The Effects of Electrical Conductivity and Oxidation Resistance On Temperatur Ise of Circuit-Breaker Contact
The Effects of Electrical Conductivity and Oxidation Resistance On Temperatur Ise of Circuit-Breaker Contact
The Effects of Electrical Conductivity and Oxidation Resistance On Temperatur Ise of Circuit-Breaker Contact
2,JUNE 1969 99
Absfraci-A study has been made of the bulk electrical con- net result is that silver is evaporated and the contact
ductivity and oxidation characteristics of several typical circuit- surfaces are oxidized. Depending upon the thickness, con-
breaker contact alloys. Results were correlated with temperature tinuity, and dielectric properties of the oxide film, surface
rise behavior of the same materials. It was concluded that surface
phenomena, particularly oxidation, exert a more potent effect on resistivity and the associated heating characteristics of
temperature characteristics than does bulk electrical conductivity. the contacts can be significantly affected.
To determine the relative importance of bulk resis-
tivity versus surface resistivity on temperature rise in
breakers, a study was conducted in which oxidation be-
LECTRICAL contacts for medium and heavy-duty
havior and electrical conductivity were measured for
circuit breakers are subjected to very severe
several contact materials. These were subsequently cor-
operating conditions. On the one hand, they are
related with temperature rise observations on breakers
required to conduct electricity without overheating the
equipped with the particular compositions in question.
contact joint. This suggests the use of a high conductivity
metal such as copper or silver. However, contacts must
EXPERIMENTAL PROCEDURE
also exhibit resistance to welding during arcing, which
dictates the selection of a high-melting-point refractory Three different types of electrical contact materials
metal or semi-metal. In practice, this dilemma has been were included in this study. These were Ag-W, Ag-Mo,
solved by the use of composite materials of silver plus and Ag-Tic. Silver contents of these grades ranged from
a refractory metal. Compositions that have proven most 30 to 50 volume percent (V/O). All of the materials re-
satisfactory over the years are Ag-W and Ag-Mo grades ported herein were prepared by powder metallurgy.
which contain approximately 50 to 65 percent Ag by Bars, ground on all surfaces, were employed for the
volume. The electrical conductivity of these materials is resistivity and oxidation tests. Resistivity was determined
generally at least 40 percent that, of copper. by passing a known current through the bar and measur-
Underwriter’s Laboratories (UL) specify that the tem- ing the voltage drop. Oxidation tests were conducted over
perature of the terminal of a circuit breaker must not the temperature range of 800-1400°F in a ceramic tube
exceed 50°C above ambient during operation at rated furnace using a flow of 0.4 ft3/hr of dry air. Weight
current and voltage. Essentially all of the heating ob- change versus time was determined by removing speci-
served at rated operating conditions arises due to the mens from the furnace, weighing them, and then returning
combined effect of two factors: resistance in the contacts them for further oxidation.
and resistance at the interface between the contacts. Cor- Temperature rise tests were conducted using 20-A
respondingly, there are two material properties that exert breakers equipped with %-inch-diameter by 0.047-inch-
a potent influence on temperature rise: the bulk (volume) thick contacts of the same compositions as the bars.
electrical resistivity of the material and its surface re- Temperature was measured with a thermocouple ce-
sistivity. mented to the terminal.
Standard circuit-breaker test conditions have little
effect on the bulk resistivity of the contacts, but the same RESULTS AND DISCUSSION
is not t,ruc for surface resistivity. During operation of Electrical resistivity values for the various types of
circuit breakers, arcing and high temperatures occur. The contact materials studied were measured over a range of
silver contents. These results are presented in Fig. 1. It
was observed that linear relationships exist in the com-
Manuscript received March 20, 1969; revised April 21, 1969. positional areas of interest. Further, for equivalent
The author is with the Stackpole Carbon Company, St. Marys,
Pa. volume percent silver, the electrical resistivities of Ag-W
IEEE ‘,!RANSACTIONS ON PARTS, MATERIALS AND PACKAGING, JUKE 1969
1
'Y-
a= Ag-TIC
O= Ag-WI 50% Ag-W
20 2.4 26 :32 36 40 44
‘% Ag
46 52 56 60
64
66
50V/oAg-Mo
and Ag-MO were low compared to those of Ag-Tic. Since \
very limited solid solution occurs in these systems, this
FiR. 2. B-hour oxidation of selected contact materials.
