Effect of Perfluoroalkyl Chain Length on Proton Conduction in

Oct 28, 2010 - Leoncio Garrido , Javier Pozuelo , Mar López-González , Gengwei Yan , Jianhua Fang , and Evaristo Riande. The Journal of Physical ...
0 downloads 0 Views 242KB Size
Download PDF
14972

J. Phys. Chem. B 2010, 114, 14972–14976

Effect of Perfluoroalkyl Chain Length on Proton Conduction in Fluoroalkylated Phosphonic, Phosphinic, and Sulfonic Acids Mahesha B. Herath, Stephen E. Creager,* Alex Kitaygorodskiy, and Darryl D. DesMarteau Department of Chemistry, Clemson UniVersity, Clemson, South Carolina 29634-0973, United States ReceiVed: July 30, 2010; ReVised Manuscript ReceiVed: October 5, 2010

The effects of increasing perfluoroalkyl chain length on the molecular properties of viscosity, diffusivity, and ionic conductivity of a series of acid model compounds analogous to comb-branch perfluorinated ionomers functionalized with phosphonic, phosphinic, and sulfonic protogenic groups are reported. Anhydrous proton transport by a Grotthuss-like hopping mechanism was observed to occur efficiently in phosphorus-based fluoroalkylated model acids but only when there is a relatively low perfluoroalkyl content. The decrease in degree of dissociation of the protogenic groups follows the order phosphonic > phosphinic > sulfonic, and the degree of dissociation and the magnitude of ion-ion correlations are approximately independent of chain length. SCHEME 1: Chemical Structures of Model Acidsa

1. Introduction The most commonly used proton exchange membrane (PEM) materials for fuel cells are based on sulfonic acid functionalized polymers such as Nafion, Flemion, and/or Aciplex.1 These polymers do not function well at temperatures above 80 °C and at low humidity. Therefore, substituting the sulfonic acid group with a phosphonic acid group as an alternative protogenic moiety to construct these polymers has been intensely discussed in the literature.2-7 We recently reported8 in a model study that involved (1) CF3SO3H, (2) CF3PO(OH)2, (3) (CF3)2PO(OH), and (4) CF3COOH that the phosphorus-based acid compounds show high protonic conductivity under anhydrous conditions over a wide range of temperature (ambient to 120 °C). Further, we showed that both of the phosphorus-based model compounds that were tested displayed a high degree of self-dissociation. Proton conduction in these two model compounds was shown to occur primarily by a Grotthuss-like proton hopping mechanism. In contrast, under anhydrous conditions the sulfonic and carboxylic acid model compounds had a relatively low ionic conductivity and showed relatively little dissociation. For PEM polymers containing side chains terminating with sulfonic acid, it has been reported9,10 that depending on the length of the side chain these polymers exhibit different macroscopic properties in terms of proton conductivity and membrane morphology. To study the combined effects of the length of the perfluoroalkyl chain and the type of protogenic group on acid properties including conductivity, we undertook a study of a new series of model acids in which the trifluoromethyl group of three of the short-chain model acids (CF3SO3H, CF3PO(OH)2, nd (CF3)2PO(OH)) was replaced with a longer perfluorobutyl chain. The chemical structures of the model compounds with this longer perfluoroalkyl chain are shown in Scheme 1. They are as follows: perfluorobutyl sulfonic acid (1), perfluorobutyl phosphonic acid (2), and bisperfluorobutyl phosphinic acid (3). These compounds were investigated for molecular properties such as diffusivity and viscosity that influence proton conductivity as described previously for the short-chain * Corresponding author. Tel.: 864-656-4995. Fax: 864-656-6613. E-mail: [email protected].

a 1, perfluorobutyl sulfonic acid; 2, perfluorobutyl phosphonic acid; 3, bisperfluorobutyl phosphinic acid.

