- Email: [email protected]
Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Short Communication
Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition Clay Hunt a,*, Marley Zachariasen a, David Driscoll a, Stephen Sofie a, Robert Walker b a b
Department of Mechanical and Industrial Engineering, Montana State University, Bozeman, MT, 59717, USA Department of Chemistry & Biochemistry, Montana State University, Bozeman, MT, 59717, USA
article info
abstract
Article history:
Degradation rates of electrical current during constant voltage operation of SOFCs with
Received 16 April 2018
anodes made using NiO precursor powders from two different manufacturers with and
Received in revised form
without the addition of aluminum titanate (ALT) added by either mechanical mixing or
10 June 2018
anode infiltration have been quantified using a novel MATLAB algorithm. Because the al-
Accepted 19 June 2018
gorithm has been used to quantify degradation rates for many different SOFC tests, it is
Available online xxx
thought that the method can be applied to most measured SOFC data to quantify the instantaneous cell degradation rate as a function of time for the entire SOFC performance
Keywords:
measurement. Degradation rates determined at different times have been plotted against
Aluminum titanate
varying concentrations of ALT addition, facilitating the estimation of optimum ALT con-
Modeling
centration for SOFC anodes made with NiO from a specific manufacturer. The algorithm
Solid oxide fuel cells
used to determine degradation rates is available upon request to the corresponding author.
Degradation rate quantification
© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Optimization
Introduction Solid oxide fuel cell (SOFC) technology is improving every day. This promising means of energy conversion is fuel efficient, clean, and approaching commercialization [1]. When both electrical energy and heat are harvested from a solid oxide fuel cell (SOFC) stack, thermodynamic efficiency can be as high as 90% [2,3]. The main problem to be overcome before this technology becomes a viable source of electricity is the degradation occurring in the electrodes of the cell in the SOFC operating environment [4e12]. Because most SOFC degradation is attributed to the SOFC anode, great efforts have been
made to stabilize traditional Ni yttria-stabilized zirconia (YSZ) anodes, and to develop entirely ceramic anodes [13e19]. Decreasing SOFC operating temperature is also a strategy in reducing cell degradation. This is pursued through the use of SOFCs made with the smallest possible electrode particles, and with metal-supported SOFCs [20]. The report of Ding and Hashida describes the synthesis of a nano-composite NiO-Ce0.8Sm0.2O1.9 powder [21]. Anode stabilization against carbon coking and nickel coarsening through alloying the Ni in the SOFC anode with another transition metal has been reported by Kim et al. [22]. Electrodes made with small particles, or even from solution infiltration of electrode scaffolds are interesting because
* Corresponding author. E-mail address: [email protected] (C. Hunt). https://doi.org/10.1016/j.ijhydene.2018.06.115 0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
the triple-phase boundary length of an electrode increases as electrode particle size decreases [23,24]. However, decreasing particle sizes are associated with more rapid nickel coarsening because the driving force of sintering is inversely proportional to particle size [25]. The work of Khan et al. has proposed a model for nickel-particle agglomeration occurring in Ni-YSZ anodes that is similar to equations concerned with grain size in sintering ceramics [26e28]. Cell degradation that is not consistent with Khan et al.'s predictions might be attributable to other means of cell degradation [29,30]. These results suggest that a means of quantifying degradation rate could be useful in identifying the way in which an SOFC is degrading. Amazing techniques exist for modeling the effects on the triple-phase boundary of nickel migration in the anode of an SOFC [31,32]. However, a method of quantification of the instantaneous degradation rate of a SOFC is not the subject of much attention despite degradation being a subject of interest. The main challenge in the quantification of SOFC degradation rates is that measured electrical power output of a SOFC is difficult to differentiate in a way that provides meaningful results. The work of Wang et al., attempts to fit first-order polynomials to cell voltage that was measured during constant-current cell operation [33]. Although the lines used fit the data in the report by Wang et al., the shape of the such a voltage curve is, in general, not linear with respect to time. Only the recent report of Welander et al. uses a method of curve-fitting that can fit a larger set of fuel-cell degradation data [34]. This report elaborates on the design and merit of the technique used in the previous work and further applies the method of degradation quantification. The impetus for developing this tool was the need to quantitatively ascertain optimum doping levels of any coarsening inhibitor for Ni-based anode catalysts [23,24,35]. Wang et al.'s method of fitting measured voltage to a line demonstrates the idea of fitting measured data to a curve to determine degradation rates from measured voltage. The method discussed in this report is simply an extension of that idea. This method can be used to quantify degradation rates of most fuel cells with data sets that record changing voltage at constant current or changing current at constant voltage. This work considers the application of this algorithm to measure the effect of the addition of Al2TiO5 (ALT) to the anodes of SOFCs made with ~3∙mm (J.T. Baker) and ~4∙mm (Alpha Aesar) initial particle size NiO on fuel cell performance. Original application of ALT to the standard SOFC anode material, the Ni-YSZ cermet, was done to engineer the CTE of said cermet. Increased resistance to fracture of cermet bars with ALT was casually observed. This observation has been the subject of the recent work of Driscoll et al. [35]. The casual observation of anode cermet resistance to fracture also motivated the 2013 study by Driscoll et al. [24]. Results of this work suggest ALT might act as a sintering inhibitor for nickel in the anode cermet under typical SOFC operating conditions. Results of the 2013 Driscoll et al. study motivated the in-depth analysis of ALT in the anode cermet reported by Driscoll et al., in 2016 [23]. Results of this work detail the effects of thermal treatment on ALT in the cermet as studied with electrical impedance spectroscopy, X-ray diffraction, electron microscopy, and voltammetry. These results suggested the effects
of ALT on cell performance might be an interesting application of the degradation-rate algorithm.
Experimental procedure Anodes Anodes were applied to 32∙mm Fuel Cell Materials electrolytes by aerosol application of homogenized suspensions of appropriate quantities of the spherical grade 8YSZ (8YS) (8YSZ (YS) Tosoh Corp., Tokyo, Japan), NiO with initial particle size of about 3∙mm (JT Baker NiO) or NiO with initial particle size of about 4∙mm (Green NiO, Alfa Aesar), and corn starch, used as a pore forming agent to yield an equal-volume mixture of 8YSZ, and nickel metal after anode reduction in the fuel cell operating environment. The anode spray consisted of about 32.4 wt% YSZ, 62.2 wt% NiO, and 5.4 wt% pore former, where total powder mass is given by the sum of the masses of the pore former, YSZ, and NiO. Sprays with ALT were made by adding into the mix sufficient quantities of ALT to make 1, 2.5, 5, and 10 wt% ALT-doped (ALT, Alfa Aesar) anode mix powder. Each suspension was thoroughly mixed in a 35 wt% xylene, 35 wt% ethanol solvent mixture using 2 wt% of copolymer dispersant (KD-1). Additionally, 2 wt% of 200 molecular weight polyethylene-glycol (PEG 200), 3 wt% of polyvinyl butyral (B98), and 1.5 wt% of butyl benzyl phthalate (S160) were added as binders and plasticizers. The dispersant was dissolved in the solvent ultrasonically (Branson Sonifier 450) before the addition of any other ingredient. The addition of the appropriate amount of ALT to the spray mixture occurred after dissolution of the dispersant, and before the addition of any other powder. The 8YSZ, NiO, and pore former were ultrasonically mixed in the suspension after the dispersant was dissolved. Binders and plasticizers were then added, the suspension was ultrasonically mixed a third time, then ball milled in a high-density polyethylene bottle with cylindrical YSZ milling media for 12 hours to facilitate complete homogenization of the mixture. Cell anodes were applied via airbrush (Badger air-brush Model No. 360-7) onto weighed (Sartorius CPA 225D), and appropriately masked electrolytes. The anode-electrolyte bilayers were then allowed to dry in air overnight before being sintered at 1250 C for two hours with heating and cooling rates of 10 C per minute. Following sintering, the mass of each bi-layer was recorded for ALT-infiltration calculations. ALT was added to the anode by either infiltration of an ALT solution, or by mechanical mixing of ALT powder into the anode spray, as described above. The ALT solution used for infiltration was prepared by mixing stoichiometric amounts of aluminum nitrate, and titanium lactate, both dissolved in deionized water, to produce Al2TiO5. The resulting ALT solution was added to the anode by pipetting a drop of the ALT solution onto the anode, allowing the water to evaporate, and removing organics by placing the infiltrated cell into a 400 C furnace for about two minutes. The mass of the doped anodeelectrolyte bilayer was then recorded to determine dopant mass. This process was repeated until desired ALT-doping concentrations were approximated. Doped and non-doped cells were then heated to 1400 C with heating and cooling
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
rates of 5 C per minute to activate the ALT, and to ensure valid comparison [23].
