ARC DISCHARGE AND EROSION BEHAVIOR OF SILVER ELECTRIC CONTACTS BETWEEN STATIC GAP

ARC DISCHARGE AND EROSION BEHAVIOR OF SILVER ELECTRIC CONTACTS BETWEEN STATIC GAP Chung, H-H. 1 Lee, R-T. 2. Chiou, Y-C. 2 1 Deprtment of Automtion Engineering Ko-Yun Institute of Technology 2 Deprtment of Mechnicl Engineering Ntionl Sun Yt-Sen University Abstrct The effects of supply voltge nd gp distnce on the rc erosion behvior of silver electricl contct by using sttic-gp electricl erosion tester with single rc dischrge. Results show tht the minimum electricl field strength t rc initition occurs t 7V/ µ m up to 175V, t 1 µ m t 33V, nd 4 µ m t 5V. Since the gp distnce becomes bigger during the rc durtion, the residul voltge is lwys positive t smller gp distnce with lower supply voltge. The negtive residul voltge occurs t lrger supply voltge due to the effect of metllic phse rc. This negtive residul voltge is cused by positive metllic ions ccumulted on the cthode. At the supply voltge less thn 2 V, the erosion re is linerly incresed with incresing gp distnce, but this reltionship becomes prbolic t the supply voltge lrger thn 2 V. At gp distnce below 5 µ m, the erosion re increses with incresing rc energy, but it decreses with incresing rc energy t gp distnce beyond 5 µ m due to the increse mount of gseous phse rc. It is seen from the eroded node surfce tht the metllic phse rc minly cuses the crter with mny strip-like metllic prticles, but the gs phse rc significntly results in mny silver powders. These results with negtive residul voltge re in good greement with the prticle sputtering-depositing model. Keyword: Arc dischrge, Gp distnce, Erosion re, Residul voltge, Arc energy, Prticle sputtering-depositing 1. INTRODUCTION The electricl contcts re widely used in control nd communiction systems. The electricl erosion of contcts due to rcing hs lwys been n importnt problem in the design of the switching system. This erosion occurs whenever contcts re closed to complete the energized circuit, or when they re pulled prt to brek it. The formtion of pips nd crters hs been recognized by Homes nd Slde [1] on the silver nd its lloy contcts. Mny other experiments hve lso investigted the erosion nd contct resistnce vrition of silver nd its lloy contcts [2-6]. In dynmic test of switching contct, the contct erosion is the result of melting nd vporizing metl when the contcts re brought together. Two types of rcs (i.e. the metllic nd gseous phses) hve been recognized by Germer [2]. They significntly influence the erosion behvior nd mss trnsfer. Generlly, if the metllic phse rc is dominnt, the mteril trnsfers from the node to the cthode, while if the gseous phse rc becomes predominnt, the mteril trnsfers from the cthode to the node. Recently, ccording to the trnsition of contct morphology, mss trnsfer, nd rc durtion of the metllic nd gseous phses, prticle sputtering nd deposition model hs been proposed by Chen nd Sw [6]. In this model, the evportion nd sputtering of the contcts were discussed to explin this mss trnsfer during rcing. It hs been known tht the rc cuses the erosion, but the surfce morphology lso influences the rc behvior during the dynmic test of switching contct. Moreover, mechnicl stresses, molten bridges, nd chunk trnsfer due to welding lso result in the erosion. Consequently, it is very difficult to develop quntittive understnding in the erosion behvior of mteril relted to the rcing condition lone. To void the influence of the complex mechnics of prticulr switching device, sttic-gp experiments hve been conducted in recent yer [7-11]. In this pper, the sttic-gp experiments with single rc dischrge re conducted to reduce the complexity due to numerous rcs striking. Furthermore, the supply voltge, the rc durtion, nd the gp distnce in this sttic gp experiment cn be precisely controlled. The effects of supply voltge nd gp distnce on the rc erosion behvior of silver contct re investigted. The minimum electricl field strength t rc initition is estblished. Using the prticle sputtering nd deposition model discuss the residul voltge nd the erosion pttern of the node surfce. 