当前位置:首页 >> 材料科学 >>

Complexes of Alkali Metals and Alkaline-Earth Metals with Carbohydrates


Advances in Carbohydrate Chemistry, Volume 21, 1967, Pages 209-271

COMPLEXES OF ALKALI METALS AND ALKALINE-EARTH METALS WITH CARBOHYDRATES BY J. A. RENDLEMAN, JR.
N o r t h h g h a l Research Laboratory, N o r t h Utilization Rsearrch and Develqnnent Divieion, Agricultural Reaarrch deroiclr, U.S.Department of Agtidure, Peoria, IUitsoia

I. Introduction.. ....................................................... 209 11. Complexes of Carbohydrates with Metal Salts.. .......................... 211 1. Proof of Existence.. ................................................ 211 2. Adducta of Alkali Metal Salts and Alkeline-earth Metal Salts.. .......... 215 220 3. Methods of Preparation.. ........................................... 222 4. IStoichiometry ..................................................... 6. hlvation.. ....................................................... 226 -6. Influence of Size of the Cation on Stability of the Complex. . . . . . . . . . . . . . 227 7. Effect of Complexing on Optical Rotation.. ........................... 228 8. Electrophoreeis..................................................... 231 9. structure of the complex.. .......................................... 236 1 1 1 . Complexw from the Interaction of Carbohydrates with Metal Bases.. . . . . . . 237 I. R$8ctions in Aqueous Media.. ....................................... 241 2. Reactions in Anhydrous, Alcoholic Media. ............................ 265 n Nonhydroxylic Solvents. ................................ 264 3. Resotione i 4. Structure of Alcoholatea and Adducts. ................................ 265 IV. Alcoholaka from Reactions, in Liquid Ammonia, of Carbohydrates with Alkali Metals, Alkaline-earth Metals, and Alkali Metal Amidee.. . . . . . . . . . . . . . . 269

I. INTRODUCTION Published information on the complexes of alkali metals and alkalineearth metals with carbohydrates has often been vague and difficult to interpret, largely because (a) ions of these metals do not readily form stable complexes in aqueous solution, and (b) the study of such weak interactions is beset with many analytical difficulties. Although the detection of complex-formation in solution is relatively easy, the determination of stability constants has not yet been accomplished. In the present article, a11 attempt is made to produce some order from the available data. The presentation on the complexes of metal salts will be largely restricted to salts having univalent anions. Complexes containing nonmetallic cations, such as quaternary ammonium ions, will be mentioned only when such information is believed relevant to the discussion of complexes of alkali metals and alkaline-earth metals. The terms “adduct” and “addition compound” are synonymous, and will be used to describe those complexes consisting of carbohydrate and metal salt (or base) bound together, prob209

210

J. A. HENDLEMAN, JR.

ably, by ion-dipole forces of attraction. In those reactions where a hydroxylic proton is removed from a carbohydrate, the term “alcoholate” will be wed instead of “alkoxide,” to differentiate between the oxyanion of a carbohydrate and that of aamonohydric alcohol. In no instance will a “carbohydrate alcoholate” signify a carbohydrate solvated with an alcohol. Von Lippmanl reviewed much of the work on carbohydrate complexes published before 1904, and, for such complexes,Vogel end Georg2 bompiled a comprehensive table based on work published before 1930. It is important to note that many of the chemical formulas suggested by early investigators for compounds formed by the interaction of carbohydrates with metal bases were mere assumptions, based on insuiilcient evidence. Similar tendencies exist in the literature even today. Scant attention has been given by chemists to possible practical applications for carbohydrate complexes and complexing behavior. Differences in the solubility or stability of complexes can, in some instances, permit the largescale separation of one polyhydroxy compound from another.a.4 Differences in electrophoretic mobility of carbohydrates in the presence of a salt can be used to separate mixtures on a emall scale and to assist in identifying a carbohydrate. The ability of metal salts to complex selectively with polyhydroxy compounds has been the basis of certain separations of carbohydrates by column chromatographyf There have been no studies to determine whether the chemical properties of carbohydrate-aalt complexes differ significantly from those of the corresponding uncomplexed carbohydrates. Should such differences exist, they may be of great benefit to the preparative chemist. An understanding of the requirements for forming complexes of polyhydroxy compounds is eseential for the biologist studying processes of cellular transport. Closely ordered arrangements of donor groups in cell membranes may be responsible for the selectivity observed for certain metals in active transport. The effect of dietary f&ctom, especially sugars and amino acids, on the gastrointestinal absorption of calcium and other alkaline-earth motah R i well documented.&’ Charley and Saltmad have proposed that thc mcchunism by which sugars promote
(1) E. 0. von Lippmann, “Die Chemie der Zuckerarten,” Friedrich Vieweg und 6ohn, Braunsahweig, 1904. (2) H. Vogel and A. Qeorg, “Tabellen der Zucker und Ihrer Derivate,” Julius Springer, Berlin, 1931, pp. 378-397. (3) A. J. Wattem, R. C. Horkeft, and C. 8. Hudson, J . Am. CAm. h., 66, 2100 (1934). (4) 8.Bey, 2 .Kldn. Md., 38, 305 (1900). (6)J. K. N. Jones and R. A. W a l l ,Can. J . Cham., 36, 2290 (1960). (6) J. T. Irving, “Calcium Metabolism,” Methuen and Co., London, 1967. (7) P. Fournier, Compt. Rad., 248, 3744 (1859). (8) P.Fournier, B. Susbielle, and Y. Dupuis, C m p t . Rend., OM), 1111 (1980). (9) P. J. Charley and P. Saltman, rscisnce, 189, 1208 (1963).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

211

the absorption of calcium involves the fonnat.ion of a sugar-calcium complex in which the CaN ion is readily available for transport.

11. COMPLEXES OF CARBOHYDRATES WITHMETAL SALTS 1 . Proof of Existence
The existence of carbohydrate-salt adducts had been tacitly assumed for many years, with little effort made to provide proof thereof. Helderman'O attempted to obtain an answer to this question by determining solubility curves, at 30°, for threecomponent system containing salt, water, and either sucrose or wglucose. Among the salts employed were sodium chloride, potassium chloride, and potassium sulfate. Unfortunately, a faulty experimental procedure led to the erroneous conclusion that adduct formation does not occur at this temperature. Later studies of the D-glucose-sodium chloride-water system by Matsuurall and Tegge,12 and of the sucrosesodium iodide-water system by Wiklund'a furnished proof of the existence

0

3
0

2

. c 0

25-

f

OO
Wt.

25%

50
of

I

NoCl

at 24".

FIO.l.-Solubility Relations11 in the System D-Glucose--Sodium Chloride-Water

(10) W. D. Xelderman, Arch. Suikerind. Ned. India, 98, 1701, 2306 (1920). (11) 8. MatBUura, Bull. C b . SOC.Japan, 9, 44 (1927); J . Chem. SOC. Japan, Pure Chem. Sect., 48, 247 (1928). (12) G. Tegge, Sfacrke, 14, 269 (1962). (13) 0. Wiklund, Zwker, 8, 266 (1956)

212

J. A. HENDLEMAN, J12.

2C,2 HZ2O,, 3NaI *3H20.

Not, moles/100 moles of H,O

FIQ.2.4k1lubility Relationsla i n the Syetem Suckm-&xh ‘umIodide-Water at 30’.

of 2 D-glucose.I\iaC1-HzO, 2 sucrose*3NaI-3 HtO, and sucrose.NaI.2 HnO. The diagrams in Figs. 1 and 2 show, respectively, the solubility relations in these two systems. They demonstrate not only the presence of sugar-salt adducts, but also how the formation of complexes affects the solubility of both the carbohydrate and the salt. have extensively studied the reTegge” and Lebedev and lation between temperature and the stability of 2 D-glucose*NaC1*HzO in aqueous systems, T h e fact that this adduct crystallizes better than pure D-glucose from aqueous solution is importmt industrially. &Glucose may be recovered from its adduct in 77.1% yield by agitating the adduct with water at 5’ for 1 hour.14 The sodium chloride diasolvee, and the D-glucose remains as a crystalline solid. Dialysis experiments16 have shown that Cn”, M P , Ba2@, and Srm form mluble chelates in aqueous alkaline solution with D-galactose, D-glucose, D-fructose, D-arabinose, D-ribose, maltose, and lactose. The absence of any precipitation of alkaline-earth metal hydroxide when an aqueous solution containing Dfructose and an alkaline-earth metal salt is made alkaline (14) N. V. Lebedev, Sb. Tr. &a. Nauchn.-Iaulsd. Inut. Wrolizn. i Sd’jitno-Spirt.
Pront., 8, 144, 202 (1980); E. L. C$lsakova and N. V. Lebedev, ibid., 8, 220, 235 (lSe0); 9, 90, 102, 110, 134 (1961); N. V. hbedev, A. A. Bannikova, and M. N. Lyuhotakayn, ihid., 8, 170, 186 (leS0); N. V. Lebedev and A. A. Bannikova, ibid., 8, 126 (1960); B, 81 (1961); 11, 68 (1963); N. V. Lebedev, B. 0. Lyubin, and A. A. Bannikova, ibid., 9, 70 (1961); Qidrolitn. i Lesokhim. Prom., 1 1 , 3 (1968); N. V. Lebedev, B. 0. Lyubin, and E. L. Glaikova, Rutwian Pat. 121,088 (1969); Chem. Abelmcia, 64, 4012 (1960); N. V. Lehedev, M. N. Lyubetskaya, and A. A. Bannikova, Rumisian Pat. 136,836 (1961); Chem. Abutmtu, 66, 14962 (1961). (16) P. D. Saltman and P. J. Charley, Frenoh Pat. 1,827,727 (1963).

A L W I I AND ALKALINE-EARTH METAL COMPLEXES

213

(pH 12) is additional evidence16 that alkaline-earth metal hydroxides form complexes with carbohydrates. Similarly, calcium carbonate is not precipitated when carbon dioxide is passed through an aqueous solution of sucrose and calcium oxide (or hydroxide).” When an aqueous solution of sucrose and calcium oxide is added to an aqueous solution of sodium aluminate, there is formed a water-soluble complex containing sucrose, aluminum, and calcium.’8 Phase studied9 of the ternary systems sucrosewater-barium oxide and sucrose-water-strontium oxide have shown the existence of sucrose-BaO, sucrose-3 BaO, sucrose-SrO,and sucrose.2’SrQ. In a phase study of the ternary system sucrose-water-sodium carbonate,20 the complex sucrose*Na&Oswas shown to exist. However, a complex of sucrose with potassium carbonate has never been observed. The specific optical rotation of many sugars and sugar derivatives is altered by the presence of metal salts, the alteration being the greater, the greater the concentration of the salt. This phenomenon, which is now generally attributed to the formation of adducts, will be considered later in more detail. Other physical phenomena that may be associated, at least partially, with complex formation are the effect of a salt on the viscosity of aqueous solutions of a sugar and the effect of carbohydrates on the electrical conductivity of aqueous solutions of electrolytes. Measurements have been made of the increase in viscosity of aqueous sucrose solutions caused by the presence of potassium acetate, potassium chloride, potassium oxalate, and the potassium and calcium salt of 54x0-2-pyrrolidinecarboxylic acid.*’ Potassium acetate has a greater effect than potassium chloride, and calcium ion is more effective than potassium ion. Conductivities of 0.010.05 N aqueous solutions of potassium chloride, sodium chloride, potassium sulfate, sodium sulfate, sodium carbonate, potassium bicarbonate, potassium hydroxide, and sodium hydroxide, ammonium hydroxide, and calcium sulfate, in both the presence and absence of sucrose, have been determined by Se1ix.n At a sucrose concentration of 15” Brix (15.9 g. of sucrose/lOO ml. of solution), an increase of 1O Brix in sucrose causes a 4% decrease in conductivity. Landt and BodeaZsstudied dilute aqueous solutions of potassium chloride, sodium chloride, barium chloride, and tetra(16) P. J. Charley, B. Sarkar, C. F. Stitt, and P. D. Saltman, Biochim. Biophp. Acb, 88, 313 (1963). (17) .I. Dubourg, “Sucrerie de BetteravcR,” J. R . Builliere, Paris, 1952, pp. 179-181. (18) E. Calvet, H. Thibon, and R. Ugo, Bull. Soc. Chim. France, 1346 (1965). (19) K. Nishizawa and Y . Hachihama, 2 .E l e k t r o c h . , 36, 385 (1929). (20):K. Nishwawa and M. Amagaaa, J . Soc. Chem. I d . , Japan, S8, Suppl. binding, 497 (1933). (21) P. Naffa and C. Frege, Sum. Franc., 100, 179, 207 (1959). (22) M.Selii, Listy Cukrouar., 88, 151 (1949-50). (23) E. Landt and C. Bodea, 2 .Ver. Deut. Zucker-Id., 81, 721 (1931).

214

J. A. HENDLEMAN, JR.

ethylammonium picrate containing high concentrations of sucrose (up to 6 7 0 j 0 ) .They, too, found that an increase in the sugar concentration causes a decrease in conductivity. To explain the influence of sucrose on conductivity, a theory was proposed in which 13ucrom molecules are assumed to associate in solution. Association would incretm as the concentration of iucrose increases and as the temperature decreases. Ions traveling between these aggregates would meet with less opposition the smaller the ion. J. M. Stokes, R. H. Stokes, and coworkers~4 extensively studied the limiting conductances of electrolytes and the limiting mobilities of ions in aqueous solutions of sucrose, wmannitol, pentaerythritol, glycerol, and the polysaccharide Ecoll (average molecular weight, about 1W). Although there was a clear correlation between the solution viscosity and the ionic mobilities, there was no simple quantitative relation between them. A similar observation was made by Landt and Bodea,N who found that ionic migration in sucrose solution ip,more rapid than that predicted by the solution viscosity. Both the size of the ion and the size of the polyhydroxy compound affect the conductivity, In general, emall ions, such as K@ and Cle, are less retarded by the presence of polyhydroxy molecules than are large ions, such as Ca*@ or N(C4Ha)4",the latter approaching more nearly the classical hydrodynamic behavior predicted by Stokes' law or Walden's rule. The viscosity of ficoll solutions is several times that of solutions containing the same percentage by weight of sucrose or other simple polyhydroxy compounds; yet, the limiting conductances of small ions in all of these solutions differ only slightly. It wm concluded f r o m these etudim that ficoll molecules form loose networks that are to a large extent permeable to small ions. Unfortunately, the possible contribution of complexing between polyhydraxy compounds and salts haa been almost completely ignored in theoreticd treatments of conductivity data. The results of a conductivity study on locust-bean gum (a neutral polysaccharide) in aqueous salt solutions led Btuy and Halseys to conclude that there is no interaction between this carbohydrate and the many ions employed (HaJ Na@, K@, Ag@,Barn, OAce, Cle, N O , ' , SOP, and HSOP). Hydroxide ion was an exception. The electrolyte concentrations ranged from 0.0005 to 0.1 N,and the maximum concentration of polysaccharide was 0.0277 M (on a monosaccharide-residue baeis). The specific conductivity of each polysaccharidegalt solution waa identical to that of the corresponding salt solution containing no polysaccharide, strongly indicating the absertce of significant adduct formation under the conditions
(24) J. M. Stokes and R. H. Stokes, J . Phye. Chm., 60, 217 (1950); 62, 497 (1958); B. J. Steel, J. M. Rtokea,,and R. H.Stokee, iW., 62, 1614 (1958); F. J. Kelly, R. Mille, and J. M. Stokoe, ibid., 64, 1448 ( I N ) ; R.H. Btokea and I. A. Weeks, Auelralian J . C'henr., 17, 304 (1864). (26) J. A. Barry and C. D.Bsbey, J . Phye. Cham., 67, 1698 (1883).

ALKALI AWD ALKALINE-EARTH METAL COMPLEXES

215

employed. Had it been possible to use higher concentrations of carbohydrate, mme complexing might, perhaps, have been observed. However, the physical limitations of low solubility and high viscosity made it impractical to exceed a polymer concentration of 0.5y0by weight. f crystals having a constant composition constitutes good The isolation o evidence of adduct formation. Crystals of sucrose*NaI*2H20, several centimeters in length and of constant composition, have been isolated.16 Were adduct formation not actually involved, sucrose and sodium iodide would have been precipitated separately. In dilute ethanolic solutions of alkali metal salts, many simple carbohydrates unite with the salt to give an isolable crystalline solid of either 1 :1 or 2: 1 ratio of carbohydrate to salt. The frequency with which these ratios are obtained argues not only for the existence of true adducts, but also for a generally preferred stoichiometry at low concentration of salt. Traces of sodium ion in samples of potassium bromide and potassium iodide used in preparing pressed discs for infrared analysis can alter the normal spectrum of D-glucose, This alteration has been attributed* to the formation of NaBr - 2 D-glucose and NaI-2 D-glucose. The normal spectra of D-xylose, D-sorbose, and sucrose are also altered by the presence of sodium ion. However, alteration of spectra is not always predictable; lactose, cellobiose, mmannitol, and erythritol give normal spectra in potassium bromide discs containing trace amounts of sodium. Tipson and Isbelln have found that the spectra of certain aldoses and alkyl glycosides in Nujol mulls differ from the corresponding spectra in discs of potassium f such behavior can be shown to be due to the formation of a iodide. I complex with the metal salt, detection of complex formation by means of infrared spectra may be a possibility. Lactose, sucrose, maltose, wgalactose, Dfructose, and many other carbohydrates are highly soluble in absolute methanol containing sufficient calcium chloride.* This high solubility strongly suggests the formation of carbohydratecalcium chloride complexes. Lactose-CaC12.4 MeOH crystallizes slowly from concentrated solutions of lactose and calcium chloride in methanol. Alcoholic salt solutions might, therefore, serve as solvents for sugar reactions, and for fractionating sugar mixtures by extraction or Crystallization.
2. Adducte of Alkali Metal Salts and Alkaline-earth Metal Salts

Tltblc! I rccords most of the known, isolated adducts of alkali metal salt8 rtnd alkaline-earth metal salts with polyhydroxy compounds. A (26)D.Gauthier, Compt. Rend., 198, 638 (1904). (27) V. C.Fsrrncr, C'hem. I d . (London), 1306 (1959). (27t~) R.S.Tipuon aiid IT. 9. Isbell, J . Res.Natl.Rur.Std.,64A,239(1960);66A,31(1962).

