3.1 Coating observations
The adopted dip-coating method resulted in a relatively thinner (~150-200 µm) wet-film layer over the substrate. As depicted in Fig.2, a short ramp-up time (2.5 h) for heat-treatment (step (b)) resulted in visible cracks along the coated length with formation of surface blisters, which can be attributed to a larger differential thermal expansion and sudden dehydration of coated suspension, respectively.
Although a longer ramp-up time (~ 10 h to 12 h) greatly minimized the cracks and blisters during first heat-cure, elimination of very fine cracks could not be achieved. Hence, additional coating layers deemed necessary repeating the application steps until no visible cracks were identified. The IR at RT for one of the probes, prepared with a single-dip coat cycle followed by full heat-curing, was found to be exceptionally high providing an insulation leakage current of < 1 nA. In view of highly brittle nature of Al2O3, preliminary non-destructive investigations were carried out using a calibrated portable USB microscope for estimation of coating thicknesses using average diametrical measurements as shown in Fig.3 (a) and Fig.3 (b). Detailed average coating thickness estimations across the cross-section (destructive examinations) using SEM for the tested probes are presented under dedicated section on metallographic investigations. Details of the probes tested for compatibility with PbLi are summarized in Table-I below. All weight measurements were done using a calibrated precision weighing balance (Sartorius make, CPA225D) with a resolution of 10 µg.
Table-I: Details of fabricated probes tested under test campaign-1 and test campaign-2
Test campaign
|
Probe identification
|
Coated length (mm)
|
Weight (g)
|
1
|
P1
|
~85
|
8.57521 ± 0.00001
|
P2
|
~65
|
7.45867 ± 0.00001
|
2
|
P3
|
~60
|
7.66220 ± 0.00001
|
3.2 Performance validations in liquid PbLi environment
3.2.1 Test campaign-1
Temperature variations with time duration details for the test campaign-1 are depicted in Fig.4. The red-markers (circle) on the plot represent instances of in-situ high-voltage IR tests. Measured IR values with time and temperature are mentioned in Table-II. All IR measurements were taken 60 s after high-voltage application to minimize the effects of capacitive charging current and dielectric absorption current [55]. A decrease in IR was observed for both P1 and P2 with an increase in temperature from 300°C to 350°C, an expected temperature-based derating phenomenon in insulators. However, for P2, during 260 h in static PbLi at 300°C, an increase in IR value was observed which could be accounted for gradual heat-curing of uppermost untreated coat layer over long operational durations. From the measured performance data, immediate observable advantages of the adopted coating method include selective coating/masking on difficult substrate geometries including the internals of pipe-sections and complex manifolds, easy application without exposure of substrate to very high temperatures as required in previous works [11, 14, 24-29, 35, 37, 49, 51-52] and a highly-dense compact coating with complete coverage conforming to electrical-insulation requirements in harsh LM environment of interest.
Table-II: IR measurements for P1 and P2 in PbLi environment during test campaign-1
Time (h)
(since t=0)
|
Temperature
(°C)
|
DC test voltage
(V)
|
IR for Probe P1
(x 109 Ω)
|
IR for Probe P2
(x 108 Ω)
|
72
|
300
|
100
|
> 6
|
> 1
|
260
|
300
|
100
|
> 6
|
> 3
|
380
|
350
|
100
|
> 1
|
> 1
|
525
|
350
|
100
|
> 1
|
> 0.9
|
After exposure of over 520 h in PbLi, thermal cycling within PbLi environment was conducted to establish insulation integrity during possible temperature gradients in a LM circuit. Considering the normal operational temperature of PbLi circulations facilities, 300°C was taken as the base temperature for test campaign-1 while IR at the start of last thermal cycle was taken as the base IR value. To gain higher accuracy in measurements, a digital IR tester (Make: Fluke, Model: 1550C) with ±5% uncertainty was used. Measured IR values are reported under Table-III and IR derating trends are shown in Fig.5. In addition to temperature effect, observed IR derating trends also include the effect of rising PbLi level surrounding the coated probes. However, for the current set-up, overall level rise per 10°C change is estimated less than 1 mm. In consideration to the chemical stability of Al2O3, temperature is assumed to be the prominent derating factor for a conservative estimation of IR derating.
The above assumption is supported by the rising trend of IR values towards the base IR value as mentioned in the last row of Table-III, taken just after achieving 300°C at the surface of tank, corroborating thermal and chemical stability of Al2O3 insulation in high-temperature PbLi environment. It should be noted that these derating factors have been estimated post-exposure of the insulation to high-temperature corrosive PbLi environment for over 690 h. In view of similarity of the fabricated probes with electrical cables, observed temperature-based IR derating is much better [55].
