APPENDIX A

TFE AND CELL TEST RESULTS

A-1

This appendix consists of:

-Device, design, and processing parameter tables (Tables A-1 to A-3)

-Operational history tables, which give test details for all recent cell, capsule, and TFE tests (Tables A-4 and A-5)

-Summaries of individual tests, test history graphs, and neutron radiograph results for in-pile tests, in the following order:

- TFE Tests

TFE 6F2 TFE 2F2 TFE 1F3
TFE 6F3 TFE 2F3 TFE 1F4
TFE 6F4 TFE 1F1 TFE 2E2
TFE 6F5
TFE 2F1

- Cell Tests

Mark VIIA IC-D3
Mark VI IC-I6

- Capsule Test

FC-3

- Out-of-Pile Test

Mark VIIA OC-A4

A-2

Hot cell examination in progress.

The TFE experienced six shutdowns at 1241, 2579, 4307, 5067, and 7685 hours and nine scrams at 116, 359, 1575, 1594, 4328, 4898, 4906, 5713, and 6756 hours.

At 765 hours the current density was raised to lower the emitter temperatures. Operating condition changes following startups at 1241, 1594, and 2579 hours are visible (Figs. A-l, 2). From 2866 to 3226 hours a high emitter temperature was maintained for diagnostic purposes. The performance loss at 4307 hours followed a shutdown. At 4767 and 5713 hours the current density was reduced in order to maintain the emitter temperature with fission gas in the inter-electrode space.

PERFORMANCE HISTORY (Figures A-1 to A-4)

The TFE was initially operated for a period of 70 hours with emitter temperatures near 1700° K during which its performance was observed to fall, apparently as a result of a dropping emitter substrate work function. Subsequent operation near 1925° K resulted in a slow improvement in performance, diagnosed to be the result on an increase in the emitter substrate work function. Consequently, the TFE was operated with an average emitter temperature of 2000°K from 2866 to 3226 hours in an attempt to clean the emitter surface. A small improvement in performance resulted, substantiating the diagnosis. Subsequent operation at 1875°K was stable until the shutdown at 4307 hours at an electrode performance approximately two thirds that predicted by CPOP under the estimated operating conditions of the TFE. The source of the contaminant which is apparently affecting the electrode work functions in the TFE has not been identified.

A-13

The power output dropped at 4307 hours following a shutdown for neutron radiography. This drop is attributed to a lowering of the emitter temperature due to the presence of fission gas in the inter-electrode space. The drop beginning at ~5200 hours has been attributed to the same cause. The TFE’s performance fell again at 6400 hours, probably due to a continued increase in the inter-electrode space fission gas pressure.

Intermittent shorting behavior has been observed beginning at ~4700 hours, probably as the result of inter-electrode shorting in the top cell. More severe shorting to the sheath tube began at 7000 hours, possibly through the top cell's sheath insulator.

NEUTRON RADIOGRAPHY RESULTS (The 7685 hour data is shown in Fig. A-5)

Radiographs were taken at 1241, 2579, 4307, 5713, and 7685 hours.

Emitter Dimensional Stability: No swelling was observed in any emitter (£ 0.10 mm [4 mils] on the diameter) prior to 5713 hours. At that time 0.53 mm (21 mils), 0.09 (4 mils), and 0.08 mm (3 mils) of diametral swelling was visible in emitters 1, 2, and 6 respectively. None (£ 0.08 mm [3 mils]) was detectable in emitter 3, 4, and 5. At 7685 hours the observed diametral swellings in emitters 1 to 6 were 0.68 mm (28 mils), 0.07 mm (3 mils), 0.11 mm (4 mils), 0.40 mm (16 mils), 0.32 mm (12 mils), and 0.50 mm (20 mils).

Fuel-Clad Reaction Layers: In emitters 1 (86% enrichment) and 6 (78%) a maximum of 0.13 mm (5 mils) of clad had reacted at 1241 hours. No significant changes were noted in later radiographs.

In emitters 2 (34%) and 5 (34%) no reaction was detected (£ 0.08 mm [3 mils]) at 1241 hours but reactions were detected in #5 at 2579 hours and #2 at 4307 hours (~ 0.13 mm [5 mils] of clad reacted). No changes were found at 5713 hours. At 7685 hours a maximum of 0.13 mm of clad had reacted in both emitters.

A-14

No reaction was detected in emitters 3 (20%) and 4 (20%) until 5713 hours. At 7685 hours a maximum of 0.13 mm of clad had reacted in both.

Fuel Behavior: Fuel sintering was observed in all emitters, with a maximum of 4% and 5% (on the diameter) in the central two emitters (3 and 4, enrichments 20%), 1% and 3% on emitters 2 and 5 (enrichments 34%), and 2% and 1% on emitters 1 and 6 (enrichments 86% and 78%). At 7685 hours the pellets in all emitters had returned to full diameter and were fuzing together. In addition, in many of the emitters the fuel was expanding into the central cavity and into the stem area.

Other: The titanium backup plate for the rupture disc in emitter 4 sagged. Those in emitters 3, 5, and 6 bowed up slightly.

At 5713 hours contaminants in the inter-electrode space were found in cells 1, 4, and 6, and at 7685 hours contamination was visible in the inter-electrode spaces of all cells.

Cracks were visible in the emitter clads of cell 1 at 5713 hours and cells 1 and 6 at 7685 hours. The seal convolution in cell 6 deformed around the emitter stem and may be cracked.

The thermionic performance of TFE 6F2 was low but fairly stable throughout the test due to the presence of an electrode contaminant of uncertain origin. A leak between the fission gas and inter-electrode spaces apparently occurred in association with a shutdown at 4307 hours and the TFE’s current density was adjusted to maintain its emitter temperature as the gas pressure built up. Probable intermittent shorting of the top cell was observed from ~ 4700 hours on. At 7000 hours a short to the sheath tube, possibly through the top cell’s trilayers, occurred.

A-15

Neutron radiograph results indicated that fuel sintering occurred to various degrees in all emitters. Emitter dimensional changes generally followed the return of the fuel pellets to full diameter after sintering. Cracks were visible in the clads of those emitters that changed the most (1 and 6) at 7685 hours. A contaminant was visible in the inter-electrode space of all cells at this time.

A-16

Hours: 8062 Relative Power: 0.92

Average Emitter Temperature: 1740°K

Performance: Dropping slowly with fission gas present in the inter-electrode space

TFE 6F3 experienced seven shutdowns at 1728, 2488, 5106, 5438, 6683, 6985, and 8062 hours, and six scrams at 1733, 2319, 2328, 3134, 4177, and 6792 hours.

