Ledda
/
MA_Platteau
1
0
Fork 0
Code Issues 0 Pull-Requests 0 Releases 0 Wiki Aktivität
Quellcode durchsuchen

update

results-discussion
Ledda vor 2 Monaten
Ursprung
43a14b0a80
Commit
70cbdd2888
1 geänderte Dateien mit 525 neuen und 12 gelöschten Zeilen
Geteilte Ansicht
Diff-Optionen
Statistiken anzeigen Patch-Datei herunterladen Vergleichs-Datei herunterladen
  1. +525
    -12
      Content/Results and Discussion.tex

+ 525
- 12
Content/Results and Discussion.tex Datei anzeigen

@@ -1,34 +1,547 @@
\chapter{Results and Discussion}
\label{chap:Ergebnisse und Diskussion}

Chapter \ref{chap:Ergebnisse und Diskussion} will present the findings of both in-situ and ex-situ investigations, along with the discussion of the results. First the in-situ results will be presented starting with the preliminary investigations in section \ref{subsec:4_Prelim} which helped develop the corrosion endurance run. The results from the corrosion endurance run as well as the high temperature endurance run will be presented in \ref{sec:4_Corrosion} and \ref{sec:4_High Temp}. The final section, \ref{sec: Ex-Situ}, provides the results of the in-situ methods will be presented and discussed.


\section{Material Characterization of SS316L}
\section{Preliminary Investigation}
\label{subsec:4_Prelim}

\subsubsection{Corrosion Parameters}
Before starting the endurance run first a preliminary investigation was conducted in order to find the optimal operating conditions for the corrosion endurance run. The table \ref{tab:versuche} found in the appendix \ref{sec:A_1} provides a chronological overview of the testing activities conducted. Although this investigations were planed on a 10 cell stack made out of type 2 cells (stainless steel 316L BP) the plan was changed due to the limited stock of these cells after the first stack broke down due to several hard shut downs and problems with the test bench. Therefore a 4 cell stack of the type 1 cells (Ti-C) was used to conduct the preliminary investigations and afterwards test the other test bench before changing to the type 2 cells again and starting the endurance run.
\subsection{pH and Electrical Conductivity Measurement}

As explained in the chapter \ref{sec: M_Preliminary} the measured product water was extracted from the condensator bottles located at the cathode and anode outlets. The results from the extracted product water are presented in the following table \ref{tab:3_ER}.

\begin{table}[h]
\centering
\begin{tabular}{cccc}
\hline
Position & Temperature [°C]&pH & Electrical Conductivity [$\mu \mathrm{S} / \mathrm{cm}$]\\
\hline
Cathode &90 & 6 & 1,36\\
Anode &90 & 5,73 & 2,07\\
Cathode &75 & 5,83 & 2,29\\
Anode &75 & 5,94 & 2,93\\
Cathode& 60 & 5,51 & 12,75\\
Anode &60 & 5,69 & 6,99\\
\end{tabular}
\caption{Measurement of the pH and electrical conductivity of the product water in the preliminary investigations.}
\label{tab:3_ER}
\end{table}

The high temperature and high humidity test (HTHH) conducted at 90°C differed from the other two tests since the water production was higher than expected and the test stopped after 45 minutes of voltage cycling between 15s at 0,85V and 10s at 0,6V while the other two tests cycled for 2 hours before shutting down. The measured electrical conductivity of the test was lower than the one of the other two test with a the lowest measurement at the cathode with 1,36 $\mu \mathrm{S} / \mathrm{cm}$. On the other hand the pH measured in the product water of the cathode was the highest with a value of 6. In all the other tests the measured pH at the cathode was lower than in the anode.

At a medium temperature of 75°C and high humidity (MTHH) the ph of the cathode was lower than in the HTHH. The electrical conductivity was higher at the anode and cathode at 75°C than at 90°C. The lowest pH of the preliminary investigations was achieved in the test with low temperature and high humidity (LTHH). The cathode side presented a pH of 5,51 and also the highest electrical conductivity with a value of 12,75 $\mu \mathrm{S} / \mathrm{cm}$ after two hours of voltage cycling at a cell and cooling temperature of 60°C.

The results measured at a high pH is in accordance with the study performed by Abdullah et al. since an increase of the pH with an increase of the temperature. Although the increase of the temperature from 50°C to 90°C at a relative humidity of 35\% showed a pH increase from 2 to 5 \citep{107_abdullah2008effect}. The increase of the pH from 5,51 to 6 at the cathode by raising the cell temperature from 60°C to 90°C although noticeable is still much smaller than the one found in the study performed by Abdullah et al.

