An investigation into sliding bearings suitable for a CO2-cooled nuclear reactor

35s

WEAR

AN INVESTIGATION

INTO SLIDING

FOR A COs-COOLED Hi. H. HEATH

BEARINGS

NUCLEAR

SUITABLE

REACTOR

ANI) Ii. F. PHILLIPS

A fonzic Poze’ev Division, Development Defiartment, EngineeringLabovatories, Whetstone, Leicester (Great Britain)

English Electric Co. Ltd.

SUMMARY In a nuclear reactor a number of mechanisms are required to operate up to about 4o0°C, often in the presence of intense radiation fluxes. Three possible solutions to the problem of making bearings for these mechanisms are discussed; the use of unlubricated metals, of massive graphite, and of thin graphite films on a metal base. Three metals, two nitrided steels and a tool steel, have been found to run well unlubricated, oath against themselves and a variety of other steels. Coefficients of friction for unlubricated metals are high, usually in the range 0.3 to 0.8 but tending on occasions towards unity. An electrographite was found which would run satisfactorily in dry COs but the performance in dry air, dry nitrogen and in a vacuum was poor. Thin graphite films have been found to have too unpredictable a life for practical purposes.

Bestrmmte Maschinenteile mtissen in einem Kernreaktor bei Arbeitstemperaturen bis au 400X funktionieren, h&fig in der Gegenwart intensiever Strahlungsstdsse. Drei mijgliche Losungen des Problems der Lagermaterialien solcher Konstruktionen werden iiberwogen : Gebrauch ungeschmierter Metalle, massiven Graphites oder diinner Graphitfilme auf einer Metallunterlage. Drei Metalle, namlich zwei nitrierte Stable und ein Werkzeugstahl, zeigten giinstige Laufeigenschaften ohne Schmierung, und zwar sowohl gegen das gleiche Material als such gegen eine Reihe anderer Stable gepaart. Die Reibungskoeffizienten fur ungeschmierte Stable sind hoch, meistens fallen sie zwischen 0.3 und 0.8 jedoch neigen sie manchmal dazu selbst den Wert I zu erreichen. Es wurde ein Elektrographit der in trockener COs Atmosphare befriedigende Betriebseigenschaften besitzt gefunden, jedoch waren seine Leistung in trockener Luft und trockenem Stickstoff unbefriedigend. Diinne Graphitfilme haben eine zu wenig zuverllssige Lebensdauer, urn praktisch brauchbar zu sein. INTRODUCTION

This paper describes work which has been done to find the friction and wear behaviour of a number of possible bearing materials for use within the pressure circuit of a graphite-moderated, COz-cooled, nuclear reactor. In these reactors a number of mechanisms, particularly those concerned with handling the fuel elements, are required to operate at temperatures up to about 400°C in both air and dry CO2 at gas pressures up to about zoo lb./in.2 (14.1 kg/cm2) and often in the presence of intense radiation fluxes. Bearings and lubricants for these environments must satisfy requirements other than the obvious one of resistance to seizure. They must be chemically and thermally stable in air and CO2 over a range of temperatures from 20 to 40o’C and they must Wear.3 (1960) 358-373

BEARINGS FOR A COa-COOLEDNUCLEARREACTOR be capable

of resisting

the radiation

fluxes they will meet in service. In addition

must be compatible

with the fuel elements;

cally with uranium

or canning

materials

they

that is to say, they must not react cheminor be such that they diffuse into the cans

to form alloys of lower melting point. In the paper the two types of test machine are described;