analogues.8 In the present paper we discuss the structure-property relationships for the shorter- vs longer-chain model compounds. This work can provide a better understanding of anhydrous proton transport and help in development of polymers with optimal properties including conductivities for fuel-cell membrane applications. 2. Experimental Section 2.1. Materials and Reagents. Unless otherwise noted, starting materials were obtained from commercial suppliers and used without further purification. Acetonitrile was dried by storing over KOH overnight and distilling from P2O5. Perfluorobutyl sulfonic acid (1) (98+ %) was obtained from Aldrich and used without further purification. 1H, 19F, and 31P NMR spectra were recorded on a JEOL NMR instrument at 300.3, 282.2, and 121.5 MHz, respectively, using CD3CN as the solvent except where noted. 2.2. Analytical Procedures. Proton conductivity, viscosity, density, pulsed-field-gradient nuclear magnetic resonance (PFGNMR) studies at 85 °C, and Hammett acidity measurements were determined as previously described by us in the experimental section of ref 8. 2.3. Synthesis. n-Perfluorobutyl Phosphonic Acid (2). Synthesis was accomplished as illustrated in Scheme 2. Under an argon atmosphere, isopropyl magnesium chloride (2 M in diethyl

10.1021/jp107190q  2010 American Chemical Society Published on Web 10/28/2010

Perfluoroalkyl Chain Length and Proton Conduction SCHEME 2: Synthesis of n-Perfluorobutyl Phosphonic Acid

J. Phys. Chem. B, Vol. 114, No. 46, 2010 14973 TABLE 1: Density G, Molecular Weight MW, Molar Concentration C0, Viscosity η, and Specific Conductivity σ of the Model Acid Compounds F

MW -3

CF3PO(OH)2a (CF3)2PO(OH)a CF3SO3Ha C4F9SO3H C4F9PO(OH)2 (C4F9)2PO(OH)

SCHEME 3: Synthesis of Bisperfluorobutyl Phosphinic Acid

ether, 5.80 mL, 11.56 mmol) was added dropwise to a perfluorobutyl iodide solution (4.0 g, 11.56 mmol) in dry diethyl ether (20 mL) at -78 °C. The solution was stirred at this temperature for 1 h to allow exchange to take place. Diethyl chlorophosphate (1.80 g, 11.56 mmol) was then added slowly by syringe at -78 °C. The mixture was maintained at -78 °C for 1 h and then allowed to warm to room temperature overnight. The reaction was quenched with 5 mL of 3 M ice-cold HCl. The ether layer was separated and dried with Na2SO4. The ether was filtered and concentrated on a rotary evaporator. The residue was mixed with excess bromotrimethylsilane (3.72 g, 24.2 mmol) and stirred at room temperature for 12 h. When the reaction was completed the flask was cooled to room temperature, and ethyl bromide and excess silylating reagent were removed by rotary evaporation at reduced pressure. The silylated ester was hydrolyzed with water (10 mL). After hydrolysis, 0.1 g of charcoal and water (2 mL) were added, and the mixture was stirred for 2 h at room temperature. The charcoal was filtered, and filtrate was extracted with ether (3 × 10 mL) and dried over Na2SO4. The ether was filtered and rotary evaporated to yield colorless oil. The oil was dried at 60 °C under dynamic vacuum for 48 h to yield (2.50 g, 8.33 mmol, 72% based on perfluorobutyl iodide) perfluorobutyl phosphonic acid. 1H NMR (300 MHz, CD3CN): 10.8 (br, 2H, PO(OH)2). 19F NMR (282.7 MHz, CD3CN): - 80.7 (s, 3F, CF3, JFCP ) 107.4 Hz), -121.4 (s, 2F, CF2), -122.9 (d, -CF2P, JPCF ) 81.8 Hz), -125.7 (s, 2F, -CF2-CF2P). 31P NMR (121.5 MHz, CD3CN): -1.92 (t, JPCF ) 81.8 Hz). Bis(n-perfluorobutyl)phosphinic Acid (3). Synthesis was accomplished as illustrated in Scheme 3. Under an argon atmosphere, isopropyl magnesium chloride (2 M in diethyl ether, 7.2 mL, 14.4 mmol) was added dropwise to a perfluorobutyl iodide solution (4.9 g, 14.4 mmol) in dry diethyl ether (20 mL) at -78 °C. The solution was stirred at this temperature for 1 h to allow exchange to take place. Dichloroethylphosphate (0.94 g, 5.78 mmol) was then added slowly by syringe at -78 °C. The mixture was maintained at -78 °C for 1 h and then allowed to warm to room temperature overnight. The reaction was quenched by adding 5 mL of 3 M ice-cold HCl, and the mixture was stirred at room temperature for 1 h. After hydrolysis was complete, 0.1 g of charcoal (2 mL) was added, and the mixture

a

η at 25 °C σ at 25 °C

C0

(g cm ) (g equiv ) (mol cm )

