-15 -4,23E-16 -7.2B-14

]32 WRIffiWR~n,+Wx~m+WBT.~ 1.38E- ]3 3.4~]~--13 4.13F,- 13 5.50E- 13 2.10E- 12 2.28E- 12 2.5E- 12
133 J~(Qp)ffiW~xer- W~'~here - 7.33E- 17 - 1.72E- 17 -4.25E- 16 -1.6")'E-16 1.26E-15 -4.23H-16 -7~E-14
134 W~z=Wlcm~+Wl~xll~+W~l~,_ 1.33E- 13 3.44E- 13 4.13E- 13 5.SflE- 13 2.10E- 12 2.28E- 12 2.5E- 12
135 f~(Qe}=W~ma,r-Wn~l~re -7.33E-17 -I.72E-17 -4.25E-17 -1.67E-16 126E-15 -4.23E-16 -7.2E-14
136 W~=W~uu+Wm-w.~+Wxva~ 1.38E- 13 3,44E- 13 4.13E- 13 5.50E- 13 2.IOE- 12 2.2SE- 12 2.5E- 12
137 NewQe~= 24675 26208 24490 30160 34880 35028.5 36132
138 NewQ~zffi 24675 26208 24490 30160 34880 35028.5 36132
139
I~0 Old Qm= 24675 26208 24490 30160 34~0 35028.5 36132
141 Old 0~= 24675 26208 24490 30160 34880 35028.5 36132
14-2 Old Q~.~ 24575 26208 24490 30160 34880 35028,5 36132
143
144 Waexvr= 1,35E- 13 3,44E- 13 4.13E- 13 5~50E-13 2.tOE- 12 2.28E- I2 2,5E- 12
145 Mildor seve~ oxidationalwear Mild Mild Mild Mixed Severe Severe Severe

change, the author assumed [ 14] that the average ArM and 8. Conclusions
Aps values [from Table ! (c) ] would be reasonable values tu
take for Av2 in Row 108 of Table 2. This paper started from an oxidational wear model involv-
Rows 88-103 of Table 2 contain the calculations in which ing the delamination of the dominant oxidized plateau of
all the above equations are used to provide the three values contact when it reaches a critical thickness after ( ! / / O
of WaTm (namely W a r . , , W~Hn and WaTHI3)obtained for passes. It is assumed that the dominant plateaux on each of
the trial values of old Ql,i, did Q ~ and old QP3- Old Qm the opposing surfanes are in contact in N discrete areas at any
equals half of the sum of old QPl and old QP2. Rows 107- given instant.
117 provide the three values of WRrH2 (namely WnTH2, The paper then summarized the experimental and theoret-
W~rm2 and WnTnzO and Rows 119-130 provide the three ical heat flow analyses and used the numerical method of
values of Wrt~s (namely WnTHSl, I~RTH32 and W~T,33) for halving the interval to find appropriate values of N to satisfy
the three trial values in Rows 89-91. Rows 131-138 of the equation ( D r n - D s) =0.
Table 2 contain the data necessary for obtaining values of Qr The original theory of oxidational weax was reviewed.
that will give values of (WR~:~m-WrtTu) dose enough to Because of the fact that static oxidation constants would not
zero to satisfy the required accuraoy. The descriptionsof these give WRI". Values anywhere close m, WnEx~, a method was
rows. as given in the table, are self-explanatory. In the inter- doseribedfordeducingAp and QF from the tribo-systamitsdf.
ests ofbeing ablate showthe results and calculationsrelevant Due to the very low values obtained for Qm and QPs
to all seven experiments on just one spreadsheet, a manual (compared with the slalicoxidation of iron values), there
iteration method was used to deduce values for old Qp~ and was a clearneed fur an alternativetheory.The generaltheury
old QP2 which rapidly converged to values that will make of oxidationalwear seoms to be thattheory.
(W~EX~-- WIn,H) in Rows 131. 133 and 135 at least three The application of the general theory to the analysis of the
orders of magnitude less than the WRTUvalues in Rows 132, oxidational wear of high-~hromium ferrilic sleet against aus-
134 and 136. tenitic stainless steel was made easierby the use ofeleelronid
From Table 2 wc see that, for the mild oxidational wear of spreadsheets, both for display and for the application of halv-
Expts 1, 3 and 5, a tdbologieal oxidation activation energy ing the interval for finding consistent values of Qp to make
(QpM)of about 25,124 kJ mol= t inserted [ntu Eq, (38) will Wnru equal to W n ~ . to within 0.1% accuracy. The average
give theoretical wear rates to within about O.1%. The contact vab.,~s of Qp are: average QvM=25.124 J tool-I; average
temperatures (TF), whic~ zzc also assumed to be the oxida- QPs =35.347 k.I reel -I.
tion temperatures, were all around 390 "C. This is about 12%
of the static oxidation of iron [ 12] value of 208 kJ tool- t for
the same temperature range. We may also sea from Table 2 Appendix A. Nomenclature
that, for the severe oxidational wear of E,-pts 9, 11 and 13, a
tribolog[eal oxidation activation energy (QPs) of about A Radius of each of the N points of contact
35.347 IO tool -~ will also give Wr~rn to within 0.1%, when Ap Parabolic Arthenius eonstant (static
inserted into Eq. (38), the contact tempeiatures (To) being oxidation)
729 °C. This is about 17% of the static oxidation of iron value Am Arrhenius constant for mild oxidations wear
~f 210 ~ tool- t for temperatures greater than 600 °C. Aes Arrhenius eonstant for ~vem oxidationa[ wear
 ZFJ, Quino/Wear 199 0996) 169-180 179

