CHAPTER 1 INTRODUCTION
1.3 Scope of Present Study
The research mainly concerns on the improvement of PEMFC CO tolerance by using oxygen-bleeding technique. In the experiment, it will inject oxidant (air) into the fuel stream at anode and investigate if it has influence on CO tolerance of PEMFC. The reason for using air instead of pure oxygen in the oxygen-bleed test is the safety concern, of course, the former can be obtained directly from the atmosphere without extra cost for supply.
The present work is divided into three parts, as shown in Fig. 1.1. The first one is to define suitable CO poisoning test condition for PEMFC that include cell voltage and current density. In such test, it will fix different cell voltages or current density, respectively, to perform poisoning experiment. Then, it will determine which one (cell voltage or current density) has a better performance in CO poisoning test and find out the difference between these two methods.
The second part is to bleed air into anode fuel stream ( ), which contains CO at different concentrations (10.1, 25, 52.7ppm). At the beginning, it should first identify the cell baseline before poisoned, and then the steady state poisoned polarization curve before the application of
CO
H /
2air bleed. The two sets of data will compare with the recovered performance obtained by air bleed. In this test, the experiments, either fixing the cell voltage or current density, are transient, and the cell performance will decay with the time. In the transient test, it can be further divided into two parts that are long- (30min.) and short-duration (3min.) poisoning, respectively. As the duration is reached, the air is injected into anode fuel stream. Then, it will determine the air injection rate for a given CO-concentration fuel stream that can effectively improve the CO tolerance and recover cell performance. Finally, the results from theses two specific times (30min. and 3min.) are compared with each other and a corresponding discussion will be given.
The third part is to use pure Pt catalyst to perform the CO poisoning tests with and without air-bleeding. The results will compare with the ones of Pt alloy catalyst.
CHAPTER2
EXPERIMENTAL APPARATUS
The experimental apparatus are set up in the Energy & Resources Laboratories of the ITRI (Industrial Technology Research Institute). It is the sole fuel cell test/research center in Taiwan. The present study uses their apparatus to carry out this research and the corresponding experiments. The apparatus consists of two major elements, which are the fuel cell test station and the test sample of PEMFC. The above-mentioned elements are described in detail as follows.
2.1 Fuel Cell Test Station
This fuel cell test station consists of four components, which are the electronic load, MFC readout power supply, power supply, gas pipelines controller. Utilizing this system can change operation conditions, like temperature, humidification, flow rate, pressure, cell potential, etc, required by PEMFC performance evaluation. The schematic configuration is shown in Fig. 2.1. The elements of test station are described as follows.
2.1.1 Electronic Load
The electronic load uses the style of HP6060b. This DC electronic load is ideal for the test and evaluation of dc power sources and components, and is well suited for applications in areas, such as research and development, production, and incoming inspection. This load box
operation mode has constant current and voltage. It measures current range between 0 and 60 A. The maximum power output that load box can support is 300 watts. Electronic load can measure cell voltage and be used to control cell voltage and current. Eventually, it will consume cell power production in the test. The load box will display the cell performance on its monitor. The HP6060b load box is shown in Fig. 2.2.
2.1.2 MFC Readout Power Supply
The PC-540 MFC Readout Power Supply (Fig. 2.3) can display and control units for precision gas control in conjunction with MFCs. This apparatus has multiple channel instrumentation that it can completely control up to four MFCs at the same time.
In the present experiment, the flow rate controller will control three flow meters, which include anode gas, cathode gas and additional gas meter. These gas meters are anode ( ), cathode ( or air) and bleeding ( or air), respectively.
H
2O
2O
22.1.3 Power Supply
This apparatus (Fig. 2.4) is a power source switch that provides power to each component in this test station. It includes mass flow controller, solenoid valve, heater (anode, cathode and cell), and three thermocouples plugs. In the power supply, it has three temperature controllers and monitors that the controlled temperature range is between 25 and 95℃.
These temperature controllers will control anode, cathode humidity bottle and cell temperature that can change temperature by direct input buttons.
When press power button, it will start all instruments of test station. On
the other hand, pressing power off will shut down all power.
