Rare Metals 2010,29(03),248-254

Arc erosion of AgSnO₂ electrical contacts at different stages of a break operation

J. Swingler A. Sumption

School of Engineering Sciences,University of Southampton

作者简介：J. Swingler E-mail:swingler@soton.ac.uk;

收稿日期：31 July 2009

基金：Barnbrook Systems, UK for their support of this work;

Arc erosion of AgSnO₂ electrical contacts at different stages of a break operation

Abstract：

Arc erosion studies are conducted on AgSnO2 contact materials at different stages in the break operation.A resistive load arrangement is used with up to 42 V DC at 24 A and a constant contact opening velocity.The arc current is terminated at different stages as the arc is drawn between the contacts enabling a study of the arcing phenomena up to that point.Surface profiling of the contacts is conducted to determine the extent of erosion at the different stages as the arc is drawn.Spectral analysis is also conducted on the arc and then related to the extent of erosion.The results show that particular features occur at different stages as the arc is drawn.As the arc is initially established,it goes through an “Arc Generation” regime where the arc roots are small and immobile on both the anode and the cathode.Material transfer occurs mainly from anode to cathode.The spectral analysis indicates that Sn and O species dominate the arc followed by the Ag species.As the arc is drawn further and enters the “Arc Degeneration” regime,the anode undergoes significantly larger erosion than the cathode.Also,both contacts indicate that multiple arc roots have formed,which are highly mobile in the later stages of the discharge.The spectral analysis indicates that Ag and N species are in high concentrations compared to other species.The mechanisms of erosion and deposition are discussed in relation to the species within the arc discharge.For the complete break operation,it is found that the anode undergoes major erosion,and it is thought that the gaseous ions species do not dominate the arc under these conditions of short arcs and voltage <42 V to cause cathode erosion.

Keyword：

contact materials;electrical contacts;arc erosion;3D surface profiles;spectral analysis;

Received： 31 July 2009

1. Introduction

Electrical contact materials used in switching devices are commonly made from silver alloys because of their excel lent properties in minimizing contact welding and arc ero sion during operation.Silver tin oxide(Ag Sn O₂)is an alloy that has been the focus of research for many years,as par ticular forms can be optimized to achieve the very best re sults for particular applications[1-4].Opening or breaking a contact pair in a low voltage switching system usually re sults in an arc being drawn between the contact pair.In the automotive application using the standard 12-14 V supply arc erosion is expected to have an insignificant effect com pared to the higher voltage systems of 36-42 V.This is be cause for materials like Ag Sn O₂,a minimum arc voltage o~12 V is required across the contact before the arc is formed[5].At these higher voltages,excessive arcing occurs,which leads to more arc erosion.

The interaction between the arc and the contact pair is a complex phenomenon involving several mechanisms of material erosion and deposition.The initial stage of drawing the arc has been described as a“metallic ion stage”and results in material being transferred from the anode to the cathode[1].It is thought that the Ag⁺and Sn⁺(M⁺)ions are attracted to the cathode and deposit there with a net flow of material from the anode to the cathode.The second stage of drawing the arc has been described as the“gaseous ion stage”and results in material transfer from the cathode to the anode.As the arc lengthens,gas molecules have a higher probability of entering the discharge,which are then ionized to g⁺.These gaseous ions are attracted to the cathode,sputter it,and then cause erosion of the cathode surface resulting in a net transfer of material from the cathode to the anode.

Often,studies into the behavior of these materials at different voltages have involved a comparison of the level of erosion on the contact surfaces[6-11].This approach describes the end result of an opening(break)operation that gives different erosion profiles depending upon materials opening characteristics,and circuit parameters.Studies at 64V,0.1 ms^-1 opening velocity between 1 and 16 A show that the cathode always lose mass[1].Others have found similar results under particular conditions[10,12-13].Whereas studies at<42 V show that the anodes have a net loss in mass[14-15].All these demonstrate that arc erosion is a complex phenomenon and needs to be investigated at every stage of the arc development as the contacts open to better understand the mechanisms involved.