A, Real area ofcoutact betweenpin and disk Fictitious~mperatures which would be

surfaces ob~ned assumingall the frictional heat g~s
C Specific heat eapacity of disk material into the pin, disk
c Heat flow/temperaturegradient along T~s, TpD General surfaceUemperat-reof pin, disk
~hermo~ouple wire measuring T^ U Linear sliding speed
DE, DTH Experimental,theoreticaldivision of heat it[to W Normal appliedload
the pin at the sliding interface wnc Wear rate coefficient (i.e. wear rate/toad)
Mass uptake of oxygen ~ r unit area Experimentallymeasured volume wear rate
Shear frictionalforce at sliding imerface (i.e. volume of material removed per unit
Fraction of oxide that is oxygen sliding distance)
Fo at II1¢conlact temperature (To) WeTe Theoretical volume wear rate
F o at the general surface temperature (Tps)
H Heat transfer ~ceffiuientfor a cylindrical steel
surface losing heat by convectionin an air
moving at approximatelythe same speed as the
mtatiug disk surface References
H, Linear heal flow enteringpin at sliding
interface [ 1 ] T.F.J. Quit~n, Tt~e role of oxid~ion in the mild wear of sleel. Be. J.
Linear heat flow leaving the exposed portion of AppL Phyx.. 13 (19621 33.
pin [2] T F L Quinn. The effect of hol.s~t lempe~en~s on file untubricaled
Linear heat flow entering the insdated portion wear,of sl~l, gSI.ETron~,, lfl (1967) 158.
Total frictionalheat dissipated at interface [3] H. Hong, R.E Hochman ~ T.FJ. Quinn, A new approach to Ihe
vxida~ioan] lhanry of mild wear. $T/.~ Tt~n$,, 31 (1988) 71,
K Factor of proportionalityin Archard's wear
[4 ] TFJ. Qdm,Tneeffec~sofo~-of-co~t~toxid~onon the oxid~ona]
law or $[ee|s,PrOd,3rd Tri~ r,~,~C~.~gtess, Wa,'~aw,]9~1.
Thermal conductivity of material of insulator, [51 T.FJ. Quinn. The dry wear ors~eel as revealed by eleclJ'c~ s " m c t ' ~
pin and X.my diffraction, Prec. lint. Mech. F~grx, 182 (1967-1q681 Pl
KoD ,
KoP Thermal conductivityof oxide on disk, pin 3N.
Parabolic oxidationrote constant [6l T.FJ. Quinnand .I.L.Woolley,The unlub6c.uedwearof 3¢,~Cr-
05%M0 steel, l..abr. Eng.. (1971) 314.
K~o Thermal conductivityof material of disk [7 ] JL. SU]llvan and S.S, Athwal, ~ mild wear of mIow-~lloy sl~l al
L; Lenglh ofexpogM portion ofpia te.lpet~tltre; up t0 5 ~ ~C Tribal Int, 16 ( 1983 ) 123.
Length of insulated poaion of the pin [8] T.FJ. Quinn aM W.O. Winer, The thenml anpects o f oxidational w~¢.
Number of contacts between pin and disk Wear, 102 (19851 67.
surface at any given instant [9l P, Gmsherg and. L M o l g a ~ Aspects of the wear of spinning
,iT 3.141593 uaveller~: the division or heat at rubbing suffuseS. Pr~l Iff$1. Much.
Eng., 181 (1966-1%7) Pc3L.
Hardness of substrata material of pin l 10] T.EJ. Quinn, The dividen of heat and surfacetempetatates at sliding
Q, Parabolic oxidation activation energy (static) sl.ce!inlerf~es ~ d their telalion to oxid~iom'~ Wear,ASLE Trdns,.2/
QrM Aclivation energy for mild oxidational wear ( 1978 [ 7g.
Q~ Activationenergy for severe oxidationatwear [ 11 ] J.L Sullivan, TFJ. Quinn and D.M. Rewson. Developmants in the
R Universal gas constant oxJdationa[ theory of mild wear. Tribol. I~t. 12 (1980) 153,
Outer radius of cylindrical insulator (see l]2l D. Copl~ and M. Cohen. The effect ofco]d wm-k on the oxida~an ¢~f
iron from 100 °C to 650 ~ . CorroS. So'., 6 ( 19661 321.
Fig. 3) [ 13] T.FJ, Quian. Oxid~fion~l wear modelling: I. Went, 153 (19921 179.
Density of oxide | 14[ T.F.J. Q,~inn,O x M ~ a l we~ m~eliin3: pat~I:(, The general
Density of oxide at coa~.ct lemperature (T~) of~xidational wear, Wear, IZT (19941 19~.
Density of disk material [ 15] J.R, Barber, "rhermoel~sfic instabilities in the sfiding of ~onfom~tng
Rlls Density of oxide at the geueml surface solids Pwe. R. So¢. ~ n ~ a t Se~ A' 312 { 1969) 381.
temperature (T~s) of the pin | 16] J.F. Ar~hard, Single ¢.o~u¢ts aa mulUplc ~..no.unlere. J. kl~pL P~.,
#2 (1961[) t420,
e~ Radius of the cylindricalpin [17] CB. Allen, T.F.J. Quinn and t.L. Sullivan, The oxidattomt wear of
t Time I~gh-chmmtum fen~lic s'l~:l cn au~enid¢ s'lalnless steel/. Tr/boL
T Temperature 107 ( i986} 172.
Thermoeouple readings at top and bottom of [ I$] D.M, Rowsenend T.FJ. qaian, F~tional heuingand the oxidatlon~
insulation (see Fig. 3) wear theory, 1 Pkys, D, 13 ( 19801 208,
Thermoeouplereading at cylindricalsurface of [19] J,F.Arcl~'d, The temgaratumof robbing s~l'eces, Wear, 2 (19591
438,
pin [20] gC, Lim and M.F. Ashby, Wens mechanism ma~, Ac-~ Meldll.. 35
r~ Contact temperature ( 19871 I.
Thickne~ of oxide on surface of disk, pin (21 ] J.V. Archmd cad W. Hi~, ~ w¢~trof m e ~ under Lu11~[ic~ed
To Temperatureof oxidation canditions, Prac, R. S~. A., 256 (19591 39,
]80 T.F.J. Qu~n/Wear 1~9(t996) 169-180

