The Interruption of High Current Vacuum Arcs
As will be discussed in Chapter 6 in this volume a vacuum circuit breaker has to successfully pass a series of certification tests before it can be offered for sale. The test series are identified in the IEEE C67 and the IEC 62271 publications. The tests include successful interruption of high, short-circuit currents. The individual vacuum interrupters in vacuum circuit breakers have to be designed to interrupt these high currents. This section discusses the performance criteria that have an impact on a vacuum interrupter’s design that will give it an acceptable high-current interruption performance. I will firstly discuss the effect of the post-arc current that occurs immediately after the vacuum arc has extinguished at the ac current zero as the TRV appears across the open contacts. I will then explore the impact this post-arc current has on the expected TRV values experienced in an ac circuit. Finally, the effect of contact temperature at current zero and the effect of the residual metal vapor on interruption performance will be examined. It will be shown that the study of post-arc current in vacuum interrupters is largely of academic interest while the discussion of the residual metal vapor density in Section 4.2.2 is highly relevant.
In order to interrupt the high current vacuum arc, it is necessary to ensure that it returns to the diffuse mode long enough before current zero that all the effects of the column mode have been forgotten. In Chapter 3, Sections 3.3.3 and 3.3.4, in this volume, I discussed two contact structures that control and/or dissipate the energy of the columnar, high current, vacuum arc. The first contact design, the transverse magnetic field structure (TMF contact), forces the columnar vacuum arc to rotate around the periphery of the contact at quite high velocities. This effectively limits the energy per unit time that can be deposited into the arc roots on the contacts so that the necessary volume of metal vapor to sustain the columnar arc is not achieved all the way to current zero. Once the arc returns to the diffuse mode, the hot spots on the cathode and anode contacts cool on a millisecond time scale. For Cu-Cr, TMF contacts, the contact faces, show fairly uniform and smooth contact erosion after high current arcing.
The second contact design, the axial magnetic field structure (AMF contact), forces the columnar vacuum arc into a diffuse mode, even at very high currents. Pieniak et al.  have measured the surface temperature of Cu-Cr(25wt.%) spiral, TMF contacts during the passage of high currents. Figure 4.20 shows their data for a 19kA (rms) current. It shows the characteristic voltage of a high current vacuum arc between spiral, TMF contacts. As the contacts open the columnar vacuum arc remains in one position until the contacts are about 3-4mm apart. It then begins to rotate around the periphery of the spiral arm: region (a). During this time, the contacts in positions 1 and 2 show an increase in temperature well above the melting temperatures of Cu and Cr. At the beginning of region (b) the columnar arc begins to break up and transition to the diffuse mode. At position 5, the temperature has now decreased to 1863K just above the melting point of Cu. If we assume that the temperature decay rate for position 4 to 5 (360 K.ms4) continues to current zero, then the contact temperature at current zero will be about 1200 K. Looking at Table 4.2, the pressure of metal vapor in the contact gap would be less than 10_2Pa (or KHmbar). This number density of less than IO-l8trr3 is much lower than would be needed for the contact gap to breakdown and reignite the vacuum arc.
FIGURE 4.20 The surface temperature of TMF contacts during the columnar arc running phase (a) and as the column breaks up and the diffuse phase begins (b) .
This data confirms the hypothesis that the transition phase (b) in Figure 4.20 does indeed result in the successful interruption of high, short-circuit currents. Again, the arc energy is spread over the contact faces, so the contact erosion is also quite uniform. Gentsch et al.  show that the distribution of the energy from a 30kA(rms) vacuum arc is equally effective for both the TMF and the AMF contacts. Once the arc goes into the diffuse mode, it remains diffuse until the current goes to zero. Each of these structures has proven to be highly effective in permitting the vacuum interrupter to control and interrupt very high currents. As will be discussed below, however, a contact with a given diameter will eventually reach a limiting level of current above which it cannot interrupt the ac circuit.