difference is directly attributable to the much lower re-
sie,tivity of tungsten (5.5 ~LR.cm) and molybdenum
(5.7 ,.&-cm) versus that of the refractory compound,
titanium carbide (200 ,.&.cm). Thus, on the basis of bulk
electrical resistivity only, we would expect the tempera-
ture rise performance of molybdenum-and tungsten-base
materials to be significantly better than that. of carbicle-
base grades.
The oxidation characteristics of pure tungsten, molyb-
denum, and titanium carbide have been reported pre-
viously [l] , [2], In brief, it has been shown that the re-
fractory metals exhibit relatively poor resistance to
oxidation at temperatures above about 600°F. In con-
tra& the compound titanium carbide exhibits excellent
oxidation resistance to temperatures of the order of 3
14.00-1800”:F. IO 20 XI405060 TO6060ICQ
TIMEtHRS)
The mechanism of oxidation is marked1.y different for
Fig. 3. lOOO”l~oxidation of selected contact materials.
the three refractories studied. The chief oxidation product
of tungsten at 8’00-1200°F is tungstic oxide, W03. This
oxidation layer grows steadily and, since the volume ratio molybdenum-base material is slightly positive at 800°F
of T?‘03 to tungsten metal is 3.35, the WOs film becomes but becomes negative at 1000°F and above. Such be-
quite thick in a short time. Further, because of the Ilarge havior is similar to that for the pure refractories in this
vohrme change, the oxide tends to crack and be porous. temperature range. As with pure molybdenum, the
The t.hick, porous layer developed is both insulating and weight loss of Ag-Mo can be attributed to the formation
nonprotective [ 11. Molybdenum, too, forms a voluminous, of the volatile oxide, MOO:,. Apparently the presence of
porous oxide at these temperatures. In addition, MoOs silver does not alter this oxidation mechanism.
volatizes at a very significant rate above 800°F. Oxida- Th.e effect of time at l.OOO”F on the absolute weight
tion of molybdenum is therefore rapid and is accom- change of the silver-refractory composites is plotted in
panied by a wei.ght loss and size reduction at tempera- Fig. 3. Again it will be noted that the oxidation of 30
tures greater than about 800’F [l]. In contrast to V/O .Ag-TIC is several orders of magnitude lower than
tungsten an.d molybdenum, titanium carbide oxidizes to that for 50 V/O Ag-W or 50 v/o Ag-Mo at 1000’F. Com-
form a tightly adherent oxide film. Further oxidation of pared with Ag-W and Ag--No, significantly less oxidation
carbide is retarded once a thin layer of the oxide has was observed for Ag-TiC over the entire temperature
formed [ 2.1. range investigated, SOO-1400°F.
The results of our studies of the oxidation character- Photographs of test specimens at various stages of
istics of silver-refractory systems paralleled those for the oxidation are shown in Fig. 4. Note that the tungsten
pure refracttories. In Fig. 2 is plotted the 6-hour oxidation composite mushrooms to several times its original size at
of 5Ov/o Ag-W, 50 v/o Ag-MO, and 30 v/o Ag-TiC at temperatures as low as 1000°F. In the same temperature
temperatures of 800-3400°F. It is readily apparent that range, the molybdenum-base specimens are reduced to
the oxidation resistance of Ag-TiC is markedly superior approximately one-half their original volume. Both of
to either of t.he other composites. It should also be noted these materials change color drastically during oxidation,
that the weight change due to oxidation of tungsten and the tungsten forming a porous green coating and the
titanium carbide composites is positive while that of the molybdenum a gray surface. In contrast to these ma-
KOSCO : TEMPERATURE RISD OF CIRCUIT-BREAKER CONTACT MATERIALS 101
100 hr at 100 hr at
800°F 1400°F
50 v/o Ag-W
6 hr
50 v/o Ag-Mo
24 hr
38 v/o Ag-TiC r
100 hr
Fig. 4. Oxidation of various contact materials.