Tmm I
Adducts o f Alkali Metal Salts and Alkaliue-earth Metal %Its
Carbohydrate (ligand)

salt

Molar ratio,. l@md:salt

Solvent of solvation, molecdes/cation

medium*

Solvent

References

potsssium bromide potassium formate potassium iodide potsssium propionate Calcmmchloride ealciumchloride

2:l

1:l 2: 1 2:l 2:l 1:l l:1 1:l 2:l 6.3 EX) 2 Hto 4 Hto
4 MeOH 2 HzO

29 29 29 29 29 30
31 31

29

P

tc

2
g

n . 4

2:3
barium iodide

32

E *

2

ealcilrmchloride

calcium bromide

2: 1 2: 1 2: 1 2:1 1:l 1:l

2:l

2:l 1:1

2

m

33 33 33 33 33
28

i;:

3 HI0 3H Z O
none

33
EtOH
EtQH
34

3

calcium chloride calcium chloride p0tasS;uumacetate sodium bromide

1:1 1:l 2: 1 2: 1 2: 1 2: 1 2: 1 2:1

3 H20
? MeOH none H Z O
0 . 1 EtOH

H20 MeOH Py-EtOH-EtOAC H20 H20 EtOH EtOH HZO HO NMP-EtOH-Et& EtOH ;H&EtOH HtO EtOH

35
28

sodium chloride sodium fluoride sodium fomte. sodium iodide
potsssium acetate.

HO

2: 1 2:l 2: 1 2: 1
1:1 2: 1 1:l 1:l 2: 1 1:l 2:1 1:l
2: 1

11, 12, 14, 38, 4 0 38

36 37 38 39 36

c ti U

36

3641

38 3,36

E

1
B
0

*

@-D-Clucopyranoside, methyl D-Gulose

calcium chloride calcium chloride calcium chloride calcium chloride

HtO

H&EtOH HtO

42
42s
43

P 3 X

a-D-culose
a-DGulopyranoside, methyl
/3

anomer

2H Z O 3H O 2 H20 none

Hz0 H&EtOH HtO-EtOH EtOH HzO HzO-EtOH

0

43
43 43

$
k 2

E l

Heptopyranoside, ~ - g Z y c e t o a - ~ calcium chloride gulo-, methyl B anomer calcium chloride

H2O
2 HtO

43a 43s

E?

TABLE I (Cdinucd)
Adducts of Alkali Metal Salts and Alkaline-earth Metal Salts Carbohydrate (ligand)

K
Solvent
mediumb

Salt

Molar ratio," ligand 4 t

Solvent of solvation, molecules/ca tion

References

Heptose, ~ - ~ L p m - ~ ~ w g t z k cdcium b chloride Heptose, D s l l J w o - & ~ htae a-wlyxopyranoside, methyl calcium chloride
calcium chloride
d c i u m chloride

1:l 1:l 1:1 1:l
1:l

3 HrO 4 Hto
2HP

HrO-EtOH H&EtDH H&-EtOH

44

44 45
28,46

f-.

?

7 HzO
2HP 4 MeOH 2 Hto 0.2 EtOH

HrO ;H&MeOH

E 2
W

1:1

Hfi-C&OH
MeOH

43a 28

5 2 :
-

E

d c i u m chloride

potsasium acetate
d u m iodide

1:l 1:1 l:1 1:l
2: 1

4H t o

HeEtQH
EtDH-EW EtOH-EW

Ha

-

36 36

47 47

.;;

%Propanone, 1, 3-dihydroxyLrsorbosi?

sodium chloride calcium cblonde
barium bromide barium chloride barium iodide

none 2 HrO

H20 Hto
HP

48
49

1:l 2: 1 2:1 2: 1

sucrose

-

-

Hso
Hto

504 50" 50'

barium thiocyanate calcium bromide calcium iodide lithium bromide lithium chloride lithium iodide potassium acetate
potassium iodide potasaium thiocyanate sodium acetate sodium bromide sodium carbonate sodium chloride sodium iodide sodium propionate sodium thiosulfate strontium chloride strontium bromide a-D-Xylose SD-Xylopyranoside, mcthyl calcium chloride potassium acetate

1:l 1:l 1:1 1:1 1:1 1:l 2: 1 1:2 1:1 1:l 1:l 1:l 1:l 1:1 1:l 1:l 2:3 2: 1 1:l 1:1 1:l 1:1
1:l

2 HtO 3 HzO 3 HtO 2 HzO 2 HtO 2 H20 -

-

2 HzO

2 HtO HZ0

-

none 2 H20 none 2 HtO HtO

HtO HtO HIO HtO H20 H20 NMP-EtOH-EhO N MP-EtOH-EhO HZO H2O HoO HoO HzO HLeEtOH

HtO

HzO 3 H20 3 HzO

N MP-EtOH-EttO
HtO HtO HtO

HnO

3 HzO

H20 EtOH

31 3

d

0

-

5
2 !

F=

a The ligand in an amylose adduct is the wglucose residue. In other adducts, it is the entire carbohydrate molecule. * Py = Zpyrrolidinone; NMP = N-methyl-Zpyrrolidinone; EtOH = ethanol; MeOH = methanol; EtOAc = ethyl acetate; EhO = ethyl ether. Gauthier did not offer any analytical data. Therefore, those results not corroborated by other investigators should be viewed with some skepticism.

Footnotes for Table I are located at the bottom of page 220.

r?

220

J. A. RENDLEMAN,

JR.

number of adducts reported in the literature have been intentionally omitted, because of either tho unreliabitity of published data or the lack of sufficient information on the combining ratio. Many investigators have neglected the possibility that their “products” were not true adducts, but actually mixtures of separate components. The water of hydration bound to carbohydrate adducts isolated f r o m aqueous media is ordinarily found by subtracting the combined weight of the carbohydrate and the salt, as determined by analysis, f r o m the total weight of a sample of the adduct. Many adducts prepared in, or recrystallized from, alcoholic media (anhydrous or aqueous) have never been analyzed for alcohol of solvation.
3, Methods of Preparation

More preparations, and attempted preparations, have been carried out in water than in any other solvent. Aqueous alcoholic media have been used occasionally. The low solubility of many carbohydrates and salts in alcohols and most other organic solvents is, perhaps, the reason why most investigators have generally avoided nonaqueous media.
(29) F.R.Senti and L. P. Witnrruer, J . Polymer Sci., 9, 116 (1962). (30) W. C.Austin and J. P. Walsh, J . Am. Chem. Soc., 60,934 (1934). (31) J. K.Dale, J . Am. C h m . Soc., 60,932 (1934). (32)A. Hybl, R. E. Rundle, and D. E. William, J . Am. C h .Soc., e7, 2779 (1966). (33) R.H. Smith pnd B. Tollena, Ber., 88, 1277 (1900). (34)R. Kuhn, H.’Baer, and A. Gsuhe, Chcm. Bet., 88, 1136 (1956). (36) R. M. Hann and C. S. Hudson, J . Am. C h m . doc., 69,2076 (1937). (36) J. A. Rendleman, Jr., J . Org. Chcm., 81, 1889 (1908). (37) J. Stenhouse, Ann., 119, 286 (1884). (38)H. Traube, N e w s Juhrb. Minenrl. Gml., Beilage Bd. VIII, 610 (1893). (39) M. Hlrnig and M. Rosenfeld, Ber., 10, 871 (1877). (40) Q. Tegge and W. Kempf, Stosrke, 10, 103 (1968). (41) J. A. WtUng, German Pat. 196,sOa (1907);Chum. Zenfr., 79, I, 1688 (1908). (42)H. [J. Isbell and W. W.Pigman, J . Re8. Natl. Bur. dfd., 18, 141 (1937). (42a) H.8 .Iabell, Bur. Srccndarde J . Reeearch 6,741 (1930). (43)H. 8.Isbell, Bur. Stundurda J . Raeurch, 8, 1 (1932). (43a) H. 8.Isbell and H. L. Fruah, J . Rse. Nu& Bur. W., 94, 126 (1940). (44)E.M, Montgomery and C. 8.Hudson, J . Am. Chem. 8oc., 04, 247 (1942). (46)H. S.Isbell and H. L. Fruah, J . &a. Nutl. Bur. dld., 81, 163 (1943). (46)B.L. Herringtan, J . Dairy Sci., 1 7 , 805 (1934). (47) J. K.Dale, J . Am. Chem. Sot., 61,2788 (1929). (48)L. M. Utkin, Biokhimiyu, 4, 800 (1939). (49) R.L.Whistler and R. M. Hixon, J . Am. Chcm. Soc., 60,‘728(1938). (50) D. Gauthier, Compt. Rend., 187, 1269 (1903). (Sl) T. M h d y and G. Vavrinecs, Zucker, 1 1 , 648 (1968). (S2) W.Cockran, Nature, 157, 231 (1946). (63) C.H. Gill, J . C h .doc., 14, 269 (1871);Ber., 4,417 (1871). (64) F.H.C.Kelly, Znlem. Sugur J., 66, 128 (1966).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

221

a. Aqueous and Aqueous Alcoholic Media.-The usual procedure for preparing adducts of carbohydrates no larger than oligosaccharides is to evaporate, slowly, a solution containing carbohydrate and salt in the molar ratio approximating the ratio expected in the desired adduct. The length of time required for crystallization to commence is often great for purely aqueous solutions, and depends upon the temperature, rate of evaporation, f nucleus formation, stirring, and adduct solubility. The presence of rate o ethanol in the medium facilitates crystallization. In the absence of ethanol, the time to commencement of crystallization may vary from 1 day to several years. Use of molar ratios of carbohydrate to salt either greater or lower than the ratio expected in the adduct being prepared sometimes leads to undesirable results. Adducts having more than one combining ratio are possible, particularly with carbohydrates of high molecular weight (oligosaccharides and polysaccharides) Furthermore, at certain ratios of carbohydrate to salt, it is possible for either the carbohydrate or the salt to begin precipitating prior to precipitation of the adduct. This could lead to erroneous conclusions concerning the actual composition of the adduct. An example of this possibility occurs in the work of Sharkov and Guzhavina,” who reported that D-glucose combines with potassium iodide, lithium chloride, lithium bromide, or lithium iodide, respectively, in the remarkably high ratio of 4:l. The ratios for the adducts of the lithium salts, particularly, are unrempable, because the surface area of the lithium ion is much too small to accommodate 4 multidonor ligands. After precipitation, many adducts may be washed with ethanol, and allowed to dry at room temperature. However, consideration must be given to the fact that, if contact of the adduct with the washing solvent is unduly prolonged, washing with ethanol can lead to a reduction in the number of molecules of hydration and, possibly, to a change in the molar ratio of carbohydrate to salt. There is very little information on the preparation of polysaccharide adducts. Senti and Witnauerm have reported the only polysaccharide adducts of a definite stoichiometric type. Their method consisted of equilibrating the alkali metal hydroxide adduct (see Section 111) of the polysacchaxide with the proper salt in 20-25% aqueous ethanolic medium. Absorption of salt, with concurrent displacement of hydroxide, increases with increase in concentration of the salt in the exchange medium. Above certain concentrations of the salt, absorption can become virtually independent of the concentration , indicating that maximum absorption has been attained and that a “stable” adduct of definite stoichiometric compo-

.

(66) V. I. Sharkov and V. Guzhavina, Zh. Prikl. Khdm., 81, 1759 (1958).

222

J. A . HINDLEMAN, JIL

sition has bcori formd. There are indications that varying the salt concentration w y , in some instances, permit the imlation of more than one “stable” adduct. Tho term “stable” is used loosely, and describes adducts whose combining ratios vary only slightly over a rather wide range of salt concentration. Neutraliaation of a potassium hydroxide adduct of amylose with gweous carbon dioxide leads to a potaseium hydrogen carbonate adduct?# This type of reaction,which is rapid in a moist atmosphere, could possibly be used in the preparation of bicarbonate adducts of other carbohydrates.

b . Anhydrous Alcoholic Media ,-Only the simple carbohydrates, including oligosaccharides,have yet been studied in anhydrous media. The best procedure to use for preparing adducts is determined largely by the respective solubility of carbohydrate, salt, and adduct; thus, no specific procedure can be outlined that would be generally applicable. The simple addition of an alcoholic salt solution to an alcoholic carbohydrate solution might be all that would be necessary to effect precipitation of an adduct. Ethanol i s better than methanol as a preparative medium, because of the lower solubility, and perhaps greater stability, of the adduct in the former. Addition of ether to an ethanolic solution of oarbohydrate and salt facilitates precipitation; however, the addition must be cautious, to prevent total precipitation. The laotams Zpyrrolidinone and N-methyl-2-pyrrolidinone may be used to incresse the solubdity of Carbohydrates in alcoholic reaction media. However, they cannot be used in the preparation of metal halide adducts, because of their tendency to associate, perhaps by complexing, with metal halide adducts.” Amides are known t o form complexes with metal halides. to For example, sddium iodide combines with N,Ndimethylformamide~ give NaI.3 N,Ndimethylformamide. Acetamide ctm form NaBr.2 acetamide and Na1.2 acetamide.“ T h e stoichiometry in anhydrous alcoholic media, aa in aqueous media, is variable. The combining ratio for a “stable” adduct prepared at a low concentration of salt may often differ from the ratio obtained at a much higher concentration. The relationship between combining ratio and salt concentration will now be discussed in more detail.
4. Stoichiometry
a. Combining Ratioe in Iaolated Adducte.-Information on the stoichiometry of carbohydrate-salt interactions is based largely upon the (66)Y.Gobillon, P.Piret, and M. van MeBresohe, BuU. 8m.Chim. F r a w , MI1 (1962);
(67)

206 (1963). Y.Gobillon and P.Piret, A&

Cjyst., 16,

1186 (1962).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

223

cmmpoeition of the complexes iwolated. The most frequently reported conibilling ratios of carbohydrate to salt are 1:l and 2:l. Other ratios are possible for certain carbohydrates, Ratios for adducts thus far prepared in aqueous media have not exceeded 2: 1. The requirements for the formation of a carbohydrate-salt adduct are those for chelation between a metal ion and a multidonor molecule. A group of two,or more, properly oriented hydroxyl groups, or a coqbination of a carbonyl group with one, or more, properly oriented hydroxyl groups is necessary. There has been no indication that the oxygen atom of an ether linkage (or of the glycosidic hemiacetal linkage) can participate as an electron donor and, therefore, serve to bind a cation. That bonding of an alkali metal cation to an alkoxyl group is possible, and may contribute, at least weakly, to chelate stability is shown by the existence of intermolecular bondinp of the type 0-Li. * -0 between molecules of lithium alkoxide in such organic solvents as ether, hexane, and p-dioxane. The greater the number of donor groups in a carbohydrate, the greater is the probability that more than one metal cation can become attached to a single donor molecule to form a po2gcation adduct. Although monosaccharides have not been shown to form polycation adducts, oligosaccharides and polysaccharides associate with two or more cations with great facility. High concentrations of salt favor high ratios of salt to carbohydrate. Ethanolic or aqueous ethanolic media favor high ratios of salt to saccharide in' oligosaccharide adducts. For example, aucmse.2 KOAc can be isolateda by adding ether to an ethanolic solution of sucrose (0.02 M ) and potassium acetate (0.4 M ). Although Mackenzie and Quinsg reported the preparation of sucrose*KCl*HgO, KellyMhas questioned its existence (as an isolable compound) after finding that sucrose and potassium chloride crystallize, separately and simultaneously, from aqueous solution. Gills3was unable to isolate any adducts of sucrose with lithium chloride, lithium iodide, potassium chloride, potassium bromide, or potassium iodide, respectively; with ammonium chloride and ammonium iodide, deliquescent crystals of variable composition were obtained. Much later, however, Wiklundls succeeded in preparing sucrose.KI.2 HaO; and GauthieldO reported the preparation of adducts of sucrose with lithium chloride, lithium bromide, and lithium iodide. Polarimetric studies of the reaction in (homogeneous) aqueous solution have definitely shown that sucrose combines with alkali metal ions in preferred stoichiometric ratio (see Section II,4b, p. 224) Thus, inability to isolate a particular adduct does not prove that the
(58) T. V. Talalaeva, G. V. Taareva, A. P. Simonov, and K. A. Kocheshkov, B d . A d . Sci. USSR, DW.Cliem. Sci. (English Trend.), 696 (1964). (59) J. E. Mackenzie and J. P. Quin, J . Chem. Soc., 961 (1929).