359

which were used for the experiments

the first of these was a pin and disc machine

and the second was a

reciprocating machine which was specially developed SO that work could be done under controlled atmospheric conditions. Results are then given of experiments on unlubricated metals, on massive graphite, and on thin graphite films upon a metal base. Unlubricated metals might offer the possibility of a very simple solution to the reactor bearing problem. The metals must not undergo phase changes or significant softening up to 400°C and they should preferably be readily machineable into quite complex bearing components. Whilst reasonable corrosion resistance is desirable it was suspected that a slow-forming oxide film might well be beneficial in reducing metalto-metal contact. Many of the common bearing metals such as copper, tin, brasses and bronzes are not admissable because they are incompatible with the fuel elements. Massive graphite (and carbon) are traditional materials for use in difficult lubrication situations but there are some disadvantages. Graphite has a coefficient of expansion a third or less than that of most steels and consequently there are problems of maintaining bearing clearances and of retaining the graphite in metallic housings when working over a temperature range of 20 to 400°C. In addition, graphites are weak materials compared with steel and their load-carrying capacities are correspondingly less. Another feature of graphite which has been given prominence is its liability to catastrophic wear in dry atmospheres. This phenomena (sometimes known as high altitude wear because of its occurrence in high-flying aircraft) has been studied amongst others by CAMPBELLAND KOZAK~, SAVAGESand SIMS~. High-altitude wear is affected by the temperature, the material against which the graphite runs, and the partial pressures of gases such as water, oxygen and carbon dioxide in the surrounding atmosphere. The characteristics of graphites can be modified This impregnation improves the strength, brings

by impregnating them with metals. the coefficient of expansion closer

to that of steel and can have beneficial effects on the wear properties in dry atmospheres. Unfortunately, few of the metals used to impregnate graphites are both compatible with the reactor fuel elements and resistant to corrosion in COz. In addition, the presence of metal often raises the coefficient of friction so that one of the advantages of a graphite bearing is lost. No work on such impregnated graphites is reported here. Graphite films laid thinly upon a metal substrate provide another well known method of dry lubrication. The main problems associated with the choice of such a lubricant for use in a reactor are that it should be chemically stable over the required temperature range, capable of operating in a dry COZ atmosphere, and adherent to the metal substrate. Wear,3 (WW

358-373

360

H. H. HEATH,

K. F. PHILLIPS

There appear to be two methods of attaching graphite films to a substrate: by burnishing or working the graphite on to the substrate, or by mixing it with an adhesive bond such as resin. Tt is difficult to find a resin which wiIf remain adhesive at 4oo’C. Nevertheless, a resin may hold the graphite in place for a period of low temperature working which is long enough to establish a good bond; this bond may, in turn, persist at high temperature. resin will not impair the lubricating

Such a tactic pre-supposes properties of the graphite

that the decomposed film.

APPARATUS

The work described in this paper was done on two types of machine; the first was a pin and disc machine, operating in undried COa, the second a reciprocating machine into which gases of known humidity could be introduced,

Fig.

1. Diagram

of pin and disc machine.

The pin and disc machine is shown diagrammatically in Fig. I. 3t consists of a number of 4 in. diameter discs driven from a common shaft and a number of dead weight loaded & in. diameter pins, which can be made to bear on the periphery of the discs, The whole is enclosed in a container which can be heated and fed with undried COa at a slight positive pressure. The rec~~ro~a~ing machine is shown in Figs. z. and 3. The reciprocating specimen is a cylinder, Q in. diameter x I~1 in. long with a 2 inn. wide flat along its upper

f*mcouple

entry

Fig. z. Diagram

of reciprocating

machine.

BEARINGS

surface. The reciprocating with 120' included

FOR A COs-COOLED NUCLEARREACTOR

cylinder

bears on a lower member in the form of a vee block

angle, which is in turn carried on a bracket

flange. The upper specimen

361

attached

to the main

is a flat plate 4 in. x 2 in. x Q in. thick which is pegged

to a carrier and pressed downwards by the loading mechanism. The test surfaces are the flat on the cylinder and the lower face of the plate attached to the upper carrier; contact between the cylinder and the vee block is regarded as a slave bearing.

Specimen carrier

.Loading weights

Fig. 3. Layout of loading mechanism for reciprocating machine.

The loading is derived from a weight and lever system. An arrangement of hardened steel knife edges transmits the load to the upper specimen carrier and constrains the upper specimen so that it is free to move only in the axial direction. This movement is resisted by a strain gauge plate which provides a measure of the frictional force. The signal from the strain gauge is amplified and displayed on a cathode ray oscilloscope. Friction values are calculated from the deflexion of the trace when the direction of specimen movement reverses; this eliminates zero drift of amplifiers and strain gauge. No difference in frictional force is noticed owing to the varying speed of the specimen

during its travel.

