Silicon Contamination

IEEE TRANWCnOh'S ON COMPOhEhTS HYBRIDS, AND MANUFACTURING TECHNOLOOY, VOL CHMT-5. NO. 2JUNE 1982 , 281
; Influence of Silicone Contamination on Brush- Commutator Contacts in Small-Size
DC Motors AKlRA OKADA AND MlNORU TODA
Abstrocl-The influenee of silicones on the brush-commutator motors are of a conventional type used in many applications. cwtnet la dirert current (dc) motors wos investigated. The surfacesof Commutator material is an alloy of Ag, Cu, and Cd (Ag:92
silicm-con(Pndnated brushes and commutators were obswved by a
seanulogelectronmicroscope(SEMI. The surfacesa hwereanalyZCd
semiquaatitstively by means of Auger doctron spctroseopy (AES).
The results
suggest that Mctioaally pdymerized silicomes at the (Ag- onto a German silver plate. Two kinds of armatures were Pd u t a l y t i c ) mctnl surface form insulative layers on both the
commutator and the brush surfaces. Such eontsmlrutioa causes e l a c t phcmomena tbat impede motor startup at low vokages. Smsll quantities of varioclsfy located siticomc can a h conlpminste corn- matator aad brush surfaces even during room temperatun storage.
I. INTRODUCTION
slLICONE contamination of electrical contacts is familiar to people concerned with switchmg relays and other devices using electrical contacts. in recent years, several studies have been reported on the effect of silicones on switch- ing mechanisms.
Kitchen and Russel reported the influence of silicone oils on electrical contacts [ I ] . Witter and Leiper investigated the effects on contact resistance. Their report includes the study on the effect of various organic and inorganic materials in- cluding silicon [2]. ALSO, Hague and Spetgler studied the effect of silicone oils on relay contacts by means of Auger electronspectroscopy(AES) [3].
The troubles with the relay contacts and appropriate countermeasures are Mdy clear as described in [I]. However there are many types of silicone products to be considered along with the peculiarities of the various devices which may be exposed to them. For example, silicone resins widely used for electrical msulation are a possible source of exposure easily encountered by electrical parts. As shall be seen below, care must be taken in using such materials near electrical contacts, particularly those at the commutator-brush inter- face in small-size direct current (dc) motors.
This paper describes the influence of silicones on the commutator-brush interface and the possible mechanisms of the formation of intervening insulative layers. This work was carried out mainly by Auger electron spectroscopy, a method well-suited for the surface analysis of materials.
The motors used in these experiments were small-size dc motors of 35-mm diameter and 45-mm length. These
ManuscriptreceivedNovember 16, 1981;revied February 24, 1982.
The authors are with RCA Research Laboratories. Jnc,, 971-2 Aza 4-go, Zushi-rnachi, Machida City, Tokyo, 194-02 Japan.
prepared for the experiments. The difference between the two types of armatures was only in their insulating materials and the means of anchoring the coils to the armature cores. The armature fundamental structure was the same in both cases. In one type armature, the insulating material used on the coils was a conventional epoxy resin, the material used on the other type was a silicone resin of a type in general use with electrical parts and classified as such for electrical insulation and other electrical purposes.
Both types of armatures were well dried in an electric oven at 8s°C for 2 h. A total of 30 motors with the two types of armatures were prepared. Fifteen of these, denoted as type A, were assembled using the armatures with the silicone resin. The other 15 motors denoted type B, were assembled with the armatures not using the silicone resin. Both A and B types were divided into three categories for the experiments.
To make clear the difference between categories A and B units, five units from A and five units from B were tested at an accelerated condition at a temperature of 50aC. Another five units from A and five units from B were tested at room temperature. The other ten motors, five of A and five of B, were stored separately at room temperature for further studies.
The rated voltage of the motors was 20 V and the no-load starting voltages of the motors were less than 1 V. To investi- gate the starting properties of the motors, a start-stop sequence was repeated with a I Hz, half-wave rectified, sinus- oidal wave having a 33 percent duty ratio. The peak voltage was kept at 6 V during the test, and all motors were driven at the same time in the forward direction. The start-stop sequence was repeated up to 100 h. A schematic diagram of the driver wave form is shown in Fig. 1.
After the start-stop sequential test, all 30 motors were disassembled. The commutators and the brushes were care- fully removed from the motors and examined by a scanning electron microscope (SEM).Their surfaces were then analyzed by -Auger electron spectroscopy. The tools used in the sample preparation processes were cleaned before each step by reo on@ TF with an ultrasonic bath. The AES measurements were made with a cylindrical mirror analyzer. The primary beam energy and the current were 5 keV and 1 0 M , respectively.