224

J. A. RENDLEMAN, JR.

adduct does not exist. Stability, solubility, and ease of crystallization or precipitation of an adduct can vary widely according to the preparative conditions. Senti and WitnaueP have provided the only information yet available on the stoiohiometry of formation of polysacoharide adducts. Their studies of addition compounds of amylose in aqueous ethanolic media showed that the combining ratio of D-glucose residue to sslt is a function of salt concentration, and that the minimum ratios are approached as the sal! concentratidn is i n c r d . Beyond a certain concentration of salt, the ratio becomes almost constant. The anion plays an important role in determining the magnitude of the minimum ratio for an amybse adduct. Potassium bromide and potassium iodide give adducts of minimum ratio 2: 1, whereas potassium acetate and potassium propionate give 1:1 adducts. A study of the composition of the potassium acetate adduct as a function of salt concentration indicated that two, relatively stable adducts axe formed, the 1 : l and the2:l.

b . Stoichiometry in Homogeneous Systems.-The phase rule is not applicable to adduct formation in homogeneous systems. It is useful only for heterogeneous systems containing multiple phases, at least one of which consists of a pure adduct. Failure to observe a phase for an adduct does not n e c e d y signify the nonexistence of the adduct; it could mean that, under the experimental conditions employed, the solubility characteristics of the adduct permit the existence of a metastable solution. This could lead to a solubility curve whose shape would indicate the absence of complex-formation. For homogeneous systems, combining ratios are difficult to determine. The exceptionally low stability of adducts in dilute solution and the rapidly reversible nature of carbohydrate-salt interactions have, thus far, precluded succeeaful electrometric analyses. WiklwxV8 and Ramaiah and Vishnu,dO by applying the method of Job" to solutions containing sucrose and alkali metal salts, have determined the favored combining ratios for various carbohydratesalt adducts in (homogeneous) aqueous solution. The principle of this method is aa follows. If equimolar solutions of two complexing solutes are mixed in different proportione, the concentration of the complex is generally a function of the proportion in which the solutions have been mixed. The Concentration of the oomplex is maximal when the solutions are mixed in the same proportion as that in which the simple componpnts are present in the complex. The position of the maximum is a function of the molar ratio of salt to complexing agent, and is independent
(60) N.A. Ramsiah and Viehnu, 19hrkma, 9,3 (1959). (61) P.Job, Ann. Chim. (Pd), 0, 113 ( 1 0 % ) .

ALKALI AND ALKALINE-EANTH METAL COMPLEXES

225

of both thc! combined concentration and the concentration of the original dutione (prior to mixing). To make use of the above principle, the complex should possess a measurable property that is unlike the corresponding property of its individual components. This is not always feasible, since, frequently, one of the components has properties very similar to those of the complex. Wiklund13 and Ramaiah and VishnusDmeasured optical rotation. Actually, they measured the combined activity of free sucrose and complexed sucrose in the presence of alkali metal salt. The rotations of these two forms were assumed to be additive. The experimental method consisted simply of mixing equimolar solutions in different proportions, and then measuring the optical rotation a for each mixture. The difference Aci = a - aro (where q,is the rotation for the same solution containing no salt) is largest when the solutions are mixed in the proportion in which the individual components are found in the actual complex. Fig. 3 shows plots of Aa against molar ratio of salt, for various sucrose-salt solutions, each of which has a combined concentration of 1 mole per liter of solution. Other studies

3
Mole fraction of sucrose

FIG. 3.-Effect of Salts on the Optical Rotrttion of Sucrose in Aqueous Solution." (Combined coucentrstion of sucrose and salt is 1 M.)

a26

J. A. RENDLEMAN, JR.

showed that the position of maximum Aa changes very little, and in some c w s not at all, when the combined concentration is raised to 2 moles per s an approximately linear funcliter. The magnitude of the maximum Aa i tion of the product of salt concentration and sucrose concentration, up to a combined concentration of 4 moles per liter. Although some of the curves are not symmetrical, their maxima lie between 0.5 and 0.6 molar ratio. The date thus provided good evidence for the existence of preponderanily 1:1 sucrowalkali metal salt adduct in aqueous solutions of low to moderate concentration of salt. Similar studies of D-fructose with alkali metal salts have given identical resu1ts.m In anhydrous ethanolic media,@potassium acetate and methyl BDglucopyranoside have been shown polarimetrically to combine in the ratio of 1:1.This ratio is ale0 found in the adduct isolated. Application of Job's principlew to aqueous solutions of sucrose and Itlkalineearth metal salts (magnesium sulfate, calcium acetate, and barium chloride), whose combined concentration waa 1 M, gave curves of Aa against concentration that exhibited two maxima, one at 0.5 M sucrose and the other at 0.66 M sucrose. These observations suggest that both 1:1 and 2: 1 carbohydrat-alt adducts can exist in solution.
5. Solvation

Adducts prepared in aqueous media generdy possess one or more molecules of water of hydration per molecule, the number being a function of cation, anion, and the combining ratio of carbbhydrate to salt. Available data on complexes of simple carbohydrates indicate that three molecules of water per molecule may be the maximum for adducts of alkali metal salts; w many as seven have been reported for those of the alkaline-earth metal salts. Most complexes, however, possese only one or two molecules per molecule. Generally, the higher the combining ratio, the smaller is the number of water molecules that can be accommodated by a molecule of the adduct. In the crystalline complex, solvent molecules can be bonded not only to the cation, but also to the anion end the carbohydrate. A detailed x-ray study has mown that, in sucrose-NaBr.2 each BF ion has bonds to one water molecule, one Nae ion, and four carbohydrate hydmxyl groups.'J2 The ability of a complex to possess solvent of crystallization, even when the free individual components themselves are incapable of doing so under the same conditioriu of temporature, is exemplified by the formation of sucrose~NaCl*2 HP at mom temperature. SucroRe and sodium chloride crystallizc from their aqueourJ soliitiotiR in the nonhydmtd form.

Ha,

(62)

W.Cochm, Nature, 167, 87'2 (1V46); C. A. Beover8 and W.Coahran, PTOC. Rog.
Soc, (London), Ber. A , 100, 267 (1947).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

227

Adducts of alkali metd salts prepared in anhydrous alcoholic media generally retain very little alcohol of solvation after being dried under vacuum at mom temperature (see Table I). The unusual ability of adducts of Dglucitol to retain alcohol is probably due largely to the great ability of Dglucitol itself to retain solvent. Adducts of alkalineearth metal salts, however, are more strongly solvated by alcohol than adducts of alkali metal MeOH is relatively stable at 60" at salts. For example,28lactose*CaCla*4 atmospheric pressure; under vacuum ( < 19 mm. of Hg) , a molecule releases only two of the four molecules of methanol. From aqueous alcoholic media, adducts of alkaline-earth metal salts tend to crystallixe as hydrates.

6. Influence of Size of the Cation on Stability of the Complex
Studies of carbohydrate complexes have not yet provided sufficient information to permit the determination of reliable relative reactivities of carbohydrates toward alkali metal ions and alkaline-earth metal ions. Relative electrophoretic mobilities of carbohydrates are a crude, and perhaps unreliable, index of relative reactivity, because the assumption must be made that the mobility is proportional only to the stability of the complex, which is not strictly true. Ease of adduct isolation and the ebulliometric behdvior 'of solutions containing both a carbohydrate and a salt are two other criteria that may be wed to judge reactivity. However, inability to effect precipitation of a complex need not be due entirely to instability of the complex. The order of decreasing ease of isolation (precipitation) is generally Na@> K@> LP. The same order is obtained for the decreasing effectiveness of these ions to promote electrophoretic migration of a carbohydrate. Ebulliometric studieaae on ethanolic solutions of carbohydrates and alkali metal salts have indicated that Na@and Ke ions have very similar reactivities, and that, at salt concentrations of 0.1 M and greater, the percentage of a carbohydrate in the complex form can be quite high. For example, in an ethanolic solution that is 0.0639 molal with respect to D-glucose and 0.1489 molal with respect to potassium iodide, the apparent fraction of mglucose in the complex form is 42%. Only carbohydrates having two or more hydroxyl groups in close proximity to each other show any tendency to complex. The Lie ion was found to possess an immeasurably small reactivity in boiling ethanolic solution. In aqueous solution, the order of decreasing effectiveness in promoting electrophoretic migration of carbohydrateseais (Ca*, Sr@,Ba*@) > Mg2@> Na@> K@> Lie > NHP. No migration occurs in the presence of tetra> Na@> K@> methylammonium ion. In methanol,sethe order is Caz@
(63) J. A. Mills, Riockm. Biophys. Res. Commun., 6, 418 (1961/1962).

228

J. A. RENDLEMAN, JR.

PiH$ > Li*.In both solvents, the ions of alkaline-earth metals appear to be more reactive than those of the alkali metals. The ineffectivenesswith which Lie ions complex with carbohydrates can possibly be attributed to the tightly bound sphere of solvation surrounding the ion. The degree of solvation of alkali metal ions in solution docroltsed4 However, the apparently with increasing atomic weight: Lie > NrP > P. greater ability of Na*, relative to K*, to complex with carbohydrates in aqueous media requires a different explanation. H e r e ,the shorter ionic radius of Nae is probably a more important factor than the sphere of solvation. In general, the strength of a bond between a donor molecule and a metal ion is greater, the shorter the radius and the greater the charge on the ion. The apparently greater stability of alkalineearth.metal complexes, relative to alkali metal complexes, is probably due largely to the double charge on alkaline-earth metal ions. The M g a ion differs appreciably from the other alkaline-earth metal ions in electrophoretic behavior, possibly for +b r e m n offered above for the behavior of Lie. In alcoholic solution, carbohydrates possibly complex with undissociated molecules (ion pairs) of salt, as well &B with free cations. Data from optical rotation experiments1ssuggest that, even in aqueous solution, undiasociated molecules of salt may be associated with the carbohydrate (see Section 11,7, p. 230), On the other hand,free aniona do not appear to complex with carbohydrates in solution (see Section 11, 8, p. 234).
7. Effect of Complexing on Optical Rotation

The ability of a dissolved salt to alter the optical rotation of a carbohydrate is now generally believed to be due to complex-formation. However, rotational phenomena have been largely ignored because of the many, often insurmountable, difficulties that can be encountered in their study. Among the factors that require consideration are multiplicity of donor groups on a polyhydroxy ligand, variability of combining ratio, stereochemical conformation and configuration, and the nature of both the mion and the cation. Because the weakness of the chelate bond has, thus far, rendered impractical a quantitative study of complex stability and equilibria in solution, there ie as yet no obvious means of relating a change in optical rotation to tho extent of complex-formation. Vavrineezmhas shown that the optical rotation of tmcrose depends upon the concentration of both tho Rugar and the salt. Empirical relationships for the chlorides, bromides, iodides, mid scetates of sodium arid potassium were determined.
(64)R.Flatt and F. Benguerel, Helu. Chim. Ach, 46, 1777 (1882). (65) G. Vsvrineoe, 2 .Zwkm'd., 12,261 (1962).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

229

Ramaiah and VishnuMhave, perhaps, come closest to determining the rotation of a pure complex. Working with aqueous solutions of sucrose and of D-fructose, they found that, at a high concentration of alkali metal salt and at a high ratio of salt to sugar (3.5 moles of salt to 40 mmoles of sugar per liter of solution), the addition of more salt to the solutions caused no further change in rotation. This observation was taken as evidence that all of the sugar had been complexed. The rotation was dependeut upon both the cation and the anion. The effect of the cation upon the rotation decreased in the order of Nae > Ke > Li@;that of the anion decreased in the order of Ie > Bre M Cle. The specific rotations for sucrose and
TABLE I1
Specific Rotations of Suoroee and of D-Fructoee in 4 M Solutionsof Alkali Metal Salts in Watep

Salt

None Lithium chloride Sodium chloride Sodium bromide Sodium iodide hdium acetate Potsssium chloride Potassium bromide Potsasium iodide Potassium acetate
a

+66.5

+63.6
+61.7 +68.4 +61.5 +63.9 +62.1 +58.8 +62.1
b

+65 .o

-87.7

-

-97.8

-97.1 -98.5 -99.9

-

From data of RBmaiah and Vishnu."

Concentration of sugar, 0.04 M.

D-fructose in 4 M solutions of alkali metal salts are given in Table 11. Unfortunately, a quantitative interpretation of these data is impossible without knowledge of the combining ratios and of the structures of the adducts. It is quite possible that more than one ion or ion pair is attached to a sugar moleaule at the high concentrations of salt employed. The effect of salt, ooncentratiom above 5 M on the optical rotation of sucrose was not Rtudied by llsmuinh a i d Vixhnu.Ro Therefore, although the sperific rotation R i it~vttriant~ over the range of 3.5 to about 5 A4 salt concentration, it may not be illvariant at higher concentrations. Bigelow and Geschwind67
(66) N. A. Ramaiah and Vishnu, Sharkara, 2, 56 (1959). (67) C. C. Bigelow arid I. I. Geschwind, Compt. Rend. Trav. Lab. Carlaberg, 19, 89 (1961).

230

J. A. RENDLEUN, JR.

have reported that the specific rotation of sucrose in 7.5 M lithium bromide solution is much lower than that in 5.1 Y lithium bromide solution. It is clearly evident from the work of Wiklund” (see Fig. 3) and of Ramdah and Vishnua* that, even at low concentrations of salt, both the cation and the anion affect the optical rotation. Wiklundl*found, at low salt concentration, essentially the same order of decreasing effectiveqess that Ramaiah and VishnuM found at high concentration, namely: Na* > Ke; and I* > BB > OAce > Cle. The pronounced effect of the anions strongly suggests” an important role for anions in the formation of complexes, Because electrophoretic studies in both water and alcohol (see Section II,8, p. 234) offer evidence that free univalent anions do not complex significantly, if at all, with carbohydrates in solution, the effect of the anion on the optical rotation may well stem from an interaction between the carbohydrate and undiseociated salt molecules. Ramaiah and Vishnu66 offered a different explanation for the anionic effect. They suggested that this difference between the anions is due to the difference in their ability to alter the refractive index of water. However, this hypothesis fails to explain why, for example, lead nitratew has no effect on the optical rotation of sucrose. The refractivity of an aqueous solution of lead nitrate is greater than that of an equimolar solution of sodium br0mide.a Salts of alkalineearth metals do not generally differ greatly f r o m salts of alkali metals in their effect upon the specific rotation of carbohydrates. The variation in the rotation of sucmse,(~~6@ D-gZpm-D-gdo-heptose:6 Cr-D-gulose,’& and methyl D-gulopyranosides,” as a function of calcium chloride concentration in aqueous solution, has been mathematically expressed in empirical equations. Aqueous mixtures of sucrose with magneeium sulfate, calcium acetate, barium chloride, and lead nitrate, respectively, have been studied polarimetrically.’O Lead nitrate differs from the other salts in having no apparent effect upon the rotation of sucrose. However, this salt does affect the rotation of Dfructose. The equilibrium rotation of ~guloeeAl. and of D-gZycero-D-gulo-heptosedb v d e s considerably with the concentration of calcium chloride (see Fig. 4). On the other hand, that of pglucose is not greatly influenced. Frush and Isbell” have adequately shown that the influence of calcium chloride is predominantly caused by an adduct formation that leads to a marked shift in the equilibrium between the a- and p-D-pyranose modifications. The higher the concentration of calcium chloride, the greater is the proportion of the u-D modification. The presence of ethanol in the system affects the position of the equilibrium rotation in such a way aa to suggest (68) “International Critical Tablee,” MoGraw-Hi11 Book Co.,Inc., New York, 1930,
(69)

VOl. VII. D. 63 ff * I f , . F. Ja&son and C. Id. Gillis, Bur. Sbndarda Sci. Pamra. No. 376. 126 (1920)

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

231

+1 0
0

82 1 0
9

- 20
0

2

4

6

8
solution

1 0

CaClp, g/IOO ml of

FIG.4.-Effect of Calcium Chloride on the Equilibrium Rotation of ~glycerct~-guloI3eptose.Q (Conaentration of sugar ia 4 g./lOO ml. of solution.)

an increme in the proportion of the CX-Danomer. The mutarotation coefficients for pure D-gulose and wgtycero-wgulo-heptose in aqueous solution are in accord with those of the corresponding calcium chloride adducts. Interpretation of the rotational data on reducing sugars in salt solution must necessarily take into account the possibility that changes in specific rotation of the sugar can be caused, at least partially, by carbohydrat-lt interactions that have no effect upon the proportion of the CY-(D or L) modification. Solutions containing both a salt and a reducing sugar contain two classes of substance: (1) free, uncomplexed sugar consisting of CY and fl (D or L) anomera in equilibrium, and (2) complex4 sugar. The latter class is possibly composed both of charged and uncharged species, formed by the interaction of the sugar with free cations and with undissociated molecules of the salt, respectively. The optical rotation of the solution is, thus, equal to the s u m of the rotations of the different complex species plus the rotation of the free anomers of the sugar. The possibility that both the pyranose and the furanose structures contribute to the overall rotation introduces additional complications. Polarimetric studies have shown that dilution of a solution of a sugar and a salt leads to a lowering in the percentage of sugar in the complex form. For example,@.in a 0.03 M solutionof cu-D-gulose*CaCl2.H2O in water, the percentage of Dgulose in the complex form, estimated from rotatiQnal measurements, was reported to be 0.9%, whereas, at a concentration of 0 . 3 4 M ,the percentage indicated waR 15%.
8. Electrophoresis

Electrophoretic migration of carbohydrates in solutions of alkali metal salts or alkaline-earth metal salts demonstrates the ability of carbohydrates

232

J. A. BENDLEMAN, JR.

to unite with free ions; however, it cannot detect the union of carbohydrate molecules with undissociated molecules of salt. The rate of migration is a function of many variables, such as stability of the complex, the stoichiometry, the concentration of salt, the cationic radius, the favored coordination geometry of the cation, and the size, configuration, and confiorniation of the polyhydroxy compound. Obviously, the electrophoretic data done would not permit the determination of the relative complexing abilities of carbohydrates; however, they do permit qualitative information to be obtained concerning the relative complexing abilities of alkali and alkalineearth metal ions (see Section 11,6, p. 227).
a. Aqueous Mdia.--Mill@ has used cellulose-paper electrophoresis to provide evidence for the existence, in dilute aqueous mlution, of complexes of polyhydroxy compounds with cations of alkali metals and alkalineearth metals. The relative effectiveness of the different cations in promoting migration toward the cathode has already been given in Section 11,6 (see p. 227). No evidence was found for complexing between the carbohydrate and the anions acetate, nitrate, and perchlorate; however, the sulfate ion appeared to have some complexing ability. Of all the polyhydroxy compounds studied in aqueous solution,"* cisinositol exhibits the greatest mobility. The epi- and allo-inositol and cis-quercitol also show considerable movement, but to a lower degree; other cyclitols are less mobile. Reducing sugars and alditols generally show very little or no movement in the presence of Mg*@ and alkali metal ions; all move in the presence of Caa, Sr@, and Baa, but the rates are only moderate. Table I11 gives the relative mobilities of several polyhydroxy compounds in aqueous solutions of various metal ions.