The difference in the pH measured in the preliminary investigations to the one found in literature could be attributed to two main factors. The first one being the duration of the test. Since the conducted tests had a relative short operating time with a 2 hours of voltage cycling compared to the minimum of 30 hours of Abdullah et al. the membrane had little to no degradation resulting in a much lower amount of H$_2$O$_2$ produced as a result of membrane degradation \citep{wallnofer2024main, 107_abdullah2008effect}. The second reason for the higher pH could be the use of the type 1 cell made out of Ti-C. The corrosion resistance of Ti is much higher than the one of stainless steel, consequently less particles will leave the BP because of corrosion \citep{Corr_Mat_wang2010electrochemical}. Since the fenton mechanism is catalysed by metal ions such as Fe$^{2+}$ the membrane degradation caused by this reaction will be less when using Ti BPs \citep{elferjani_coupling_2021}.

\section{Corrosion Endurance Run}
\label{sec:4_Corrosion}

Following the results of the preliminary investigation the decision was made to lower the temperature from the endurance run from 90 to 66°C in order to produce a more acid product water and therefore reinforce the corrosion mechanism. For both endurance runs cells type 2 were used since the type 1 cells made out of Ti-C have a higher corrosion resistance. After the test Bench was successfully tested with a 4 cell stack made out of type 1 cells to avoid any damage to the 4 cell stack made out of type 2 (that could have been caused by a malfunction of a test bench) the corrosion endurance run could be started. The results of the in in-situ characterization of the cells will be presented in a more detailed way in the following section.

\subsection{Polarization Curves at 60°C}
\label{subsec: 60_PolCurve}

After the test specimen was activated by repeating the 80°C polarization curve four times in a row until the performance of the cell was optimal for the endurance run to start. Before beginning the endurance run the cell underwent a beginning-of-life (BoL) characterization which included the polarization curves at 60°C and 80°C. The results of the BoL characterization as well as those after 12500 , 37500 and 81000 cycles are shown in figure\ref{fig:60_Pol}. It is also worth mentioning, that the stoichiometry of cathode and anode was constant at $\lambda_{air}$ = 2 and $\lambda_{H_2}$ = 1,5 throughout the 60°C and 80°C polarization curves after the current density was increased above 0,3 A/cm$^2$ and therefore the minimum volume flow of the test bench was also surpassed (see appendix \ref{sec:A_Sto}).

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/60_PolCurve_Korrosion.pdf}
\caption{60°C polarization curve at BoL after 12500 VC, 37500 VC and 81000 VC.}
\label{fig:60_Pol}
\end{figure}

At low current densities the BoL curve shows the highest Voltage and therefore the lowest activation polarization. The second best curve at low current densities is the polarization curve after 81000 VC at the end of the endurance run. When the current density is increased to over 0,6 A/cm$^2$ the the curve after 81000 VC shows a higher loss compared to the other ones. At high current dentities the curve after 81000 VC presents higher losses than the other curves. The degradation of the test specimen over the period of the 81000 VC is made clear since the voltage of the curves decreases after each characterization with the BoL curve being the one with the least losses followed by the 12500 VC as well as the 37500 VC and the highest loss at high current densities after the last characterization at 81000 VC.


\subsubsection{High-Frequency Resistance (HFR)}

The high-frequency resistance (HFR) was also measured during the polarization curves since it gives an insight on the performance of the cell. A lower HFR can indicate a higher performance as well as better hydration of the membrane \citep{108_lin2021prediction}. The results of the measurements during the 60°C polarization curves can be found in the following figure \ref{fig:60_HFR}.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/60_HFR.pdf}
\caption{HFR during the 60°C polarization curve at the BoL, after 12500 VC, 37500 VC and 81000 VC.}
\label{fig:60_HFR}
\end{figure}

The curves of the BoL and after 12500 cycles show a peak in the HFR at the beginning of the test and at very low current densities (under 0,2 A/cm$^2$). Moreover, the after 37500 cycles the HFR is still low at very low current densities when compared to the curve after 81000 VC. When current densities are increased in the polarization curves the HFR of the curves after 12500 VC and after 37500 VC decreases while the curve after 81000 VC shows an very stable and low value of around 50,5 $m\Omega\cdot\text{cm}^2$ per cell.

The peaks at the beginning are likely to be caused by the high cathode and anode stoichiometry at low current densities until the volume flow can be increased from the minimum (5,6 Nl/min at anode and 11,36 Nl/min cathode until 0,3 A/cm$^2$) with the increasing current densities. The initial increase in HFR was also detected by Ma et al. until the cell reaches a final stable HFR \citep{109_ma2022effect}. This can also be seen in the last polarization curve after 81000 VC when a stable HFR value of around 50,5 is reached and maintained throughout the curve.


\subsubsection{60°C Polarization Curves of the Cells after 81000 VC}

Since the voltage decrease was the highest in the EoL polarization curve after 81000 VC the voltage and polarization curves was also analysed to be able to get a better overview of the degradation in the cells. The results of the 60°C polarization curve of each of the 4 cells in the stack is presented in Figure \ref{fig:60_Cells}.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/60_Cells.pdf}
\caption{60°C polarization curves of the 4 cells after 81000 VC.}
\label{fig:60_Cells}
\end{figure}

At very low current densities cells 1-4 have roughly the same voltage loss. As current density increases the voltage loss of the cells increases as well with the first cell showing the highest voltage at 0,6V and the others decreasing in numerical order until the lowest voltage detected in the cell 4 with 0,53V. Due to the significant difference between the voltage in cell 1 and the cell 4 at high current densities both cells will be analysed in greater detail in the ex-situ investigations in section \ref{sec: Ex-Situ}.