2.EXPERIMENTAL APPARATUS nd PROCDURES 2.1 Experimentl pprtus The experiments re conducted on horizontl sttic-gp nd single-rc dischrge tester with mesuring system shown in Fig. 1. In this tester, two micrometer heds of non-rotting spindle type re employed. In ech micrometer hed, one revolution of the thimble moves the spindle the distnce of.5 mm. In order to djust the gp distnce in the grdution of.25 µm, 1:5 worm ger reducer is employed to drive the micrometer hed. A stepping motor with the resolution of.9 o /step drives the worm nd PC progrm cn control the stepping motor. A disk specimen is ttched to the spindle of 1st micrometer hed. Hence, it cn be moved right nd left through the spindle of 1st micrometer hed. To eliminte the effect of the bcklsh of micrometer hed on the precision of the micro-movement of the disk specimen, 2nd micrometer hed is used to clibrte the micro-movement of the disk 1

specimen through the reference plte ttched to the spindle of the 1st micrometer, s shown in Fig.1. A crtridge hed probe mounted on 2nd micrometer hed is connected to n electricl comprtor or differentil type nlog mu-checker with grdution of.1 µm. Hence, the micro-level shift of the disk specimen through this device cn be clibrted by the nlog mu-checker. The sequence of opertion for this device is described in section 2.3 experimentl procedure. Electricl Test specimens comprtor Mircometer hed, 1st Mircometer hed, 2nd DC power supply Chrging Circuit Vrible resistor.1μω C.94µF Switches Dischrging Circuit Anode: Ag Cthode: Fe (Needle) 1Ω Ditigl Ocilloscope CH1 CH2 Fig. 2 Schemtic digrm of DC R-C circuit with chrging nd dischrging brnches. G Worm ger reducer 1.8 3. 1. Stnd Isoltion + Reference plte Stepping motor φ.6 2 φ5. 2.5 φ Personl computer Dt cquisition system Oscilloscope Personl computer Fig. 1 Schemtic digrm of horizontl sttic-gp nd single-rc dischrge tester with mesuring system. Fig. 2 shows DC R-C circuit with chrging nd dischrging brnches. The chrging brnch of the circuit includes high voltge power supply (1 6 V DC), vrible resistor (.1 MΩ), nd cpcitor bnk (.94 µf). The dischrging brnch consists of the cpcitor bnk, ppropritely spced electrodes, nd 1 Ω resistor. The vrible resistor is used to djust the chrging time, nd the 1 Ω resistor is used to mesure the rc current. When the cpcitor bnk is chrged by the power supply to certin vlue of voltge, the chrging circuit is mnully switched to the dischrging circuit. A digitl oscilloscope HP 54645A records the voltge nd current cross the electrode pir during rcing. Since it hs two chnnels, smpling ll chnnels simultneously t mximum rte of 2 MS/s with 1 M smples of memory depth per chnnel, the vrition of the voltge nd the current cross the electrode pir cn be observed with high precision. 2.2 Test specimens The contcts re mounted horizontlly nd isolted from the body of the tester, with the cthode t the left nd the node t the right, s shown in Fig. 1. The cthode is n iron needle to confine the erosion of nodes to smll designed re. The node specimen is mde of pure silver. The size nd the shpe of the electrode specimens re shown in Fig. 3. All experiments re conducted in common lbortory tmosphere. Before ech experiment, the node specimen is polished with grde 4, 1, 15, nd 2 emery ppers in order, nd then with lumin slurry, so tht its surfce roughness R is bout.1 µm or R mx of.5 µm. Rdius=.5 Cthode: Fe (Needle) () Anode: Ag Fig. 3 The size nd shpe of the contcts () cthode specimen, (b) node specimen. 2.3 Experimentl procedure Before ech test, the electrodes re clened with cetone in ultrsonic clener. The node is fixed t the jig of the 1st micrometer hed, nd the cthode is fixed t the stnd through n insultor. When the node comes into contct with the cthode slightly with the grdution of.25 µm, digitl multi-meter lso monitors the contct condition between the electrodes. Generlly, when the node do not contct the cthode, the contct resistnce chieves infinity. Hence, if the contct resistnce is in the rnge of 1 to 3Ω, then the gp distnce between electrodes is ssumed to be zero. Moreover, to eliminte the effect of the bcklsh of micrometer hed on the precision of the gp distnce between the electrodes, n nlog mu-checker with grdution of.1 µm is employed to clibrte it. When the gp distnce hs been djusted to certin vlue, the supply voltge is preset from 4 V to 5V through DC power supply. The test prmeters used re.1-4 µm gp distnce t increments of.1 µm. When the gp distnce is djusted to certin vlue, turn the switch on to supply certin vlue of voltge. During the test, the digitl oscilloscope nd the dt cquisition system record the wveforms for the interfce voltge nd current between electrodes. The eroded re cn be clculted by using powerful softwre progrm on the photomicrogrph of the scnning electronic microscope. The experiments re limited to contcts in tmospheric ir. The verge room temperture for the test is 25 o C nd the verge reltive humidity is 85%. 