A STUDY OF 1, 1, 3, 3-TETRAMETHYLGUANIDINE AS A NONAQUEOUS SOLVENT Thesis I09 I50 Dag?“ of DI}. D. MICHIGAN STATE UNIVERSITY Melvin Lee Anderson 1965 THESIS LIBRARY Michigan Stan University This is to certify that the thesis entitled A STUDY OF 1, 1, 3, B-TETRAMETHYLGUANIDINE AS A NONAQUEOUS SOLVENT presented by Melvin Lee Anderson has been accepted towards fulfillment of the requirements for _Eh..__D-_ degree in _Cb.emist r y WW Major professor 0-169 ABSTRACT A STUDY OF 1, 1, 3, 3-TETRAMETHYLGUANIDINE AS A NONAQUEOUS SOLVENT by Melvin Lee Anderson Although I, l, 3, 3-tetramethylguanidine has been known for a long time, it has only recently been studied as a nonaqueous solvent. In order to evaluate its solvent properties, a number of general areas were investigated: physical properties, salt solubilities, coordinating characteristics, general reactivity, and acid-base character. The high heat of vaporization of tetramethylguanidine and its transformation to a glass when cooled indicate considerable association in the liquid. Its infrared spectrum gives evidence for hydrogen bond- ing at the imine nitrogen. The proton magnetic resonance spectrum is consistent with the conventional molecular structure. Tetramethylguanidine was found to have such a low dielectric constant (11. 5) that ionic solutes are probably only partially dissociated in solution. Solubilities of a number of sodium and potassium salts were measured and were found to have the same relative order with respect to one another as solubilities of inorganic salts in liquid ammonia, but tetramethylguanidine is a somewhat poorer solvent. Tetramethylguanidine is completely miscible with most common organic solvents. Transition metal salts form highly colored solutions in tetramethyl- guanidine but the colored complexes are difficult to isolate. The solvent functions as a monodentate ligand, apparently by coordination through the imine nitrogen. Cobalt(lI) complexes in tetramethylguanidine seem to be tetrahedral, and the concentration or identity of the anion has little Melvin Lee Ande rson effect on the visible absorption spectrum of the complex. In general, Beer's law is not obeyed by cobalt(II) salts in tetramethylguanidine. The following complex species are postulated in these systems: Co(TMG)4++, Co(TMG)4++;X, and Co(TMG)4++;ZX-, where the latter two are ion-pairs. Tetramethylguanidine hydrolyzes slowly at room temperature to form 1, l ~dirnethylurea and dimethylamine: NH 0 (CH3)3N-c'i-MCI-13,)Z + H20 ——> (CH3)2N—&—NHZ + (CH3)ZNH Moist solvent reacts rapidly with carbon dioxide to precipitate tetra- methylguanidinium bicarbonate: fin HZN+Hco3’ (CH3)ZN- -N(CH3)2 + H20 + co2 —-> (CH3)zN-&-N(CH3)2 Tetramethylguanidine does not react with metallic sodium or potassium, but lithium decomposes it. The liquid is oxidized by permanganate, dichromate, periodate, or silver(I). Mercury(I) salts undergo disproportionation. Tetramethylguanidine titrates as a strong base in either aqueous or nonaqueous medium. The chloride, bromide, bicarbonate, and acetate salts were isolated and their melting points determined. Only monoprotonation occurs (at the imine nitrogen), followed by charge localization on the carbon atom: WHZ (CH3I2N‘E‘ NiCH3)z Eight visual indicators for acid-base titrations were studied in tetramethylguanidine as a solvent. Their color changes correspond to conductometric end points if the indicator is a weaker acid than the Melvin Lee Ander son acid being titrated. Good conductance curves were obtained for p-toluenesulfonic, benzoic, and salicylic acids as well as for phenol, ammonium bromide, and tetramethylguanidinium bromide. Titration with tetra-n-butylammonium hydroxide gave two conductance breaks for g-nitrophenol and l, 3-dinitrobenzene, whereas 1, 3, S-trinitrobenzene exhibits three distinct end points. These phenomena can be interpreted as addition of one, two, and three hydroxyl ions to the respective nitro- aromatic ring. The fact that maleic and citric acids show two and three conductance breaks, respectively, in their titrations, indicates that differential titration of mixed acids in tetramethylguanidine is possible. No acidic properties were found for urea, acetamide, benzene, or nitrobenzene. A STUDY OF 1, l, 3, 3~TETRAMETHYLGUANIDINE AS A NONAQUEOUS SOLVENT BY Melvin Lee Anderson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1965 AC KNOW LED GME NTS The author wishes to express his appreciation to Drs° Robert N. Hammer and Alexander I. Popov under whose guidance this investigation was accomplished. To the Dow Chemical Company, Midland, Michigan, the author extends his gratitude for a leave of absence, for a one-year fellowship and a summer research grant, and for the toxicological investigation of l, l, 3, 3-tetramethylguanidine. Thanks go to the American Cyanamid Company, Wayne, New Jersey, for their gift of several gallons of tetramethylguanidine. A heartfelt expression of gratitude is extended to the author's wife, Fern, for many years of patience and understanding throughout the pursuit of the Ph. D. degree and for typing the preliminary drafts of this dissertation. ii TABLE OF CONTENTS INTRODUCTION . HISTORY . Preparation ofl, 1,3, 3-Tetramethylguanidine and Related Compounds. . . Properties ofl, 1,3, 3—Tetramethylguanidine and Related Compounds. Infrared Spectra of Guanidines Protonation of Guanidines . . . . . Guanidines as Ligands in Metallic Complexes EXPERIMENTAL SECTION . Purification ofl, 1,3, 3-Tetramethylguanidine . Techniques for Handling 1, l, 3, 3— Tetramethylguani dine Instrumental Methods. . . . . . . Derivatives ofl, 1, 3, 3- Tetramethylguanidine Chemicals . RESULTS AND DISCUSSION . Physical Constants of l, l, 3, 3-Tetramethylguanidine Solubilities inl, l, 3, 3- Tetramethylguanidine Reactions ofl, l, 3, 3— Tetramethylguanidine . Transition Metal Complexes of 1,1 3, 3 Tetramethxl— guanidine . . Acid- Base Titrations in1 1 3°,3-Tetramethy1guan1d1ne CONCLUSIONS . RECOMMENDATIONS . REFERENCES . APPENDICES iii Page 10 12 12 16 16 19 20 34 35 37 37 48 60 76 94 121 125 126 132 TABLE II. III. IV. VI. VII. VIII. IX. XI. XII. XIII. LIST OF TABLES Basic Strengths of Guanidines . Melting Points of Guanidinium Salts . Physical Properties of 1, l, 3, 3-Tetramethy1— guanidine . Purification of Alkali Metal Salts . Physical Constants for 1, 1, 3, 3-Tetramethyl— guanidine . Comparison of Properties of 1, 1, 3, 3-Tetramethyl- guanidine with Related Compounds. Quantitative Solubilities in 1, l, 3, 3-Tetramethyl- guanidine . Comparison of Solubilities in Various Solvents . Semiquantitative Solubilities in 1, 1, 3, 3-Tetra— methylguanidine . Miscibility of l, 1,3, 3-Tetramethylguanidine with Organic Liquids Melting Points of Salts of 1, 1, 3, 3-Tetramethyl- guanidine . Miscellaneous Reactions of 1, l, 3, 3-Tetramethyl- guanidine . Transition Metal Complexes Obtained by Evaporation of Solvent . iv Page 11 24 38 4O 51 53 55 59 bib 73 78 LIST OF TABLES — Continued TABLE Page XIV. Effect of Anions in a Tenfold Excess on the 600 mu Absorption of Co(CzH3OZ)Z°4HZO in Tetramethyl- guanidine....................... 82 XV'. Spectra of Cobalt(II) Chloride, Sodium Iodide Solu- tions in Tetramethylguanidine . . . . . . . . . . . . 93 XVI. Colors of Indicators in 1, 1,3, 3-Tetramethyl- guanidine....................... 96 XVII. Wavelength Maxima: Visible Absorption Spectra of Indicators.......................100 XVIII. Compounds Incapable of Titration as Acids in l,1,3,3-Tetramethylguanidine. . . . . . . . . . . . 117 XIX. Recovery of Acids by Conductometric Titrations . . 119 FIGURE 10. ll. 12. 13. LIST OF FIGURES Infrared Spectra of l, 1, 3, 3-Tetramethylguanidine . Apparatus for Distillation of 1, 1, 3, 3-Tetramethy1- guanidine . Dielectric Constant Apparatus . Apparatus for Solubility Equilibrations Standard Flame Photometric Curve for Sodium Hydrogen Sulfate Solutions . Conductometric Titration Cells Conductance of Methanol in l, 1, 3, 3-Tetramethyl- guanidine . Effect of Temperature on the Specific Gravity and the Refractive Index of 1, 1, 3, 3—Tetramethy1- guanidine . Standard Curve: Dielectric Constant Measurement. Infrared Spectrum of Liquid 1, l, 3, 3-Tetramethyl- guanidine . Comparison of Infrared Spectra of 1, 1, 3, 3-Tetra- methylguanidine Proton Magnetic Resonance Spectrum of 1, 1,3, 3- Tetramethylguanidine Carbon Dioxide Absorption by 1, 1, 3, 3—Tetramethy1- guanidine . vi Page 13 17 22 26 29 32 33 42 44 46 49 SO 63 LIST OF FIGURES - Continued FIGURE 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Titration of 1, 1, 3, 3-Tetramethylguanidinium Bicarbonate . Infrared Spectra of l, 1, 3, 3-Tetramethy1guani- dinium Bromide . Titration of l, 1, 3, 3-Tetramethylguanidine in Aqueous Solution with Standard Acid . Potentiometric Titrations of 1, 1, 3, 3—Tetramethyl~ guanidine . Visible Absorption Spectra of Hydrated Cobalt(ll) Salts in l, 1, 3, 3-Tetramethylguanidine . Visible Absorption Spectra of Anhydrous Cobalt(ll) Salts in 1, I, 3, 3-Tetramethylguanidine . Effect of Water on the Absorption Spectrum of CO(C2H302)2' 4Hzo . Beer' 5 Law Plot for Anhydrous CoClZ Solutions Beer's Law Plot for Anhydrous Co(C2H3Oz)2 Solutions . Effect of Iodide Ion on the Spectrum of Co(CZH3OZ)Z-- 4HZO . Effect of Iodide Ion on the Spectrum of CoClz': 6HZO. Visible Absorption Spectra of [Co(TMG)4](ClO4)2 . . Visible Absorption Spectra of Curcumin, An Acid- Base Indicator . Visible Absorption Spectra of Clayton Yellow, An Acid-Base Indicator . Page 64 68 7O 72 8O 83 84 86 87 88 89 97 98 LIST OF FIGURES - Continued FIGURE 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. Conductometric Conductometric Conductometric Conductometric Conductometric Conductometric Conductometric Conductometric Titration of p-Toluenesulfonic Acid Titration of Benzoic Acid. Titration of Salicylic Acid . Titration of g-Nitrophenol Titration of Phenol . Titration of 1, 3-Dinitrobenzene Titration of 1, 3, 5-Trinitrobenzene Tit rations: (a) Ammonium Bromide; (b) Tetramethylguanidinium Bromide . Conductometric Conductometric Titration of Citric Acid Titration of Maleic Acid viii Page 101 103 105 106 108 110 116 LIST OF APPENDIC ES APPENDIX Page I. Toxicological Investigation of 1, 1, 3, 3-tetra- methylguanidine . . . . . . . . . . . . . . . . . . 133 11. Data and Calculations: Solubility of NaHSO4 . . . 135 ix INTRODUCTION Since the solvent properties of l, 1, 3, 3-tetramethy1guanidine, NH H3 C\ H CH3 N - C - N / \ 1-13C CH3 are essentially unknown and because the compound has become avail- able from the American Cyanamid Company ( 1), this investigation was undertaken to determine the utility of tetramethylguanidine as a nonaqueous ionizing solvent. Five general areas of study were investigated to a limited extent: physical properties of tetramethylguanidine, solubilities in the solvent, inorganic reactions of tetramethylguanidine, acid-base equilibria, and coordination characteristics of tetramethylguanidine toward transition—metal ions. Tetramethylguanidine contains two tertiary amine nitrogens and one imine group, each with a lone pair of electrons. Therefore it should be a strong donor ligand in the formation of metal ion complexes (2). Because of the closeness of the nitrogen atoms, however, chelation seems unlikely. Thus, if tetramethylguanidine were bidentate, the resulting complex would contain an improbable four-membe red ring. On the other hand, tetramethylguanidine may act as a bridging group in a polynuclear complex (2) . Tetramethylguanidine has a conveniently wide liquid range extend- ing from well below room temperature to a boiling point of 159-1600 C. (l). The exceptionally strong basic nature of tetramethylguanidine (pKa in water at 25° = 13. 6) (1) should make it very useful as a solvent for the study of weak acids. Hydrogen-containing compounds which are only weakly acidic or even basic in water should have their acidic character sufficiently enhanced for titration in tetramethylguanidine (3). Tetramethylguanidine is a strong proton acceptor which levels acids . + to the acidity of the solvent cation, TMGH (4): TMG + HA ———> TMGH+ + A' . (1) This reaction, however, does not imply complete dissociation of the resulting electrolyte. The degree of dissociation depends upon the dielectric constant of the solvent, the nature of the anion A-, etc. (3). As a basis for further work, it was also desirable to determine some solubilities of alkali metal salts in tetramethylguanidine and to characterize certain reactions of the solvent. HISTORY Preparation of 1, l, 3, 3-Tetramethylguanidine and Related Compounds The earliest known reference to 1, 1, 3, 3-tetramethylguanidine is the work reported by Berg (5) in 1893 in which he found that tetra- methylguanidine is formed by the action of dimethylcyanamide on dim ethylamine hydrochlo ride: H3C\ /CH3 H3C\ NH /CH3 /N-CEN + ClI—l-HN\ —-—+ /N-C - N\ + HCl (g) In a fairly complete study of the preparation and properties of a number of the methylated guanidines, Schenck (6) in 1912 prepared 1, 1, 3, 3-tetramethylguanidine by heating ammonia and pentamethyl- thiuronium iodide in ethanol solution: + _ H3C\ ISCH31 I,CH, H3C\ NH /CH3 /N-C — N\ + NH3 ———> N-C — N + H1 + CH3SH (3) / \ — H,c CH, H3C CH3 The isomeric compound 1, 1,2, 3-tetramethy1guanidine was prepared by Schenck from dimethylamine and trimethylpseudothiourea: I /N-d - SH + HN\ ———> N- - N + (3143811 (3) Schenck also prepared the completely methylated compound pentamethyl- guanidine by heating methyliminodiethylcarboxylic acid and dimethylamine: 3 H5C2\ NCH3 /CH3 H3C\ NCH3/CH3 /C3H5 N-C-OH+2HN\ ——> N- -N +HN +1120 / / \ \ Hsc2 CH, H,c CH, C2115 (5) as well as by the action of niethylamine on pentamethylthiuronium iodide: + _ H3C\ SCHH CH, /H H3C\ NCH,/CH, /N-C - N\ + H,CN\ —7\ /N— - N\ + CH,SI—1 + 141(6) H3C CH3 H H3C CH3 In 1924 Schenck and von Graevenitz (7) reported another prepa- ration of 1, 1, 3, 3-tetramethy1guanidine. 1, l—Dimethylguanidine was first prepared by treating dimethylcyanamide with alcoholic ammonia: H3C\ H,c- NH /N-CEN + NH, ——>~ N - c - NH,. (1) H,c H,c The solution was then saturated with hydrogen sulfide to give 1, l -dimethy1thiourea: \ H3C\ NH 11,3 . j 1,,N - c — NH, + H,S -—~>~ /N — c — NH, 4- NH, .