Turn I11
Relative Mobilitles. of Polyhydmxy Compound8 in Aquaour S O ~ U ~ ~ of OXIB Metallic Ionra Compound
Bag
82 26

Md'
20 2 0

Na (s
1 0 3
1 1 2

K '
6 3
1 2

&Inoaitol 6pi-Inodtol GIditol Allitol ~-Talom

13
1s

5

1 1

1

4 Cstionio movements are given as peroentagsa of the anionio movement (about 10 cm.)of p-tnitropenienmlfonic acid on the mme strip, with 2,3,Btri-O-methyl-~-glucose M I the marker for Bero migration. The eleotrolyte was a 0.1 M solution of the metal acetate in 0.2 A4 Boetio acid. Electrophoreds wae performed for 1 hr. at a potential gradient of about 20 v./om. on Whatman No.4 paper under a uniform preseure of 0.4 atm., with oooling by tep water. Compounds were applied as 0.1 M aqueous solutione.

ALKALI AND ALKALINBEARTH METAL COMPLEXES

233

OH I

FIG.6.4-Inoaitol.

Mills6*attributed the outstanding complexing power of cis-inositol to the presence of three axial hydroxyl groups in a chair conformation (see Fig. 5 ) ; these groups are suitably oriented for close approach to a cation. Arrangements of three hydroxyl groups that are close enough together to be associated with a cation can be discerned in certain of the conformations of the other cyclitols that show catiohc migration. Charley and Saltmane studied the migration of radioactive Cate in the presence and absence of lactose in aqueous solution at pH 7.0. The solutions were buffered with sodium hydrogen carbonate. The inability of calcium to migrate in the presence of lactose indicated that Cat” had reacted with the sugar to form a soluble, uncharged complex.

b. Alcoholic Media.-Studies of the interaction of polyhydroxy compounds with alkali metal salts in alcoholic solutiona6have shown that electrophoretic migration is much faster in methanolic and ethanolic solution than in aqueous solution. This observation indicates that the stability of a complex is much greater in alcoholic than in aqueous systems. Even 2,3,4 ,6-tetra-0-methyl-D-glucose migrates at a small, but measurable, rate. Glass-fiber paper, instead of cellulose paper, was employed in the studies, in order to eliminate the possibility of errors that could arise from the formation of complexes between cellulose and metal ions. The effectiveness with which the solvent promotes migration decreases > water. The extremely low stability of in the order: methanol > ethanol > complexes in water can be explained by the relatively great tendency of metal ions to associate with water molecules.70The difference between the rate in ethanol and that in methanol can be attributed, at least partly, to the fact that salts we more highly dissociated into free ions in methanoP; a higher concentration of free cations would permit a higher concentration of positively charged carbohydrate species. Relative rates of carbohydrate migration are spread over a wider range of values in methanol than in ethanol: this can be attributed to a lower
(70) A.

E. Martell and M. Calvin, “Chemistry of the Metal Chelate Compounds,” PrenticA3aI1, New York, 1962, p. 239. d ,and N. N. Salstnikov. Uch.Zap. Kb’kouak. Qos. (71) N. A. I z d o v , E. I. V Univ. Tr. Khim. Fak. i Nauchn.-Isaled. Znut. Khim., 71, No. 14, 29 (1956).

234

J. A. RENDLEMAN, JR.

stability of complexes in methuriolic media. Stability o f chelates is known to bo greater in solvents of lower than in those of high The relative rates in an alcoholic medium are virtually independent of the nature of the anion (see Table IV) ; this constitutes a strong indication that free anions do not complex with carbohydrates to a n y significaiit extent. On the other hand, anions do play an important role in determining absolute
TABLIC IV
Rater of Eleatmphoreticr MigratJona of ~ - X y l ~ r and e D-GIUOO i n ~ Methanolic %duthnr of Varloua Electmly te#

Salt

M Temp., Migration "C.' rate,
mm./hr.

Maib

Temp., Migration "CSb rate, mm./hr.

M~lb

NaI NaBr NaCl NsOAo

0.03

0.16 0.16 0.03 0.16

38

38 34

-

4.0

0.61 0.57 0.68 0.66
0.60

L i c l

NI4C1 CaCl,

0.30 0.30
0.16 0.16

0.16

34 38 43
37 39

M s 1 ~ Ha0 6

44

7.2 3.8 6.7 4.7 0.1 3.0 14.0
0.8

-

47 38

34

10.1 8.2

-

0.64
0.66

-

-

0.06 0.67 0.56
0.2

38 -

6.6 4.3

39 44

-

0.64 0.40

-

10.2 1.7

0.40
0.41

-

a Zone electrophoresis WM on glass-fiber paper at a potentiel gradient of 16.7 v./in.; referenae (nonmigrating) oompouad waa ohrysene, whioh ie visible under ultraviolet lieht when dry, but not when wet; oompounds were applied aa alooholio solutions (0 .O& 0.08 M,solubility permitting); all carbohydrates migrated towerd the oathode; Mart. rate of migration relative to that of ~-ribose.'Maximum temperature to whioh the system rose.

-

rates of migration. In promoting migration, metal halides are more effective than the corresponding acetates. In methanol, magnesium acetate causes no migration, whereas magnesium chloride (hexahydrate) effects a memurable movement that is roughly comparable to that of ammonium chloride and lithium chloride. This differencein the ability of different salts (poseessing a common cation) to promote migration is possibly due to a difference in the degree of ionic dissociation, and not to complexing between the carbohydrate and the free anion.
(72) A. Brtlndstr6m, Arkiu Kemi, 7, 81 (1964).

ALKALI APSD ALKALINE-EARTH METAL COMPLEXES

235

Table V contains a list of the absolute rates and relative rates of niigration for various carbohydrates in methanolic solutions of potassium acet'ate and sodium acetate. Table I V shows the effect of different salts on the absolute rates and relatlve rates of wxylose and wglucose in methanol. The ability
TABLE V
Rate8 of Electrophoretic Migrationa of Polyhydroxy Compounds in 0.3 M Solution8 of Potaadum Acetate and Sodium Acetate in Methanol= KOAc Compound Temp.? Migration "C. rate, mm./hr.
42 48 42 48 7.3 10 .a 4.6 6 .O 7.4 9.1 4.8 10.2 4.1 4.8
MRib

NaOAc Temp.? Migration "C. rate, mm./hr.
37 38 38 9.4 9.3 4.7
MRib

D-Xylose

0.60
0.58 0.73 0.90 0.64 1 0 -68 0.64

1.oo 1.oo

38 38 38 a7 37 37 37

-

0.60

1.oo 1.oo

-

D-Lyxose D-Arabinose a-D-Glucose fructose Sucrose Maltose Ibffinose Meleaitoee 1,6-Anhydro-p-D-g1uc0pyranose Methyl a-D-glucopyranoside B momer Methyl a - ~ msnnopy ranoside D-Glucitol Erythritol
Q

48
48 42 42 42 42

6 .O

.oo

42 42 42

-

-

-

-

7.8 4.3 7.9

0 . 4 6
0.84

0.65 0.83

-

-

-

4.3 3.4 12.3

0.36 1.31

0.4

4.6 4.4 6.7 6.2 4 .O

0.63 0.62 0.94

-

-

-

42 42

0.69

0.66

-

-

See Table IV, footnote a.

* Maximum temperatwe to which the system rose.

of a cation to promote migration in methanolic salt solution decreases in > Lie, the order: Ca2e > Na" > Ke > NH4@ There is no clear correlation between the electrophoretic data available and the geometry of polyhydroxy compounds. The unusually high, relative may be due to the rate of migration of 1,6anhydro-~-wglucopyrrtnose presence of the two axially oriented cis hydroxyl groups on G 2 and C-4 of a boat conformation. Such an orientation in a rigid molecule would be

236

J. A. RBINDLDMAN,

JR.

uxpoc:tcd to givo u uo,omplex of stability higher than average, aimilar to that of cia-inoeitol in aqueous media, Electrophoresis in nonaqueous media may be an effective means of separating polyhydroxy compounds from mixtures which would otherwise be difficult to resolve. Subsequent separation of salt f r o m carbohydrate could then be agcon@ished by means of ion-exchange techniquee.

9. Structure of the Complex
The true structure of a carbohydrate-salt adduct can be determined only from detailed x-ray diffraction studies, a few of which have been made. Such studies enabled Beevers and Cochrad2 to determine the complete structure of suprose-NaBr.2 HSO. Each Na@ion was found to have sixfold coordination, with almost regular octahedral symmetry. The Bre ion also has sixfold coordination, but the coordination group has no regular shape. Each Na@ ion is close to one Bre ion, two water molecules, and three carbohydrate hydroxyl groups. There are only two direct intermolecular bonds between hydroxyl groups themselves; the remaining hydroxyl groups are linked through Na@ and Bre ions and the water molecules. The hydroxyl groups of both the D-glucose sjld the D-fructose moiety participate in the bonding with oation and anion. The separation between Na@and B e in the complex (2.94 A.) is actually maller than that in pure, crystalline sodium bromide (2.98 A,). Senti and WitnauerSO made similar x-ray studies of potassium salt adducts of amylose. Adducts having a 2:l ratio of D-glucose residue to salt (iodide, bromide, formate, acetate, and bicarbonate) were found to have tetragonal lattices with fourfold screw symmetry; 1 : l adducts (acetate and propionate) were orthorhombic. The structure of the tetragonal adducts is determined, not by amylose-amylose contacts, but lsrgely by amylose-cation contacts. Amylose-anion contacts appear to be of minor importance, for the unit-cell dimensions are relatively insensitive to the size of the anion in an isomorphous series of potassium salt adducts. The elements of symmetry in these complexes indicate that the D-glucose residues in the amylose chain, or, at least, those chains in the cry-stalline portions (of the complex) that are responsible for the discrete diffraction patterns, are equivalent. The conformation of the &glucose residues was not determined. By means of three-dimensional, x-ray diffraction data, Hybl, Rundle, and W i l l i w solved the crystal and molecular structure of the potassium acetate adduct of cyclohexsamylose, a Schardinger dextrin. Cyclohexaamylose is a macro-ring consisting of six D-glucopyranose rings connected by a - ~(1-4) -glucosidic linkages. Because there are six D-glucopyranose

ALKALI AND ALKALINHbIMRTH METAL COMPLEXEB

237

residuw in each turn of the helical structure of amylose, cyclohexaamylose is an ideal model compound on which to base a study of both the conformation of the D-glucopyranose rwiduea and the geometry of the CY-D( 1 4 ) -gluoosidic linkage in amylose. The analysis of the complex 2 cyclohexaamylose*3.08KOAc-19.4 HtO showed a translational stacking of the cyclohexaamylose molecules, to yield a cylindrical, carbohydrate canal structure. The a-n-glucose residues are all in the pyranose form, and this i s in,the C1 (D) cgnfonnation (la283e4e5e). Each molecule of cyclohexaamylose has six pocket positions along its surface that are occupied by water molecules and potaasium ions. The potassium ions are in distorted octahedral environments, outside the carbohydrate channels. Two of the three acetate ions in each unit cell are at highly anisotropic, disordered sites inside the cyclohexaamylose macro-rings. The third acetate ion has not yet been accounted for. There are intermolecular hydrogen bonds between the hydroxyl groups at C-2 and C 3 of each pair of contiguous P-glucose residues.

111. COMPLEXES FROM THE INTERACTION OF CAF~BOHYDRATES WITHMETAL BASES
The interaction of strong bases with polyhydroxy compounds, a 1though extensively studied, has not yet been fully clarified. The available evidence indicates that the removal of a proton by a basic anion, to give an alcoholate (reaction 1 ), and the formation of an adduct (reaction 2 ) can both occur. In alcoholic media, reaction 1 has been definitely shown to occur. However, in aqueous media, differentiation between reactions 1 and S has not yet been possible. ROH + MB ROM + HB (11
ROH

+ MB

ROH*MB

0)

where M = a metal ion; and B = OH*, CNe, or an alkoxide ion. Alcoholates of polyhydroxy compounds will be included in the category of complexes, because of the probability that most, if not all, of them are stabilized by inner chelation of the metal ion with neighboring hydroxyl groups, similar to that illustrated in Fig. 6, Adducts, a h , should be stabilized by chelation.

FIQ.&-Chelate Structure of

the Sodium Alcohblste of a 1 ,ZDiol.

238

J. A. RENDLEMAN, JR.

There is considerable supporting evidence for the existence of undissociated inner chelates in aqueous solution. Potentiometric pH meaaurementsn and nuclear magnetic resonance studies74.76 of tiqueous solutions of a- and phydroxy carboxylic acid salta indicate an appreciable association between the alkali metal cation and the organic anion; alkaliie-earth metal ions associate even more strongly than do alkali metal ions. The stability of the complex increases a,s the radius of the cation decreases." All alkali metal cations form complexes with malate ion. The observation that the tetramethylammonium ion is much less strongly bound than an alkali metal ion is understandable, in view of the fact that quaternary ammonium ions are known to resist being solvated, even by water. The constants of formation of various metal kojates, including that of calcium, have been determined by potentiometric titration?' Possibly, the &membered, chelate ring of the metal kojate (see Fig. 7) contributes significantly to ita stability. Chsistepenn has suggested that chelation between alkali metals and

FIQ. 7.--Chehte Structure of Calcium Kojate.

pyridoxal plays a role in biologic4 transport; he studied both the alkali metals and the alkaline-earth metals. The formation of complexes of wgluconate ion with alkaline-earth metal ions has been studied." Supporting the concept of the existence of alkali metal hydroxide adducts is the isolation of highly crystalline, 1:1 m o l a r adducts of potassium hydroxide with certain tertiary acetylenic carbinole and glycols.~~ However, these complexes cannot be strictly compared to alkali metal hydroxidecarbohydrate adducts, because of the probable involvement of the r shell of the carbon-carbon triple bond. Weizmann,@on the other hand, has reported the formation of potassium hydroxide complexes of acetala and of
(73) L. E. Erickmn and J. A. Denbo, J . Phyu. C b m . , 67, 707 (1983). (74) 0. Jsrdetriky and J. E. W e & , J . Am. Chem. Soc., 84, 318 (1880). (76) L. E. Erickson and R. A. Alberty, J . Phy8. Chem., 66, 1702 (1062). (76) B. E. Bryant and W. C. Fernelius, J . Am. Chem. Soc., 76, 6361 (1964). 1087 (1066). (77) H. N. ChristenMn, dlcimcc,ll, (78) R. K. Canand A. Kibrick, J . Am. Chem. Soc., 60, 2314 (1038). (70) R. J. Tedeschi, M. F. W i l s o n ,J. Scanlon, M. Pawlak, and V. Cunicella, J . &g. C h . ,48, 2480 (1963). (80)C. Weimmann, B r i t i s h Pat. 682,191 (1948); C h .AMroatb, 41, a630 (1047).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

239

dialkyl ethers of glycols; this indicates that alkoxyl groups call act 89 electron donors, in the same capacity a s hydroxyl and keto groups. However, the electrondonating ability of an alkoxyl group would undoubtedly be much less than that of a hydroxyl or keto group. Both the pH and the conductivity of an aqueous solution of sodium hydroxide are markedly decreased by the presence of polyhydroxy compounds. By assuming that the reactions are due entirely to the removal of one or more hydroxylic protons, several inveatigabrs have calculated ionization constants for a number of simple carbohydrates.a1-’J4 A tabulation of most of these constants (as p&) is availab1e.w Hirsch and Schlage showed that the reactions are reversible, and that the extraordinarily high “acidity” of reducing sugars is not due to enolization of the aldehydo form. An aldehydo or keto group is not essential for high reactivity toward alkali metal hydroxide, as shown by a comparison of selected p& (18’) , sucrose (-12.6) , and ~-glucitol (13.57). values for D-glucose (12.43) The lesults of calculations in which these values were used indicate that, for aqueous solutions that are 0.16 M in carbohydrate and 0.16 M in sodium hydroxide, the extents to which D - ~ ~ U C O Ssucrose, ~, and wglucitol react to give alcoholate and water (reaction 1 ) are, roughly, 70%, So%, and 20010, respectively. These percentages must, of course, be viewed with a s based on the assumption that some skepticism, since their calculation w the contribution of reaction 8 is negligible. Apparent ionization constants wheat starchJW and alginate.” have also been determined for In Fig. 8, tha number of hydroxide ions consumed per molecule of carbohydrate, in an aqueous solution that is 0.25 M in sodium hydroxide, has been plotted against the initial concentration of carbohydrste. The ease with which a single molecule of disaccharide can effect the disappearance of more than one hydroxide ion is readily apparent. Starch and dextrin are unable to consume more than 0.5 hydroxide ion per wglucose residue. s molecules of Should the &values for oligosaccharidea be expressed a sodium hydroxide consumed per monosaccharide residue, it would at once be apparent that the difference between oligosaccharideand plysaccharide is not very large. Furthermore, on this basis, pure *glucose is more reactive than any of the oligosaccharides studied. Smoleiiski and Porejko8astudied the pH of aqueous solutions containing (81)L. Michaelie and P. Ilona, Biocham. Z . , 4@,232 (1913). (82)J. Thamsen, Ada C b m . S m d . 6, 270 (1952).