The study presented here investigates the progression of erosion as the contacts open as the arc is drawn.The technique was first published in Ref.[16]and is designed to identify which erosion mechanism contributes at what particular point during the opening operation.The technique uses the arc voltage as the trigger for termination of the discharge at particular points.Previous studies by others have used both contact gap and arcing time as the parameters for investigation[7-9].However,the physical processes that dominate the arc are closely related to the voltage across the discharge.The independent parameters that govern the arc discharge are the contact velocity and the supply current,assuming an atmospheric arc with conventional contact materials.In many studies on these phenomena,the contact velocity is a nonlinear function with time,as the contacts accelerate apart from zero to a maximum contact gap.In the study presented here,as with other studies[1],the velocity is controlled such that the characteristic displacement is a linear function with time.Even with constant velocity,the relationship to the arcing process will be related to the absolute value of velocity;hence,with a slow opening operation,the arc physics will differ from that of a fast opening operation at a given arcing time.

The identification of arc voltage as the main parameter for these experiments is preferred as it will generate data that is independent of both the supply current and the contact velocity.This enables detailed observations of the physical processes underlying the arcing process.

2. Experimental method

2.1. Experimental arrangement

The experimental apparatus was designed to controllably open and close a pair of contacts under particular conditions as previously reported[1,16-17].A main feature of the apparatus was to trigger the termination of the load current at particular points during the opening(break)operation.Additionally,a PC-based spectral analysis system(supplied by Ocean Optics)was used to monitor the arc discharge formed between the contact pair.

Fig.1 is a schematic diagram of the break apparatus and the arrangement used.Three standard automotive 12 V lead-acid batteries were used for the load current that was connected in series with the load resistance,the“load current control circuit,”and the contacts under test.The load resistance was noninductive and arranged to allow 24 A current loading when the contacts were closed.The“load current control circuit”was used to terminate the current at a particular arc voltage as the contacts opened.The contacts opened at a constant velocity of 0.050±0.001 m/s and were controlled by the“opening control mechanism.”

Fig.1.Schematic diagram of experimental arrangement.

A constant opening velocity was used(even though actual relay devices open with variable velocity)to keep the system variables to a minimum.The velocity chosen was considerably slower than relay devices(typically>0.3 ms^-1)but this was required to achieve longer arcs at this power level for higher light intensity for spectral analysis.

The spectrometer was triggered at a particular point in time to capture data as the arc was drawn between the contact pair.The data was captured in a 0.8 ms window up until the current was terminated.The data consisted of the intensity of light produced by the arc for wavelengths between200-800 nm.

The load current was initiated 35±5 ms before the contacts were opened.The spectrometer was triggered to capture data 0.8 ms before the current was terminated.The termination of the current was triggered by the arc voltage attaining a particular value,known as the“trigger voltage level.”The arc voltage increased as the contact pair was opened,as shown in Fig.2.

The silver tin oxides,contact samples investigated were of a rivet geometry with a hemisphere contact surface of 10mm in radius.These were manufactured by an extrusion process with a medium oxide grain size.The composition was 8 wt.%tin oxide with 92 wt.%silver.

Fig.2.Voltage and power profiles for a break operation.

2.2.3 D surface profiling

The surface erosion analysis uses a volume change method by measuring the samples before and after testing[18-19].The evaluation of volume changes has been demonstrated as advantageous since accurate mass change measurement is not possible where contacts are housed in relays and permanently fixed to components of the relay.Volume measurement techniques from 3D profiling also have the advantage of determining deposition as well as material removal.

The tested contact samples were measured using a high precision laser measurement system[11]to provide 3D surface profiles.These profiles were further processed using the software package BODDIES[20]to give volumetric analysis.Deposition volumes above a datum surface and volumes removed below this datum were evaluated.

In a review of the subject[21],the noncontact method is shown to have been developed further for increased scanning speed and for increased accuracy.It is also shown that the method could be used to measure low erosion levels.