Reactor control rods near the TFE were raised at 1922 hours to preferentially raise the temperature of the TFE’s upper cells, in an effort to achieve a better emitter temperature profile and to reduce performance cycling. The operating current density was reduced at 2170 hours to raise the emitter temperatures.

Testing was terminated at 8062 hours for programmatic reasons.

PERFORMANCE HISTORY (Fig. A-6 to A-9)

The performance of the TFE dropped during the initial 30 hours of testing before stabilizing. At ~ 120 hours the TFE’s operating correlations were measured. They indicated the average collector temperatures were nearly optimum and there was no evidence that the collectors were running significantly hotter than the TFE sheath. Individual correlations between emitter and sheath temperatures were obtained.

A-25

The TFE’s initial relative power was slightly better than expected considering the device’s wide spacing (14 mils, P_R ~ 1.05). This, plus the fact that TFE 6F3’s optimum cesium reservoir temperature was ~ 10°K lower than that observed in TFE 1F1 for similar operating conditions indicates an oxygen reservoir of some sort may have been present in the TFE.

Beginning at 829 hours the TFE experienced cyclic performance decreases during which its output power dropped ~ 40% for 1/2 hour periods. These cycles have been attributed to the existence of a leak between the TFE’s cesium and fission gas spaces which allows the accumulation of cesium in the cooler portions of the fission gas spaces in the TFE’s lead. This cesium contacts hot regions of the lead at irregular intervals resulting in brief periods of high cesium pressure operation with concurrent low TFE power output. A total of 134 periods of cycling had been experienced as of 8062 hours. Operation with a cesium reservoir temperature below optimum has been found effective in reducing the performance cycles repetition rate, and the TFE has been operated slightly below optimum from ~ 1000 hours. Operating low in the core to increase the power to the TFE’s upper cells has also proved helpful. The TFE design was modified to correct this deficiency.

Following a shutdown at 1728 hours evidence of fission gas in the inter-electrode gap was found. This is probably gas vented from the fuel at the time of the shutdown. The fission gas pressure acts to reduce emitter temperature in the TFE, but is not high enough to significantly affect thermionic performance.

Following the shutdown at 6683 hours, an internal short to the sheath tube, probably occurring through a sheath insulator in one of the TFE’s upper cells, was observed. The resistance of the short appeared to increase slowly after the 6985 hour shutdown.

Radiographs were taken at 0, 1728, 3134, 5106, 6683, and 8062 hours. The last three sets are shown in Figs. A-10 to A-13.

A-26

Maximum Emitter Dimensional Change (on Diameter)

Cell 1728 hrs 3134 hrs 5106 hrs 6683 hrs 8062 hrs

1 (mm) £ 0.08 £ 0.05 0.06 0.14 0.17
(mils) £ 3 £ 2 2 6 7
2 (mm) £ 0.08 £ 0.05 0.09 0.08 0.24
(mils) £ 3 £ 2 4 3 9
3 (mm) £ 0.08 0.08 0.11 0.17 0.25
(mils) £ 3 3 4 7 10
4 (mm) £ 0.08 0.08 0.13 0.20 0.18
(mils) £ 3 3 5 8 7
5 (mm) £ 0.08 0.13 0.18 0.24 0.27
(mils) £ 3 5 7 9 11
6 (mm) £ 0.08 0.12 0.19 0.14 0.20
(mils) £ 3 5 7 5 8

Distortion was observed in the emitter bottoms of cells 1, 2, and 4 at 3134 hours and thereafter. The maximum distortion consisted of ~ 0.6 mm (25 mils) deflection downward at the center at 8062 hours (cell 2). No distortion was observed in cells 3 or 6 (£ 0.1 mm [5 mils]). A downward deflection of ~ 0.3 mm (10 mils) was visible in cell 5 at 8062 hours.

Fuel-Clad Reaction Layers: No evidence of fuel-clad reactions has been observed.

Fuel Behavior: The fuel had fully redeposited by 1728 hours except possibly at the top and bottom corners. Isothermal cavities are visible in all emitters. The fuel cover plates in most emitters were distorted slightly (~ 0.5 mm [20 mils]) by the fuel pushing upward at 3134 hours and thereafter.

Other: The titanium backup plates in emitters 2, 4, and 6 were pushed up 3-4 mm at 1728 hours, indicating that the titanium rupture disc had bonded to the plate and that fission gas pressure had then pushed the assembly up.

A-27

Despite the presence of a leak between the fission gas and cesium spaces in the TFE, its thermionic performance prior to 6683 hours was high and stable. The TFE is subject to power loss cycles of ~ 1/2 hour duration at irregular intervals due to cesium behavior in the fission gas spaces. This condition occurs as a result of capsule heat transfer characteristics and would not occur in a thermionic reactor. There is evidence that a partial sheath insulator short after the 6683 hour shutdown has reduced the TFE’s performance, and the short resistance seems to be increasing.

Neutron radiographs indicate the presence of swelling in all emitters, with the maximum 0.27 mm (11 mils) in emitter 5 at 8062 hours.

A-28

Hours: 2956 Relative Power: 1.00

Average Emitter Temperature: 1700°K

Performance: Stable

This test has experienced four shutdowns at 332, 1577, 1879, and 2956 hours and one scram at 1686 hours. Operating correlation data was obtained during the first 250 hours of testing. The reactor power was raised at 238 hours. After 311 hours the current was adjusted to provide the lowest emitter temperature consistent with a lead power output of 700 watts. An emitter temperature of 1700°K was maintained after 668 hours.

The TFE’s input power was temporarily low following the 1577, 1686, and 1879 hour startups.

Testing was terminated at 2956 hours as a result of the program closeout.

PERFORMANCE HISTORY (Fig. A-14 to A-17)

The TFE’s performance initially was quite high with a P_R > 1.1, dropping during the first one thousand hours to a P_R = .95. During this period the efficiency of the second cell down in the TFE decreased to a degree which explains much of the performance loss. Its optimum cesium reservoir temperature is high indicating an emitter contaminant may be present. The high performance of the rest of the TFE indicates an oxygen source may also be present. Performance has been stable from 1000 hours, with lead power output of ~ 700 watts with an average emitter temperature of 1700°K.

A-39

NEUTRON RADIOGRAPHY RESULTS (Figs. A-18 to A-23)

Radiographs were taken at 0, 1577, and 2956 hours.

Emitter Dimensional Stability: No emitter dimensional changes were observed at 2956 hours (£ 0.05 mm [2 mils] on the diameter).

Fuel-Clad Reaction Layers: No evidence of’ fuel-clad reactions was observed at 1577 hours. Four mils (0.10 mm) of’ cladding had reacted at 2956 hours in cells 1 and 6.