The higher RH of the 60°C polarization curve could have caused a high concentration polarization at high current densities. Since the RH is at 50,6 \% the higher water content could have limited the transport of the reactants to the reaction site in the cell 4. Another reason could have been the corrosion of the plates. If the MEA at the cell 4 was degraded due to the Fe$^{2+}$ catalysing the fenton reaction when set free due to the corrosion of the BP leading the cell 4 to a lower performance than the other cells \citep{elferjani_coupling_2021, frensch2019impact}.

\subsection{Polarization Curves at 80°C}
\label{subsec: 80_PolCurve}

After the 60° polarization curve the temperature of the cell was increased to 80°C by increasing the coolant temperature ($T_{coolant,in}$), then the dew point temperatures ($T_{dp,A}$, $T_{dp,C}$) were increased to 53°C which results in a RH of 30,1\% for cathode and anode. The results from the 80°C polarization curves can be seen in Figure \ref{fig:80_Pol}.


\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/80_PolCurve_Korrosion.pdf}
\caption{80°C polarization curve at BoL after 12500 VC, 37500 VC and 81000 VC.}
\label{fig:80_Pol}
\end{figure}


The 80°C polarization curve show almost no signs of degradation of the cells. The voltage at high current densities of 2,2 A/cm$^2$ are almost identical for all the curves at 0,59V which is the same as the value for the BoL voltage for the 60°C curve. At a low current density of 0,1 A/cm$^2$ the BoL curve is slightly better than the other ones. It is also noticeable, that the 81000VC curve is better than the 37500 VC and the 12500 VC curve has the lowest value at this current density. At a higher current density of 1,4 - 1,6 A/cm$^2$ the the BoL curve has a slightly higher average voltage than the other curves.

Overall the 80°C polarization curves present much lower degradation than the 60°C polarization curve. One reason for this could be the lower RH value of 30,1\% of this polarization curve when compared to the RH 50,6\% which can be found in the 60°C curve. The higher humidity of the 60°C curve could lead to a much higher activation polarization which is heavily influenced by the hydration of the membrane \citep{jouin2016}. The ohmic loss at higher current densities is very similar for all the curves since they are almost parallel to each other until the end.

In addition, the effects of the mass transport losses that could have made a great difference at very high current densities cannot be significantly detected in the figures. A reason for this could be the high stoichiometry used in both cathode and anode. Since there is more reactant than needed at the active sites even the a high reaction rate at high current densities will not consume all the reactants and limit the reaction like it would at a lower stoichiometry
\citep{Loss_mazzeo2024assessing}.



\subsubsection{High-Frequency Resistance (HFR)}

The HFR measured during the 80° polarization curves is presented in Figure \ref{fig:80_HFR}. At very low current densities (under 0,2 A/cm$^2$) the BoL, 12500 and 37500 curves show a peak of 160 $m\Omega\cdot\text{cm}^2$ per cell. For the 81000 VC curve the peak is not as high with a start value of around 62 $m\Omega\cdot\text{cm}^2$ per cell.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/80_HFR.pdf}
\caption{HFR during the 80°C polarization curve at the BoL, after 12500 VC, 37500 VC and 81000 VC.}
\label{fig:80_HFR}
\end{figure}

The 81000 VC curve is the first one to reach a stable HFR at 51 $m\Omega\cdot\text{cm}^2$ per cell. The 37500 VC and the 12500 VC curves reach a stable HFR at a current density of 0,3 A/cm$^2$. At a current density of 2,2 A/cm$^2$ the BoL curve has the highest HFR followed by the 37500 VC and the 12500 VC curves.

Just like in the 60°C polarization curve the lowest HFR was measured in the 81000 VC curve. A reason for this could be that the membrane has reached an optimal level of humidity to reach a lower HFR at 51. In the investigations by Ma et al. the peak at the beginning was also measured and then quickly reached its stable HFR \citep{109_ma2022effect}.

\subsubsection{80°C Polarization Curves of the Cells after 81000 VC}

Figure \ref{fig:80_Cells} displays the 80°C polarization curve after 81000 VC of each cell separately. Cell 1 is again the best performing cell just like in the 60°C curve. Cells 2, 3 and 4 show a very similar voltage at high current densities of 2,2 A/cm$^2$ with the highest voltage after cell 1 belonging to cell 4 although the voltages of cells 2, 3 and 4 are very similar and do not show a significant decrease in the performance compared to each other.
\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/80_Cells.pdf}
\caption{80°C polarization curves of the 4 cells after 81000 VC.}
\label{fig:80_Cells}
\end{figure}

%Discussion Overall performance at 80°C better than at 60°C discuss??