(cP)

(S cm-1)

1.23 × 10-2 8.66 × 10-3 1.13 × 10-2 6.0 × 10-3 6.2 × 10-3 3.5 × 10-3

-69 4.0 21.8 84.0 198.2

--1.60 × 10-2 6.60 × 10-2 9.20 × 10-4 7.90 × 10-4 5.86 × 10-5

1.85 1.75 1.70 1.80 1.85 1.76

-1

150 202 150 300 300 502

-3

The data for F, MW, C0, η, and σ are taken from ref 8.

was further stirred for 2 h at room temperature. The charcoal was filtered, and filtrate was extracted with ether (3 × 10 mL) and dried over Na2SO4. The ether was filtered and rotary evaporated to yield colorless oil. The oil was dried at 60 °C under dynamic vacuum for 48 h to yield (2.17 g, 4.32 mmol, 60% based on perfluorobutyl iodide) bis(n-perfluorobutyl) phosphinic acid. 1H NMR (300 MHz, CD3CN): 10.8 (br, 1H, PO(OH)). 19F NMR (282.7 MHz, CD3CN): -81.6 (s, 3F, CF3), -121.2 (s, 2F, CF2), -122.0 (d, -CF2P, JFCP ) 81.8 Hz), -126.3 (s, 2F, -CF2-CF2P). 31P NMR (121.5 MHz, CD3CN): -0.07 (t, JPCF ) 88.8 Hz). 3. Results and Discussion Table 1 presents a selected set of physical properties of the three model compounds at 25 °C. C0 represents the molar concentration of molecular acids calculated using the measured density and calculated molar mass for the acids. 3.1. Ionic Conductivity and Fluidity. The conductivity and fluidity data shown in Figure 1 (top and bottom, respectively) include data for the model compounds listed in Table 1 and also for those studied in our prior work (ref 8). The conductivities and fluidities of all the model compounds, irrespective of their perfluoroalkyl chain length, follow Arrhenius behavior at all investigated temperatures. The best-fit parameters for fits to the data for the model compounds may be found in the Supporting Information. Increasing the perfluoroalkyl chain length of the short-chain model compounds always leads to a decrease in conductivity. For example, the conductivity of longer-chain C4F9P(O)(OH)2 has decreased by approximately an order of magnitude compared to its shorter-chain analogue CF3PO(OH)2, and the conductivity of (C4F9)2PO(OH) has decreased by approximately 2 orders of magnitude compared to (CF3)2PO(OH). This decrease could be caused by several factors, including decreased proton concentration, decreased acid dissociation, and decreased proton (ionic) mobility which in turn could be caused by a decreased fluidity or different degrees of proton hopping. Figure 2 shows how proton conductivity and fluidity each vary with fluorine to proton ratio (F/H ratio) for the model compounds. All the phosphorus-based acids except for (CF3)2PO(OH) show a monotonically decreasing proton conductivity with increasing fluorine to proton ratio. Figure 2 (bottom) shows that the fluidity of phosphorus-based compounds is fairly independent of the length of the perfluoroalkyl chain. Thus, it appears that decreased conductivity with increasing fluoroalkyl content is not directly correlated with fluidity, as it would be for ionic transport by a simple vehicle mechanism. In contrast, for the sulfonic acids, Figure 2 shows that both proton conductivity and fluidity are decreasing with the increasing perfluoroalkyl chain length. In ref 8, we showed that the proton conductivity of anhydrous trifluoromethyl sulfonic acid

14974

J. Phys. Chem. B, Vol. 114, No. 46, 2010

Figure 1. (top) Anhydrous conductivity data as a function of temperature. (bottom) Fluidity vs temperature studies for the model compounds 1-3 compared with their short-chain analogue.