3. Results

Three sets of results are collected,which include(a)voltage/current characteristics as the arc is drawn,(b)3D surface profiling of the eroded contact from which volume analyses and erosion surface diameters are made,and(c)spectral data from the arc to identify ionic species dominant within the discharge.

3.1. Voltage and power characteristics

Fig.2 shows the typical voltage and power characteristics of the drawn arc.The power characteristic indicates a maximum power dissipation at~2.7 ms into the discharge(voltage trigger level of 20 V).By drawing the arc between the contact pair,the arc appears to progress through three regimes.First,an arc generation regime where the power dissipation increases to a maximum,followed secondly by an arc degeneration regime where the power dissipating reduces until the arc enters a regime where it extinguishes.

3.2.3 D contact surface profiles during break erosion

Three-dimensional surface profiles of contact pairs tha have the current terminated at particular voltage levels on opening are shown in Fig.3.Three example pairs of contacts are shown for different trigger voltages.The top view of the contacts is shown to scale with each contact being 3mm in diameter.Each contact pair has undergone 100 break operations.Eroded areas can be clearly seen,which increase in diameter with higher trigger voltages.

Fig.3.Surface profile images of Ag Sn O2 contacts.

Fig.4 shows the diameter of the affected area of all contact surfaces investigated.This data is determined from the3D profiles as exemplified in Fig.3.It can be seen that there is an initial increase in the diameter of the eroded areas on both the anode and cathode.After the initial increase,the affected areas remain approximately the same irrespective of the trigger voltage.The anode has a larger eroded area compared to the cathode.

Fig.4.Diameter of eroded area on contacts.

Volume analysis of the contact pairs are given in Fig.5.This shows that as the arc is allowed to extend,there is an increase in erosive activity on the anode.At the trigger point of 12 V(~0.1 ms into the arcing processes,see Fig.2),0.8×10^-3 mm³ of material has been removed,whereas at the trigger point of 38 V(~8 ms into the arcing process),3.5×10^-3 mm³ of material has been removed.The cathode undergoes deposition that,after an initial high deposition rate(2×10^-3 mm³ of material at a trigger point of 13 V),remains constant throughout the arcing process thereafter with a deposition of 0.5×10^-3 mm³.A constant deposition and erosion seem to be indicated between 19 and 22 V,whereas the arc duration between voltage trigger levels is not constant.Between voltage trigger levels of 12-18 V,the arc duration is~2.72 ms,of 19-22 V the arc duration is~0.98 ms,and of 24-38 V the arc duration is~4.27 ms.

Fig.5.Net volume change on contacts.

3.3. Spectral analysis of the arc discharge

Fig.6 shows examples of the spectra acquired during the opening operation.The plots are the intensity of the arc emission against the wavelength between 179 and 861 nm.The spectral lines plotted are identified as species that are likely to be in the arc discharge.A datum line is chosen which appears in all spectra to enable a comparison of the relative levels of the other spectral lines.This is thought to be the Ag(I)line at 224.6 nm.It is found that there are three distinct groups of species that can be identified from the spectral lines.These groups of species are found to be related to the arc generation,arc degeneration,and arc extinction regimes as the arc is drawn.The first group of spectral lines are thought to be mainly generated by Ag(II)and some Sn(I)species.These are produced in the initial stage of drawing the arc between the contacts.The second group is thought to be attributed to Ag(I)species.These species are indicated throughout the arcing process.The third group is thought to be gaseous species,such as N(II),which are indicated toward the final stages of the arc.

Fig.6.Arc spectra during break operations.

All the spectral data acquired for different trigger voltages have been analysed relative to the datum spectral line Fig.7 shows this analysis with two graphs relating to the Ag(I)and Ag(II)species and the other graph relating to the Sn(I)and Sn(II)species.Each graph is a relative intensity of emission relative to the datum spectra line plotted against the trigger voltage that is equivalent to the arc duration before the current is terminated.The graph of the Ag species shows that the Ag(I)relative intensity remains constant throughout the arc discharge.This is not surprising as it is compared to the 224.6 nm spectral line of the same species(the datum line).The Ag(II)relative intensity,however,increases toward the end of the arc discharge,indicating an increase in the ratio of Ag(II)to Ag(I)species as the arc is lengthened.The Sn(I)and Sn(II)species have high intensities at the beginning and the end of the arcing process.