- (8) H, c H, c“ Methylation with dinnethyl sulfate gave 1, 1, 2-trin1ethylpseudothiourea, H,c 11,C\ SCH, \ \ /N ' i ‘ NH2 '1” (H3C)zSO4 ”fi‘ /N " C 3 NH '1‘ CH3HSO4 ' (2) H3C H3C’ This product was treated with alcoholic dimethylamine and me rcuric chloride to give the desired symmetrical guanidine derivative which was isolated as the chloroaurate. H3C\ CH, /CH, 2 /N - =NH+2HN\ + HgCi,-———-——->- H,C CH, H,C\ NH CH, .. ' / 2 /N - C — N\ + Hg(SCH,)2 + z HCl (19) H,C CH, Schenck and von Graevenitz (7) also obtained 1, 1, 3, 3—tetramethy1- guanidine by treatment of cyanogen iodide with dimethylamine: /CH, /,,.CH, H,C\ NEG - 1+ 2 HN\ —-9~ NEG — N\ + /NH-Hl (L1) CH, CH, H,C /CH3 /CH3 /N(CH3)Z NEG - N + HN —-—> HNPC (12) \ \ \ -—- CH3 CH3 N(CH3)2 At about the same time a number of papers were published by Lecher and co-workers (8a, b, c) on the preparation and characteri— zation of completely alkylated guanidines such as pentamethylguanidine and hexamethylguanidinium iodide. As an example, the former was prepared by treating dimethylamine with 1, 1, 2, 3—tetramethy1pseudo— thiourea and removing methyl mercaptan from the reaction with mercuric chloride: H,C\ NCH, ,CH, /N - E - SCH, + HN\ + HgCi, ——-—> H,C CH, H3C\ NCH3 /CH, 2 /N — ti - N\ - HC1 + Hg (SCH,), (13) H,C CH, The free base was then obtained by heating the hydrochloride salt. and driving Off hydrochloric acid. Only a few scattered and incidental preparative methods for alkyl-substituted guanidines can be found in the literature after 1925. In 1959 a patent was issued to Cain (9) on the preparation of l, 1, 2, 3— tetraethylguanidine and its salts, but the method is generally the same as that of Lecher gt a_1_1. (8). A genuinely new method for the prepa- ration of pentaalkylguanidines was reported by H. Z. Lecher and co- workers in a 1958 patent (10). In this case, the hydrochloride of a C-chloro-N, N, N-trialkylformamidine reacts with a dialkylamine in nitrobenzene solution: NR R R R NR R I / / I Cl-C-N -HC1+HN --——>- N-C-N +2HC1 (14) \ / \ R R R The method can be used to prepare either iso- or hetero-pentaalkyl- guanidines such as pentamethylguanidine or 1, 1, 2-trimethy1-3, 3- diethylguanidine . Properties of 1, l, 3, 3-Tetramethy1guanidine and Related Compounds Until the work of Lecher and Graf (8b), it was generally assumed that in a guanidinium salt such as the hydrochloride, the proton is bound by one of the amine, or -NH;,, groups. These author“ showed, however, that the reaction of methyl iodide with l, I, 3, 3-tetramethyl- 2-ethy1guanidine as well as of ethyl iodide with pentamethyliguanr-dmc yields identical reaction products: CH,1 H3C\ 1'\'IC,H5 CH, H,C\ +NC,H5 /CH_, _ - ' ' y i I _ _ l /N C N\ T CH,1 /1\ C N\ (15, H,C CH, H,C CH, Cszi H,C NCH, /CH3 H3C\ +NCH, /CH3 >N-h -N\ ~+C2H51+ / -6 -N\ (1Q) H,C CH, H,C CH, As shown in equations 15 and _1_6, this could only be explained on the assumption that salt formation involved the imine, or :NH, group. Lecher and Graf (8b) also pointed out that guanidinium hydroxide and its alkyl derivatives are bases of about the same strength as the alkalies. These hydroxides were not isolated but by using acid-base indicators, it was shown that aqueous solutions of the various guanidin- ium iodides do not hydrolyze but behave as salts of strong bases and strong acids. In 1951 Angyal and Warburton (l 1) did a fairly complete study of the basic strength of guanidine and several of the methylated guanidines. Most of the salts of the methyl-substituted guanidines were prepared by the methods of Schenck (6, 7) and Lecher (8), and the respective pKa values were measured by potentiometric titration in aqueous solution. The results of this work are Shown in Table I where it is seen that all of the guanidines have pKa' s of the same order of magnitude. The value found (11) for free guanidine was 13. 6 at 250, in good agreement with an older value of 13. 65 (12) and a newer one of 13. 54 (13). Recently a value of 14. 06 has been reported (14) for the pKa of tetra- methylguanidine (isomer not reported) in aqueous solution. Angyal and Warburton (11) state that the absolute error of their results (Table I) may be as much as 0. 2 pKa unit, but since all values were obtained under identical conditions, they believe that differences in relative values are Significant. This conclusion seems open to question. Melting points of a number of salts of the methyl~substituted guanidines have been reported by Schenck (6), Angyal and Warburton (11), and Kitawaki (15). These are given in Table II. Table I. Basic Strengths of Guanidines Guanidines pKa Reference Guanidine 13 . 6 11 Guanidine 13. 65 12 Guanidine l3 . 54 13 1-Methy1guanidine 13 . 4 11 1, 1-Dimethy1guanidine l3. 4 11 1, 3-Dimethylguanidine 13.6 11 1, l, 3-Trimethy1guanidine 13. 6 11 I, 2, 3-Trimethylguanidine 13. 9 11 l, 1,2, 3—Tetra1r1ethy1guanidine 13. 9 I 1 I, 1, 3, 3-Tetramethylguanidine 1.3. 6 11 Tetramethylguanidine (isomer not reported) 14. 06 14 Pentaniethylguanidine 13 . 8 1 1 N53 932 932 -EfieEeEem 212.2 m 6-32 3.32 LEEEeEeSM .m .2 .2 Tm: m.2-.2m2 2-32 -2.Eeefiebee-m.~.2.2 8? Te: 8? 233825.; .N .2 21m .NON m 2.23 23362;; .2 .2 2182 2132 m-m .82 me: seems? 2.3385; .2 92.3. Tmew 2.38852 .2 02% 2332-2 oCCOHJU oumndmonoHJO ofimHfiZ owfiUoH opmnufinfi nudism QEUHQMDO Co .328 852531130 20 meson mass: .22 eBeH 10 E: s sentially the only available values of physical constants for 1 , l, 3, 3 -tetramethylguanidine come from Market Development Depart— ment reports (1,16) of the American Cyanamid Company. These data are given in Table 111 along with an additional boiling point measurement from the literature (17a). Recently Williams and co-workers (17a, b) used 1,1,3, 3-tetra- methylguanidine as a solvent for the titration of weak acids. They found that phenol and several substituted phenols could be titrated potentio- metric ally with tetramethyl— or tetrabutylammonium hydroxides using a glas S electrode and a modified calomel electrode. Good equivalence point inflections were obtained in all titration curves. Other electrode system s as well as indicators for the titration of weak acids in tetra-- methylguanidine also were investigated. A glass electrode coupled with a s ilver—silver bromide reference electrode was most satisfactory The indicators alizarin yellow and azo violet gave results in agreement with potentiometric end-point values. Infrared Spectra of Guanidines It wa S not until approximately fifteen years ago that the infrared spectra of some of the guanidines were reported .‘1‘1 the literature. The extens ive work by Randall e_t 9L1: (18) includes the infrared spectra for guanidinium acetate, carbonate. and thiocyanate; methylguanidhmrn chloride and sulfate; and triphenyl— and symmetrical diphenylguanidthe. FEW frequency assignments have been made for guanidines Fabian it a, (19), in their review, report wavelengths for the C:N stretching absorption in the range of 5. 92-6.17 u. The Sadtler compil— ation 120) Contains the infrared Spectrum from 2-15 ,1 of liquid 1, 1, 3, 3- tetramethylguanidine (Figure 1, curve a). Curve 1) in Figure. 1 is the infrared Spectrum of a five percent solution of tetramethylguanidine in cyclohexane obtained by Drago (21). Table III. Physical Properties of l, 1, 3, 3-Tetramethylguanidine Property Value Reference Boihng pouu, OC 159-60 1 Boihng ponu, OC (738rnnfl 159—61 17a pKa, 250 13.6 1 pH (1% solution) 12.7 1 Solubility Soluble in water and common organic solvents Vapor pressure, mm at1600 760 at 1000 100 at 65° 20 at 250 0.2 16 16 16 16 12 Protonation of Guanidine s Angyal and Warburton (11) in 1951 stated that "guanidines are theoretically not expected and have never been found to be diacid bases. " However, two years later, Williams and Hardy (22) found 1. 28 protons per guanidine molecule in 99. 9 percent sulfuric acid solu- tion. This is the only evidence in the literature for anything greater than monoprotonation of guanidines. Most of the research within the past few years has dealt with the position of protonation rather than the amount. From positions of the infrared absorption bands of guanidine salts and some methyl-substituted guanidine salts, Goto and co—workers (23) concluded that the positive charge resides mainly on the central carbon atom rather than being distributed over the entire guanidinium group. From molecular orbital calculations, Paoloni. (24) concluded that the guanidinium ion Should be considered as a tertiary positive carbon or triaminocarbonium ion, (NHZ)3C+, analogous to the quaternary nitrogen ion. Kotera e_t al. (25) attempted to determine the location of proton addition in guanidine by analyzing the proton magnetic resonance spectrum of crystalline guanidiniuln iodide. They concluded that pro» tonation occurs predominantly on the imine (:NH) nitrogen but that amine (~NH2) addition could not be ruled out. Guanidines as Ligands in Metallic Complexes There is a dearth of information in the literature on the use of guanidine as a complexing agent for metallic ions. The related com— pounds, urea and thiourea and their derivatives, on the other hand, have been extensively investigated as ligands. The metal complexes of the a1kyl~substituted ureas haVe been studied to a limited extent. 13 ¢. .ocfiowcmsmaxsuoEmHuoHé .m .2 .H mo ouuooam condemn: 4 unaware N. 1.502335: 0. m w ¢ N / >71 2;. .ummv “223310486 2. 20:33 £3 .9 2 \l 3~.uu5 m5{: See pp. 69—70, 28 Willard, Merritt, and Dean (38). To minimize any anion effect on the flame, separate standard curves were constructed for each salt from measurements on standard aqueous solutions. A typical standard curve is shown in Figure 5. After each residue was titrated for tetramethylguanidine, the solution was diluted with water to a convenient volume (usually 25 ml) in a volumetric flask. Samples of these solutions were then aspirated in the flame photometer several times and an average flame intensity calculated. The actual flame intensity due to the alkali metal content was obtained by subtraction of the flame background and the average intensity of a blank. The blank solutions contained tetramethylguanidine (titrated to pH 5) in roughly similar concentrations to the samples. The amount of alkali metal in the sample was obtained directly from the standard curve, and salt solubilities were then calculated in grams of solute per hundred grams of solvent. A typical tabulation of data and calculation of solubility (for NaHSO4) is given in Appendix II. A second method of solute analysis in the residue from samples of saturated solution was by titration of the anion in question. For instance, sodium bromide was determined by titration with standard silver nitrate solution. The third method of solute analysis was simply that of determining the weight of the residue after evaporation of excess solvent. Any weight due to residual tetramethylguanidine was subtracted from the total weight of the residue after titration for the solvent. This method was used only in cases where flame photometry was unsuitable and where the pre- cision of the results of at least two of the triplicate samples was five percent or less. The solubilities of KBrO3 and KClO, were obtained by the weight-residue method. Semiquantitative solubilities of a large number of inorganic com- pounds were also obtained by the weight-residue method. In addition, SCALE DIVISIONS 29 90 80 70 60 50 40 30 20 0 0.2 0.4 0.6 0.8 1.0 PPM No Figure 5. Standard Flame Photometric Curve for Sodium Hydrogen Sulfate Solutions. l.2 3O ' solubility ranges or minimums were obtained in conjunction with ' experiments in which tetramethylguanidine was used as a solvent. C. Spectra Ultraviolet, visible, and near infrared spectra of tetramethyl— dine solutions were obtained with a Beckrnan Model DK-Z recording rophotometer. Glass-stoppered 1 cm silica cells were used in measurements to minimize exposure to air. Infrared spectra from 2-15 u were measured with a Perkin-Elmer l 21 spectrometer using cells with either sodium chloride or sium bromide windows. Proton magnetic resonance spectra were led with a Varian Model A-60 spectrometer. D. Titrations All potentiometric titrations in aqueous solution for acids or bases done with a Beckman Model G pH meter using a glass-«calomel ode system. Aglass-—silver, silver chloride electrode system sed in argentometric titrations for halide ion. Tetramethylguanidine etermined by nonaqueous titrations in glacial acetic acid and in Iitrile using glass-—silver, silver chloride electrodes and perchloric 1 acetic acid as the titrant. Dry nitrogen was bubbled into these ieous solutions during titration, and the titrant was protected from :h a tube of Drierite on top of the buret. Conductometric titrations in tetramethylguanidine as a solvent were sing a Serfass Model RC M15 Conductivity Bridge. Solutions ontained in a 48 x 87 mm beaker closed with a rubber stopper h which passed the tip of a 10 ml buret. a nitrogen inlet, and a 2n outlet. Sample and titrant solutions were protected from the >here with drying tubes containing Ascarite and Drierite. During In, solutions were mixed with a small magnetic stirrer. No 'ature control was used. 31 For the series of titrations reported in this thesis, two conductance 3113 with slightly different cell constants were used. The electrodes .cell number I were those of a commercial dip cell fitted into the rubber opper and sufficiently immersed in the sample solution to cover the ectrode surfaces. The two electrodes were of shiny platinum, each Ix 13.5 mm, and positioned 2 mm apart. The measured constant for is cell using 0. 0100 I\_/l KCl solution was 0.1168 cm‘l. A drawing of e cell is shown in Figure 6a. Cell number 11 had a constant of 0. 05143 cm-l. It was constructed th electrodes fastened to the inside wall of the cell as shown in gure 6b. The electrodes were 15 x 2.0 mm shiny platinum separated 3. 5 mm. Conductance titrations were done by a standard procedure. For imple, in the titration of weak acids with a base in tetramethylguanidine ,ution, sufficient solid acid to give about a 0.01 M solution was weighed and dissolved in tetramethylguanidine in a 100 ml volumetric flask. ese operations were done in a dry box. Fifty or sixty milliliters solution was pipetted into the conductance cell which was immediately ppered, dry nitrogen was bubbled through the solution, and stirring un. The titration was then carried out. The titrant (base) used in these conductance titrations was a :hanol solution of tetra-n—butylammonium hydroxide (Eastman). This ,erial was obtained as a 25 percent solution of the base in methanol was diluted further with dry methanol to give titrant solutions rang- from 0. 