Fuel-Behavior: No fuel sintering has been observed (£ 0.5%)

Following an initial performance drop apparently associated with the second cell down in the TFE, the test has operated stably with at PR @ 1.0.

No emitter swelling or fuel sintering has been observed.

A-40

Hours: 1379 Relative Power: 0.98

Average Emitter Temperature: 1800°K

Performance: Stable

This test has experienced two shutdowns, at 302 and 1379 hours, and one scram at 109 hours. Between startup and 109 hours, it was operated in reactor hole C-10. Between 109 and 302 hours, it was operated in F-5, and since 302 hours in E-19. Operating correlations data was taken at 400 hours.

Testing was terminated at 1379 hours as a result of the program closeout.

PERFORMANCE HISTORY (Fig. A-24 to A-27)

The TFE’s performance has been stable with a P_R @ 1.0 since its operation was optimized at 400 hours.

NEUTRON RADIOGRAPHY RESULTS (Fig. A-28 to A-33)

Radiographs were taken at 0 and 1379 hours.

A-51

Cell 1379 hrs

1 0.11 (mm)

2 #0.05 (mm)

#2 (mils)

3 0.20 (mm)

4 #0.05 (mm)

#2 (mils)
5 0.07 (mm)
3 (mils)
6 0.14 (mm)
5 (mils)

#

The four fuel pellets in emitters 1, 2, and 5 are slightly cocked and the emitter clad has swollen to partially conform to the resulting staggered structure. Some swelling of these emitters was anticipated due to the designed interference fit between fuel and clad at operating temperatures. The slightly staggered nature of the fuel stack has apparently accentuated this swelling. There was less swelling in emitters 2 and 4, which was probably due to the specially shaped fuel pellets which were used to reduce the stress on the clad.

Fuel-Clad Reaction Layers: Three mils (0.08 mm) of clad had reacted in cells 1 and 6 at 1379 hours.

Fuel Behavior: The fuel in emitters 2 and 3 has sintered ~ 1/2%.

Other: One open fission gas vent hole is visible in both cells 5 and 6.

A-52

The performance of TFE 6F5 has been stable with a P_R ~ 1.0 since operating conditions were stabilized at 400 hours. The use of fuel pellets designed for interference fit at temperature in combination with stacking irregularities had generated some emitter swelling at 1379 hours.

A-53

Awaiting hot cell examination.

TFE 2F1 has experienced six shutdowns at 1338, 3066, 3826, 6444, 6776 and 8021 hours, and seven scrams at 334, 352, 3086, 3657, 3663, 4422, and 5515 hours. Following the shutdown at 1338 hours the TFE was moved in core to allow testing at higher current densities. It was moved again at the 3066 hour shutdown to slightly reduce the power input. High emitter temperature tests for diagnostic purposes were performed at 250, 290, 1176 hours, and from 3976 to 4120 hours. At 5451 hours the current density was reduced to allow higher emitter temperature operation. Testing was terminated at 8021 hours due to emitter temperature uncertainties and low thermionic performance.

PERFORMANCE HISTORY (Figures A-32 to A-35)

During the initial 70 hours of testing with an average emitter temperature of ~ 1400°K the TFE’s performance dropped rapidly and simultaneously the optimum cesium reservoir temperature rose, indicating a change in electrode work functions. The rate decreased following a raise in emitter temperature. The TFE was then twice operated open circuit with emitter temperatures of ~ 1930°K each time, resulting in improved performance and a lower optimum cesium reservoir temperature. The second test resulted in a relative performance compared to CPOP of 95%, a 20% improvement over the TFE’s pre-heat treatment performance. These tests strongly indicated the presence of an emitter contaminant.

The performance improved further with another raise in operating emitter temperature (~ 1850° avg.) to 105% of CPOP and remained at that level until

A-63

1176 hours when a third high temperature test to 1970° was performed. The third test resulted in a performance drop, associated with indications of cesium instabilities and possible shorting during the open circuit period of operation. The precise reason for the performance drop has not been ascertained. The post-test relative power of 90% improved until the reactor was shutdown for neutron radiographs at 1338 hours.

Operation with higher input power resulted in stable performance with a calculated Pr ~ 1.15. The emitter temperatures may actually have been hotter than calculated during this period, further explaining the high stable performance.

A comparison of IV curves taken during the test (Figure A-35) indicates that an emitter work function change, not a gas effect, was producing the performance loss between 3000 and 5150 hours, but fission gas in the inter-electrode space may have been contributing to the performance loss after 5150 hours. The IV curves and the TFE’s power output characteristic indicate an inter-electrode short occurred in the top cell at ~ 6700 hours.

A-64

NEUTRON RADIOGRAPHY RESULTS (The last two sets are shown in Figs. A-36, 37)

Radiographs were taken at 0, 1338, 3066, 4472, 6444, and 8021 hours.

Emitter Dimensional Stability: No swelling was observed in either emitter until 8021 hours, when ~ 0.51 mm (20 mils) was observed on the diameter of the top emitter. No swelling was observed in the bottom emitter (£ 0.05 mm) at that time.

Fuel-Clad Reaction Layers: A maximum of 0.08 mm (3 mils) of cladding had reacted in both emitters at 1338 hours (81% enrichment and 80% enrichment). No change was observed at 3066 or 4472 hours. A maximum of 0.13 mm (5 mils) of clad had reacted at 6444 and 8021 hours.

Fuel Behavior: Both sets of fuel pellets were sintered a maximum of 2% at 1338 hours; the bottom cell pellets had returned essentially to full diameter at 3066 hours, but some of those in the top remained sintered at 4472 hours. At 6444 hours all pellets had returned to full diameter. The central fuel cavity was almost closed in both emitters at 6444 and 8021 hours, and the fuel had risen into the stem area 2 mm in the top and 3 mm in the bottom emitter at 8021 hours.

Other: The spring locators in both emitters were broken at startup. (The design was changed as a result of this observation). After 4472 hours a contaminant was visible in the inter-electrode space of both cells. The quantity of contaminant visible at 8021 hours makes it very probable that both emitters are cracked. The collector bus bar tabs broke off in the first 1338 hours of operation. The TFE and bus bar configuration has remained stable since that time, however.

The performance of TFE 2F1, initially high, has suffered throughout testing due to the presence of an emitter temperature sensitive contaminant. Some

A-65

of this contaminant is visible in the neutron radiographs. There is evidence that a leak between the cesium and fission gas spaces occurred at 5150 hours, and an inter-electrode short in the top cell at 6700 hours. No swelling was observed in the bottom emitter during the test. While the top emitter showed no swelling prior to 6444 it swelled ~ 20 mils between 6444 and 8021 hours. A maximum of 0.13 mm (5 mils) of cladding had reacted in both emitters by 8021 hours. The collector bus bar tabs broke early in the test, probably due to the high temperatures that these leads were operated at, but the lead structure retained its current carrying capability.