Cathodes The cathode spray was made by mixing appropriate quantities of 8YS, lanthanum-doped strontium manganite (LSM), and pore former (corn starch) to form an equal-volume mixture of 8YS, LSM, and pore former. Based on the total mass of the powders comprising the cathode, 3 wt% KD-1 was added as a dispersant, 3 wt% of PEG 200, 4.7 wt% of B98, and 2.4 wt% of S160 were added as binders and plasticizers. The solvent for the suspension was an equal-parts mixture of about 45 wt% xylene, and 45 wt% ethanol. Powders were added to the solvent, ultrasonically mixed, then ball milled in the same manner as the anode spray. Cathodes were added to the cells by airbrush application of the cathode spray. After drying, the cells were heated to and held at 900 C for two hours with heating and cooling rates of 10 C per minute.
Electrochemical degradation testing and quantification Completed fuel cells were tested in a custom-built fuel-cell test rig, which is depicted in the work of Lussier et al. [36]. Cells were loaded between clamshell-furnace-enclosed test rig platens. Silver-wire weave and nickel foam served as current collectors for the cathode and anode sides of the cell, respectively. Electrical connection between the cell and the current collectors, and the current collectors and the platens was made with silver-oxide paint. Once assembled, the entire fuel cell test rig was heated to 800 C. The operating temperature of the test rig yields thermodynamic reduction of silver oxide on the both the air and fuel sides of the cell, facilitating a porous, electrically conductive layer of silver to connect each component of the fuel-cell test setup to its adjacent component for excellent current collection. Note that the test rig is connected to a computer that controls and measures the flow rates of the hydrogen and nitrogen that are delivered to the anode side of the cell, the air that is supplied to the cathodeside of the cell, and the electrical current output of the cell via a LabVIEW program. The fuel cell current I as a function of time t was fit to Equation (1) using an in-house MATLAB algorithm. rffiffiffiffiffiffiffiffiffiffiffiffi 1 t 1 þ a1 c1 $ þ a2 c2 $tanðat bÞ tþ1 t0 t IMax þ a3 c3 $ 1 þ expðat þ bÞ
IðtÞ ¼ a0 c0 $
(1)
The user-defined constants a0 , a1 , a2 , and a3 were chosen as either 1 or 0 and served as “on” or “off” switches for each term in Equation (1). If a2 was set to 1, a3 was set to zero, and vice versa. The constant t0 is increased by a set interval each time the appropriate loop of the algorithm is run. The algorithm then determines a value of the coefficient of determination ðR2 Þ for each value of t0 . If the R2 value of any fit was found to be better than the current, best R2 value, then the associated value of t0 was kept as the best value of t0 . If the R2 value associated with any value of t0 was found to be less than the best R2 value, then the value of t0 was assigned to be the value of t0 associated with the best recorded R2 value.
3
The best combination of terms for fitting Equation (1) to a data set was determined via trial and error. Particularly, the selection of the inverse-tangent function or the logistic equation required more than one attempt at fitting each data set. The constants a and b were determined using an iterative MATLAB scheme. This scheme first determines an optimum value of b using a fixed value of a and five different values of b. The least squares method of curve-fitting was then used to determine values of the constants c0 , c1 , and c2 or c3 for each value of b. The corresponding R2 values for each value of b were found. The relationship defined by the ðb; R2 Þ points was used to define the equation of a parabola given by R2 ¼ A2 $b2 þ A1 $b þ A0 . Once determined, the coefficients of this parabola were used to define an optimum value of b in Equation (1), given by b ¼ A1 =ð2$A2 Þ. This optimum value of b was then used as the reference point for defining the next set of five b values. This process was iterated until the change in the R2 value became less than a user-specified value. This process was then used to determine an optimum value of a. Because changing the value of a changed the optimum value of b, the whole process was iterated several times. The degradation rate, RðtÞ ¼ dIðtÞ dt , of the current is given as Equation (2). RðtÞ ¼
pffiffiffiffiffiffiffiffiffiffiffiffi a0 c0 a1 c1 $t0 t0 t a2 c2 $a þ pffiffi þ 2 2 t 2 t$ðt t0 Þ2 ðb atÞ þ 1 þ
a3 c3 $a$IMax $expðb atÞ 2
ð1 þ expðb atÞÞ
(2)
The degradation rate of each cell was then normalized with respect to measured current, Imeas ðtÞ. This normalized degradation rate, RN ðtÞ, is described mathematically as Equation (3). RN ðtÞ ¼
RðtÞ Imeas ðtÞ
(3)
Results Equation (1) was fit to fuel cell data from all cells run. Degradation rates were determined according to Equation (2), and recorded at 5, 10, 15, and 20 hours for the cells made with the ~4∙mm precursor powder with ALT addition via either infiltration or mechanical-mixing. Degradation rates were recorded at 12.5, 25, 37.5, and 50 hours, and at 25, 50, 75, and 100 hours for cells made with the ~3∙mm precursor powder. Degradation rates taken at the indicated times were plotted against dopant concentration. Positive and negative values indicate performance gains and losses, respectively. Comparison of performance of cells with ALT added via infiltration and mechanical mixing indicated that, in all cases, ALT added via infiltration resulted in favorable degradation rates. Microstructure images (not shown) suggest better preservation of microstructure of cells with ALT added via infiltration when compared to that of cells with ALT added via mechanical mixing. Based on those results, study of the effect of ALT addition via mechanical mixing between 1 and 5∙Wt.% was deemed unnecessary.
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
Fig. 1 e Left: Degradation rates of cells with ALT-infiltrated, ~4mm initial-particle-size NiO anodes taken at 5, 10, 15, and 20 h with (inset) plots of second-order polynomial fits of each data set with optimum ALT amount indicated as 5.7 ± 0.7 wt% ALT. Right: Degradation rates of cells with ALT addition via mechanical mixing.
