Reversible, Room-Temperature Ionic Liquids. Amidinium Carbamates

Taisuke Yamada, Paul Joseph Lukac, Tao Yu, and Richard G. Weiss ... Nikolay Tumanov , Balázs Zsirka , Ágnes Gömöry , László Kollár , Rita Skoda-Földes...
0 downloads 0 Views 64KB Size
Chem. Mater. 2007, 19, 967-969

Reversible, Room-Temperature Ionic Liquids. Amidinium Carbamates Derived from Amidines and Aliphatic Primary Amines with Carbon Dioxide Taisuke Yamada, Paul Joseph Lukac, Mathew George, and Richard G. Weiss*

967 Scheme 1

Department of Chemistry, Georgetown UniVersity, Washington, D.C. 20057-1227 ReceiVed NoVember 2, 2006 ReVised Manuscript ReceiVed January 16, 2007

Interest in ionic liquids (ILs), especially room-temperature ionic liquids (RTILs),1 has increased enormously during the past decade because, among other applications, they may be used to replace less environmentally friendly (“green”) solvents.2,3 Supercritical carbon dioxide has also been employed as a green solvent, although it requires specialized equipment and reaction vessels.4 Here, we report a new class of solvents which can be cycled between a RTIL state and a non-ionic solvent by exposing the liquid mixtures sequentially to 1 atm of carbon dioxide and molecular nitrogen (Scheme 1). These systems obviate the necessity of employing specialized equipment, including maintaining absolutely dry atmospheres, and offer the distinct advantage of creating environments which can be made either to dissolve or phase separate solutes and added liquid components depending on their polarity. Although it is known that amidines and alcohols5 react very rapidly with CO2 to form amidinium carbonate salts, some of which are liquids at room temperature, they are stable only under scrupulously dry conditions. Jessop and co-workers6 have demonstrated that the fraction of the amidinium carbonate made by bubbling CO2 through a carefully dried 1:1 mixture of 1,8-diazabicyclo[5.4.0]undec7-ene (DBU) and 1-hexanol7 can be returned to its original state by bubbling through N2 or Ar gas. Also, bubbling CO2 through a dilute solution of DBU8 or an N′-alkyl-N,Ndimethylacetamidine9 in water yields an acyclic amidinium bicarbonate. The acyclic amidinium bicarbonate is a reversible surfactant.9 (1) (a) Welton, T. Chem. ReV. 1999, 99, 2071-2083. (2) (a) Short, P. L. Chem. Eng. News 2006 (April 24), 15-21. (b) Weyershausen, B.; Hell, K.; Hesse, U. Green Chem. 2005, 7, 283287. (c) Seddon, K. R. Nat. Mater. 2003, 2, 363-365. (3) (a) Ionic Liquids: Industrial Applications to Green Chemistry; Rogers, R. D., Seddon, K. R., Eds.; ACS Symposium Series 818; American Chemical Society: Washington, DC, 2002. (b) Ionic Liquids IIIA: Fundamentals, Progress, Challenges, and Opportunities; Rogers, R. D., Seddon, K. R., Eds.; ACS Symposium Series 901; American Chemical Society: Washington, DC, 2005. (4) DeSimone, J. M., Tumas, W., Eds. Green chemistry using liquid and supercritical carbon dioxide; Oxford University Press: New York, 2003. (5) Hori, Y.; Nagano, Y.; Nakao, J.; Fukuhara, T.; Taniguchi, T. Chem. Express 1986, 1, 224-227. (6) Jessop, P. G.; Heldebrant, D. J.; Li, X.; Eckert, C. A.; Liotta, C. L. Nature 2005, 436, 1102. (7) The uptake of CO2 by DBU/alcohol in toluene solutions is not quantitative at 1 atm of pressure. The percent uptake of CO2 by neat DBU/alcohol systems was not mentioned by Jessop et al.6 Hori, Y.; Nagano, Y.; Nakao, J.; Taniguchi, H. Chem. Express 1986, 1, 173-176. (8) (a) Hori, Y.; Nagano, Y.; Miyake, S.; Teramoto, S.; Taniguchi, T. Chem. Express 1986, 1, 311-314. (b) Heldebrant, D. J.; Jessop, P. G.; Thomas, C. A.; Eckert, C. A.; Liotta, C. L. J. Org. Chem. 2005, 70, 5335-5338.