Table-III: Temperature derating factors for IR during last thermal cycle of test campaign-1
Temperature
(°C)
|
DC test voltage
(V)
|
IR for Probe P1 (x 109 Ω)
|
Temperature derating for P1
|
IR for Probe P2 (x 108 Ω)
|
Temperature derating for P2
|
300 (base)
|
275
|
6.18
|
1.00
|
3.15
|
1.00
|
310
|
275
|
3.75
|
0.61
|
2.09
|
0.66
|
320
|
275
|
2.42
|
0.39
|
1.60
|
0.51
|
330
|
275
|
1.73
|
0.28
|
1.35
|
0.43
|
340
|
275
|
1.20
|
0.19
|
1.08
|
0.34
|
350
|
275
|
0.822
|
0.13
|
0.835
|
0.27
|
300
|
275
|
6.08
|
0.98
|
2.39
|
0.76
|
Fig.6 presents images of P1 and P2 after removal and post-chemical cleaning, depicting only partial cleaning achieved even after sufficient immersion time in a 1:1:1 (volume ratio) solution of acetic acid, hydrogen-peroxide and ethyl alcohol. Such a change in appearance was not observed during an earlier short-duration experimental study, where complete cleaning could be achieved for one of the sample probes exposed to PbLi at 300ºC for continuous 250 h. Observed discoloration in present case could be attributed to partial ingress of PbLi in the coated layer over longer exposure durations. After chemical cleaning, no weight loss was observed for P1 while a normalized weight loss of ~ 1.56 mg/cm2 over the coated surface area was observed for P2. This loss could be primarily accounted for dehydration of topmost untreated coat layer of P2 during facility heating and could also include partial thinning of the coating section immersed in PbLi. However, in view of no variation in the order of magnitude for IR within uncertainty limits at relevant temperatures (refer Table-II and Table-III), dehydration could be ascribed as the prominent cause for observed weight loss.
3.2.2 Test campaign-2
Electrical-insulation performance achieved for probe P2 (test campaign-1) through gradual heat-curing of uppermost layer during system operation seems promising to achieve coating on substrates with sharp bends, multiple/parallel internal flow sections, complex geometries etc. To establish the repeatability and reliability of this coating technique, a more rigorous test campaign-2 was conducted using probe P3 as per details shown in Fig.7, while the measured IR values with time and temperature are reported under Table-IV. The variation in IR followed a pattern similar to that observed during test campaign-1, accounting for gradual heat-curing of the untreated coat layer. After continuous high-temperature PbLi exposure of over 1340 h, temperature based derating factors were estimated using a base temperature of 350°C as reported under Table-V and Fig.8. For insulation health estimations, IR readings were taken at 60 s and 120 s after application of high voltage for all the cases. As observed from Table-V, at any given temperature, a rising value of IR from 60 s to 120 s substantiates the health of insulation. A comparison of data from Table-IV and Table-V, alongwith consideration to the fact that capacitive charging current normally decays within 30 seconds, signifies that dielectric polarization is taking more time which is also indicative of insulation health [55]. However, as the order of magnitude for IR readings at 60 s and 120 s remains the same, this effect is not of any significant implication for practical purposes of achieving electrical-insulation in a LM environment.
Fig.8 also affirms that temperature deratings estimated using IR data at 120 s are higher than corresponding deratings using IR data at 60 s. However, degradation in the IR over a temperature gradient of 50ºC is relatively low compared to both the cases under test campaign-1. Also, the IR value essentially did not deteriorate between 350°C-370°C in contrast to the pattern observed under test campaign-1. These improvements could all be ascribed to long duration exposure near the required curing temperature of 427°C. The IR after cool down to the base temperature was observed higher than the corresponding value at start of the cycle, which is assumed to be the result of an effective dielectric polarization due to repetitive IR tests at high voltage leading to a polarization current component to zero.
This assumption is also in coherence to the nearly identical IR values observed at 60 s and 120 s after a few IR measurements (refer Table-V). However, this needs to be investigated in more detail with similar tests. The IR at base temperature of 350ºC measured at the end of last thermal cycle (1350 h of exposure) tends towards the measured IR observed at same temperature between 500 h - 800 h of exposure. Similarly, the measured IR values at 300°C (1367 h of PbLi exposure) were 1.89 x 108 Ω and 3.32 x 108 Ω after 60 s and 120 s, respectively. All these observations corroborate high integrity and healthiness of electrical-insulation over complete test duration.