Alfa Aesar NiO Fig. 1 compares the performance of cells with anodes made from the ~4∙mm NiO precursor powder. Degradation rates of cells with ALT added via infiltration have been shown on the left side of Fig. 1, whereas the degradation rates of cells with ALT added via mechanical mixing of ALT powder into the spray suspension have been shown on the right side of Fig. 1. Although the addition of ALT to a fuel-cell anode via mechanical mixing is seen to decrease cell degradation rate, performance of cells with ALT added via mechanical mixing are not seen to improve at any point during testing. Comparison of degradation rates of cells with ALT added by mechanical mixing with degradation rates of cells with ALT added by infiltration indicates ALT addition to cell anodes via infiltration is more effective at improving cell-degradation rates than ALT addition via mechanical mixing of powder
precursors. The explanation of this observation is currently the subject of additional work. The plot shown on the left-hand side of Fig. 1 suggests that a parabolic relationship between ALT concentration and cell degradation rates exist. Second-order polynomial curve-fits of degradation rates against dopant concentration at 5, 10, 15, and 20 h is shown as the inset of the left-hand side of Fig. 1. This result indicates that there is an optimum amount of ALT for electrochemical stabilization. The average peak location of the 5, 10, 15, and 20-h fits was found to be 5.7 ± 0.7 wt% ALT, within the 95% confidence interval.
J.T. Baker NiO Fuel cell performance of cells made with green NiO with initial particle size of ~3∙mm (J.T. Baker) is summarized in Fig. 2. The left-side plot of Fig. 2 shows cell degradation rates at 12.5, 25,
Fig. 2 e Variation of degradation rates of cells with anodes made from NiO precursors with ~3mm initial particle size at (left) 12.5, 25, 37.5, and 50 h and (right) 25, 50, 75, and 100 hours Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
37.5, and 50 hours. These results suggest that cell improvement rates of about 4.5% per hour were measured at 12.5 h of cell life, that degradation rates with the cells made from ~3∙mm precursor particle size NiO are substantially different from the degradation rates of the cells made from ~4∙mm precursor particle size NiO, and that the addition of ALT via infiltration to cell anodes with ~3∙mm precursor particle size is an effective means of stabilization of cell performance. The right-side plot of Fig. 2 indicates cell improvement rates at 50 h with the addition of 4 wt% ALT to the cell anode. This is consistent with the results of the left side of Fig. 2.
[2]
[3]
[4]
Mechanism of stabilization Fukui et al. report on a method of stabilization of Ni in YSZ anodes that is thought to be similar to the mechanism of stabilization considered in this work [37]. The coarsening of nickel that occurs in SOFC anodes is a result of the growth of larger nickel metal particles at the expense of smaller nickel metal particles. Decoration of nickel metal particles with smaller ceramic particles is currently thought to inhibit coarsening by requiring any mass transfer to result in the formation of dimples on the donor particle. Formation of these dimples is thought to increase surface area of the donor particle. Increasing surface area implies increasing surface energy. The spontaneous increase of energy is thermodynamically forbidden. Details of this stabilization mechanism are the subject of a forthcoming report.
[5]
[6]
[7]
[8]
[9]
Conclusions Results suggest that fuel cell degradation rate depends greatly on the precursor materials used in cell fabrication. It is also observed that the addition of ALT to the traditional NiO-YSZ anode modifies fuel cell performance in cells made with each precursor NiO powder. The method of ALT introduction is seen to influence SOFC performance, where measurements of a cell's output of current at a constant voltage indicate ALT addition via infiltration is a more effective means of improving cell performance than ALT addition via the mechanical mixing of ALT powder with other anode precursors before anode application. Finally, an optimum amount of ALT in SOFCs made with ~4∙mm NiO precursor particle size is proposed as 5.7 ± 0.7 wt.% ALT. Estimation of this optimum ALT amount was facilitated by a novel method of cell degradation-rate calculation. The algorithm used to determine degradation rates is available upon request.
[10]
[11]
[12]
[13]
[14]
Acknowledgements [15]
This work was supported by the DOE under grant #FE000261.
references
[1] Park SY, Ahn JH, Jeong CW, Na CW, Song RH, Lee JH. Ni-YSZsupported tubular solid oxide fuel cells with GDC interlayer
[16]
5
between YSZ electrolyte and LSCF cathode. Int J Hydrogen Energy 2014;39:12894e903. https://doi.org/10.1016/ j.ijhydene.2014.06.103. Powell M, Meinhardt K, Sprenkle V, Chick L, McVay G. Demonstration of a highly efficient solid oxide fuel cell power system using adiabatic steam reforming and anode gas recirculation. J Power Sources 2012;205:377e84. https:// doi.org/10.1016/j.jpowsour.2012.01.098. Choudhury A, Chandra H, Arora A. Application of solid oxide fuel cell technology for power generation e a review. Renew Sustain Energy Rev 2013;20:430e42. https://doi.org/10.1016/ j.rser.2012.11.031. da Silva FS, de Souza TM. Novel materials for solid oxide fuel cell technologies: a literature review. Int J Hydrogen Energy 2017;42:26020e36. https://doi.org/10.1016/ j.ijhydene.2017.08.105. Iwata T. Characterization of Ni-YSZ anode degradation for substrate-type solid oxide fuel cells. J Electrochem Soc 1996;143:225e625. https://doi.org/10.1149/1.1836673. Montenegro-Hernandez A, Vega-Castillo J, Mogni L, Caneiro A. Thermal stability of Ln2NiO4þ?? (Ln: La, Pr, Nd) and their chemical compatibility with YSZ and CGO solid electrolytes. Int J Hydrogen Energy 2011;36:15704e14. https:// doi.org/10.1016/j.ijhydene.2011.08.105. Ananyev MV, Farlenkov AS, Eremin VA, Kurumchin EK. Degradation kinetics of LSMeYSZ cathode materials for SOFC. Int J Hydrogen Energy 2018;43:951e9. https://doi.org/ 10.1016/j.ijhydene.2017.11.107. Kim SJ, Kim KJ, Dayaghi AM, Choi GM. Polarization and stability of La2NiO4þd in comparison with La0.6Sr0.4Co0.2Fe0.8O3d as air electrode of solid oxide electrolysis cell. Int J Hydrogen Energy 2016;41:14498e506. https://doi.org/10.1016/j.ijhydene.2016.05.284. Bo Y, Wenqiang Z, Jingming X, Jing C. Microstructural characterization and electrochemical properties of Ba0.5Sr0.5Co0.8Fe0.2O3-dand its application for anode of SOEC. Int J Hydrogen Energy 2008;33:6873e7. https://doi.org/ 10.1016/j.ijhydene.2008.07.066. Yokokawa H, Tu H, Iwanschitz B, Mai A. Fundamental mechanisms limiting solid oxide fuel cell durability. J Power Sources 2008;182:400e12. https://doi.org/10.1016/ j.jpowsour.2008.02.016. Shri Prakash B, Senthil Kumar S, Aruna ST. Properties and development of Ni/YSZ as an anode material in solid oxide fuel cell: a review. Renew Sustain Energy Rev 2014;36:149e79. https://doi.org/10.1016/j.rser.2014.04.043. Khan MS, Lee SB, Song RH, Lee JW, Lim TH, Park SJ. Fundamental mechanisms involved in the degradation of nickel-yttria stabilized zirconia (Ni-YSZ) anode during solid oxide fuel cells operation: a review. Ceram Int 2015;42:35e48. https://doi.org/10.1016/j.ceramint.2015.09.006. Shaikh SPS, Muchtar A, Somalu MR. A review on the selection of anode materials for solid-oxide fuel cells. Renew Sustain Energy Rev 2015;51:1e8. https://doi.org/10.1016/ j.rser.2015.05.069. Presto S, Kumar P, Varma S, Viviani M, Singh P. Electrical conductivity of NiMoebased double perovskites under SOFC anodic conditions. Int J Hydrogen Energy 2018;43:4528e33. https://doi.org/10.1016/j.ijhydene.2018.01.066. Weisentein AJ, Childs N, Amendola R, Driscoll D, Sofie SW, Gannon P, et al. Processing and characterization of Sr2xVMoO6-y double perovskites. Mater Chem Phys 2013;139:706e18. https://doi.org/10.1016/ j.matchemphys.2013.02.021. Childs NB, Wesinstein A, Smith R, Sofie S, Key C. Electrical conductivity of Sr2-xVMoO6-y (x ¼ 0.0, 0.1, 0.2) double perovskites. J Appl Phys 2013;113:0e9. https://doi.org/ 10.1063/1.4811715.
