Chapter 3. HTPEM operation with reformed fuel
when operated with reformed methanol in the investigated range of temper-atures. The hydration and dehydration phenomena is more pronounced at 180 ◦C operation when dry hydrogen is replaced with wet hydrogen having 15 % H2O vapour. The investigation of methanol and water suggests that the presence of water mitigates some degradation effects in the presence of methanol. The presence of methanol may lead to the following reaction as shown in Eqn. 3.1 or form esters with PA as reported in [132]. The ester formation would lead to reduced H3PO4 available for the proton conduc-tivity. However, the presence of H2O improves the mobility of H3PO4 and thereby the proton conductivity. Thus, the overall proton conductivity is not compromised.
CH3OH →CO+4 H++4 e− (3.1) The CO generated could poison the catalyst to some extent. However, in the presence of water the following reaction shown by Eqn.3.2is highly possible as also reported in [54,55]. The converted CO2 has only dilution effect and can be easily removed from the cell.
H2O →H++OH−
CO+OH− →CO2+H++e− (3.2) Thus, it was concluded that direct integration of reformed methanol with no intermediate purification system is viable with HT-PEM fuel cells. A CH3OH slip of < 3% does not affect the performance as the water in the reformed fuel mitigates some of the effects of methanol slip.This could serve as an important finding to decide the operating temperature which would be optimal for the reformer with lower CO.
3.2. Different break-in procedures
Thus, to address this cost factor different break-in processes were investi-gated to speed up the break-in process. First, break-in was done for different duration in order to understand whether the break-in process could be elim-inated. Secondly, break-in with reformed fuel instead of pure hydrogen was investigated. The analysis was done to estimate the durability and compare the performance. The MEAs used were supplied by Serenergy and are simi-lar to BASF MEAs.
The break-in time was divide into three: No break-in (0 h), Intermediate break-in (30 & 50 h) and normal break-in (100 h). The cells were cycled between 0.2 A cm−2 and 0.6 A cm−2after the respective break-in times. The cycling is carried out based on two assumptions:
• Break-in involves the redistribution of PA uniformly across the MEA
• PA migration and diffusion is a function of current density
The voltage over time for cells with different break-in time is shown in Fig.3.3. The comparison shows minimal difference in the voltage inventory over time for all the cells and it is confirmed by two sets of tests (Test 1 and Test 2). Test 2 was a repetition of test 1 with only one change in the intermediate break-in time. The intermediate time for test 1 was 30 h and it was 50 h for test 2.
0 100 200 300 400 500 600 700 800 900
0.4 0.5 0.6
0.7 Test 1 (break-in time- 0, 30 and 100 h)
0 100 200 300 400 500 600 700 800
Time [h]
0.4 0.5 0.6
0.7 Test2 (break-in time- 0, 50 and 100 h)
Voltage [V]
Fig. 3.3:Voltage comparison with different break-in time, Source: Paper B.
Chapter 3. HTPEM operation with reformed fuel
Table 3.2:Test description with break-in time and corresponding degradation at 0.2 A cm−2and 0.6 A cm−2.
Test 1 with 0h, 30hand 100h Cell name Break-in
time
@ 0.2A cm−2 @ 0.6A cm−2 Operation time
- [h] [µV h−1] [µV h−1] [h]
Cell 1 0 -26 -39 730
Cell 2 30 -13 -40 803
Cell 3 100 -31 -52 791
Test 2 with 0h, 50hand 100hbreak-in
Cell 4 0 -33 -30 711
Cell 5 50 -42 -53 710
Cell 6 100 -44 -42 710
The test time and corresponding degradation rates at 0.2 A cm−2 and 0.6 A cm−2 were calculated and is shown in Table3.2. The different degra-dation rates suggest that the break-in time does not affect the performance and the long term durability. This could be related to load cycling results re-distribution of acid and is not related to constant current operation. Similar, degradation’s were also reported in other studies under load cycling oper-ation [133–135]. The cause of degradoper-ation under load cycling was reported to be related to hydration and dehydration of PA. One improves the proton conductivity while the other degrades.
To understand the different break-in operational strategy the EIS recorded over time was analyzed. The EIS data was fitted with an equivalent circuit model and the different resistances calculated. A significant variation among the different MEAs was a problem to carry out comparisons. Therefore, the resistances were compared based on the inventory over time and not the absolute numbers.