\section{High Temperature Endurance Run}
\label{sec:4_High Temp}

The high temperature endurance run was conducted on the same model of test bench as the corrosion endurance run and also at the same time. The same activation sequence with four 80°C polarization curves was performed for the activation of the 4 cell stack. The cell temperature was ramped up to 103°C with a dew point temperature at 72°C at the anode and 85°C at the cathode. This leads to a RH of 30,1\% at the anode and 51,3\% at the cathode and consequently lower than the 36,65\% at the anode and 54,66\% of the cathode of the corrosion endurance run. The high temperature endurance run could not be completed due to the formation of pinholes in the cell which created a leak between anode anode cathode. However, the BoL and the characterization after 12500 VC was completed and its findings will be explained in the following.

\subsection{Polarization Curves at 60°C}
\label{subsec: 60_PolCurve_HT}

The results of the BoL characterization and the characterization of the cell after 12500 VC can be found in the Figure \ref{fig:60_Pol_HT}. The BoL curve has a slightly higher average cell voltage at current densities ranging from 0,1 to 1 A/cm$^2$ as well as from 1,6 to 2,2 A/cm$^2$.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/60_PolCurve_HT.pdf}
\caption{60°C polarization curve at BoL and after 12500 VC.}
\label{fig:60_Pol_HT}
\end{figure}

Between 1 and 1,6 A/cm$^2$ the 60° polarization curve after 12500 VC has a slightly better performance as the BoL curve. This could be attributed to a better humidity level of the membrane leading to a lower ohmic voltage loss and a better performance in that section \citep{108_lin2021prediction}


\subsubsection{High-Frequency Resistance (HFR)}

Figure \ref{fig:60_HFR_HT} illustrates the measured HFR during the two 60°C polarization curves. Both curves show a peak in the HFR of 64,4 $m\Omega\cdot\text{cm}^2$ per cell at the lowest current densities (<0,1 A/cm$^2$). After the initial peak the HFR of the 12500 VC curve stays lower than then HFR of the BoL curve.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/60_HFR_HT.pdf}
\caption{HFR during the 60°C polarization curve at the BoL.}
\label{fig:60_HFR_HT}
\end{figure}

The 12500 VC curve has the lowest HFR between the current densities of 1 and 1,6 A/cm$^2$ and reaches the lowest point at an HFR of around 50,7 $m\Omega\cdot\text{cm}^2$ per cell. At a current density of 2,2 A/cm$^2$ the BoL curve has a HFR of 56,7 while the 12500 VC curve has a HFR of 51,8 $m\Omega\cdot\text{cm}^2$ per cell.

The sudden drop of the HFR of the 12500 VC curve to the lowest points at the current densities between 1 and 1,6 A/cm$^2$ could explain the performance boost as the activation losses are compensated at this point due to a lower ohmic loss in this section \citep{108_lin2021prediction}.

\subsubsection{60°C Polarization Curves of the Cells after 12500 VC}

Figure \ref{fig:60_Cells_HT} shows the voltage of the 4 cells at the different current densities during the 60°C polarization curve. Cell 4 presents the lowest voltage with a value of 0,57V while cell 1 has the highest voltage of the cells with a value of 0,6V at a current density of 2,2 A/cm$^2$ after 12500 VC.

Meanwhile the cell 4 of the corrosion endurance run after the 12500 VC at 2,2 A/cm$^2$ had a voltage of 0,58V. Therefore, a higher degradation of the cell can already be seen at this early stage after just 12500 VC.


\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/60_Cells_HT.pdf}
\caption{60°C polarization curves of the 4 cells after 12500 VC.}
\label{fig:60_Cells_HT}
\end{figure}

This increased degradation of the cell 4 compared to the corrosion endurance run after the same number of cycles could be caused by the Pt dissolution and redeposition resulting in a loss of the electrochemical surface area (ECSA). This mechanism is influenced by high temperatures as well as high water content in the ionomer which are the exact conditions in after the first cycle of the high temperature endurance run as well as in the 60°C polarization curve with the high RH \citep{wallnofer2024main}.


\subsection{Polarization Curves at 80°C}
\label{subsec: 80_PolCurve_HT}

The results of the in-situ characterization with the 80°C polarization curve of the high temperature are presented in the following Figure \ref{fig:80_Pol_HT}. With a decrease in the RH due to the 80°C characterization method both curves show almost the same voltage at the highest tested current density of 2,2 A/cm$^2$.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/80_PolCurve_HT.pdf}
\caption{80°C polarization curves at BoL after 12500 VC.}
\label{fig:80_Pol_HT}
\end{figure}

At very low current densities the BoL curve has a slightly higher voltage than the 12500 VC curve. The higher activation loss at the beginning could be caused by a degradation of the Pt catalyst when compared to the BoL \citep{jouin2016}.

At higher current densities the 12500 VC curve shows a lower ohmic resistance and is therefore able to compensate the first losses until at the highest current desensitise it has the same voltage as the BoL curve. The HFR which will be shown in the following could explain this behaviour.