Fenski et al.  have investigated the effect of a 50Hz half-cycle vacuum arc between AMF and TMF contacts for currents in the range 5kA (rms) to 40kA (rms). Figure 4.21 shows typical traces for a 25kA (rms.) current for both AMF and TMF contact structures. Both contact structures have a diameter of 90mm and are made from a Cu-Cr(25wt.%) material. In each case, the contacts open about 1ms after current initiation where the current is greater than 10kA. Thus, a well-established bridge column will exist immediately after the rupture of the molten metal bridge. In Figure 4.21 it can be seen that once the AMF contacts open, the arc voltage increases to a peak value at about lms. From the discussion in Section 3.3.4 in this volume, it is only after this period that the arc goes into the fully diffuse mode. As the arc voltage here remains at quite a low value, about 30V, it can be assumed (also from the discussion in Sections 2.6.3 and 3.3.4 in this volume) that there is sufficient cathode area for the cathode spots to spread with very little overlapping of the plasma plumes above them. This fully diffused arc would be expected to distribute its energy over most of the cathode and anode contact areas.
The arc voltage for the TMF contact in Figure 4.21, by contrast, is quite different. It does have, however, the exact structure that would be expected from my discussion in Section 3.3.3 in this volume. There is a dwell time of about 3ms during which time the contact gap opens to about 4.5mm. At this time, the columnar vacuum arc begins to move rapidly around the periphery of the contacts with the characteristic arc voltage spikes of up to 200V. About lms before current zero, the arc
FIGURE 4.21 The arc voltages for a 25kA (rms) vacuum arc between AMF and TMF contact structures .
motion ceases. This is when the vacuum arc would be expected to return to the diffuse mode. The slight bumps in the voltage traces (circled), result from the injected current from the Weil-Dobke synthetic circuit. It is interesting to note that for the 500ps before this voltage “bump,” both voltage traces have similar structures. Since we know that the vacuum arc between the AMF contacts is fully diffuse during this period before the current zero, it is reasonable to assume that this is good evidence that the arc between the TMF contacts is also fully diffuse at current zero.
It is now instructive to extend the post-arc current discussion that I began in Section 4.2.1 when I described the interruption of the diffuse vacuum arc. In fact, Fenski et al.  show' that the postarc currents (ipac) for both the AMF and the TMF contacts after the 25 kA (rms) vacuum arcs shown in Figure 4.21 are almost identical; see Figures 4.22(a) and (b). Here you can see that the shapes and peaks of the post-arc currents are nearly equal, even though an extremely fast rising TRV is used (60kHz, 8.5kV/ps). An abrupt reduction in the post-arc current occurs in both cases about 3ps after current zero: i.e., w'hen the ion sheath reaches the new anode. After this, there is a much slower decay, which results from the decay of the residual, low-density plasma remaining in the intercontact gap. I believe that these data present a very strong confirmation that in this experiment the columnar vacuum arc between the TMF contacts has transitioned into the diffuse mode long enough before current zero that all memory of the high current mode is forgotten during the post-arc
FIGURE 4.22 The post arc current after the interruption of a 25kA (rms) current with about 9 ms of arcing for (a) AMF contacts and (b) TMF contacts .
period. Figure 4.23 presents a compilation of high current interruption performance for TMF and AMF contacts as a function of vacuum interrupter diameter, which is generally proportional to the contact diameter, and also the circuit voltage. Flere it can be seen that, in general, the greater the vacuum interrupter’s diameter, the higher the current it can interrupt; also, the higher the circuit voltage, the greater the required vacuum interrupter’s diameter to interrupt the same current.
Let us return to the main circuit current and see what effect it has on the level of post-arc current. In an interesting set of experiments, Flakamata et al.  have taken large diameter contacts (170mm) with a multi-pole AMF structure behind them (see Figure 3.44). These open at the beginning of a 52FIz, half-cycle of current from the primary circuit. A Weil-Dobke synthetic circuit injects a high-frequency current just before current zero and then supplies the TRV. The di/dt of the injected current just before current zero, is related to the peak TRV and its RRRV. An oscillogram of a typical post-arc current is shown in Figure 4.24. The delay of lps-2ps in the critical rise of the TRV after current zero is clearly seen. Post-arc current data for a 57kA(rms.) vacuum arc are shown in Figure 4.25; here the influences of the di/dt, dUR(t)/dt and URlpeak) on the peak post-arc current are
FIGURE 4.23 A compilation of data taken from 12 vacuum interrupter manufacturers, for both TMF and AMF contacts, of the interruption limit as a function of current, circuit voltage and vacuum interrupter diameter (i.e., generally proportional to contact diameter).
presented. The most interesting data, however, are shown in Figure 4.26. Here the post-arc current is shown to strongly depend upon the di/dt just before the current zero and the dUR(t)/dt and t/PfJ.„tJjust after the current zero, but seems to have little memory of the high current arc that occurred before it. These results are of paramount importance to the vacuum interrupter designer. They mean that with a properly designed contact structure and with a properly sized contact to ensure that the vacuum arc is diffuse at current zero, the vacuum interrupter will successfully interrupt the ac circuit. In fact, contact structures have been suggested for the interruption of currents as high as 200kA .