is strongly correlated with fluidity suggesting that the charge carriers move in this acid by a vehicle mechanism. In such a case, increasing perfluoroalkyl chain length lowers fluidity, and therefore proton conductivity is expected to decrease by the same factor that was observed for fluidity if the percent dissociation of acid remains the same for both these sulfonic acid model compounds. Comparison of the conductivity and fluidity data for the sulfonic acids at 85 °C shows that the fluidity of C4F9SO3H has decreased by about 2.2 times and the conductivity has decreased by about 4.2 times compared to CF3SO3H. The greater decrease in conductivity for C4F9SO3H than expected from the change in fluidity could be due to a lower carrier charge concentration, perhaps due to a lower degree of dissociation. 3.2. Walden Plots. The relationship between conductivity and viscosity may be further analyzed through the use of a Walden plot which is a log-log plot of specific conductivity vs fluidity. For a data set involving just one electrolyte, the plot usually includes data acquired at different temperatures. As discussed in ref 8, the Walden plot of a system that obeys Walden’s rule (Λη ) constant), will have a slope of unity, and its Walden plot line will pass through a point for an ideal electrolyte, often approximated by a dilute KCl solution in water. Figure 3 presents Walden plot lines for all the short- and longer-chain model compounds of sulfonic- and phosphorusbased acids discussed in the previous section. Considering just the sulfonic acids, we note that the Walden lines for both acids are located far below the ideal line, are on almost the same line, and have slopes very near to unity. The fact that the lines

Herath et al.

Figure 2. (top) Conductivity data as a function of F/H ratio. (bottom) Fluidity vs F/H ratio; for all the short- and longer-chain model compounds.

Figure 3. Walden plot for the model acids.

are far below the ideal line suggests that both of these acids undergo only a small degree of dissociation, and the fact that they are on almost the same line suggests that they both have approximately the same small degree of dissociation. (Other factors can also affect the Walden line position, as discussed in more depth below in the section describing Nernst-Einstein calculations; however, when the deviation is as large as it is here, the reason is almost always a low degree of dissociation.) The fact that their slopes are both near unity suggests that the activation energies for viscous transport and ionic transport are similar to each other. Taken together these points lead us to

Perfluoroalkyl Chain Length and Proton Conduction

J. Phys. Chem. B, Vol. 114, No. 46, 2010 14975

TABLE 2: PFG-NMR Diffusion Coefficients for the Model Acids at 85 °C at 85 °C DH (m2/s) compound CF3SO3H CF3PO(OH)2 (CF3)2PO(OH) C4F9SO3H C4F9PO(OH)2 (C4F9)2PO(OH)

× 10

10

10.80 1.12 9.47 4.87 0.26 0.14

DF (m2/s) × 1010

DH/DF

10.70 0.50 5.99 4.87 0.14 0.12

1.00 2.25 1.58 1.00 1.85 1.17

conclude that under anhydrous conditions both of these fluoroalkyl sulfonic acids transport protons by a vehicle mechanism with a relatively low conductivity that is a consequence of a relatively low degree of dissociation. The Walden plots for C4F9PO(OH)2 and (C4F9)2PO(OH) are located above the lines for the sulfonic acids but still well below both the ideal-electrolyte line and the lines for the short-chain analogues, CF3PO(OH)2 and (CF3)2PO(OH), which suggests that the degree of dissociation of these acids is probably quite low (though perhaps not as low as for the sulfonic acids). The Walden line slopes for the two long-chain phosphorus acids are 0.77 and 0.92, respectively. The slope for C4F9PO(OH)2 is approximately similar to what was observed previously8 for CF3PO(OH)2 (0.74) and for H3PO411(0.67). In the case of these two acids, deviation of slope from 1 was taken to mean that hopping/structure diffusion, which has an activation energy lower than that for proton transport by a vehicle mechanism, was making a significant contribution to proton transport. The observation that C4F9PO(OH)2 also exhibits a low Walden plot slope suggests that hopping transport is also important for this acid. It is quite interesting that hopping transport can occur in an acid that also has a relatively low degree of dissociation. One possible explanation of this is that the acid undergoes a substantial degree of aggregation, with fluorinated groups and protogenic groups self-associating in a manner that allows for hopping transport even the liquid phase and even when most acids are not dissociated. In contrast, the Walden slope for (C4F9)2PO(OH) (0.92) is significantly higher than what was observed for its short-chain analogue (CF3)2PO(OH) (0.67). A Walden slope value closer to unity for (C4F9)2PO(OH) suggests that proton transport in this acid occurs primarily by a vehicle transport mechanism. Perhaps the very high fraction of fluorocarbon content in this acid prevents the kind of protogenic group self-association that can occur in C4F9PO(OH)2 and (apparently) also in (CF3)2PO(OH). 3.3. Pulsed-Field-Gradient Nuclear Magnetic Resonance Spectroscopy Studies. 1H and 19F pulsed-field-gradient NMR data were used to determine diffusion coefficients for the proton (DH) and its conjugate base (DF) in the model acids. PFG-NMR experiments were performed at 85 °C. The results for the model acids studied in this work and also for the short-chain analogues from ref 8 are given in Table 2. The decreasing order of the DF values follows the order CF3SO3H > (CF3)2PO(OH) > C4F9SO3H > CF3PO(OH)2 > C4F9PO(OH)2 > (C4F9)2PO(OH)2. These decreasing values of DF follow almost the same order as the decreasing fluidities of the model acids. The decreasing fluidities of the model compounds are CF3SO3H > C4F9SO3H > (CF3)2PO(OH) > CF3PO(OH)2 > C4F9PO(OH)2 > (C4F9)2PO(OH). The agreement between these properties is expected since unlike the protons the anions are transported by the vehicle mechanism. The difference in interchanging order of the (CF3)2PO(OH) and