Table-IV: IR measurements for P3 in PbLi environment during test campaign-2
Time (h)
(since t=0)
|
Temperature
(°C)
|
DC test voltage
(V)
|
IR for Probe P3
(x 108 Ω)
|
72
|
300
|
275
|
2.90
|
261
|
300
|
275
|
4.57
|
338
|
350
|
275
|
1.30
|
526
|
350
|
275
|
2.18
|
766
|
350
|
275
|
2.99
|
838
|
350
|
275
|
2.89
|
917
|
400
|
275
|
0.583
|
1106
|
400
|
275
|
0.767
|
1109
|
350
|
275
|
2.99
|
Table-V: Temperature derating factors for IR during last thermal cycle of test campaign-2
Temperature (°C)
|
DC test voltage
(V)
|
IR for P3 after 60 s
(x 108 Ω)
|
Temperature derating for P3 (as per IR data at 60 s)
|
IR for P3 after 120 s
(x 108 Ω)
|
Temperature derating for P3 (as per IR data at 120 s)
|
350 (base)
|
275
|
0.945
|
1.00
|
1.29
|
1.00
|
360
|
275
|
1.06
|
1.12
|
1.39
|
1.08
|
370
|
275
|
1.26
|
1.33
|
1.22
|
0.95
|
380
|
275
|
0.804
|
0.85
|
0.98
|
0.76
|
390
|
275
|
0.735
|
0.78
|
0.821
|
0.64
|
400
|
275
|
0.638
|
0.68
|
0.768
|
0.60
|
350
|
275
|
2.28
|
2.41
|
2.37
|
1.84
|
Image of the probe after removal from the tank and after chemical cleaning is shown in Fig.9. The PbLi exposed coated section appeared blackish after cleaning, which is in close agreement to the findings reported in [56]. Fig.9 shows major visible cracks just after completion of chemical cleaning. Spontaneous chipping of a few coat-sections was noticed during chemical cleaning followed by detachment of almost complete coating during subsequent drying at RT. Therefore, weight measurements for P3 post-exposure could not be performed. A similar observation with an AIN-BN sample immersed in Li was reported [21], where Li covered surface broke into pieces during subsequent cleaning in water, accounting for induced stresses due to reaction between water and Li. However, in present study, Li activity is negligible owing to the low weight percentage of Li (0.62% to 0.68%). Additionally, neither such observations have been reported for alumina coated samples in previous corrosion studies with PbLi nor observed during test campign-1 for probe P2, which had a similar fabrication method and PbLi immersion process as that of P3. The possibility of crack-generation/initiation during thermal cycling cannot be completely ruled out owing to large difference in CTEs. However, the unexposed top-section of coated probe (~20 mm), which is also expected to experience similar thermal gradients, remained well-adhered to the substrate with no presence of cracks or indication of brittleness. A crack-generation within LM would have led to a sharp decline in the measured IR, which was not observed. Therefore, in view of IR integrity observed from Table-IV and Table-V, it is inferred that the through-crack generation occurred post-removal of the probe from LM.
One of the possibilities could be the cracking due to compressive stresses generated over LM exposed portion of coated substrate because of temperature-dependent volumetric contraction of the surface-adhered PbLi during cool-down period. Under present experimental constraints, such compressive stresses are inevitable. In contrast, during the flowing PbLi conditions in a coolant/breeder system of a fusion reactor, LM will be surrounded by ceramic insulating coatings, leading to an absence of such compressive stresses. However, as the coating cracks were not observed during test campaign-1, the contribution of PbLi exposure duration and temperature is of interest for further compatibility studies. In view of similarity of the fabricated probes to electrical cables, volumetric electrical-resistivity was estimated for the enveloping cases using IR relation for an electrical cable. It should be noted that in the present experimental study, a conservative estimation of volumetric electrical-resistivity has been performed taking the tip-portion (blob) as a perfect electrical-insulation in view of its relative higher thickness as compared to the coating thickness along the length. Therefore, complete insulation leakage current is assumed to flow only through thickness of coating along the probe length, leading to an under-estimation of calculated volumetric electrical-resistivity for the probe(s) under reported operating conditions. Calculated volumetric electrical-resistivity remained of the order of 109 – 1011 Ω-cm between 300°C-400°C. The coating resistance, measured as a product of volumetric electrical-resistivity and insulation thickness, remained of the order of 104 – 106 Ω-m2, many orders of magnitude higher than required for successful operation of LM based coolant circuits with acceptable MHD pressure drops [2, 24, 57]. Calculations for volumetric electrical-resistivity and coating resistance are provided under Appendix-A. This further substantiates the applicability of Al2O3 coatings for high-temperature PbLi coolant/breeder circuits in fusion power plants. It can be observed that the volumetric electrical-resistivity for fully treated probe P1 is one order of magnitude higher than that of P2 at same temperature, an effect that can be attributed to higher degree of compaction achieved with full heat-curing. However, sufficient volumetric electrical-resistivity achieved and ease of coat-formations using gradual heat-curing through system operations makes the second method more promising and attractive.