3.EXPERIMENTAL RESULTS AND DISCUSSION 3.1 Arc dischrge digrm In this experiment, digitizing oscilloscope records the voltge nd the current cross the electrode pir. Since it hs two chnnels, smpling ll chnnels simultneously t mximum rte of 2 MS/s with 1 M smples of memory depth per chnnel, the vrition of the voltge nd Ag Cu (b) 2

the current cross the electrode pir cn be observed in detil. Fig. 4 shows the typicl wveforms of voltge nd current during rcing cross the electrode pir. It hs been known tht the short rc occurs when the gp distnce between the electrodes is in the sub-micron rnge in low voltge system. Interfce voltge, V (V) Interfce voltge, V (V) 4 3 2 1-1 -2-3 -4 4 3 2 1-1 -2-3 -4 V b V I p -2 2 4 6 8 Time ( µ sec) V b V Ip -2 2 4 6 8 Time ( µ sec) Fig. 4 Typicl voltge nd current wveforms of vrious rc dischrges. () Supply voltge of 4 V nd gp distnce of.2 µm. (b) Supply voltge of 5 V nd gp distnce of.2 µm. The electric field t rc initition is of the order of 1 6 V/cm. Figure 4() shows the voltge nd current wveforms cross the electrode pir t the gp distnce of.2 µm nd the supply voltge of 4 V. The electric field strength in this cse is 6 2 1 V / cm. This vlue is enough to initite n rc. In this experiment, the cpcitor bnk is chrged by the power supply to certin vlue of voltge, nd then it dischrges in fixed spced electrodes. Since the electric field strength is enough to initite the rc, the current strts from zero nd increses to pek vlue I p of 7 mperes during the single-rc dischrge period. During this period, the voltge cross the electrode pir decreses from the brekdown voltge V b (bout 2 V) to the residul voltge V r until the end of rcing. This residul voltge results from the residul electricl chrge in the cpcitor bnk when the rc is expired. The rc durtion in this experiment is bout 5.8 µs. It hs been found by Boyle nd Germer [12] tht the node crter grows very slowly for the rc durtion below the microsecond rnge. Hence, in this study, the cpcitnce in the cpcitor bnk is given constnt of.94 µf. This cse is typicl result I I V r V r 3 2 1 5 4 3 2 1 Interfce current, I (A) Interfce current, I (A) for smll gp distnce nd low supply voltge. Contrry to this cse, Fig. 4(b) shows similr result for bigger gp distnce nd higher supply voltge. It is seen from this figure tht the voltge cross the electrode pir suddenly drops from the brekdown voltge V b (bout 26 V) to negtive vlue of residul voltge V r until the end of rcing. It is seen from Figs. 4() nd (b) tht this residul voltge cn be positive or negtive vlue. Hence, the residul voltge is influenced not only by the cpcitor bnk, but lso by the test conditions. Supply voltge, V (V) s 5 4 3 2 175 1 Arcing Slope = 7 V/ µ m V =12 ln[d] + 62 s No-Arcing 2.5 1 2 3 4 Gp distnce, d ( µ m) Fig. 5 Boundry between the rcing nd no-rcing regions in terms of the supply voltge nd the gp distnce. Since the ir between n electrode pir possesses high resistnce t certin gp distnce, the current does not strt to flow until the brekdown voltge is chieved. Hence, t certin supply voltge nd certin gp distnce, the rcing effect on the electrodes cn be esily distinguished by observing the voltge nd current wveforms cross the electrode pir. Moreover, the scnning electron photomicrogrphs cn lso be used to certify the eroded surfce. Fig. 5 shows the boundry between the rcing nd no-rcing regions under wide rnge of supply voltge nd gp distnce. It is seen from this rc dischrge digrm tht s the gp distnce is less thn 2.5 µm, the criticl supply voltge between the rcing nd no-rcing regions is linerly proportionl to the gp distnce with the slope of 7 V/µm. It ws indicted by Germer [2] tht the slope