Fig.7.Relative spectra intensity of metallic ion species.

Fig.8 shows two graphs relating to the O(I)and O(II)species and the N(II)species.The O(I)and O(II)species have a high relative intensity at the beginning of the arc discharge,only indicating a large presence of species up to a trigger voltage of 15 V.After a trigger voltage of 15V,there is little presence of these species.The N(II)spices have a high relative intensity at the end of the arc discharge process only.However,the N(II)species can be seen to be present after a trigger voltage of 15 V.

4. Discussion

Fig.9 shows generalized profiles of the power and voltage over an opening operation taken from Fig.2.The power profile has been pided into three regimes corresponding to“arc generation,”“arc degeneration,”and“arc extinction.”In the“arc generation”regime,the power dissipation generally increases as the arc is drawn.This shows that the conductivity of the arc increases during this regime as apposed to the arc having constant conductivity that would result in increased resistance during lengthening.Sufficient conducting species are being formed to develop(or generate)the arc.In the“arc degeneration”regime,there is an insufficient amount of conducting species being formed to maintain an increasing or constant level of conductivity.Thus,the power dissipation lowers as the contact gap lengthens.In the“arc extinction”regime,the conductivity decreases at an accelerating rate as conductive species are lost.Distinct stages can be identified from the experimental data within the three arc regimes,which are illustrated with contact pairs in Fig.9.

Fig.8.Relative spectra intensity of gaseous ion species.

Fig.9.Regimes of the arc discharge during a break opera-tion.

4.1.“Arc generation”regime

Stage 1.Bridge rupture and metallic ion initialization The rupture of the molten metal bridge of the separating contact pair is indicated by the sudden rise in voltage as the arc discharge is initiated(see Figs.2 and 9).Material removed from both anode and cathode to form the bridge is then used in the metallic ion arc[22]to for conductive species.The contact surface profiles at~0.1 ms into the arcing process(where the trigger voltage level is 12 V)indicate that material is removed from the anode and deposited on the cathode once the bridge is ruptured.This is consistent with the model proposed by Chen et al.[10]and previous studies[1].

Stage 2.Growth of the metallic ion arc.The growth of the metallic ion arc is indicated on the voltage characteristic with an initial drop in arc voltage as the arc is established,followed by an increase in voltage as the arc is drawn.This is shown on Figs.2 and 9 during the first 0.4 ms of the arcing process.Increased anode erosion and increasing cathode deposition are seen in Fig.4 up to a trigger voltage level of13 V.

Stage 3.Anodic to cathodic erosion transition.At the 14and 15 trigger voltage level(0.8 to 1.6 ms into the arc discharge process),the surface analysis data shows a decrease in the rate of cathode deposition and anode erosion.At this stage in the arcing process,the relative amount of O spices and Sn species has started to decrease.It is thought that these species originate from the Sn O₂ in the contact surface.At the start of this stage,high amounts of O and Sn dominate the arc compared to Ag,and this is thought to be due to the lower latent heat of vaporisation.Toward the end of this stage,the contact surface is depleted of Sn O₂,and therefore,less Sn and O species are found in the arc,as seen in Figs.7and 8.

Stage 4.Arc conductance instability.At the higher trigger voltages between 16-20 V(1.6 to 2.6 ms into the arcing process),instabilities in the arc conductance causes the voltage characteristic to show increased levels of fluctuations.This is illustrated in Fig.9 by the widening of the voltage profile.At the same time,as the onset of these fluctuations,the affected area of the contact surface has enlarged to its maximum.The arc roots have become mobile,damaging the surface area and affecting the voltage level as they move.The anode has a maximum affected area with a diameter of 1.2 mm,and the cathode has a maximum effected area with a diameter of 0.5 mm.