TABLE 3: Nernst-Einstein Calculated Conductivity σnmr (S cm-1), Experimental Conductivity Obtained from an Arrehnius Plot σimp (S cm-1), and Haven Ratio σnmr/σimp for the Model Acid Compounds at 85 °C compound CF3SO3H CF3PO(OH)2 (CF3)2PO(OH) C4F9SO3H C4F9PO(OH)2 (C4F9)2PO(OH)

σNMR

HR ) σNMR/σimp

σimp -1

7.62 × 10 6.25 × 10-2 4.13 × 10-1 1.83 × 10-1 7.72 × 10-3 2.85 × 10-3

-2

1.47 × 10 4.33 × 10-2 6.53 × 10-2 3.46 × 10-3 4.90 × 10-3 6.14 × 10-4

51.8 1.4 6.3 52.8 1.6 4.6

C4F9SO3H in DF values could be due to their differences in the size of the anions. At 85 °C the fluidity of (CF3)2PO(OH) is 11.7 (Poise)-1, while the fluidity of C4F9SO3H is 24.3 (Poise)-1. However, the C4F9 group is larger than the (CF3)2 by 2 carbons and 3 fluorine groups. This may have caused a slight decrease in DF value of C4F9SO3H compared to (CF3)2PO(OH). 3.4. Nernst-Einstein Relation and Haven Ratio. When both ion diffusivity and conductivity data are available for a particular electrolyte, it can be instructive to use the Nernst-Einstein relation and the Haven ratio (HR) to analyze the data. The Haven ratio is the ratio of the ionic conductivity estimated from PFGNMR diffusivity data (σNMR) using the Nernst-Einstein equation (given by eq 1 below),12 to that measured experimentally (σimp), in this case by an ac impedance method. Thus, HR ) σNMR/σimp.

σNMR )

z2F2 C (H)[DH + DF] RT 0

(1)

In this equation, σNMR is the total ionic conductivity expected from self-diffusion coefficients, and C0 denotes the molar concentration of the model acid (C0 ) F/MW in mol cm-3 for which F is the density and MW is the molecular weight of the model acid), assuming that only the complete first ionization of the model acid contributes as a source of [H+] ions. DF and DH are the diffusion coefficients of the conjugate base and proton in the model acids; z is the valence of the ions which in this case is 1; and F, R, and T have their usual meaning. Table 3 presents values for σNMR calculated using eq 1, σimp measured by impedance spectroscopy, and Haven ratios calculated using σNMR and σimp. Table 3 shows two clear trends. First, the Haven ratio correlates well with the type of protogenic group, with values being quite large (approximately 50) for fluoroalkylsulfonic acids, intermediate (4 to 6) for the bis(fluoroalkyl)phosphinic acids, and low (just above 1) for the fluoroalkylphosphonic acids. For these strong acid electrolytes, a Haven ratio close to unity indicates that the acids are fully dissociated (C0 is an accurate representation of the ionic concentration), and all the dissociated ions are free from ion-ion interactions and contribute completely to the ionic conductivity. A Haven ratio greater than 1 indicates deviation of the experimental conductivity from the calculated conductivity due to ion-ion interactions or due to low dissociation. In this regard, the results are consistent with conclusions derived from Walden plots, namely, that the extent of acid dissociation varies in the order phosphonic > phosphinic > sulfonic. The second trend is that for all three acid types the Haven ratio is approximately independent of fluoroalkyl chain length. This is in some ways a surprising result since changing the fluoroalkyl chain length can have large effects on acid properties, particularly on polarity/dielectric constant, viscosity,

