Data Loading...
sAn -Methyl-3-Ethylimidazolium Bromide Aluminum Bromide ... Flipbook PDF
466 J. Electrochem. Soc., Vol. 138, No. 2, February 1991 * The Electrochemical Society, Inc. was continually monitored b
106 Views
16 Downloads
FLIP PDF 443.63KB
AD-A234 345 Reprinted from JOURNAL OF THE ELECrlTx:HEMIAI. S((IETY
Vol. 138, No. 2, February 1991 Printed in U.S.A. Copyright 1991
-;
J 1.
Not
-- "
vI
II _'..._
Electrochemical Investigation of the
sAn
. .. .
-
1-Methyl-3-Ethylimidazolium Bromide Aluminum Bromide Molten Salt System Jeffrey A. Boon and John S. Wilkes* The FrankJ. Seiler Research Laboratory,USAF Academy, Colorado 80840-6528
John A. Lonning University of Colorado at Denver, Denver, Colorado 80204
ABSTRACT The electrochemical behavior of the room temperature molten salt made by combining 1-methyl-3-ethylimidazolium bromide and aluminum bromide was studied. It was found that the bromide system behaved, for the most part, analogously to the chloride counterpart. Bromide solvated in a room temperature molten salt is electrochemically similar to bromide in other nonaqueous solvent systems. The diffusion coefficients and the heterogeneous rate constants were determined for the bromide and tribromide species in the molten salt.
Although first used more than 40 years ago as an electrolytic bath for plating aluminum (1), room temperature molten salts have received renewed interest in recent years as electrolytes for battery applications. A molten salt electrolyte will theoretically increase the power density of a battery by not only decreasing the weight involved in the electrolyte system but also through the use of new couples which become feasible in a molten salt system. We have described previously a battery cell that employs a rechargeable bromine-bromide electrochemical couple in the 1-methyl-3-ethylimidazolium chloride-aluminum chloride low-temperature molten salt (2). Preliminary experiments show that the electrochemistry of the brominebromide couple is complex, possibly due to halide exchange in the chloroaluminate melt. We report here a detailed electrochemical study of bromine and bromide in the chemically similar 1-methyl-3-ethylimidazolium bromide-aluminum bromide system. Although Popov and Geske (3), among others, have characterized the nonaqueous electrochemical behavior of the different bromide-containing species, the chemical and electrochemical properties of the room temperature molten salts are different enough to warrant the investigation of bromide when it is solvated in the melt system. Experimental 1-methyl-3-ethylimidazolium chloride (MEIC).-The MEIC was prepared by a method similar to that presented by Wilkes et al, (4). The procedure was changed to eliminate the high pressures involved in the reaction vessel by Electrochemical
e
m
LE C4py crife
running the reaction at a lower temperature (ca. 30'C) for a longer time (ca. 2 weeks). The purity of the MEIC was confirmed by noting the lack of any unreacted 1-methylimidazole in the MEIC by 'H NMR, and by noting the absence of the electrochemical reduction of any protonic impurity in a melt prepared from the MEIC. 1-methyl-3-ethylimidazolium bromide (MEIB).-The MEIB was prepared by a method similar to that described by Sanders et al. (5, 6). The only modifications was that the reaction was carried out in a round bottom flask with a reflux condenser attached. The reaction mixture was heated to keep the ethyl bromide refluxing. The purity of the MEIB was checked in a manner similar to that of the MEIC. Aluminum bromide.-The aluminum bromide (Aldrich, 98+%) was purified by vacuum sublimation. The aluminum bromide was repeatedly sublimed until no color was observed in the collected material (a total of 3-5 sublimations). Aluminum chloriu.-The aluminum chloride was purfled by the method described by Wilkes et at. (7). Tetramethylammonium tribomide.-The tetramethylammonium tribromide was prepared by the method described bv Chattaway and Hoyle (8) and characterized by UV-VIS spectroscopy as described by Buckles et al. (9). The electrochemistry was carried out in a Vacuum Atmospheres Company dry box with a combined H.O/O, concentration of less than 10 ppm. The atmosphere in the dry box was UHP helium purified by a Vacium Atmospheres MO-40 Dri-Train. The quality of the atmosphere
4
12
466
J. Electrochem. Soc., Vol. 138, No. 2, February 1991
* The Electrochemical Society, Inc.