\subsubsection{High-Frequency Resistance (HFR)}

The HFR of the cells during the 80°C polarization curve of the high temperature endurance run are depicted in the following Figure \ref{fig:80_HFR_HT}. At low current densities the HFR of both curves peaks at a maximum value of 168 $m\Omega\cdot\text{cm}^2$ per cell and then quickly stabilizes at a lower HFR value of around 57 $m\Omega\cdot\text{cm}^2$ per cell.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/80_HFR_HT.pdf}
\caption{HFR during the 80°C polarization curve at the BoL, after 12500 VC.}
\label{fig:80_HFR_HT}
\end{figure}

The phenomena seen in the 80°C polarization curve at high current densities (above 1,6 until 2,2 A/cm$^2$) could be attributed to the lower HFR presented by the 12500 VC curve at this current densities. During the 12500 VC curve the membrane has a slightly better humidification an therefore is able to be more efficient and produce a better power output due to a lower ohmic voltage loss \citep{108_lin2021prediction}.


\subsubsection{80°C Polarization Curves of the Cells after 12500 VC}

Figure \ref{fig:80_Cells_HT} shows the voltage of the 4 cells at the different current densities during the 80°C polarization curve after 12500 VC. Cell 1 has the highest voltage of the cells at high current densities (2,2 A/cm$^2$) while the other cells (2, 3, 4) have almost the same voltage at 0,58V.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.9\textwidth]{Figures/Results/PolCurves/80_Cells_HT.pdf}
\caption{80°C polarization curves of the 4 cells after 12500 VC.}
\label{fig:80_Cells_HT}
\end{figure}

Although the activation losses of all the cells are almost the same, ohmic and polarization losses of the cells differ from each other. Therefore the voltage of cells 2,3 and 4 is under 0,6V.


Ma et. al. suggests that higher the initial HFR peak and the temperature of the cell are more time will be needed to reach a stable HFR \citep{109_ma2022effect}. This was validated with the results of the HFR on the high temperature endurance run and corrosion endurance run. The HFR of the high temperature endurance run had a higher peak at 168 $m\Omega\cdot\text{cm}^2$ per cell on the 80°C polarization curve and reached a stable HFR after a current density of 0,3 A/cm$^2$. Meanwhile the corrosion endurance run and its 60°C polarization curve reached a lower HFR much faster since the peak was only at 64,4 $m\Omega\cdot\text{cm}^2$ per cell and the 81000 VC did not have a peak at all.

Furthermore the significant decrease in the voltage at high current densities of the cell 4 from the corrosion endurance run at the 60°C polarization curve after 81000 VC will be further investigated using the ex situ-methods in the next section.

\newpage

\subsubsection{Department of Energy Target}

\section{pH Measurement}

\section{Endurance Run}

Betriebsbedingungnen soll ist vergleich

\subsubsection{Polarization Curves}

\subsubsection{Effects of Relative Humidity in the Air}

\subsection{Product Water Analysis}


\section{Ex-Situ Analysis}

After the corrosion endurance run the cells were brought to the laboratory for further investigation. To avoid contamination the cells were opened in a clean room and then underwent a microscopical, LIBS and REM/EDX analysis. Since the cell 4 presented a significantly higher voltage loss than the cell 1 in the 60°C polarization curve after 81000 VC of the corrosion endurance run this two cells will be compared to a reference cell. The results of this analysis will be presented in the following section starting with the microscopical analysis in Section \ref{subsec:4_Microscope}.


\subsection{Microscopical Analysis}
\label{subsec:4_Microscope}

The microscopical analysis focuses its attention on the positions A6 and M1 of the BPs 1 and 4 as well as the cathode outlet. A6 refers to the position in the matrix shown in Figure \ref{fig:Matrix} at the cathode outlet while M1 is located at the cathode inlet.

\subsubsection{Analysis of the BPs}
% Vergleich von Referenz zu BP1
\begin{figure}[htbp]
\centering
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_A6_1.pdf}
\caption{Reference BP position A6.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_A6_2.pdf}
\caption{Reference BP position A6 x 100 ($\mu m$).}
\end{subfigure}
\vspace{0.2cm}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_M1_1.pdf}
\caption{Reference BP position M1.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_M1_2.pdf}
\caption{Reference BP position M1 x 100 ($\mu m$).}
\end{subfigure}
\caption{Microscopy Analysis of the BP: Stainless Steel 316L Reference BP.}
\label{fig:4_Micro_ref}
\end{figure}

Figure \ref{fig:4_Micro_ref} shows the results of the microscopy of the reference BP. An overview of the position A6 can be seen in the part (a) of the figure and for M1 in the part (c) of the figure. In the overview almost no discolourations were found except for the welding seams which presented a slight red and brown discolouration in the reference plate. A close-up of the channels between the C coated ribs in positions A6 and M1 did not show any particular defects in the metal.