(83)P. Hirech and R. Sohlags, 1.Phyeilc. Chern. (Leipzig), A, 141, 387 (1929). (84)P.Souchay and R. &had, BuU. Soc. Chim. Fronts, 819 (19M)). (85) B.Capon and W. G. Overend, Advan. Ca&oh&ate Chsm., 16,32 (1960). (80) 8 .M. Neale, J . Teztile Inst., 2 0 , 373T (1929). (87)8.P.Sari0 and R. K. Schofield, Proc. Roy. Soc. (London), S s d .A, 186, 431 (1948). (88)K.Smolehki and 8. Porejko, Roczniki Chem., 16,281 (1936).

240

J. A. RIJNDLEMAN, J R .

8FIO.8.--Relation between Consumption of Sodium Hydroxide and Conoentration of Carbohydrrh i n Aqueous Solution." (Concentration of polymcaharidea h a x p r e g e e d i n mole of p.plucose &due per liter of solution. Initial conaentration of sodium hydroxide is 0.26 M.e Imole of sodium hydroxide oonsumed psr m o l e of carbohydrate. Carbohydratea: D-glucoee 0, D-fruotOrw 0, BUDIOBB X, leotoee 0, meltose A, staroh @, and dextrin A.)

both calcium hydroxide and sucro~e.Their reaults were similar to those obtained with sodium hydroxide and ~umom, in that sucrose behaves a~ a weak acid. Measurementsof the pH of alcoholic solutions containing alksli metal hydroxide and carbohydrate have not yet been reported. A pH meter is capable of giving reproducible e.m.f. values in alcohol or alcoholwater solvents,m but the pH numbers read from the inetrument ere subject to no simple, clear interpretation in term of chemical equilibrium. There isJ at this time, no universal scale of acidity for mlventa differing in w a t e ~ alcohol ratio. Studies of the heats of reaction of sucrose with sodium hydroxide, and of sucrose with barium hydtoxide, in aqueous solution have shown that neither reaction ie as simple aa that of an wid-base neutralismtion; the experimental reeults indicated the possible involvement o f hydrogen bonding.88. After devoting considerable thought to the problem of the chemical structure of alkali cellulose, B l e a M I and Lositsksyaw have suggested (89) R.Gi. Bab, M.Pesbo, and R. A. Robiaeon, J . Phya. Chenr., 67, 1838 (1983). (SQa) E.Calvet, H.Thibon, and P. Leydet, Bull. Soc. Chim. Fmncs, 2187 (1%). (90)8. V, BleehinskiI and 8. F. Lositateya, Tr. I&. Khim. A W . Not& Kim. 88R.
I , 73 (1961).

ALJCALI AND ALKALINE-EARTH METAL COMPLEXES

241

that alkali cellulose exists as a mixture of metal hydroxide adduct and metal alcoholate, the adduct preponderating. They hypothesized that the ratio of adduct to alcoholate vanes with’thesize of the metal cation; the stronger the ion-dipole bonding between a cation and a carbohydrate hydroxyl group, the greater should be the proportion of the adduct. However, there is no unequivocal evidence to confirm their hypothesis. The same attractive force between cation and hydroxyl group that would stabilize an adduct might just as readily stabilize an alcoholate. The metal cation of a carbohydrate alcoholate is, very probably, chelated to neighboring hydroxyl groups.
1. Reactions in Aqueous Media

In homogeneous, aqueous solution, alkali metal hydroxides react with carbohydrates to produce negatively charged carbohydrate species. Although the general feeling among chemists is that these species are free alcoholate anions, the possibility that they are composed, at least partially, of carbohydratehydroxide ion adducts cannot be dismissed. There is electrophoretic evidence that the S0,S ion is capable of weak bonding with polyhydroxy compounds.” If this is true, perhaps the hydroxide ion and certain other oxyanions are likewise capable of weak bonding. Detailed x-ray studies of crystalline lithium hydroxide monohydratee’ and of sodium hydroxide tetrahydratee* have shown that hydrogen bonds can form between water molecules and hydroxide ions. Because contact of a metal ion with a hydroxide ion can increase the dipole moment of a hydroxide ion, and enhance its ability to form hydrogen bonds with hydroxyl groups,g* it is conceivable that undissociated, metal hydroxide molecules should form hydrogen bonds with carbohydrate hydmxyl groups more readily than can free hydroxide ions. In either situation, bonding would be expected to be stronger, the greater the partial positive character of the hydrogen atom on the hydroxyl group of the carbohydrate. Evidence for the existence of OHe(ROH)s, where ROH is an alkyl alcohol, in solutions of alcohols in benzene and in nitrobenzene, has been obtained by Agarwal and Diamond.” Their studies, which involved the use of various quaternary ammonium hydroxides, indicated that the hydroxide ion is capable of associating with three alcohol molecules, regardless of whether it IR in the free form or nmciated with a cation.
(91) R. Pepinsky, 2 . Krist., la, 119 (1940). (92) G. Beurskena and G. A. Jeffrey, J . Chem. Phge., 41, 924 (1904). (Q3) A. F. W e l l s ,“Structural Inorganic Chemistry,” University Press, Oxford, 1947, p. (94) B. R. Agarwal and R. M. Diamond, J . Phys. Chem., 67, 2785 (1963).

360.

242

J. A. RBNDLEMAN, JR.

Makolking6studied the rate of exchange of l 8 0 between labeled water and alkali cellulose, and between labeled water and the trisodium alcoholate of cellulose, and concluded that merceriration proceeds by reaction 8 ; that is, alkali cellulose is an adduct. He b d his conchpion on the fact that theN is no measurable exchange between water and the alooholate, whereas there is measurable exchange between alkali cellulose and water. H i s conclusion, which is based upon difference of rate, is, however, not necessarily valid. The trisodium alcoholate may not possess a structure that is as accessible to water aa is that of alkali cellulose. The complex 2 sucrose-NaOH, prepared in aqueous alcoholic media, loses the elements of w a t e r at 110"under vacuum, t o give the corresponding alcoho1ate.m Treatment of the alcoholate with glacial acetic acid permits B U C ~ Y to ) ~ be ~ recovered in 90% yield. Ahali metal complexes may be analyred for their metal content by simple acidimetric titration. Analysis for adduct (hydroxide) content is more involved, and entails the aseumption that there can be no water of hydration attached to an alcoholate anion. The methodm involves: first, dissolving the complex in anhydrous methanol, and then, treating the resulting solution with an appropriate anhydrous acid, such as tartaric acid. The acid servea to convert any hydroxide ion into water (reaction S), which can then be quantitatively determined by titration with the Karl Fischer reagent. The amount of water thereby measured is assumed to equal the amount of hydroxide originally preaent i n the adduct. By similar means, it has been shown@@. that, when alkali cellulose (prepared in an aqueous sodium hydroxide medium) is dried under vacuum at 6 5 O , the reaulting material consists of both alcoholate and hydroxide adduct. There is the poseibility that complexesieolated from aqueous (or aqueous alcoholic) media are not hydroxide adducts, but, instead, hydrated alcoa r l Fisoher analysis does not, holatea. The above adaptation of the K unfortunately, distinguish between a hydroxide ion and a water molecule. Such a distinction could posaibly be made by detailed x-ray analysis; however, neither alcoholates nor adducts have, as yet, been obtained in a form suitable for such a study. Tablo VI contains a comprehensive list of known alkali metal hydroxide ndducts. Complexm prepared by the interaction of carbohydrates or naotiitcn of ctirhohydmtewwith nlkdi motd hydroxide in anh!/drrgusalcoholic
(95) I. A. Mskolkin, Zh. Ob8hch. Khim. 12, 306 (1042). (96)J. A. Nendleinn, Jr., J . Org. C h . ,81, 1845 (1966). (!IOU) E.(ifligerand H.Noh, Helu. Chim. A&, 40, 660 (1987).

TABLE VI
Adducta of Carbohydrates with Alkali Metal Hydroxides Carbohydrate &and)

MOH

Molar r ~ t i o , ~ ligand-MOH

Solvent of solvation, molecules/cation

Solvent m e d i u m

References

s
z

Amylose

CsOH LiOH KOH NaOH KOH

3: 1 3:l 1:l 3:l 1:l 1:l 1:2 3: 1 1:l 2: 1 2: 1 3:2 4:3 3: 1 1:l 1:l 1:l 2: 1 2:l 3:l 4:3
f:2

-

-

H&EtOH H&EtOH H&-EtOH H&EtOH H&EtOH H&EtOH HaEtOH Hto H&EtOH HeEtOH H Z O H&EtOH H Z O H@-EtOH H+EtOH H&EtOH H&EtOH

97 97

T r
c (

*

98

97

U

*

~Arabiiose

96

k

cellobiose
Celluld

99 100
9% 102

CsOH
KOH

101

E T ss
c1

-

HaEtOH

RbOH

NaOH

Z O 3H H Z O

102 101 102 102 101 103 104 104 101 103 104 104

i ! b
0

103

-

5 Fi

L

3 HtO

-

-

HZO-EtOH HeEtOH HeEtOH

El20

h3

8

Tmuc VI (Continued)
Adducts o f Gubohyhtca w i t h Alkali Metal H y M d e s
Carbohydrate

(ligand)

MOH

Molar mttio,ligand:MOH

Solvent of
&ation,

Solvent m e d i u m

Refere-

molec&/~tion

D-GlUcoSe

LiOH KOH NaOH KOH

1.S:l 1 : l 1:1*2 2.0:l
(2-3) KOH

no &O
0 . 2 &o 0.1 H a , 0 . 0 8 EtOH

-

H&-EtQH H&WH H&EtOH H&WH H&EtOH HJO-EtOH H&-EtQH H&-EtOH

96 99 96 96

KOH

NaOH NaOH KOH NaOH

1: 3 1 : l 1 : 1 1:1 1:l 1:l 1:2 1 : 1 2 1: 2 :1 2:1 1:l 1:2 1:3

H&-EtOH H&WH
H&EtOH H&-EtOH

106 106 107 107
108,109

-r
Q

NaOH
Starch

NaOH

Hto

H Z O

Hso
H&EtOH HeEtOH H&EtOH H&EtQH

108, 109 108,109

SUCroSe

CsOH KOH

96 110,111 110 110

NaOH

1:l 1 . 7 : l 1 : 1 . 8 1:2.5 1:5.1

0.3 HzO

-

no HtO 0.22 HsO, 0.04EtOH

HeEtOH HeEtOH HeEtOH H&EtOH HtO-EtOH

111 96 96 96

96

r

The Jigand in a polysaccharide adduct is the D-glucose residue. In other adducts, it is the entire carbohydrate molecule. b For adducts of cellulose with lithium hydroxide, 6ee the discussion in Section III, la (p. 2 5 0 ) .
6

r

E
U

(9i)F.R. Senti and L. P. Witnauer, J. Am. Chern. Soc., 7 0 , 1438 (1948). (98)W. J. Heddle and E. G. V. Percival, J. Chin. Soc., 1690 (1938). (99)E. G.V.Percival, J. C h .Soc., 1160 ( 1 9 3 4 ) . (100) E.G.V. Percival and G. G. Richie, J. Chem. Soc., 1765 (1936). (101) E .Heuser and R. Bartunek, CeUu2oseehemie, 6,19 (1925). (102) K . G.Ashar, J . Cltim. Phys., 48, 583 (1951). (103) I. Sakurada and S. Okamura, Kouoid-Z., 81, 199 ( 1 9 3 7 ) . (104) G.Champtier and J. Nhl, BuU. Soc. Chim. FMW, 930 ( 1 9 4 9 ) . (105) A. Herzfeld, Ann., aa0,206 ( 1 8 8 3 ) . (106) A. Bau, Z .Vet. Deoct. Zzccker-lnd., M,481 (1904). (107)K.Beythien and B. Tollens, Ann., 266, 195 (1889). (108) G.Champtier and 0. Yovanovitch, J. Chim. Phys., 48,587 (1951J. (109) 0. Yovanovitch, Compt. Rend.., 292, 1833 ( 1 9 5 1 ) . ( 1 1 0 )E .G.V. Percival, J. Chem. Soc., 648 (1935). (111) E. Soubeiran, Ann., 43, 223 (1842).

-.

>

E
E
3 r

246

J. A. RENDLEMAN, Jll.

TABLE m1
Complexea of Carbohydrates with Alkaline-earth Metal Hydroxides Carbohydrate (ligand) Hydroxide Molar ratio, ligandiaation

medium
HeEtOH HgO-EtOH HlO HZO-EtOH HQ-EtOH HsO Ha-EtOH Ha-EtOH Ha-EtOK HaO-EtOH H&EtOH HiO-EtOH HsO-EtOH HlO HiO-EtOH HaEtOH Hi0 HlO HiO-EtOH HlO Ha0 HgO-EtOH H&-EtOE HlO Hi0 HqO HsO-EtOH H&EtOH H&EtOH

Solvent

References

drabhose
p.F~ctose D-G~UCOW

Lactose Maltose Manninotrioso
RaffiIlOee

Be Sr Ca B E Ca Ca Ba Ca Sr Blb Ba Ca

2:1 2:1 1:l 2:1
1.1

4

1:l
1:1 1:l 1:1

Stachyose Sucrose

Sr Be

S r

1:l 1:l 1:2 1:l 1 : 3 1:2 2:3 1 :2 1 : 6

Ba
C+

S r
Q

,P-Trehalose

CR

D-XJ’hi?

BE Sr

1:1 1:l 3:l 1:1 1:l 1 : 2 1 : 3 1:l 1:2 2:3 2:1 2: 1

HlO

4 69,112,113 114 59 69 105 59, 105 105 115 107 107 116 107 107 115 117 117 59 19 19 69 69 59, 118 69 19,59 19,69 119 4 4

(112) E. Poligot, Compt. Rand., 90, 163 (1880). (113) H.Wintw, Ann., 144, 296 (1888). (114) H. Will, Arch. Phrni., 816, 812 (1887). (116) C.Tanret, Bull. Soc. Chim. Frame, 87, 947 (1802). (118) L.Lindet, J . Fabr. S w e , 81, 19 (1880). (117) G. Tanret, Compt. Rend.,166, 1620 (1912). (118)P. Horsin-DOon, Bull. doc. Chim. Frunce, 17, 166 (1872). (119) I. Schukow, 2. Ver. Dad. ZuckerZnd., 50,818 (1800).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

247

rneclia*'WJ'O are not listed, b e c a w of the probability that they are, preponderantly ,alcoholates. With the exception of magnesium hydroxide, alkaline-earth metal hydroxides are similar to alkali metal hydroxides in that both types are strongly basic and highly dissociated in aqueous solution. Interaction of an alkaline-earth metal hydroxide or oxide with a carbohydrate results in an increased solubility of the hydroxide or oxide, apparently through the formation of either an alcoholate, a carbohydratemetal hydroxide adduct, or a carbohydrate-metal oxide adduct. Mackenaie and Quin"' have suggested that, in the compounds of reducing sugars with calcium hydroxide, the calcium is united with the hydroxyl group at C-1, possibly in the form
-CH

/ \
0

0

\ /

Ca

or

\

HC-O-Ca-OH,

/

and that, in the compounds of nonreducing sugars, the calcium is present simply aa calcium hydroxide bound to the carbohydrate in a manner similar to that of a salt in a salt-carbohydrate adduct. Because there has been little experimental work with either alkaline-earth metal hydroxides or oxides since that of Mackenzie and Quin:Q any attempt to draw conclusions concerning the composition and structure of the complexes must await further experimentation. For convenience, therefore, all complexes formed from hydroxides or oxides will be called adducts of alkaline-earth metal hydroxides, and they are listed aa such in Table VII. Water of hydration is omitted from Table VII for several reasons. (1) Many investigators did not consider water of hydration in determining the formula for a complex. (2) In all instances where water was determined, it w a s determined indirectly. The difference between the weight of a complex before and after the complex had been subjected to dehydrating conditions (for example, 0 0 ' ) was often assumed to equal the weight of water under vacuum at 1 of hydration. (3) Dehydration of a complex could involve either the removal of water of hydration or the removal of water formed by chemical reaction, or both. And (4) , there is considerable question concerning the chemical structure of the complex. There has been no reported isolation of a magnesium hydroxide complex. Attempts by BenedikP to prepare such a complex of sucrose were unsuccessful. A barium hydroxide adduct of amylose was prepared by Senti and WitnaueP7 by means of an exchange of barium hydroxide for potassium hydroxide in a potassium hydroxide-amylose adduct; however, the exact
(120) R.Benedikt, Ber., 6, 413 (1873).