was continually monitored by a light bulb whose lilament was exposed to the dry box atmosphere. All electrochemical experiments were performed with a Princeton Applied Research (PAR) Model 173 potentiostat'galvanostat equipped with a Model 179 Digital Coulometer plug-in module. Any required potential or current programming was provided by a PAR Model 175 Universal Programmer. The potentials were monitored with a DANA Model 5330 digital voltmeter. The electrode used in these experiments was a platinum disk, 6 mm diam, imbedded in a Teflon sleeve, purchased from Pine Instruments Coinpany. Rotation of the working electrodes was accomplished with a Pine Instruments Company MSR electrode rotator. The reference electrode currently used in our lab is mac"- observing the Al :,'Al couple formed by placing an aluminum wire in an N = 0.600 chloride melt, with an asbestos fiber separating the reference electrode from the working compartment of the cell. To investigate the possible interferences due to the chloride migration across the reference electrode junction into the bromide melt, a cell was constructed with a chloride reference electrode and a neutral bromide melt. A cyclic voltammogram was obtained on the cell immediately after construction and. again, after six days of continuous contact between the chloride electrode and the bromide melt. No effect of the migration of the chloride ion across the junction was observed at the current limits with which we were interested. Because of the negligible effect of the chloride ions crossing the reference electrode junction and the longterm instability of the bromide melts, we used the chloride reference electrode throughout this work. Scans taken at less than a volt per second were recorded
Raman spectroscopic studies (11) indicate that AlBr- and AIBr4 exist in bromoaluminate melts in an analogous fashion to AI.,CI- and AICI4 in chloroaluminate melts. The dissolution reactions for the MEIB-AlBr4 melts are thus
on a Houston Instruments Model 2000 Omnigraphic X-Y
cesses on the other electrodes, platinum was chosen as the
recorder. S,-ans taken at greater than 1 V/s were recorded
working electrode material for the rest of this work.
on a Nicolet Model 204A storage oscilloscope and then plotted on the Omnigraphic recorder. Cell heating was performed by a resistive block heater. The temperatures were monitored by a Doric Model 412A Trendicator, using a type K thermocouple. Sanders and Hussey (5,6, 10) determined the physical properties of the AlBr:,-MEIB melts used throughout this woik. The physical properties are similar to those of the chloride system: but the melting points of the bromide
Some workers (5. 10) have worked with the bromide melts assuming the melts behave the same as the chloride counterparts. This work initialy made, but then verified this assumption. To gain an initial understanding of the electrochemistry involved in the bromide system, cyclic voltammograms of an acidic, basic, and neutral melt were obtained. Neutral melt.-The cyclic voltammogram of a neutral bromide and a neutral chloride melt are shown in Fig. 3. If
system are approximately 50' higher than the chloride,
Because of the presence of bromide, which acts as a Lewis base, any melt witn N less than 0.5 is referred to as a basic melt. Heptabromoaluminate (AI.Br I)acts as a Lewis acid so any melt with an N greater than 0.5 is termed acidic. Because of the lack of the Lewis base (bromide) and the Lewis acid (heptabromoaluminate), a melt with N equal to 0.5 is called a neutral melt. Results and Discussion 11,;.-as found that, unlike the chloride system, the formation reactions (Eq. [! 1 and 121) of the bromoaluminate melts are appreciably slow. It was also found that bromide melts tend to turn the distinctive orange-brown color associated with bromine. This coloration is probably due to the oxidation of the bromide. In order to ensure the completion of the dissolution reactions, once a melt was made, it was allowed to stand at 60'C for at least 12h prior to use. The melt was then used for no longer than three days before being destroyed. in order to minimize any effects of changes in the melt composition. Figure 2 shows the cyclic voltammogram of an N = 0.490 bromide melt at glassy carbon, tungsten, and platinum working electrodes. Because of the apparent high degree of electrochemical irreversibility of the oxidative pro-
in the neutral bromoaluminate molten salt are the tetrabromoaluminate anion and the imidazolium cation. The cathodic limit of the melt is the irreversible reduction of the imidazolium cation and is approximately the same for
*-20
•
13r--
0
1l1 121
the initial assumption is correct, the only species present
The same definitions which are used for the chloroaluminate molten salt systems were used in this work. Since there is no unreacted aluminum bromide present in the melt, upon completion of the dissolution reactions the term N is used to indicate the apparent mole fraction of aluminum bromide present in the binary melt. As with the chloride system, the experimental working range of N was limited to the region between N = 0.333 and N = 0.667 because of the higher temperatures required to make a homogeneous melt outside of this range. The stoichiometry of the molten salt determines its anionic composition, which is shown as a function of the apparent mole fraction of aluminum bromide in Fig. 1.
