'Iv' . O‘r‘ my A n- n 'Iv “A ”‘1! .n a.“ ‘ I v. - ,'. i w w. 'u n v (w (my. , V . . . . ., . . . , . ‘ . Bravo. if. I‘ iv :Il.‘«'l O ‘ .. ‘ . .. . 1" tr. Jvffluih-IIIKEQ... up. . cflbufi_ . . .. . , v ‘ ‘ . . .. J??? Econ... .4. .. I. . «mmhuummfin ow. . .IJ'D“: . Sea. .5 (I! . . V ‘ . « haqfiz a ct. . l. ("3? .1, , v. $1 5.1:... 2...... . 9 .flo...‘ . . loilyvllivu. I‘D-fl! a . ‘1! In. I 5.1.552].- ..II" .e a v v: I... 9m! '12. - I. I . $1 I.“ :I. 7 1.9L. . v {If It»?! {Ii-i: I I: 1!}... .1: 'pvhc" I (If? I? - ’rrl: Etc! V : i):- .- I .. 0.4V! «I u to»! II. .h-nrhnlrvf. ..'g y” way. u. ii! , . 5 Evil: rs..." II‘ I! . . .. . 40 a. . (1 7|.) inn-575:! .UD I},NJJ.5-fid:lrr ‘ 1.... .erP. .' How” A ['05: .36? . .vt ....“ .19... . En“. t w L‘. .z' 11 .39 1:} , .[tri' .péc: O u It". i :0 -P 4"" . . . nu...t..:.f¢. 1.3.14.2“... ., 4 3.30.. . I. I {v .“I: uni ~»_ 6 Iss- \" at. . .I fill-5h. flinttp Ail Iii. evifv .I. .réttltéisilhhtzlvcb I 7h“ s". I ? r!!!..(ll€ll|l x. 39h..." . 5.3... up «Hand? . smut... 1.1... . a... _ . i ‘ u I}: 6.9, l. .5133“. (.1 t! '1 l O ‘A r. 1...: . rill I luv“.- 009.51; 1.9.47144‘ rvithvl 23’ Pbiv‘u. .‘rvlcdd.rflfl.huaajizw . . . . ll , .. . J. .. Fizfi l . 2. DY » . . uL\ _.f E! p r WMWWWMM 3(Ggg%%b 3 1293 005503 LIBRARY w r Michigan State University This is to certify that the thesis entitled SPECTROSCOPIC STUDIES OF N-METHYLFORMAMIDE(NMF) AND BINARY NMF/MIXED SOLVENT SOLUTIONS OF ELECTROLYTES presented by Chuang, Huey-Jan has been accepted towards fulfillment of the requirements for Master Science degree in if? dMajor professor Date August 9, 1988 0-7539 MS U is an Affirmative Action/Equal Opportunity Institution ‘TVIESI_] BEIURNING MATERIALS: Place in book drop to LJBRARJES remove this checkout from .—c-- your record. FINES will be charged if book is returned after the date stamped below. SPECTROSCOPIC STUDIES OF N-METHYLFORMAMIDE(NMF) AND BINARY NMF/MIXED SOLVENT SOLUTIONS OF ELECTROLYTES by Chuang, Huey-Jan A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1988 ABSTRACT SPECTROSCOPIC STUDIES OF N-METHYLFORMANIDE(NMP) AND BINARY NEE/MIXED SOLVENT SOLUTIONS OF ELECTROLYTES BY Chuang, Huey-Jan Spectroscopoic methods were used in this study of alkali salts in pure N-methylformamide(NMF) and NMF- containing binary solvent mixtures. UV-VIS spectra of alkali picrates (LiPi, NaPi and KPi) indicate that these salts are completely dissociated in NMF solution. In contrast to some other solvents, Na+'C222 complex in NMF has a relatively low formation constant (log K = 2.25 i 0.02). Preferential solvation for Na+ and Cs+ ions was investigated in twelve binary solvent mixtures by NMR chemical shift measurements. The relative solvating ability toward Na+ is HMPA >> H20 > DMSO > DMF ~ NMF ~ FA ~ neon > EtOH > PY >> THF > AC ~ AN >> NM ; and that for the Cs+ ion is DMF > HMPA > DMSO ~ NMF ~ FA >> PY. The order is related to the donicity of the solvent except for pyridine and ethanol. Linewidth measurements show that the environment of the Na+ ion in NMF-HMPA and NMF-NM systems varies widely with solvent composition. The dependence of the 23Na chemical shift and linewidth on the nature of the anion and the salt concentration was measured in NMF-AN, NMF-DMSO, NMF-HZO and NMF-HMPA mixtures. Only in AN-rich Chuang, Huey-Jan mixtures was any effect observed. Picrate anion solvation was studied in NMF-AC and NMF-DMF systems by UV-VIS spectroscopy. The results indicate that picrate anions are preferentially solvated by NMF. (iv 0; ‘1 4‘ 13L \ffiggfl (iii)? i warm-av ii ACKNOWLEDGEMENTS The author would like to thank Dr. Alexander Popov and Dr. George Leroi for their guidance and encouragement during the past year. Also, she would like to thank Dr. Jehn Gaudiello for his comments and corrections during the writing of this thesis. Gratitude must be extended to Dr. Shen-Wen Ko for his support, good suggestions and encouragement in past years. Thanks also to Chung-Shang Institute of Science and Technology for financial support. Appreciation is expressed to the members of Dr. Popov’s research group, Yongshen Ron and Lee-Lin Soong, for their help and friendship during this year. Special thanks are given to Claudia Turro for her friendship and expenditure of time in proofreading this thesis. Deep appreciation must be extended to my family for their understanding, encouragement and love which helped me through this year and the writing of this thesis. iii TABLE OF CONTENTS Page LIST OF TABILESooooooooooooo ..... 000000000000 000000 0000 Vi LIST OF FIGURES0000000000000000000000000000000000 00000 Viii I0 II. INTRODUCTION AND HISTORICAL REVIEW A. Introduction................................ 1 B. Historical Review........................... 2 1. Interactions in Electrolyte Solutions.. 2 2. Interactions in N-methylformamide SOlutions.0000O000000000000000000000000 11 3. Preferential Solvation in Binary SOlventMixtures00000000000000000000000 19 4. Preferential Solvation in Binary Solvent Mixtures Containing N-methylformamide................. ..... 31 EXPERIMENTAL PART A. Material Purification .......... ............. 36 1. Reagents.. ............ ................. 36 2. Solvents...... ...... ................... 37 B. Sample Preparation.......................... 38 C. Spectroscopic Measurements.................. 39 1. IR................... .......... . ...... . 39 2. Raman...... ............. ............... 39 3. UV-VIS.... ................ ............. 40 4. NMR.................................... 40 iv Page III. RESULTS AND DISCUSSION A.- Ion Association Studies in Pure N-methYI fomamide 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 2 10 Picrates0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 42 2. Na+C222 Formation Constant............ 48 B. Ion Solvation Studies in Binary Solvent Mixtures0000000000000.000000000000000000000 51 1. NMR Studies of Preferential Solvation............................. 51 (1) Isosolvation Point Measurements.. 51 (2) Linewidth Measurements........... 87 2. UV-VIS Studies of Sodium Picrate...... 93 C. Vibrational Spectroscopic Studies.......... 95 D. Suggestions for Future Work....... ...... ... 98 REFERENCES............. ..... ......................... 100 Table 10 11 12 13 14 LIST OF TABLES Page Physical Properties of Solvents......... 4 Infrared and Raman Frequencies (cm-1) of Alkali Thiocyanate Ion Pairs in Tetrahydrofuran......................... 8 Raman Stretching Frequencies and Assignments............................. 8 Wave Numbers (cm_1) of the Main Absorp- tion Bands of FA, NMF, and DMF Mole- cules in a Medium of Acetonitrile, and the Variation Under the Influence of 018801ved salts.000000000000000000000000 12 Raman u- and v - Frequency Maxima so so of the NH Stretéying Band of NMF: the frequency shifts related to the pure solvent and the isotropic-anisotropic separations/Auiso-aniSO/k-denotes the salt/amide molar concentration ratio.... 17 Preferential Solvation for Cations...... 27 Preferential Solvation for Anions....... 28 Concentration Dependence of UV-VIS Absorbance (365 nm) of Picrate Salts.... 44 UV-VIS Peak Absorbance Variation of Picrate Salts in NMF Solution........... 46 Formation Constant of Na+.C222 Complex.. 50 Variation of 23Na Chemical Shift and Linewidth in Mixed Solvents............. 58 Variation of 133Cs Chemical Shift and Linewidth in Mixed Solvents............. 69 Isosolvation Point of Na+ and Cs+ Ions Binary Solvent Mixtures................. 71 Anion Dependence on 23Na Chemical Shift and Linewidth..... ..... ................. 73 vi 15 16 17 Anion Effect on Isosolvation Point...... 79 Concentration Dependence on 23Na Chemical Shift and Linewidth............ 80 Peak Position of UV-VIS Absorption of Sodium Picrate in Mixed Solvents........ 96 vii Figure 7a 7b 10 LIST OF FIGURES Page 1/T of Li+ vs. salt concentration in solutions of LiCl and LiClo4 in formamide, N-methylformamide N,N- dimethylformamide, and dimethyl- sulfoxide at 25°C........................ 6 1/T1 of Li+ vs. mole fraction of the mixed solvent, x, in 1 aqm solutions of LiClO in the following mixed solvents: HZO- PA, HZO-DMF, HZO-F, and F-DMF..... 6 Smoothed R(u) curves of formamide, water and aqueous solutions of formamide00000000000.000000000000000000000 7 Spectra in 2-butanone at 25°C............. 10 Structure of N-methylformamide............ 