248

J. A. RENDLEMAN, JH.

composition of the adduct was not reported. Dextran, also, reacts with hydroxides of calcium, strontium, and barium to give complexes; these complexes have pravidg a method of fractionating dextran for molecular size,'21 the process being based upon the fact that fractional acidification of an aqueous suspension of the complex results first in the dissolution of the component having the highest molecular weight. Dialysis studies'" of d u l o s e and dextran in aqueous solutions of barium hydroxide, sodium hydroxide, and "cadoxene" (cadmium hydroxide in ethylenediamine) showed no difference in complexing ability between the two polysaccharides. Furthermore, in solutions having equal base normality, Ba2@, Nae, qnd Cd(e$hylenedismine)P had loughly equal complexing abilities.
a. Stoichiometry of Alkali-Metal Hydroxide Reactions.-The combining ratio for alkali metal hydroxide adducts is variable, as with that for alkali metal salt adducts. The ratio is dependent on the concentration both of hydroxide ion and carbohydrate, and on the size both of alkali metal ion and carbohydrate. At low concentrations (<0.1 M ) of hydroxide, there is a tendency for simple carbohydrates to form adducts having a 2: 1 ratio of carbohydrate h i i ratio is also found in many alkdi metal salt adducts prepared to metal. T at low concentrationsof salt (seeSection 11,4, p. 222) and in most alcoholate complexes prepared at low concentrationsof base (see Section 1142, p. 259). With increasing concentration of hydroxide, particularly where the ratio of carbohydrate to hydroxide in the reaction mixture is small, oligosaccharide8 tend to form adducts having one or more molecules of metal hydfoxide per molecule of carbohydrate.wnsg The greater the hydroxide concentration, the greater is the hydroxide content of the resulting adduct. One molecule of sucrose appears able to support as many as five molecules of sodium hydroxide." One molecule of cellobiose can accommodate at least two molecules of potassium hydroxide; maltose can accommodate three.8gLactose is intermediato, forming an adduct possessing between two and three molecules of potassium hydroxide per molecule." Hydroxide adducts of nonsugars and of sugar derivatives have been given but little attention. Heddle and E. G. V Percival" reported the preparation of 1:l adducts of potassium hydroxide with methyl &-Dglucopyranoside and methyl 8-Sglucopyranosidein alcohol-ether containing only a few percent of water. However, Rendleman has repeated the preparation of the 8-D-glucopyranosidecomplex, using Percival's procedure, and has found that the product actually consists of 29% of adduct and
(121) E.L. Wimmer, U.8.Pat. 2,686,679; Chem. Abefmcte, 4% 7878 (1066). (122) H. Vink, Maktomol. Chem., 76, 86 (1904).

ALKALI AXD ALKALINE-EARTH METAL COMPLEXES

249

71% of alcoho1ate.m The reportm of the preparation of the p o t m i m i hydroxide adduct of methyl &D-glucopyranosidein an anhydrous ethanolic system is also largely incorrect. Konaqueous systems, &s will be discussed later, favor proton removal and, therefore, alcoholate formation (see , Section 111,2, p. 255). Senti and Witnauerg’ treated amylose with lithium hydroxide, potassium hydroxide, and cesium hydroxide, respectively, in 25% aqueous ethanol, to obtain products having a 1:3 ratio of hydroxide to D-glucose residue. However, adducts having this ratio are stable, in the preparative medium, over only a limited range of hydroxide concentration. The stoichiometry is variable, just, as it is with simple oligosaccharides. Adducts of amylose with sodium hydroxide and with ammonium hydroxide were also prepared. The latter could be formed only by exchange with the alkali metal hydroxide-amylose adduct in ethanolic ammonia solution. Heddle and Percivals reported that they had obtained a 1:1 potassium hydroxideamylose adduct by treating amylose acetate with ethanolic potassium hydroxide. In view of the fact that monosaccharides react with potassium hydroxide in nonaqueous media to give mainly alcoholate, the alleged composition of this amylose complex is questionable. Champetier and Yovanovitchlw treated corn starch with aqueous sodium hydroxide solutions of various concentrations, and obtained three adducts having 1:2, 1 :1, and 2: 1 ratios of sodium hydroxide to D-glucose residue. No 2:3 or 3:4 adducts, such as those reported for cellulose, could be obtained. As the hydroxide concentration is changed, the transition from one addition compound to another is abrupt, thus showing an absence of topochemical fixation. Champetier and NBelIo4also studied cellulose in aqueous sodium hydroxide solutions, and found the following ratios of sodium hydroxide to D-glucoseresidue: 1:2,2:3,3:4, and 1 : l . In theabsenceof metal hydroxide, natural cellulose and mercerized cellulose are composed of units of 2 C~H10O5*H2O and C&Ilo05*H20, respectively. Champetier and NBel suggested that the formation of a sodium hydroxide-cellulose adduct occurs through tl stepwise displacement of water from cellulose hydrate by sodium hydroxide molecules. On the other hand, Chuin, Petitpas, and Marsaudon12a have explained the phenomenon as a displacement of the water molecules of sodium hydroxide hydrates (NaOH .m H2O) by hydroxyl , , ( HOcellulose)n. At concengroups of cellulose, to give NaOH (H20)m trations of less than 15% of sodium hydroxide, accessibility of the hydroxyl groups of native cellulose to sodium hydroxide is less than that of mercerized cellulose; but, a t concentrations greater than about 15% (the threshold
(123) J. ChBdin, G. Petitpas, and A. Marmudon, Mem. Serv. Chim. Eta1 (Paris), 40, 811 (1955).

250

J. A. RENDLEMAN, JR.

concentration for complete mercerization) , the accessibility is the same for both types. Complete merceriration occurs when 85% of the D-glucose residues have each fixed one molecule of sodium hydroxide, equilibrium .’~~ believed that, being reached in less than 48 h o ~ r ~Abadie-Maumert126 when cellulose is immersed in 18% sodium hydroxide, only half of the sodium hydroxide in the complex is combined with the cellulose. He postulated that the other half is present in the intermicellar solution trapped within the adduct. The formula given for the complex was
2 C ~ I O O(NaOH)aombimed* S. (NaOH).oh*ll(H:O),,t.

Ashar’02treated cellulose with aqueous potassium hydroxide, and ob.served the formation of adducts having the same ratios of hydroxide to D-glucose residue as those observed by Champetier and N & P for sodium hydroxide and cellulose, namely: 1:2, 2:3, 3:4, and 1:l. Bleached cotton linters were shaken with aqueous potassium hydroxide solutions of different concentrations at 20°, and the composition of the resultant solids was plotted against the hydroxide concentration. Plateaus corresponding to compounds possessing the aforementioned ratios were obtained. Heddle and PercivaP had also reported the preparation of a 1 :1 potassium hydroxide adduct. The reaction of cellulose with aqueous lithium hydroxide is a continuous function of the hydroxide concentration, and gives no adducts of greatly favored combining ratio. Heuser and Bartunek’Ol isolated an adduct having a 1:2 ratio of lithium hydroxide to D-glucose residue. At the high concentration of 5 M lithium hydroxide,‘* the ratio is 0.75:l. Less attention has been paid to the reaction of cellulose with rubidium hydroxide and with cesium hydroxide. Heuser and Bartunek’O’ isolated adducts of rubidium hydroxide and of cesium hydroxide that had the general formula MOH-3 CsHloOa. Their studies showed that the concentration, in weight percent, of alkali metal hydroxide required for forming a “stable” adduct of the lowest alkali content increases with incresse in the atomic weight of the metal: Li < Na < K < Rb < Cs. However, on a molar basis, this relationship does not hold. No simple relationship exists between the size of cation and the concentration of hydroxide necessary for the formation of a “stable” adduct. Treatment of an amylose-alkali metal hydroxide adduct with a concentrated solution of an inorganic salt (such EMpotassium iodide or potassium acetate) in aqueous ethanol can result in the total displacement of hydroxide This displacement indicates the existence of an equilibrium
(124) A. Webei, K.G. h h r , and G . Champetier, Compt. Rend., 288, 1318 (1054). (126) F.A. Abadie-Maumert, Compt. Rend., PI, 1957 (1961). (120) K.G.Aahar, Compl. W .262, , 734 (1961).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

251

bctwecp ail alkali mctal hydroxide tdduct and a salt adduct. Displacement o f the rnettll cation by another cation can alRo bccur." In aquem alcoholic solution, the combining ratio for sodium hydroxide cellulose is a function of the kind of alcohol, the concentration of the sodium hydroxide, and the concentration of In aqueous ethanol, f sodium hydroxide occurs at 18% of water, to maximum consumption o give R compound containing 1.15 molecules of sodium hydroxide per glucose residue. A smaller percentage of water results in a lower consumption of sodium hydroxide. Higher alcohols are similar to ethanol in behavior; however, the limited miscibility of most of the higher alcohols with water prevents the occurrence of a maximum. In contrast to the relatively high consumption of sodium hydroxide in aqueous ethanol and higher alcohols, consumption in aqueous methanol is very low. Nicoll , Cox,and Conawayln have critically reviewed the different methods generally employed for determining the alkali content in cellulosealkali metal hydrqxide,adducts.

b. Stoichiometry of Alkaline-earth Metal Hydroxide Reactions.The only information available on combining ratios for alkaline-earth metal hydroxide adducts is derived from studies of mono- and oligosaccharides. Various attemptssBto isolate adducts of methyl glycosides have failed. Polysaccharide adducts have thus far been almost entirely neglected. PrzyleckiU8 has prepared a chloride-free calcium compound of starch by treating an aqueous, alkaline solution of starch with calcium chloride. The composition of the product was reportedly constant. It is known with reasonable certainty that monosaccharides combine with alkaline-earth metal hydroxides in 2 :1, 1 :1, and , perhaps, 1:2 ratios of monosaccharideto metal. With oligosaccharides, only adducts containing one or more molecules of metal hydroxide per carbohydrate ligand are known. The ease with which more than one molecule of metal hydroxide becomes fixed to a molecule of oligosaccharide incrertses with increase in atomic weight of the alkaline-earth metal: Ca < Sr < Ba. The maximum f molecules of calcium hydroxide accommodated by a molecule number o of sucrose is probably only threepg although certain investigators have reported as many as four1%and six."* The tetrasaccharide stachyose reportedly accommodatessix molecules of strontium hydroxide per m0lecule.~~7
(127) W.

D.Nicoll, N. L. Cox, and R. F. Conaway, in "Cellulose and Cellulose Derivatives," M. W. Gra&, E. Ott, and H. M. Spurlin, eds., Interecience Publishers, Inc., New York, N. Y., 1964, p. 825 ff. (128) 8 . J. Przylecki, Sprawozdania Posisdzen Towarz. Nu&. Warmw. W&id ZV, Naxk Biol., SD, 186 (1936); Chem. Zen.&., 109, I, 1992 (1938). (129) W. Wolters, N. 2 .Rhuclee7-Znd., 10, 287 (1883).

252

J. A. RENDLEMAN, JR.

c. Electrophoresis.-Although there have been no electrophoretic studies on aqueous solutions of alkaline-earth metal hydroxides, the behavior of a large number of polyhydroxy comy w d s in 0.1 M sodium hydroxide wlution has been studied. Frahn a d Mills1ao found definite movemznt toward the anode (positive electrode) of nearly all compounds having more than two hydroxyl groups. This movement can be attributed either to the ionization of hydroxyl groups (alcoholste formation) or to the formation of adducts with free hydroxide ions, or both. Both species are negatively charged and would move toward the anode. In no case was there migration toward the cathode (negative electrode). With certain exceptions, the general order of decreasing rate of migration is: aldoses, ketoses 2 reducing oligosaccharides > nonreducing oligosaccharides > polyhydric alcohols and methyl glycosides. The mobilities of tetra-0methyh-glucopyranose and 2,3,6- and 2 ,4 ,6-tri-O-methyl-~-glucopyrmose are only slightly less than that of unsubstituted D-glucose. Among the monosaccharides, there is a trend toward the highest rates for the conformationally least stable sugars. However, the relatively high rates for D-glucose and D-mannose, both of which occur in the stable C1 (D) conformation, show that conformational instability is not the decisive factor operative. The acidic, angmeric hydroxyl group of reducing sugars probably contributes much to their overall high mobility. The rates for methyl glycosides are very low compared to those for the corresponding sugars. Diols do not migrate, indicating little or no reaction with hydroxide. The higher polyhydric alcohols, except for glycerol, do exhibit migration. In fact, certain heptitols behave as if, in acidity, they are comparable to reducing sugars. The fact that tetra-0-methyl-D-glucopyranoseis very similar to D-glucose in mobility may mean, in some instances, that simple ionization of a hydroxyl group, without concomitant formation of a cyclic complex, is responsible for the migration. Furthermore, cis-inositol, which is well known for its outstanding ability to form complexes with borate ion1" and with the ions of alkali metals and alkaline-earth metals,'Jahas B relatively low mobility, in aqueous sodium hydroxide] that is almost identical to that of D-glucitol. However, the possibility that the cyclitol preferentially forms uncharged complexes, composed of un-ionized polyhydric alcohol and undissociated sodium hydroxide, must not be overlooked. Such electrically neutral entities, should they exist, would be electrophoretically undetectable. Also, the fact that free, alkali metal ions can interact with polyhydroxy compounds in aqueous solution to give positively charged species6*cannot be disregarded in interpreting electrophoretic data on
(130) J. L.Frabn and J. A. Mills, Awlrdhn J . C h . , 12, 66 (1959).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

253

hydroxide solutions. Migration of these positively charged species toward the cathode would decrease the overall migration rate of the carbohydrate toward the anode. d. Optical Rotation.-Reeves arid Blouinlg’ measured the rotation of various glycosides in 1 M sodium hydroxide, hoping to gain information that would lead to an understanding of the effect of h l i metal hydroxide upon the optical rotation of amylose.1a2 The optical rotation of many glycosides in neutral aqueous solution is essentially the same as that in 1 M sodium hydroxide. Some glycosides, however, show relatively large, reversible changes in optical rotation at this concentration of hydroxide. To explain the changes, Reeves and Blouinls1suggested that alkali-sensitive glycosides undergo a change in conformation. The removal of a proton from a glycosidic hydroxyl group was postulated to give a charged species h i s more highly solvated than the corresponding, uncharged molecule. T (highly solvated) species would conceivably adopt a conformation in which as many aa possible of the ring hydroxyl groups would assume equatorial positions, where there is less steric hindrance than is afforded by axial positions. Thomsen,laa Smolehski and Kozlowski,’% and Reeves and B l ~ u i n ~ ob*~ served that sodium hydroxide has a relatively large effect on the optical rotation of sucrose. a,a-TrehaloseIu1 on the other hand, is affected only slightly. T h o m m P and Reeves and Blouinl*lmade no attempt to interpret the unusual behavior of sucrose; Smolexkki and Kozlowski,la however, assumed that the reaction was that of alcoholate formation, and they calculated dissociation constants for sucrose. The possibility that either chelation of sodium hydroxide (or OHe ion) with glycosides or inner chelation of undissociated sodium alcoholates is responsible for part or all of the abovediscussed changes in rotation had not been considered. Support for a hypothesis based on chelation in which at least two neighboring hydroxyl groups are involved is provided by the work of Lindberg and Swan.la6These investigators studied the changes in optical rotatiop of various glycosides effected by “cadoxene,” a solution of cadmium hydroxide in 28% aqueous ethylenediamine; they concluded that complexes we formed between cadmium and adjacent hydroxyl groups in a molecule of the glycosides. Methyl /3+glucopyranoside shows no electrophoretic mobility in “cadoxene” and lessens the conductivity of “cadoxene,” which proves that the complexes are uncharged. Except for the magnitude
(131) 132) 133) (134) (135)

t

R. E. Reeves and F. A. Blouin, J . Am. Chem. Soc., 79, 2261 (1957). R. E. Reeves, J . Am. Chem. Soc., 76, 4595 (1954). T. Thom’n, Ber., 14, 1647 (1881). K. Smolehski and W. Koslowski, Roczniki Chem., 16, 270 (1936). B. Lindberg and B. Swan, Acta Chem. Scad., 17, 913 (1963).