The reciprocating drive is obtained from an eccentric and is introduced into the machine through a stainless steel bellows. The movement per cycle is I in. and the speed is controlled by an infinitely variable gearbox. A mild steel hood is bolted over the assembled specimens and the whole made vacuum tgiht. A water-cooled extension of the hood carries the strain gauge, which thus operates at a sensibly constant temperature and avoids awkward compensating arrangements. Heating elements are fitted to the outside of the hood. Temperature is controlled by a thermocouple which is placed in the gas near the specimens; another thermocouple is attached to the upper specimen. These read to within 15°C at temperatures up to 400°C. For runs other than in air the gases are of commercial quality obtained from bottles. For runs at low humidity the machine is first evacuated for 12 h at 70°C at an absolute pressure of less than I ,u of mercury; leak rates are kept below IO ,U in IO min. The machine is cooled under vacuum before allowing the gas to flow through at approxiWear, 3 0960) 358-373

362 mately

H. H. HEATH,

0.5

I(. F. PHILLIPS

l/min. The gas is dried by passing through silica gel and magnesium per-

chlorate from which it emerges with a humidity of better than IO p.p.m. by volume; thereafter the gas picks up moisture slightly from the walls of the apparatus. The humidity of the gas is monitored as it emerges from the machine on an electrolytic hygrometer and the reading is taken as the humidity at the specimens. In practice, humidities of better than 25 p.p.m. by volume can be regularly obtained at the hygrometer. When humidities higher than this are required a part of the gas flow is obtained from a by-pass line round the driers. The pressure of the gas is maintained constant at IO lb./in.” gauge. RESULTS

The results obtained on the pin and disc machine and on the reciprocating machine are given below. Experiments on unlubricated metals are discussed first, followed by those on massive graphite and lastly those on graphite films. Unlubricated metals A number of metals was tested in the pin and disc machine under an applied load of 12 lb. (5.44 kg), a speed of 62.5 ft./min (19.1 m/min) (60 rev.[min) and an atmosphere of undried COZ gas. Pins and discs were ground to surface finishes of 8-15 pin. c.1.a. Specimens were degreased in carbon tetrachloride and then heated for 12 h in undried CO2 at 400°C. This was done so that any tempering could take place before running TABLE CHARACTERISTICS

En41 A En 40 B T.S. P. 1 D.S. s. 1 En 56 D P. 2 s. En

2

57

A.I. H.C. S.G.I. En 3 A Ti En 31 S.G.N. En58B G.I. Ni.R.

Nitriding steel* * 1000* goo* Nitriding steel** 85o* High speed tool steel Sprayed hard alloy plating 740 Die steel 675* Martensitic stainless steel 640* Martensitic stainless steel 540* Sprayed hard alloy plating 480 460’ Stainless steel 315 460” Martensitic stainless steel 305 Acicular iron 455 Hardchrome plate 450 S.G. Iron BS 2789 Type I 3oo Mild steel 220 Titanium 215 Low chrome steel 210 S.G. Ductile Niresist 200 Austenitic stainless steel 185 Grey iron 180 Type 1 Niresist 165

I

OF METALS

TESTED

1.6

0.3 0.3 0.75 4.25 1.5 0.8 0.3 1.25

16.5 13.0 23.0

0.07

16.0

0.16

16.5

3.0

4.25 16.0 12.0

0.35 0.3 0.15

1.5 0.3 0.3 0.4

0.4

3.0

2.3

3.6 0.2

2.0 0.2

1.0

.I,4

3.0

0.75

0.1

3.0 3.0

18 2.1

0.5 Mn 0.45 Mn 18 W 8.0 MO 1.0 0.9

1.1

V

MO

Mn 0.3 Mn 8.0 Mo 1.5 MO 1.5 cu

0.8

0.5

Mn

1.1 Mn 0.9 MO 0.5 0.7

Mn Mn

0.4 1.4 2.0 0.9 0.5 10.0

Al MO V Ni v MO CO

6.0 Ni 0.3 Ti 2.5 Ni 1.2 Ni 0.8 Ni

0.45 Mn 0.8 Mn 20.5 Ni 1.7 Mn 8.5 Ni 0.6 Ti 0.8 Mn 0.15 P 0.1 s 1.6 6.5 Cu I.75 1.25 Mn 15.5 Ni

0.25 2.6 0.8

* After hardening and tempering, Other hardnesses are for metals as manufactured. ** After nitriding, the friable white layer was ground from the test surfaces. Wear. 3 (196o) 358-373

BEARINGS

commenced

FORA C02-~~~~~~ NUCLEAR REACTOR

and also in an attempt

to obtain

some standard

oxide film condition

the specimen surfaces. The wear tests were carried out with the end faces of the pins running same tracks

on the discs for 35,000 revolutions.