The calculated ohmic resistances are shown in Fig.3.4. A close look into the ohmic resistance suggests that the cell 3 and cell 6 have similar trend. An increase is seen in the beginning and then around ≈300 h the become less steeper. The no break-in cells 1 and 4 and intermediate break-in cells 2 and 5 do not show similarities in the change of resistance over time.
The analysis suggests that the performance in the present study is not af-fected by the different break-in times. However, the uniformity of cell ohmic resistance is ensured by the normal break-in process. Further, investigation with minimal variations in the MEA is required to suggest the optimal break-in procedure.
The test 3 was carried out with reformed fuel having the following fuel composition, H2- 64.7 %, H2O- 12 %, CO2- 21.3 %, CH3OH- 2 %. The test
3.2. Different break-in procedures
0 200 400 600 800 1000
Time [h]
0.1 0.11 0.12 0.13 0.14 0.15 0.16
Resistance [ .cm
-2]
cell1 cell2 cell3 cell4 cell5 cell6
Fig. 3.4: Comparison of fitted ohmic resistance over time for different break-in times, Source:
Paper B.
was performed based on Sahlin et al. [1] work that reported a low CO less than 0.5 % when the methanol slip is 2 %. As discussed in Paper A and also by others [64], HT-PEMFC performance is not significantly affected with a methanol slip of 3 %.
Thus, two cells were tested, Cell 1 with pure hydrogen break-in followed by reformed fuel and Cell 2 break-in was using reformed methanol compo-sition as mentioned before. The voltage inventory over time shows a faster degrading cell when break-in was carried out using reformed fuel, Fig.3.5.
The different sections of the voltage inventory were analyzed to understand the cause.
In Table 3.3, the voltage inventory at different sections of the voltage pro-file is shown. The first section is the break-in process, where we see a positive effect on voltage with pure hydrogen as reported also in the literature. How-ever, the break-in with methanol in the fuel resulted in a decrease in voltage by 30 mV over 100 h. The degradation is attributed to some acid loss due to the water in the feed. This was further confirmed from EIS data recorded before and after the break-in as reported in Paper B. The interesting
transi-Chapter 3. HTPEM operation with reformed fuel
0 100 200 300 400 500 600 700
Time (hours) 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Voltage
Reformed Methanol breakin Pure Hydrogen Breakin
0 50 100
0.4 0.5 0.6 0.7
Fig. 3.5: Voltage comparison of break-in with reformed fuel and break-in with pure hydrogen operation, Source: Paper B.
tion was observed when the current density was changed from 0.2 A cm−2to 0.6 A cm−2, the voltage change with current density was lower for the cell 2 compared to Cell 1. This was attributed to sudden change in the fuel compo-sition in Cell 1 from pure hydrogen to reformed fuel. The degradation was much higher for Cell 2 compared to Cell 1 as seen in Table3.3.
To understand the cause of higher degradation in Cell 2, the EIS recorded over time was analyzed using equivalent circuit modelling and a significant change in the low frequency resistance and ohmic resistance is seen in Fig.3.6.
The ohmic resistance as explained in Chapter 2 is related to the membrane re-sistance and low frequency rere-sistance is a attributed to mass transport issues.
Thus, it is evident that reformed fuel operation enhances the acid loss from the membrane which reduces the proton conductivity. The acid which gets lost over time is assumed to be covering the GDL pore and some are trapped in the flow channel, which enhances the mass transport issues. The
3.2. Different break-in procedures
Table 3.3:The voltage changes and degradation rate calculated from the voltage profile
Cell No.
Voltage change during break-in
Voltage drop on current change
Degradation
@0.6A cm−2
[mV] [mV] [µV h−1]
Cell 1 +60 15 158
Cell 2 -30 26 897
0 50 100 150 200 250 300 350
Time [h]
0 0.1 0.2 0.3 0.4 0.5 0.6
Resistance [ .cm
2]
ROHM
RHF
RIF
RLF
Fig. 3.6:Comparison of different resistances for cells operated on reformed fuel during break-in, Source: Paper B
tests suggests that break-in times do not affect the performance and degra-dation significantly provided the cells are cycled between 0.2 A cm−2 and 0.6 A cm−2. However, the uniform change in resistance could be possible only after normal break-in. The break-in with reformed fuel is not suitable for longer durability. Thus, it is suggested to avoid supplying H2O into the cell till the acid redistribution is ensured.
Chapter 3. HTPEM operation with reformed fuel