254

J. A. RINDLEMAN, JR.

of the rotational changes, the results of the polarimetric studies in “cadoxene” are identical with those obtained by Reeves1” in his studiea of glycosides in cuprammonium hydroxide solution. The complexes formed with “cadoxene” are, probably, very similar to those formed with cuprammonium hydroxide: that is, 5-membered, cyclic complexes involving adjacent hydroxyl groups. It has not been shown whether the “cadoxene” complex is an alcoholate or a hydroxide adduct. e. Preparation of Adducts. (1) Alkali Metal Hydroxide Adducts. (a) Mono- and Oligo-saccharides.-The usual method of preparation

is to dissolve the sugar in a small amount of water, dilute with ethanol, and then, with rapid stirring, add either ethanolic or aqueous ethanolic alkali metal hydroxide. Addition of a dilute solution of the hydroxide gives low ratios of hydroxide to carbohydrate; addition of concentrated hydroxide favors high ratios. Fornation of adducts of saccharides having a low molecular weight may require the addition of ether to the reaction mixture to effect their precipitation. Adducts having a very high alkali content may be prepared by dissolving the sugar in a concentrated solution of the alkali metal hydroxide (aqueous), and then adding ethanol to bring about precipitation. In all cases, the water content of the final mixture should be sufficient (10% or more) to prevent the formation of alcoholate. The precipitated adduct ia quickly separated from its mother liquor by the most convenient method (filtration, centrifugation, or decantation) , washed w i t h ethanol (or a mixture of ethanol and ether) ,and then, perhaps, with ether, and dried under vacuum. C a r e must be taken to protect the adduct from air (which contains carbon dioxide, oxygen, and water vapor, a l l of which can attack the adduct). Adducts of nonreducing sugars may be stored under nitrogen for many days. However, adducts of reducing sugars are not very stable, and should be used soon after preparation. The adducts are colorless and amorphous. Mtindy and Vavrineczsl reported that they obtained crystalline needles of sucrose*NaOH by slowly evaporating an aqueous solution containing equimolar proportions of the components. However, there has been no repetition of this experiment to confirm the report. Adducts prepared from reducing sugars are exceptionally sensitive to heat, and, even at room temperature, gradually turn yellow, possibly from deep-seated reactions involving enoliration.

(b) Poly8accharides.-Water-insoluble carbohydrates, such aa cellulose, must be equilibrated with aqueous, or aqueous alcoholic, hydroxide solutions. The effects of the concentration of hydroxide and of alcohol on the combining ratio have already been disauased. When equilibrium has
(136)

R.E.Reevee, Aduan. Carbohyd*ds Chcna., 6, 107 (1961).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

255

been attained, the product is either pressed free of adhering solution or wiped dry with absorbent tissue. Adsorbed alcohol, which is difficult to remove by vacuum drying, can be removed by humidification over water in a vacuum desiccator. Use of a high concentration of sodium hydroxide’” in water (about 5 M ) permits the formation of a 1:l adduct of cellulose. Lower concentrations lead to adducts having a lower content of alkali. Acetates can be used instead of the free polysaccharide for preparing polysaccharide adducts. For example,” at concentrationa of 0.1 M alkali metal (lithium, sodium, and potassium) hydroxide in 25y0aqueous ethanol, amylose triacetate is deacetylated to form adducts containing approximately one molecule of alkali metal hydroxide per three D-glucose residues. Partially crystalline amylose adducts can be obtained by immersing stretched filaments of amylose triacetate in the deacetylating solution.
(2) Alkaline-earth Metal Hydroxide Adducts.-Precipitation of an adduct from an aqueous solution containing both a sugar and a hydroxide is usually effected at room temperature by treating the solution with ethacol. The greater the concentration of the hydroxide with respect to that of tho sugar, the greater is the ease of fixing more than one metal atom upon a molecule of the sugar. High temperatures favor a high ratio of metal to sugar. Prolonged treatment with aqueous ethanol of a complex having a low metal content is known to favor conversion of the complex into one having a greater content of the metal. The isolated complex is washed with aqueous ethanol and dried, either between filter papers or in a vacuum. Amorphous and crystalline complexes have both been prepared by this general method. A review by Mackeneie and Q ~ i non ,~ complexes ~ of alkaline-earth metal hydroxides, is recommended to those who desire a more thorough description of the methods of preparation.
2. Reactions in Anhydrous, Alcoholic Media

In anhydrous, alcoholic media at 25”, both the hydroxide and the cyanide of alkali metals react with “nonacidic” carbohydrates, to give colorless, amorphous, hygroscopic precipitates that are preponderantly mono- (alkali metal) alcoh01ate.g~ Under the proper conditions of concentration, most of the metal alcoholates combine with an additional molecule of carbohydrate per molecule, to give products whose molar ratios of carbohydrate to alkali metal are greater than 1:l. A small proportion of the hydroxide adduct or the cyanide adduct accompanies the alcoholate as the latter is
(137)

E. G . V. Percival, A. C. Cuthbertson, and H. Hibbert, J . Am. Chem. Soe., 6a, 3267 (1930).

256

J. A. RENDLEMAN, JR.

precipitated from aolution. The hydroxide adduct may, actually, be a hydrate of the alcoholatebut no meaw for distinguishing between a hydroxide addyct and an alcoholate hydrate has yet been devised. (For the sake of simplicity, the term adduct is employed throughout this review.) Table VIII gives the composition of a number of complexes formed by the interaction of &ali metal hydroxide with various carbohydrates in
TAB- VIII
Composition of Variour Alcoholate Complexes Precipitated from Anhydrour Ethanolia Mdiaw Ligand Metal Molar ratio, Cornpodtion of complep Ethanol of so1vation.b cation ligand/cation % Alcohol- % Hydrox- molecules ate ide adduct per cation

n-Arabinose D-Glucose D-MUUIW Maltola Methyl a+ glucopyranoside 3 , anomer D-Glucitol Sucrose~

Na Na

1.18 2 .o 1.9 1.18 2 .o 1.8 2.3 1.6 1 .o 2.1 1.4

87

13
10 15

0.2 0.3 0.3 0.06 0 -07 0.4

K

90 8s

Na

88
97

12 3 12 30 18 12

X
Na Na Na

88
70

undet.
1.6 1.4
0.2 0.2

Li

88
80
77

82

Na

20 23

0 The aemmption ie made that the combining ratio for the dcoholate is identical to that for the hydroxide adduct. Percenteges are in mole percent. Solvent remaining after vacuum treatment at 26" fox 10 hr. 0 3-Hydroxu-zmethyl~-pyran~ne. d Uae of methanol inatead of ethanol gave a 1.8: 1 oomplex that consisted of 70% of alcoholate and 30% of adduct.

anhydropa ethanolic medium. The alcoholate content was determined by the modified Karl Fisoher analysis discussed in Section IIIJ (see p. 242). It is interesting that the percentage of alcoholate in the &glucose complex is approximately the same as that in the D-glucitol complex, even though ~gluco~m because , of its free reducing group, is more acidic than D-glucitol. Ordinarily, it would have been expected that this difference in acidity would have a much greater effect on the relative proportions of alcoholate

A L W L I AND ALKALINE-EARTH METAL COMPLEXES

257

in the two complexes. One explanation for the Similarity in alcoholate content is that alcoholates are, possibly, less soluble than adducts and are, therefore, the first to be precipitated. Another explanation is that inner chelation is such an exceptionally strong driving force in the formation of alcoholates that any difference in carbohydrate acidity is of minor importance. Methanol is rarely used as a solvent in the preparation of alcoholates. A monopotassio derivative of 1,2-anhydro-a-~-glucopyranose,lasand a sodio derivative of sucrose (see Table VIII), have been isolated from methanolic media. The facile'removal of hydroxylic protons from 2olyhydroxy compounds was first demonstrated by Sugihara and WolfrornJy8@ who treated cellulose with sodium hydroxide in boiling 1-butanol, and removed water of formation by concurrent azeotropic distillation. Later, Wolfrom and El-Taraboulsi'" similarly treated methyl a-D-ghcopyranoside, and obtained the monosodium alcoholate (reaction 4). Subsequent investigations by
GHMOS NaOH

+

* GHISOiNa + HzO

(4)

RendlemanDB have shown that the monosodium alcoholate of methyl CY-Dglucopyranoside can also be obtained by treatment of the ~-glucoside with sodium 1-butoxide in boiling 1-butanol (reaction 6). Heat is probably not
GHuOS

essential for effecting the formation of the alcoholate from sodium hydroxide in 1-butanol, in view of the ease with which alcoholates are formed in ethanolic sodium hydroxide solution at room temperature. However, to ensure complete removal of hydroxide ion and any possible water of hydration from the isolated complex, azeotropic distillation is a necessary operation. In a refluxing solution of sodium hydroxide and methyl a-D-glucOpyranoside in 1-butanol, possibly both the hydroxide ion and the butoxide ion react directly with the D-glucoside to produce an alcoholate. Sodium 1-butoxide is formed readily, in quantitative yield, by refluxing a solution of sodium hydroxide in an excess of 1-butanol (reaction 6) under dry
NaOH

+ NaOBu * C&O~Na + BuOH

(6)

nitrogen while water of formation is being removed by azeotropic distillation?' The product isolated from the reaction of methyl a-D-glucopyranoside with either sodium hydroxide or sodium 1-butoxide in refluxing 1-butanol
(138) A. Pictet arid P. Castan, Helu. Chim. Ada, 4, 319 (1921). (139) J. M. Sugihttra and M. L. Wolfrom, J . Am. Chem. Soc., 71, 3509 (1949). (140) M. L. Wolfrom and M. A. El-Taraboulsi, J . Am. Chem. Soc., 76, 5350 (1953).

+ BuOH

NaOBu

+ HtO

(6)

258

J. A. RENDLEIIMN, JB.

contains a small proportion of butoxyl p u p , a which suggests that reaction (7), the formation of a butoxide adduct, contributes significantly to the
C1HlrO~ NaOBu

overall reaction. Reducing sugars are decomposed by sodium hydroxide in boiling butanol. Treatment of reducing sugars with alcoholic alkali metal hydroxide must, therefore, bo conducted at room temperature, in order to avoid this decomposition of the ctirbohydrate. There have been few investigations of the reactions of polyhydroxy compounds with alkali metal &oxides in nonaqumus media at room temperature. Percivalw reported the preperstion of ogluoose*NaOEt and D-glucose*NaOMeby treating Dglucoee with sodium ethoxide, and penta0-acetyl-D-glucopyranosewith sodium methoxide, respectively The adducts were dried under vacuum at 00” for 24 hours, a treatment which would make the presence of any significant proportion of alcohol of solvation unlikely. Any traces of moisture in the preparative medium preclude the formation of an alkoxide adduct. Early reports that D-glucose,* ~-fructose,”~ and lactosdu react with sodium ethoxide in anhydrous ethanol at room temperature to give alcoholates axe questionable, because the products were not analy5ed for their ethoxyl content. hmpl6n and Kunz”* treated *glucose with sodium in absolute ethanol to obtain a compound that they believed to be a 1:l D-glucose-sodium ethoxide adduct; and Pringaheim and DernikosIu treated the monoaoetate of a datrin with sodium ethoxide, and obtained a compound having one sodium atom per wglucose d u e . However, the analyses performed were insufficientto permit identificationof the products. Alekhinelu treated turanose with sodium ethoxide in ethanol, and obtained a product containing one sodium atom per turanose moiety; however, he did not specify that he used anhydrous conditions, and he did not record an ethoxyl analysis. Pacsu and coworkersla have reported the preparation of the sodium alcoholatc of cellulose by treating either cotton or viscose rayon with either sodium methoxide or sodium l-butoxide in anhydrous alcoholic media, at temperatures ranging from 25 to 120’. No proof of formation of alcoholate waa offered. However, there is little reason to doubt its occurrence at the higher reaction temperatures. The ratio of sodium to *glucose (141) M.H b i g and M. Roeenfeld, Bet., 14, 45 (1879).

+

ClrH1dOvNeOBu

(7)

.

(143) 0. Zemplcln and A. KUM, Bet., MI, 1705 (leas). (143) Pllingsheim I . and D. Dernikoa, Ber., 66, 1433 ( l e a ) . (144) A. Alekhine, Ann. Chirn. (Paria), l8, 632 (1889). (146) R. F.Bohwenker, T. Kinwhite, K. B e u r h , and E. Pamu, J . Polyma Sci., 51, lSti:(lesl).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

259

residue varied with the time of reaction, the extractant (alcohol or acetone) used in extracting residual sodium alkoxide from the reaction product, the length of time for the extraction, and, to a small extent, the temperature. Sodium 3-methyl-l-butoxide, dispersed in xylene, has dao been used in preparing sodio derivatives of cellulose.1a In summary, the reaction between an alkali metal &oxide and a polyhydroxy compound in hot alcoholic media produces an alcoholate and, possibly, a small proportion of alkoxide adduct; however, the conditions governing the ratio of alcoholate to adduct have not yet been well defined. Reactions with alkali metal hydroxides and cyanides produce mixtures (of alcoholate and adduct) that consist mainly of alcoholate. Occurrence of reactions between alkaline-earth metal hydroxides and polyhydroxy compounds in anhydrous alcoholic media has not been reported.
a. Stoichiometry of Alcoholate Formation.-The ratio of carbohydrate to metal in an alcoholate complex is a function of several factors: the cation radius, the concentration of metal hydroxide in the preparative medium, and the configuration of the carbohydrate.w With monosaccharides, oligosaccharides, and many of their derivatives, at hydroxide concentrations of about 0.05 M or less, and with the carbohydrate present in either a stoichiometric proportion or an excess, the ratios approximate to the maximum values (see Table IX); these are, usually, 2:l for alkali metals other than lithium, although there are exceptions where the maximum ratios are 1:1 and 3: 1. Lithium alcoholate complexes do not show ratios greater than 1.5: 1; many me 1:1. High concentrations of hydroxide l e d to smaller ratios. At about 2 M concentration, the ratios generally approach a minimum of 1:1 for monosaccharides and their derivatives. For oligosaccharides, smaller ratios are possible. However, a large fraction of the metal in these low-ratio complexes is probably present as alkali metal hydroxide. Studies of sucrose in the presence of a high concentration of potassium hydroxideg6 indicate that, although the resulting complex contains more than one atom of potassium per carbohydrate moiety, no more than 1 hydroxylic proton is removed from the sugar molecule. Ethanolic potassium hydroxide (20%) reacts with sucrose and with sucrose octaacetate to give complexes having a carbohydrate-to-metal ratio of 1:2. For each complex, a modified Karl Fischer analysis indicated that slightly less than 1 proton had been given up by each carbohydrate moiety. The complex obtained from pure sucrose had the approximate composition CIZH~O-(OH *KOH)1.8(OK)0.7. Conceivably, large molecules having reaction sites (hydroxyl groups) separated
(146) V. A. Derevithya, M. Prokof’eva, and 2 .A. Rogovin, Zh. Obshch. Khim., 28, 1368 (1958).

Tmm I X
Maximum combining Ratios o f Alcoholate cOmploteso Isolated from Anhydrous Ethanolic MediaMaximum combining ratio, ligand/cation Alcoholate

z

Li'

Na "
1 . 2( 0 . 2EtOH)

K"

G 3 '

-

1 . 0 (1.4EtOH) 1.5 1.1 1 . 5 (no EtOH)

-

1 . 5 2 . 3 1 . 5( 1 . 6 EtOH) 2 . 2 1 . 9 (0.4EtOH) 2.0 ( 0 . 3 EtOH) 2 . 0 1 . 2( 0 . 1EtOH) 2 . 2 2 . 1 ( 0 . 2EtOH) 1 . 9

1 . 4 1 . 4 1 . 4

-

1 . 9 (0.5 EtOH) 2 . 3 b 1 . 9 2 . 3 1 . 9 e 3.8 (0.3 EtOH) 2 . 2 2 . 0 (0.3EtOH) 2 . 3 2.0 2 . 1( 0 . 4EtOH) 2 . 0 2 . 0 2 . 1 1 . 9 (0.3EtOH)

2 . 3 2.2 2 . 6

t . .