FIGURE 4.24 The post arc current and the TRY after the interruption of a 57kA (rms) .
FIGURE 4.25 The effect of di/dt and TRV on the post arc current .
FIGURE 4.26 The effect on the peak of the post arc current as a function of di/dt before current zero, showing that memory of the peak arcing current has little effect .
Another aspect of these data that is important for the vacuum interrupter designer to consider is that a lower value of di/dt before current zero and a lower value of the TRV peak give a lower post-arc current. So, it might be expected that a contact with a given diameter would be able to interrupt a higher current in a 5kV circuit than it would in a 38kV circuit. Certainly in Figures 3.30, 3.58 and 4.22 the data for both TMF and AMF contacts seem to confirm this statement. A typical interruption characteristic for a well-designed vacuum interrupter, with either TMF or AMF Cu-Cr contacts, interrupting a symmetrical ac current is shown in Figure 4.27. In my experience this characteristic is valid for all circuit voltages and all symmetrical currents. Up to a limiting short circuit current (1.0 x ilim) the probability of interrupting a given short circuit current is 100%. After this ilim the probability of successful interruption rapidly falls to zero. Figure 4.28 shows the variation that
FIGURE 4.27 The probability of interruption of a symmetrical short circuit current as a function of a vacuum interrupter’s limiting current (iUm).
FIGURE 4.28 Variation in the probability of interruption of a symmetrical short circuit current as a function of a vacuum interrupter’s contact material (note: VIM is vacuum induction melting) .
can result in this characteristic. Here Godechot et al.  have evaluated the performance for Cu-Cr contact material with different compositions and manufacturing methods.
There is no clear relationship between post-arc current and the high current interruption performance of a given vacuum interrupter. That being said, some researchers have tried to find a direct correlation between the post-arc current and a vacuum interrupter’s ability to interrupt a given ac circuit [9,46,47]. For example, it might be expected that the greater the post-arc current, the greater will be the probability that the vacuum arc will reignite after current zero. However, in order to study the post-arc phenomenon in practical vacuum interrupters most researchers have had to investigate the vacuum interrupter’s performance beyond its natural design limit. They have done this in a number of ways; e.g., by increasing the current interrupted beyond the vacuum interrupter’s designed current interruption limit, by impressing a very fast di/dt just before the natural current zero, by having a very fast RRRV after the current zero and by increasing the peak value of the TRV beyond the design limit. In fact, some studies have used all four. The resulting data, however, have not given a definite correlation between the post-arc current and a given vacuum interrupter’s ability to interrupt the circuit. A good example is the work by Van Lanen et al. . Here the post-arc current is shown to vary considerably from one operation to another when a 25 kA (rms) arc is interrupted with a RRRV of about 6kV.ps-1. In these experiments the peak values of the post-arc current are shown to differ by a factor of 2, the time of the total post-arc current pulse by a factor of 1.75 and the total charge passed by a factor of 2.5. In all cases the circuit is successfully interrupted with no sign of stress. Also, most studies have shown (see, for example, Binnendijk et al. ) that the reignition can occur well after the post-arc current has dropped to a very low level, but the total TRV voltage across the contact gap has become large enough to cause a dielectric breakdown of the whole contact gap. One aspect of the post-arc current studies that is very interesting is that the model presented in Section 4.1.2 can be successfully applied to the case of high current vacuum arcs when those arcs return to the diffuse mode a millisecond or so before the natural current zero [9, 14, 20,48].