363

against

on the

This 35,000 revolutions was made

up as follows: 5,000 revolutions at room temperature, followed by 5,000 revolutions at 360°C to determine the effect of temperature, and finally 25,000 revolutions at 360°C to determine the effect of running time. A number of spot checks running specimens in the reverse order, i.e. first at 360°C and finally at room temperature, have shown that the observed effects were due to temperature and not to the specimens having run in. A metal tray was placed beneath each disc so that the wear debris could be collected. The weight of the debris and the weight lost from the pin were recorded, as was the nature of the wear track. The weight lost from the pin was an immediate measure of its wear rate. It was impracticable, however, to measure the 4 in. diameter disc to the necessary precision for its wear rate to be determined and the weight of debris was taken as indicative of the wear behaviour of the complete pin/disc combination from which an assessment of the disc wear could be obtained. Table I gives the main properties of the materials tested, and the wear rates are given in Table II. In order to present the information compactly the wear rates in Table II have been coded; the first three digits, (which are to be read separately) represent respectively the wear rate of the pin during 5,000 revolutions at room temperature, the wear rate during 5,000 revolutions at 360°C and lastly, the wear rate during 25,000 revolutions at 360°C. The second three digits which follow the stroke represent the amount of debris collected under the same conditions. The assessment of the wear scars after 35,000 revolutions is similarly coded in Table III, pin first, followed by the disc. In both Tables II and III low numbers represent the best results. By combining the information given in Tables II and III it has been found possible to select metals which will run successfully unlubricated in CO2 gas. Tests on the reciprocating machine were carried out under the conditions stated in Table IV. The specimens used for tests on the reciprocating machine had a surface finish of from 8-12 pin. c.1.a. and were cleaned with carbon tetrachloride immediately before assembly. No pre-baking or tempering was carried out. The results are given in Table IV and are in general agreement with those found on the pin and disc machine. From these results it can be seen that the coefficients of friction at 28o’C to 400°C in dry CO2 were generally lower than for room temperature. Two materials tested, S2 and Hardchrome plate, suffered gross wear. The effect of temperature on the coefficient of friction of En 41 A running against itself unlubricated in dry CO2 can be seen in Fig. 4; in spite of fluctuations there is a decided reduction in the coefficient of friction at and above 300°C. No change in the coefficient of friction with CO2 moisture content has been detected over the moisture range IO-1,000 p.p.m. when running nitrided steels for a period of 38 h (9,600 ft.) both at 2oT and 360°C. Wear,3 (I960) 358-373

364

H. H. HEATH.

K. F. PHILLIPS

Wear,

3 (r#o)

358-373

% m c? 2

s

z

3 e W

2

I

Ni.R.

211

112

312

En 58 B

G.I.

I/I

312

2/I

312

I/I

I/2

S.G.N.

En 31

Ti

En 3 A

S.G.I.

s.

En 57 Not hardened

Not hardened

A.I.

57

2

s.

En

2

P.

En56D

D.S.

I

312

=I= 112 112

s. I

P.

T.S.

En 40 B

=I2 212 112 =I2 =I=

En4rA

En 41 A

pkisc

3-

2 -

Code. I -

III

=I=

212 212 211

En 40 B

211

311

312

=I=

=I= 211

=I=

=I= =I= =I=

T.S.

OF TEST

41

III

112

P. I

SURFACES

TABLE

111

515

211

215

213

z/3

315

III

3/5

512

2/I

213

211

212

215

D.S. s. I

OF THE

LISTED

515

415

213

211

112

213

~~

En56D

TESTS

4 - Badly scored. 5 -Scuffed.

515

213

3/I

=I= 212

212

~-

III

ON COMPLETION

Polished wear. Polished wear with grooves. Scored.