A-66

Hours: 5574 Relative Power: 0

Average Emitter Temperature: 1820°K

Performance: Unstable due to gas and electrode contamination

The test has experienced five shutdowns at 2618, 2950, 4195, 4497, and 5574 hours and three scrams at 646, 1689, and 4304 hours. Following an initial period of reactor and TFE stabilization, the TFE’s current was adjusted to provide a lead power of 320 watts. This resulted in an average TFE lead power production equivalent to that required in a 6F TFE producing a lead power of 1000 watts.

Evidence of a fission gas leak to the inter-electrode space was observed after the 4195 hour shutdown, and the current density was dropped to maximize output power. At the 4497 hour shutdown the TFE was moved in core.

PERFORMANCE HISTORY (Figs. A-38 to A-41)

The relative performance of this TFE was high initially but had stabilized at ~ 1.0 by 100 hours. Collector optimum data taken at 190 hours indicated the TFE had an optimum collector temperature higher than OC-A4 and it was operating with a collector temperature below optimum. For calculating purposes, the optimum has been assumed to be 130° higher than that found in OC-A4.

A-75

Prior to the gas leak at 4195 hours, the performance of the TFE was stable within + 10%. Following the gas leak, its performance became more erratic with performance cycles with periods of the cycle of 1 1/2 days occurring near the end of the test. Higher optimum cesium reservoir temperatures were observed during the lower portion of the cycle than during the higher. Whether these cycles and the general loss of performance is due to high gas pressures or electrode contamination has not been ascertained.

The cycles and performance loss are not entirely due to cesium instabilities, though they may be a contributing factor.

NEUTRON RADIOGRAPHY RESULTS (Final three sets, Figs. A-42 to A-44)

Radiographs were taken at 0, 646, 2618, 4195, and 5574 hours.

Emitter Dimensional Stability: No swelling was detected at 5574 hours in the top emitter (£ 0.05 mm [2 mils] on the diameter). Swelling amounting to 0.08 mm (3 mils) on the diameter in the bottom emitter was noted at 646, 2618, and 5574 hours, but was not detected at 4195 hours.

Fuel-Clad Reaction Layers: No reaction layer was detected in either emitter at 646 hours but ~ 0.08 (3 mils) of clad had reacted at 2618, 4195, and 5574 hours.

Fuel Behavior: Fuel sintering was observed in both emitters at 2618 and 4195 hours, and amounted to ~ 1% on the diameter in both. At 5574 hours the fuel had swollen back in both cells. The fuel was bulging upward ~ 1.8 mm (80 mils) in the top cell and ~ 1.0 mm (40 mils) in the lower at this time.

Others: The fuel hold down devices appear to be distorting slightly due to the upward swelling of the fuel.

A-76

The performance of TFE 2F2 was stable and as predicted for 4195 hours. It produced power equivalent to a lead power of 1000 watts in a 6F TFE during that period. Following a gas leak after a shutdown at 4195 hours its performance rapidly fell as the result of gas and electrode contamination and testing was terminated at 5574 hours.

No swelling of the top emitter was observed during the test and the bottom emitter swelled ~ 0.08 mm during the first 646 hours of operation and stabilized there. The fuel in both emitters sintered and had swollen back by 5574 hours.

A-77

Hours: 2956 Relative Power: 0.8

Average Emitter Temperature: 1750-1850°K

Performance: Dropping slowly. Fission gas in inter-electrode space. Partial lower cell short.

This test has experienced four shutdowns at 332, 1577, 1879, and 2956 hours and one scram at 1686 hours. It was moved to a new reactor hole at 1879 hours.

Testing was terminated at 2956 hours as a result of program closeout.

PERFORMANCE HISTORY (Figs. A-45 to A-48)

Based on ignition current temperatures, the performance of TFE 2F3 was initially high with P_R ~ 1.3 and dropped smoothly prior to stabilizing at ~ 1100 hours. Sheath temperature-heat flux correlations and ignition current temperature estimates indicate the performance loss was due to fission gas in the inter-electrode space. Slow response to cesium temperature changes confirms this diagnosis. The equilibrium thermal conductance in the gas was k = 3.2 x 10^-4 watts/cm²°K. For purposes of calculating emitter temperature and P_R, the conductance of this gas was assumed to have increased linearly in time between 0 and 1070 hours.

Following the 1879 hour shutdown, collector bus currents indicated a partial short in the bottom cell of the TFE. This has been confirmed by neutron radiographs taken at 2956 hours. The resistance of the short appears to have decreased further around 2500 hours.

A-88

The rapid swelling visible in the bottom emitter of the TFE is not consistent with the low temperatures at which ignition currents indicate this device has operated. It is quite possible that the presence of the gas has affected the ignition current correlation. The high calculated relative power also indicates emitter temperature estimates may be low. An emitter temperature 75-100°K hotter than those calculated would result in relative powers nearer unity. If that is the case the average emitter temperature during the test may be more nearly 1850°K than the 1760°K calculated.

NEUTRON RADIOGRAPHY RESULTS (Figs. A-49 to A-51)

Radiographs were taken at 0, 1577, and 2956 hours.

Maximum Emitter Dimensional Change (on Diameter)

Cell 1577 2956

1 mm £ 0.05 £ 0.05
mils £ 2 £ 2
2 mm 0.10 0.33
mils 4 13

Fuel-Clad Reaction Layers: No evidence of fuel clad reactions has been observed.

Fuel Behavior: The fuel in both emitters had completely redeposited by 1577 hours. It was bulging upward ~ 1 mm in the top emitter and 0.8 mm in the bottom emitter at 2956 hours.

Other: The titanium backup plate for the rupture disc has risen up slightly in the top emitter at 1577 hours and was up 1.1 mm at 2956 hours. It had sagged 4 mm in the bottom emitter at 1577 hours and 4.8 mm at 2956 hours.

A-89

This test has suffered from inter-electrode gas throughout testing, making estimates of the emitter temperature difficult. An average based on ignition current calculations is 1760°K. TFE performance would be more consistent with an 1850°K average.

The two emitters behaved quite differently. The top emitter, with a 2 mm clad, showed no signs of swelling although the fuel had bulged up ~ 1 mm and the rupture disk backup plate had swollen upward by 2956 hours, indicating possible fission gas containment. The lower emitter, with 1 mm clad swelled and partially shorted at 1879 hours. The fuel had bulged upward less than that in the top emitter at 2956 hours and a sagging rupture disk backup plate indicates the fission gases have probably been released from the emitter structure. Thus, despite the uncertainties in the average operating temperature, one can conclude that the 2 mm clad has significantly less clad swelling.