13 Isotropic and anisotropic components for NaClO and LiClO4 solutions in NMF and chelate configuration of the Li(HCONH2)2 complexes................ 15 Apparent molar volumes of DMF, ¢V(DMF) in mixtures with DMSO, AN, NMF, water and methanol at 25° C.................... 18 Apparent molar volumes of A, ¢v(A), (A - DMSO, AN, NMF, and water) in mixtures with DMF at 25°C............... 18 Free energies of transfer of NaCl and AgCl from water to AN-HZO mix- tures at 25°C00000000000000000000000000 21 u(Cl-0) absorption bands of C104" anion in solutions of lithium per- chlorate in AN, in NM, and in NM/AN binary solvent............... 22 The variation of 25Cl chemical shifts of chloride salts in mixtures of CHZCN and DMSO with water..................... 25 viii 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Fractional change in u.v. peak maxi- mum for N03, 300 nm band in DMSO + Hzomixtures0000000000000000000000000000 26 23Na magnetic relaxation rates extra- polated to zero salt concentration in MeOH + H20 mixtures..................... 30 Plot of infinite dilution sodium-23 chemical shifts vs. the donor number of the solvent: nitromethane, aceto- nitrile, sulfolane, propylene car- bonate, acetone, ethylacetate, tetra- hydrofuran, dimethylformamide, tetra- methylurea, dimethylsulfoxide, water, pyridine, hexamethylphosphoramide, formamide, methanol and ethanol......... 32 23Na+ magnetic relaxation rates extrapolated to zero salt concentra- tion in formamide-Hzo and in NMF- H20 mixtures as a function of mole fraction of H20 in the solvent.......... 33 Concentration dependence of UV-VIS absorbance of picrate salts in NMF...... 43 Structure of picrate anion.............. 46 23Na resonance lines of Na+.C222 complexation in NMF..................... 49 23Na resonance lines of Na+.C222 complexation in DMF..................... 52 The curve of chemical shift vs. sol- vent composition....... ..... ............ 54 Sodium-23 chemical shift in mixed solve in mixed solvent.................. 55 Sodium-23 chemical shift in mixed salvents0000000000000000000000000000000 56 Sodium-23 chemical shift in mixed salventSeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee 57 Concentration dependence of sodium-23 chemical shift in NMF-EtOH mixtures..... 63 Structure of HMPA....................... 64 Cesium-133 chemical shift in mixed selvents0000000000000000000.000.000.000. 67 ix 26 Cesium-133 chemical shift in mixed solvents...... .......................... 27 Anion effect on sodium-23 chemical Shift in NMF-”0000000000000000000000000 28 Concentration dependence of sodium-23 Chemical Shift in NMF-AN. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 29 Hydrogen bonding between NMF and water.. 30 H-bonding between solvents.............. 31 Cesium-133 linewidth in mixed solvents.. 32a Sodium-23 linewidth in mixed solvents... 32b Sodium-23 linewidth in mixed solvents... 33 Sodium-23 linewidth variation in NMF- AN mixtures. ........ .................... 34 Excitation energy of picrate anion in mixed selventSeeeeeeeeeeeeeeeeeeeeeeeeee 68 77 82 84 85 89 91 92 94 97 CHAPTER I INTRODUCTION AND HISTORICAL REVIEW A . INTRODUCTION While electrolyte solutions have been studied intensively for well over a century, our knowledge of the ion-ion and ion-solvent interactions still remains very rudimentary. This ignorance is particularly true for solvents with dielectric constants higher than that of water and which have high solvating abilities. One of these solvents is N-methylformamide (NMF) which has a dielectric constant of 182 at room temperature and 300 at ~40° C; it is an interesting and little-studied solvent. Dilute solutions of electrolytes in NMF should not show any measurable degree of ionic association. Yet, as we shall note below, several investigators have reported cation-anion interactions in this solvent, and have calculated ion pair formation constants for a number of alkali halides by using electrical conductance or heats of dilution measurements. Others have established, but in a qualitative way, that N- methylformamide is an effective solvating agent for alkali and alkaline earth cations, and that the degree of solvation depends on the properties of the solvated cation. The lack of adequate models for describing electrolyte solutions is strongly magnified when we deal with solvent mixtures. Because of solvent-solvent interactions, physicochemical properties of solvents in mixtures very often are quite different from the properties of the neat liquids. Only a few fragmentary studies on binary mixtures 1 involving N-methylformamide have been reported in the literature. It would be highly desirable to establish the extent of ion-ion interactions in. a medium ‘with such a large dielectric constant and to clarify the nature of ion-solvent interactions. Additional studies are also necessary for a better understanding of electrolyte solutions in binary solvent mixtures. B. HISTORICAL REVIEW 1. Interactions in Electrolyte Solutions In an electrolyte solution, three types of interactions take place, whose relative importance depends on. the nature of the electrolyte and. the participating solvent. The first is the ion-solvent interaction. When an ion is surrounded solely by solvent molecules, it is called a solvated ion or a free ion. These are prevalent especially in solvents with high dielectric constant and high solvating ability. The second type of interactions are ion-ion interactions, which play an important role in concentrated solutions and/or in solvents with low dielectric constants. When the cation and the anion form an ion pair, but separated by several solvent molecules, the species is called a solvent-separated ion pair. When the cation and anion are in direct contact, we have a contact ion pair. The type of ion pair formed in electrolyte solutions is strongly affected by the nature of the solvent. Usually, contact ion pairs are formed in solvents with low dielectric constant and low solvating ability; solvent- separated ion pairs are formed in solvents with low dielectric constant but high solvating ability. The physical properties of the solvents studied in this thesis are collected in table 1. The third type of interactions are solvent-solvent interactions. 131 the liquid. phase, some solvents form dimers, trimers or oligomers by hydrogen bonding or dipole- dipole interactions between solvent molecules, and the solvent can be somewhat or highly structured. Electrical conductance measurements can distinguish free ions from ion pairs, but cannot distinguish contact ion pairs from solvent-separated ion pairs. Spectroscopic methods probe the nearest neighbors of an ion and provide information about the inner solvation shell of the ion. These techniques can distinguish contact ion pairs from free ions and solvent-separated ion pairs, but are unable to distinguish between free ions and solvent-separated ion pairs. However, spectroscopic methods are essentially powerful techniques for studying the interactions in electrolyte solutions. Mishustin and Kessler studied ion-solvent interactions and ion associations by' measuring the spin-lattice relaxation rate (1/T1) of the lithium cation in pure solvents 3 ZI+3 ( equ ) 2 ... .________ x .