@ Registered trademark of Du Pont.
0148-6411/82/0600-0281$00.75 O 1982 IEEE
percent, Cu:6 percent, Cd:2 percent). The brush material is a Ag-Pd alloy (Ag:70 percent, Pd:30 percent) cladded
282 1EEE TRANMCJlOhLS ON COMPONEhTS, HYBRIDS, AND MANUFACWlUNG TWHKOLOGY, VOL CHM 1.5, NO. 2 J W E1982
Fig. 1. Driver output voltage wavefonn(startstop sequential test). "On" cycle continues for 330 ms, and "off' cycle, 670 ms.
armture resin
Jnstable ' U.T. 1 N.T. after4 hrs 1
1
3 Unstable
I Unstable N.T. N.T.
5 Jnstrble 1 N.T. ( N.T. 1 N.T. 1 ifter 9 hrs
I1 N.T. denotes no-trouble during 100 h operation.
Commutator-brush structure. Commutator is alloy of Ag, Cu,
Fig. 2.
and Cd. Brush is Ag-Pd alloy Brush trace is commutator surface area where the brush contact make its tram. Beside trace is com. mutator arca immediately adjacent to thc brush trace. Contact area is area where electrical contact between the brush and commutator is formed on brush surface. Beside contact is area adjacent to con. tact area on the brush.
Throughout the experiments, the scanning speed and the time constant of $e lock-in amplifier were held constant at 2.5 eVls and 300 ms, respectively. The vacuum was held in the range 1.3X lo-' to 1.5X lo-' Pa.
To make semiquantitative analysis of sample surfaces, three AES measurements were repeated in the same surface area of each sample. The motor surface areas analyzed are shown in Fig. 2. The AES analysis on the as-assembled units were made over the whole area of the commutator and the brush surfaces.
To check possible human error and cross-contamination in the vacuum chamber, a small plate of Ag was investigated with each sample in the same analytical sequence. The Ag plates were prepared similarly to the samples. An Ag plate was mounted on the sample holder and monitored before each analysis of a sample. A new Ag plate was prepared for each sample preparation process for each AES measurement cycle.
III. RESULTS AND DISCUSSION
The results of the start-stop sequential test are summarized i~ Table I. All the motors in A group at 50°C showed startup difficulties within 9 h. One motor never rotated again after stopping 2 h from the start of the test. The other four units showed instabilities in starting properties during the test. During the "on" part of the drive cycle, these units sometimes rotated and sometimes not.
The A motors at room temperature, and ali of the B group at both temperatures, worked well during the 100h operation. In summary, startup trouble was only observed in the accel- erated tests on the motors using the silicone resin.
To distinguish the difference between the surface condi-
Typical SEM photograph of surface of brush trace on corn-
TABLE I
RESULTS OFSTART-STOP SEQUENTIAL TESTONBOTH TYPESOFMOTORS
Temperature 50'C Room teap. (21.C) II
InsulaLo: on S~Ylrane m x y resin Srllcone
reslll
1EPOXY resin
Fig. 3 .
mutator of A unit tested at 50°C. Many small powder-likedeposits can be seen.
tions on the commutarors and the brushes, SEM observation and AES analysis were made. Figs. 3 and 4 show typical SEM photographs of the commutator surfaces of A and B groups at 50°C,respectively.
As can be clearly seen, the surface of A is contaminated by many small particles while this tends not to be the case with the surface of B. The commutator used for the SEM observa- tion in Fig. 3 was removed from the A unit that showed insta- bility after 4 h from the start of the test. Figs. 5 and 6 are
typical SEM photographs of the commutator surfaces of room temperature A and B units. In Fig. 5 it is evident that con- tamination was similarly formed on the surface of A group, even at room temperature. The surface of B in Fig. 6 is much cleaner than that of the A unit. The SEMphotographs on the brush surfaces of both A and'B units at both temperatures did not show such clear differences. -
The AES analyses on both commutator and brush surfaces of the A and B units of each temperature type showed strong differences between groups A and B. Typical AES spectra are shown in Figs. 7 and 8. Fig. 7 is a typical spectrum from the brush trace of a commutator of a B unit operated at 50°C.A corresponding SEM photograph is shown in Fig, 4. It can be seen there are S, C1, C, 0,and Si contaminants on the surface. The change in the sensitivities with energy is given in the
N.T.
OKADA AND TODA SILICON COhTAMIN4nON ON BRUSHCOMMUTATOR
Fig. . Typical SEM photograph of surface of brush trace on n ~iatorof B unit tested at W C . Small amount of wwde deposit can be seen.