Figure \ref{fig:4_Micro_BP1} shows the microscopy of the BP1 of the cell from the corrosion endurance run. The channels show no clear defects or signs of degradation after the 81000 VC conducted. On the other hand the discolouration at the welding seams is now more visible than in the reference plate. Over the flow field in the upper right corner of position M1 (\ref{fig:4_Micro_BP1} (c)) the BP also presents some scratches in the surface.


\begin{figure}[htbp]
\centering
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP1_A6_1.pdf}
\caption{BP1 position A6.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP1_A6_2.pdf}
\caption{BP1 position A6 x 100 ($\mu m$).}
\end{subfigure}

\vspace{0.2cm}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP1_M1_1.pdf}
\caption{BP1 position M1}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP1_M1_2.pdf}
\caption{BP1 position M1 x 100($\mu m$)}
\end{subfigure}
\caption{Microscopic Analysis of the BP: Stainless Steel 316L BP1.}
\label{fig:4_Micro_BP1}
\end{figure}

The cell 4 of the stack had the highest voltage drop at high current densities through the 60°C polarization curve. Therefore the BP 4 was analysed under the microscope as well and the results can be seen in the following Figure \ref{fig:4_Micro_BP4}. The red and brown discolourations at the welding seam on the positions M1 and A6 are more visible than in the reference BP. The close-up of the channel in \ref{fig:4_Micro_BP4} (d) also shows defects in the material on the side of the channels. The top right corner of the BP 4 position M1 ( Figure 4.15 (c)) also shows scratches in the metal.

\begin{figure}[h]
\centering
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_A6_1.pdf}
\caption{BP4 position A6.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_A6_2.pdf}
\caption{BP4 position A6 x 100 ($\mu m$).}
\end{subfigure}
\vspace{0.2cm}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_M1_1.pdf}
\caption{BP4 position M1.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_M1_2.pdf}
\caption{BP4 position M1 x 100 ($\mu m$).}
\end{subfigure}
\caption{Microscopic Analysis of the BP: Stainless Steel 316L BP4.}
\label{fig:4_Micro_BP4}
\end{figure}

However, the most significant difference between the BPs 1 and 4 can be seen at the cathode outlet in Figure \ref{fig:4_Micro_SP}. The BP 4 (\ref{fig:4_Micro_SP}(d)) has a severe red and brown discolouration whereas the outlet in the BP 1 (\ref{fig:4_Micro_SP} (b)) remains without discolourations. The edges of the cathode outlet in BP 4 also show signs of metal dissolution due to corrosion.

\begin{figure}[htbp]
\centering
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP1_Co_1.pdf}
\caption{BP1 Cathode Outlet.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP1_Co_2.pdf}
\caption{BP1 Cathode Outlet x 100 ($\mu m$).}
\end{subfigure}
\vspace{0.2cm}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_Co_1.pdf}
\caption{BP4 Cathode Outlet.}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_Co_2.pdf}
\caption{BP4 Cathode Outlet x 100 ($\mu m$).}
\end{subfigure}
\vspace{0.2cm}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_A6_S.pdf}
\caption{BP4 A6 Welding Seam x 10 ($\mu m$).}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/M_BP4_M1_S.pdf}
\caption{BP4 M1 Welding Seam x 100 ($\mu m$).}
\end{subfigure}
\caption{Microscopic Analysis of the BP: Stainless Steel 316L at the Cathode Outlet and Welding Seam.}
\label{fig:4_Micro_SP}
\end{figure}

Since the welding seam already showed signs of discolourations and corrosion in the reference plate without having conducted an endurance run this positions were analysed again in the BP 4 to look for damages. Figure \ref{fig:4_Micro_SP} (e) shows the welding seam in the position A6 of the BP 4. Across the lower part of this image three light coloured points stand out in the brown discoloured part, this could indicate three sites where pitting corrosion has already started.

Welding seams are known to be a weak point for corrosion. Since the temperatures of the process causes a decomposition of the austenitic matrix. Cr and Mo are consumed by the formation of precipitates like carbides and cause Cr-depleted areas which are more vulnerable to corrosion \citep{yan2019effect}.


\newpage
\subsubsection{Analysis of the CCM}

Furthermore, the CCM will also be analysed to look for corrosion or Pt agglomeration, dissolution or any other damages. Since the CCM showed no increased damage at the positions A6 and M1 when compared to the other positions in the cell matrix this time the positions M6 and A1 will be analysed in greater detail. Figure \ref{fig:4_Micro_CCM} depicts the microscopical analysis of the CCMs from the cathode side of BPs 1 and 4.