2—0. 8 I\_/_I. The concentration of base was such that only it three milliliters was required for sixty milliliters of acid solution. conductance of pure methanol in tetramethylguanidine was determined a standard conductance curve was drawn (Figure 7). Methanol con- ance values were then subtracted from measured conductances in it rations . l' BURET ASS INSULATION Figure 6a. 32 LEADS TO IRIDGE —— VENT HOLE —- PT ELECTRODES MAGNETIC STIRRING BAR Conductometric Titration Cell Number I. LEADS TO BRIDGE RUBBER INSULATION ELECTRODES I / It/{II/a \. n BURET MAGNETIC STIRRING BAR Figure 6b. Conductometric Titration Cell Number 11. 33 .ocdvficgwgauoEdun—orfium .m J .H :fi Bad—3oz mo oocduodvcoo zonzu n . N .N. ouswdh Q0. N.. ON. 'N. ON. Nn. 0| I _NNO ' JDNVLDDONOD G I 34 When indicators were used in titrations, they were generally . o . . . . . . added as oven-dried (110 ) solids. Initial indicator concentrations were approximately 10‘4 M except in a few cases where a somewhat higher concentration was necessary to develop sufficient color intensity. Derivatives of 1, 1, 3, 3-Tetramethylguanidine Tetramethylguanidinium salts were prepared as precipitates by addition of acid to an excess of tetramethylguanidine, followed by fil- tration. The precipitate was washed with pure solvent. Occluded solvent was then removed under vacuum at room temperature. The chloride and bromide salts were prepared in this manner from anhydrous hydrogen halides, while the acetate was obtained by addition of glacial acetic acid to tetramethylguanidine. No precipitate formed when seventy percent aqueous perchloric acid was added to tetramethylguanidine; neither did precipitation occur when the mixture stood in air. Tetra- methylguanidinium bicarbonate was prepared by addition of solid or gaseous carbon dioxide to a 1:1 TMGzHZO mixture. All the tetramethyl - guanidinium salts were found to be hygroscopic and were stored over Mg(ClO4)z or P205 in a vacuum desiccator. Colored solid transition metal compounds, presumed to be com- plexes, of tetramethylguanidine with the acetates of cobalt(Il), nickel(Il), and chromium(lll) were prepared and analyzed. The method involved preparation of an essentially saturated solution of the hydrated metal acetate and vacuum evaporation of excess tetramethylguanidine at temperatures below the point of decomposition. After the resultant filmy solid was ground in a dry atmosphere, it was dried further under vacuum at the temperature previously used. The cobalt and nickel products were heated to 115-200 while only a 65-700 temperature was used for the chromium compound. Higher temperatures decomposed or melted the substances. 35 rocedures similar to these were tried using hydrated copper(ll) in attempts to prepare a solid copper-tetramethylguanidine An intensely blue solution was obtained which could be ted to a dark blue tacky material which showed little change olonged heating under vacuum at 65-700, but which decomposed ated at 800. Thus, no analysis of the copper-tetramethyl— ie reaction product was made. as . Purification he purification method developed for liter quantities of tetra— ;uanidine has been described above (p. 16 ff. ). In addition, hods used for the purification of the alkali metal salts used in ititative solubility measurements have been previously described V). The inorganic salts used for the semiquantitative solu- ;udies and other miscellaneous applications were simply dried to t weight. Acid-base indicators used in some conductance 718 were dried at 1100. ransition metal salt hydrates were generally stock material used further purification. Anhydrous CoClZ and Co(C2l-13OZ)Z were ad by vacuum dehydration of the hydrates at temperatures (1300 0, respectively) below their hydrolysis or decomposition points 20(ClO4)2~ 6HZO was purified by recrystallization from water and in air. The acids used in the conductometric titrations were laterials purified, if necessary, by sublimation or distillation eir melting or boiling points agreed closely with literature values. anesulfonic acid was used as the stock monohydrate since it was tto dehydrate completely (40). Subsequently, experiments showed ter of hydration from the monohydrate had no effect on the results .cid-base titrations. 36 >rganic solvents such as methanol or acetonitrile used in this ere purified by distillation from 4A Linde molecular sieves or rying agents (41). These distillations were customarily done 1 small (g. 1 cm x 20 cm) columns packed with Pyrex glass 5. Elemental Analyses Sommercial analysis of stock and distilled tetramethylguanidine ne of its derivatives was done by Alfred Bernhardt Micro- :al Laboratory, Mulheim (Ruhr), West Germany. Analysis of f the tetramethylguanidine derivatives was also done by Spang nalytical Laboratory, Ann Arbor, Michigan. The elemental s of transition metal complexes of tetramethylguanidine was 1 out by Schwarzkopf Microanalytical Laboratory, Woodside, >rk. RESU LTS AND DISCUSSION Physical Constants of l, 1, 3, 3-Tetramethy1guanidine Several physical constants of pure, distilled 1, l, 3, 3—tetramethyl- guanidine were determined. A list of measured as well as some calcu~ lated constants is given in Table V. A. Heat of Vaporization The heat of vaporization of tetramethylguanidine was calculated from the vapor pressure data of Table III published by the American Cyanamid Company (16). In order for all the data to describe a linear relationship between log p and l/T, the assumption was made that the vapor pressure of tetramethylguanidine at 250 is Z. 0 mm rather than 0. 2 mm as reported (typographical error?). This assumption is sup— ported by measurements made in this laboratory. The Trouton con- stant W8. 8 determined from the calculated value of AHV and the measured boiling point. The relatively high value for the heat of vaporization indicate 8 considerable association in the liquid state. The value of 11' Z kc a1 mole'1 for the heat of vaporization of tetramethylguanidine is quite Close to that of other nitrogen—containing solvents. For instance, the heat of vaporization of ethylenediamine is 11. Z kcal mole‘1 (42), While t1lat of dimethylformamide is nearly the same (11.37 kcal mole-1) (43)' The high Trouton constant of 25. 9, on the order of that for water at 26. 1 and methanol at 24. 9 (44), is an additional indication of a reaLSOIiably high degree of self-association in liquid tetramethyl- guanidine. 37 38 Table V. Physical Constants for 1, l, 3, 3-Tetramethylguanidine Measured Boiling point (745 mm) 159. 5° 0 Density, d” 0.9136 g ml-1 Dielectric Constant, €25 11.5 Refractive Index, n3 1.4658 Specific Conductance (250) 9. 9 x 10'7 ohm'1 Viscosity (250) 1.40 cp Infrared Absorption: v (N-H) 3311 cm‘1 v (C=N) 1594 cm-1 Proton Magnetic Resonance: CH3 2. 62. ppm NI-l 5.17 ppm Calculated Heat of Vaporization, AHV 11.2 kcal mole'l Trouton Constant 2.5. 9 cal mole‘l deg"l Molar Refraction, Mr?)5 34. 91 cm3 mole'1 39 B. Freezing Point It was not possible to measure the freezing point of tetramethyl- guanidine. All attempts to cool samples of tetramethylguanidine to very low temperatures resulted only in "glass" formation. At -75 to -700 C, tetramethylguanidine flows as a viscous liquid which becomes glassy at temperatures below about —800. No crystallization occurs down to -1900 C. Apparently there are intermolecular forces in liquid tetramethylguanidine that are too strong to permit rearrangement into the ordered crystalline state. C. Molar Refraction The molar refraction of tetramethylguanidine was calculated by the Lorenz—Lorentz equation using the measured value of the refractive index, 113. The value for Mr?)5 of 34. 91 cm3 mole“1 is smaller by about 1. 6 units than the molar refraction estimated from tables of atomic and bond refractions (45). This fact may indicate some contraction of the tetramethylguanidine molecule by intramolecular hydrogen bond- ing since the molar refraction represents the actual volume of the molecules in one mole of substance. D. Boiling Point The high boiling point of 1, 1, 3, 3-tetramethylguanidine together with its glass temperature of about —800 give this substance an ex— tremely wide liquid range. Thus, solutions in tetramethylguanidine could perhaps reasonably be studied over a two-hundred degree tempera- ture range. A few other compounds containing C—I-l and N-I—l bonds with similar wide liquid ranges are known. As indicated in Table VI, tetramethylguanidine appears to lie intermediate between N-methyl- aniline and diisobutylamine in a number of fundamental physical properties. This may indicate that the true melting point of tetramethyl- guanidine is somewhere between —57 and -700 C. mama mama 5mm; 0 .ofioa mason O E- A: as- Awmflmv 8-2 S- 0o .ofioa 92:32 m. mid Edd owoo Ta Em .asdcoo 88.4 $3.; Noam .a 5?: ozooodom N 43.: N .m: N .2: armada 523282 mzdimomoazmo: Aarovzfzavozrmmov aroamvzmrao Santos oggmfifionOmEQ ofigoddwafluogdu “0H. ocflficmgfiozt Z moddomgoo poumaom flog? ofipwcmomifiuoadgofitm .m J .H mo moSHoQOMnH mo GOwHHmmEoD .H> 3an 41 The boiling points of the free base forms of only two of the eleven theoretically possible methyl-substituted guanidines have been reported. These are 1,1, 3, 3—tetramethylguanidine, b. 159. 5 (this work, 1, 16, 17) and pentamethylguanidine, b. 155-160 (8a), both liquids at ordinary temperatures. Nevertheless, it is felt that at least four methyl groups are required per guanidine moiety in order that the compound be a liquid at room temperature. This would be similar to the corresponding family of methyl-substituted ureas where tetramethylurea is the only member which is liquid at room temperature (46, 47). Luettringhaus and Dirksen (47) point out that the normally strong association of amides due to hydrogen bridges involving the hydrogen on the amide nitrogen is precluded in tetramethylurea. The material thus can act only as a hydrogen—bond acceptor. The same argument can indeed be used for pentamethylguanidine, a liquid at room temperature (8a), and also for l, l, 3, 3-tetramethylguanidine if the hydrogen on the imide nitrogen in the latter shows a lesser tendency to form hydrogen bridges than a corresponding amide hydrogen. It would be interesting to know some of the common physical properties of all the other methylated guanidines. Even though most of them have been prepared as salts (6, 8a), apparently it is difficult to isolate the free bases (11). E. Density The density of l, l, 3, 3-tetramethylguanidine appears normal for a C, H, N- compound of its molecular weight (see Table VI). The plot of specific gravity of tetramethylguanidine vs. temperature from 12 to 320 C is given in Figure 8a. The refractive index of tetramethylguanidine over the range from 17 to 330 is shown in curve b of the same figure. The nearly linear relationships found for these properties is an indi- cation of ideal behavior over the given temperature ranges which were studied . 42 L470 L460 L400 L464 L402 0324 ‘\ 0920 “ 7. 1"E o.— O E ‘ x' C 3 ; 8 QDIC P" - ._, x m o , a 2 \ ~ t 2 o a t t‘ .. — m 2': QOI2 0.00. \‘\ IO ll 22 20 30 34 Turnnun: ,° c. Figure 8. Effect of Temperature on the Specific Gravity and the Refractive Index of l, 1, 3, 3-Tetramethylguenidine. 43 F. Dielectric Constant The standard curve used for the determination of the dielectric constant of l, l, 3, 3—tetramethylguanidine is shown in Figure 9. As indicated in the Experimental Section and shown in the figure, a number of reagent grade solvents of known dielectric constants were measured to construct the standard curve of measured frequency 2‘ dielectric constant. The frequency of 20. 2 x 106 c.p. s. for tetramethyl- guanidine corresponds to a dielectric constant of 11. 5. It is felt that the error involved in this figure, all factors considered, is i 5%. This would mean a resultant e for tetramethylguanidine of 11. 5 i 0. 6. Many of the more familiar basic nonaqueous solvents have similar dielectric constants. For instance, that of ethylenediamine at 200 is 14. 2 while that of pyridine at 250 is 12. 3 (48). By analogy, therefore, it is hoped that tetramethylguanidine may prove equally as useful a nonaqueous solvent as ethylenediamine or pyridine. The solvent properties of the latter two, of course, depend not only upon their dielectric constants but also on their strong electron-pair donating ability. The fact that tetramethylguanidine possesses a lone pair of electrons on each of its three nitrogen atoms also should make it a good donor solvent. G. Specific Conductanc e The measured specific conductance for 1, l, 3, 3-tetramethyl- guanidine of 9. 9 x 10'7 ohm'1 is a tentative upper limit. This value might be lowered by extremely careful purification. The observed specific conductance is about an order of magnitude greater than that for ethylenediamine and larger by a factor of 103 than the lowest value obtained for pyridine (48). However, since there is some indication (elemental analysis, sharpness of boiling point) that relatively pure tetramethylguanidine was used in this work, then the “high" specific 44 on Nn ON .ucoEouameoz acmumcoo 033339 "9650 pumpomum .o ousmfih w.»z<»mzou o.¢hou4u.o QN ON w. zenzo ronruo / 10¢sz ——_—_— —_——-— Br" > C1—. As expected, salts of the large anions such as ClO4_, NCS_, and I— possess high solubility while more ionic salts (HSO4-, C1_, etc.) are quite insoluble. Perhaps surprisingly, of the sodium salts the chlorate and nitrate show reasonably good solubilities while the acetate is low. Some idea of how tetramethylguanidine compares as a solvent with other nonaqueous solvents and with water appears in order. Table VIII lists the solubilities of a few common sodium salts in water, ammonia, and ethylenediamine (en) (55) and compares these values with the corresponding solubilities found in tetramethylguanidine in this work. In general, solubilities in tetramethylguanidine parallel those in liquid ammonia and in ethylenediamine, but tetramethylguanidine is a somewhat poorer solvent. It may be therefore that tetramethyl- guanidine does not solvate the cation or the anion of the solute as readily as does ammonia or ethylenediamine. It might also be expected that the entropy of solution in tetramethylguanidine would be quite large due to disruption of the hydrogen-bonded liquid structure. The greater dissolving power of ammonia over tetramethylguanidine is likely due to the smaller size of the former. In the case of ethylenediamine, its chelating ability could confer a stronger solvating tendency upon it than is possible for tetramethylguanidine. Under the conditions of analysis used in the quantitative solubility measurements, stable tetramethylguanidine solvates of only two of the salts in Table VII, sodium acetate and potassium bromide, were iso- lated. In other words, vacuum desiccation at room temperature used to remove excess solvent from a sample of saturated solution also may have decomposed any solvate species. In the case of sodium acetate and potassium bromide, residue analysis corresponded approximately to NaCZH3OZ~ l. 5 TMG and KBr- 3. 