Typical SEM photograph of surface of brush trace on com-
Fig. 5.
mutator of A unit tested at room temperature. Powder-Uke deposits similar to those in Fig. 3 can be seen.
Fig. 6. Typical SEM photograph of surface of brush trace on commu- tator of B unit tested at room temperature. Note lesser quantity of deposits contrasted to those m Fig. 5.
figure. Note the very small peak at 84 eV. This peak is due t o silicon LMM transitions which undergo a chemical shift due t o the siloxane bond 141. Fig. 8 is a typical spectrum from a brush trace on a commutator of an A unit. A corresponding SEM photograph is shown in Fig. 3. A clear difference can be seen in the contrast with Fig. 7. The peak height o f the Si KLL transition at 1620 eV is much larger than that in Fig. 7. A higher oxygen concentration is also revealed. Note the
1-
0
400
800 1200 1600 Energy (eV)
Fig. 7.
o n B unit tested at 50°C. There are S, C1, C, 0, and Si contaminants. Sensitivity multipliers are given along with energy. Note small peak at 84 eV.
Typical Auger spectrum from the brush trace on commutator
400
800 1200 Energy (eV)
I600
Typical Auger spectrum from brush trace on commutator in A unit tested at 50°C. Surface has same contaminants as in Fig. 7. Extremely largeSipeakat1620eVisseen.Notethat84eVpeakis larger than that in Fig. 7.
Fig. 8.
284
]EEE TRANMClIUhS Oh COMPOhENTS. HYBRIDS AND MANUFACTURING 1 KHNOLOGY. VOL CHMT-5. '40 2,JUNE 1982
T ABLE I1
SURF ACECONCENTRATIONS OF ELEMENTS ON BRUSH
TRACE ON THE COMMUT A TORS
A B
S.D. 5.0. CX S.D. CX S.D.
A dash in the S.D. column means that the element is observed only inoneof the spectra.
T A B L E I11 SURFACECONCENTRATIONSOFTHEELEMENTSON THE BESlDE TRACE AREA OF THE COMMUTATORS
Temp. I 5D°C i IMOR tenperature SampleABAB
Elementlev) Cx S.D. Cr S.D. Cx S.D. Cx S.D.
TABLE IV
SURFACE CONCENTRATIONS OF THE ELEMENTS ON THE CONTACT AREA OF THE BRUSHES
S 1j2
Pig 357
1 ,016
( ,106
5 ,004 -002 ,016
.DO1 , -006 002
,006 1 ,108 .015
Temp.
sdmple A Element lev) Cx S.D.
Roan temperature
CY S.D. ,021 ,006
,001 ,000 ,353 ,019 .a65 ,016 ,099 .036 .003 -002 .C39 .OD6 .C24 .OD2
,013 / ,095
,014 / .281
5 152 C1 182 C 272 Pb 330 Ay 357 cd 376 0 510 Fe 703 Cu 920
.Old ,003 .004 .004 ,396 ,008 .05C ,017 .079 ,019 ,004 ,001 .05C ,025 ,016 ,001
extremely large peak height at 84 eV. The AES spectra on the brushes and the commutators tested at room temperature showed similar tendencies for motors in A and B groups.
No Si trace was found on the Ag plates which had been attached to the sample holder as references for possible contamhation from handling and the vacuum chamber itself.
Typical results of AES measurements on the samples are tabulared in Table I1 through Table VII. The concentrations of the elements and their standard deviations were calculated from the data on three points analyzed in the same areas. Tables 11 and 111 are typical results on "brush trace" and "beside trace" areas on the commutators in both groups of the motors at both temperatures.
. 3 1 9 - 0 1 3 .C76 , 0 0 0
Tables IV and V list typical results for "contact area" and "beside contact" areas on the brushes in both groups at both temperatures. The notations used above are shown in Fig. 2. In the tables C, is the surface concentration and
,.S.D.is its standard deviation. As a reference the surface concentrations of the elements on the as-assembled motors in both groups are given in Tables VI and VII.
In the analyses, elemental sensitivities were calculated from handbook data [ 5 ] . For Si the sensitivity factor for the KLL transition was used. The calculated values of the
Temp. 1 Sample
50°C
002 .039 .005
(109
.DO7
.OK2
---
.OOi .010
. 0 1 -9
C1 182 C 272 Pd 330 Ag 357 Cd 376 0 510 Fe 703 cu 920 Zn 994 51 1620
001 .a32 .Old
029 001 064
1011 I 023
1 .053
,371
TABLE V
SURFACE CONCENTRATIONSOF THE ELEMENTS ON THE BESIDE CONTACT AREA OF THE BRUSHES
TABLE V! SURFACECONCENTRATIONSON COMMUTATORSOF BOTH TYPES OF MOTORS STORED AT ROOM TEMPERATURE
A
Cr S.D.