4.2.“Arc degeneration”regime

Stage 5.Growth of gaseous ion arc.At the trigger voltage level of 20 V(3.2 ms into the arcing process),the gradient of the voltage characteristic becomes less steep.This gradient change is not necessarily due to gaseous ions from gases penetrating the arc,as the spectral data does not show N species dominating until later in the arcing process.However,as the arc lengthens,gaseous ions are expected to form in the arcing process at a voltage level of 24 V(5.6 ms into the arcing process)but not necessarily to dominate the discharge.A net transfer of mass from cathode to anode that is expected with the gaseous ion arc[1]is not seen in this study,which is consistent with other work at these voltage levels[15].Although after 24-26 V,N species are seen in the arc,and the anode does have an increasing loss of volume(to the environment).Spectra shows increase of Ag(I)and Sn,indicating vaporized anode material in the arc.It is thought that the gaseous ions do not dominate the discharge process at these voltage levels due to the lower energies involved and shorter arcs.As a consequence,there is no transfer of mass from the cathode to the anode.

At the middle to latter stages of the arcing process,the affected areas on each contact reaches a plateau in their sizes as illustrated in Fig.4.These areas are thought to be caused by the arc roots moving around on the contact surface.It is thought that the arc roots move toward the point of the highest electric field at the contact interface as material is eroded or deposited.Once the material is eroded and/or deposited,the profile of the electric field charges,causing the arc to move further.Thus,a cycle of erosion/deposition at particular positions,leading to changes in the electric field profile,leading to arc root movement,and leading to erosion/deposition at a new position,is set-up.

4.3.“Arc extinction”regime

The extinction process starts between trigger voltages of28-38 V(at~7.5 ms into the arcing process)as the arc is unable to support itself.

5. Conclusions

The erosion and arcing characteristics of a Ag Sn O₂ contact pair during an opening operation are studied up to 42 V under resistive load conditions.The contact pair is opened to several predetermined points in the arcing process determined by the arc voltage.At these particular arc voltages,the load current is terminated,and the effect of the arcing processes up to that point is investigated.

It is found that the initial stages of the arcing process with a voltage level<16 V(during the“arc generation”regime)are dominated by Sn and O species thought to originate from the Sn O₂ in the contact.Ag(I)also dominates in these initial stages.The deposition of the cathode is at a maximum at a voltage of 13 V(0.8 ms into the arcing process).It is found that at the middle stages of the arcing process with a voltage level of 24 V,as the arcing processes move from an“arc generating”regime to a“degenerating regime,”the affected area on the contact surface reaches a maximum(1.6-26 ms into the arcing process).Also,there are fluctuations in the arc voltage probably caused by the mobility of the arc.It is found that at the final stages of the arcing process with a voltage level of>24 V(during the“arc extinc-tion”regime),N and Ag(II)species dominate.

It is found that the anode undergoes major erosion com pared to the cathode.It is thought that this is due to the gaseous ions species not dominating the arcing process un der these conditions of short arcs and voltages<42 V.The extent of the larger areas of erosion,particularly seen in the latter part of the arcing process,is thought to be due to mul tiple mobile arc roots.It is proposed that this mobility is de pendent upon the changing surface profile as erosion and deposition takes place.

参考文献

[1] Swingler J. and McBride J.W., The erosion and arc character- istics of Ag/CdO and Ag/SnO2 contact materials under DC break conditions, IEEE Trans. Compon. Packag. Manuf. Technol. Part A, 1996, 19 (3): 404.

[2] Meng F.B., Lu J.G., Zhu Y.Q., Lu N.Y., Chen Z.C., and Huang Q.T., Failure characteristics of Ag-SnO2 contact mate- rial in AC level, Rare Met., 2008, 27 (1): 18.

[3] Du Z.J., Yang T.Z., Gu Y.Y., Qiu X.Y., and Dou A.C., Preparation of silver tin oxide powders by hydrothermal re- duction and crystallization, Rare Met., 2007, 26 (5): 470.