5 TMG. However, the precision of analysis for tetramethylguanidine in these residues was not good enough 53 Table VIII. Comparison of Solubilities (g/100 g) in Various Solvents at 25 H20 NH3 en TMG NaClO4 209. 6 30.1 52. NaNCS 142.6 205.5 93.5 31. Nal 183.7 161.9 34.6 21. NaNO3 91.5 97.6 33.5 1.4 NaCl 35.98 3.02 0.33 0.0041 54 to determine the above formulas with certainty. In addition, it seems unreasonable that these two salts would form solvates with tetramethyl- guanidine while all the others listed in Table VII do not. If this were the case, sodium acetate and potassium bromide should possess en— hanced solubilities in tetramethylguanidine, but they do not appear to be out of their expected order in the table. A definite tetramethylguanidine solvate stable at room tempera- ture and above appears to be formed, however, by lithium chloride. Although only semiquantitative solubility measurements on this salt (S = ~10 g LiCl/lOO g TMG) were made, the residue analysis corres- ponded quite well to the formula of a hemisolvate, LiCl- 0. 5 TMG. In addition, no weight loss nor change in analysis occurred on heating . . 0 this reeldue to 60 under vacuum. B . Semiquantitative Solubilitie s Semiquantitative solubilities of a large number of compounds were obtained generally by single equilibrations followed by the weight- residue method of analysis or indirectly from attempts to prepare stock solutions as concentrated as possible. Table IX categorizes these com- pounds from the very soluble to the relatively insoluble with their approxi- mate solubility ranges in terms of grams of solute per 100 grams solvent. As mentioned below in the section on reactions, the solubilities of a number of the compounds is no doubt enhanced by reaction with the solvent. For instance, ammonium salts undergo solvolysis, transition metal salts form complexes with tetramethylguanidine, KZCrzO7 is slowly re— duced at room temperature, presumably to some Cr(III) species (a green solution was formed), etc. When a group of related compounds, E'E" salts of the same cation or same anion, are listed in Table IX, they are given in order of decreasing solubility. In other words, NH4NCS is more soluble than NH4ClO4, etc., and LiNCS is more soluble than LiNO3, etc. However, 55 Table IX. Semiquantitative Solubilities in l, 1, 3, 3-Tetramethylguanidine Very Soluble (~> 10 g/100 g TMG): NH4NCS, NI-I4C104, NI—I4NO3, NI—l4l, LiNCS, LiNO3, 1510104, Lil, LiBr, LiCl, KNCS, Mg(ClO4)2, Ca(NO3)Z, Pb(NO3)Z, Co(CzH3OZ)z-4HZO, Cu(CZH3OZ)z-HZO, Cr(CzH3OZ)3-HZO Moderately Soluble (—- l — 10 g/100 g TMG): NH4C2H3OZ, Nail—1503, KI, chrzo7, SI‘(NO3)2, COC12'6H20, COCIZ, CUzclz, CUNCS, NI(C2H3OZ)Z'4H20, TMG'HBT, B'CH3‘C6H4'SO3H Slightly Soluble (~ 0.1 - 1 g/100 g TMG): NH4Br, NaZS, NaNOZ, KCIO4, KNCO, H3BO3, Ba(NO3)Z, NaOCl—I3, NaOCZHS, (C61-15)4AsC1, (C6H5)4AsClO4, TMG-HGl, (CI-13).,NOH, Co(NO3)z-6HZO, CoSO4-7HZO, Co(C104)2-6HZO, Cu(NCS)Z, citric acid monohydrate Relatively Insoluble (“ < 0.1 g/100 g TMG): (NH4)ZSZOS, NH4C1, NH,Hso,, (NI-1.92504, (NH4)ZCO3, NH4F, LiCzH3OZ, 1.12504, LiF, LiOH, LiZCO3, Lizczo,, NaOl—l, NaHCO3, KNO3, KC3H3OZ, K2804. Ca(OH)2, Ba(OH)Z, BaCO3, BaO, maleic acid 56 this does not neccessarily mean that the ammonium salts listed in the first line of the table are all more soluble than the lithium salts which follow in the second line. The very soluble ammonium salts (the thio- cyanate, perchlorate, nitrate and iodide) have appreciably higher solu- bilities than the corresponding alkali metal salts while the reverse appears true for the bromides and chlorides. For the latter halides, the lithium salts are the most soluble, followed approximately by sodium, then the ammonium and potassium salts. These anomalies are likely due to a combination of opposing factors such as the high reactivity by solvolysis of NH4+ salts, the high polarizability of the large anions in- volved, and the high charge density, or solvation energy of small cations such as the lithium ion. For the above reasons, therefore, the order of solubilities of the alkali metal and ammonium salts would have to list the ammonium and lithium salts as approximately equal, iii" NH4‘V Li > Na > K. In the case of the anions of these monopositive cations, the order of solubilities is approximately NCS > C104 > N03 > I > Br > CZH3OZ > C1 > $04, with some shifting of this order for each of the cations. An examination of Table 1X leads to a number of additional general comments: As one would expect, the order of solubility of the alkaline earth metal salts is Ca > Sr > Ba, although only the nitrates of these were measured. Of the hydrated transition metal salts, the acetates are the most soluble, followed by the chlorides. Perhaps surprisingly, the hydrated nitrate and perchlorate salts of cobalt are no more soluble than the sulfate. The results obtained in this case, however, may be non- equilibrium solubilities. For instance, in the preparation of stock solu— tions of Co(ClO4)z- 6HZO in tetramethylguanidine it was never possible to dissolve all the solute at any of the three concentrations 10'3, 10‘2, or 10'1 molar. Apparently the resulting cobalt perchlorate-tetramethyl- guanidine complex, or the solvent itself, forms a coating on the solute 57 particles which prevents further dissolution. Since the anhydrous cobalt chloride has about the same solubility as the hexahydrate, it may be that the solvent displaces all of the water molecules from their positions around the Co++ ion. Further elaboration of this supposition will be made in the section on transition metal complexes. For titrations in tetramethylguanidine the only acids found to have sufficiently high solubilities for use as standards or titrants were tetramethylguanidinium bromide and p—toluenesulfonic acid. Possible bases such as sodium methoxide, sodium ethoxide, and tetramethyl— ammonium hydroxide were not sufficiently soluble for use in titrations. The slight solubilities of the above three bases and of chloride and perchlorate salts of the tetraphenylarsonium ion indicate that tetramethyl— guanidine does not tend to solvate either large organic cations or anions very well. Since the salts of the ammonium and lithium ions tend to be the most soluble, their salts listed in the relatively insoluble group of Table IX give a good indication of the inorganic anions which confer low solubility. Thus, probably all salts of peroxydisulfate, bisulfate, sulfate, fluoride, hydroxide, carbonate, and oxalate are quite insoluble. The usefulness of this list may lie in its applicability to synthetic chemistry in tetramethylguanidine as a solvent where the precipitation of an insoluble salt is used to drive a reaction to completion. Barium oxide is negligibly soluble in tetramethylguanidine. It therefore found use in this work as a drying agent since it, as well as Ba(OH)Z or BaCO3, formed on reaction of the oxide with water or carbon dioxide, respectively, are all insoluble. The low solubilities of citric and maleic acid may simply be an indication of the slight degree of dissolving power tetra- methylguanidine has for large organic polyprotic acids. 58 C. Solubility and Miscibility of Orianic Compounds In the acid-base titration studies discussed later, numerous organic acids were readily dissolved in tetramethylguanidine. However, since it was most practical in these studies to titrate only dilute solu- tions of the acids (NO. 01 molar), their solubilities were not reached and thus were not measured. It can only be stated that their solubilities are greater than approximately 0. l g/100 g of tetramethylguanidine. Some of the compounds in question are listed below in Tables XVIII and XIX. Since these compounds dissolved readily and quite rapidly in tetramethylguanidine, it is felt that their actual solubilities may be fairly high and that many other organic compounds should also be very soluble. Another fact that points to the high dissolving power of tetramethyl- guanidine for organic compounds is its complete miscibility with many of the common organic solvents. Table X lists the solvents tested for miscibility with tetramethylguanidine as a sampling of various classes of organic liquids. The table also indicates that some of the solvents undergo exothermic reaction with tetramethylguanidine, but the possible nature of these reactions will be discussed in the next section. The results in the table do show, however, that tetramethylguanidine is completely miscible with all common organic solvents, with the possible exception of saturated hydrocarbons and compounds containing long alkane groups. Apparently a number of characteristics of tetramethyl- guanidine contribute to its miscibility with so many liquids: the reactions noted in the table, the tendency for hydrogen-bonding to occur, and the symmetry and similarity in structure between the outer sheath of CH3-groups in tetramethylguanidine and aliphatic or aromatic compounds. Table X. Miscibility of l, 1, 3, 3-Tetramethylguanidine with Organic 59 Liquids Liquid Miscibility Visible Reaction Acetone Complete None Carbon tetrachloride Complete Slow Dichloromethane Complete Slow p- Dioxane Complete None Methanol Complete Some heat evolution Benzene Complete None Pyridine Complete None Cyclohexanol Complete Slight heat evolution Ethyl acetate Complete None Acetic anhydride Complete Much heat evolution N, N—Dimethylformamide Complete None Ethyl Cellosolve Complete Some heat evolution Petroleum ether Complete None Cyclohexane Complete None n—Pentane Partial None 2- Butyl ether Partial None Tetramethylurea Complete None Nitrobenzene Complete None Ethylenediamine Complete None Acetonitrile Complete None Tetrahydrofuran Complete None Ethyl ether Complete None 60 Reactions of 1, 1, 3, 3-Tetramethylguanidine One of the problems associated with the use of l, l, 3, 3-tetra- methylguanidine as a nonaqueous solvent is its reactivity with a number of compounds. Therefore, to more fully understand these problems and to circumvent them where possible, a number of reactions were studied. A. Reaction with Water The end product of the hydrolysis of tetramethylguanidine is 1, l—dimethylurea which slowly precipitates out of a tetramethyl- guanidine-water solution at ordinary temperatures but which forms rapidly as an insoluble residue on evaporation of a similar solution. The same product, 1, l-dimethylurea, forms as a suspension of finely divided particles on addition of a small volume of tetramethylguanidine to a large volume of water (volume ratio 1 TMG24O H20). 1, 1— Dimethyl— urea was identified by elemental analysis, melting point, mixed melting point, and infrared spectra. The over-all hydrolysis reaction can perhaps be written: NH (H3C)ZN — 6 - N(CH3)Z + H20 —> (H3C)2N - - NH2 + I-lN(CI-l3)z- (1_z_3_) -0 Although dimethylamine was not identified, it is well—known that amidines are easily hydrolyzed to carbonyl compounds (56) and that guanidine is hydrolyzed in basic solution to urea and ammonia. Thus, dimethylurea and dimethylamine appear to be reasonable tetramethylguanidine hydrolysis products. Attempts to prepare tetramethylurea by the reaction of tetramethyl- guanidine and water gave only qualitative evidence for its formation. A stoichiometric mixture of tetramethylguanidine and water was refluxed for two hours and then distilled. Although only a small amount of distil- late was obtained, considerable white solid collected in the air—cooled condenser. This solid was deliquescent. When heated at 1100 for one 61 hour, it changed to a liquid having the very characteristic odor of tetramethylurea. Perhaps the white hygroscopic solid was tetra- methylguanidinium hydroxide, which has never been isolated (8b). The hygroscopicity of this solid is similar to that of other tetramethyl- guanidinium salts prepared in this work and described below. Heating the aqueous solution may have eliminated ammonia to form tetramethyl- urea, as indicated in Equation _1_9: + _ HalfiIOH fl (H3C)ZN - c - N(CH3)Z %ZBOo—>- (H3C)2N — c — N(CH3)2 + NH3 4‘ (1_9) If this assumption is true, the mechanism of tetramethylguanidine hydrolysis may be more complicated than is indicated by Equation (E). A simpler mechanism of hydrolysis may involve addition of water to the C=N group in tetramethylguanidine followed by a rearrangement of the intermediate to l, l-dimethylurea and dimethylamine: H (H3C)ZN-E — N(CH3)2 + HOH ————>~ NH_2 _________ NHZ (H3C)2N'gi -N(CH3)Z E—-—-> (H3C)ZN - (I: + HN(CH3)Z :5 _________ I 65 (Q) B. Reaction with Carbon Dioxide Basic nonaqueous solvents readily absorb carbon dioxide to form carbonates, carbamates, etc. (57). Only the Bicarbonate salt, on the other hand, has been identified in the reaction of tetramethylguanidine with moist carbon dioxide (see below): + - NH HZNHCO3 (H3C)ZN — c — N(CH,)Z + coZ + 1420 —>- (H3C)ZN - ('i - N(Cl—13)z <2_1) The need for water in the reaction was determined by measuring the relative weight increases on addition of carbon dioxide to tetramethyl— guanidine with or without added water. These data are plotted in 62 Figure 13; the upper curve shows the weight increase with time when dry carbon dioxide was bubbled into a 1:1 mole ratio of tetramethyl- guanidine and water. The lower curve is that for an equal amount of anhydrous tetramethylguanidine. It was possible to obtain 92 percent of the theoretical weight increase with the tetramethylguanidine:water mixture but less than two percent increase with anhydrous tetramethyl- guanidine. Evaporation probably caused the yield of bicarbonate to be less than theoretical in the first case. A constant weight increase of l. 9 :1; 0.1% was obtained for carbon dioxide addition to various samples of tetramethylguanidine dried by different methods. It may be, there- fore, that carbon dioxide is soluble to this extent in tetramethylguanidine, i_.g., equivalent to a 1. 