The questions, “Is there a relationship between the value of the post-arc current and a vacuum interrupter’s interruption ability?” and “What is the impact of contact design on the vacuum interrupter’s interruption ability ?” still need to be explored. Some insight to answer these questions can be found in the work of Yanabu et al. . These researchers have worked with a synthetic experimental circuit that gives them a high ac current for the arcing phase and then injects a high-frequency current just before current zero, which gives them the needed TRV to observe the post-arc current. Like other researchers , they initially open the contacts with a low dc current. Thus, all effects of the high current molten metal bridge and the subsequent development of the high current arc from the bridge column arc that I described in Chapter 2 in this volume are negated. In these experiments, a low current diffused vacuum arc will have been fully established for about 10ms as the contacts open. It is only then that the high ac current (~50Hz) is passed through the arc. Thus, this high current vacuum arc begins in the fully diffused mode with widely distributed cathode spots. The fact that there is no initial bridge column arc is especially important when interpreting the interruption data. For example, the high current arc between the butt contacts will take time to form into the columnar mode from the initially diffuse vacuum arc and the high current arc between the AMF contacts will immediately form into the high current diffuse mode. In spite of these differences from the normal operation of a vacuum interrupter in a practical ac circuit, the results of the experiments are instructive. Figure 4.29 gives the peak value of the post-arc current (ipac) for Cu butt contacts and for Cu AMF contacts. For the butt contacts, the f only increases slowly for circuit currents 4kA(rms.) to 18kA(rms.). After about 18kA(rms.), there is a sharp increase as the circuit current increases and failure to interrupt the circuit begins at about 21kA(rms.). With the AMF contacts, the slow increase of the ipac now continues to about 25kA and then increases more rapidly. In this case there are no failures to interrupt the circuit, even though the ipac reaches higher values than those measured for the butt contact where interruption does not occur. Thus, just measuring the ipac alone does not necessarily give an indication of how easily a given contact structure interrupts a given high current circuit.
FIGURE 4.29 The post arc current after the interruption of a vacuum arc as a function of circuit current for Cu butt and Cu AMF contacts .
It is also important to understand the importance of the contact structure. These data can be interpreted using the discussion presented in Section 3.3 in this volume and our knowledge of the butt and the AMF contact structures. The vacuum arc between the butt contacts will develop into the columnar mode once the circuit current exceeds 14-15 kA(rms). As this columnar vacuum arc burns through current maximum, it will release a copious quantity of metal vapor into the contact gap in order to maintain the high-pressure arc column. It will also heat the contacts at the arc’s roots to a temperature higher than the melting point of the Cu. At current zero this stationary columnar vacuum arc will have well-developed residual hot spots on the new cathode and anode as well as an increase in the residual metal vapor and plasma in the intercontact gap. Thus, not only will the contacts continue to supply metal vapor to the intercontact region, but it will also be possible to have electron emission from the new cathode even for relatively low fields impressed on it by the rising TRV. In these experiments, the columnar arc on the butt contacts remains in place until just before the current zero for a circuit current of about 20kA(rms) when the density of the metal vapor in the contact gap is high enough for the gap to break down after the current zero. The Cu contacts with an AMF impressed across them seem to have a large enough diameter for a fully diffuse high current vacuum arc to be developed for the whole current half cycle; even for currents as high as 35kA(rms). As I have discussed in Section 2.6.3 in this volume, the AMF does trap the plasma between the contacts, but for the fully diffuse high vacuum arc, the energy into the contact is spread over the faces of the contacts and is not concentrated in one small area. The AMF contact, with its diffuse arc at current zero, will have no residual hot spots. Consequently, even though there may be the same plasma density between the contacts at current zero, there is a much lower probability of electron emission from the new cathode and any failure to interrupt the circuit will occur at much higher currents. Figure 4.30 presents the effect of Cu content for Cr-Cu contacts on the ipac for increasing arcing current: here the highest Cu content (75 wt.%) gives the lowest ipac. Again, no interruption failures are recorded even with a ipac.& 18A.
While Hakamata et al.  have shown that for a well-designed contact only the di/dt of the ac current before current zero influences the ipac and there is little memory of the peak of the arcing current, it is possible to demonstrate the effects of the vacuum arc’s current especially if butt contacts are used. Of course, in a normal ac circuit the di/dt just before the current zero is directly proportional to the peak of the ac current; see Equation (4.1). For experiments with butt contacts
FIGURE 4.30 The post arc current after the interruption of a vacuum arc as a function of circuit current for Cu-Cr AMF contacts with different Cr content .