[4] Lu J.G., Wang J.Q., Zhao J.Y., Wen M., Wang B.Z., and Wang X.H., A new contact material?Ag/SnO2-La2O3-Bi2O3, Rare Met., 2002, 21 (4): 289.

[5] Slade P.G., Electrical Contacts: Principles and Applications, Marcel Dekker, New York, 1999.

[6] Ben Jemaa N., Doublet L., Morin L., and Jeannot D., Break arc study for the new electrical level of 42 V in automotive applications, [in] Proceedings of the 47th IEEE Holm Con- ference on Electrical Contacts, Montreal, 2001: 50.

[7] Ben Jemaa N., Doublet L., Jeannot D., Sone H., and Morin L., Transition from the anodic arc phase to the cathodic metallic arc phase in vacuum at low DC electrical level, [in] Proceed- ings of the 47th IEEE Holm Conference on Electrical Con- tacts, Montreal, 2001: 88.

[8] Doublet L., Ben Jemaa N., Jeannot D., and Hauner F., Make arc erosion and welding tendency under 42 V DC in automo- tive area, [in] Proceedings of the 49th IEEE Holm Conference on Electrical Contacts, Washington, 2003: 158.

[9] Doublet L., Ben Jemaa N., Hauner F., and Jeannot D., Elec- trical arc phenomena and its interaction on contact material at 42 V DC for automotive applications, [in] Proceedings of the 50th IEEE Holm Conference on Electrical Contacts, Seattle, 2004: 8.

[10] Chen Z.K., Material Transfer and Contact Resistance Dete- rioration in Light Duty Electrical Contacts [Dissertation], Keio University, 1995.

[11] McBride J.W., Witter G.J., and Chen Z.K., Volumetric erosion measurements of contacts tested in dc automotive relays, [in] Proceedings of the 47th IEEE Holm Conference on Electrical Contacts, Montreal, 2001: 73.

[12] McBride J.W. and Sharkh S.M.A., The effect of contact opening velocity and the moment of contact opening on the ac erosion of Ag/CdO contacts, IEEE Trans. Compon. Packag. Manuf. Technol. Part A, 1994, 17 (1): 2.

[13] Chen Z.K. and Sawa K., Effect of arc behavior on material transfer: A review, IEEE Trans. Compon. Packag. Manuf. Technol. Part A, 1998, 21 (2): 310.

[14] Chen Z.K. and Witter G.A., Comparison of contact erosion for opening velocity variations for 13 volt circuits, [in] The Proceedings of the 52nd IEEE Holm Conference on Electrical Contacts, Montreal, 2006: 15.

[15] Sallais D., Ben Jemaa N., and Carvou E., An arc study at high DC current levels in automotive applications, IEEE Trans. Compon. Packag. Technol., 2007, 30 (3): 540.

[16] Swingler J., Sumption A., and McBride J.W., The evolution of contact erosion during an opening operation at 42 V, [in] Proceedings of the 51st IEEE Holm Conference on Electrical Contacts, Montreal, 2005: 346.

[17] Swingler J. and McBride J.W., The net zero mass loss phe- nomenon on opening switching contacts with AC loading, IEEE Trans. Compon. Packag. Technol., 1999, 22 (1): 27.

[18] McBride J.W., The volumetric erosion of electrical contacts, IEEE Trans. Compon. Packag. Technol., 2000, 23 (2): 211.

[19] Ben Jemaa N., Doublet L., Jeannot D., Hauner F., and Morin L., Make arc parameters and subsequent erosion under 42 VDC in Automotive area electrical contacts, [in] Proceedings of the 48th IEEE Holm Conference on Electrical Contacts, 2002: 132.

[20] Taicaan Technologies, www.taicaan.com (accessed 30th July 2009).

[21] McBride J.W., A review of volumetric erosion studies in low voltage electrical contacts, IEICE Trans. Electron., 2003, E86-C (6): 908.

[22] Utsumi T., Theoretical and experimental investigations of the dynamic molten bridge, IEEE Trans. Parts Mater. Packag., 1969, PMP-5 (1): 62.