9 percent weight increase. Another possibility would be that a like amount of the hydroxide, TMGH+OH-, was present and unremovable from the tetramethylguanidine by any dehydrating methods and simply reacted with the carbon dioxide to form the bi- carbonate. Because there is little or no difference in the conductance of pure tetramethylguanidine and of solvent saturated with tetramethylguanidin- ium bicarbonate, it is concluded that tetramethylguanidinium bicarbonate is essentially insoluble. Addition of carbon dioxide therefore could be used to remove water from tetramethylguanidine. Some evidence for the identity of the bicarbonate salt was obtained by titration of its aqueous solution with acid. The shape of a tip: .4. ,“ré titration curve (Figure 14) corresponds to that of sodium bicarbcnaie. Elemental analysis supports the formulation TMGH+HCO3I reasonasl; well. Calculated: c, 40.7; H, 8. 53; N, 23.7% Found: c, 38.5; H, 8.70; N, 22.2% It is possible that the analytical sample became wet because of the hygro- scopic nature of the salt. The observed composition corresponds to 63 ON .oc«p«:§m~>£uo5duuo.fitm .m J J >3 coeumu0m3< QEXOHQ conumU m. V. mu»:z_a.ua.» o_ w v \ W o~xa o: 141d .\ Ia Ll 1‘ "lVIII >¢Ouzh mo A«$00. p .2 223$ O NO (.0 0.0 .LH9I3M suvuo ‘3sv3u3NI 64 .oumdonudoflm EdagpmomaamlfioEMHuohtm .m J J mo :oCmHfiH .vd ouswfih moz: 1 06.0 .x a. w r. c n N _ o neuzoz mooz+zoze IA o n e n u . o moz: H 06.0 .2 65 TMGH+Hco,'-o.6 H20 (Calculated: c, 38.3; H, 8.68; N, 22.4%). Additional physical properties of the bicarbonate are reported and discussed below. C. Reaction with Acids A strong base like 1, l, 3, 3-tetramethy1guanidine should react vigorously with strong acids to form TMGH+ salts. (No evidence was found in this work for TMGHZ++ salts.) The reactions proved indeed to be vigorous and strongly exothermic, but in some cases no salt could be isolated. The chloride, bromide, and acetate precipitated when formed by addition of aqueous acids to tetramethylguanidine. Reaction of gaseous HCl or HBr with tetramethylguanidine also led to precipitation of the halide salt. These white, somewhat hygroscopic solids were dried by vacuum desiccation over sulfuric acid, and their melting points were determined (Table XI). These salts are not reported in the literature. Tetramethylguanidinium iodide is reported to melt at 1200 (11) and to be very hygroscopic. The chloride and bromide were analyzed by aqueous argentometric titration and found to be 1:1 TMGzHX salts. C5H14N3Cl, calculated: Cl, 23.37%; found: Cl, 23.13%. C5H14N3Br, calculated: Br, 40. 75%; found: Br, 40.40%. No insoluble products formed on reaction of tetramethylguanidine with such concentrated aqueous acid solutions as 98% H2804, 72% HClO4, or 48% HF. Decomposition apparently occurred on addition of concen- trated sulfuric acid to tetramethylguanidine as indicated by charring. No solid product was formed even on long standing of 1:1 base2acid mixtures of tetramethylguanidine and concentrated perchloric or hydro- fluoric acid. It is felt, however, that the 504—: HSO4_, C104: or F- salts could be prepared by other methods if desired. 66 Table XI. Melting Points of Salts of 1, l, 3, 3-Tetramethylguanidine + , _ o TMGH Salt Melting Pomt, C of 211—212 Br- 184-185 HCO,’ 110-113 (321—1302: 90-93 67 D. Spectra and Structure of Protonated l, 1, 3, 3-Tetramethjlguanidine Since tetramethylguanidine apparently undergoes only mono- protonation (11, this work), it is of interest to identify the basic site. Although it is not certain whether protonation of ureas and amides occurs on the oxygen or the nitrogen (58, 59), it appears that guanidines protonate predominantly at the imine nitrogen (2 5), with the resultant positive charge localized on the central carbon atom (23, 24). In this study, the infrared spectrum of 1, 1, 3, 3-tetramethyl- guanidinium bromide (KBr pellet) was recorded (Figure 15). Compari- son of this spectrum of protonated tetramethylguanidine with that for the free base (Figure 10) indicates that some frequency shifts occur on protonation. Most notable is that attributed to the N-H stretching vibration at 3311 cm’1 in tetramethylguanidine which is no longer found in tetramethylguanidinium bromide. The explanation may lie, therefore, in protonation of the :NH group to either an -NH2+ or a :C+-NHZ group. The latter localized charge structure is analogous to that for the guanidinium ion. From the tables of Nakanishi (51) it is found that the -NH?‘+ group absorbs in the “ammonium band" region from 2250-2700 cm'1 while the trialninocarbonium ion has a broad band at approximately 3300 cm“1 in the free amino region. The spectrum of TMGH+Br- in Figure 15 shows a broad band from about 27 50-3250 cm‘l, which includes, of course, the -CH3 stretching absorptions, with no bands in the 2250- 2700 cm'1 region. Thus, the structure of the tetramethylguanidinium ion can perhaps best be represented as (H3C)ZN\+ / C - N\ (H3C)ZN/ H . Assuming that the NH absorption bands of tetramethylguanidine are the only ones which change appreciably on protonation, some tentative 68 Q. .oEEoum Edfiodgcdamanuofiawuaofium .m J J mo @30on ponds»: 1.152351, N. o. a o . e .3 ooami aouvaaosev —+ 69 band assignments in the spectrum of the free base (Figure 10) can be made. A strong band at 1493 cm‘1 which disappears may therefore be assigned to an NH bending vibration and the broad band from about 765-800 cm"1 may be due to an NH rocking deformation. Neither tetramethylguanidine or any of its salts were found to absorb in the visible region of the spectrum, but small shoulder peaks were found in the ultraviolet below 300 mu. However, these peaks are too small (Emaxj‘v/ 1—10) to be useful in quantitative spectrophotometric analysis. E. Titration of l, l, 3, 3-Tetramethylguanidine with Acid Solutions l, 1, 3, 3-Tetramethylguanidine can be titrated potentiometrically with standard acid solutions in either aqueous or nonaqueous media using a glass-calomel electrode system. A typical titration of tetramethylguanidine in water solution with aqueous standard acid is shown in Figure 16. In this case, 1. 00 ml of tetramethylguanidine was dissolved in water and titrated with 0.1027 1\_/I HClO4 using a Beckrnann Model G pH meter. Tetramethylguanidine titrates like a strong mono- protonic base with one steep inflection in the curve. Over a series of six titrations, the end points were constant within i 2% of 1:1 stoichiometry. Titrations of tetramethylguanidine in water as a solvent must be done with a freshly prepared solution to avoid hydrolysis (see Equation l_8). Evidence supporting Equation l_8, and especially dimethylamine as a product, is found in the titration of a tetramethylguanidine-water solution after allowing it to stand for several hours. In this case a second break at pH — 5 is obtained which can be attributed to the presence of dimethylamine. Additional evidence for only monoprotonation of 1, 1, 3, 3-tetra- methylguanidine was obtained by potentiometric titrations in nonaqueous 70 .33.. E Sodom £3 :ofidfiom msooDU/w 5 ocfigcodmanuogmhohtm .m .E .H mo :ofleufih 4: oudmfih . ¢o_ozfl~.~o_.o .2 Nn om v.0. o~ o. N. w v o o l '/ N A c A w x 3 tn o I! o. N. 71 media. Two typical titrations are shown in curves 1 and 2 of Figure 17. Curve 1 is the titration in glacial acetic acid with perchloric acid as titrant; curve 2 is similar except for use ofacetonitrile as a solvent. In each case, the steep potentiometric inflection is only about two per- cent lower than the calculated l:l acidzbase ratio. Addition of excess acid in each of these titrations gave no second inflections in the titration curves. Perhaps even better evidence for tetramethylguanidine undergoing only monoprotonation is the titration of pure, undiluted tetramethyl— guanidine with 4.0 M aqueous HCl (Figure 17, curve 3). Again, there is only one potentiometric inflection with the measured end point about one percent below calculated 1:1 stoichiometry. Initially in this titration, TMGH+C1_ precipitated, but on further addition of titrant sufficient water was present to keep the salt in solution. F. Qualitative Study of Miscellaneous Reactions of l, 1, 3, 3—Tetramethylguanidine A number of other substances were found (Table XII) to react with tetramethylguanidine but in general were studied only on a qualitative or semiquantitative basis. Some gas evolution occurs on addition of metallic sodium or potassium to tetramethylguanidine but even in the presence of a large excess of solvent, the alkali metal was not consumed. Although the evolved gas was not identified, there appeared to be little or no reaction. Lithium metal, lithium amide, and sodium amide, on the other hand, decompose the solvent to give unknown red-brown and orange solid products. Sodium amide was also treated with tetramethylguanidine in liquid ammonia with similar results. Little evidence is found in the literature for alkali metal guanidide salts. The one reference found, that of a 1927 British patent (60), claims the formation of the mono- and di-sodium compounds of urea and of guanidine by reaction with VOLTS Emf. 72 0.0 IN HOAc 0.6 2 8.. J. 0.2 IN CHSCN O NO DILUENT ‘0.2 1 —0.6 A0010 ACID ——-> Figure 17. Potentiometric Titrations of 1, l, 3, 3-Tetramethyl- guanidine. 73 Table XII. Miscellaneous Reactions of l, 1, 3, 3-Tetramethy1guanidine Reactant Results, products Na; K No reaction ('3) Li; LiNHZ; NaNHZ Decomposition, orange solid LiA1H4 Vigorous reaction LiH; CaHZ Slow reaction KMnO4 KZMnO4(?) + MnOZ (vigorous, pyrophoric) KZCrZO7 Cr+3 (slow) NaII_O4 Yellow solution Ag ++ Agg ++ ng+ Hg + Hg NH4 NH3 + _ CC14; CHC13; CHZCIZ TMGH Cl (:52 Decomposition; orange solid 74 sodiuln hydride, but the work has apparently never been substantiated. The reactions of tetramethylguanidine with the hydrides listed in the table again gave gas evolution but no products were isolated. The reaction of tetramethylguanidine with KMnO4 is pyrophoric if the permanganate is finely ground. A mixture of green and brown solids is formed which suggested the reduction of MnO4- to MnO4-- and MnOz. Potassium dichromate also oxidized tetramethylguanidine as indicated by the resultant green (Cr+3) solution, but the reaction is quite slow at room temperature. Sodium periodate apparently oxidizes tetramethyl- guanidine and may itself be reduced to 13 as evidenced by the yellow solution which is formed. Silver(I) salts, if soluble in tetramethylguanidine, are reduced to free silver, especially when heated. On first addition of a solution such as aqueous silver nitrate to tetramethylguanidine, however, a brown precipitate resembling silver hydroxide is formed. When heated, this mixture forms a silver mirror on the walls of the container. The disproportionation of Hg(l) in tetramethylguanidine to Hg and Hg(II) is not surprising since a similar disproportionation occurs in the case of mercury(I) chloride in the analogous solvent, ethylenediamine (61). Ammonium salts on dissolution in tetramethylguanidine invariably evolve ammonia gas. This shows that tetramethylguanidine is a stronger base than ammonia and that the ammonium ion acts as an acid in tetra- m ethylguanidine: + + TMG + NH4 —?~ TMGH + NH3. (2 ) However, the presence of the solvo—cation, TMGH+, was not directly demonstrated in any of the ammonium salt solutions in tetramethyl- guanidine. In fact, if the solvo-cation does form, one should obtain similar solubilities for the NH,+ and TMGH+ salts of the same anion. This, however, does not hold true in this work where it was found that 75 the solubilities of NH4Br and NH4Cl were about an order of magnitude greater than the solubilities of corresponding TMGH+ salts. These solubility differences are essentially unexplainable but it must be re- called that the solubility measurements in question here were made only semiquantitatively. Reactions analogous to that of tetramethylguanidine with chlorinated hydrocarbons to form tetramethylguanidinium chloride are also found with other basic nonaqueous solvents. For instance, ammonium chloride is one of the products in the reaction of anhydrous ammonia with di- chloromethane (62). It is also known (63) that such organic chlorides as chloroform, carbon tetrachloride, ethylene dichloride, butyl chloride, and dichloroethyl ether react at 1000 C with an excess of ethylene— diamine to form secondary amines and amine hydrochlorides. The decomposition reaction of carbon disulfide with tetramethyl- guanidine is vigorous and exothermic. For instance, on addition of carbon disulfide to tetramethylguanidine, an immediate reaction takes place in which a yellow gas is evolved, the tetramethylguanidine becomes yellow-red, and a red—orange solid is formed. The complexity of this reaction is obvious and thus was not investigated further. The reactions of 1, 1,3, 3—tetramethylguanidine which were studied qualitatively and listed in Table XII can perhaps best be summarized in terms of some of the problems which they cause in the study of tetramethylguanidine as a solvent. Thus, it was not. possible to prepare alkali metal salts of tetramethylguanidine by reaction with the elements, their amides, hydrides, etc. If any of these reactions had been successful, the strongest possible base in the solvent would, of course, have been prepared. The reactions of tetramethylguanidine with permanganate, dichromate, periodate, and Silver(I) indicate that the solvent can be oxidized readily. The reduction of Ag+ to free silver by tetramethylguanidine precludes the use of silver salts in titrations, 76 syntheses, etc. in tetramethylguanidine. The reaction of tetramethyl- guanidine with Hg(I) as well as Ag(l) could also cause problems in the use of salts of these species in construction of electrodes for use in the solvent. The reaction of tetramethylguanidine with carbon tetrachloride, carbon disulfide, etc. rule out their use as spectrophotometric solvents in the presence of tetramethylguanidine. Transition Metal Complexes of 1, 1, 3, 3-Tetramethylguanidine A. Isolation and Analysis of Complexes As has been mentioned previously and shown in Table IX, it was found that a number of the common transition metal salts, both hydrated and anhydrous, exhibit appreciable solubility in tetramethyl— guanidine. This was especially true with the