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
[17] Yu R, Guan W, Wang F, Han F. Quantitative assessment of anode contribution to cell degradation under various polarization conditions using industrial size planar solid oxide fuel cells. Int J Hydrogen Energy 2018;43:2429e35. https://doi.org/10.1016/j.ijhydene.2017.12.030. [18] Kennouche D, Chen-Wiegart YCK, Yakal-Kremski KJ, Wang J, Gibbs JW, Voorhees PW, et al. Observing the microstructural evolution of Ni-Yttria-stabilized zirconia solid oxide fuel cell anodes. Acta Mater 2016;103:204e10. https://doi.org/10.1016/ j.actamat.2015.09.055. [19] Lee YH, Muroyama H, Matsui T, Eguchi K. Degradation of nickel-yttria-stabilized zirconia anode in solid oxide fuel cells under changing temperature and humidity conditions. J Power Sources 2014;262:451e6. https://doi.org/10.1016/ j.jpowsour.2014.03.031. [20] Lee K, Kang J, Lee J, Lee S, Bae J. Evaluation of metalsupported solid oxide fuel cells (MS-SOFCs) fabricated at low temperature (~1,000 C) using wet chemical coating processes and a catalyst wet impregnation method. Int J Hydrogen Energy 2018;43:3786e96. https://doi.org/10.1016/ j.ijhydene.2018.01.027. [21] Ding C, Hashida T. Synthesis and evaluation of NiOeCe0.8Sm0.2O1.9 nanocomposite powders for lowtemperature solid oxide fuel cells. Int J Hydrogen Energy 2011;36:5567e73. https://doi.org/10.1016/ j.ijhydene.2011.02.073. [22] Kim S, Kim C, Lee JH, Shin J, Lim TH, Kim G. Tailoring Nibased catalyst by alloying with transition metals (M ¼ Ni, Co, Cu, and Fe) for direct hydrocarbon utilization of energy conversion devices. Electrochim Acta 2017;225:399e406. https://doi.org/10.1016/j.electacta.2016.12.178. [23] Driscoll DR, McIntyre MD, Welander MM, Sofie SW, Walker RA. Enhancement of high temperature metallic catalysts: aluminum titanate in the nickel-zirconia system. Appl Catal A Gen 2016;527:36e44. https://doi.org/10.1016/ j.apcata.2016.08.020. [24] Driscoll D, Hunt C, Muretta J, Sofie S. Thermally stabilized nickel electro-catalyst introduced by infiltration for high temperature electrochemical energy conversion. ECS Trans 2013;53:1e36. [25] German RM. Sintering theory and practice, vol. 1; 1996. [26] Zubair M, Taqi M, Song R, Lee J, Lee S, Lim T. A simplified approach to predict performance degradation of a solid oxide fuel cell anode. J Power Sources 2018;391:94e105. https:// doi.org/10.1016/j.jpowsour.2018.04.080. [27] Chaim R. Activation energy and grain growth in nanocrystalline Y-TZP ceramics. Mater Sci Eng A 2008;486:439e46. https://doi.org/10.1016/j.msea.2007.09.022.
[28] Hossain S, Abdalla AM, Radenahmad N, Zakaria AKM, Zaini JH, Rahman SMH, et al. Highly dense and chemically stable proton conducting electrolyte sintered at 1200 C. Int J Hydrogen Energy 2018;43:894e907. https://doi.org/10.1016/ j.ijhydene.2017.11.111. [29] Osinkin DA. Degradation of Ni-Zr0.9Sc0.1O1.95anode in H2þ H2O at low temperature: influence of nickel surface charge. Int J Hydrogen Energy 2018;43:943e50. https://doi.org/ 10.1016/j.ijhydene.2017.11.071. [30] Laurencin J, Hubert M, Sanchez DF, Pylypko S, Morales M, Morata A, et al. Degradation mechanism of La0.6Sr0.4Co0.2Fe0.8O3-d/Gd0.1Ce0.9O2-dcomposite electrode operated under solid oxide electrolysis and fuel cell conditions. Electrochim Acta 2017;241:459e76. https:// doi.org/10.1016/j.electacta.2017.05.011. [31] Iwai H, Shikazono N, Matsui T, Teshima H, Kishimoto M, Kishida R, et al. Quantification of SOFC anode microstructure based on dual beam FIB-SEM technique. J Power Sources 2010;195:955e61. https://doi.org/10.1016/ j.jpowsour.2009.09.005. [32] Wilson JR, Cronin JS, Barnett SA. Linking the microstructure, performance and durability of Ni-yttria-stabilized zirconia solid oxide fuel cell anodes using three-dimensional focused ion beam-scanning electron microscopy imaging. Scripta Mater 2011;65:67e72. https://doi.org/10.1016/ j.scriptamat.2010.09.025. [33] Wang F, Wang W, Ran R, Tade MO, Shao Z. Aluminum oxide as a dual-functional modifier of Ni-based anodes of solid oxide fuel cells for operation on simulated biogas. J Power Sources 2014;268:787e93. https://doi.org/10.1016/ j.jpowsour.2014.06.087. [34] Welander MM, Zachariasen MS, Hunt CD, Sofie SW, Walker RA. Operando studies of redox resilience in ALT enhanced NiO-YSZ SOFC anodes. J Electrochem Soc 2018;165:F152e7. https://doi.org/10.1149/2.0531803jes. [35] Driscoll DR, Tokmakian BW, Johnson NP, Hunt CD, Sofie SW. Catalyst enhancing aluminum titanate for increasing strength of nickel-zirconia cermets. Mater Lett 2017;209:307e10. https://doi.org/10.1016/j.matlet.2017.08.035. [36] Lussier A, Sofie S, Dvorak J, Idzerda YU. Mechanism for SOFC anode degradation from hydrogen sulfide exposure. Int J Hydrogen Energy 2008;33:3945e51. https://doi.org/10.1016/ j.ijhydene.2007.11.033. [37] Fukui T, Ohara S, Naito M, Nogi K. Performance and stability of {SOFC} anode fabricated from {NiO{ytextendash}YSZ} composite particles. J Power Sources 2002;110:91e5. https:// doi.org/10.1016/S0378-7753(02)00218-5.