CONDITIONS

3/4

213

2/3

P. 2

IN TABLE

11

513

313

213

215

Hardened En 57

515

III

415

212

211

212

S.G.I.

415

515

515

212

I/5

215

S.G.N.

750 750 24.750

En 41 A

S.G.I.

S.G.N.

H.C.

En 40 B

En 41 A

En41 A

En 40 B

10

25 15

COZ Air CO? co2

18

r,joo

750

75O

750 750 12,000

400

L8

18

Air CO2 CO2

18 18 4oo

750 750 5,000 20

35

45 40

Air CO2 CO2

750 750 49.500 20

30 40

18 18 330

Air CO2 CO2

by vol.)

content

Moisture (fi.~.m.

10

~_~___

GIL5

ON THE

RECIPROCATING

MACHINE

0.8

0.64 0.4 Aver. 0.44 Max. 0.6 Min. 0.3 0.64 0.64

0.6 Aver. 0.4 Max. 0.5 Min. 0.3 0.7

L5,OOO _~~ -______

8,000

26,250

51,000

0.2 -0.8 0.6 Aver. 0.5 Max. 0.6 Min. 0.36 0.5 0.4 Aver. 0.3 Max. 0.4 Min. 0.2

734,750

0.6 0.6 Aver. 0.3 Max. 0.55 Min. 0.15

-

1.13 pg/ft.

0.87 ,ug/ft.

0.43 pg/ft.

0.88 C”g/ft.

I.02 ,ug/ft.

Speed 4.2 ft./min (1.28 m/min) (50 c/min) -

Air CO2 CO2

18 18 330

(“C)

Tem+mture

IV

METALS

TABLE UNLUBRICATED

18 18 330

1,000 750 41,000 +92,ooo at 120 c.p.m.

En 4~ A

En 41 A

(ft.)

Distance

Plate

CYlindeV

Material

OF RUNNING

Load 13.6 lb. (146 lb./in.2 10.3 kg/cmz)

RESULTS

Plate track corroded Cylinder polished

Cylinder polished

Plate slight pick-up

Plate polished and grooved Cylinder corroded

Plate polished with slight grooves Cylinder corroded

Plate and cylinder polished with slight grooves

s. 2

2

Not hardened

s.

H.C.

s. 2

2

Not hardened

s.

H.C.

Not hardened

s. I

I

s.

s. I

Not hardened

Air CO2 CO2

Air CO2 CO2 CO2 Air coa

230 330

18 18

1,250 625

18

18

18 330

25

4oo

100

IO

30 50

z::

18

x5 40

Air

25 45

18

Air CO2 CO2

0.6 0.6

0.8

0.9 0.9 0.7

0.4 0.45 Aver. 0.35 Max. 0.5 Min. 0.3

0.35 0.48 Aver. 0.32 Max. 0.4 Min. 0.3

0.5 0.55 Aver. 0.3 Max. 0.4 Min. 0.25

IV (continued)

18 330

330

18

18

125 250 87.5 750

750 750 33,550

750 750 22,375

750 7.50 22,250

TABLE

1.875

2,000

35,ooo

23.875

23.750

pglft.

pg/ft.

3oo

yglft.

0.01 in. Wear on both faces

445 rglft.

0.5

2.8

Plate and cylinder track scuffing

Plate and cylinder hardchrome track worn through

Plate track pick-up Cylinder track corroded

Plate and cylinder track powdery corrosion

Plate track pick-up Cylinder track scoring

368

H. H. HEATH,

I<. F. PHILLIPS

All the hardened metals used in these experiments have been found to retain their cold hardnesses when baked for

2,000

h at 400°C.

I

2

Running time, h

Fig. 4a. The effect of temperature on the friction of En 41 A running against itself: lower temperatures first. Reciprocating machine, 4.2 ft./mm, 13.6 lb. load, COZ humidity approx. 20 p.p.m.

4

6 8 Running time, h

10

12

Fig. 4b. The effect of temperature on the friction of En 41 A running against itself: higher temperatures first. Reciprocating machine, 4.2 ft./min, 13.6 lb. load, CO2 humidity approx. 20 p.p.m.