hydrated acetate salts of Co(II), Cu(II), and Cr(III), which all formed highly colored solutions in the solvent. Therefore, in order to identify the complex species giving rise to the colors, various attempts were made to isolate solid material from tetramethylguanidine solutions of these metal acetates and of some other transition metal salts listed in Table IX. Isolation of tetra- methylguanidine complexes was attempted by solvent evaporation, addition of another solvent to precipitate an insoluble complex, addition of another solvent to extract a soluble complex, etc. However, all these methods were generally unsuccessful. Some limited success was achieved by evaporation of excess solvent. in a vacuum oven at tempera- tures 10-200 below the predetermined atmospheric melting points of the products ultimately obtained. The resulting solids were ground in a mortar under anhydrous conditions and were analyzed commercially. In this manner, the following tetramethylguanidine "complexes" (or solvates) of some of the transition metal acetates were isolated: Co(C2H3OZ)2c 0, 33TMG; Ni(C3HSOZ)3- O. 5TMG; and Cr(C3H3OZ)3° 1.0TMG. No ”complex” of copper(II) acetate could be isolated due. to 77 decomposition below its melting point. Analytical and other data for the materials which were obtained are given in Table XIII. Except for the analytical percentages for carbon, the observed analyses correspond to the given formulas reasonably well. Analytical discrepancies prob- ably can be ascribed to uncertainties in the amount of solvent removal in the isolation of these compounds. B. Structure of Complexes In addition, and subsequent to the above studies, the more definite complexes of tetramethylguanidine mentioned in the historical section of this thesis were prepared and analyzed. Thus, in work in this laboratory and elsewhere (29, 30, 31) , the following complexes have been reported: [Co(TMG1411CIO4lzi [Cu 0. l M) of cobalt(II) acetate tetrahydrate in tetramethyl- guanidine, the intense color “max” 500) of coba1t(II) solutions, etc. Other metal acetate hydrates were not as soluble as the cobalt salt and did not produce as intense a color. For instance, for Cu(C2H3OZ)2onO, emaxxj75. Spectra of tetramethylguanidine solutions of the latter salt were studied by Kennedy (30) in this laboratory in cooperation with the author. She found that the spectra of copper(ll) acetate solutions were Gaussian-shaped with xmax = 680 mu and that deviations from Beer' 8 law between 0. 002 M and 0.008 M were positive. An additional argument favoring the selection of the cobalt(Il) ion for spectral study in tetramethylguanidine is that it probably has been investigated more than any other system in other nonaqueous solvents. Notable among these investigations is the work of Katzin and co-workers (64a-d) dealing with the absorption spectra of cobalt(II) salts in alcohols and acetone. Other studies include the series of papers by Libus, e__t a_1. (65a-c) using alcohols and acetonitrile as solvents for cobalt complexes as well as the spectral studies by Buffagni and Dunn (66) of the CoClz and CoC14_— species in nitromethane, N, N—dimethylformamide, and other solvents. Considerable suc7ess has been achieved by the above authors in the characterization of various complex species in these solvents by visible absorption spectroscopy. D. Effect of Anions on the Spectrum of Cobalt(II) Acetate Absorption spectra in tetramethylguanidine of several hydrated cobalt(ll) salts Show a maximum around 600 mu (Figure 18). ABSORBANCE 80 C0 (SCle ‘ 3.120 COCIZ ‘ 6.120 ,Co(OAc)2~4HzO C0 (CI04)2‘ 6.120 \ / / ‘ 500 550 600 550 700 WAVELENGTH , my. Figure 18. Visible Absorption Spectra of Hydrated Cobaltfll) Salts in l, l, 3, 3-Tetramethylguanidine. 81 The similarities in these spectra appear to indicate that tetramethyl- guanidine forms a stable complex with coba1t( II) which is little in- fluenced by the particular anion present. Further illustration of this ‘ is given in Table XIV which shows the rather small wavelength and absorbance changes in the spectrum of 10'3 l\_/I Co(C2H3Oz)Z~4HzO solution in tetramethylguanidine on addition of a ten-fold excess of an anion. For solubility reasons, the anions were added as their lithium salts except for acetate ion, in which case ammonium acetate was used. Anions are listed in Table XIV in order of decreasing effect on the position of the absorption maximum. E. Effect of Water on the Spectrum of Cobalt(ll) Acetate The effect of water of hydration on the spectra of the cobalt(ll) salts was determined by recording the spectra of the anhydrous acetate and chloride in tetramethylguanidine. These spectra are shown in Figure 19, and a comparison with the spectra of the corresponding hydrated salts in Figure 18 shows no difference in curve shape. There is, however, a small shift of the maxima (< 5 mp) to lower wavelengths in the case of the anhydrous salts and an increase in absorbance. In contrast to water of hydration, large amounts of added water greatly alter the absorbance in tetramethylguanidine of cobalt(ll) salts. Figure 20 illustrates this effect for a 10'3 I\_/_i solution of Co(C2H3 Oz)2' 4HZO. From the figure it is seen that something above three percent H30 is necessary to change significantly the absorption spectrum; at nine per- cent H20 the absorbance has decreased by nearly one-half and the curve shape has changed; finally, at 25 per cent H20 there was zero absorbance and the solution was colorless. Because of the known hydrolysis of tetramethylguanidine, it is not surprising that the 25 percent HZO solu- tion was colorless since this is more than enough water to hydrolyze all the tetramethylguanidine present. 82 Table XIV. Effect of Anions in a Tenfold Excess on the 600 mp. Absorption of Co(C2H3 Oz)z-4HZO in Tetramethylguanidine Anion Axmax AAmax C1- +15 mu +0.147 NO3— +13 +0.14? Br- +12 +0. 186 C10.“ +11 +0. 030 SCN' + 3 +0.128 camoz' —10 -0.016 ABSORBANCE 83 9 / Co (OAc)2 500 Figure 19. 550 600 650 700 WAVELENGTH, mlu Visible Absorption Spectra of Anhydrous Cobalt(II) Salts in l, 1, 3, 3-Tetramethylguanidine. ABSORBANCE 84 0/0 "20 a o b 3 a c 9 d 25 b C A 475 500 550 600 700 WAVELENGTH, my. Figure 20. Effect of Water on the Absorption Spectrum Of CO(C2H302)3' 41—110. 85 F. Beer's Law Study of Cobalt(II) Salts A Beer's law study of cobalt(II) salts in tetramethylguanidine was made to obtain information on solvent-solute interaction. Hydrated and anhydrous cobalt salts were investigated over concentration ranges of approximately 10‘4 to 10'3 IXI. Only in the case of O to 5 x 10-4 M anhydrous CoCIz did any of the salts obey Beer's law. The Beer's law plot for anhydrous cobalt chloride is shown in Figure 21, at a wave- length near the main peak at 600 mu and near the shoulder at 550 mu. Deviation may occur at concentrations higher than 5 x 10’4 I\_/_f CoClz. A typical Beer' 3 law plot for other cobalt(II) salts in tetramethyl- guanidine, all of which showed deviation, is given in Figure 22. At 590 mu, Beer's law appears to be obeyed up to ”4 x 10'4 IV_1 Co(CzI-I3Oz)z, with negative deviations at higher concentrations. However, at wave- lengths on either side of the main peak, 550 and 620 mu, the deviations occur over the entire concentration range. Thus the Beer‘s law studies are a good indication of reaction between most coba1t(II) salts and tetra- m ethylguanidine . G. Possible Cobalt(II) Complex Species in Tetramethyl- Eanidine Solution Since the nature of the anion present has some influence on the absorption spectrum of coba1t(II) in tetramethylguanidine (Table XIV), it was decided to study this effect in greater detail. Iodide ion was selected for study because of the high solubility of NaI (Table VII). The effect of increasing concentrations of iodide ion on the spectrum of Co(CzH3OZ)2'4HZO and CoClz' 61—120 solutions in tetramethylguanidine is shown in Figures 23 and 24. The figures show that: (I) there are no major changes in the spectrum of either cobalt salt even at a thousandfold excess of iodide ion, (2) all spectra have a left-shoulder inflection near 550 mu but only the cobalt chloride spectrum at 0 and 10‘3 l\_/I I- shows a right-shoulder ABSORBANCE 86 06 Q4 soomi/ 550mg / / o‘ 1 2 3 MOLES 1" x 104 Figure 21. Beer‘s Law Plot for Anhydrous CoClz Solutions. ABSORBANCE 87 v I.2 I.0 / 590ml; /’ 0.8 / / 620M}; Q6 4/’/1 550 me 0.4 0.2 O 0 2 4 6 8 I0 M0LE$i"xuo‘ Figure 22. Beer‘s Law Plot for Anhydrous Co(C2H3O;)z Solutions . 88 I I Co(OAc)2 -4H20 = 9.5 x 10-4 g CURVES No I I 0M 6 2 10-4 3 4 x 10‘3 5 4 IO'2 5 IO" 6 (1.4 3 4 I m o z < m n: 0 g ‘— < w J /4 2 l \\ N h 475 500 550 600 650 700 WAVELENGTH , m,“ Figure 2.3. Effect of Iodide Ion on the Spectrum of CO(C1H302)3' 4Hzo. ABSORBANCE 89 CURVE NoI CoCl2~6H20 = 5 x 10“ fl _1_ 0T 2 10-3 3 IO'2 4 5x10'2 5 5:: IO" 5% ___// WAVELENGTH,mp Figure 24. Effect of Iodide Ion on the Spectrum of CoClz- 6HZO. 90 inflection near 650 mg, and (3) there are no isosbestic points in either series of spectra. The latter fact indicates that these systems may be quite complicated, e.g., there may be three or more complex species in equilibrium with each other. These complications are perhaps not surprising when one considers the number of possible ligands in either of the systems: TMG, I_, H20, and Cl~ or Cal-130;. It is unfortunate that low solubilities in tetramethylguanidine limited these studies at the time to so few possible Co(Il) saltzadded salt combinations. Hindsight now indicates, however, that solubilities may have been sufficiently large for analogous spectral studies of such systems as anhydrous cobalt chloride plus lithium chloride or anhydrous cobalt iodide plus sodium iodide. Conversely, inorganic halides such as lithium chloride prob- ably form ion pairs (66) in a low dielectric-constant solvent like tetra- methylguanidine. To avoid ion-pairing, chloride ion can be added as the salt of a large organic cation such as tetramethylammonium- or tetraphenylarsonium chloride, but these compounds also possess low solubility in tetramethylguanidine (Table IX). Despite the above mentioned difficulties in interpretation of such spectra as those of Figures 23 and 24, it is felt that some additional comments can be made in regard to possible cobalt(II) complex species present in tetramethylguanidine solution. There is little doubt that only tetrahedral, or four-coordinate, complexes of cobalt(II) are involved; all cobalt(II) complexes exhibiting strong absorption bands (6 = 102 to 103) in the spectral region above about 550 mp are those of four— coordinated cobalt (67, 68). Longhi and Drago (2.9) have published the visible absorption spectrum of the complex salt, [Co(TMG)4](CIO4)Z as obtained in dichloro- methane solution and by solid reflectance (Figure 25). There is con- ]+ siderable similarity between their [Co(TMG)4 + spectra and those Observed in this work (Figures 18-20, 23, 24). Small differences in 91 8.0 o a. . 0 s . o . o . s . s . O 6.0 e ° b s O u o o z 0 < m o a: 8 . m < ’ . m 4.0 > ° 0 - s ’— s S s .1 I.“ . O a: o s . o . o 0 ° C ~ 2.0 ‘ i, ' o /‘—” / VI 4 \\ \ \~—’ ‘ \-‘ O 400 500 600 700 * WAVELENGTH,my. Figure 25. b. c. 1.65 x 10-314 in calmz Visible Absorption Spectra of [Co(TMGh} (ClOJZ. a. Solid Reflectance 1.0 x10‘2 M in calm2 92 xmax values may be instrumental errors; the fact that Drago obtained a second maximum near 530 mp as opposed to the shoulder inflections of this work may be due to the solvent effect of dichloromethane on the one hand and tetramethylguanidine on the other. In the absence of a solvent (the reflectance spectrum of solid [Co(TMG)4](ClO4)2, Figure 25), only a shoulder inflection is found near 550 mp. It is felt that the spectra recorded in this study of cobalt(II) in tetramethylguanidine under varying conditions indicate that the principal complex species present is the completely solvated cation, [Co(TMG)4]++. The small wavelength shifts and absorbance changes which were found (Table XIV) on addition of anions to Co(CzH3OZ)Z'4I-IZO can perhaps be rationalized in terms of different degrees of ion-pairing depending on the anion X‘: [Co(TMG),]+++X' : [Co(TMG)4]++;X' (2_3_) _ _ + _ [Co(TMG)4]++;X + X ——m—\ [Co(TMG)4] +;2X (g) The spectra used to obtain the data of Table XIV would therefore represent varying concentrations of the three species, [Co(TMG)4]++ ++ — - [Co(TMGm ;X , and [Co(TMG)4]++;2X . J The spectra of Co(II) solutions with added iodide ion (Figures 23 and 24) perhaps also can be explained in terms of ion-pairing (Equations Q and _2_4). Although small, still the biggest spectral changes in these experiments are found in the cobalt chloride plus iodide systems shown in Figure 24. Here it is seen that there are two and only two types of spectra. A summary of the principal wavelengths is given in Table XV. The spectra of low I_/Co++ ratios may therefore be due to the [C0(TMG)4]++;ZC1- or [Co(TMG)4]++;Cl_ species since chloride ion readily undergoes ion—pairing. As the ratio of I- to Co++ is increased to twenty and above, the right shoulder inflection disappears and the 93 Table XV. Spectra of Cobalt(II) Chloride-Sodium Iodide Solutions in Tetramethylguanidine Initial Concentration = 5 x 10'4 l\_/_I CoCIZ- 61-130 Wavelength, mp I-/CO++ Shoulder Maximum Shoulder 0 564 612 640 2 564 614 643 20 553 602 None 100 553 600 None 1000 553 600 None 94 spectra become constant in shape and quite similar to that reported (29) for [Co(TMG)4](ClO4)2. Therefore it is felt that curves 3—5 in Figure 24 represent the [Co(TMG)4]++ species. Conductivity measure- ments as a function of added iodide ion concentration could perhaps shed some light on degrees of ion—pairing in these solutions. The fact that anhydrous CoClz obeys Beer's law up to 5 x 10‘4 M (Figure 21) supports the postulation of a single ion-paired complex, probably [Co(TMG)4]++;2Cl-, at these concentrations. Perhaps in the case of the other Co(II) salts in tetramethylguanidine the deviation from Beer' 5 Law is a measure of the dissociation or association equilibria represented by Equations 23 and a (69). Acid-Base Titrations in I, l, 3, 3—Tetramethylguanidine A. Selection of Useful Indicators At the time this work was done, nothing was known of the possibility of using visual indicators to detect end points in acid-base titrations in tetramethylguanidine. Williams and co-workers (17a) had already demonstrated that potentiometric acid-base titrations could be made in tetramethylguanidine, but they did not report use of indicators until later (17b). They found that alizarin yellow and azoviolet gave visual end points in agreement with potentiometric values in the titration of benzoic acid. Less satisfactory indicators were g-nitroaniline, thymol blue, phenolphthalein, and g-cresolphthalein. In the work reported in this thesis, phenolphthalein was found to be a useful visual indicator. None of the other indicators mentioned by Williams were studied here. Some twenty-one indicators were tested in the present work, each in pure (neutral) tetramethylguanidine as well as in acidic and basic solutions. The acidic solutions were 10‘3 111 in TMGH+Br-, although the colors of each of the potentially useful indicators were also verified in + - 0.01—0.021v_1 TMGH solutions. The basic solutions used were/v IO 2 1:1 95 in tetrabutylammonium hydroxide, added to tetramethylguanidine as a twenty-five percent methanol solution. The presence of the small amount of resultant methanol had no effect on the color of the indicators. The indicators and their colors are listed in Table XVI. Eight of them, marked with an asterisk in the table, are potentially useful in acid-base titrations. Agreement between the color changes of some of these indicators and the end points in conductometric titrations will be discussed below. Metacresol purple and 2, 4-dinitrophenylhydrazine are listed in the table as having questionable colors because of changes in these colors with time and non-reproducibility of the colors. Perhaps these two indicators react with the solvent. The indicators from crystal violet to basic fuchsin in Table XVI cover the aqueous pH transformation range from about pH 1 to pH 13. The remaining compounds, all nitro-containing dyes, have pKa values from about 14 to 18. 