A-90

Awaiting hot cell examination.

The TFE’s operating current has been adjusted through the test to maintain an emitter temperature of ~ 1800°K despite input power changes as required by other test devices. Seven shutdowns have been experienced at 543, 1784, 3122, 4850, 5610, 8228 and 8560 hours. Ten scrams have occurred at 138, 660, 902, 2118, 2135, 4870, 5450, 5456, 6256, and 7299 hours. Periods of changing operating conditions primarily due to reactor power stabilization are visible following most startups. The reactor startup following the 543 hour shutdown extended over several days, as required for test initiation of TFE 6F2.

PERFORMANCE HISTORY (Figs. A-52 to A-55)

The initial operating conditions of the TFE were established at 117 hours. The emitter temperature was estimated at that time on the basis of ignition current data, and the collector temperature was estimated on the basis of OC-A4 data and collector optimum data taken with the TFE at 1160 hours. This data indicated the collector was running 200° hotter than the sheath tube. The electrode power performance of TFE 1F1 based on these estimates was 7% better than that predicted by CPOP. Between 4000 and 8228 hours, the thermionic performance of 1F1 decreased slowly at the rate of 6%/10,000 hours. The decrease may have been the result of nonuniform inter-electrode spacings resulting from emitter swelling. During the 8228 hour shutdown the TFE developed a partial inter-electrode short which did not clear on start up. Testing was consequently terminated at 8560 hours.

A-98

SHEATH INSULATOR TEST HISTORY (Figs. A-56 to A-57)

At 1352 hours testing of the TFE’s sheath insulator was initiated with the application of a 5.0 volt potential between the emitter and sheath tube. The sheath current rose at that potential from an initial value of approximately 8 ma to 40 ma, where it appeared to stabilize. At 5300 hours the voltage was raised to 7.5 with a resultant current of 180 ma. At 5900 hours the voltage was raised to 10.0, resulting in operation with some ignited currents at 600 ma. At 6400 hours the current ignited further and stabilized with a current near one ampere. The test history and samples of the sheath tube-emitter IV characteristics at different times during the test are shown in the figures. The source or sources of the discharge currents (which dissipate a total of ~ 10 watts) was not ascertained.

NEUTRON RADIOGRAPHY RESULTS (Final 2 sets, Fig. A-58)

Radiographs were taken at 0, 1784, 3122, 4850, 6256, 8228, and 8560 hours.

Emitter Dimensional Stability: No emitter swelling was observed (#0.10mm [4 mils]) on the diameter at 1784 or 3122 hours, but the clad had swollen a maximum of 0.14 mm (6 mils) at 4850 hours, 0.27 mm (11 mils) at 6256 hours, 0.32 mm (13 mils) at 8228 hours, and 0.37 mm (14 mils) at 8560 hours.

Fuel-Clad Reaction Layers: No evidence of fuel clad reactions was observed.

Fuel Behavior: The fuel had fully redeposited by 1784 hours. Cracks along the fuel edges have been visible since 4850 hours, but no venting cracks have been seen.

A-99

Other: The titanium backup plate for the rupture disc appears to have fallen on one side between 4850 and 6256 hours and it was as far down as the upper radiation shields at 8560 hours.

TFE 1F1 performed stably at 7% above its predicted level until 4000 hours, at which time it started decreasing at a rate of 6%/10,000 hours. This decrease may be due to the observed emitter swelling. Such swelling could result in both off cesium optimum operation of portions of the emitter and increased conduction cooling by inter-electrode cesium vapor, resulting in reduced efficiency. Testing was terminated at 8560 hours following an inter-electrode short which occurred during a shutdown at 8228 hours.

Sheath insulator testing was conducted from 1352 hours and at test termination the TFE was supporting a stable discharge current of ~ 1 amp at 10 volts.

A-100

Awaiting hot cell examination.

TFE 1F3 experienced two shutdowns at 760 and 3378 hours. It has also experienced five scrams, at 20, 591, 598, 1406 and 2449 hours. The operating current density was lowered at 140 hours from 7.0 to 6.0 A/cm to raise the emitter temperature. The sheath temperature was lowered at 615 hours from 1150° K to 1090° K to allow operation nearer the optimum collector temperature. From 910 hours to 1055 hours the emitter temperature was raised from 1860°K to 1930° K in order to ascertain its affect on performance. At 2290 hours the TFE was lowered in the core in an attempt to stop cesium refluxing. Testing was terminated at 3378 hours.

PERFORMANCE HISTORY (Fig. A-59 to A-62)

TFE 1F3 is unique in that it was fabricated with a hole between the fission gas and cesium spaces. Fission gas effects must therefore be a prime suspect for any performance changes seen.

The performance of the TFE has varied in time from the beginning of testing. While it initially had a relative power of ~ 1, it experienced a large loss of performance between 250 and 450 hours. The prime cause of the performance loss appears to be a change in collector work function. Emitter work function changes would be reflected in optimum cesium reservoir changes which have not been observed. Emitter temperature loss due to gas would result in a loss in open circuit voltage, as would a raise in collector temperature due to trilayer unbonding. The changes in V_oc which were observed could be attributed almost entirely to changes in sheath temperatures. In addition, the horizontal shift nature of the IV curves is indicative of j _c changes as opposed to emitter temperature or work function changes. Changes have also been observed during the test in the TFE’s response to collector temperature variations, suggesting changes either in collector sheath tube temperature differences or collector work function. The presence of fission gas must

A-110

remain suspect since the inter-electrode and fission gas spaces in this TFE are connected. It should be noted that TFE 1F2 had fuel similar to TFE 1F3, and preliminary hot cell results indicate a very low fission gas release rate for that device.

Cesium refluxing in the vicinity of the TFE lead seal was observed beginning at 2000 hours. The refluxing resulted in intermittent shorts lasting typically for tens of hours. At 2290 hours the TFE was lowered in the core and the current density increased to maintain the emitter temperature. The resultant increased heating in the stem area cleared the shorting. Intermittent shorting, possibly through the sheath insulator, began at 3200 hours, however, and the TFE was partially but stably shorted when the test was terminated at 3378 hours.

SHEATH INSULATOR TEST HISTORY (Figures A-63, 64)

Discharge testing of TFE 1F3 was initiated at 64 hours with a voltage of 5.0 volts and a current of 275 ± 25 ma. This current remained steady until 615 hours. At that time the collector temperature was lowered 160°K and the sheath current fell from 260 ma to as low as 150 ma. At 621 hours the current rose rapidly as a result of a collector-sheath tube short, and testing was terminated.