is of order of 5 V/µm for clen or inctive surfce, nd it is of the order of 6 V/µm for ctive surfce. Hence, t gp distnce less thn 2.5 µm, the slope or the minimum electricl field strength t rc initition obtined in this study is in the rnge between ctive nd inctive surfces obtined by Germer [2]. Contrrily, when the gp distnce is greter thn 2.5 µm, the criticl supply voltge is grdully incresed with incresing gp distnce. This criticl supply voltge, V s (V), for the gp distnce greter thn 2.5 µm expressed in term of gp distnce d(µm) is given by V s = 12ln[d] + 62. Hence, it is seen from Fig. 5 tht t the smll gp distnce, the rcing phenomenon is dominted by the gp distnce, 3

while it is significntly influenced by the supply voltge t the bigger gp distnce. The ir brekdown occurs t 7 V/µm up to 175 V,t 1 µm t 33 V, nd 4 µm t 5 V. -4-3 5V 3.2 Arc dischrge behvior on the interfce Fig. 6 shows the effects of supply voltge nd gp distnce on the residul voltge. It is seen from this figure tht the residul voltge is zero for the supply voltge close to 2 V. When the supply voltge is less thn 2 V, the residul voltge keeps positive vlue, nd it decreses with incresing supply voltge nd gp distnce. When the supply voltge is greter thn 2 V, the residul voltge becomes negtive nd its bsolute vlue decreses with incresing supply voltge nd gp distnce. Hence, the residul voltge is influenced not only by the cpcitor bnk, but lso by the rc dischrge behvior. When the residul voltge is zero, it implies tht ll the energy in the cpcitor bnk is relesed into the rc energy. As the residul voltge is positive, it implies tht prt of energy in the cpcitor bnk is relesed into the rc energy. These cses occur t lower supply voltge (below 2 V) nd smll gp distnce (below 2.5 µm). Generlly, the gp distnce becomes bigger during the rc durtion. As result, the energy in the cpcitor bnk cn not completely be relesed into the rc energy due to the bigger gp distnce during the rcing. Hence, the residul voltge is lwys positive t smller gp distnce nd lower supply voltge. As the residul voltge is negtive, it implies tht not only ll the energy in the cpcitor bnk is relesed into the rc energy, but lso the interfce between the electrodes produces the opposite voltge. This phenomenon occurs t high supply voltge (lrger thn 2 V), nd it cn be explined by the prticle sputtering-depositing model [6]. It is seen from this model tht the negtive residul voltge results from the metllic phse rc. In the metllic phse, electrons emit from the cthode to the node due to the electricl field. During the process, the metl vpor is ionized by electron bombrdment, resulting in the formtion of metllic ions nd new electrons. The metllic ions impct in the cthode, nd then deposit on the cthode due to the electricl field nd the higher cohesive bility between the ionized metl vpor prticles nd the cthode. As result, lot of positive metllic ions ccumulte on the cthode. Furthermore, ll the energy in the cpcitor bnk hs been relesed into the rc energy. Consequently, the residul voltge becomes negtive due to metllic ions ccumulted on the cthode. At the sme time, this voltge is not enough to initite the rc due to the bigger gp. Residul voltge, V (V) r -2-1 1 3V 25V 2V 16V 1V 6V 4V 4V 2 1 2 3 4 µ Gp distnce, d ( m) Fig. 6 Effects of supply voltge nd gp distnce on residul voltge. Arc energy, E ( mj ) 3 25 2 15 1 3V 4V 5V 25V 5 2V 16V 1V 4V 1 2 3 4 Gp distnce, d ( µ m) Fig. 7 Effects of supply voltge nd gp distnce on rc dischrge energy. Fig. 7 shows the effects of supply voltge nd gp distnce on the rc energy E. In this pper, the rc energy is clculted by using the following form E = t V(t)I(t) dt. Here V nd I re the voltge nd the current cross the electrode pir, respectively, nd t is rc durtion. It is seen from Fig. 7 tht the rc energy increses with incresing supply voltge, but it vries in errtic fshion with incresing gp distnce. This implies tht the higher the supply voltge, the higher the pek current or the energy of the cpcitor bnk, thus resulting in the increse in rc energy. Moreover, when the gp distnce increses, the metl vpor density is so low tht the rc trnsfers from the metllic phse to the gseous phse, thus resulting in the rc energy vried in errtic fshion. 3.3 Erosion re on the surfce of silver electrode It is very difficult to mesure precisely the vrition of gp distnce, pip re on the iron needle, nd the crter depth on the cthode fter ech test. Moreover, it hs been found by Wng et l. [1] tht the rtio of the crter dimeter to the crter depth is greter thn 1 under the 4