5 as determined in mixed aqueous- nonaqueous systems (70). Thus it is seen that except for tropeoline 00, the useful indicators all have high pKa values (or possess isoelectric points above pH 7 in aqueous solution). This usefulness is likely due to the strongly basic nature of the solvent itself. In fact, practically all the indicators that undergo any color change in tetramethylguanidine show this change only on addition of base (see Table XVI) with the corresponding acidic and neutral solutions possessing the same color. B. Visible Absorption Spectra of Indicators Two indicators that gave an appreciable visual color change even at 10‘3 111 OH— were curcumin and Clayton yellow. Visible absorption spectra of acidic, neutral, and basic solutions of these indicators are reported in Figures 26 and 27. As is evident, when an indicator solu— tion in tetramethylguanidine is made acidic there is little or no change in its absorption spectrum. When made basic, however, a marked 96 .Hofimoflofi 125C; Hdwomb 3030.». Bodowncooko BOSQWICoqu ocfldqmoHfiZuw uofiow> 30.20% 3026? oGMEminGoLQflUOHflZ - No". omeHO 30:0? Edna 30:9» madam opfiflcmuoomofiizu No“. a. a e oceeohosfiscoiofiefioé .N uoflofl>upom xcwnw Mcfimuowcduo ocflflcmOHfiEQIv .Now cgoumlowcdho omcdnonpom owawhoupom ocflEancoflaprHfiEQuw .N Boflkuomcmyo BodowuomCNHO Boflowuowcwfio Emnosm 06mm uoHoC/lpom pomlomcmHO pomlmwcwno 30:6? ~513va xodeLK/oflow. mmoHHoHoO mmoi oHoO 91mm 15.32 xomHmTCooHO GSOHmH-omc.m.HO Esoumupom osfim Sexism 30:on “got/1651mm “BEN/Loom seasonsovw ~6de ~5on 3an ~5on onnH Cflofimspamfiogcwnbn “got/IMEQ mmoHHoHoO mmwioaoO Eofimaufiafioaosms. m. m. m. oaandnm HOmeHomu—og 30:0? 30:0? 30:6.» pom Houudoz odfim odfim Sim odam HoE>£quum~ 051m uofiofi>uosfim uoflofl>uodam m pom EHMNZ/w 6.3m 631m edfim 25m Honosmccohm 30:0? 303on 3026». odouConondocfiE/Vfl -N uofiofi> Bozo? 30:0? 00 ocfloomohhow 30:9» nomcmHO 3026.», 30:6? uoaofl> 139: O Fmov 62.: $3025 eooooaofi ofimmm #93562 339% oGfiEQMDwfmefioEmhuofium .m .H .H e3 mHOumoflde mo muoHoO .H>X oHndH £932ch ommméfiolw c< .riESuuSU Ho «33on :oCQuOmneq‘ wanfimfi> .oN vuzwah 12.x»ozu4u33 03 can can a: on: n~¢ ooc 97 < < 8215-336. "a J<¢h3w2 "z z -5231» v.70. u< . _ . 33NVBHOSSV 98 cow .uoumfipfi ommm-30< c< .3020.” c0330 wo whomaw :oCQHOmQ< ozflmfi> Km. ousmfih Onn 000 is “1:23:43 ms? 00¢ 00? 19.3.6-3 mob. J<¢h3wz Latins: fine. ”m ”z N 4 {———— souveuosav 99 spectral change occurs. Absorption maxima for curcumin and Clayton yellow are given in Table XVII. It is therefore believed that acid-base behavior in tetramethylguanidine could be studied spectrophotometrically by proper selection of indicators such as curcumin, Clayton yellow, etc. It should be added that the spectra of Figures 26 and 27 were recorded using indicator concentrations which were unknown but sufficient to give reasonable abso rbanc e readings . C. Conductometric Titrations of Organic Acids Conductometric methods were used to study the applicability of tetramethylguanidine in the titration of substances which are very weak acids or even weak bases in water solution. Organic acids which can be titrated in water should be even stronger acids in a basic solvent like tetramethylguanidine. This indeed appears to be the case with such representative acids as p-toluenesulfonic, benzoic, and salicylic acids. The latter two acids have been titrated potentiometrically in tetramethylguanidine by Williams e_t a}. (17b). Conductance titration curves are shown in Figure 28 for p-tolu'enesulfonic acid at two different sets of acid-base concentrations. The end points are exceedingly sharp and appear within experimental error of 1:1 stoichiometry. Quantitative acid recovery percentages for this and some succeeding titrations will be given below (Table XIX) sub- sequent to discussion of the remainder of the titrations. Figure 28 also contains the color transformation range for the three indicators phenolphthalein, 2-nitrodipheny1amine, and curcumin which were used in separate titrations of p-toluenesulfonic acid. The first change in indicator color corresponds almost exactly with the conductometric end point. Apparently the base reacts first with the strong acid p-toluenesulfonic acid (or TMGH+, in this case) and then with the weaker indicator acid. Thus, in the acid solution the equilibrium 100 Table XVII. Absorption Maxima: Visible Absorption Spectra of Indicators Wavelength Maxima, mu Indicator Acid Neutral Base Curcumin: 563 565 592 461 467 479 -- 417 409 362 368 369 Clayton Yellow: 529 533 542 423 421 388 101 .Eo< uEOESmocogoH 1N Mo «53933. 0338305950 .wN “:th mmzu:n_oom»_z-~ _fl_ z_u.. TMG + HOH + (C4H9)4N (26) occurs. After the last TMGH+ is consumed at the end point, the indi- cator acid reacts with the base: - + and the color changes from that of HIn to that of In_. Beyond the end point, the rapidly increasing conductance of the solution must be due to addition of base, which is ionized as + _ Thus, the base could probably be written simply as the OH- ion in Equations 26 and 2;] since the large (C4H9)4N+ cation does not take part in the reactions and probably contributes little to the total conductance of a given solution. The conductometric titration curve for benzoic acid is given in Figure 29. This particular titration was actually one in which a solu— tion of the base was standardized using a weighed quantity of the acid. In addition, the concentration of the base was determined by an aqueous titration and found to agree with the titration in tetramethylguanidine. Figure 29 also shows the behavior of the two indicators thymolphthalein and Clayton yellow in separate titrations of benzoic acid. Apparently Clayton yellow is a stronger acid than benzoic acid since its color change occurs immediately upon addition of base. Thymolphthalein, on the other hand, could be used as a visual indicator for benzoic acid, 103 .Eu< 2055mm «0 203.933. ocuoEouodpcoO .ow ouamarm .3235 1 ~36 ..s N . o N v Q 3 0 N 0 n 3 I. o m D 3 0 H u. o. 1 ‘ m g N. ziizkiaosgk _N_ 33.5» 2022.8 2 "324:6 «Sou .3285... m zuounzou .5036 ” 323m _ P _ I 104 although the initial color change occurs slightly before the measured conductometric end point. A separate experiment extending the addition of base beyond a 2:1 basezacid ratio showed no additional break in the conductance curve. This single end point is important in connection with the titration of nitroaromatic compounds described below. Salicylic acid was of interest because of the possibility of titrating both the first [K1(aq) 2%10'3] and second [Kz(aq) £5 10-13] protons in a strongly basic solvent such as tetramethylguanidine. Figure 30 shows that this was not possible; only the first strong-acid proton was definitely titratable. The curvature in the conductance plot after the end point was reproducible, but it was not possible to draw this curve as two straight lines intersecting at a second end point. D. Titrations of Nitroaromatic ComJgounds Since a nitro group on an aromatic ring is electron-withdrawing, protons on the benzene ring itself or on the substituents tend to be acidic. This acidity should be even more pronounced in basic solvents, and Williams and Custer (17a) have already found 2, 4-dinitrophenol as well as p_- and m-nitrOphenol to behave as acids in tetramethylguanidine. In the present work g-nitrophenol was titrated conductometrically (Figure 31). The expected end point occurs at a 1:1 mole ratio, but there is also a second sharp break in the curve at a 1:2 acid-base stoichiometry. That the second end point is real was shown by reproduci- ble results using titration cells with different constants (see Figure 6) and by variation of the concentration of basic titrant. Possibly, after titration of the hydroxyl hydrogen, one of the ring protons is sufficiently acidic to be titrated. A better explanation is believed to lie in the addition of a hydroxyl ion to the ring. The reasoning behind this comes primarily from some of the work of Fritz gt al.(71), in which it was found that some nit roaromatic amines gave two end points in titrations 105 .Bod. 3131mm mo :ofiomkfirfi oCooEouospcoU .0m opzwflrm :02 3m N {I 2 m_n.0 ._2 :NoUESvIwU o manod“ minim _ _ O _ gOIXl-WHO‘33Nv1300NOD 106 .AOCoLQoufiZ um. mo :oflmbffi oCNoEooospcoU . am 3.3;th .3235 w. woovd ._z N . \ mozzp zoczmozNo m 506.0 250.8 _ _ _ ”w4a24m _ O. N. ¢_ gOI x l_WHO ‘BONVianONoa 107 in pyridine with the base triethylbutylammonium hydroxide. With 1, 3, S-trinitrobenzene, for example, they attribute the second end point to OH- addition to the benzene ring. Referring again to Figure 31, it is seen that the color change for thymolphthalein corresponds to the first end point of g-nitrophenol. The consecutive reactions occurring in the titration can be written as follows: OH 0' D N02 +OH' —->~ ‘NOZ +HOH (g2) HIn+0H‘ ——> In- + HOH (3.9.) o‘ - N02 + on‘ No, - (_1_) OH Further insight into reactions such as Equation 11 may be possible through ultraviolet or visible absorption studies. The fact that water is a product in titrations of acids with (C.H9)4NOH apparently is not a problem due to the slowness of the tetramethylguanidine hydrolysis reaction. Further validation of the second break in the conductance titration of g-nitrophenol was obtained by titration of phenol where only a single end point at 1:1 stoichiometry was found. Titration data are given in Figure 32 and, within experimental error, the conductance is linear from the 1:1 end point to as high as a 1:3 acidzbase ratio. Thus, ring addition is more facile in the case of nitro-substituted aromatics. Perhaps the potentiometric titrations of the nitrophenols by Williams (17a) would have given second end points had base addition been continued. 40:0an mo £03.95“? uCuoEouospcoU .Mm ouswflh 10236 z 30° .5 m. N . o Z.— . \\ 108 zonzwu .2 3066 2.0.3 ”3513 gO|X|_WH0l33NV13nON03 109 An attempt to compare the titration of g-nitrophenol with that of nitrobenzene met with little success. Apparently nitrobenzene is too weak an acid in tetramethylguanidine to give a satisfactory conducto- metric titration. Two polynitrobenzenes, l, 3-dinitro- and 1, 3, S-trinitrobenzene, were titrated. Two conductometric titrations of l, 3-dinitrobenzene were carried out under different conditions (Figure 33). The marked slope changes at 1:1 and 1:2 acidzbase ratios are evidence for the re- action of each mole of the dinitrobenzene with two moles of hydroxyl ion. Since there likely are no acidic protons in 1, 3-dinitrobenzene, the hydroxyl ion may again be adding to the ring or perhaps there is re~ action of hydroxyl ion with each of the nitro groups. That the latter may very well be true is shown by the titration of l, 3, S-trinitrobenzene in Figure 34. Although the constructed end points in this titration do not agree very well with the integral acid: base ratios, there are three breaks in the conductance curve near the 1:1, 1:2 and 1:3 ratios. Thus, the number of nitro groups substituted on a benzene ring corresponds to the number of end points which obtain in these titration systems. The literature is helpful here in interpreting the titration behavior of the polynitrobenzenes. Similar multiple end points have been found using ethylenediamine as a solvent. Favini and Bellobono in both potentiometric (72a) and conductometric (72b) titrations with sodium aminoethoxide in ethylenedimaine found 1, 3-dinitrobenzene to have two end points and l, 3, 5-trinitrobenzene to exhibit three. They also studied other polynitrobenzenes and toluenes and obtained generally similar results. It is also of interest that only single end points were found for phenol and for benzoic ac id, as in the present work. By combination of the titration results with some spectrophotometric studies on the same systems, Favini and Bellobono (72a) interpret the titration behavior as 110 .ocouconoHficfiflum J mo :oCmHfiB 07308803950 :ozcam.fl.0Nnd r: 0 n v m .2 «tam; NA N02 .vxwo .ON0.0 .6 0.00“ w4a1 Na > K, Ca > Sr > Ba, and NCS > ClO4 > N03 > I > Br > CZH3OZ > C1 > 804. Many organic compounds possess reasonable solubility in tetramethyl- guanidine and complete miscibility is found with most common organic solvents. C . Reactions Tetramethylguanidine undergoes slow hydrolysis at room tempera- ture to l, 1 -dimethylurea and dimethylamine. Carbon dioxide is 121 122 readily absorbed by tetramethylguanidine in the presence of water or atmospheric moisture to form tetramethylguanidinium bicarbonate. Water could perhaps be removed from tetramethylguanidine by addition of carbon dioxide to form the insoluble bicarbonate salt. Reactions of tetramethylguanidine