We have explored reactions of primary and secondary alkylamines with CO2 to obtain alkylammonium alkyl carbamates which are gelators10 of a variety of organic liquids.11 In those studies, the objective was to make solid ammonium carbamates which self-assemble into solid fibrillar networks and trap microscopically very large volumes of the organic liquids. Reactions between amines and CO212 have also been used to effect the self-assembly of ammonium carbamates into various scaffolds and to produce sensors.13 Of the 48 amidine/amine combinations shown in Table 1, 23 can be interconverted repeatedly at room temperature between their RTIL (Ln-C) and non-ionic liquid forms (L+n) by the treatments shown in Scheme 1. Primary amines were chosen for this study because they react more readily with CO2 than secondary (or tertiary) amines.14 In general, the RTILs were more viscous than the L+n mixtures from which they were made. Qualitatively, the RTILs with 1 and 2 were less viscous than the other Ln-C RTILs made from the same amidine. Evidence for the presence of amidinium carbamates comes from thermogravimetric analyses (TGA) of the weight loss (presumably of CO2) upon heating, from following the quantitative uptake of CO2 by the L+n mixtures as a function of time, and from comparisons of IR spectra of the amidine/amine mixtures before and after addition of CO2; experimental details may be found in Supporting Information. As opposed to DBU/alkanol solutions, the uptake of CO2 by neat amidine/amine solutions (including those with DBU as the amidine) is quantitative within experimental error when the resulting amidinium carbamate is a liquid at room temperature.7 In the absence of CO2, all of the amidine/amine combinations in Table 1, except those with 1,6-diaminohexane (8), (9) Liu, Y.; Jessop, P. G.; Cunningham, M.; Eckert, C. A.; Liotta, C. L. Science 2006, 313, 958-960. (10) (a) Terech, P.; Weiss, R. G. Chem. ReV. 1997, 97, 3133-3159. (b) Weiss, R. G., Terech, P., Eds. Molecular Gels. Materials with SelfAssembled Fibrillar Networks; Springer: Dordrecht, The Netherlands, 2006. (c) George, M.; Weiss, R. G. Acc. Chem Res. 2006, 39, 489-497. (11) (a) George, M.; Weiss, R. G. J. Am. Chem. Soc. 2001, 123, 1239310394. (b) George, M.; Weiss, R. G. Langmuir 2002, 18, 7124-7135. (c) George, M.; Weiss, R. G. Langmuir 2003, 19, 1017-1025. (12) Rudkevich, D. M.; Xu, H. Chem. Commun. 2005, 2651-2659. (13) Hampe, E. M.; Rudkevich, D. M. Tetrahedron 2003, 59, 9619-9625. (14) (a) Hoerr, C. W.; Harwood, H. J.; Ralston, A. W. J. Org. Chem. 1944, 9, 201-210. (b) Lallau, J.-P.; Masson, J.; Guerin, H. Bull. Soc. Chim. Fr. 1972, 3111-3112. (c) Nakamura, N.; Okada, M.; Okada, Y.; Sugita, K. Mol. Cryst. Liq. Cryst. 1985, 116, 181-186.

10.1021/cm062622a CCC: $37.00 © 2007 American Chemical Society Published on Web 02/06/2007

968 Chem. Mater., Vol. 19, No. 5, 2007

Communications

Table 1. Phases of Neat 1:1 (mol/mol) Amidine/Amine Mixtures before (B) and after (A) CO2 Bubblinga amine 1

2

3

4

5

7b

6

8b

amidine

B

A

B

A

B

A

B

A

B

A

B

A

B

A

B

A

DBUc A B C D E none

l l l l l l l

l l l l l l s

l l l l l l l

l l l l l l s

l l l l l l l

s s s s s m s

l l l l l l l

l l s s s s s

l l l l l l l

l l l l l l s

l l l l l l l

l m l m m s s

l l l l l l l

l m m m m s g

m m m m m m s

s s s s s s s

Figure 1. TGA curves for neat (a) E+5 and (b) E5-C heated from room temperature. Curve c is curve a minus curve b.