2.1.4 Gas Pipelines Controller
The gas pipelines controller (Fig. 2.5) can manage what kind gas can feed into anode and cathode fuel stream. It also can control gas humidity, temperature, back pressure and flow rate. There are five gas inlets behind this apparatus that both anode and cathode have two gas channels respectively, the last one is nitrogen channel. The anode channels always feed with and reformation gas ( ). The cathode channels always feed with and air. On the operation board, it has anode and cathode globe valve, which can change the gas that test cell needs and can shut the valve that the gas cannot pass through it into the test sample. In addition to globe valve, there have three solenoid valves in the instrumentation. They are the safe valves. When this test station is in danger or overload condition, the solenoid valves will shut off anode and cathode gas channel and the other valve will open to purge nitrogen into test cell. Before fuel gas enters anode and cathode sides, it must go through humidity bottle first. This bottle will fill up water and has a heater and thermocouple inside that can heat the water when gas pass through it to increase the humidity of fuel gas before it enters the fuel cell.
Finally, the board also has two adjust pressure valves that can change outlet backpressures for both anode and cathode fuel streams.
H
2H /
2CO
O
22.1.5 Liquid-Gas Separator
After fuel gas passes through test cell, there will have some remnant gas and water discharged from the cell fuel outlet. If there is too much water in the waste gas channel, then it will prevent remnant gas from
entering into the atmosphere. Therefore, it must enter a liquid-gas separator first so that the water can be kept in a bottle and only let the gas go out. Then the waste gas will discharge to the atmosphere by an exhaust fan.
Finally, the photo of complete test station is shown in Fig. 2.6. In this photo, each apparatus from top to bottom is MFC readout power supply, electronic load, power output, gas pipelines controller and liquid-gas separator, respectively.
2.2 Test Sample of PEMFC
The PEMFC has six major components, which include MEA (membrane electrode assembly), GDL (gas diffusion layer), gasket, gas flow channel and current collector. The MEA (Fig. 2.7) is Gore’s commercial product, PRIMEA series 5561 MEA. The membrane in the middle has a thickness of 35μm and catalyst loads are of 0.45 2
cm
2.7) used in the experiment is CARBEL CL GDL and its thickness is of 0.4 mm. The Ucar carbon is used as the material of flow field channels. It will be processed serpentine flow channels (Fig. 2.8) on the carbon board.
These flow fields consist of 26 equally spaced channels of 1 mm width, 1 mm height and 1mm width. The current collector (Fig. 2.9) composes of copper plating gold. This current collector board can conduct electric current from test cell. The end plank (Fig. 2.10) is made by nickel-plating steel, used as fixing and protecting this fuel cell structure. The
cm2 cm2
dimensions of carbon board, current collector and end plank all are 100 mm × 100 mm.
Fig. 2.11 shows each constituting component of PEMFC. There is a sequence to compose a cell. The MEA is always in middle of the cell and both anode and cathode electrode have a GDL. In the outer circle of GDL, it places a gasket (Fig. 2.7) to prevent the gas leakage to environment. This composing sequence is shown in Fig. 2.12. Finally, this PEMFC test sample is shown in Fig.2.13.
2.3 Test conditions
The experimental conditions for these tests are fixed. The fuel flow rates of anode and cathode are calculated from the theoretic volume, which can produce one ampere of current. They are 7.6 and 3.8cc/min/Amps for hydrogen and oxygen, respectively. Then, for the anode, it is multiplied by the stoichiometry of 1.37 to obtain a value of 10.4 cc/min/Amps), whereas for the cathode, it is 1.84 to get 7 cc/min/Amps. This can guarantee the fuel flow rate being sufficient to initiate the electrochemistry reaction. The reason for the cathode stoichiometry higher than the anode one is the lower oxygen activity to reaction. The fuel cell temperature always fixes at 65 . The pressures at the outlet of fuel stream for both anode and cathode are 101kPa. The humidification temperatures are 80 and 70 for anode and cathode fuel streams, respectively. These temperatures can let fuel possess enough humidity to crossover membrane and obtain the optimum cell performance with a cell temperature of 65 .
οC
οC
οC
οC
2.4 Procedure of the Experimental Operation
1. Connect all lines of test station with fuel cell before the experiment. For instance, connect potential sensors to the anode and cathode current collectors; connect positive pole loading line to the cathode current collector. On the other hand, connect anode with negative pole, insert the heater into the end plank hole, insert the thermocouple into the carbon board hole, and connect pipelines to the fuel cell.