______ x h 1 T 40 I2(2I-1) 1 Ham.OI mmm.ou omv.on hb©.0l Hmm.OI 0mm.OI oom.OI 0mm.OI mum.on ONS.OI OH®.OI 0®®.OI mam.OI pHfinm AwlOHv .mH oesmflm no mafia one ow HaoHEooo azmm aoaosaao opaaaaaa one maaopaa so oooaauoo was osfla> mane * oo.m oHo.o m.om s.m zz ooasposoaoaz oa.m Hom.o m.sm ea za ofiaapaoopooa mo.m oom.o oe.om SH o< ooooooa oe.H ooo.o on.» om mme aaaaooaosoaApoe sm.m om.m o.moH s.om «m moasasaom sm.m moo.o s.mm s.om moo: Hoaaooos ow.m mow.o He.om o.om esp ooaeaEAoaasoooeHouz.z o.m ooo.H mo.oa w.om ooze ooaxooaomfiaoooEHo oo.H who.H om.am o.Hm scam Hooaoom amw.H omw.o om.ms mm on: goods sm.m vww.o m.mH H.mm so oaaoaaso oo.m mam.m oo.o m.mm ease ooasaaooomoooflaooosaxom ow.m oo.a o.mmH om mzz ooasasaoaasaoosuz AQV Am.dm MIOHV afiIHOE.Hdoxv a . c o .z.n .hna4 unopaom basebaom Ho nowauonoum HaoHMSAA H OHQGB and in binary solvent mixturesl. The concentration behavior of T1'1 for solutions of LiCl and LiClO4 in FA, NMF, DMF and DMSO, depended on the anion (Fig. 1); in mixed solvents a maximum in the value of T1'1 was caused by changing the symmetry of inner solvation shell (Fig. 2). With the same technique, Weingartner studied the interionic interactions of LiCl and NaBr in NMF solutionsz. His results indicated that the concentration dependence of the magnetic relaxation rate of alkali metal and halide nuclei was caused by ion-ion interactions in the solutions. Chabanel and Wang showed by using IR and Raman 3 that the dimerization of the KNCS and NaNCS spectroscopies ion pairs was more pronounced than for LiNCS in tetrahydrofuran. The frequencies assigned to ”C-N' ”C-S and VSCN vibration of the salts are listed in Table 2. Irish and Davis4'5 have inferred the hydration of the nitrate ion from the near IR frequencies of the anion. Huang and coworkers6 have shown through vibrational spectroscopy of acetonitrile solution that there are specific interactions between the halide ion and the CH3 group of acetonitrile. Raman studies of liquid NH3 and of NaI and NaClO4 in liquid N113 by Lagowski gt; a. showed the presence of ion-solvent and solvent-solvent interactions by the splitting and shifting of the NH band of ammonia7 (Table.3). Low frequency Raman spectra (in the 0-400 cm'l) of FA-HZO mixtures were used to study the solvent-solvent interactions Figure 1. 1/T of Li+ vs. salt concentration in solutions of LiCI(O) and LiClO (0) in formamide (1), N—methyl- formamide (2), N, -dimethylformamide (3), and di- methylsulfoxide (4) at 25‘C. Figure 2. 1/T1 of Li+ vs. mole fraction of the mixed solvent ,x, in 1 aqm solutions of LiClO in the following mixed solvents: H O—HMPA(0), MZO-DMSO(V), HZO-DMF (a), HZO-F(A), HZO-NMF(O), and F-DMF(0). From reference 1. INTENSITY .4 ~Ic1'nmoncn) llJJJllllJLlllllJllllJJlllllllllllllll 0 100 200 300 cm“ 400 Figure 3 . Smoothed R07) curves of (A) formamide, (I) water and (B) -(H) aqueous solutions of formamide. All spectra were obtained at room temperature. The volume percentages (v/v) of formamide in water are, (A) 100, (B) 95. (C) 90, (D) 70, (E) 50, (F) 30, (G) 10. (H) 5 and (I) 0. From reference 8. Table 2. Infrared and Raman frequencies (cm'l) ofaalkali thiocyanate ion pairs in tetrahydrofuran u(CN) v(CS) 6(SCN) Free Ion SCN' 2052, 2053 (R) 738, 735 (R) 465 Ion Pairs LiNCS 2065, 2066 (R) 781, 782 (R) 480 NaNCS . 2056; 2056 (R) 760, 760 (R) 475 KNCS 2050, 2051 (R) 752 473 Dimers (LiNCS), 2034, 2051 (R) 788, 813 (R) 503 and 489° (NaNCS)2 2042, 2052 (R) 769, 772 (R) 489 and 478 (KNCS),d 2044, 2052 (R) LiNamCS), 2040 758, 755 (R) 483 and 473 496 and 483 ‘1 Raman frequencies are indicated by (R). b From Bu,NSCN solutions. c In a mixture 20% THF + 80% C6H6. d In dioxolane. From reference 3. Table 3. Raman stretching frequenciesa and assignments Symmetry type Assignment C3 V; 91 ”5 Solution C“- 2v. 9‘ ”a N0: (25°) 2347(p)° 2373(p) 2403(p) 2521(dp) N02H (25°) 3342(p).3367(p) NH3(25°) 3214(p) 3271(p) 3300(0) 3385(dp) NH3(-71°) 3209(p) 3261(p) 3298(p) 3380(dp) NHg/Nal“ 3212(9) 3252(p) 3292(0) 3370(dp) NH;/NaClO.d 3222(9) 3280(9) 3306(9) 3393“”), a In cm‘ ‘. 0 p = polarized; dp = depolarized. C Mole ratio = 12. “ Mole ratio = 9. From referenence 7. region over the whole range of solvent compositions8 (Fig. 3). Hydrogen bonding involving the amide functional group has been of particular interest to investigators, and many theoretical and experimental results have been publishedg'15 on this topic. The nuclear magnetic relaxation rate studies by Hertz gt a1. provided structural information about FA and NMF in the pure liquid and in mixtures with CD3CN16. Gilkerson e_t_ _a_1. studied the concentration dependence of the ion pair formation equilibrium of lithium picrate and tetrabutylammonium picrate in 2-butanol by UV-VIS spectroscopy17 (Fig.4). The formation constants were calculated and compared to those obtained from conductance measurements. He interpreted the difference between KA (spec) and KA(cond) for lithium picrate in 2-butanol (KA (spec)=9440 2: 480; RA (cond)=6610 : 40) and cesium tetraphenylborate in acetonitrile (KA(spec)=40 :10: RA (cond)=17.7: 0.2) in terms of the ion pair formation. The electric field of the approaching Li cation perturbs the electronic energy level of the picrate anion at distances significantly greater than the contact distance, so that the UV-VIS spectra indicate a greater fraction of ions as paired than does the conductance techniques. It is the same condition for cesium tetraphenylborate in acetonitrile. In this laboratory, many studies of interactions in electrolyte solutions by IR , Raman and NMR 10 Figure 4. Spectra in 2- utanone at 25 C. Molar absorptivity (e, in cm‘lM- ) vs. wavelength(A, in nm):(a)0.176 mM tetra—n-butylammonium picrate; (b) 0.114 mM lithium picrate; (c) 1.02 mM lithium picrate; (d) 0.67 mM lithium picrate plus 15.3 mM lithium tetraphenylborate. From reference 17. 11 spectroscopieslsu22 have been carried out. The results provided very useful information about alkali metal ions and chloride ions solvated in non-aqueous solvents. Although many studies have been made in this field, unfortunately, sometimes the results of various experimental methods do not give the same answer. For example, the 23Na chemical shifts23 as well as IR and Raman spectroscopic results24 from nonaqueous solutions of NaBr, NaI have been interpreted in terms of contact ion pairs. On the contrary, the 1H magnetic relaxation rates of the solvents gave no evidence for ion pair formationzs. 2. ct o s ' -Me h ' t on From IR studies26 in the liquid phase it has been determined that 90% -92% of the N-methylformamide exists in the 'trans form (Fig' 5). Theoretical calculations also indicate that the trans form is more stable than the cis conformation. X-ray diffraction measurements and ab initio MO-SCF calculations27 show that the liquid structure of NMF is linear and flexible, and. that H-bonding' between NMF molecules produces strong interaction. As mentioned before, NMF has a very high dielectric constant, and the physical properties listed in Table 1 indicate that NMF is also a powerful dissociating solvent. As a matter of fact, IR studies of lithium, sodium and magnesium perchlorate and sodium iodide by Perelygin and coworkers showed that the ”C=O and ”C-N band positions of NMF depended only on the 12 a .3 ::::a:.:: E32. can .Anvx .m¢a_ .Ewmm.nu:;.m .:x ..s at .m._ .:_x>_c;c; mw_- onmm was- some ow“- comm new. A. mso_ __. new. 17. mar" cm_- meme cm. new. a- ::s_ s- mama A. w- oamm o- osvm .az em- omnm . : mom. a- seo_ mn+ mom. a". :so_ _m- mean we. amm_ m_i mac. 2m- moan m_xeo_;cuz a"- mean a mom. m- :so. .N. mem_ :_- rso_ mm- were .m. x:n_ s- has. ms; area ¢:_;_; s- some a mom. .c use. 3.. sxm_ m- mxc_ x- cams on. was. e1 cas— o- seem e:_;:z . soon 1 New. . use. . sums . are. . ween - new“ i v:s_ i teem . >< : 9< a :4 > be > e< a >< a >< > >4 3 . . e_am 2.0 Out 2.0 o o =1: 2.6 Ono =iz -2 i:::.:: , i - .ii--.iii.-;i--ti..i i,:,- . .i.i.:i.ii as: as: _ommw: he casesuue— one some: sceaauue> use use .o_ataa=oaoo< «6 Esmeoa m cm mo~=oc~oa has one .maz .mu v Oman? 13 Figure 5. Structure of N-methylformamide (A) monomer CH 0 H / c —— N \ c — N < H H H / CH- 0 trans cis (B) dimer ?”'I fl\.—c I 3c=° Cso-m-I—l’c" ”'5’ \o — u-l\ ' \c-o I" \I c'" '- ‘Crl l/ u/ \I I “I ' ' .... a \ n-emo u\ /c-o u—c, \\c_' eae—e—u I l-tt/ \c—o-a P.—.\ \ "a r“ 4"?” ' "ii I I a n ,- ‘.-./ ‘4. .-_c’ \om-a—n/ I" h \c-o I/ (C) polymer Linear Flexible Structure in Liquid Phase H ._.. -l:H3 { N< omwee>< owm>< amuse: on“: sewas_om .omuah eoqaauucoonoo Ld—OE ocaad \ u—aw one neuo:oc l a \ amuseIOMw >< \ meomaahanom omoouaomwsaiOHQOuaomm use use aeo>~ow been osu cu twaanou mauusm hoeochLu ose ”ml: Co use: newcoaotam :2 one .0 define: heecscoa: cease: and 0mm: case: n O 1.“... 18 780+ . L J A 1 l l 1 1 1 O 0.2 0.4 0.6 0.8 1 XDMF Figure 7(a) . Apparent molar volume: of DMF. WMF). in mixtures with DMSO. AN, NMF. water. end methanol nt 25°C. Y I 1 v fl 1 T 1 Y 72.0 " 7104: A=DMSO ... ’A :i :: a l- 3 59.5 E P‘E 58-5‘D A=NMF a u is :E ( 530T 9’ 52-0 _ ' A=AN 1301 16.0 Figure 7 (b) . Apparent molar volumes of A. MA). (A -= DMSO, AN. NMF. and water), in mixtures with DMF at 25°C. Fro. reference 38. 