supply voltge of 1 V. This indictes tht the erosion crters re shllow. Hence, in this pper, only the crter re on the eroded node is clculted by using powerful softwre progrm. Erosion Are, A ( 1 µ m 2 ) 3 e 18 16 14 12 1 8 6 4 2 25V 3V 2V 1V 6V 4V 4V 5V 1 2 3 4 Gp distnce, d ( µ m) Fig. 8 Effects of supply voltge nd gp distnce on erosion re. Figure 8 shows the reltionship between the erosion re nd the gp distnce t vrious supply voltges. It is seen from this figure tht t the supply voltge less thn 2 V, the erosion re is linerly incresed with incresing the gp distnce, but the reltionship between the erosion re nd the gp distnce becomes prbolic t the supply voltge lrger thn 2 V. It is found from Figs. 7 nd 8 tht t smller gp distnce (below 5 µm) the erosion re increses with incresing rc energy, but it decreses with incresing rc energy t bigger gp distnce (beyond 5 µm). This phenomenon cn be resonbly explined by the mount of metllic nd gseous phse rcs, nd the morphology of the eroded surfce. 4.CONCLUSIONS The effects of supply voltge nd gp distnce on the rc erosion behvior of silver electricl contct by using sttic-gp electricl erosion tester with single rc dischrge. From the experimentl results, the following conclusions cn be drwn. (1) The boundry between the rcing nd no-rcing regions hs been obtined t supply voltge (up to 5 V) nd gp distnce (up to 4 µm with the increment of.1 µm). The minimum supply voltge, V s (V), t rc initition expressed in term of gp distnce d(µm) is given by V s =7d for d 2.5µ m V s =12ln[d]+62 for 4 µ m d 2.5µ m. (2) The residul voltges is lwys positive t smller gp distnce nd lower supply voltge becuse the energy in the cpcitor bnk cn not completely be relesed into the rc energy. The negtive residul voltge occurs t supply voltge lrger thn 2 V due to the ction of the metllic phse rc. This negtive residul voltge is cused by positive metllic ions ccumulted on the cthode, it cn be explined by the prticle sputtering-depositing model [11]. (3) At the supply voltge less thn 2 V, the erosion re is linerly incresed with incresing the gp distnce, but the reltionship between the erosion re nd the gp distnce becomes prbolic t the supply voltge lrger thn 2 V. At gp distnce below 5 µm, the erosion re increses with incresing rc energy, but it decreses with incresing rc energy t gp distnce beyond 5 µm due to the increse in the mount of gseous phse rc. References [1] P. G. Slde nd F. A. Holmes, Pip nd crter formtion during interruption of lternting current circuits, in Proc. 8 th int. Conf. Electric contct Phenomen, Tokyo, (1976) 29-214. [2] L. H. Germer, Physicl processes in contct erosion, J. Appl. Phys., 29 (1958) 167-182. [3] F. A. Holmes nd P. G. Slde, The erosion chrcteristics of Ag contcts nd the effect of dding smll percentge of W, IEEE Trns. Components, Hybrids, nd Mnufct. Technol., PHP-13 (1977) 23-3. [4] H. Sone nd T. Tkgi, Role of the metllic phse rc dischrge on rc erosion in Ag contcts, IEEE Trns. Components, Hybrids, nd Mnufct. Technol., 13 (199), 13-19. [5] Z. K. Chen nd K. Sw, Chrcteristics of Ag contct morphology in breking rcs, Wer 199 (1996) 237-244. [6] Z. K. Chen nd K. Sw, Prticle sputtering nd deposition mechnism for mteril trnsfer in breking rcs, J. Appl. Phys. 76 (1994) 3326-3331. [7] H. J. Kim, J. C. Gustfson, nd P. C. Wingert, Improvement of the rc erosion resistnceof Ag-CdO mteril with mtrix-strengthing dditive, in Proc. 11 th Int. Conf. on Electircl Contct Phenomen, (1982) 212-216. [8] P. C. Wingert nd H. J. Kim, Development of the rc-induced erosion surfce in silver-cdmium oxide, IEEE Trns. Components, Hybrids, nd Mnufct. Technol., vol. CHMT-8, (1985) 64-69. [9] B. J. Wng, N. Sk nd E. Rbinowicz, Sttic gp erosion of Ag-CdO electrodes, IEEE Trns. Components, Hybrids, nd Mnufct. Technol., 14 (2) (1991), 374-385. [1] B. J. Wng, N. Sk nd E. Rbinowicz, Sttic gp, single-sprk erosion of Ag-CdO nd pure metl electrodes, Wer, 157(1992), 31-49. 5