opening a 50Hz circuit and interrupting а ЮкА current Lindmayer et al.  show that the ipac. pulse has a total time of about lOps and a peak value of about 2.5A. At this time, in their experiment, the recovery voltage, UR(t), has a value of about 50kV and it is increasing, at a rate UR(t)/dt of about 5kV/ps, to its peak value of 160kV. After a half cycle of arcing with a 15kA (rms) current the ipac continues to increase. When UR(t) reaches 75kV, the contact gap breaks down and the circuit is not interrupted. An example of the continued increase of the ipac and the breakdown of the contact gap is shown in Figure 4.31. Here the contrast between the ipac values after the passage of а ЮкА current and the passage of a 15kA current is well illustrated. At 15kA, the high current stationary columnar vacuum arc would have formed. At current zero, there would have been hot spots on the new
FIGURE 4.31 The post arc current after the interruption of a vacuum arc as a function of circuit current for Cu-Cr butt contacts for two circuit currents, ЮкА and 15kA .
cathode and a high enough density of metal vapor in the contact gap to initiate a gas breakdown. Smeets et al.  show that the post-arc current on its own is not a measure of a vacuum interrupter’s high-current interruption performance. While observing certification testing of vacuum circuit breakers, they observe the following instances of reignition and failure to interrupt high-currents:
- 1. Immediate reignition during the rising edge of the post-arc current. This type of reignition is extremely rare and is only observed when the current to be interrupted is much higher than the value for which the vacuum interrupter had been designed
- 2. Early reignition during the post-arc current stage. Again, this rarely occurs and indicates that the current is beyond the vacuum interrupter’s design current
- 3. Delayed reignition or dielectric reignition which occurs after the post-arc current has ceased and the TRV across the open contacts increases to its peak value. This is the most common form of reignition in vacuum interrupters and infers that the current to be interrupted is greater than the ilimshown in Figure 4.27
These observations infer that there has to be sufficient metal vapor in the contact gap after current zero for a Townsend avalanche to take place. This in turn suggests that for successful current interruption to take place, the contacts at current zero have to have cooled enough to prevent excessive evaporation of metal vapor into the contact gap. Temperature measurements by Donen et al.  of a TMF spiral contact interrupting 16 kA (rms) gives a final contact temperature at current zero of 1675K which is similar to that measured by Pieniak et al. . This will give a metal vapor density in the contact gap of less than 1020.nr3. Poluyanova et al.  record the anode temperature at current zero as a function iljm for an AMF contacts: see Figure 4.32. The anode temperature at current zero for the current i,im is 1660K. So, the temperature of the vacuum interrupter’s contacts at current zero appears to be the most important parameter for successful interruption of high currents. Also, the post-arc current plays little or no part in determining a vacuum interrupter’s performance. As
FIGURE 4.32 The anode temperature of the anode for an AMF contact after current zero as a function of .
Lindmayer et al.  state, there can be several current enhancement mechanisms working together to a level that the circuit cannot be interrupted. These include:
- 1. The ipac resulting from the ion sheath expansion towards the new anode and the ion current flowing towards the new cathode
- 2. The emission of electrons from the new cathode liberated by the field at the cathode resulting from UR(t)/s(t). If there were a residual hot spot on the cathode after the current zero, the electron emission would be enhanced by T-E emission
- 3. The emission of electrons will also be enhanced by the bombardment of the new cathode by ions traveling from the edge of the space charge sheath
- 4. If there is a high enough residual metal vapor density after the current zero and/or if the contact surfaces remain at a high enough temperature to continue to evaporate metal vapor, the electrons can begin to ionize this metal vapor. The resulting higher current would also continue the heat input into the contact surfaces and result in the continued evaporation of metal atoms
- 5. The cooling of the contacts by conduction or from the evaporation of the metal, as well as the dispersion of the residual metal vapor from the contact gap, would limit the effects of this ionization
Again, the structure of the contact plays a major role at high currents. For example, Lindmayer et al.  show that the peak of the /p„(.for the butt contact when it fails to interrupt the 15k A circuit is the same as that observed for both a TMF and an AMF contact structure after a half cycle of arcing at 25kA, where they both successfully interrupt the circuit at current zero. This again emphasizes the importance of these contact structures for distributing the energy from the high current vacuum arc more or less uniformly over the contacts’ surfaces.