\begin{figure}[h]
\centering
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/CCM_M_1_M6.pdf}
\caption{CCM cathode BP1 M6 x 200 ($\mu m$).}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/CCM_M_1_A1.pdf}
\caption{CCM cathode BP1 A1 x 200($\mu m$).}
\end{subfigure}
\vspace{0.2cm}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/CCM_M_4_M6.pdf}
\caption{CCM cathode BP4 M6 x 200($\mu m$).}
\end{subfigure}
\begin{subfigure}{0.45\textwidth}
\centering
\includegraphics[width=\linewidth]{Figures/Results/CCM_M_4_A1.pdf}
\caption{CCM cathode BP 4 A1 x 200 ($\mu m$).}
\end{subfigure}
\caption{Microscopic Analysis of the CCM: Sx 200 ($\mu m$).}
\label{fig:4_Micro_CCM}
\end{figure}

Position M6 of the CCM from the BP 1 shows signs of Pt-agglomeration. The three small darker lines in that can be seen in this position are cracks in the CCM. Furthermore wave structures are also visible across all the positions. When looking at the position A1 of the CCM shown in \ref{fig:4_Micro_CCM} (b)
the Pt agglomeration is even more visible in the bright spots.

The CCM from the cathode in BP 4 shown in \ref{fig:4_Micro_CCM} (c) and (d) also shows signs of Pt agglomeration. Especially the bright parts in the position A1. Both position M6 and A4 show the same wave structures previously observed in the CCM of the cathode from the BP 1. Unlike BP 1 the image \ref{fig:4_Micro_CCM} (c) also present 4 darker spots which could be a part of the MPL which adhered to the CCM.

The ORR reaction relies heavily on the Pt catalyst of the CCM which highlights the important role of Pt in the performance of the cell \citep{PEM_MEA_parekh2022recent}. The agglomeration process can increase the hydrophilicity of the carbon support which consequently can lead to to more water in the cell and in some cases flooding of the cell. This could also lead to a limited amount of oxygen reaching the active sites and a decrease of the cell performance \citep{okonkwo2021platinum}. The Pt agglomeration will be further investigated using the REM and EDX.



\subsection{LIBS Measurement}
\label{subsec:4_LIBS}

In order to further evaluate the discolourations found in the BP4 and to provide a better insight into the conditions of the stainless steel 316L plate and its possible corrosion and loss in performance, an additional LIBS analysis was conducted. First the positions M1 and A6 of the BP4, where the relative atomic concentrations of its components were compared to those of the reference BP. The results of this analysis can be found in the Figures \ref{fig:LIBS_M1} and \ref{fig:LIBS_A6}.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.7\textwidth]{Figures/Results/LIBS_M1.pdf}
\caption{LIBS: Comparison of the position M1 of BP4 with the reference plate.}
\label{fig:LIBS_M1}
\end{figure}

When comparing the position M1 of the BP4 with the reference plate the BP4 shows a light increase of the oxygen concentration as well as a decrease in carbon. Although the increase in oxygen is clear, the standard deviation of the 16 points measured on the carbon is to high on as to be able to draw any conclusions. The standard deviation of Fe, Ni and Cr is as large as the difference measured from the reference plate to the BP making it imposable to draw a conclusion from this change. Since the position M1 did not present any discolourations it could be possible that only a light oxide layer started to appear on it.

Figure \ref{fig:LIBS_A6} compares the position A6 from the BP 4 to the reference plate. Since the microscopy of position A6 did not present any discolourations or alterations of the BP this result is also reflected in the LIBS measurement.

\begin{figure}[H]
\centering
\includegraphics[width=0.7\textwidth]{Figures/Results/LIBS_A6.pdf}
\caption{LIBS: Comparison of the position A6 of BP4 with the reference plate.}
\label{fig:LIBS_A6}
\end{figure}

The increase on oxygen is even less noticeable than in the comparison of M1 to the reference plate. The other elements stay have the same concentrations as in the reference plate which does not indicate corrosion in this part of the cell.

However, the cathode outlet showed a significant difference when analysed with the microscope. Therefore the cathode outlet of the BP 1 and BP 4 were compared with the one of the reference plate. The result of this comparison can be found in the following Figure \ref{fig:LIBS_Cathode}.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.7\textwidth]{Figures/Results/LIBS_Cathode Outlet.pdf}
\caption{LIBS: Comparison of the cathode outlet of the BP 1, BP 4 and the reference BP.}
\label{fig:LIBS_Cathode}
\end{figure}

BP 4 shows a significant decrease in carbon of almost 23\% compared to the reference plate as well as the BP 1. In addition, a decrease in Cr and Ni of the BP 4 is also detected by the LIBS. BP 4 also shows an increase of 12\% in the oxygen concentration compared to the BP 1. This increase shows clear signs of the formation from an oxide layer which could explain the discolouration found in Figure \ref{fig:4_Micro_SP}. This oxide layer is also a clear sign for corrosion in the cell since the corrosion reaction causes the formation of a passive oxide layer.