Massive graphite

Two electrographites “A” and “B” and a carbon “Cl’, which is the parent material from which “A” is graphitized, were tested. The experiments were done on the reciprocating machine with an applied load of 13.6 lb. (146 lb.lin.2, 10.3 k&ma) and a mean speed of 4.2 ft./min (1.28 m/min). Table V gives some of the properties of the test materials. The finely machined graphite surfaces were not cleaned or treated in any way prior to assembly. TABLE

V

CHARACTERISTICS OF GRAPHITEAND Designation

A B C

Description

Electrographite Electrographite Ungraphitized version of .4

CARBON

Shore hardness

Bulk density

Young’s modulus (1b.h~)

50 50 go/go

I.66 I.75 I.59

0.8 .I00 0.65.106 r.6g.106

Coeffii& of tkmal expansion per “C

4.5.10-6 2.8.10-G 4.7.10-G

-

Tests which consisted of 750 ft. in air and 750 ft. in CO2 both at room temperature, followed by 12,800 ft. in CO2 at 330°C were made on graphites “A” and “B”. The CO2 moisture content was maintained at less than 50 p.p.m. It was found that the coefficient of friction of “A” tended to be lower than that of “B”, whilst the wear rate of “A” was significantly lower (0.36 ,ug/ft. as against 5.6 rg/ft.). As a result of these tests work was discontinued on graphite “B”. Wear, 3 (1960) 358-373

BEARINGS

FORA C02-~~~~~~ NUCLEAR REACTOR

369

Graphite “A” when run against itself had a coefficient of friction of 0.3 at room temperature in air; at room temperature in dry COZ the coefficient of friction fell from 0.25 to 0.06 during a prolonged run. At 330°C in dry COZ the coefficient of friction was 0.06. Between 330°C and 400°C high friction and gross wear occurred in dry COZ. When a graphite “A” plate was run against an En 41 A nitrided efficient steady

of friction

at room temperature

cylinder

the co-

in air was 0.3; at 280% in dry CO2 it was

at 0.05. In both cases the wear rates were too small to measure.

Between

and 400°C in dry CO2 the wear rate remained low but the coefficient of friction a run of 1,000 ft. alternated between approximately 0.05 and 0.4.

300

during

A graphite “A” plate was also run against an En 58 B stainless steel cylinder. The coefficient of friction at room temperature in air rose from 0.16 to 0.24 during a run of 750 ft. At 28o“C in dry CO2 the coefficient of friction fluctuated between 0.05 and 0.3. Gross wear occurred during the run at 28o’C.

randomly

In an attempt to discover whether dry CO2 had an effect on the friction and wear of graphite the “A” material was run in a variety of atmospheres. The results are

Fig. 5. The effect of alternating dry CO2 with vacuum on the friction of graphite “A” running against itself. Reciprocating machine, 4.2 ft./min, 13.6 lb. load, 18°C. CO2 humidity approx. 20 p.p.m., vacuum below I p of mercury.

2

4

6

8 10 12 14 16 l8 20 22 Running time, h

Fig. 6. The effect of alternatingdry CO2 with dry nitrogen on the friction of graphite “A” running against itself. Reciprocating machine, 4.2 ft./min, 13.6 lb. load, 18’C.

Running time, h

Fig. 7. The effect of alternating dry CO% with dry air on the friction of graphite“A” runningagainstitself.Reciprocatingmachine, 4.2 ft./min, 13.6 lb. load, 18°C. Wear,3 (1960)358-373

Ii. H. HEATH, Ic. 1;. PHILLIPS

370 illustrated

in Figs. 5-7. In COZ a lower coefficient

of friction

was found

than

in a

vacuum of less than I ~1 of mercury, but repeated rise in the friction level observed in CO2 indicating

runs in vacuum led to a steady progressive deterioration of the

surface, which was attributed

A similar effect was found when

dry nitrogen

alternated

when dry air alternated

to the runs in vacuum.

with dry CO2. A somewhat with dry COz: friction

but the total wear and the friction

anomalous

result

was obtained

in dry air was lower than in dry CO2

level were higher than would have been expected

from a run in dry CO2 alone. This last result was attributed to disruption of thesurfaces which occurred during the runs in dry air even though the friction levels were low, a reminder that high wear rates are not necessarily associated with high friction. It is apparent that dry CO2 can inhibit wear on graphite, a result which is in agreement with other workers, e.g. CAMPBELL AND KOZAK~. A comparison between the two related materials “A” (graphitized) and “C” (ungraphitized) was also made. Initially the “C” material was found to be much less sensitive

to variations

in atmosphere

than

the “A”.