19 3. e 'a Solvat o ' B n 8 ve ' u In a mixed solvent solution that contains an electrolyte, a special case of ion solvation arises when the solvent composition of the inner solvation shell of the ion is different from that of the bulk solution. If the compOsition of the inner solvation shell of a solute has a higher proportion of one solvent than does the bulk solution, it is said to be preferentially solvated by that t39. Homoselective solvation indicates that both componen cation and anion of the electrolyte are preferentially solvated by the same solvent component; on the other hand, heteroselective solvation indicates that the cation is preferentially solvated by one component and anion is preferentially’ solvated by' the other29. An. example of homoselective solvation is calcium chloride in a water- methanol mixture, where both Ca2+ and Cl- ions are preferentially solvated by water. Heterselective solvation is observed with silver bromate in a water-acetonitrile mixture, where Ag+ ions prefer acetonitrile whereas Br03' is preferentially hydrated4o'41. There are many methods used to study preferential solvation, physicochemical methods being the most popular. Janardhanan and Kalidas reviewed electrochemical methods29 including (a) the use of Gibbs free energies of transfer of the ions in different solvents, (b) solvent transport numbers, and (c) the Walden product variation of the ions 20 with solvent composition in the mixed solvents of interest. By using free energies of ion transfer measurements, Cox et _;.42 studied Ag+, Na+ and Li... in AN + H20 mixtures. It was found that the overall free energies of transfer and their dependence on solvent composition can be explained simply in terms of the stronger interactions between the electrolyte ions and water. An exception is the silver ion, which prefers AN (Fig. 8). This method has been extended to the study of other mixtures43'44. Other physicochemical methods to study the preferential solvating ability of a solvent in a mixture are viscosity and electrical conductance measurements. These two physical properties vary with the composition of the solution and are affected by solute participation. Gill and coworkers studied the behavior of electrolytes in many binary solvent mixtures using various physicochemical techniques, from which the same 'preferential solvation results were obtained45'50. They evaluated the actual solvated radii of ions in mixed solvents by the measurement of limiting conductance. Their results also showed evidence of the interactions describable by hard-soft chemistry; pyridine (a soft base) formed strong coordinate bonds, particularly with Ag+, Cu+ and Cu2+ ions (soft acids). Walden studied ion- solvent interactions of alkali halides in water-amide mixtures using viscosity data51. His results indicated that the ions were preferentially solvated by water in the water- 21 '16 4 7 1>NaCH 12* f O C‘- T J ~ I o 5. " E 0’! 3 /, ...... g5 /. ""’ 1“ Na.- < 0 ..-=..-’.-’§.':§:Z::: ..... "/ ‘ *‘§-‘L~.‘~+—'——+/+ -4» ................ Ag‘ 9503mm Figure 8. Free energies of transfer of NaCl and AgCl from water to AN-HZO mixtures at 25°C. From reference 42. Absorption 132 \ _ ~__---’ I I l 1050 1100 "50 \‘. cm“ Fig.7 . v(Cl—O) absorption bands of C10.“ anion in solutions of lithium perchlorate in AN (2), in NM (7), and in NM/AN bi- nary solvent (3-6). Salt concentration, M: 2-7) 0.5; 8,9) 2.0. Ratio of concentrations of molecules of solvents: 3) 1:1; 6) 3:1; 5) 10:1; 6) 60:1; 8-10) 2:1. Solution tempera- ture, ’K: 1-7, 9) 293; 8) 243; 10) 353. Dashed curve (1) is for absorption by NM/AN solvent with 1:2 mole ratio. From reference 54. 23 rich region, and that the influence of various amides upon solvating the ions are similar. Spectroscopic methods including IR, Raman, UV-VIS, and NMR measurements provide a detailed probe on a molecular level to investigate ion solvation in mixed solvents. These techniques give information about the inner solvation shell of the ion studied. Emons and coworkers studied Al3+ solvation in MeOH-HZO mixtures using Raman spectroscopy. The changes of the MeOH Raman band and the isotopic shifts of the -OD symmetric stretching vibrations of the hexasolvated complexes52 indicated that Al3+ is prefentially solvated by water. Akopyan used IR spectroscopy to study Li+ in AN-AC mixture553. He used the CaN stretch vibration of AN and the asymmetric stretching vibrations of the C-C-C bonds of AC to show that the solvating power of AC toward Li+ is stronger than that of AN. LiClO4 was also studied in NM-AN mixtures by Perelygin using IR spectroscopy54 (Fig. 9). Nitromethane and acetonitrile have similar dielectric constants, but have considerable differences in electron donor power. Several possible types of solvated complexes at different solvent compositions distinguished, according to the wavenumber of IR vibration bands, were proposed to explain the solvating properties of this mixed solvent system. NMR spectroscopy, through the measurment chemical shifts and relaxation rates of magnetic nuclei in mixed solvents, is widely used in this area. The isosolvation 24 point has been defined as the chemical shift corresponding to the midpoint between the respective chemical shifts in the two pure solvents which comprise a given binary solvent mixture. The curvature of a plot of the chemical shift of a nucleus as a function of solvent composition.*was first interpreted in terms of preferential solvation by Frankel g; 11.5556. These workers used NMR spectroscopy to study Co(III) in a mixture of chloroform with benzene or carbon tetrachloride. The results showed that Co(III) was preferentially solvated by chloroform in all the mixtures studied. The isosolvation point was calculated from both linewidth and chemical shift variations of the Co(III) nucleus. The authors also measured the 35Cl chemical shift of NaCl, LiCl and (c2115) 4Nc1 in DMSO-H20 and CH3CN-H20 systems, and found. that the Cl- ion ‘was preferentially solvated by H20 in the CH3CN-H20 mixture, but that there was a nearly even competition for solvation of C1" in the DMSO- H20 mixture57 (Fig. 10). Covington and Thain studied the same mixtures, but with different nuclei58. They concluded that Cs+, N03" and 1' ions were preferentially solvated by' DMSO and that Li+ showed slight preference for H20 (Fig.11, Tables 6.7). In AN-Hzo mixtures, Na+, Cl’, Br'and I" ions are preferentially hydrated. The results of investigations of HZOZ'HZO mixtures indicated that Rb+, Cs"' and F' ions are preferentially solvated by H202 and the Li+ ion by water41. These conclusions were based on the assumption that the 25 -50 O Chemical shill, PPM from 1M aqueous NoCI. O 0.5 1.0 Volume fraction organic solvent or P/P0(H20). Figure 10. The variation of "Cl chemical shifts of chloride salts in mixtures of CH3CN and DMSO with water. Open circles represent 0.25 M LiCl in DMSO—water. Closed cirlces represent 0.25 M (QH;).NCI in DMSO—water. Closed triangles represent 0.25 M (QH;).NC| in CH3CN-water. All of these curves are chemical shift as a function of volume fraction of the organic component. The last curve (open squares) represents 0.25 M (Csz).NCl in CH.CN—water, this time showing chemical shift as a function of the relative partial pressure of water (approximately relative activity) compared to pure water (P/Po). From reference 57. 26 .wm OUCQNmHmH EOHnH .2 n a». 3 55. 59.2983 2: m_ .2. can .0 ...... a». Ewen—933 on. 82.. ex 80:02.. 308 ..a 8258 .538 .8 Ewen—263 838288 33a .8 :2... on. m a 8053 £14 .3 coon—nos a._ At .32 fl 4 £9. u .5! 5:3 AC :3 89.. pea—3.8 ..l.l ” . 2.38 .ScoEtoaxm .0 ..Ha shaman :3 E .53 $2258 Cum +0220 5 p83 8: 8n .mOZ 8.. 838388 339 5.: E omega .mcocufim A8293. o.