a l, s, m, and g indicate respectively liquid, solid, liquid-solid mixture, and glassy transparent material which crystallized partially over a period of weeks. b Diamines were mixed with 2 molar equiv of amidines. Amidine/8 mixtures were heated slightly to dissolve 8, and then CO2 was bubbled through the warm solutions. c DBU was dried by the procedure outlined in ref 6 and distilled. However, a small amount of water remained, as indicated by a cloudy appearance after CO2 was bubbled through a neat sample.8b

yielded liquid solutions at room temperature. Solutions of 8 were made by heating it gently in the presence of an amidine partner. All of the combinations employing “dried” DBU remained liquids after addition of CO2 except those with 8 or the bulkiest amine, t-butylamine (3). The sensitivity of the systems to moisture depended on the nature of the primary amine. When commercial DBU was used as received, only n-butyl amine (1) and 1,2-diaminoethylene (7) yielded RTILs upon addition of CO2. Some RTILs employing even “dried” DBU (DBU1-C, DBU2-C, DBU5C, and DBU6-C) became cloudy after standing for 1 day, indicating the aggregation of amidinium bicarbonates from residual water. However, when CO2 was bubbled through C+1 containing 3 wt % of added water, a liquid phase was obtained and persisted for long periods.15 Thus, the aliphatic amidine/amine systems should be easily adapted for many applications without maintaining scrupulously dry conditions. A comparison of the results from the acyclic amidines with the isomeric butyl amines (1-3) provides additional insights into the dependence of the phase formed upon CO 2 bubbling and the molecular structures of the constituent molecules. The unbranched n-butyl and sec-butyl amines yielded ILs with all of the acyclic amidines in Table 1; t-butyl amine produced only solids. Additionally, results from 1, 5, and 6 with the acyclic amidines suggest that longer chain lengths may discourage RTIL formation. This tentative conclusion is bolstered somewhat by the observation that only the acyclic amidines with the shortest n-alkyl chain lengths yield RTILs with cyclohexyl amine (4) and n-octyl amine (6). Additional studies with both shorter and longer chain lengths will be conducted in the future to determine whether there is an optimal intermediate chain length for the RTILs. Results from TGA measurements of E+5 and E5-C are included here as an example (Figure 1). The plateau in Figure 1c above 50 °C corresponds to a loss of 12.4% sample weight, in excellent agreement with the calculated CO2 weight loss, 12.8%. The TGA measurements also indicate that some of the amine and amidine components are being evaporated and entrained in the nitrogen wind which the samples experience during the experiments. Regardless, it (15) We have not tested yet all of the RTILs with acyclic aliphatic amidines for their ability to remain liquids in the presence of water. However, we believe that they will behave like the C1-C system.

Figure 2. Plots of percent CO2 uptake by E+5 (b), C+2 (∆), E+6 (2), and 5 (O) as a function of time. See eq 3 in Supporting Information.

should be possible to employ the RTIL amidinium carbamates above room temperature while retaining their reversibility with the non-ionic components in a closed vessel.16 The time dependence of the uptake of CO2 by stirred mixtures of C+2, E+5, and E+6, as well as by neat 5, was followed using a gas burette filled with 1 atm pressure of CO2 (Figure 2).17 As expected,11 about one-half an equivalent of CO2 (47%) was taken up by neat 5; the formation of hexylammonium N-hexylcarbamate was virtually quantitative.11 Although the time profiles for uptake of the CO2 gas must depend to some extent on the rate of stirring and the liquid-gas surface area, it is clear from the data that the two mixtures which form RTILs, C+2 and E+5, behave very differently from the mixture which leads to a solid, E+6 (Table 1). Uptake of CO2 by C+2 and E+5 was rapid initially and then continued very slowly, reaching a plateau value after approximately 30 min. The eventual uptake exceeded the theoretical amount by approximately 4 and 10%, respectively. The solubility of CO2 gas in many ILs is known to be very high.18 Both the excess amount and the time profile for its uptake indicate that the rapid (ca. 100%) part is due to formation of the RTIL; thereafter, more CO2 becomes dissolved in the ionic phase. The rate of uptake by E+6, leading to E6-C, was initially as rapid as with the other amidine/amine samples, but it slowed after a short period of time and reached an eventual plateau corresponding to only approximately 73% of the theoretical value because solidification of E6-C traps the remaining liquid of E+6 and isolates it from contact with CO2. Formation of amidinium carbamate is also evident from the Fourier tranform infrared spectra of E+5 and E5-C. After (16) An obvious strategy for increasing the useful temperature ranges of both the amidinium carbamates (as well as carbonates) is to place them under several atmospheres of CO2 pressure or to replace CO2 with CS2. The temperature ranges and pressures over which they remain stable will be determined in future experiments. (17) House, H. O. Modern Synthetic Reaction, 2nd ed.; W. A. Benjamin: Menlo Park, CA, 1972; p 2. (18) (a) Cadena, C.; Anthony, J. L.; Shah, J. K.; Morrow, T. I.; Brenneck, J. F.; Maginn, E. J. J. Am. Chem. Soc. 2004, 126, 5300-5308. (b) Shiflett, M.; Yokozaki, A. Ind. Eng. Chem. Res. 2005, 44, 44534464.