After

a few hundred

feet of

rubbing, however, “C” began to behave in the same way as “A”, from which it was originally assumed that a graphoid surface layer had been formed during the experiments. Electron diffraction failed to show the presence of a graphoid layer, and the cause of the behaviour

remains

uncertain.

Graphite jilms Preliminary experiments were made to assess the usefulness of a range of resinbonded colloidal graphite films. It was known that the resins used would not withstand 400°C for long periods but it was felt that the initial bond might help the establishment of a burnished graphoid layer on the metal substrate. Results were disappointing and comparison between resin-bonded films and films deposited directly from colloidal graphite

in water showed no significant

difference.

In all cases life in hot, dry CO2

was limited to a few thousand feet of rubbing. The life in air, both at room and elevated temperatures, was considerably longer than in dry CO2 but failure usually occurred after temperature cycling. This effect, which is most apparent on surfaces which have been run, is attributed to thermal stresses at the graphite metal interface. Following the work of BISSON, JOHNSON AND ANDERSOX~ and of PETERSON AND JOHNSON~, who found that the performance of graphite films could be improved by incorporating metallic salts, experiments were carried out using a variety of additives. The tests in air were carried out on both machines and in those in dry COZ on the reciprocating machine only. For each experiment an untreated surface was run against a treated surface, An assessment of the results obtained with each additive is given in Table VI. Failure was defined as the onset of high friction or of a high rate of wear. Most of the additives improved the performance of graphite films to some extent both in air and CO2, giving a longer life and a more polished type of wear. In all cases, however, failure occurred after a few temperature cycles up to 400°C. None of the additives tested was considered to give sufficient all-round improvement

2

ti ?

J$

z+

3 g z w

-

In air

TABLE

VI

on the ru‘iprccating

machine

H,O

at13.6 tb.

< IOO [email protected]. load r.zjt./mifi

SUBSTRATEAGAINSTUNLUBRICATEDE~

In dry CO,

En3A

(1.28

m/tnin)

3A

As for graphite alone.

Silica At 18°C inferior to graphite alone. aerogels At 360°C inferior to graphite alone.

At 18°C similar to graphite alone. At 380°C run for tens of thousands of feet without breakdown. Increasing gas moisture content increases the coefficient of friction.

Not tested.

Na&Oa

At 18°C similar to graphite alone. At 360°C similar to graphite alone.

At 380°C run for few thousand feet without breakdown. Breakdown after temperature cycling.

CdCOa At 18’C run for hundreds of thousands of feet before failure. At 360°C run for tens of thousands of feet before failure.

Not tested.

Not tested.

At 18°C inferior to graphite alone.

At 18°C run for tens of thousands of feet before failure. At 380°C as at 18X. PbSOd found to form slowly at 400°C.

At 18°C nature of wear more polished than graphite alone. Run for hundreds of thousands of feet before failure. At 360°C as at 18°C but with slightly shorter life.

At 18°C run for hundreds of thousands of feet before failure. At 360°C as at 18°C. Slightly better than graphite in temperature cycling. Inhibits wear for short periods after lubricant film breakdown. Found to form PbS04 after baking at 4oo°C.

Cd(OH)z At 18°C inferior to graphite alone.

Cd0

PbS04

PbS

Not tested.

At 18°C not tested. At 380°C run for few thousand feet only before failure.

Not tested.

18°C distance to breakdown longer than for graphite alone. 360°C as at 18°C. As graphite at temperature cycling. 18°C similar to PbC03. 360°C similar to PbCOa.

At 18°C inferior to graphite alone.

Pba0.r

PbCOa

At At PbMo04 At At

At 18°C run for few thousand feet without breakdown. At 300°C similar to performance at 18°C. Lubrication properties deteriorate with baking at 400% in COZ. Suspected change to PbsO, or to PbCOa.