Ce;-
MEIBr + AlBr, - MEl AIBr4 MEI *AIBr4 + AIBr:, MEI AI,Br-
uA
13
sal. sn) 0C 5..54
*
0
a
7
A
0+a
.
.
..
_
Fig. 1. Anionic composition diogram for the bromoaluminote molten salt system.
Fig. 2. Cyclic voltammogrom of an N :-0.490 bromide melt ot a (a) glossy carbon, (b) tungsten, ond (c) platinum electrode.
J. Electrochem. Soc., Vol. 138, No. 2, February 1991 0 The Electrochemical Society, Inc.
467
a a f
e--
200
A
50 4uA
I
--
f
b
b
I
I
2.0
1.0
I
0.0 -1.0 Potential/V
I
-2.0
2.0
I
1.0
I
I
0.0 -1.0 Potential/V
-2.0
Fig. 3. Cyclic voltammogrom of a neutral (a) bromide and (b) chloride melt.
Fig. 5. Cyclic voltammogram of a slightly basic (N = 0.490) (a) bromide and (b) chloride melt.
the chloride and bromide melts. The anodic limit of the melt is the oxidation of the tetrabromoaluminate by the half reaction
the potential limits of the melt remain the same as those of the neutral melt, but now there is a reduction due to the heptabromoaluminate anion. The magnitude of the reductive process increases from 0 in the neutral melt, as the apparent mole fraction of aluminum bromide increases. The sharp break at the start of the reduction peak is commonly seen in nucleation2 phenomena preceding a deposition process. The non tl decay seen on the related oxidation wave is indicative of an insoluble species being stripped off of the electrode surface. The half reaction for this reductive process is 4AI 2Br 7 + 3e -- A10 + 7AlBr 4 [4]
[3] 4A1Br4 --, 2ABr7 - Br., + 2e beobserved couples redox no are there though Even tween the melt limits, when the potential is swept slowly, if the potential is swept into the anodic region quickly, a deposition and stripping wave is seen which is similar to that seen in an acidic melt. These observations argue for the validity of the initial assumption that the neutral bromide and chloride melts are analogous to one another. Acidic melts.-Upon the addition of a small amount of aluminum bromide to the neutral melt, the melt becomes slightly acidic (resulting in a small concentration of the heptabromoaluminate anion). The cyclic voltammograms of a slightly acidic (N = 0.510) bromide and chloride melt are shown in Fig. 4. The half reactions which determine
a
T 200 Atermined Celectrode (9
The voltainmogram in Fig. 4 demonstrates that, in general, the analogy between the bromoaluminate and chloroaluminate melts holds up well with respect to the behavior of acidic species in the bromoaluminate melts. Basic melts.-Upon the addition of a small amount of MEIB to a neutral melt, the melt becomes slightly basic and a small concentration of bromide results. The cyclic voltammograms of an N = 0.490 bromide and chloride melt are shown in Fig. 5. Unlike the chloride system, the bromide system shows two electrochemical processes for the oxidation of the bromide. A major focus of this work was to assign and characterize these two oxidative processes. As is common, the formal electrode potentials of the two oxidative processes in the basic bromide system were defrom the cyclic voltammograms by the method described by Kissinger and Heineman (12). The formal potential of the first oxidative process was found to be independent of the rate at which the electrode potential was scanned and is 0.684V vs. the Al"-'/Al reference electrode described earlier. The second oxidative process Table I. Transfer coefficients and number of electrons transferred Wave I Technique
_______
2.0
_
ILSV
0.0 -1.0 1.0 Potential/V
-2.0
Fig. 4. Cyclic voltammogrom of a slightly acidic (N = 0.510) (a) bromide and (b) chloride melt.