NEUTRON RADIOGRAPH RESULTS (The last set is shown in Fig. A-65)

Radiographs were taken at 0, 1406, and 3378 hours.

Emitter Dimensional Stability: No detectable swelling (£ 0.05 mm [2 mils] on the diameter) was observed at 1406 hours. At 3378 hours the emitter diameter had increased a maximum of 0.07 mm (3 mils).

Fuel Clad Reaction Layers: Approximately 0.08 mm (3 mils) of clad had reacted by 1406 hours. The amount reacted had not increased by 3378 hours.

A-111

Fuel Behavior: There is some fuel deposition on the clad I.D. visible opposite all four fuel pellets. The fuel was sintered a maximum of 1% on the diameter at 1406 and 3378 hours.

The performance of TFE 1F3 has been unstable and poor due at least in part to collector work function changes. No definite evidence of gas in the inter-electrode space has been found. Sheath insulator testing was terminated at 621 hours because of shorting, primarily to the collector. The test was terminated at 3378 hours with possible shorting through the sheath insulators.

The emitter diameter had increased a maximum of 0.07 mm (3 mils) at 3378 hours.

A-112

Hours: 4928 Relative Power: 0.40

Average Emitter Temperature: 1840°K

Performance: Dropping

Testing of TFE 1F4 was initiated on June 15, 1972. The secondary containment gas was changed at 140 hours to lower the collector temperature. The TFE has experienced five shutdowns at 1972, 2304, 3549, 3851, and 4928 hours, and two scrams at 1043 and 3658 hours.

Testing was terminated at 4928 hours as a result of program closeout.

PERFORMANCE HISTORY (Fig. A-66 to A-70)

The performance of this TFE was stable from startup to ~ 3500 hours. The apparent change in emitter temperature, input power, and relative performance at 140 hours are the result of changing operating correlations concurrent with the secondary gas change.

The initial relative power and optimum cesium reservoir temperature of this TFE were comparable to those expected of a nonoxygenated device. The low optimum collector temperature suggested additive performance. Collector temperatures were varied in an attempt to observe an oxygen additive effect on the IV curves as shown in the figures. No effect was found, and consequently, the presence of additive performance remained unconfirmed.

A-121

Following the 3549 hour shutdown, the performance dropped and it continued to fall until test termination at 4928 hours. Analysis of the device’s IV curves indicate the probable reason for the performance loss was an emitter work function change. A less likely alternative is fission gas in the inter-electrode space, which cannot be completely ruled out.

It is interesting to note that the performance loss occurred in approximately the same time some of the fuel pellets returned to full diameter.

SHEATH INSULATOR TEST HISTORY (Fig. A-71, 72)

Discharge testing of TFE 1F4 was initiated at 193 hours with a voltage of 5.0 volts. The current remained stable at 1 ma. At 1700 hours the voltage was raised to 10.0 volts with a stable current resulting of 4 ma. At 1980 hours the polarity was reversed and the sheath tube made positive with respect to the emitter. The insulator resistance fell, and at 2320 hours it developed a bistable mode as shown in Fig. A-72. Currents could be driven stably for short periods on either IV curve. The insulator would switch from one mode to another for reasons which have not been ascertained. Tests indicate the current flow was through the collector sheath insulator. After 2400 hours testing was terminated and current was driven by the TFE potential alone.

NEUTRON RADIOGRAPHY RESULTS (Figs. A-73, 74)

Radiographs were taken at 0, 1972, 3549, and 4928 hours.

Emitter Dimensional Stability: No detectable change in emitter diameter (£ 0.05 mm [2 mils]) was measured at 1972, 3549, or 4928 hours.

Fuel Clad Reaction Layers: Approximately 0.08 mm (3 mils) of clad had reacted by 1972 hours. No change was noted at 3549 hours or 4928 hours.

A-122

Fuel Behavior: The fuel was sintered a maximum of 1/2% on the diameter at 1972 hours, and 1% on the diameter at 3549 hours. The top two pellets remained sintered ~ 1% at 4928 hours, but the bottom two had swollen back to approximately the clad I.D.

TFE 1F4 operated stably with performance comparable to a similar but non-oxygenated TFE for ~ 3500 hours. It is possible that additive performance was masked by electrode contaminants in the device. It is also possible the processing through the 1520°K sheath tube bonding cycle during fabrication had a detrimental effect on the oxygenated molybdenum layer on the collector. The performance of the TFE dropped after 3500 hours, probably due to electrode contamination effects.

Sintering of the carbide fuel was a maximum of 1/2% at 1972 hours and 1% at 3549 hours. The bottom two pellets had swollen back by 4928 hours.

A-123

Hours: 11084 Relative Power: 0.94

Average Emitter Temperature: 1840°K

Performance: Dropping slowly

The TFE’s operating current has been adjusted throughout the test to maintain an emitter temperature average of 1850°K. Ten shutdowns have been experienced, at 543, 1784, 3122, 4850, 5610, 8228, 8560, 9805, 10107, and 11084 hours. Eleven scrams have occurred at 138, 660, 902, 2118, 1(2?)235, 4870, 5450, 5456, 6256, 7299, and 9914 hours. Periods of operating condition change, primarily due to reactor power stabilization, are visible following most startups. The TFE was moved in core following shutdown at 543 hours.

Testing was terminated at 11084 hours as a result of the program closeout.

PERFORMANCE HISTORY (Fig. A-75 to A-78)

The TFE’s initial conditions were established at 116 hours. Emitter temperatures calculated from ignition current measurements indicate the two emitters in TFE 2E2 are operating at two temperatures ~ 130°K apart with the bottom probably hotter. The reported temperatures are the averages of these two emitter temperatures. Its collector temperature was estimated using collector optimum data taken at 1150 hours in conjunction with OC-A4 data. On the basis of these measurements the collectors of TFE 2E2 are estimated to be running about 130°K hotter than the collector heat rejector. Based on these estimates at 116 hours the TFE 2E2 had an electrode performance 6% better than predicted by CPOP.

A-134

The TFE’s performance was stable within ± 10% throughout the test, but since 8000 hours there were indications of a slow performance drop possibly due to emitter swelling.

NEUTRON RADIOGRAPHY RESULTS (Last three sets, Figs. A-79 to A-81)

Radiographs were taken at 0, 1784, 3122, 4850, 6256, 8228, 9805, and 11084 hours.