--
OUADA AND TODA: SILICON CONlAMLNATlON ON BRUSHCOMMUTA70R
TABLE VII
SURFACE CONCENTRATIONS ON
BRUSHESOF BOTH TYPES OF
Brushes
about the same order as those on the brushes of the as- assembled motors in the B group.
The analyses on the as-assembled motors of both A and B groups indicate that there is no difference in the surface compositions of the elements on both commutators and brushes. Quite good agreement in the as-assembled motors of the groups A and B is seen on both commutators and brushes (see Tables VI and VII).
The analytical results in Table I1 through Table VII are somewhat semiquantitative. There is no good method to obtajn the actual elemental sensitivities of a complex matrix with unknown concentrations of the elements. Therefore, the accuracy of the calculated values for the elemental sensitivities are, unfortunately, unknown. Nevertheless, elemental sensi- tivities calculated from handbook data usually give a good approximation in similar cases. In any case there is a strong contrast between the A and B groups.
The SiLMM transition was observed at 84 eV in all the AES spectra from siliconecontaminated surfaces. The Si on the surfaces is regarded to be in the form of some siloxane
[ 4 ] . Such siloxanes are considered t o be polymerized at the "contact area" of the brushes and to remain on both commutator and brush surfaca. The polymerization of organic vapor on Ag-Pd contacts is a well-known matter. Experiments on A type motors with ventilation holes in the casing to prevent buildup of vapors, did not show the startup trouble experienced here in the 50'C operation [ 6 ] . There- fore we deduce that the siloxanes that reached the surfaces were frictionally polymerized on the "contact area" of the brushes as is seen in the production processes of the so- called brown powders, which result from polymerization of organic vapor by the mechanochemicd reaction at a catalytic metal surface.
The source of the silicone contamination in the A group is clearly the silicone resin used on the armature coils. The silicone resin contains a small fraction of low molecular weight cychc siloxane as unreacted material 171. Therefore it is reasonable to deduce that these cyclic doxanes with low molecular weights are vaporized from the resin even at room temperature. The resin itself has a quite low vapor pressure after it has been well dried [7].We thus consider that the siloxanes evaporated from the resin were polymerized between the electric contact at the brush-commutator inter- face. The polymerized siloxanes deposited onto the brush and the commutator surfaces subsequently formed an in- sulative layer.
Some fraction of the deposits of the polymerized siloxanes might be converted into silicon dioxide by arcing at the edges of the commutator segments. However if a major portion of the contaminants were silicon dioxide, the transition of Si LMM should be at 74 eV IS] instead of at the level 84 eV for silicones (41. Therefore it is clear form Fig. 8 that the AES result strongly indicates that the majority of the sdicones remained as polymers. We conclude that the polymerized siloxane layers caused startup troubles in the A group devices at 50°C.
No problem in the starting properties in the A group at room temperature was observed through 100h of operation.
TABLE Vlli CALCULATEDRELATIVE SENSITIVITIES FOR THE ELEMENTS
-
FROM 151
Transrtion Relative
Element Transitlor.
S LM
Cl LW C KLL Pd .\WH A4 H3H
energy Srniiiivl;y 152 .1014
182 1.137 272 .14i2 330 .81J1 357 1.333
Cd UHN 376 0 KLL 510 Fe LIm 703 Ca LHIl 920
Zn iM 994 51 KLL 1620
.91d6 .9951 .207& .23fi6 .15t3 .03PJ,,
Valuesare normalized to 1.000 by the Ag MNN transition a t 357 eV
elemental sensitivities are listed in Table VIII with notations for the transitions and their energies.
In the case of the A group at 50°C, the commutator surfaces have extremely hlgh Si concentrations on both "brush trace" and "beside trace" areas. However the commutator surfaces of the A group at room temperature showed a &f- ferent character. Their "brush trace" areas have high Si'con- centrations while the "beside trace" areas have nearly equal concentrations compared to those of the commutator surfaces of the asmassembled units of A group.
In the case of the B group, the Si concentratio'ns on the commutators at both temperatures are low and nearly equal to those on the as-assembled units of B group.
The "contact area" of the brushes of the A group at both temperatures have high Si concentrations, while the "beside contact" areas have concentrations nearly equal to those on the brushes in the asassembled units in group A .