with acids are vigorous and exothermic; the chloride, bromide, acetate, and bicarbonate salts have been isolated. Protonation of tetramethylguanidine occurs at the imine nitrogen followed by a localization of charge on the central carbon atom, [(CH3)2N]2C+NH2. Only monoprotonation is possible, however. Alkali metal guanidide salts such as Na+NC[N(CI-l3)z]z- could not be iso- lated. Tetramethylguanidine is oxidized by agents such as KMnO4 but none of its oxidation products have been characterized. Reaction of tetramethylguanidine with ions such as Ag+ and ng++ may limit the use of silver(I)ormercury(I) salts in electrode construction for electro- chemical studies. Reaction of tetramethylguanidine with CC14 and C82 preclude their use as solvents in spectrophotometry. D . Metal Compl exe s A maximum of four molecules of tetramethylguanidine can co- ordinate with a typical transition metal ion such as cobalt( II). Tetra- methylguanidine is monodentate but could function, perhaps, as a bridging ligand. Insufficient evidence was obtained in this study to determine whether tetramethylguanidine coordinates via the imine or amine nitrogen, but coordination at the imine position recently has been claimed (29). Tetramethylguanidine is a strong donor ligand as indicated partially by the minor spectral changes with added anions, water, etc. Beer' 8 law studies of cobalt(II) systems in tetramethylguanidine indicate con- siderable solvent-solute interaction. The visible absorption spectra in 123 these systems are consistent with the presence of one or more of the species [Co(TMG)4]++, [Co(TMG)4]++;X-, or [Co(TMG)4]++;2X- in tetramethylguanidine solution. The latter two species are ion-pairs between the fully complexed cobalt(II) ion and a mononegative anion such as chloride ion. Only tetrahedral Co(II) complexes seem to be formed in tetramethylguanidine . E . Ac id- Base Titrations Acids can be titrated conductometrically in tetramethylguanidine as a solvent using (2-C4H9)4NOH in methanol as titrant. Several visual indicators (see Table XVI) exhibit color changes in tetramethylguanidine which could lead to their use in spectrophotometric studies of acid- base behavior in the solvent. The indicator color changes agree with the conductometric end points providing the indicator acid is weaker than the sample acid. Acids likely are leveled to the acidity of the solvo-ca’tion, tetra- methylguanidinium ion, but no measure of the degree of leveling was made. In addition, the degree of dissociation of any of the resultant TMGH+A— salts formed on acid addition were not determined. Titration of nitroaromatic compounds in tetramethylguanidine results in multiple end points through hydroxyl ion addition to the aromatic ring. Thus, g—nitrophenol and m-dinitrobenzene gave two con- ductometric end points, while the titration of 1, 3, 5-trinitrobenzene resulted in three breaks in the conductance curve. Excellent conductometric titration of all three protons in citric acid was obtained. Since the consecutive ionization constants for citric acid in aqueous solution only differ by about tenfold from each other (8.7 X 10-4, 1.8 x 10‘s, and 4.0 x 10'6), differential titration of mixed acids in tetramethylguanidine should be possible. 124 Ammonium ion titrates as an acid in tetramethylguanidine; in other words, tetramethylguanidine is a stronger base than ammonia. A number of compounds such as urea, water, acetamide, and nitro- benzene, however, do not possess sufficiently acidic properties to be titrated conductometrically in tetramethylguanidine. F. General In the introduction it was stated that an over-all purpose of this work was to determine the utility of 1, l, 3, 3-tetramethylguanidine as a solvent. The above conclusions have demonstrated a number of potential areas of usefulness for tetramethylguanidine in terms of its properties. However, despite the findings of the present studies, a great deal more study is necessary before tetramethylguanidine will become a commonly used nonaqueous solvent. It is hoped that this thesis will generate new interest in 1, 1, 3, 3-tetramethylguanidine. RECOMMENDATIONS As mentioned in the conclusions, additional research on 1, 1,3, 3- tetramethylguanidine would be helpful indeed. Some of the more im- portant studies are: 1. Determination of the self-ionization constant. 2. Thermodynamics and kinetics of reactions in or of tet ram ethylguanidine . 3. Comparison of potentiometric and conductometric titrations with high-frequency, thermometric, and coulometric methods. 4. Differential titrations of two or more acids. 5. Ion-pair formation studies. 10. REFERENCES . American Cyanamid Company, Market Development Department Report, IC-0219R-500, May, 1960. . J. Kleinberg, W. G. Argersinger, Jr., and E. Griswold, Inorganic Chemistry, D. C. Heath and Co., Boston, 1960, pp. 218-220. . L. F. Audrieth and J. Kleinberg, Nonaqueous Solvents, John Wiley and Sons, Inc., New York, 1953, pp. 33-34. . H. H. Sisler, Chemistry in Nonaqueous Solvents, Reinhold Pub- lishing Corp., New York, 1961, pp. 6—10. . A. Berg, Compt. rend.,116, 887 (1893). . M. Schenck, Z. physiol. Chem. E, 328 (1912). . M. Schenck and F. von Graevenitz, Z. physiol. Chem. 141, 132 (1924). (a) H. Lecher and F. Graf, Ber. 56B, 1326 (1923). (b) H. Lecher and F. Graf, Ann. Chem. 438, 154 (1924). (c) H. Lecher, F. Graf, and F. Gn'a'dinger, ibid. 445, 61 (1925). C. K. Cain, U. S. 2,898,369, Aug. 4, 1959. C. A. E, 4395i. H. Z. Lecher, E. M. Hardy, and C. L. Kosloski, U. S. 2,845,459, July 29, 1958. C. A. 53, 231e. . S. J. Angyal and W. K. Warburton, J. Chem. Soc. 1951, 2492. . N. F. Hall and M. R. Sprinkle, J. Am. Chem. Soc. 2, 3469 (1932). . T. H. Wirth and N. Davidson, J. Am. Chem. Soc. 86, 4325 (1964). . T. E. Mead, J. Phys. Chem. Q, 2149 (1962). 126 15. l6. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 127 R. Kitawaki, Nippon Kagaku Zasshi, 7_8_, ’1435 (1957). c. A. 542 4382b. American Cyanamid Company, Market Development Department Report, ICOZ34—R-1000, Sept., 1960. (a) T. R. Williams and J. Custer, Talanta 2, 175 (1962). (b) T. R. Williams and M. Lautenschleger, ibid. 19, 804 (1963). H. M. Randall, R. G. Fowler, N. Fuson, and J. R. Dangle, Infrared Determination of Organic Substances, D. Van Nostrand Co., Inc., New York, 1949. J.. Fabian, M. Legrand, and P. Poirier, Bull. soc. chim. France 1956, 1499. Sadtler Spectrum No. 16115, S. P. Sadtler and Son, Inc., Philadelphia, 1962. R. S. Drago, private communication. G. Williams and M. L. Hardy, J. Chem. Soc. 1953, 2560. T. Goto, N. Nakanishi, and M. Ohashi, Bull. Chem. Soc. Japan 2, 723 (1957). L. Paoloni, Gazz. chim. ital. Q, 957 (1959). A. Kotera, T. Morita, S. Aoyagi, Y. Kakiuchi, S. Nagakura, and K. Kume, Nippon Kagaku Zasshi _8_2, 302 (1961). C. A. _5_5, 15128g. P. Ray, Chem. Revs. e_1, 313 (1961). P. S. Gentile and L. H. Talley, J. Am. Chem. Soc. 19, 4296 (1957). S. Akihama and S. Toyoshima, Chem. Pharm. Bull. (Tokyo) 2, 1254 (1962). C. A. g, 12885c. . R. Longhi and R. S. Drago, Inorg. Chem. 4, 11 (1965). 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 128 B. A. Kennedy, unpublished data. J. H. Bloor, unpublished data. P. Job, Ann. chim. 2, 113 (1928). R. L. Shriner and R. C. Fuson, The Systematic Identification of Organic Compounds, John Wiley and Sons, Inc., New York, 1948, p. 27. A. Weissberger and E. Proskauer, Organic Solvents, Oxford University Press, London, 1935, p. 94. F. Daniels, J. H. Mathews, J. W. Williams, and Staff, Experimental Physical Chemistry, McGraw-Hill Book Co., Inc., New York, 1949, p. 557. A. A. Maryott and E. R. Smith, Table of Dielectric Constants of Pure Liquids, National Bureau of Standards Circular 514, U. S. Government Printing Office, Washington, D. C., 1951, pp. 5-13. W. J. Mader, R. D. Vold, and M. J. Vold, in A. Weissberger, ed., Physical Methods of Organic Chemistry, 3rd ed., Vol. 1, Interscience Publishers, Inc., New York, 1959, pp. 655-688. H. H. Willard, L. L. Merritt, Jr., and J. A. Dean, Instrumental Methods of Analysis, 3rd ed., D. Van Nostrand Co., Inc., New York, 1958, pp. 218—257. K. T. Wilke and W. Opferman, Z. Physik. Chem. (Leipzig) 224, 237 (1963). A. I. Popov and J. C. Marshall, J. Inorg. Nucl. Chem. B, 340 (1961). A. Weissberger, E. S. Proskauer, J. A. Riddick, and E. E. Toops, Jr., Organic Solvents, Interscience Publishers, Inc., New York, 1955, pp. 289-295. G. L. Putnam and K. A. Kobe, Trans. Electrochem. Soc. E, 609 (1938). B. E. Geller, Zh. Fiz. Khim. .3_5, 2210 (1961). 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 129 F. Daniels, Outlines of Physical Chemistry, John Wiley and Sons, Inc., New York, 1948, p. 182. . F. Daniels, loc. c_i_t., p. 71. John Deere Chemical Company, Product Data Bulletin, Tetramethylurea. A. Luettringhaus and H. W. Dirksen, Angew. Chem. Internat. Ed. 3, 260 (1964). A. I. Popov in H. B. Jonassen and A. Weissberger, eds., Techniques of Inorganic Chemistry, Vol. 1, Interscience Pub— . lishers, Inc., New York, 1963, p. 45. L. F. Audrieth and J. Kleinberg, 10C. <23, pp. 119, 123. S. Glasstone, Textbook of Physical Chemistry, D. Van Nostrand Co., Inc., New York, 1946, p. 502. K. Nakanishi, Infrared Absorption Spectroscopy——Practica1, Holden-Day, Inc., San Francisco, 1962, pp. 38-39. E. Lieber, D. R. Levering, and L. J. Patterson, Anal. Chem. 23, 1594 (1951). K. Nakamoto, Infrared Spectra of Inorganic and Coordination Compounds, John Wiley and Sons, Inc., New York, 1963, pp. 71, 145. J. H. Hibben, J. Chem. Phys. 5, 166 (1937). L. F. Audrieth and J. Kleinberg, LCE. (_:_i£., p. 120. V. Grignard, G. Dupont, and R. Locquin, Traité de Chimie Organique, Masson et cie, ed., Paris, Vol. 13, 1941, pp. 901-902. C. A. Jacobson, ed., Encyclopedia of Chemical Reactions, Reinhold Publishing Corp., New York, Vol. 2, 1948, pp. 355, 365-369. R. Huisgen and H. Brade, Chem. Ber. 9_0_, 1432 (1957). E. Spinner, J. Phys. Chem. 64, 275 (1960). 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 130 Anonymous, Brit. 293,040, July 1, 1927, C. A. 23, 14185, W. B. Schaap, A. E. Messner, and F. C. Schmidt, J. Am. Chem. Soc. 11, 2683 (1955). C. A. Jacobson, ed., loc. Ci" pp. 227-238. L. F. Audrieth and J. Kleinberg, lo__c. 223., p. 122, (a) L. I. Katzin and J. R. Ferraro, J. Am. Chem. Soc. 7_2 5451 (1950). (b) L. I. Katzin and E. Gebert, ibid. E, 5455, 5464(1950). (c) L. I. Katzin, J. R. Ferraro, and E. Gebert, i_b_i_£1._7__2_, 5471 (1950). (d) L. I. Katzin and J. R. Ferraro, 11351.14, 2752, 6040 (1952). (a) W. Libus’, A. Ugniewska, and S. Minc, Roczniki Chem. 3_4, 29(1960L (1)) W. Libué, ibid. 35, 411 (1961). _— (c) w. Libué, ibid. 39, 999 (1962). S. Buffagni and T. M. Dunn, J. Chem. Soc. 1961, 5105. C. J. Ballhausen and C. K. Jgfrgensen, Acta Chem. Scand. 2, 397 (1955). L. I. Katzin, J. Inorg. Nucl. Chem. 4, 187 (1957). R. D. Hill, Ph. D. Thesis, Michigan State University, East Lansing, Michigan, 1962. R. Stewart and J. P. O’Donnell, J. Am. Chem. Soc. 83, 493 (1962). J. S. Fritz, A. J. Moye, and M. J. Richard, Anal. Chem. Q, 1685 (1957). (a) G. Favini and l. R. Bellobono, Ann. chim. (Rome) 50, 825 (1960). '_7 (b) G. Favini and l. R. Bellobono, _1_131_d. _5__1_, 841 (1961). (a) N. van Meurs and E. A. 1V1. F. Dahmen, Anal. Chim. Acta _1__9, 64 (1958). (b) N. van Meurs and E. A. M. F. Dahmen. ibid. 2.1, 443 (1959). 131 74. L. F. Audrieth and J. Kleinberg, l_og. Ci” p. 77. 75. A. I. Popov, private communication. APPENDICES 132 r.____________._.__ -1 1 APPENDIX I Toxicological Investigation of 1, l, 3, 3-Tetramethylguanidine Range finding toxicological tests on 1, 1, 3, 3-tetramethylguanidine have been done by courtesy of the Biochemical Research Laboratory of Dow Chemical Company, Midland, Michigan. The following conclu- sions are quoted verbatim from the report of the tests: "1, 1, 3, 3—Tetramethylguanidine has a moderate acute oral toxicity, but should present no problem from ingestion incidental to industrial handling. It should be borne in mind, however, that serious injury may occur from the accidental or willful swallowing of relatively small amounts. Containers should be clearly and carefully labeled so that accidental swallowing due to mistaken identity cannot occur. ”The undiluted material is severely irritating to the eye. Direct eye contact would likely result in extensive conjunctival and corneal injury sufficient to produce at least some permanent impairment of vision and possibly total loss of sight. Washing, if it is to be effective, must be immediate and thorough. Unless first aid measures and pre- cautionary handling recommendations suggested in this report are followed, permanent impairment of vision may be sustained. Chemical workers goggles are recommended for safe industrial handling when— ever the likelihood of eye contact exists. ”1, 1, 3, 3-Tetramethy1guanidine is also severely irritating to the skin. Short exposures of 15 seconds or so are likely to produce a severe burn. Precautions must be taken to avoid all skin contact. Protective clothing should be worn whenever the likelihood of skin contact exists. 133 134 "Inhalation studies conducted on rats indicate that single exposures to the vapors of l, l, 3, 3-tetramethylguanidine at room temperature should present no problem. The effect of vapors of the material main- tained at elevated temperatures were not studied. ”These conclusions are based on range finding toxicological tests and are limited to precautions for industrial handling of the material. Development of specific uses will require consideration of the health problems presented and of the need for further toxicological studies. " APPENDIX 11 Data and Calculations: Solubility of NaHSO4 Data: Weighings (a) Wt. Satd. Soln. Sample No. for Flame Anal. 7 4.49399 g 8 4.46883 9 4.46882 Data: Flame Photometry ppm Scale Divisions (ave.) 1.00Na 82.8 0.80 66.4 0.60 52.1 0 40 38. 9 O 20 14.2 0 10 7. 9 These data are plotted in Figure 5. (b) ppm Na minus No. Scale Div. ppm Na Blank 7 56.7 0.653 0.524 8 49.6 0.612 0.483 9 53.1 0.698 0.569 Blank 10.4 0.129 0.00 135 136 Calculations: Flame Photometry (b) (C) (d) No. mg Na/l mg Na/sample vol. mg NaHSO4 7 0.524 0.01310 0.06840 8 0.483 0.01208 0.06307 9 0.569 0.01422 0.07424 Calculations: Solubili_ty (d) (e) = (a-d) (ti/106) No. Wt. Solute Wt. Solvent g Solute/100 g Solvent 7 0.06840 mg 4.49392 g 0.001522 8 0.06307 4.46877 0.001411 9 0.07424 4.46875 0.001661 Ave . Solubility 0. 00153 ATE N l“1111111111111111111115 0 3 0 8 2 19 2 4 93 “will