1784 3122 4850 6256 8228 9805 11084

Top Emitter (mm) £ 0.10 £ 0.10 £ 0.10 0.11 0.14 0.17 0.22
(mils) £ 4 £ 4 £ 4 4 6 7 9
Bottom Emitter (mm) £ 0.10 0.10 0.18 0.31 0.33 0.36 0.46
(mils) £ 4 4 7 12 13 14 18

Fuel-Clad Reaction Layers: No evidence of fuel-clad reactions has been observed.

Fuel Behavior: The fuel was fully redeposited at 1784 hours. None had moved above the radiation shields in either emitter or below the pedestal in the bottom emitter, but a small amount of fuel had moved below the pedestal in the top emitter at 3122 hours. The fuel in the top emitter was cracked through in all radiographs after startup except those at 11084 hours (presumably due to differential thermal expansion on cooling). Cracks across the fuel have also been observed in the bottom cell. Cracks appear at the upper top corners of the fuel in the top emitter.

Other: The fission gas vent holes are visible in both cells and appear open.

A-135

The performance of the TFE has been stable within ± 10% throughout 11084 hours of testing, although there is evidence of a slow drop after 8000 hours. Swelling is visible in both emitters, but is significantly greater in the bottom emitter which operated at a higher temperature. The swelling may be contributing to the slow performance loss. No evidence of shorting behavior was observed.

A-136

Awaiting hot cell examination.

Following an initial period of operation at 9 A/cm², the cell was taken open circuit at 280 hours. The reactor was shut down briefly at 332 hours. Between 337 and 341 hours, the emitter temperature fell several hundred degrees. Testing was terminated at 1577 hours.

PERFORMANCE HISTORY (Fig. A-82, 83)

The initial performance of IC-D3 was difficult to determine accurately because of uncertainties in ignition current determined emitter temperature resulting from uncertainties in cesium pressure. These uncertainties arise from the fact that the cesium reservoir is graphite-cesium lamellar compound. The best estimate is a value of P_R = 1.0 ± 0.2. These estimates of the emitter temperature based on ignition currents indicate it was running approximately 150°K colder than the thermocouple temperatures.

At 280 hours the cell was taken open circuit and the cesium reservoir temperature reduced in order to maximize the emitter temperature rise. The emitter thermocouple temperature rose to ~ 2450°K, indicating the emitter surface temperature was near 2300°K. The emitter thermocouple temperature dropped slowly until 337 hours, probably as the result of degradation. Between 337 and 341 hours the temperature fell to ~ 1200°K. The cesium reservoir temperature was raised between 500 and 1000 hours to determine if inter-electrode gas might have been the source of the cooling. No such evidence was found. The probably reason for the drop, substantiated by neutron radiograph pictures, was contact between the emitter and collector as a result of emitter swelling.

A-146

NEUTRON RADIOGRAPHY RESULTS (Fig. A-84)

Radiographs were taken at 0 and 1577 hours.

Emitter Dimensional Stability: The emitter diameter had increased a maximum of 0.30 mm (12 mils) at 1577 hours.

Fuel-Clad Reaction Layers: No evidence of fuel clad reaction layers was observed.

Fuel Behavior: The fuel was fully redeposited at 1577 hours.

Following 280 hours of normal operation IC-D3 was taken open circuit to determine the effect of high emitter temperature operation. Such operation, with an emitter temperature near 2300°K, resulted in emitter swelling sufficient to short out and cool the emitter in 57 hours. The pressure exerted on the emitter came both from the UO₂ fuel and the 1 atmosphere helium pressure in the fuel cavity.

A-147

Hours: 6334 Relative Power: 0.87

Average Emitter Temperature: 1830°K

Performance: Dropping slowly

IC-I6 has experienced six shutdowns at 760, 3378, 3710, 4955, 5257, and 6334 hours. It has also experienced six scrams at 20, 591, 598, 1406, 2449, and 5064 hours. Its performance characteristics were mapped between 3270 and 3320 hours. The emitter temperature was cycled once a week to a temperature 100°K above the operating point for a period of ~ 4 hours beginning at 3378 hours.

Testing was terminated at 6334 hours for programmatic reasons.

PERFORMANCE HISTORY (Figs. A-85 to A-88)

The initial performance of the cell was high with a P_R of ~ 1.2, but it dropped and stabilized at a P_R ~ 1.0 during the first 400 hours of testing. A slow downward trend is evident after ~ 3000 hours. This may be due to inter-electrode gap changes caused by emitter swelling. The scatter in the calculated data is the result of scatter in collector heat rejector power data which is used in the calculations, and does not represent real variations in the cell’s operating conditions.

NEUTRON RADIOGRAPHY (Last three sets, Figs. A-89 to A-91)

Radiographs were taken at 0, 1406, 3378, 4955, and 6334 hours.

A-151

Emitter Dimensional Stability: No swelling was observed until 4955 hours, at which time the emitter diameter had increased by 0.11 mm (4 mils) relative to the emitter bottom. At 6334 hours it measured 0.08 mm (3 mils) relative to the bottom, but there was evidence the bottom had swollen 3-4 mils, making a total of 6-7 mils. The emitter temperature was cycled to a value 100°K higher than nominal for ~ 4 hours once a week. Eight such cycles were performed between 3378 and 4955 hours and seven between 4955 and 6334 hours. Therefore, these temperature cycles increased the emitter diameter about 0.5 mils (0.12 mm) per cycle.

Fuel-Clad Reaction Layers: No evidence of fuel clad reactions has been observed.

Fuel Behavior: As of 1406 hours the fuel had fully redeposited and had risen upward ~ 0.3 mm (10 mils) on one side, cocking the radiation shields slightly, at 3378 hours the radiation shields had been tilted up ~ 0.6 mm (25 mils).

Following an initial period of high performance, IC-I6 stabilized with relative power nearly that predicted. The performance has dropped slowly following the initiation of emitter temperature cycling, probably as the result of concurrent gap size variations.

Neutron radiograph measurements indicate the emitter temperature cycling has resulted in a faster swelling rate than observed in IC-I4, a test operated under similar conditions.

A-152

Testing was terminated at 1379 hours for programmatic reasons.

CHRONOLOGICAL TEST HISTORY (Fig. A-92 to A-95)

This test experienced a scram at 109 hours, shutdowns at 302, and 1379 hours.

The initial 108 hours of testing were conducted with helium as the secondary containment gas. Neon gas was introduced into the secondary containment at this time in order to bring the emitter temperatures closer to their design values (1800°K). The data taken during the next 150 hours was used to derive the correlation between the emitter thermocouples and the average of the two thermocouples in the inconel sleeve (collector) surrounding each emitter. A correlation was also derived between the emitter thermocouples and the thermocouples placed on each tantalum emitter cap. In both of these correlations, it was assumed that the average emitter surface temperature was 50°K hotter than the emitter thermocouple temperature (based on 2D calculations). All measured and calculated temperatures are shown on the performance plots.