Communications

Figure 3. Chemical shifts of protons (a) and 13C (NdCsN(CH3)2) (b) in a 175 mM E+5 in CDCl3 solution upon alternation between CO2 and N2 bubbling. For part a: NdC(CH3)sN (b, 1.85-2.1 ppm), NdCsN(CH3)2 (2, 2.85-3.15 ppm), and sCH2NdCsN (9, 3.15-3.35 ppm). Photographs are of a 4:2 (v/v) decane/E+5 solution (left) and a decane/E5-C two-phase mixture (right) with Eosin Y (0.26 mM).

CO2 bubbling, the NdC stretching frequencies of E+5 at 1629 cm-1 19 were replaced by bands for E5-C at 1646 and 1575 cm-1 which can be assigned to protonated amidine19 and carbamate stretches11b (Supporting Information, Figure S1). The chemical shifts of selected 1H and 13C resonances in NMR spectra of Ln-C but not of the corresponding nuclei in the L+n mixtures were sensitive to concentration in the range explored, 10-350 mM. A detailed study of this concentration effect was conducted in CDCl3 solutions of the E+5/E5-C system (Figure 3 and Supporting Information, Figures S2 and S3). The nuclei of E most affected, the protons of the geminal N-dimethyl groups (NdCsN(CH3)2), the C-methyl protons (NdC(CH3)sN), the R-methylene protons (sCH2NdCsN), and the amidine carbon atom (Nd C(CH3)sN), exhibited only small peak shifts upon addition of CO2. However, infrared spectral studies conducted with a different amidine-amine pair at 10 mM, showing the presence of the distinctive IR stretching frequencies of amidinium and carbamate, and our previous NMR study of the concentration dependence on chemical shifts of protons and carbon atoms in alkylammonium alkylcarbamates20 provide strong indirect evidence that E5-C was present after bubbling CO2 even through the 10 mM solution. The shifts, primarily to lower field, increased with increasing solute concentrations upon addition of CO2 and reached plateau values at approximately 100 mM. By bubbling N2 gas through the solutions with higher concentrations of E5-C (N.B. 175 mM; Figure 3), the proton signals could be returned to their E+5 values; the system was cycled between E+5 and E5-C, with little or no degradation of the sample. The solvatochromic dye,21 1-(p-dimethylaminophenyl)-2nitroethylene (DAPNE),22 has been used to estimate the (19) The wavenumber assignment of the protonated amidine was confirmed by the IR spectrum of a 1:1 E/trifluoroacetic acid mixture which showed a sCdN stretch at 1648 cm-1. Protonation of the amidines has been reported: Corset, C.; Froment, F. J. Phys. Chem. 1990, 94, 6908-6911. (20) George, M.; Weiss, R. G. Langmuir 2003, 19, 8168-8176. (21) (a) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358. (b) Reichardt, C. Chem. Soc. ReV. 1992, 21, 147-153. (22) Richter-Egger, D. L.; Tesfal, A.; Flamm, S. J.; Tucker, S. A. J. Chem. Educ. 2001, 78, 1375-1378.

Chem. Mater., Vol. 19, No. 5, 2007 969

polarity of our reversible RTIL systems. Its 423 nm absorption maximum in E+5 indicates an environment slightly less polar than toluene (λmax ) 425 nm).22 After CO2 exposure, the E5-C absorption is shifted to 438 nm, indicating an environment more polar than acetone (λmax ) 433 nm)22 and less polar than N,N-dimethylformamide (DMF; λmax ) 446 nm;22 (See Table S2 and Figure S4 in Supporting Information). In general, E5-C appears to be less polar than other ILs. For example, 1-butyl-3-methyl imidazolium hexafluorophosphate is much more polar than DMF based upon ET(30) values.23 As such, it is possible to dissolve reversibly less polar materials into the L+n and separate them from Ln-C. For example, after bubbling CO2 through for