RDLSV Amperometry Coulometry
Wave 2 n
7n
0.41 0.55
1 II
0.54
1 a
1.0
'Could not be determined due to the proximity of the two waves. " Overall number of electrons transferred for the combined first and second oxidative processes.
468
J. Electrochem. Soc., Vol. 138, No. 2, February 1991 (c)The Electrochemical Society, Inc. Table II. Diffusion coefficients and Stokes-Einstein products of different species solvated in molten salts and acetonitrile Acetonitrile
Melt Species
Diem's
Br Br CI
7.4 - 10 9.4 x 10 9.6 x 10?
Dri -gcm T
s
2.5 10 3.2 x10 4.9 x 10''
was found to be dependent on the scan rate. The formal electrode potential as a function of the square root of the scan rate was extrapolated to a scan rate of zero. The resuiting formal electrode potential is 1.01V vs. the Al 3/A 5 reference. The number of electrons transferred. n,and the transfer oefficients. a were determined for each of the two o idn
tire proc:.... by variety of electrochemical techniques. The product an was determined using rotating disk linear sweep voltammetry and stationary electrode linear sweep voltammetry based on the slope of the leading edge of the voltammetric wave. The an product was also calculated using constant current amperometry and the linear portion of the resulting Allen-Hickling plots. Under the assumption that quasireversible reactions have a transfer coefficient near 0.5 allows the assignment of n = I for each of the two oxidation processes. The number of electrons was determined for the overall oxidation process by carrying out exhaustive oxidation of bromide, present in a slightly basic melt, at a potential anodic of the second oxidative process. The value of n for the overall process is 1.0. The proximity of the two oxidation waves precluded the use of coulometry in determining the number of electrons involved with the individual oxidation processes. Table I shows a summary of results for the number of electrons and transfer coefficients for the oxidation processes. Finally. tribromide in the form of tetramethylammonium tribromide was added to an N = 0.490 bromide melt. The cyclic voltammograms of the melt run before and after the addition of the tribromide are shown in Fig. 6. Although the technique is not a definitive proof of the identity of the species being oxidized, it stiongly supports the assignment of the tribromide anion to the second oxidative process. The addition of known tribromide also clarifies the reductive process. The anodic peak at L.1V and the cathodic peak at 0.7V greatly increase, thus sub-
D-n -/gcms T
Diem" s
K
1.2, 10 2.0 , l0 2.4 K10
12.0 mA
K
1.6, 10 2.7 , 1011 2.8 10"
stantiating that tribromide is both oxidized and reduced during the voltammetry. The assignments pruposed by Popov and Geske (3) for the oxidation of bromide in other nonaqueous electrolytes apply, also, to the molten salt electrolyte. The first anodic process is the oxidation of the bromide to bromine by the 1-!f rcactioii
[5] 2Br --Br 2 + 2e after which the resulting bromine undergoes a chemical reaction with the remaining bromide to form the tribromide anion Br 2 + Br --Br3
[61
The second anodic process isthe oxidation of the tribromide by the half reaction 2Br3 - 3Br2 + 2e [7] The diffusion coefficients of the bromide and tribromide anions were determined by rotating disk linear sweep voltammetry. The values determined for the diffusion coefficients and the corresponding Stokes-Einstein products are listed in Table II along with the literature values determined by Iwasita and Giordano (13) in acetonitrile. Also listed in Table II are the values for the diffusion of chloride through a chloride melt (14) and acetonitrile as dctermined by Serano et al. (15). The lower values of the diffusion coefficients in the molten salts