2. Add the water to the humidification bottle by the atmosphere style water bottle.
3. Open the valves of , , and fuel cylinders and retain the inlet pressure up to 80 psi.
2 2 2 2
2
2
2
H O N H / CO
4. Turn on the globe valves of the gas pipelines on the controller operation board; anode side turns to or pipeline, cathode side turns to pipeline.
H
2H /
2CO O
5. Turn on the power source of the exhaust fan. Because this laboratory is in the airtight space and the experimental gases contain CO and
H
, so it is a necessary procedure.6. Check if any fuel gases leakage from pipeline’s connection by applying suds on them. It is a very important procedure, especially for the uses of CO and,
H
.7. Push down the power button of the test station to star this system.
8. Activate the software of the test system.
9. Set the minimum fuel flow rate and the flow rate per ampere current of anode and cathode. The minimum flow rates of anode and cathode are 104cc/min and 70cc/min, respectively, and the flow rates per ampere current are 10.4cc/min/Amp and 7cc/min/Amp, respectively.
10. Set anode and cathode humidification temperature.
11. Push down the gas reset button on the power supply board. At the same time, solenoid valve will shut off , which stops to purge the fuel cell.
N
212. Push down the applying fuel icon of software window. Now, anode and cathode fuel pipeline solenoid valve will open and the fuel gas is fed into fuel cell.
13. Set the cell temperature to star the cell heater as the humidification temperature is up. It is to protect MEA from drying to damage the MEA.
14. Push down the “apply load” icon of the window that the electronic load will start loading from the test sample as the OCV (open-circuit voltage) reaches the steady value.
15. Set the same cell overpotential to activate the test fuel cell. In general, a new cell needs to activate several hours until it achieves optimum or steady performance.
16. Define proper CO poisoning test condition for PEMFC that include cell voltage and current density for the first part of the present study.
17. For the second and third parts, hold the same cell potential to record current-time curve. At first, feed pure as the anode fuel, then observe the current change by transferring the fuel to
mixed gas.
H
2CO H /
22
18. Then inject air into anode fuel stream by the equivalent or periodic method; or change fuel gas back to the pure to observe current recovery rate.
H
19. In present study, replace CO concentration of the anode fuel and try different air bleeding ratio or frequency. Repeat the procedure from (17)-(18) steps.
CHAPTER 3
UNCERTAINTY ANALYSIS
All of the data from experimental results may not be equally good to adopt. Their accuracy should be confirmed before the analyses of experimental results are carried out. Uncertainty analysis (or error analysis) is a procedure used to quantify data validity and accuracy [21].
Errors always are presented in experimental measuring. Experimental errors can be categorized into the fixed (systematic) error and random (non-repeatability) error, respectively [21]. Fixed error is the same for each reading and can be removed by proper calibration and correction.
Random error is different for every reading and hence cannot be removed.
The objective of uncertainty analysis is to estimate the probable random error in experimental results.
From the viewpoint of reliable estimation, it can be categorized into single-sample and multi-sample experiments. If experiments could be repeated enough times by enough observers and diverse instruments, then the reliability of the results could be assured by the use of statistics [22].
Like such, repetitive experiments would be called multi-sample ones.
Experiments of the type, in which uncertainties are not found by repetition because of time and costs, would be called single-sample experiments.
3.1 Analyses of the Propagation of Uncertainty in Calculations
Uncertainty analysis is carried out here to estimate the uncertainty levels in the experiment. Formulas for evaluating the uncertainty levels in the experiment can be found in many papers [22, 23] and textbooks [21, 24 and 25]. They are presented as follows:
Suppose that there are n independent variables, , ,…, , of experimental measurements, and the relative uncertainty of each independently measured quantity is estimated as u
x
1x
2x
ni. The measurements are used to calculate some experimental result,
R
, which is a function of independent variables,x
1,x
2,…,x
n;R
=R ( x
1,x
2 ,...,x
n)
.Normalize above equation by dividing
R
to obtaini Eq. (3-2) can be used to estimate the uncertainty interval in the result
due to the variation in xi. Substitute the uncertainty interval for xi,
To estimate the uncertainty in
R
due to the combined effects of uncertainty intervals in all the xi’s, it can be shown that the best representation for the uncertainty interval of the result is [23]12
3.2 The Analysis of CO Concentration
In the present experiment, it will discuss the poisoning influence with the lower CO concentrations, which are 50, 25 and 10ppm, respectively.