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Short Communication
Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition Clay Hunt a,*, Marley Zachariasen a, David Driscoll a, Stephen Sofie a, Robert Walker b a b
Department of Mechanical and Industrial Engineering, Montana State University, Bozeman, MT, 59717, USA Department of Chemistry & Biochemistry, Montana State University, Bozeman, MT, 59717, USA
article info
abstract
Article history:
Degradation rates of electrical current during constant voltage operation of SOFCs with
Received 16 April 2018
anodes made using NiO precursor powders from two different manufacturers with and
Received in revised form
without the addition of aluminum titanate (ALT) added by either mechanical mixing or
10 June 2018
anode infiltration have been quantified using a novel MATLAB algorithm. Because the al-
Accepted 19 June 2018
gorithm has been used to quantify degradation rates for many different SOFC tests, it is
Available online xxx
thought that the method can be applied to most measured SOFC data to quantify the instantaneous cell degradation rate as a function of time for the entire SOFC performance
Keywords:
measurement. Degradation rates determined at different times have been plotted against
Aluminum titanate
varying concentrations of ALT addition, facilitating the estimation of optimum ALT con-
Modeling
centration for SOFC anodes made with NiO from a specific manufacturer. The algorithm
Solid oxide fuel cells
used to determine degradation rates is available upon request to the corresponding author.
Degradation rate quantification
© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Optimization
Introduction Solid oxide fuel cell (SOFC) technology is improving every day. This promising means of energy conversion is fuel efficient, clean, and approaching commercialization [1]. When both electrical energy and heat are harvested from a solid oxide fuel cell (SOFC) stack, thermodynamic efficiency can be as high as 90% [2,3]. The main problem to be overcome before this technology becomes a viable source of electricity is the degradation occurring in the electrodes of the cell in the SOFC operating environment [4e12]. Because most SOFC degradation is attributed to the SOFC anode, great efforts have been
made to stabilize traditional Ni yttria-stabilized zirconia (YSZ) anodes, and to develop entirely ceramic anodes [13e19]. Decreasing SOFC operating temperature is also a strategy in reducing cell degradation. This is pursued through the use of SOFCs made with the smallest possible electrode particles, and with metal-supported SOFCs [20]. The report of Ding and Hashida describes the synthesis of a nano-composite NiO-Ce0.8Sm0.2O1.9 powder [21]. Anode stabilization against carbon coking and nickel coarsening through alloying the Ni in the SOFC anode with another transition metal has been reported by Kim et al. [22]. Electrodes made with small particles, or even from solution infiltration of electrode scaffolds are interesting because
* Corresponding author. E-mail address: [email protected] (C. Hunt). https://doi.org/10.1016/j.ijhydene.2018.06.115 0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
the triple-phase boundary length of an electrode increases as electrode particle size decreases [23,24]. However, decreasing particle sizes are associated with more rapid nickel coarsening because the driving force of sintering is inversely proportional to particle size [25]. The work of Khan et al. has proposed a model for nickel-particle agglomeration occurring in Ni-YSZ anodes that is similar to equations concerned with grain size in sintering ceramics [26e28]. Cell degradation that is not consistent with Khan et al.'s predictions might be attributable to other means of cell degradation [29,30]. These results suggest that a means of quantifying degradation rate could be useful in identifying the way in which an SOFC is degrading. Amazing techniques exist for modeling the effects on the triple-phase boundary of nickel migration in the anode of an SOFC [31,32]. However, a method of quantification of the instantaneous degradation rate of a SOFC is not the subject of much attention despite degradation being a subject of interest. The main challenge in the quantification of SOFC degradation rates is that measured electrical power output of a SOFC is difficult to differentiate in a way that provides meaningful results. The work of Wang et al., attempts to fit first-order polynomials to cell voltage that was measured during constant-current cell operation [33]. Although the lines used fit the data in the report by Wang et al., the shape of the such a voltage curve is, in general, not linear with respect to time. Only the recent report of Welander et al. uses a method of curve-fitting that can fit a larger set of fuel-cell degradation data [34]. This report elaborates on the design and merit of the technique used in the previous work and further applies the method of degradation quantification. The impetus for developing this tool was the need to quantitatively ascertain optimum doping levels of any coarsening inhibitor for Ni-based anode catalysts [23,24,35]. Wang et al.'s method of fitting measured voltage to a line demonstrates the idea of fitting measured data to a curve to determine degradation rates from measured voltage. The method discussed in this report is simply an extension of that idea. This method can be used to quantify degradation rates of most fuel cells with data sets that record changing voltage at constant current or changing current at constant voltage. This work considers the application of this algorithm to measure the effect of the addition of Al2TiO5 (ALT) to the anodes of SOFCs made with ~3∙mm (J.T. Baker) and ~4∙mm (Alpha Aesar) initial particle size NiO on fuel cell performance. Original application of ALT to the standard SOFC anode material, the Ni-YSZ cermet, was done to engineer the CTE of said cermet. Increased resistance to fracture of cermet bars with ALT was casually observed. This observation has been the subject of the recent work of Driscoll et al. [35]. The casual observation of anode cermet resistance to fracture also motivated the 2013 study by Driscoll et al. [24]. Results of this work suggest ALT might act as a sintering inhibitor for nickel in the anode cermet under typical SOFC operating conditions. Results of the 2013 Driscoll et al. study motivated the in-depth analysis of ALT in the anode cermet reported by Driscoll et al., in 2016 [23]. Results of this work detail the effects of thermal treatment on ALT in the cermet as studied with electrical impedance spectroscopy, X-ray diffraction, electron microscopy, and voltammetry. These results suggested the effects
of ALT on cell performance might be an interesting application of the degradation-rate algorithm.
Experimental procedure Anodes Anodes were applied to 32∙mm Fuel Cell Materials electrolytes by aerosol application of homogenized suspensions of appropriate quantities of the spherical grade 8YSZ (8YS) (8YSZ (YS) Tosoh Corp., Tokyo, Japan), NiO with initial particle size of about 3∙mm (JT Baker NiO) or NiO with initial particle size of about 4∙mm (Green NiO, Alfa Aesar), and corn starch, used as a pore forming agent to yield an equal-volume mixture of 8YSZ, and nickel metal after anode reduction in the fuel cell operating environment. The anode spray consisted of about 32.4 wt% YSZ, 62.2 wt% NiO, and 5.4 wt% pore former, where total powder mass is given by the sum of the masses of the pore former, YSZ, and NiO. Sprays with ALT were made by adding into the mix sufficient quantities of ALT to make 1, 2.5, 5, and 10 wt% ALT-doped (ALT, Alfa Aesar) anode mix powder. Each suspension was thoroughly mixed in a 35 wt% xylene, 35 wt% ethanol solvent mixture using 2 wt% of copolymer dispersant (KD-1). Additionally, 2 wt% of 200 molecular weight polyethylene-glycol (PEG 200), 3 wt% of polyvinyl butyral (B98), and 1.5 wt% of butyl benzyl phthalate (S160) were added as binders and plasticizers. The dispersant was dissolved in the solvent ultrasonically (Branson Sonifier 450) before the addition of any other ingredient. The addition of the appropriate amount of ALT to the spray mixture occurred after dissolution of the dispersant, and before the addition of any other powder. The 8YSZ, NiO, and pore former were ultrasonically mixed in the suspension after the dispersant was dissolved. Binders and plasticizers were then added, the suspension was ultrasonically mixed a third time, then ball milled in a high-density polyethylene bottle with cylindrical YSZ milling media for 12 hours to facilitate complete homogenization of the mixture. Cell anodes were applied via airbrush (Badger air-brush Model No. 360-7) onto weighed (Sartorius CPA 225D), and appropriately masked electrolytes. The anode-electrolyte bilayers were then allowed to dry in air overnight before being sintered at 1250 C for two hours with heating and cooling rates of 10 C per minute. Following sintering, the mass of each bi-layer was recorded for ALT-infiltration calculations. ALT was added to the anode by either infiltration of an ALT solution, or by mechanical mixing of ALT powder into the anode spray, as described above. The ALT solution used for infiltration was prepared by mixing stoichiometric amounts of aluminum nitrate, and titanium lactate, both dissolved in deionized water, to produce Al2TiO5. The resulting ALT solution was added to the anode by pipetting a drop of the ALT solution onto the anode, allowing the water to evaporate, and removing organics by placing the infiltrated cell into a 400 C furnace for about two minutes. The mass of the doped anodeelectrolyte bilayer was then recorded to determine dopant mass. This process was repeated until desired ALT-doping concentrations were approximated. Doped and non-doped cells were then heated to 1400 C with heating and cooling
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
rates of 5 C per minute to activate the ALT, and to ensure valid comparison [23].