170
160
150
I40
z Ix,
z 120
0
0 110
r
2 100
90
80
70
60
50
40
30
20
IO
0
1000 I100 12.00 1300 1400 0 DO0 2000 3000 4000 5ooO 6000
600 900
TEMPERATURE OF CYCLES
Fig. 5. 6-hour oxidzkion of selected contact materials. Fig. 6. Temperature rise during soak versus cycles before soak.
terials, titanium carbide-base composites retain both their 2) temperature rise was measured at start-up and after
size and silvery color even after 100 hours at 1400°F. each 1000 cycles by closing the breaker and soaking at
The highest temperature at which molybdenum and rated current and voltage until an equilibrium tempera-
tungsten composit,es were oxidized was 12OO”F, and even ture was attained. Arcing is visible during cycling so we
here the tests had to be discontinued after 6 to 24 hours can assume high surface temperatures which would cause
because of severe deterioration of the compacts. some degree of oxidation. Results are reported in Fig.
6-hour oxidation data at 800-1400°F for the three types 6. It was observed that Ag-W and Ag-Mo grades de-
of coniposites are presented in Fig. 5. Once again, the teriorated rapidly under the test conditions so that the
obvious superiority of carbide-base grades is apparent. acceptable temperature rise of 50°C was exceeded after
For example, the oxidation resistance of titanium carbide 1000 to 3000 cycles. In contrast, titanium carbide-base
at 1200°F is 2-3 orders of magnitude better than that of materials exhibited only a slight increase in temperature
either Ag-W or Ag-MO. Two Ag-TiC compositions, 30 rise in the course of the 6000 cycles. For example, the
and 38 V/D Ag, are reported in Fig. 5. Oxidation charac- temperature rise prior to any cycling for 38 V/O Ag-TiC
teristics were essentially the same for both. was equivalent to that for 50 v,/o Ag-W, about 2O”C,
To determine the relative importance of bulk con- despit,e the fact that the resistivity of the titanium car-
ductivity and oxidation resistance on temperature rise bide-base grade was three times higher than the tung-
performance, breakers equipped with contacts of the de- sten-base grade. After 6000 cycles, during which the con-
sired composition were subjected to a modified endurance tacts arced and oxidized, the temperature rise of Ag-TiC
test designed to accelerate surface deterioration. The test materia.1 was 30°C while that of Ag-W exceeded 150°C.
procedure was similar to the standard UL endurance test- In effect, the temperature rise characteristics of each
with two exceptions : 1) breakers were not overloaded and material occupy a particular temperature band. These
102 IEEE ‘TRIhNSACTIONS ON PARTS, MATERIALS AND PACKAGING, JUNE 1969
THICK INSUL-
ATING FILM
-EXTREMELY THIN
OXIDE FILM ON TIC
ONLY, Ag NOT
INSULATED
ba::tds overlap iin the region 0 to 1000 cycles; but after tact and causing high drops. We might expect, then, that
this, breakers equipped with tungsten-and molybdenum- temperature rise would be igreatest for this system, as has
base materials experienced rapid temperature increases beer. observed (Fig. 6).
while those with carbide contacts remz,ined relal;ively The oxide formed on Ag-MO is also voluminous and
s-table. The actual temperatures reached with Ag-W were covers the surface including nonoxidized silver areas.
the highest recorded, Ag-Mo was intermediate, and Ag- However, its maximum thickness is limited by volatiliza-
TiC! lowest. tion of MoOH so a thinne:r insulating film results than for
The evidence suggests that temperature rise character- silver tungsten. For this reason, the temperature rise of
istics of various contact materials can be correlated to Ag-MO in these tests is slightly less than Ag-W although
their oxidation behavior. Specifically, we believe that, the both are excessive.