compared to in acetonitrile are generally explained by the high viscosities of the molten salts. These results support the model of a highly organized crystal structure in the melts, which has been reported previously by our laboratory (16). The higher diffusion coefficient for the tribromide, compared to the bromide and chloride, is puzzling at first glance. The explanation probably lies in the fact that chloride (and probably bromide) hydrogen bonds to the organic cation to a greater extent than the tribromide. The heterogeneous rate constants were determined for the bromide and the tribromide ions by the peak separation method of Nicholson (17) and are shown in Table III. The magnitude of the rate constant (10 2-10
a
2
3)
indicates
that there is an appreciable degree of electrochemical irreversibility for these two processes, even at the platinum electrode. Conclusion The oxidation of bromide in the low-temperature bromoaluminate melts is more complex than chloride oxidation in the similar chloroaluminate melts. The two-step electrochemical oxidation is similar to that seen in acetonitrile. The electrochemical oxidation occurs in two steps involving tribromide ion, similar to the processes observed in other nonaqueous solvents. Otherwise, the melts behave similarly to the analogous chloroaluminate system. Table III. Heterogeneous rate constant as a functio- of sweep rate
L
L
2.0
1
1.0
r... -1.0 Potential/V 0.0
_
I -2.0
Fig. 6. Cyclic voltammogrom of an N = 0.490 bromide melt (a) before and (b) after the addition of tetromethylommonium tribromide.
vim V/s
AEpmV
50 100 200 500
77 81 96 117
Wave 1 k'/cms
10'
2.78 - 0.68 3.30 - 0.71 2.68 0.42 2.38 0.22
AEp/mV 86 107 79 94
Wave 2 k'/cmis - 101 6.56 * 1.04 4.87 - 0.71 17.6 - 3.5 16.0 - 2.6
J. Electrochem. Soc., Vol. 138, No. 2, February 1991 © The Electrochemical Society, Inc. 469 Acknowledgments 7. D. W. Seegmiller, G. W. Rhodes. and L. A. King, Inorg. Nucl. Chem. Letters. 6, 885 (1970). The authors thank Mr. J. Lloyd Pflug and Mr. F. C. Kib8. F. D. Chattaway and G. J. Hoyle, J. Am. Chem. Soc., 45, ler for their expert technical assistance. This work was 654 t1923). supported by the Air Force Office of Scientific Research. 9. R. E. Buckles, A. I. Popov, W. F. Zelezny, and R. J. Smith, ibid., 73, 4525 (1951). Manuscript submitted Dec. 29, 1989; revised manuscript 10. C. L. Hussey and J. R. Sanders, This Journal.134, 1977 received Sept. 4, 1990. (1987). The Frank J. Seiler Research Laboratory assisted in 11. B. P. Gilbert, in "Molten Salt Chemistry: An IntroducTeFang the Seblicaeron Ressachi abrt. ation and Selected Applications," G. Mamantov and meeting the publication costs of this article. R. Marassi, Editors, p. 213, NATO ASI Series, D. Reidel Publishing, Dordrccht, Holland (1986). REFERENCES 12. "Laboratory Techniques in Electroanalytical Chem1. F. H. Hurley, U.S. Pat. 2,446,331 (1948). istry," P. T. Kissinger and W. R. Heineman, Editors, 2. C. J. Dymek, G. F. Reynolds, and J. S. Wilkes. This Marcel Dekker, Inc., New York (1984). Journal.134, 1658 (1987). 13. T. Iwasita and M. C. Giordano, Electrochim. Acta, 14, 3. A. I. Popov and D. H. Geske, J. Am. Chem. Soc.. 80, 1045 (1969). 5346 (1958). 14. M. Lipsztajn and R. A. Osteryoung, Inorg. Chem., 23, 4. J. S. Wilkes, J A. Levisky, R. A. Wilson, and C. L. Hus1735 (1984). sey, lnorg. Chem.. 21, 1263 (1982). 15. L. Serano, V. A. Macagno, and M. C. Giordano, Electro5. J. R. Sanders, E. H. Ward, and C. L. Hussey, This Jourchim. Acta, 17, 561 (1972). nal, 133,325 (1987). 16. K. M. Dieter, C. J. Dymek, N. E. Heimer, J. W. Rovang, 6. Erratum: J. R. Sanders, E. H. Ward, and C. L Hussey, and J. S. Wilkes, J. Am. Chem. Soc., 110, 2722 (1988). ibid.. 133, 1526 (1987). 17. R. S. Nicholson, Anal. Chem., 37, 1351 (1965).