Cathodes The cathode spray was made by mixing appropriate quantities of 8YS, lanthanum-doped strontium manganite (LSM), and pore former (corn starch) to form an equal-volume mixture of 8YS, LSM, and pore former. Based on the total mass of the powders comprising the cathode, 3 wt% KD-1 was added as a dispersant, 3 wt% of PEG 200, 4.7 wt% of B98, and 2.4 wt% of S160 were added as binders and plasticizers. The solvent for the suspension was an equal-parts mixture of about 45 wt% xylene, and 45 wt% ethanol. Powders were added to the solvent, ultrasonically mixed, then ball milled in the same manner as the anode spray. Cathodes were added to the cells by airbrush application of the cathode spray. After drying, the cells were heated to and held at 900 C for two hours with heating and cooling rates of 10 C per minute.
Electrochemical degradation testing and quantification Completed fuel cells were tested in a custom-built fuel-cell test rig, which is depicted in the work of Lussier et al. [36]. Cells were loaded between clamshell-furnace-enclosed test rig platens. Silver-wire weave and nickel foam served as current collectors for the cathode and anode sides of the cell, respectively. Electrical connection between the cell and the current collectors, and the current collectors and the platens was made with silver-oxide paint. Once assembled, the entire fuel cell test rig was heated to 800 C. The operating temperature of the test rig yields thermodynamic reduction of silver oxide on the both the air and fuel sides of the cell, facilitating a porous, electrically conductive layer of silver to connect each component of the fuel-cell test setup to its adjacent component for excellent current collection. Note that the test rig is connected to a computer that controls and measures the flow rates of the hydrogen and nitrogen that are delivered to the anode side of the cell, the air that is supplied to the cathodeside of the cell, and the electrical current output of the cell via a LabVIEW program. The fuel cell current I as a function of time t was fit to Equation (1) using an in-house MATLAB algorithm. rffiffiffiffiffiffiffiffiffiffiffiffi 1 t 1 þ a1 c1 $ þ a2 c2 $tanðat bÞ tþ1 t0 t IMax þ a3 c3 $ 1 þ expðat þ bÞ
IðtÞ ¼ a0 c0 $
(1)
The user-defined constants a0 , a1 , a2 , and a3 were chosen as either 1 or 0 and served as “on” or “off” switches for each term in Equation (1). If a2 was set to 1, a3 was set to zero, and vice versa. The constant t0 is increased by a set interval each time the appropriate loop of the algorithm is run. The algorithm then determines a value of the coefficient of determination ðR2 Þ for each value of t0 . If the R2 value of any fit was found to be better than the current, best R2 value, then the associated value of t0 was kept as the best value of t0 . If the R2 value associated with any value of t0 was found to be less than the best R2 value, then the value of t0 was assigned to be the value of t0 associated with the best recorded R2 value.
3
The best combination of terms for fitting Equation (1) to a data set was determined via trial and error. Particularly, the selection of the inverse-tangent function or the logistic equation required more than one attempt at fitting each data set. The constants a and b were determined using an iterative MATLAB scheme. This scheme first determines an optimum value of b using a fixed value of a and five different values of b. The least squares method of curve-fitting was then used to determine values of the constants c0 , c1 , and c2 or c3 for each value of b. The corresponding R2 values for each value of b were found. The relationship defined by the ðb; R2 Þ points was used to define the equation of a parabola given by R2 ¼ A2 $b2 þ A1 $b þ A0 . Once determined, the coefficients of this parabola were used to define an optimum value of b in Equation (1), given by b ¼ A1 =ð2$A2 Þ. This optimum value of b was then used as the reference point for defining the next set of five b values. This process was iterated until the change in the R2 value became less than a user-specified value. This process was then used to determine an optimum value of a. Because changing the value of a changed the optimum value of b, the whole process was iterated several times. The degradation rate, RðtÞ ¼ dIðtÞ dt , of the current is given as Equation (2). RðtÞ ¼
pffiffiffiffiffiffiffiffiffiffiffiffi a0 c0 a1 c1 $t0 t0 t a2 c2 $a þ pffiffi þ 2 2 t 2 t$ðt t0 Þ2 ðb atÞ þ 1 þ
a3 c3 $a$IMax $expðb atÞ 2
ð1 þ expðb atÞÞ
(2)
The degradation rate of each cell was then normalized with respect to measured current, Imeas ðtÞ. This normalized degradation rate, RN ðtÞ, is described mathematically as Equation (3). RN ðtÞ ¼
RðtÞ Imeas ðtÞ
(3)
Results Equation (1) was fit to fuel cell data from all cells run. Degradation rates were determined according to Equation (2), and recorded at 5, 10, 15, and 20 hours for the cells made with the ~4∙mm precursor powder with ALT addition via either infiltration or mechanical-mixing. Degradation rates were recorded at 12.5, 25, 37.5, and 50 hours, and at 25, 50, 75, and 100 hours for cells made with the ~3∙mm precursor powder. Degradation rates taken at the indicated times were plotted against dopant concentration. Positive and negative values indicate performance gains and losses, respectively. Comparison of performance of cells with ALT added via infiltration and mechanical mixing indicated that, in all cases, ALT added via infiltration resulted in favorable degradation rates. Microstructure images (not shown) suggest better preservation of microstructure of cells with ALT added via infiltration when compared to that of cells with ALT added via mechanical mixing. Based on those results, study of the effect of ALT addition via mechanical mixing between 1 and 5∙Wt.% was deemed unnecessary.
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
Fig. 1 e Left: Degradation rates of cells with ALT-infiltrated, ~4mm initial-particle-size NiO anodes taken at 5, 10, 15, and 20 h with (inset) plots of second-order polynomial fits of each data set with optimum ALT amount indicated as 5.7 ± 0.7 wt% ALT. Right: Degradation rates of cells with ALT addition via mechanical mixing.