Most of the thermocouple changes observed in the first 200 hours of testing were probably the result of redistribution of the UO₂ fuel. At 236 hours the emitter thermocouple in the top emitter dropped 80°K, probably due to some flaw in the thermocouple itself. While the slow drop in the emitter thermocouple temperature visible in all the emitters later in the test may be evidence of fuel redeposition, it is more likely that it is the result of thermocouple degradation.

A-162

The input power values shown were derived from the correlation presented in Section 1 and the emitter temperature values.

NEUTRON RADIOGRAPH RESULTS (Figs. A-96 to A-99)

Radiographs were taken at 0 and 1379 hours.

Emitter 1379 Hours
1 £ 0.08 (mm)

£ 3 (mils)
2 £ 0.05 (mm)
£ 2 (mils)
3 0.05 (mm)
2 (mils)
4 0.10 (mm)
4 (mils)

Fuel-Clad Reaction Layers: No evidence of fuel-clad reactions were observed.

Fuel Behavior: The fuel had extensively redeposited in the top two emitters and appeared to be completely redeposited in the bottom two emitters at 1379 hours. Some top corner cracks were visible in the fuel in the top emitter.

Other: The fuel voids visible in the top two emitters are quite complex and the actual fuel internal geometry following redepositions is not clear.

There is evidence that some emitter swelling may be occurring in the bottom two emitters at 1379 hours. There is not enough data yet to confirm that the different fuel geometry in the top two emitters are reducing the swelling rate.

All four emitter thermocouples were working at 1379 hours, although there is evidence they may be suffering degradation.

A-163

The test of Mark VIIA OC-A4 included performance mapping with a liquid cesium reservoir, performance mapping with a graphite sorption reservoir-cell combination, and life testing. Testing was terminated after a total of 26,296 hours of operation and a post-operative examination was completed.

The performance mapping of OC-A4, with both a liquid and. a graphite sorption reservoir, was completed on March 13, 1969 after 814 hours of operation and life testing was initiated with an input power of 1535 watts and a current density of 10 A/cm². Degradation of the emitter thermocouples occurred during the final mapping of the cell, and the initial emitter temperature (2050°K) was estimated on the basis of the initial cell mapping. After 24,536 hours of operation a remotely programmable driving supply was attached and the cell operated to test that unit. The testing was terminated after a total of 26,296 hours of operation and a post-operative examination conducted.

PERFORMANCE HISTORY (Figs. A-100, A-101)

The initial electrode power density was 6.8 W/cm². The small abrupt changes in output power which occurred during the test are attributed to filament aging with concurrent changes in the emitter temperature profile. The cell’s output appears to have slowly dropped a few per cent between initiation of the testing and 15,229 hours. At that time some low power diagnostic measurements were taken and the cell returned to an output power about 15% lower than previously obtained. The power output stabilized at a value 35% lower than the steady-state value. The optimum cesium reservoir temperature increased 25°C prior to the performance change at 15,229 hours. From that time to 24,056 hours, it increased another 20°.

A-172

The slow drop in performance during the first 15,229 hours of life testing has been attributed primarily to emissivity changes due to carbon on the collector. Changes in emitter and collector work function probably also contributed to the loss. The rise in optimum cesium reservoir temperature may have been the result of emitter work function changes, changes in the graphite reservoir sorption characteristics, or losses of cesium. A loss of ~ 110 mg of cesium is necessary to explain the full optimum temperature rise if it were the sole mechanism. In reality a combination of these mechanisms probably occurred.

The sudden loss of power at 15,229 hours was due to emitter contamination and a consequent work function change. The work function change following startup was sensitive to the cesium pressure history during the startup, indicating a very small amount of contaminant was present on the emitter and its loss rate to the collector was affected by cesium pressure. The most probable contaminant which explains the performance loss is titanium (from the Cu-Ti final closure braze).

Post-Operational Examination (Figs. A-102 to A-105)

A cross sectional view of OC-A4 is shown in Fig. A-102. Following mapping with the liquid reservoir, the cesium was transferred into the graphite and the liquid reservoir assembly was pinched off in the secondary pinch off area.

During disassembly the cell easily separated at the final closure braze joint where the collector meets the insulator seal. Much of the braze material appears to have been lost from the joint.

The collector and the I.D. of the insulator seal were found to be coated with a layer of carbon or carbide (Fig. A-103, 104). The coating over the collector appeared to be uniform and £ 5m thick, suggesting carbon transport probably occurred in a gaseous phase perhaps in a CO-CO₂ cycle rather than by surface migration. No coating was found on the emitter (Fig. A-103). An emission spectrographic analysis of the emitter surface showed the presence of niobium (0.1m g/cm²), aluminum (0.06m g/cm²) and silicon in some spots (0.1m g/cm²).

A-173

A similar analysis of the collector revealed copper (4.0m g/cm²), silicon (0.4m g/cm²), silver (0.2m g/cm²), titanium (0.1m g/cm²), iron (0.08m g/cm²), and aluminum (0.08m g/cm²). The presence of copper and titanium confirm the final closure braze as a source of contamination. The small amounts of other impurities, particularly silicon, were probably introduced with the cesium and graphite. Microhardness measurements indicate no carbonization of the emitter or collector deeper than 50m had occurred. Similar results were found for the niobium pedestal which supported the graphite reservoir, although a carbide layer ~ 3m thick is visible (Fig. A-105).

The graphite (Fig. A-104) sorption reservoir appeared to be in good condition, although it had partially delaminated along the layers. Such delaminations have been observed in the past shortly following cesium loading to levels equivalent to those in A4’s reservoir, so delamination was probably not responsible for the changes during the test life.

Mark VIIA OC-A4 was operated a total of 26,296 hours, demonstrating the integrity of the envelope for that period. The use of a graphite sorption reservoir was demonstrated. A full performance map of the cell was obtained.

Post operational examination and test diagnostics show carbon was transported to the collector from the sorption reservoir, probably in a gaseous phase. Performance losses due to this were probably less than 5%. Carbide layer formation on the collector was £ 5m thick. The niobium pedestal used to support the graphite reservoir had a carbide layer ~ 3m thick. The significant thermionic performance loss observed after 15,229 hours of life testing appears to have been associated with emitter contamination by the final closure braze material rather than the graphite reservoir.

Thus, the weakest portion of the cell was the final closure braze. TFE designs are of all welded construction. The use of a graphite sorption reservoir for periods as long as 20,000 hours would appear feasible, and a test of its affect on performance in an all welded system would be useful.

A-174