REPORT DOCUMENTATION PAGE
FOM
A Nrm 70pro 18d
reocirtnCi currdeo for tni ciection of noornmatlonm~go to a~eriqe I ioer estoorse, rlcwa~irg the time for reviewng instruacrs. 5ear~minq existing -jta soa,'',, p3tr.ef'ng and m~aintaini~ng thedata needed. and con-pieting ano re,.C-mg :-e c ;ilet~on of information Senld comments regarcing this buarden estimate or any cther asoert of tmis collection of lnfOmatjon. rictucir.9 srugostion%for reducing this orcrd . to reaadauarter5s 'ervices. Direciorate for information Qoerations and Peo'orl". 121 ffso oa~is Ht-y. ae 1204, Aringtcn., J,! 22202-4302. and to (me -?foce 0r Management and Budget. Papernork Reduc*,cnProme-t (0704-0183)1 VnasMngton, DC 20503 Pubfic
~vastnntOn'
1. AGENCY USE ONLY (Leave blank) 4. TITLE AND SUBTITLE
A
12. REPORT DATE
Elcrceia
In
13. REPORT TYPE AND DATES COVERED stg
io
ofhe
1-Methyl-3-Ethylimidazolium Bromide Aluminum Bromide Molten Salt System
5. FUNDING NUMBERS
F2-10
6. AUTHOR(S)
Jeffrey A. Boon John S. Wilkes John A. Lanning______
_______
7. PERFORMING ORGANIZATION NAME(S) AND AORESS(ES)
Frank J.
8. PERFORMING ORGANIZATION
REPORT NUMBER
Seiler Research Laboratory
USAF Academy Colorado
80840-6528
FJSRL/JR-91-0004
9. SPONSORINGi/MONITORING AGENC( NAME(S) AND ADDRESS(ES)
10. SPONSORING ,MONITOqING
AGENCY REPORT NUMBER
AF Office of Scientific Research Building 410 Boiling AFB, DC 20332
11. SUPPLEMENTARY NOTES
I12b. DISTRIBUTION
12a. DISTRIBUTION / AVAILABILITY STATEMENT
CODE
DISTRIBUTION UNLIMITED
13. ABSTRACT (Maximum 2C0 words)
The electrochemical behavior of the room temperature molten salt made by combining 1-methyl-3-ethylimidazolium bromide and aluminum bromide was studied. It was found that the bromide system behaved, for the most part, analogously to the chloride counterpart. Bromide solvated in a room temperature molten salt is electrochemiclaly similar to bromide in other nonaqueous solvent systems. The diffusion coefficients and the heterogeneous rate constants were determined for the bromide and tribromide species in the molten salt.
14. SUBJECT TERMS
15. NUMBER OF PAGES
5 16. PRICF CODE 17.
SECURITY CLASSIFICATION
OF REPORT
UNCLASSIFiED NSN 75,:0-011280-5500
18.
SECURITY CLASSIFICATION
OF THIS PAGE
UNCLASSIFIED
j
1.SECURITY CLASSIFICATION
20. LIMITATION OF ABSTRACT
OF ABSTRACT
UNCLASSIFIED
UNLIMITED Stardard Form 298 (Rpv 2-89) P''%cr,ond by -IS - 3A
..... .041 ... .... ____
Sto Z318