14976

J. Phys. Chem. B, Vol. 114, No. 46, 2010

and conductivity. These results suggest that, despite these changes, the ratio between the expected and observed conductivity is determined principally by the nature of the protogenic group itself, and less so by the substituents on that group. In this paper, we have investigated the transport properties of model acids under anhydrous conditions. However, during the operation of a fuel cell under low humidity conditions some water will always be generated. In preliminary work we have noted that adding relatively small amounts of water can dramatically increase ionic conductivities of these acids. For example, we find that addition of 3 mol of water to 1 mol of anhydrous acid increases the proton conductivity of the phosphonic and phosphinic acids by 1 order of magnitude from the anhydrous conductivities. We plan to describe these findings in more detail in a follow-up paper to be published later. 4. Conclusions Ionic conductivity in perfluorinated phosphorus-based acids depends on the fluorine content of acids. Lower fluorine content is necessary to form a high concentration of hydrogen bonds and for efficient transfer of proton by the Grotthuss mechanism. Perfluoroalkyl phosphonic acids have a higher percent dissociation and give a higher conductivity upon increasing perfluoroalkyl chain length compared to phosphinic or sulfonic acids under anhydrous conditions. The conductivity of perfluorinated sulfonic acids depends on carrier mobility, and under anhydrous conditions they have relatively low carrier concentration due to the low degree of acid dissociation.

Herath et al. Acknowledgment. This work was funded by the United Sates Department of Energy (US DOE) hydrogen program grant number DE-FG36-06GO16031. Supporting Information Available: Hammett acidities and results of the Arrhenius analysis of the best-fit parameters for the longer-chain model acids. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Li, Q.; He, R.; Jensen, J. O.; Bjerrum, N. J. Chem. Mater. 2003, 15, 4896. (2) Kanamura, K.; Tanaka, A.; Gervasio, D.; Kennedy, V.; Adzic, R.; Yeager, E. B. J. Electrochem. Soc. 1996, 143, 2765. (3) Kreuer, K. D. Chem. Mater. 1996, 8, 610. (4) Souzy, R.; Ameduri, B. Prog. Polym. Sci. 2005, 30, 644. (5) Souzy, R.; Ameduri, B.; Boutevin, B.; Gebel, G.; Capron, P. Solid State Ionics 2005, 176, 2839. (6) Rager, T.; Schuster, M.; Steininger, H.; Kreuer, K. D. AdV. Mater. 2007, 19, 3317. (7) Steininger, H.; Schuster, M.; Kreuer, K. D.; Kaltbeitzel, A.; Bingol, B.; Meyer, W. H.; Schauff, S.; Brunklaus, G.; Maier, J.; Spiess, H. W. Phys. Chem. Chem. Phys. 2007, 9, 1764. (8) Herath, M. B.; Creager, S. E.; Kitaygorodskiy, A.; Desmarteau, D. D. ChemPhysChem 2010, 11, 2871 (invited article). (9) Arico, A. S.; Baglio, V.; Di Blasi, A.; Antonucci, V.; Cirillo, L.; Ghielmi, A.; Arcella, V. Desalination 2006, 199, 271. (10) Miyatake, K.; Yasuda, T.; Hirai, M.; Nanasawa, M.; Watanabe, M. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 157. (11) Chin, D.; Chang, H. J. Appl. Electrochem. 1989, 19, 95. (12) Jost, W. J. Diffusion in solids, Liquids, Gases; Academic Press: NY, 1960.

JP107190Q