Biographies as the effects of extreme pressure lubricants on the scuffing

and pitting behaviour of bearing steels. He still mainlained
Terry Quinn: is a aibologist with a wide experience in apply- his interest in unlubriuated wear, however, which eventually
ing physical and computational methods to problems in lubri- led him, in 1981, to become Professor of Mechanical Engi-
cation, friction and wear. His first research projects were n~rlng at Georgia Instituteof Technology in Atlanta,Geor-
concerned with graphite lubrication and wear in the Surface gia, USA. This was where his interestin computational
Physi~ Department at the Associated Electrical In~!ustries modelling uf wear processes firstbegan. Dr.Quinn returned
Research Laboratories at Aldermaston in the UK. In the 1965 to the UK in 1986 where he becarac Professor of Engineering
he moved to the Aerospace Propulsion Research Institute at at the London Campus of the United S~cs International
San Antonio in Texas, USA, where his ideas on Oxidational University in Bushey, Hertfordshire until his retirement in
Wear were first given a mathamatical form. Upon moving 1993. During his time at Bushey, he wrote a book on "Phys-
back to the UK, he became Reader in Tribology(in 1974) at ical Analysis for Tribology" (1991) and published his first
Aston University, leading a Tribolngy Group involved in papers on Oxidational Wear Meddling. Hc continues with
cooperative research into various industrial problems, such his interest in computational methods applied to wear,