At 18°C nature of wear more polished than graphite alone. At 380°C similar to graphite alone. Initially better than graphite at temperature cycling. Lubrication properties deteriorate with baking at 4oo’C. At 400% forms Pb304.

PbO

At r8OC prevents metal to metal contact for a few thousand feet only. At 380°C inferior to performance at 18°C.

oft&s

ADDITIVESONA

Assessment

GRAPHITEFILMSINCORPORATING

Assessment oft&s on the 9% and disc machine at IZ lb. load 62.5 ft./m& (19.1 m/min) and m the reciprocating machine at 13.6lb.lmd 4.*jt./min(r.a8m/min)

FRESULTSOBTAINEDRUNNINGTHIN

Graphite At I 8’C prevents metal-to-metal contact for tens of thousands of feet. At 380°C breaks down after a few thousand feet unable to withalone stand temperature cycling after running.

Additive N ~~I.50by dry weight

ASSESSMENTS•

372

H. H. HEATH,K. F. PHILLIPS

to make its inclusion worthwhile for the range of conditions of interest. The additives which gave the best friction and wear characteristics stable in the required environments.

were found not to be chemically

It became apparent during these experiments that thin graphite lubricant films were liable to break down locally to allow metal-to-metal

contact after an unpre-

dictable interval, which could be quite short. If the substrates had poor resistance to scuffing such local contact spread to disrupt the whole bearing surface. It is considered advisable to limit the application of thin lubricant films to substrates which are able to run without scuffing in the absence of lubrication.

DISCUSSIONANDCONCLUSIONS A number of tests have been carried out on a selection of metals, all but one of which (titanium) are conventional. The results obtained are presented in a series of tables from which it is possible to obtain the relative merits of a number of combinations of these metals, when subjected to sliding unlubricated in COz. Three metals, En 40 B and En 41 A nitriding steels and a tool steel, have been found to give good all-round service when run against each other, against themselves and against most of the other metals tested. While there is little to choose between the nitriding steels and the tool steel on grounds of wear resistance, ease of machining and lack of distortion make the nitrided steels the preferred materials for many applications. In general, the coefficients of friction of unlubricated metals were less between 300 and 400°C than they were at room temperature when run in dry COs. The wear rate was often less at 360°C than at room temperature. It is not, therefore, safe to assume that a mechanism which has been proved hot will run satisfactorily cold. No correlation between friction and wear and the humidity of the CO2 has been found when running unlubricated En 41 A against itself for a period of 38 hours. The coefficient of friction of unlubricated metals varied widely during a run and on occasions approached unity, a point of some importance when link mechanisms are to be designed. A feature of all the experiments was the amount of oxidized debris formed even when wear was small. It is clear that practical mechanisms need careful attention to the disposal of this debris to prevent it accumulating and jamming the bearings. One of the electrographites tested was also found to behave well in dry COz, but this performance could not be maintained in a vacuum nor in dry air or nitrogen. The coefficient of friction of this electrographite in dry CO2 was found to be below 0.1 both when running against itself and against En 41 A. Above 300°C its behaviour tended to be erratic. Graphite films alone and with additives have not been found reliable. They appear to break down after an indeterminate length of run and it is recommended that for practical purposes the substrates should be able to run well unlubricated so that film failure need not be disastrous. Wear, 3 (MO) 358-373

BEARINGS

FOR A CopCOOLED

373

NUCLEAR REACTOR

ACKNOWLEDGEMENTS

The authors wish to thank the English Electric Co. Limited for permission to present this paper. REFERENCES 1 W. B. CAMPBELL AND R. KOZAK, Trans. Am. Sot. Mech. Engrs., 70 (1948) 491. z R. H. SAVAGE, Gen. Edec. Rev., 48 (x945) 13. 8 R. F. SIMS, Proc. Isst. Elec. Engrs. (Lmdon), IOI, Pt. II frg54) 213. Q E. B. BISSON, R. L. JOHNSON AND W. J. ANDERSON, Proc. Conf. Luby~ca~~~ and Wear, ~957~ 348 5 M. B. PETERXIN AND R. L. JORNSON, Lubricafion Eng., 13 (1957) 203. Received

March 16, 1960

Wear, 3 (w60)

358-373