As Kaumanns  has shown, the interruption of high currents can result in somewhat random behavior in the amplitude and duration of the post-arc current. So, while the level of circuit current to be interrupted can have an effect on the shape and duration of the post-arc current pulse as well as its magnitude, other parameters have to be considered when analyzing ipac data. One overwhelming influence on the ipac is, of course, the design of the contact structure. The effects of current level and contact structure are well illustrated in Figure 4.33. Here the 90mm diameter, Cu-Cr contacts used to develop the ipac data after interrupting 25kA (rms), Figure 4.22, are now subjected to a 40kA (rms) vacuum arc . For both the AMF and TMF contacts the greater di/dt before the current zero gives rise to a higher ipac before the readjustment of the plasma and the formation of positive space charge sheath at the new cathode. Once the TRV begins to appear across this sheath, the appearance of the ipac differs for the AMF and TMF contact structures. The ipac for the AMF contact has a shape similar to that shown in Figure 4.22. The maximum value of ipac is higher after the 40kA arc, but its duration is about the same. The larger total charge density in the contact gap is to be expected. However, the same ds/dt (Equation 4.7) of the sheath is not expected, i.e., the sheath moves at the same average speed from the new cathode to the new anode after both the 25kA rms arc and the 40kA rms arc between the AMF contacts. The TMF contact structure shows quite a different ipac pulse after the 40kA arc. Once the TRV begins to appear across the sheath, the ipac increases to a value of 10A about 4ps after the current zero, before falling to a low value at about 7ps.
Interestingly enough, the shield voltage begins to appear at the same time for the TMF contact as it does for the AMF contact. This implies that the sheath reaches the anode at about the same time for both contact structures even though the ipac pulses are quite different. Figure 4.21 shows that the vacuum arc between TMF contacts returns to the diffuse mode about 1ms before the current zero for the 25kA rms current. At 40kA rms I would expect this return to be perhaps 0.5ms or less before the current zero. Thus, there could well be a higher temperature region on the new cathode as well as a higher residual metal vapor density between the contacts after current zero. The voltage appearing across the sheath would then be sufficient to liberate electrons from the new cathode by T-E
FIGURE 4.33 The post arc current between (a) AMF contacts and (b) TMF contacts from Figure 4.17 after the interruption of a 40kA (rms) current circuit and a 9 ms arcing time .
emission. If the metal vapor density is high enough ionization could also take place. The increase in the ipac indicates that both of these phenomena are occurring. However, in this example neither of these enhancement processes is sustained. Once the TRV appears across the full contact gap, the contact surface cools and the residual metal vapor dissipates, the ipac then begins to decrease to a very low value and eventually ceases to exist. It is interesting to note that in this case the much higher i does not result in a failure to interrupt the circuit. Both the TMF contact and the AMF contact withstand the full TRV.
As I will present in Chapter 6 in this volume, vacuum circuit breakers that usually have an opening delay of a few current cycles will still have a considerable AMF across the contacts at current zero. This lag in the AMF results from the eddy currents in the AMF contact structure. The AMF across the contact gap after the current zero will keep the residual plasma trapped between the contacts. Most experimental studies of the post current zero phenomena, however, minimize this effect by opening the contacts during the first half cycle of current. Even so, Steinke et al.  have shown that even when AMF contacts open on the first half cycle of current there can be a considerable AMF across the contacts at the first current zero. Depending upon the AMF contact structure, it can be 30% to 60% of its peak value. The effect of this AMF on the residual plasma between the contacts at current zero is well illustrated by the work of Arai et al. . Using a 500Hz current pulse with a peak value of 3kA and a di/dt я -9.5 A.ps~‘at current zero, they show that for a residual AMF from an AMF contact structure with 100mm diameter, OFIFC Cu contacts and a contact gap of 30mm the ion density after current zero is about 2 x 1018 nr3. When additional dc AMFs of lOOmT and 200mT are added, the ion density increases to about 4 x 10l8nr3 and 8 x 10l8nr3, respectively. Ge et al.  using an 20kA (rms) vacuum arc show that an additional AMF pulse of up to lOOmT at current zero can increase the intercontact gap ion density by as much as 3.5 times. Steinke et al.  have investigated ipac for two AMF contact designs. The first has an AMF 2.2 times that of the second at current zero. They show that the ipac of the first AMF design is always greater than that of the second after current zero of a 50kA and 60kA arc. They show that while the higher ipac and residual charge may impact the eventual breakdown of the contact gap, the dielectric breakdown of the contact gap occurs long after the ipac has gone to zero. They conclude that the difference in performance between the two contact systems can be attributed primarily to differences in the residual metal vapor density and not the ipac. In this case the contact design with the higher AMF confines a higher density of the plasma, which is a mixture of neutral vapor, electrons and metal ions, longer within the contact gap. While this does show a higher ipac, it is the higher density of the neutral vapor that eventually results in the breakdown of the contact gap during the recovery phase after current zero.