Alfa Aesar NiO Fig. 1 compares the performance of cells with anodes made from the ~4∙mm NiO precursor powder. Degradation rates of cells with ALT added via infiltration have been shown on the left side of Fig. 1, whereas the degradation rates of cells with ALT added via mechanical mixing of ALT powder into the spray suspension have been shown on the right side of Fig. 1. Although the addition of ALT to a fuel-cell anode via mechanical mixing is seen to decrease cell degradation rate, performance of cells with ALT added via mechanical mixing are not seen to improve at any point during testing. Comparison of degradation rates of cells with ALT added by mechanical mixing with degradation rates of cells with ALT added by infiltration indicates ALT addition to cell anodes via infiltration is more effective at improving cell-degradation rates than ALT addition via mechanical mixing of powder
precursors. The explanation of this observation is currently the subject of additional work. The plot shown on the left-hand side of Fig. 1 suggests that a parabolic relationship between ALT concentration and cell degradation rates exist. Second-order polynomial curve-fits of degradation rates against dopant concentration at 5, 10, 15, and 20 h is shown as the inset of the left-hand side of Fig. 1. This result indicates that there is an optimum amount of ALT for electrochemical stabilization. The average peak location of the 5, 10, 15, and 20-h fits was found to be 5.7 ± 0.7 wt% ALT, within the 95% confidence interval.
J.T. Baker NiO Fuel cell performance of cells made with green NiO with initial particle size of ~3∙mm (J.T. Baker) is summarized in Fig. 2. The left-side plot of Fig. 2 shows cell degradation rates at 12.5, 25,
Fig. 2 e Variation of degradation rates of cells with anodes made from NiO precursors with ~3mm initial particle size at (left) 12.5, 25, 37.5, and 50 h and (right) 25, 50, 75, and 100 hours Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
37.5, and 50 hours. These results suggest that cell improvement rates of about 4.5% per hour were measured at 12.5 h of cell life, that degradation rates with the cells made from ~3∙mm precursor particle size NiO are substantially different from the degradation rates of the cells made from ~4∙mm precursor particle size NiO, and that the addition of ALT via infiltration to cell anodes with ~3∙mm precursor particle size is an effective means of stabilization of cell performance. The right-side plot of Fig. 2 indicates cell improvement rates at 50 h with the addition of 4 wt% ALT to the cell anode. This is consistent with the results of the left side of Fig. 2.
[2]
[3]
[4]
Mechanism of stabilization Fukui et al. report on a method of stabilization of Ni in YSZ anodes that is thought to be similar to the mechanism of stabilization considered in this work [37]. The coarsening of nickel that occurs in SOFC anodes is a result of the growth of larger nickel metal particles at the expense of smaller nickel metal particles. Decoration of nickel metal particles with smaller ceramic particles is currently thought to inhibit coarsening by requiring any mass transfer to result in the formation of dimples on the donor particle. Formation of these dimples is thought to increase surface area of the donor particle. Increasing surface area implies increasing surface energy. The spontaneous increase of energy is thermodynamically forbidden. Details of this stabilization mechanism are the subject of a forthcoming report.
[5]
[6]
[7]
[8]
[9]
Conclusions Results suggest that fuel cell degradation rate depends greatly on the precursor materials used in cell fabrication. It is also observed that the addition of ALT to the traditional NiO-YSZ anode modifies fuel cell performance in cells made with each precursor NiO powder. The method of ALT introduction is seen to influence SOFC performance, where measurements of a cell's output of current at a constant voltage indicate ALT addition via infiltration is a more effective means of improving cell performance than ALT addition via the mechanical mixing of ALT powder with other anode precursors before anode application. Finally, an optimum amount of ALT in SOFCs made with ~4∙mm NiO precursor particle size is proposed as 5.7 ± 0.7 wt.% ALT. Estimation of this optimum ALT amount was facilitated by a novel method of cell degradation-rate calculation. The algorithm used to determine degradation rates is available upon request.
[10]
[11]
[12]
[13]
[14]
Acknowledgements [15]
This work was supported by the DOE under grant #FE000261.
references
[1] Park SY, Ahn JH, Jeong CW, Na CW, Song RH, Lee JH. Ni-YSZsupported tubular solid oxide fuel cells with GDC interlayer
[16]
5
between YSZ electrolyte and LSCF cathode. Int J Hydrogen Energy 2014;39:12894e903. https://doi.org/10.1016/ j.ijhydene.2014.06.103. Powell M, Meinhardt K, Sprenkle V, Chick L, McVay G. Demonstration of a highly efficient solid oxide fuel cell power system using adiabatic steam reforming and anode gas recirculation. J Power Sources 2012;205:377e84. https:// doi.org/10.1016/j.jpowsour.2012.01.098. Choudhury A, Chandra H, Arora A. Application of solid oxide fuel cell technology for power generation e a review. Renew Sustain Energy Rev 2013;20:430e42. https://doi.org/10.1016/ j.rser.2012.11.031. da Silva FS, de Souza TM. Novel materials for solid oxide fuel cell technologies: a literature review. Int J Hydrogen Energy 2017;42:26020e36. https://doi.org/10.1016/ j.ijhydene.2017.08.105. Iwata T. Characterization of Ni-YSZ anode degradation for substrate-type solid oxide fuel cells. J Electrochem Soc 1996;143:225e625. https://doi.org/10.1149/1.1836673. Montenegro-Hernandez A, Vega-Castillo J, Mogni L, Caneiro A. Thermal stability of Ln2NiO4þ?? (Ln: La, Pr, Nd) and their chemical compatibility with YSZ and CGO solid electrolytes. Int J Hydrogen Energy 2011;36:15704e14. https:// doi.org/10.1016/j.ijhydene.2011.08.105. Ananyev MV, Farlenkov AS, Eremin VA, Kurumchin EK. Degradation kinetics of LSMeYSZ cathode materials for SOFC. Int J Hydrogen Energy 2018;43:951e9. https://doi.org/ 10.1016/j.ijhydene.2017.11.107. Kim SJ, Kim KJ, Dayaghi AM, Choi GM. Polarization and stability of La2NiO4þd in comparison with La0.6Sr0.4Co0.2Fe0.8O3d as air electrode of solid oxide electrolysis cell. Int J Hydrogen Energy 2016;41:14498e506. https://doi.org/10.1016/j.ijhydene.2016.05.284. Bo Y, Wenqiang Z, Jingming X, Jing C. Microstructural characterization and electrochemical properties of Ba0.5Sr0.5Co0.8Fe0.2O3-dand its application for anode of SOEC. Int J Hydrogen Energy 2008;33:6873e7. https://doi.org/ 10.1016/j.ijhydene.2008.07.066. Yokokawa H, Tu H, Iwanschitz B, Mai A. Fundamental mechanisms limiting solid oxide fuel cell durability. J Power Sources 2008;182:400e12. https://doi.org/10.1016/ j.jpowsour.2008.02.016. Shri Prakash B, Senthil Kumar S, Aruna ST. Properties and development of Ni/YSZ as an anode material in solid oxide fuel cell: a review. Renew Sustain Energy Rev 2014;36:149e79. https://doi.org/10.1016/j.rser.2014.04.043. Khan MS, Lee SB, Song RH, Lee JW, Lim TH, Park SJ. Fundamental mechanisms involved in the degradation of nickel-yttria stabilized zirconia (Ni-YSZ) anode during solid oxide fuel cells operation: a review. Ceram Int 2015;42:35e48. https://doi.org/10.1016/j.ceramint.2015.09.006. Shaikh SPS, Muchtar A, Somalu MR. A review on the selection of anode materials for solid-oxide fuel cells. Renew Sustain Energy Rev 2015;51:1e8. https://doi.org/10.1016/ j.rser.2015.05.069. Presto S, Kumar P, Varma S, Viviani M, Singh P. Electrical conductivity of NiMoebased double perovskites under SOFC anodic conditions. Int J Hydrogen Energy 2018;43:4528e33. https://doi.org/10.1016/j.ijhydene.2018.01.066. Weisentein AJ, Childs N, Amendola R, Driscoll D, Sofie SW, Gannon P, et al. Processing and characterization of Sr2xVMoO6-y double perovskites. Mater Chem Phys 2013;139:706e18. https://doi.org/10.1016/ j.matchemphys.2013.02.021. Childs NB, Wesinstein A, Smith R, Sofie S, Key C. Electrical conductivity of Sr2-xVMoO6-y (x ¼ 0.0, 0.1, 0.2) double perovskites. J Appl Phys 2013;113:0e9. https://doi.org/ 10.1063/1.4811715.