— 90 0.0 v.0 «.0 o a _ _ 2 2 _ _ 2 A \ \ [I \ ll \ \ l \ l \ \ \ II \ l. \ \ \ \ x N l \\ no. mot \ \ In \ IL \ \ \ II \ ll \ \ \ I. \ l \ \ \ l .I. \ \ \ _ _ _ _ _ _ _ p _ o— 27 Table 6. Preferential salvation for cations Isosolvation Point Slovent Mixture Cation or prefered solvent Ref. (1) DMSO NM Na+ 0.05 M.F. DMso 67 (2) * pr NM Na+ 0.12 M.F. FY 67 (3) TMU NM Na+ 0.06 M.F. TMU 67 (4) HMPA NM Na+ 0.05 M.F. HMPA 67 (5) AN NM Na+ 0.15 M.F. AN 67 (6) PY AN Na+ 0.29 M.F. FY 67 (7) DMso AN Na+ 0.10 M.F. DMSO 67 (8) TMU AN Na+ 0.11 M.F. TMU 67 (9) HMPA AN Na+ 0.06 M.F. HMPA 67 (10) TMU HMPA Na+ 0.23 M.F. HMPA 67 (11) DMso HMPA Na+ 0.15 M.F. HMPA 67 (12) PY HMPA Na+ 0.10 M.F. HMPA 67 (13) FY TMU Na+ 0.16 M.F. TMU 67 (14) DMSO TMU Na+ 0.39 M.F. DMSO 67 (15) DMso PY Na+ 0.10 M.F. DMso 67 (16) AN H20 Na+ H20 58 (17) *EN H20 Na+ EN 58 (18) DMF H20 Na+ DMF 63 (19) HMPT H20 Na+ HMPT 65 (20) AN H20 Na+ H20 65 (21) MeOH H20 Na+ H20 62 (22) 3202 H20 Cs+ H202 41 (23) DMSO H20 05* DMSO 58 (24) su H20 05* su 58 (25) *NEF H o T1+ 0.44 M.F. NEF 71 2 + (26) NMF H20 Tl 0.46 M.F. NMF 71 (27) FA H o T1+ 0.39 M.F. FA 71 2 + (28) NEF NMF Tl 0.40 M.F. NEF 71 (29) NEF FA T1+ 0.42 M.F. NEF 71 (30) NMF FA T1+ 0.50 M.F. NMF 71 (31) MeOH H20 A13+ H20 52 (32) Ac AN Li+ Ac 53 (33) NM AN Li+ AN 54 (34) H202 H20 Li+ H202 41 (35) MeOH H20 Rb+ H20 62 (36) H202 HZO Rb+ H202 41 (37) NMF H20 ouz+ NMF 44 * TMU, EN, 50, NEF are abbriviations for tetramethylurea, ethylene, N-ethylformamide, respectively. sulpholan and Table 7 Solvent Mixture DMso H20 AN MeOH AN H20 MeOH H20 DMso H20 H202 H20 MeOH H20 AN H20 AN H20 DMSO H20 Preferential salvation for anions N03 N03 01' c1' C1 F Anion Preferentially Solvated Ref. by DMSO 58 MeOH 58 H20 58 MeOH 62 H20 57 H202 41 MeOH 62 H20 58 H20 58 DMSO 58 29 salvation changes can be represented by a series of n competitive equilibria (n is the salvation number), assumed to be the same for pure water and hydrogen peroxide. The variations of the free energy of transfer of a salt from water to mixed aqueous solvents derived from e.m.f. measurements has been applied to alkali metal salts in MeOH- H20 mixtures, and the results were in agreement with spectroscopic measurementssg. Nuclear magnetic relaxation techniques were used to study preferential salvation of ions in solvent mixtures. There are three different techniques employed: (i) ion nucleus-solvent nucleus intermolecular dipole-dipole relaxatianéo, (ii) the magnetic relaxation by quadrupolar interaction of the ionic nuclei of the solute with the solvent, and (iii) isotope effect on the magnetic relaxation by quadrupolar interaction of the ionic nuclei of the solute with the solvent. H012 and coworkers applied the last two techniques to the study of some alkali halide salts in MeOH- H20 mixtures61 ' 62 (Fig. 12). By extrapolating the relaxation rates of 23Na, 87Rb, 35Cl, and 81Br to infinite dilution over the whale mixture range, they calculated that the relative solvating ability of MeOH is stronger toward Cl’, Br', but weaker toward Na+, Rb+. These techniques have been extened to the study of many other salts in different systemss3’55. In this laboratory, ion salvation properties have been studied by the combination of IR, Raman and NMR 30 .No mocanauau fioum .388 03 £83 208 com 82828 8:58.38 .80: so. as: . an: ..o 50323 “020358 2: a. mvcaamotao can macucaaEaa .3: on. ..o 838, ...»: . as: 03. on. 300580 on: 828:. 2C. .883 egg—08$ Em— .833 was: 23.5 823 up: . 3h: u x .GV .83 a. 86388 88.. 5:938 .. 80:202.: .. Hoe: pennan— AOV 8538 O~I+IOoZ E 3:888:00 =8 93 a. pose—anabxu 8.2 802838 0:2.me «Z: .. NH auamfim moo: x 68 8. 8 8 o... 8 ‘11. OM A o o . o 0 l a 1% O— N. H\ mm: L; i n o? I. ah x m? w om. 2 yr ... ) 1 m NT S. 31 techniquesa6'68. The isosolvation points of many solvent mixtures toward 23Na were obtained by NMR chemical shift measurements (Table 6). The relative solvating ability (or the relative donor ability) toward 23Na is HMPA 2 DMSO =TMU > PY > AN >> NM. An empirical scale of solvent donor ability is based on the enthalpy of the complexation reaction between SbCl5 and a given solvent in dilute 1,2- dichlaroethane solutian69: $15015 + 5 se- 3 - 316015 The donor number (or donicity) of the solvent S is equal to the ne tive value of nth : = - in g e alpy DN (3) AHS . SbCls kcal'mol'l. It has been shown that there is a linear relationship between solvent donor number and the relative chemical shift of 23Na at infinite dilution in these solutions70 (Fig. 13). 4. Pzefetential Salvation in Binaty Solvent uixtureg Containing N-Methylfarmamide H012 and cowokers studied preferential salvation of the Na+ ion in NMF-H20 mixtures by measuring the 23Na nuclear magnetic relaxation rates, extrapolated to zero salt cancentratian63'64. The results showed no preferential hydration of Na+ in the composition range 1 > x > 0.5, where X was the mole fraction of water, and weak preferential hydration in the remaining region (Fig. 14). The relaxation rate of Dr“ reflected the large field gradients at the anion 16-1 12* 32 ()._ 13. ‘4 l I I 1 0 1O 20 30 40 DONOR NUMBER Figure 13. Plot of Infinite Dilution Sodium-23 Chemical Shifts versus the Donor Number of the Solvent: (1) Nitromethane, (2) Acetonitrile, (3) Sulfolane, (4) Propylene Carbonate, (5) Acetone, (6) Ethylacetate, (7) Tetrahydrofuran, (8) Dimethylformamide, (9) Tetramethylurea, (10) Dimethyl- sulfoxide, (11) Water, (12) Pyridine, (13) Hexamethyl- phosphoramide, (l4) Formamide, (15) Methanol and (16) Ethanol (Greenberg,M.S., Ph.D Thesis, 1975) 33 :L '2 s" (7' 23Na. 0 00’ ' 0 “2° . NMF my mL - 1» 60' I ‘0” . H{)°fiunnmflde m . I .0 I: it. . 6‘ W— ' 6 {f Figure 14- ”Na“ magnetic relaxation rate: extrap- olated to zero salt concentration in form- amide-H30 and in NMF-H20 mixtures as a function of 1:; (mole fraction of H30 in the solvent). Experimental values are given as filled symbols: I, formamide + H30; 0, NMF + H30. The corresponding theoretical values according to Eq. (3a) are given as open squares and open circles. From reference 63. 34 nucleus. Gibbs free energies for Cu(II) ion transfer from H20 to NMF-H20 mixture were calculated from the emf of the cell“: Cu / Cu2+ (H20 )// Cu2+ (H20 + NMF) / Cu The results indicate that NMF had stronger solvating ability toward Cu(II) than water. Briggs and Hinton measured the 205T1 chemical shifts in binary’ solvent. mixtures of NMF-FA, NMF-NEF and. NMF-H20. These showed that the solvating abilities toward 205T1 were NMF = FA, NMF < NEF, NMF > H20. The isosolvation points of these systems were obtained from experimental data and theoretical calculations71 (Table 6). Kalidas and Schneider determined the solvent transference number of AgIO3 and Agzso4 using emf measurements in NMF-MeOH mixtures43. The transference number of AgIO3 showed two maximal values throughout the whole mole fraction range. These workers explained the phenomenon as heteroselective solvation(ref. page 19) of the salt with low mole fraction of NMF and homoselective solvation(ref. page 19) by NMF at other compositions. The transfer energy of the Ag+ ion, however, showed that it was preferentially solvated by NMF. 35 The studies of NMF-containing binary solvent mixtures, as reviewed above, were only on a few systems. Information regarding this solvent in mixtures is deficient. In this work, twelve NMF containing-binary solvent mixture systems have been studied. CHAPTER II EXPERIMENTAL PART A. MATERIAL PURIFICATION 1. Reagents Sodium perchlorate and cesium perchlorate were of reagent grade quality (Alfa) and were used without further purification except for drying at 120°C for two days. Sodium thiocyanate, cesium thiocyanate (Mallinckrodt) and potassium thiocyanate (MCB) were recrystallized from methanol and dried under vacuum at 70°C for three days before use. Sodium pictrate was prepared by neutralizing picric acid with sodium hydroxide in ethanol solution, then it was recrystallized twice from absolute ethanol (7 g sodium picrate in 100 ml ethanol), and once from water (4 g sodium picrate in 100 ml water). The product was dried at 150°C for five days. Potassium picrate was prepared by neutralizing picric acid with potassium hydroxide in ethanol solution then it was recrystallized three times from water (10 g potassium picrate in 100 ml water). The product was dried at 120°C for 24 hours. Lithium picrate was prepared from lithium carbonate and picric acid in acetone solution then it was recrystallized four times from an acetone- benzene mixture and dried at 150°C for five days. Tetrabutylammonium picrate (m.p. 91.6-91.9°C) was prepared from the aqueous solution of tetrabutylammonium iodide with sodium picrate, then it was recrystallized from a water- ethylacetate mixture and dried under vacuum at 50°C for three days. Sodium tetraphenylborate (Aldrich, Gold Label) was dried under vacuum for three days at room temperature. 