Since the welding seam also showed a clear discolouration across the reference plate and was even more visible on the BP 1 and BP 4 after the corrosion endurance run it was also analysed using LIBS. The results of the comparison from the welding seam in position A6 of the BP 4 with the reference plate can be seen in the following Figure \ref{fig:LIBS_Sonderpositionen}.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.7\textwidth]{Figures/Results/LIBS_Sonderposition.pdf}
\caption{LIBS: Comparison of the welding seam in position A6 of the BP 4 with the reference plate.}
\label{fig:LIBS_Sonderpositionen}
\end{figure}

When comparing the welding seam to the reference plate it is visible, that the oxygen percentage increases around 8\% showing sings of the formation of an oxide layer \citep{laedre2017materials}. Furthermore the decrease in carbon can also be detected at the welding seam. The high standard deviation of Fe makes it impossible to evaluate the decrease in the concentration. Since the concentration change in Cr, Ni and Mo was barely noticeable, it is also not possible to make a final conclusion about this metals.

Due to the fact, that welding seams are more susceptible to corrosion as mentioned before the discolouration seen at in this spots as well as the increased oxygen percentage measured by the LIBS the corrosion can be confirmed \citep{yan2019effect}. The protective oxide layer created by the metal could also lead to a higher ohmic resistance due to a less reactive oxide layer formed to protect the metal from corrosion \citep{105_novalin2022concepts}.


\subsection{SEM/EDX Measurement}
\label{subsec:4_SEM/EDX}

The final section of this chapter will present the results of the SEM and EDX measurements of the CCM. The Fenton mechanism mentioned in the theoretical background \ref{subsec:membrane degradation} is responsible for the membrane degradation and can be catalysed with Fe$^{2+}$ as well as other metal ions from the BP. Therefore, the CCM of the cathode was analysed using SEM and EDX in order to find traces of metal which could have come from the BP and travelled from there to the MPL, GDL or CCM and into the membrane during the corrosion process of the BP and whit its dissolution from the BP.

\subsubsection{CCM BP 1}

The results of the SEM and EDX analysis of the cathode CCM from BP 1 is presented in the following Figure \ref{fig:REM_4_A1}. Since the microscopical analysis performed in \ref{subsec:4_Microscope} showed a Pt agglomeration it was further analysed with SEM and EDX.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.8\textwidth]{Figures/Results/CCM_REM_BP1_A1.pdf}
\caption{REM/EDX of the CCM at the position A1.}
\label{fig:REM_1_A6}
\end{figure}

As seen in Figure \ref{fig:REM_4_A1} no traces of Fe, Ni or Cr could be found in the CCM of the BP 1. The Pt agglomeration could be proved with the spectrum since the Pt peak was found in the position A1 of the CCM. The results of the position M6 did not confirm any traces of Fe, Ni or Cr and can be found in the appendix \ref{sec: A_REM/EDX}.


\subsubsection{CCM BP 4}

Since the BP 4 showed clear signs of corrosion in the microscopy and in the LIBS analysis as well as a Pt-agglomeration in the CCM this was also furher analysed with SEM and EDX. The results of both SEM and EDX analysis of the cathode CCM from the BP 4 at the position A1 are presented in the following Figure \ref{fig:REM_4_A1}.

\begin{figure}[htbp]
\centering
\includegraphics[width=0.8\textwidth]{Figures/Results/CCM_REM_BP4_A1.pdf}
\caption{REM/EDX of the CCM 4 at the position A1.}
\label{fig:REM_4_A1}
\end{figure}

Even though the BP 4 corroded, still no traces of Fe, Ni or Cr could be found in the CCM and therefore the EDX showed no peak of any of the aforementioned elements. However, the Pt- agglomeration was also visible with a peak in Pt shown in the EDX analysis.

A reason for this could be the duration of the endurance run. The corrosion endurance run only lasted 400 hours which compared to the 8000 hours the lifetime could have resulted in a much lower degradation of the BP due to corrosion and consequently a much lower concentration of metal ions being released from the BP to the membrane through the CCM.

As mentioned by Low et al. the the MEA is very vulnerable to metallic poisoning. When the oxidation process starts the metallic cations migrate from the BP through the cell and cause a greater membrane degradation and also cause a higher ohmic resistance due to the reduced ionic conductivity \citep{low2024understanding}.

A study performed by Novalin et al. analysed also analysed the metal ion dissolution from the BPs as a result of cycling measurement. They were able to find traces of Fe, Ni and Cr in the GDL and MEA after 700 cycles and a RH of 66\% and a stoichiometry of 2 at the anode and 2,5 at the cathode (higher than in the corrosion run). Novalin concluded, that dryer conditions enhance repassivation mechanisms of the metal BP which makes the transport of the metal ions to the MEA less probably \citep{105_novalin2022concepts}.

\subsection{EDX Measurement}
Lastly, since the Pt agglomeration was identified by the microscopy and the EDX analysis it is also worth discussing. Pt agglomeration could also be a factor influencing the activation losses due to the possible loss of ECSA produced by the voltage cycling
\citep{Pol_thiele2024realistic}.

\subsection{SEM Measurement}
%\subsection{SEM Measurement}

\subsection{ICP Measurement}
%\subsection{ICP Measurement}


Verfassen Vorschau
Laden…
Abbrechen
Speichern