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e6
[17] Yu R, Guan W, Wang F, Han F. Quantitative assessment of anode contribution to cell degradation under various polarization conditions using industrial size planar solid oxide fuel cells. Int J Hydrogen Energy 2018;43:2429e35. https://doi.org/10.1016/j.ijhydene.2017.12.030. [18] Kennouche D, Chen-Wiegart YCK, Yakal-Kremski KJ, Wang J, Gibbs JW, Voorhees PW, et al. Observing the microstructural evolution of Ni-Yttria-stabilized zirconia solid oxide fuel cell anodes. Acta Mater 2016;103:204e10. https://doi.org/10.1016/ j.actamat.2015.09.055. [19] Lee YH, Muroyama H, Matsui T, Eguchi K. Degradation of nickel-yttria-stabilized zirconia anode in solid oxide fuel cells under changing temperature and humidity conditions. J Power Sources 2014;262:451e6. https://doi.org/10.1016/ j.jpowsour.2014.03.031. [20] Lee K, Kang J, Lee J, Lee S, Bae J. Evaluation of metalsupported solid oxide fuel cells (MS-SOFCs) fabricated at low temperature (~1,000 C) using wet chemical coating processes and a catalyst wet impregnation method. Int J Hydrogen Energy 2018;43:3786e96. https://doi.org/10.1016/ j.ijhydene.2018.01.027. [21] Ding C, Hashida T. Synthesis and evaluation of NiOeCe0.8Sm0.2O1.9 nanocomposite powders for lowtemperature solid oxide fuel cells. Int J Hydrogen Energy 2011;36:5567e73. https://doi.org/10.1016/ j.ijhydene.2011.02.073. [22] Kim S, Kim C, Lee JH, Shin J, Lim TH, Kim G. Tailoring Nibased catalyst by alloying with transition metals (M ¼ Ni, Co, Cu, and Fe) for direct hydrocarbon utilization of energy conversion devices. Electrochim Acta 2017;225:399e406. https://doi.org/10.1016/j.electacta.2016.12.178. [23] Driscoll DR, McIntyre MD, Welander MM, Sofie SW, Walker RA. Enhancement of high temperature metallic catalysts: aluminum titanate in the nickel-zirconia system. Appl Catal A Gen 2016;527:36e44. https://doi.org/10.1016/ j.apcata.2016.08.020. [24] Driscoll D, Hunt C, Muretta J, Sofie S. Thermally stabilized nickel electro-catalyst introduced by infiltration for high temperature electrochemical energy conversion. ECS Trans 2013;53:1e36. [25] German RM. Sintering theory and practice, vol. 1; 1996. [26] Zubair M, Taqi M, Song R, Lee J, Lee S, Lim T. A simplified approach to predict performance degradation of a solid oxide fuel cell anode. J Power Sources 2018;391:94e105. https:// doi.org/10.1016/j.jpowsour.2018.04.080. [27] Chaim R. Activation energy and grain growth in nanocrystalline Y-TZP ceramics. Mater Sci Eng A 2008;486:439e46. https://doi.org/10.1016/j.msea.2007.09.022.
[28] Hossain S, Abdalla AM, Radenahmad N, Zakaria AKM, Zaini JH, Rahman SMH, et al. Highly dense and chemically stable proton conducting electrolyte sintered at 1200 C. Int J Hydrogen Energy 2018;43:894e907. https://doi.org/10.1016/ j.ijhydene.2017.11.111. [29] Osinkin DA. Degradation of Ni-Zr0.9Sc0.1O1.95anode in H2þ H2O at low temperature: influence of nickel surface charge. Int J Hydrogen Energy 2018;43:943e50. https://doi.org/ 10.1016/j.ijhydene.2017.11.071. [30] Laurencin J, Hubert M, Sanchez DF, Pylypko S, Morales M, Morata A, et al. Degradation mechanism of La0.6Sr0.4Co0.2Fe0.8O3-d/Gd0.1Ce0.9O2-dcomposite electrode operated under solid oxide electrolysis and fuel cell conditions. Electrochim Acta 2017;241:459e76. https:// doi.org/10.1016/j.electacta.2017.05.011. [31] Iwai H, Shikazono N, Matsui T, Teshima H, Kishimoto M, Kishida R, et al. Quantification of SOFC anode microstructure based on dual beam FIB-SEM technique. J Power Sources 2010;195:955e61. https://doi.org/10.1016/ j.jpowsour.2009.09.005. [32] Wilson JR, Cronin JS, Barnett SA. Linking the microstructure, performance and durability of Ni-yttria-stabilized zirconia solid oxide fuel cell anodes using three-dimensional focused ion beam-scanning electron microscopy imaging. Scripta Mater 2011;65:67e72. https://doi.org/10.1016/ j.scriptamat.2010.09.025. [33] Wang F, Wang W, Ran R, Tade MO, Shao Z. Aluminum oxide as a dual-functional modifier of Ni-based anodes of solid oxide fuel cells for operation on simulated biogas. J Power Sources 2014;268:787e93. https://doi.org/10.1016/ j.jpowsour.2014.06.087. [34] Welander MM, Zachariasen MS, Hunt CD, Sofie SW, Walker RA. Operando studies of redox resilience in ALT enhanced NiO-YSZ SOFC anodes. J Electrochem Soc 2018;165:F152e7. https://doi.org/10.1149/2.0531803jes. [35] Driscoll DR, Tokmakian BW, Johnson NP, Hunt CD, Sofie SW. Catalyst enhancing aluminum titanate for increasing strength of nickel-zirconia cermets. Mater Lett 2017;209:307e10. https://doi.org/10.1016/j.matlet.2017.08.035. [36] Lussier A, Sofie S, Dvorak J, Idzerda YU. Mechanism for SOFC anode degradation from hydrogen sulfide exposure. Int J Hydrogen Energy 2008;33:3945e51. https://doi.org/10.1016/ j.ijhydene.2007.11.033. [37] Fukui T, Ohara S, Naito M, Nogi K. Performance and stability of {SOFC} anode fabricated from {NiO{ytextendash}YSZ} composite particles. J Power Sources 2002;110:91e5. https:// doi.org/10.1016/S0378-7753(02)00218-5.
Please cite this article in press as: Hunt C, et al., Degradation rate quantification of solid oxide fuel cell performance with and without Al2TiO5 addition, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.06.115