36 37 Sodium iodide (Alfa) and sodium chloride (E.M.) were dried at 120°C for three days. Cryptand-zzz (E.M.) was used as received. 2. Solvents N-methylformamide (NMF, Aldrich) was dried over freshly activated Linde 3 A molecular sieves for two days then distilled twice over granular barium oxide at 55°C under reduced pressure. Acetone (AC, J .T. Baker) was fractionally distilled over calcium sulfate (Drierite). Formamide (FA, Fischer) was purified by fractional distillation over barium oxide. N,N-dimethylformamide (DMF, J .T. Baker) was dried over Linde 3 A molecular sieves, followed by vacuum distillation over calcium hydride. Dimethylsulfoxide (DMSO, E.M.) was refluxed over barium oxide, followed by fractional distillation under reduced pressure. Pyridine (Py, E.M.) was refluxed over barium oxide and fractionally distilled. Nitromethane (NM, Aldrich) was refluxed over calcium hydride for two days then fractionally distilled under reduced pressure. Tetrahydrofuran (THF, E.M.) was refluxed over a mixture of potassium metal and benzophenone for one day and fractionally distilled. Methanol (MeOH, J.T. Baker) was refluxed over magnesium turnings and iodine for three days and then fractionally distilled. Acetonitrile (AN, J.T. Baker) was refluxed over calcium hydride for two days, followed by fractional distillation. Hexamethylphos- 38 phoramide (HMPA, Aldrich) was refluxed over barium oxide for one day then fractionally distilled under reduced pressure. Ethanol (EtOH, J.T. Baker) was refluxed over magnesium turnings for two days then fractionally distilled. All purified solvents were stored over Linde 3 A molecular sieves in brown bottles sealed with teflon tape and placed in a dry box under helium atmosphere. The content of water was measured by NMR and IR spectroscopy, and no on band was observed. 3. SAMPLE PREPARATION The binary solvent mixtures were prepared by adding the desired volume of NMF (using a 100-1000 #1 Eppendorf digital micropipette) to a 2 ml or 5 m1 weighed flask, reweighing the flask, then adding the second solvent of interest and weighing it again. From these weights and volumes the mole fraction of each solvent in the solution and the volume susceptibility of the solution can be calculated. To prepare salt solutions, a weighed amount of the salt of interest was added to a weighed 2 ml flask which was then filled to the mark with the desired solvent or solvent mixture. The solutions used for formation constant measurements of the Na+-C222 complex were prepared by placing the desired amount of cryptand-222 (C222) into a 2 m1 flask, which was then filled to the mark with 0.05 M sodium picrate-NMF 39 solution. The ratio of the concentrations of C222 and sodium ion are calculated. Because of the high absorbance of picrate salts in the UV-VIS region, the concentration of picrate solutions were chosen to be in 0.1-1.0 mM range. C. SPECTROSCOPIC MEASUREMENTS 1. IE Infrared spectra were measured with a Bomen DA3.01 FTIR spectrometer. A standard demountable liquid cell with ZnSe or CaFZ windows and 0.05 mm spacers were used. Some samples had to be measured without the spacer, but only with a drop of the solution pressed between windows. Measurements were carried out in the 800 cm'1~2500 cm"1 1 resolution. spectral range under 2 cm' 2. Karen Raman spectra were collected with a Jarrell-Ash model 25/100 spectrometer following excitation with a Spectra Physics model 165 Kr+ or' model 164 Ar+ laser. Solution samples were placed in a Pyrex cell. Frequent calibrations were made by using the 459 cm"1 peak of carbon tetrachloride (J.T. Baker, "photrex"). The width at half- height of the Rayleigh peak (with slits 10/20/10 pm) was 3 cm'l. The scan rate was 0.5 cm'l/sec. PMT voltages up to 1300 volts were utilized, and the slits were always set to 100/200/100 pm when sample solutions were measured. The 4O resolution depends on the slit widths and the wavelength of excitation line (normally 2-4 cm"1 resolution). The excitation lines used were 4880 A and 5145 A for the Ar+ laser, and 6471 A for the Kr+ laser. 3. UV-VIS UV-VIS spectra were obtained with a Cary 17 spectrophotometer in the 350-560 nm range. The samples were placed in a 1 mm pathlength quartz cell and measurements were made at room temperature. Because of potential photodecomposition of the picrate salts, the UV-VIS measurements were carried out within a few minutes after the samples were prepared. 4. Egg Sodium-23, cesium-133 and chlorine-35 NMR measurements were made on a Bruker WH—180 multinuclear NMR spectrometer at 47.62 MHz, 23.61 MHz, 17.64 MHz respectivly , in the pulsed Fourier transform mode. All the samples measured were placed in 10 mm"5 NMR tubes (Wilmad). A 4 mm¢ coaxial insert (Wilmad), containing lock solvent solution, was used as external reference. The reference solutions were 0.5 M NaCl in D20 for sodium-23 and chlorine-35, and 0.5 M CsBr in 020 for cesium-133 measurements. All the chemical shift data were corrected for the differences in the bulk diamagnetic susceptibility between 41 the sample and the reference according to the equations of Martin g; gl. 5corr = sobs + 4“/3 (Xref ' Xsample) where Xref and Xsample are the unitless volume susceptibilities of the reference and the sample respectively; scorr and °obs are the corrected and observed chemical shifts respectively. The susceptibilities of salts are negligible. In the studies of binary solvent mixtures, the volume susceptibility of a given mixture is calculated by using the Wiedmann’s law. V V calc = A» .X(A) + 3 VA + VB VA + VB (B) where chalc is the calculated volumetric susceptibility of the solution. In this equation x(A) and X(B) are the volume susceptibilities of pure solvent A and B, and VA, VB are their respective volumes. Chemical shifts were measured is. infinitely dilute aqueous solutionsof the respectived sales. The widths of resonance lines were obtained by the Gaussian-Lorentzian fit provided by computer program NTCCAP. CHAPTER III RESULTS AND DISCUSSION A. Ion Association Studies in Pure N-methylformamide 1- 212m Picrate salts (such as alkali picrates and alkylammonium picrate) have been studied in 2-butanone by UV-VIS spectroscopy7. The variations in the band shape and wavelength of the picrate anion have been interpreted in terms of ion pair formation; it has been suggested that ion pair formation changes the energy of the electronic excited state and affects the absorption frequency. Since there are some salts which show incomplete dissociation in NMF, the study of ion association properties of picrate salts in NMF is of great interest. Sodium picrate (NaPi) and potassium picrate (KPi) were first chosen to carry out this study. The UV-VIS spectra gave interesting results. All the absorption bands had a maximum at 365 nm with a shoulder at ~420 nm and the absorbances followed Beer’s Law; that is, the absorbance was proportional to the salt concentration. However the intensities of NaPi and KPi bands were different at the same salt concentration (Fig.15). These results seem to indicate some kind of interaction between the cation and the anion. In order to prevent cation-anion interactions, 18-crown-6 (18C6) and cryptand-222 (C222) were added separately (with mole ratio of Na+:c222 (or 18C6) = 1:2) to complex with the cation. To our surprise, there were no changes in the band shape and/or the wavelength of absorption. The electronic spectrum of pure picric acid solution indicated that the 42 43 502 ED KI Ev. .122 E wrom 382d *0 mocontowbo m_>l>3 so mocmccmamo cozobcmocoo .m_ 930E AEEV mtom 3803 cc cozobcmocoo o; . m6 md to «.0 0.0 _ r _ . p 3 _ . _ t 0.0 C .. \ C \ a \ led \ W 1.0.0 S 0 J . Q 0 ..N; w a 1m.— 44 Table 8. Concentration dependence of UV-VIS absorbance (365 nm) of picrate salts W Concentration LiPi HPi 0.103 (mM) 0.169 t 0.003 0.175 :t 0.003 0.206 0.332 0.344 0.309 0.499 0.525 0.412 0.664 0.700 0.515 0.823 0.878 0.618 1.001 1.060 0.721 1.168 1.232 0.824 1.332 1.417 0.927 1.495 1.557 linear coeff. 