There are no significant differences between "contact area" and "beside contact" areas of the group B motors at both temperatures. The concentrations of Si in these samples are
286
IEEE TKANSAf3IONS ON COMPONENTS, HYBRIDS. AND MANUFACTURING TECHNOLOGY. VOL CHhlT-5. NO. L JUNE 1982
The volatility of the siloxanes at room temperature is much less than that at 50°C. Thus it takes longer to form insulative layers at the interface between the brush and the commutator at room temperature. However it is expected that these motors, the A group at room temperature, will eventually become faulty when a thicker layer of the silicones is formed after longer operation.
Silicone contaminations of 5-8 percent were detected on the B group at both temperatures and both types of the as-assembled motors. The source of this silicone contamination was the epoxy adhesive used to attach a small piece of rubber on the brushes behind the brush contact areas. This rubber effectively damped out brush charter. The dampers were glued by the adhesive and backed with a separator paper which had been coated by one of the silicone rubbers. The AES spectra of the glue residuals on the backside of the brushes under the damper showed Si traces with the 84 eV LMM transition. The due itself is considered to have no silicones. It appears likely that the silicones in the separator paper migrated into the glue. The quantities of the silicone in the due seem not enough to cause a serious hazard as in the case of the silicone resin. Actually, the authors did not observe any serious effeci from this silicone source. It is clear that no unit in the 3 group at both temperatures showed the trouble. However it has to be noted that this small quantity of silicone detectably contam- inated the devices even in room temperature storage.
The AES analyses on the sample motors assembled without the damper on the backside of the brushes showed very low Si concentrations close to the detection limit of our AES system (about 0 3 percent for Si in this case).
The C concentrations in the results are widely scattered. This is mainly due to the production of a black powder material near the brush contact areas. Also s o m e of the ana- lytical values for the elements have large standard deviations. In the case of such nonuniform surfaces, it is quite difficult to obtain a result within a small standard deviation. Accord- ingly, the results shown in Tables I1 to VII seem reasonable. Other analytical results on other motors in both groups from the three categories show good agreement w~ththe results described above.
IV. CONCLUSION
The silicone resin used as an insulating material caused a serious problem with the starting properties of small size
dc motors. The motors with the silicone resin at the accel- erated condition showed instabilities in starting properties. The motors tested at room temperature did not show the problem at least during 100 h operation. Nevertheless, the devxces with the silicone resin were heavily contaminated by the silicones. These silicone contaminants are concluded to be produced by the frictional polymerization of the unreacted low molecular weight siloxanes included in the resin as irn- purlties. The siloxanes are polymerized at the Ag-Pd brush surface and remained on the commutator and brush surfaces.
A very small quantity of the silicones can easily cause the contamination seen in the room temperaturestored as-assem- bled units. Therefore one must be careful if there is a possib~l- ity that some electrical part encounters silicone products. The insulative layers formed on the contact surfaces may cause unreliable operation not only on relay contacts, but also in motor brush-commutator contacts as described in this paper.
It is important to make a'survey of the materials used for the parts, tools, processes, and the environment of pro- duction processes. The reliability of the motors or the electri- cal contacts can be strongly affected by careless use of silicone materials.
ACKNOWLEDGMENT
The authors wish to express their appreciation to E. 0. Johnson, 'RCA Research Laboratories, Tokyo, for his encour- agement through the work, and for editing of the manuscript.
REFERENCES
N . M. Kitchen
contacts-effects,
Purrs, Hybrids, Packaging, vol. PHP-12, no. I. pp. 24-28. Mar.
and C . A . Russel. ..Silicone oils on electrical sources, and countermeasures." IEEE Trans.
1976.
G.J. WitterandR. A. Le~pr,"A comparicnnfortheeffectsof various funns ul sillcon contaminatton on contact prrI"orrnance," IEEE Trans. Components. Hybrih, Monufacr. Technol.. vol. CHMT-2. no. I , pp. 56-41; Mnr. 1979.
C. A . Hague and A . K. Sptegler. "lnvesl~gationo l silicone oil contamination on relay contacrs using Auger electron rpectros- copy ." Applicurions Surface Sci.. (41, pp. 2 14-220. 1980.
Ci. W. Stuptan. "Auger specfroscopy of silicones." J. Appl. Physics, vol. 45, no. 12. pp. 5278-5282, Dec. 1974.
L.
E. Davis. N.C. MacDonald. P. W. Palmberg. G . E. R~acha.nd R. E. Weber, Handbook of Auger Electron Spectroscopy. Eden Prairie. MN. Physical Electronics Industries. USA. 1976. 2nd ed S . Osaka and M. Toda. private communlcatlon.
J. Koizuml. private communication. 

No comments:

Post a Comment