The SAN FU GAS company provides the specific mixing gases according to the experimental requirements. It analyzes these gases by using the GC-DID method so that the accurate value of CO concentration contained in the gas can be ensured. These concentrations are 52.7, 25 and 10.1ppm, respectively. Finally, these mixing gases are stuffed into the steel cylinders by the high pressure of 120 .
CO H /
2H
2/ cm2
Kg
3.3 The uncertainty of test station apparatus
The apparatus must correct with other standard instruments to make sure that it can normally operate and let the inaccuracy of the experimental result reduce to the minimum. However, the test apparatus used in present experiment are in the Energy & Resource Laboratories of ITRI and they have been corrected by their own researchers periodically.
(1) The uncertainty of HP 6060B electronic load: u
V, u
AThe HP 6060B electronic load in the test station had corrected its potential and current meter before experiment. The research uses FLUKE 8060A Digital Multimeter and Chroma Smart N300-040 Electronic Load to correct HP load box. These correction instruments had corrected by the Center for Measurement Standards of ITRI. At the beginning, set the
potential of the load box and the readout value will show on its monitor.
Then, use FLUKE digital meter to measure the actual potential. If the inaccuracy of potential is lower than ±2%, it can consider the load box being normal. In Table 3.1, it shows the error for different potentials.
The standard value is the load box readout, whereas the digital value is the actual measured value.
Next, the researchers correct the DC current meter of HP load box.
They use Chroma Smart electronic load and FLUKE digital meter to find the impedance of the shunt. After that, they connect the shunt between HP load box and DC power source. And adjust different potentials of power source so it can change the measurement current of load box meter.
At the same time, the shunt will measure a signal of current. After converting this signal, it can define the actual current of this circuit. In Table 3.2, it shows the error for different current. The conversion value is the actual current, the other is load box measurement value.
(1) The uncertainty of mass flow controller
In this study, there have three MFCs in this test station that includes anode, cathode and oxygen bleeding flow meter. The researchers correct these MFC according to MFG handbook. The specified error is shown as follows:
The ranges of MFC specified accuracy are 1000±5%with anode MFC, with cathode MFC and
% 5
2000± 500±1% with bleeding MFC. They use the same company instrument, series 5850 MFC, as the standard correction apparatus to correct these MFCs. The results are listed in next
three tables (Table 3.3, 3.4, 3.5). They are anode, cathode and air-bleeding flow meter respectively. In these tables, the standard value means the setting flow rate, the Brooks MFC read value means the test station MFC readout value, the measurement value is the actual measured value. Then, these data can define the errors in different flow rates.
(2) The uncertainty of temperature controller
The correction standard is base on MFG handbook. They use the standard temperature controller, corrected by the Center for Measurement Standards, as the correction apparatus. The error of this must be lower than 5%. There are three temperature controllers in the test station.
They include anode, cathode humidifier and fuel cell, respectively. The results of analyses are listed in Table 3.6, 3.7, 3.8. In these tables, standard value means the setting temperature and the measure value means actual value, measured by the correction apparatus.
3.4 The Experimental Repeatability
In general, the life of the commercial MEA is about 300hr, so the fuel cell performance will decline with an increase of test time. Because CO makes anode catalyst decaying and ageing in the poisoning test, the life of fuel cell is expected to be shorter. In order to improve the cell CO tolerance, it uses air-bleeding technique as the CO oxidant to remove the adsorbed CO from catalyst surface. However, the major amount of will react with to form water and resultant reaction heat is quite high.
This is the other factor that may cause anode catalyst decay. Therefore, it must complete experiment as quick as possible to reduce the time effect,
O
2H
2which influences the experiment results. Due to these factors, it is difficult to the perform repeatability test. In order to confirm the accuracy and confidence of the experiment, the cell performance must recover to the base performance before it carries out the next poisoning experiment. The present experimental works can be divided into three parts, therefore, there are three test samples available to perform tests.
which influences the experiment results. Due to these factors, it is difficult to the perform repeatability test. In order to confirm the accuracy and confidence of the experiment, the cell performance must recover to the base performance before it carries out the next poisoning experiment. The present experimental works can be divided into three parts, therefore, there are three test samples available to perform tests.