0.99995 0.9998 slope 1.62 1.68 Alb—822m Concentration NaPi KPi 0.100(mM) 0.156 1 0.003 0.194 i 0.003 0.200 0.269 0.385 0.300 0.384 0.576 0.500 0.699 0.907 0.700 0.925 1.315 0.900 1.194 1.714 1.000 1.280 1.892 linear coeff. 0.9994 0.9996 slope 1.285 1.889 45 absorption band is due to the picrate anion in NMF solution. These observations indicate that there were only anion- solvent interactions, but then the difference in the absorption intensities between KPi and NaPi is difficult to explain. Many possible causes of this difference have been considered. Are impurities contained in these salts? If the weighed solute contains some water or non-picrate salts, the intensity of the picrate absorption would decrease; on the other hand, if picric acid or picrate salts of lighter metal cations contaminate the sample, the absorbance would increase. The two salts were analyzed for the alkali metal content by metal ion plasma spectrascopy. They were found to be 2 98% pure, and the 5 2% impurities cannot induce such a large variation in the UV-VIS absorption intensities. Proton NMR is another tool often utilized to check sample impurities, especially for organic contaminants and water. Unfortunately, we cannot use the proton NMR results to check the impurities of other picrate salts contained, because the two protons on the picrate anion molecule are strongly shielded by the neighboring nitro groups on the benzene ring (Fig. 16) but from the spectrum we know that there were no other organic compounds or water contained. The picrate salts (KPi and NaPi) and picric acid have the same in resonance lines at 7.45 ppm in pure NMF, which are the chemical shifts of the protons of the picrate anion. 46 Table 9. UV-VIS peak absorbance variation of picrate salts in NMF solution Peak absorbance after Picrate salts 0 hr 36 hrs 52 hrs 70 hrs HPi 1.685 1.673 1.641 1.623 NaPi 1.397 1.376 1.354 1.332 KPi 1.892 1.887 1.879 1.858 * the uncertainty of peak intensity is i 0.003 9N N01 Mk Figure 16. Stucture of picrate anion 47 A second possibility is the decomposition of picrate salts. According to the UV photolysis study of picric acid in aqueous solution by Mitteilung72, picric acid was entirely decomposed under intense UV light (380 V, 3.8 A) in 180 min. To test this point, 1 mM KPi, NaPi and HPi in NMF solution were placed under normal light in the laboratory, and the intensity of their UV-VIS absorption was measured at 0, 36, 52 and 70 hrs after preparation. From the results listed in Table 9, there was very little intensity reduction with increasing time. This observation indicates that the decomposition may have occurred during synthesis and/or storage. Actually, freshly synthesized NaPi and LiPi showed intensities (at 365 nm) of 1.63 and 1.62 in 1 mM solutions of NMF respectively, which were higher than stocked NaPi (1.3), but still lower than stocked KPi (1.89). It seems that KPi is more stable than NaPi. Although there was decomposition of the Pi' ion, according to the above study, we may still conclude that: (i) the picrate salts are completely dissociated in NMF, and the position, intensity and shape of the UV-VIS absorption band is controlled by picrate anion-solvent interactions, which are independent of the cation; (ii) The decomposition of picrate ion may not change the metal ion concentration in solution, but changes the concentration of the picrate anion; (iii) the {dimerization. of jpicrate anion (Pizz'), which has been studied in aqueous solution73, does not occur in NMF solutions. The absorbance gs. picrate concentration 48 obeyed Beer’s Law, which indicates that only Pi'1 is present in NMF solution. 2. §a*.§;;; Formation Constant In the above study, cryptand (C222) was used to complex the Na+ ion of the NaPi. The formation constant of the Na+C222 complex can provide useful information about how well the Na+ ion is encaged in NMF. NMR chemical shift techniques were utilized in this study. The chemical shift of 23Na was measured as a function of the concentration ratio [C222]/[Na+]. The results show that, at room temperature, there is slow exchange between free and complexed Na+. As shown in Fig.17, the free 23Na resonance line does not disappear until [C222]/[Na+] = 1.27, which indicates that C222 is not a strong complexing ligand toward the Na+ ion in NaPi when NMF is the solvent. The formation constant, which was calculated from the area of the 23Na resonance lines, is 177 z 10 (log Kf = 2.25 i 0.02). Na+ + c222 <— Na+'C222 Kf = [Na+'C222] / [Na+][C222] It is surprising to note that the formation constant of the Na+C222 complex in other solvents is much higher than in NMF (Table 10). Soong measured in NMF solution the Na+'C222 complexation constant using NaTPB with C222 , and obtained a formation constant of 152 : 8. Our data indicate that the weak complexation of Na+'C222 is caused by the solvent, NMF. 49 (.222 ’Nd - 1274 Mr - v fi fi c222 IN6'0952 :zzzme': 0722 \ fi~ M“. . NW Figure 17. 23Na Resonance Lines of Na+.0222 Complexation in NMF. 50 Table 10. Formation constant of Na+ c222 complex Solvent log Rf N-Methylformamide 2.25 t 0.02 Water 3.98 Methanol 7.9 Ethanol 8.5 Acetonitrile 9.63 Propylene carbonate 10.5 N-Methylpropionamide 5.8 N,N-Dimethylformamide 6.1 Dimethylsofoxide 5.3 Except for NMF, 911 data are taken from Cox, B.G. g; a1. , J.Am.Chem.Soc., 1981, 193, 1384 51 The formation constant of the Na+'C222 complexs in DMF solution was measured by NMR spectroscopy. The exchange rate of 23Na nuclei between free and complexed sites is slow at room 'temperature, and two resonance lines were observed until the ratio of [C222]/[Na+] was 0.95 (Fig. 18). These results indicate that in DMF the Na+'C222 complex is quite stable. Because of instrumental limitations, the formation constant cannot be calculated from the area of the rasonace signals. The above results indicate, however, that the complex formation is much stronger in DMF than in NMF. This difference probably results from a stronger solvation of the ligand by NMF than by DMF. If C222 is solvated by solvent molecules, the actual "free" concentration which remains to complex with Na+ ‘would be reduced, and the apparent formation. constant of Na+C222 in this solvent. would be small. B. Ion Solvation Studies in Binary Solvent Mixtures 1. Eng Studies of Preferential Solvation (1) Isosolvation point measurements This technique, which is used to determine preferential solvation of certain ions in binary solvent mixtures, is based on two assumptions: (i) the solvation number of the ion under investigation is constant at all compositions of a given binary solvent system. (ii) there are no ion pairs formed over the whole composition range studied. The chemical shift of a metal cation reflects the 52 No. 5 C 222 NI= 0,7 N0 4 63%." = 0508 AAA "-—v_i ‘ ‘ ‘jLw—w YV’ ‘fi?;' 1 ~V:‘V":;VAV Y‘ _"‘_‘-A—‘ A N02 C222/Iv“f : 02 pure DMF L 1 l L l 1 L 1 1 1 l 1 l 1 L 1 l 40 30 20 :0 0 -10 -20 ~30 ~40 ow“ Figure 18. 23Na Resonance Lines of Na+.C222 Complexation in DMF. 53 electron density in its immediate environment, and is related to the electron donicity of the solvent. When a cation is solvated by pure solvent A it shows a chemical shift 5A; the addition of a second solvent B usually changes the chemical shift. It has been observed that if solvent B has higher electron donicity than A, the resonance line of the nucleus shifts to a lower field. 0n the other hand, if solvent B has lower electron donicity than A, a diamagnetic chemical shift would be observed. The plot of chemical shift as a function of solvent mole fraction should be a straight line when the solvents A and B have the same solvating ability toward the cation. The isosolvation point of this system would occure at equal mole fractions, XA = 0.5 = XB. Usually, the two solvents in a mixture have different solvating abilities; therefore, the plot is a curve as shown in Fig 19. a. Preferential Solvation Towards the Na+ Ions Sodium-23 chemical shifts of 0.1 M NaClO4 were measured as a function of solvent composition in twelve binary solvent mixtures composed of NMF with HMPA, PY, H20, EtOH, DMSO, DMF, MeOH, FA, THF, AC, AN, and NM. The results are shown in Table 11 and Fig 20-22. The isosolvation point is given as the mole fraction of NMF in a mixture (Table 12). There is a relationship between the solvent donicity and isosolvation point. The higher the donicity of the solvent the better is its solvating ability toward Na+ ion. No exceptions are found in the results, namely, PY and Chemical shift 54 It‘s— .