In another paper, Steinke et al.  show that the opening time before the current zero has a marked effect on the value and duration of the / . An example of their experimental data and a simulation using the model described in Section 4.1.2 is shown in Figure 4.34. Here the effect of shorter arcing time (i.e., smaller contact gaps), lower arcing currents and the state of the vacuum arc
FIGURE 4.34 The post arc current pulse as a function of opening time before current zero for an AMF contact structure, comparing experimental data with a calculation using the sheath model .
- (i.e., at what stage it is in when becoming fully diffuse) on the ipac is clearly observable. The first ipac peak is proportional to the product of the initial ion density, the initial ion velocity and the effective area of the contact between the residual plasma and the new cathode after the current zero. The increase in the ipac duration is explained by the time it takes for the sheath to cross the contact gap to the new anode. It is interesting to note that for the shortest contact gap and the shortest arcing time, the vacuum arc would most probably still be in the bridge column mode. Thus, even though the ipac has the lowest peak value and has the shortest duration, there would be a high density of metal vapor trapped between the closely space contacts and there would be a high probability of reigniting the arc once the recovery voltage had reached a high enough value. This study also shows that by increasing the RRRV, both the magnitude and the duration of the ipac are increased. Their simulations indicate that this results from secondary electron emission caused by positive ions impacting the new cathode. These data imply that the residual metal vapor between the contacts after current zero is the major cause of arc reignition and not the ipac. For this reason, Dullni et al.’s conclusion [37, 38] that the residual metal vapor and not the residual charge has the greater influence on whether or not a contact gap will recover its dielectric strength after a current zero is valid even though there is plasma trapped by the residual AMF at current zero. This is indeed fortunate: as long as the total density of metal vapor between the contacts remains below about 1022 nr3, then the vacuum interrupter will have an excellent chance of interrupting an ac circuit’s current. The recovery of the contact gap can be seen in three stages:
- 1) The residual plasma in the contact gap dominates the first stage after the current zero. It can be described using the sheath model, which is determined by the TRV (i.e., the dUR(t)/ dt ) and the velocity (ds/dt) of the sheath as it travels from the new cathode to the new anode
- 2) The vapor density dominates the second stage as the TRV increases after the post arc current has decreased to zero. If the metal vapor density between the contacts is too high a gas-assisted breakdown of the recovering contact gap is possible
- 3) The third stage is the full recovery of the contact gap to its full design high voltage, withstand value
What does this mean for the vacuum interrupter designer? First of all, when designing a contact structure to interrupt ac currents, it is important to ensure that the vacuum arc reverts to the diffuse mode before the current zero is reached. For low' currents, the recovery process is so fast that for all practical ac circuits it is almost impossible not to interrupt it. The sheath model of the contact gap recovery provides a good description of the physical phenomena involved in changes in the residual plasma during the initial stages of the recovery process, but it is not an indicator of the vacuum interrupter’s ability to interrupt current. At high currents, if a columnar vacuum arc persists until current zero, the vacuum interrupter wall fail to interrupt the ac circuit. This results primarily from the ionization of the increased density of the neutral metal vapor in the intercontact gap by electrons emitted from the hot spots formed on the new' cathode.
The TMF and AMF contact structures that have been developed to control the high current vacuum arc have both demonstrated an excellent ability to distribute the vacuum arc’s energy over the surface of the contacts. They also ensure the diffuse vacuum arc mode at current zero. In fact, if the diffuse vacuum arc is present at current zero, then the sheath model of the post-arc current is still valid even for high-current vacuum arc . If the contact structure is designed correctly, then the value prior to the current zero of the arcing current seems not to be important: i.e., the contact structure’s memory of the high current arc is not the primary driver during the recovery process. Thus, for the vacuum interrupter designer, it is important to size the contact for the expected level of current to be interrupted and to make sure the diffused vacuum arc is present 1 or 2ms before the expected current zero. Figures 3.33 and 4.35 gives examples of the vacuum arc appearance at peak current and then just before current zero for well-designed TMF and AMF contacts.
FIGURE 4.35 Photographs of a high current vacuum arc (31.5kA rms) between TMF, Cu-Cr (25 wt.%) contacts and a high current vacuum arc (18kA rms) between and Cu-Cr (20 wt.%) contacts with an AMF contact structure at current maximum and just before current zero.