ratio of the volume of oxide produced to the volume of Titanium carbide forms a thin, tight, adherent film
meLa which generates it, VO/VVZ, is a particularl;y im- with vo/v?n equal to about 1.0. Such an oxide film ex-
portant parameter. When this ratio approximates. 1.0, the hibits two advantages: ‘I) because it is very thin, it is
oxide is thin and tightly adherent, and prevents further low in resistance; and 2’) because t,he oxide volume ap-
oxidation. Titanium carbide exhibits such a vo/vm ratio, proxirnates that. of the metal which generated it, ad-
therefore only a thin oxide layer forms, and that only on joining silver areas are not covered with an insulating
the carbide. The silver is unaffected, structural integrity oxide layer. Both of these factors favor good electrical.
is retained, and surface resistivity remains low. In con- coma& and low drops. Thus, only a small temperature
trast, when vo/vm is of the order of 1.5 or grea.ter, a rise develops.
spongy, volumflnous oxide, which grows and covers the From the results of our investigation, we have con-
sil\,er, is formed. Such a porous layer tends to have a cluded that bulk electrical resistivity is of minor im-
strong insulating properties. For tungsten, voJvm equals porl;ance in determining the temperature rise charac-
2.11to 3.35, and we believe this accounts in large part for teristics of these contact materials. Instead, surface
tungsten’s poor temperature rise performance. resistjvity, which is strongly related to the oxidation be-
The effect of the oxidation of molybdenum is more havior of the composites, is the dominant factor.
difficult to explain. At 800”F, a slight weight gain is ob- A. comment might also be made regarding the silver
served, but this situation reverses as temperature in- con-tent of Ag-W and Ag--MO grades which have proven
creases so that a net weight loss occurs at higher tempera- sa,tisfactory. Commercial circuit breaker contacts usually
tures. The oxidation process here is complex and involves contain from 50 to 60 v/o Ag. As indicated in Fig. 1, in-
the generation of several sub-oxides of molybdenum creasing the silver content of Ag-W and Ag-Mo com-
(z~o/v?n q = 1.6 to 2.3) as well as the volatilization of posites from 25 to 60 vl)o effects only a relatively small
Mo03. The sub-oxides constitute an insulating film, the change in bulk resistivil;y. On the other hand, previoua
thickness of which is limited by the loss of MOO:< vapor work demonstrates that silver does not oxidize in air at
at, the air interface. Nonetheless, the sub-oxide film in- temperatures of SOO-1400°F [3]. In light of this, it seems
creases surface resistivity and this manifests itself as a likely that the major purpose of high-silver compositions
.temperature rise during breaker operation. in present contact materials is to enhance oxidation re-
Fig. 7 is a. schematic model of oxidation in the systems sistance and delay overgrowth of an insulating oxide
Ag-W, Ag-Mo, and Ag-Tic. As indicated, the oxide layer rather thsn to provide any useful improvement in
generated on A.g-W is more massive than for either of the bulk conductivity. As a corollary to this, lower silver
other systems. In fact: the volume of the oxide is so great content materials should prove satisfactory in circuit
t,h:tt it grows and covers the nonoxidized silver areas on breaker applications provided the refractory hard phase
the contact surface, thus preventing good electrical con- is sufficiently resistant to oxidation.
K”SC0 : TEMPERATURE RISE OF CIRCUIT-BREAKER CONTACT MATERIALS 103
John C. Kosco was born in Du Bois, Pa., in 1932. He received the B.S. degree in chemistry
from Notre Dame University, Notre Dame, Ind., in 1954, the M.S. degree in plastics
engineering from Princeton University, Princeton, N. J., in 1956, and the Ph.D. degree
in metallurgy from Pennsylvania State University, University Park, in 1958.
He has been employed by the Stackpole Carbon Company, St. Marys, Pa., since
1958, first as a Research Metallurgist and currently as the Director of Metallurgical
Research. His work has included problems in such fields as cermets, high-temperature
intermetallic materials, electrical contacts, tungsten and molybdenum refractory mate-
rials, and the wettability of metal ceramic systems. He has several publications and patents
in the fields of interest mentioned.
Dr. Kosco is a member of the American Society for Metals, the American Chemical
Society, and the American Powder Metallurgy Institute.