- I I kw \ Chemical shift I I | I ' I ' I | I I I I I I I I I I I I I I I l 1 B A B Solvent composition Solvent composition A==B AI>B Chemical shift Solvent composition A‘0 luzz Omzoluzz I‘M. 32023.. 005:: E :Em 62:35 nNIEEUom .5 8.6: .122 00 cozoot 0.92 0., m6 who 30 N0 0.0 (wdd) mus lovlwauo cz—wnipos 57 quluIZ Z> H20 > DMSO > DMF % NMF ~ FA ~ MeOH > EtOH > PY >> THF > AC ~ AN >> NM. 66 b. Preferential solvation toward Cs+ ions Five binary solvent mixtures (NMF-HMPA, NMF— PY, NMF-DMSO, NMF-DMF, and NMF-FA) were used in this study, the concentration of the Cs+ was 0.1 M (CsClO4). The small solubility of CsC104 in other solvents limited our studies. The results (Fig.25) show that we can rate the solvating ability toward Cs+ as, DMF > HMPA > DMSO == NMF a FA >> PY Several interesting points arise from these results. The first is the slightly preferential solvation of HMPA toward the Cs+ ion which is very different from that toward the Na+ ion. Both Cs+ and Na+ are univalent alkali cations, the Na+ ion has a higher charge density. We can expect that the interaction of the P=O group of HMPA to be stronger with Na+ than with Cs+. Moreover, in this case the bulky structure of HMPA seems to play an important role, and could prevent it from substituting NMF molecules in the inner solvation shell of the Cs+ ion. Another surprise is that DMF shows much better solvating ability toward Cs+ than toward the Na+ ion. The possible explanation is that DMF is not an associated solvent, therefore, it can solvate cations more easily than NMF. 67 300200 0055 E are 62820 9.78380 .3 839.1 .122 to 8:8: .....02 0.. 0.0 who .30 N0 0.0 Fl. . . p . . . I >mlu22 Clo 0— N— c + H20 / \ / \ H CH3 H H3C c - CN 85 Figure 30. H-bonding between solvents M H H ...... >c_./ 0/ \CH3 mu: - Ac H\ H ...... /C——N/ 0/ \CH3 NM - THF H ...... \ /H /C._N\ 0/ CH3 mg}: - AN 5...../ 0/ CH3 mg: - PY H C--N H ...... 86 /N (C ”5.):— N (6'13);- N (CH2); 0 O O I ECMHB 87 A similar situation occures for the NMF—NM mixture, where after 30 hours storage the solution becomes yellow. A possible reaction is: O H H H H \ / I / C-—-N + H-—C -—N02 ---> N——-C + H20 / \ | \ H These predictions are reasonable, because both the -CsN and -N02 frequencies are strong electron-withdrawing groups, so that their neighboring -CH3 group is reactive. Fortunately, the reaction rates are slow, so the effects on the preferential solvation measurements describe in this study are minimal. (2) Linewidth measurements The linewidth of a quadrupolar nucleus (I > 1/2) is proportional to the spin-spin relaxation rate (1/T2), which is equal to the spin-lattice relaxation rate (l/Tl) in the "extreme narrowing" region (dilute solution, rapid molecular motion). Thus, it can be represented by the equation, 1 1 3x2 21 + 3 (equ )2 s2 l 2 ( ) m c / T2 T1 10 12(21-1) h 3 88 4xna3 To =‘--- (This is Debye’s expression which has been 3kT successfully applied to many systems) where Wl/Z is the linewidth at half height, I is the spin of the nucleus, q is the electric field gradient at the nucleus, Q is its quadrupole moment, s is the asymmetry parameter rco is the Sternheimer antishielding factor, n is the viscosity of the solution and r is the correlation time. (Rotational correlation time is the time for molecular rotation through one radius and translational correlation time is the time for moving one diameter.) In view of the equation, the addition of a second solvent to an electrolyte solution should decrease the symmetry of the solvation shell of the metal ion, thus increasing the electric field gradient and broadening the resonance line. Viscosity may also change in the solvent mixture, which changes the correlation time and the linewidth. Therefore, the variation of the resonance linewidth of a nucleus reflects the (environment of the nucleus and. provides information regarding ion solvation in the given solvent mixture. a. 133Cs linewidth variation in mixed solvents The 'variations in 'the 133Cs resonance linewidth in three mixed solvent systems (NMF-HMPA, NMF-DMF and NMF—DMSO) are less than 10 Hz over the whole range of measured solvent compositions (Fig. 31). The large 133Cs chemical shift has no influence on the resonance linewidth. 0.00200 0000.0 0_ 0.2300: nn.|E:_000 ..n 0590 .122 .0 00:00... 0.02 0;. DA. 050 #Av NH. .90 _ . _ . . . F) . _ . 0 0m20|m22 I ma (m2zlu2z I 9 :20qu2 4 . m. w . -2 ._. by. I I I I I I I e! CL 9 e ee.q e e e u e e H» 8 e e e e W . M 10m mH 1| U1 . mm 2 ( ron 90, In NMF-DMF and NMF-DMSO systems, the largest linewidth is exhibited at 0.9 mole fraction of NMF, while that of the NMF-HMPA system is for the equimolar solution. There is no relationship of these observations to the l33Cs chemical shift results reported in section A. b. 23Na linewidth variation in mixed solvents In contrast to the small variation in the 133Cs resonance linewidth, relatively large changes in 23Na linewidths are observed with changes in solvent mole fraction. The linear changes of linewidth ‘vs. solvent composition in NMF-H20, NMF-MeOH and NMF-Eton systems (shown in Fig. 32(a)) seem to be controlled by the variation in viscosity. In the NMF-FA system, the maximum linewidth is observed at 0.6 mole fraction of NMF, indicative of the most asymmetric environment for the Na+ ion. Although FA has higher viscosity than NMF, the 23Na linewidth in pure FA is narrower than in NMF, which may suggest that solvation is more symmetric in FA. Little change is observed in the 23Na linewidth as a function of solvent composition in NMF-DMSO mixtures. This is not surprising, considering the similar viscosity and donicity of the two solvents. The curvature of 23Na linewidth variation in NMF-AN mixtures clearly indicates that the Na+ ion is preferentially solvated by NMF. There are two interesting systems, NMF-NM and NMF-HMPA, which show extremely broad lines in the mixture, with 91 0.00200 005E E 52300: nmlEEoom .onmnenoi ....22 .0 00:00... 0.02 0.. 0.0 0.0 .v.O Nd 0.0 p b p p — b b s L 5 ON (01022 a 0 1901.122 ... 1 8 10.2122 4 . W Ton I. e n . m_w 70 e e I I I I I “10¢ as. I I I I O "I. I e e an“ . . . . a... m. e e W e e e 0 I00 \nm 2 ( row 92 2520... 828 5 £23.05. 818220 3030560 .122 .0 cones: e_oz 0.. 0.0 0.0 To «.0 0.0 — pl - s p s . L — . —0 221022 .. a 8 (0221.122 . .0 0 . . . . . . ion m. I I I e w I I I . .jOOP 7—U I I C. . . won. w. . .. m. w loom u e \u/ H . . Tons 72k 93 maximum. breadth at 0.07 and 0.6 mole fraction of NMF, respectively; These two systems have the largest 23Na chemical shift 'variation. among' all the systems ‘measured (Fig. 32(b)): thus the solvent components are quite dissimilar, and their mixture may produce large electric field gradients. It is noteworthy that the broadest line is located near the isosolvation points of these two system, but not exactly at those points. From these two curves of the linewidth variation, it is easy to discern the relative solvating abilities of the two solvents in the mixture. The, dependence of the 23Na linewidth on the concentration of dissolved salts shows reasonable results in that an increment of the salt concentration increases the viscosity of the solution and broadens the resonance line. Possible anion effects on the 23Na linewidth were also investigated in NMF-AN, NMF-DMSO, NMF-H20 and NMF-HMPA systems. The participating anions exert no influence in the latter three systems. However, in the NMF-AN mixture, an anion effect exists at NMF mole fraction less than 0.1. This is additional evidence of ion pair formation in AN-rich mixtures (Fig. 33). 2. gz-VIS Studies of Sodium Picrate ip gimed Solvenpe UV-VIS studies of picrate salts in pure NMF demonstrate that the absorption band is sensitive to the interaction between solvent and the picrate anion. It is interesting to study the anion solvation in mixed solvents. 94 200.0 00.22 200.0 200(2 200.0 _