As Figures 3.30, 3.58, and 4.23 indicate, there is a limit of current that can be interrupted by a contact with a given diameter for both the TMF and AMF designs. The discussion in this Section is relevant to explain this phenomenon. If at current zero the gas density of metal vapor between the contacts is too high, the gap between the contacts will breakdown at a voltage lower than its true vacuum break dowm level, because electrons crossing the gap will ionize the gas. A metal vapor density can result from two interacting processes on the contact face.
The first process is the time before current zero that the vacuum arc goes into the diffuse mode. The greater this time, the more completely will the metal vapor from the high current arc phase dissipate from the contact gap. It will also permit the contact surfaces to cool down to a temperature low enough that the metal vapor released from the contact surfaces after current zero does not add significantly to the gas density between the contacts. Thus, the probability of the gas density remaining below about 1022 nr3 will increase.
The second and related process is the high current arc phase. If the heat input into a contact face is too great, it may liberate so much metal vapor that it can still have a high density before the current zero even if the vacuum arc has gone into the diffuse mode. The TMF contact is successful because the high current columnar vacuum arc moves rapidly around the periphery of the contact once the contacts exceed a gap of 3-4mm: see Section 3.3.3 in this volume. As the ac current falls to zero, the motion of the arc roots results in a dearth of metal vapor released and this eventually results in a failure to maintain the high-pressure arc column. This then leads to a transition of the vacuum arc into the diffuse mode and a cooling of the residual arc spots. A good example is presented by Donen et al.  who show that during the columnar arc motion between TMF spiral contacts the maximum temperature is 2000C. As the 16kA current reaches zero, the surface temperature of the contacts has dropped to about 1400C. As the arc moves around the contact the arc root regions on the contact surfaces cool once the arc has moved on. If now the current is increased for a given contact diameter, not only does the heat input into the contact surface increase, but also the velocity of the column increases (at least up to about 95kA: see Figure 3.27). Thus, not only is more metal vapor released into the contact gap, but also the arc roots return to their starting point faster, shortening the cooling time of the contact surfaces. Consequently, a high enough current will eventually be reached beyond which a TMF contact of a given diameter will fail to interrupt at a current zero.
At first sight, it would appear that the AMF contact would not have the same limitation discussed above for the TMF contact. Experimental evidence, however, does show an interruption limit for a given contact diameter. This limit results from the nature of the diffuse high current vacuum arc discussed in Sections 2.6.3 and 3.3.4 in this volume. The cathode spots will spread over the contact. If the plasma plumes from these spots do not overlap before they reach the anode, then the heating of the anode will occur over most of the anode’s surface. Once the number of cathode spots (which is a function of the current) exceeds a given number for a given contact diameter (i.e., area), the plasma plumes will overlap and a diffuse plasma column will develop. The current flowing through this column will subject it to a magnetic pinch force, which will limit the area where it attaches to the anode. Thus, the heat input to the anode will be increased in this area. If the current is high enough, the heat into this anode spot will also be high enough to permit too high a density of metal vapor at current zero.
Watanabe et al.  using AMF contact structures show that for a metal vapor density greater than 1020 nr3, the contact gap will fail to interrupt the current. They state that this occurs when the contact surfaces reach a temperature of about 1750K. Table 4.2 shows that the vapor pressure of Cu plus Cr at this temperature is about 33Pa. Niwa et al.  indicate that a somewhat higher temperature of 2150K is required. Here Table 4.2 shows that the vapor pressure of Cu plus Cr is now about 2 x 103Pa. At this temperature, Table 4.1 shows that there is a gas density of 7 x 1022 nr3 and a small electron mean free path (about 10‘2mm). Thus, it is possible for a Townsend avalanche to be established and the vacuum arc to reignite. It can therefore be seen that for both the TMF and AMF contact structures there will be a limited current that can be interrupted for a given contact diameter.
Other factors can influence the density of metal vapor at current zero. One, of course, is the nature of the contact material itself. For example, its thermal conductivity, its electrical conductivity and its vapor pressure at given temperatures will all influence the interruption performance at current zero. A second and perhaps not so obvious factor, is the total volume inside the vacuum interrupter. The metal vapor requires a volume into which it can move away from the contacts themselves. If this volume is limited, it is possible for a higher pressure of metal vapor to exist at current zero. The third effect is the contact opening time, the rate of rise of the recovery voltage and its peak value. All will affect the interruption ability of the butt, TMF and AMF contact structures. This will be discussed further in Chapters 5 and 6 in this volume.