Z

E

*

2 . 0( 0 . 1 EtOH)

-

Z

E

2 . 0 1 . 9 2 . 1

P

-

a A minor, but siepificsnt, fraction of each isolated complex i an alkali metal hydroxide adduct. All complexes were t r e a t e d 86 pure alcoholate in the calculations of the combining ratio. Except where indicated by superscript c, the ratios are corrected, for ethanol content, only in those casee where the solvent content is known (ethanol of solvation, if known, is reported as molecules per cation in parentheam next to the combining ratio). b Based on the dimeric form of the aldose. The complex is actually composed of approximately s t i m a t e d on the basis of a b b l e molecule of solvation (ethanol) Der cation. four aldose residues. c E

ALKALX AND ALKALINE-EARTH METAL COMPLEXES

261

by relatively great distances should function as polybasic acids that release 3 or more protons. Gaver’” claimed to have prepared sodium “starchate” by treating starch with alcoholic sodium hydroxide solution; a polysodium alcoholate was presumed to have been formed by the release of a proton from each D-glucose residue. No conclusive proof that each sodium atom is associated with a carbohydrate oxyanion was, however, given. Although the stoichiometry varies with the concentration of hydroxide, favored combining ratios do exist. Certain ratios predominate, over a wide range of concentration. The ability of a metal alcoholate to accommodate an additional molecule of carbohydrate increases with increasing ionic radius:gB Li < Na < K < Cs. The difference in stoichiometry between lithium and sodium is much greater than that between either sodium and potassium, or potassium and cesium. The coordination number of an alkali metal is known to increase with increasing ionic radius. Brewerla reported that the maximum number of donor groups oriented about an alkali metal cation is four for lithium, and as many as six for sodium, potassium, rubidium, or cesium. A greater surface area would allow accommodation of more than one carbohydrate moiety; but, in addition, solvent molecules are more strongly attachpd to cations of smaller radius, and these may not be readily displaced by carbohydrate molecules. The importance of carbohydrate configuration in the formation of complexes is exemplifiede6 by the behavior of certain isomeric sugars. For example, each of the four D-aldopentoses (D-arabinose, D-xylose, D - I ~ X O S ~ , and D-ribose) reacts with potassium hydroxide or cesium hydroxide to form a complex having a carbohydrate-to-metal ratio of 2:l. However, : l with sodium hydroxide, only D-xylose, D-lyxose, and D-ribose form 2 complexes; n-arabinose gives a 1:1 complex. Similarly, D-mannose reacts with sodium hydroxide to form a complex having a maximum ratio of 1: 1, whereas both D-glucose and D-galactose form 2 :1 complexes. With potassium hydroxide, all three of these hexoses give 2: l complexes, A 3: l complex is formed by the reaction of levoglucosan (lt6-anhydmj3-Dglucopyranose) with potassium hydroxide. Weber, Ashar, and Champetier*24 treated cellulose (dried cotton linters) with 1 M sodium hydroxide in various anhydrous alcohols at room temperature. Although no analyses were performed to determine whether the products were alcoholates or adducts, our present knowledge concerning simple carbohydrates in anhydrous media would strongly suggest that the products contained a large, if not preponderant, proportion of alcoholate. When equilibrium wm established in alcoholic systems that were 1 M in (147) K.M. Gaver, U.S. Pat. 2,397,732 (1946); Chem. Abstracts, 40, 3620 (1946). (148) F.M. Brewer, J . Chem. Soe., 361 (1931).

262

J. A. RENDLEMAN, JR.

sodium hydroxide, the following frtlctions of a mole of sodium hydroxide were found to have reacted per mole of D-glucose residue: in methanol, 0.124; in ethanol, 0.26; in l-propanol, 0.58; and in l-butanol, 0.78. In water, only 0.07 mole of sodium hydroxide reacted. Equilibrium was established very rapidly in methanol ( <24 hours). In the higher alcohols, a period of more than 400 days was required. reacts with butanolic sodium hydroxide The e m with which at the reflux temperature (with concurrent azeotropic distillation to remove water of formation) to give alcoholates containing one metal ion per ~-glucose residue suggests that similar treatment of nonreducing oligosaccharides would, likewise, give alcoholates containing one metal ion per monosaccharide residue. No such reaction with oligosaccharides has, as yet, been reported. The formation of adducts of polyhydroxy compounds either with alkali metal alkoxides or with cyanides has not been sufliciently studied to permit any generalizations to be made regarding its stoichiometry. Table I X (see p. 260) lists many of the known alcoholate complexes prepared from alkali metal hydroxides in snhydroua ethanolic media. For additional complexes from ethanolic media, the reader is referred to Table VIII (p. 256) and to Rsf. 96.

b . Electrophoresis.-Many fmtors other than alcoholate formation might influence the rate and direction of electrophoretic migration: (1) the degree of dissociation of the metal alcoholate into free ions, (2) the size of the carbohydrate anion, (3) the degree to which the metal cations combine with the carbohydrate to give free, positively charged, carbohydrate species, and (4) the carbohydrate-to-metalcombining ratio. The competition between hydroxide ion and metal ion for carbohydrate has been demonstratedM in a combination of electrophoretic experiments involving Dglucom, D-xylose, sucrose, and D-glucitol. In methanolic lithium h y d r m ' h solution, all movements were toward the anode. In methanolic sodium hydToxide solution, movement of D-xylose and D-glucosewas toward the anode, whereas movement of sucrose and D-glucitol was in the oppoeite direction. A reasonable interpretation of these phenomena rests solely on the assumption that a competition exists between the metal cation and the hydroxide ion for the carbohydrate. Free alcoholate anions, as well as any free carbohydrate-hydroxide ion species, would migrate toward the anode. Cationic species, produced by the chelation of a free, metal cation,with one or more carbohydrate molecules, would migrate toward the cathode. Negative and positive species should both exist; however, the mobility of each species and the extent to which each is formed determine the rate and

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

263

direction of carbohydrate migration actually observed. I n alcoholic sodium hydroxide solution, the two opposing movements almost cancel each other, and migration is slow. The somewhat acidic sugars, D-xylose and D-glucose, s an anion, but the much less acidic carbohydrates, sucrose migrate slowly a and D-glucitol, which do not have an acidic, hemiacetal hydroxyl group, maintain a net positive charge and migrate toward the cathode. In methanolic lithium hydroxide solution, all carbohydrates migrate toward the anode, because. of the extremely weak ability of lithium cations to form positively charged carbohydrate species (see Section 11,6, p. 227). Electrophoresis does not show the presence of uncharged species, such as undissociated metal alcoholate or carbohydrate-metal hydroxide sdducts. These species are probably present in alcoholic solutions, but their concentration has not yet been ascertained. Their presence is suggested by the relatively low mobility of carbohydrates in alcoholic solutions of alkali metal hydroxide. In aqueous media, where greater dissociation of ion pairs should occur, the mobility is extremely high. The possible existence of free carbohydratehydroxide ion species cannot be disregarded, because of the hydrogen-bonding properties of the hydroxide ion.
c. Preparation of Alcoho1ates.-Only a brief description of the general procedure for preparing alcoholates will be given. The solubility of both the carbohydrate and the alcoholate often determines the proportion of solvent necessary. To a solution of carbohydrate in either ethanol or N-methyl-2-pyrrolidinone is rapidly added dropwise, with stirring, an f the product tends to remain ethanolic solution of alkali metal hydroxide. I in solution under these conditions, ether is added, either before or after addition of the hydroxide, in order to facilitate precipitation. The concentration of the hydroxide solution being added is usually a.very important factor in determining the combining ratio of the alcoholate (see Section III,2a, p. 259). After the product has been separated from its mother liquor, it is quickly washed with ethanol or with a mixture of ethanol and ether, and then dried under vacuum. Alcoholates of carbohydrates are colorless, amorphous, and, like the corresponding alkali metal hydroxide adducts, very hygroscopic. Alcoholates of nonreducing saccharides (soluble or insoluble) and of saccharide derivatives may also be prepared by the method of Wolfrom and coworkers.g2." A mixture of the carbohydrate and sodium hydroxide in 1-butanol is refluxed, preferably under a nitrogen atmosphere, Pnd all of the water of formation is removed by concurrent azeotropic distillation. This method is preferred when no trace of hydroxide adduct or water is desired in the final product. Butoxide ion is a possible contaminant, how-

264

J. A. RBNDLBMAN, JR.

ever. The use of a 1:1 molar ratio of sodium hydroxide to monosaccharide residue in the remtion medium leads mainly to the monosodium alcoholate ( b d on the monosaccharide residue). The method of GaveP7 for nonreducing carbohydrates (refluxing the carbohydrate in alooholic alkali metal hydroxide solution without concomitant amtropic distillation) night give results similar to thoae obtained by the Wolfrom method.m*" However, the presence of water and hydroxide ion in the final reaction mixture would probably cause a. small fraction of the product to be the hydroxide adduct.
3. Reaction# in Nonhydroxylic Solvents

Media other than water and alcohol have thue far been virtually ignored. However, ethylenediamiaehas received some attention, apparently because of its ability to facilitate the release of protons from compounds whose acidity is ordinarily considered to be very weak. Derevitskaya, Smirnova, and bgovin;4@ in a study of the degree of dissociation of the hydroxyl groups of Dglucose, methyl cu-D-glucopyranoside,methyl @-D-glucopyranoaide, cellobiose, and maltose, electrometrically titrated 0 . 0 0 1 M solutions of these compounds in e t h y l e n d i e with 0.1 M potassium hydroxide (alcoholic), The volume of the titrant was plotted against the derivative of the current with respect to the volume of the titrant. The resulting curves contained a number of peaks that were interpreted as being the equivalence points of the different hydroxyl groups. The first peak on the maltose, and cellobiose was the highest, and it differed curves for D-~~UCOBB, greatly in height from one carbohydrate to another. I f the assumption is made that the heights of these first peaks characterize the degree of dissociation of the respective anomeric hydroxyl group, these compounds c m be arranged in the following order of decreasing acidity of the anomeric hydroxyl group: D-glucose > cellobiose > maltose. A comparison of the curves of the two ~glucopyranosides indicates that the @-D Bnomer is More acidic than the WD anomer. Five peak8 w e r e obtained for pglucose, and four for each of the two D-glucopyranasides. There is some question, however, as to whether each of these peaks corresponds to the titwtion of a different hydroxyl group. Rendlema,n160has treated D-glucose with potassium hydroxide and with sodium hydroxide in ethylenediamine under the and has found that conditions employed by Derevitskaya and ooworkers,14@ (1) addition of either potassium hydroxide or sodium hydroxide to the Dglucose solution results in the precipitation of a product (an almost
(140)

Chnn. Sect. (Englkh Trend.),U6,1254 (1981). (160) J. A. Rendleman, Jr,, unpubliehed observations.

V. A. Derevitakaya, G. S. Bmirnova, and 2 ; . A. Rogovin, Proc. A d . 815.USSR,

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

265

invisible gel) before half of the stoichiometric proportion of the hydroxide (the amount reqgred for reaction with five hydroxyl groups per molecule) has been dded, and (2) the product that is isolated after a stoichiometric proportion of hydroxide has been added contains essentially all of the sugar originally present in solution, but no more than two atoms of alkali metal per D-glucose moiety. Complete identification of the derivative has not yet been attempted. Treatment of D-mannose with lithium hydroxide, and with potassium hydroxide, in liquid ammonia161gives a crystalline di-lithio derivative and a crystalline monopotassio derivative, respectively. The structures of these products have not been determined. The use of metal alkyls, metal aryls, metal hydrrdes, and metal carbides for preparing alcoholates of carbohydrates in inert, aprotic solvents has not yet been reported.
4. Structure of Alcoholates and Adducts

The general structure of alkali metal alcoholates of polyhydroxy compounds is probably very similar to those proposed by Martell and CalvixP for the alkali metal chelates of o-salicylaldehyde (see Figs. 9 and 10).laJmJM Unfortunately, because of the highly amorphous nature of nearly aJl pf the alcoholates and adducts formed by the interaction of metal hydroxides with carbohydrates, x-ray diffraction studies have failed to furnish information regarding the precise location of the metal in them complexes. The ease of formation of alcoholates and adducts can be related to both the acidity and the geometry of the polyhydroxy compound. The geometry is important, in that the chelate ring must possess a minimum of strain in order to allow the complex to possess the maximum stability. Reaction of a cyclic 1,2diol with a metal hydroxide is physically impossible if the hydroxyl groups are oriented in directions that are exactly opposite (180') to each other. Cyclic 1,&diols can form chelates if both of the hydroxyl groups are cis and axial. An example of this is the reaction of 1,&anhydro-

Fro. %-Lithium Alcoholate of Salicylaldehyde.
(151) (162) (153) (154)

K. Shimo and R. Tada, Sci. Rept. Re8. Inst. Toholcu Uniu., Ssr. A , 7, 235 (1955). Ref. 70, p. 242. N. V. Sidgwick and F. M. Brewer, J . Chem. Soc., 127, 2379 (1925). 0. L. Brad\ and W. H. Bodger, J . Chem. Soc., 952 (1932).

266

J. A. RENDLEMAN, JR.

Fro. lO.-Sodium Alcoholate of Salicylsldehyde.

PD-glucopyranose with potassium hydroxide in ethanol to give an alcoholate, whose structure is possibly that shown in Fig. 11. The location of the metal cation (or cations) in alcoholates has been investigated by the method of substitutive methylation for the sodium derivatives of methyl a-wglucopyranoside, amylose, and cellulose. The method assumes that methylation by either dimethyl sulfate or methyl iodide occurs only at an anionic oxygen stom. Lendu methylated the monosodium alcoholstes (prepmed in boiling, butanolic sodium hydroxide by the

FIU.Il.-PoeSible glucopyrenoee.
(165)

Structure of the Potsesiurn Alcoholate of l,BAnhydro-&n-

R.W.Lena, J . Am. C h .Soc., 89, 182 (1960).

ALKALI AND ALKALINE-EARTH METAL COMPLEXES

267

method of Sugihara and Wolfram*@) of methyl a-o-glucopynmoside and cellulose with methyl iodide, and obtained producta that he subsequently hydrolyzed and analyzed by quantitative, paper chromatography. The hydrolyzates contained mono-, di-, and tri-0-methyl-D-glucose and unsubstituted ~glucose. The mono-O-methyl fraction contained the principal isorsers, of which 2-O-methyl-~-glucose was the ether present in the highest percentage. From the distribution of the methoxyl groups in the monosubstituted wglucosea, the relative rate constants for substitution at the hydroxyl groups on C-2, (2-3, and C-6 were calculated to be 5, 1, and 2.5, respktively. It should be pointed out that the failure to observe monosubstitution at the hydroxyl group on C-4 of methyl eD-glucopyranoside from could have been due to a lack of separation of 3-O-methyl-~-glucose the 4-methyl ether by paper chromatography. Thus, the relative rate for substitution calculated for the hydroxyl group on C-3 is, possibly, actually equal to the combined relative rates for substitution at the hydroxyl groups on G 3 and (2-4. To explain the formation of the di- and tri-methyl ethers, Lena1" suggested two possible mechanisms: either (1) methylation occurs on' un-ionized hydroxyl groups, or (2) the distribution of ionized hydroxyl groups (anionic oxygen atoms) changes continually during the course of the reaction. Bines and WhelaxP methylated the monosodium "alcoholate" of amylose (prepared by heating amylose with butanolic d i u m hydroxide for 2 hours in a closed v 1 at 85-87") by means of a method16' which involves heating a mixture of the alcoholate and a methyl halide in toluene at 100" in a sealed reaction vessel for 4 hours. Hydrolysis of the product gave mono-0methyl-D-glucoses (36.4%), di-0-methyl-D-glucoses (14.6%), unsubstituted D-glucose (43.573, and 5.5% of remaining material that had been incompletely hydrolyzed. The 2-, 3-, and 6-O-methyl-wglucoses were present in the proportions of 2.6: 1:1.2. Doane and coworkers1@methylated starch by the method of Hodge, Karjala, and Hilbertl" (sodium metal-liquid ammonia-methyl iodide) to degrees of substitution (D.S.) of 0.34 and 0.90 methoxyl per D-glucose residue. Hydrolysis of the products gave D-glucose, and mono-, di-, and tri-0-methyl-*glucose in the approximate molar ratios of 36: 6 :2 :1 and . S .0.34 and 0.90 products, respectively. Resolution of 6:3:2:i1, for the D the monomethyl ethers into their isomers showed a 1:1ratio of 2-O-methylin the D.S. 0.90 hydrolyzate, and a 3:2 D-glucose to BO-methyl-~-glucose
(150) B. J. Bmea and W. J. Whelm, J . Chenl. Soc., 4222 (1982). (157) K. M. Gaver, E. P. Lasure, and D. V. Timen, U. 8. Pat. 2,871,779 (1964); f%sm. Abeirmt8, 48, 8589 (1954). (158) W.M.Doane, N.L.Smith, C. R. Rumll, and C. E. Rist, Staerke, 17,225 (1966). (159) J. E.Hodge, 8 .A. Karjala, and G.E.Hilbert, J . Am. Chem. Soc., 78, 3312 (1951).

268

J. 4. J t W D L E W ,

JR.

ratio in the D . S . 0.34 hydrolyzate. Significantly, there wae no 3-O-methylD-glucose in either hydrolyoate. The sodium alcoholate of cellulose prepared from a sodium alkoxide is probably similar to, if not identical with, that prepared from sodium hydroxide. Manomethylationof the alcoholate prepared with either sodium occura preferentially a t the hydroxyl methoxide or sodium l-b~toxidel4~ groups on C-2 and C-6. Very little methylation occurs a t the hydroxyl group on C-3. In adducts of alkali metal hydroxides, the hydroxide ion is probably chelated to two or more hydroxyl groups and attached most strongly to the hydroxyl group of highest acidity. Percival@-lwJ1O methylated a number of potassium hydroxide adducts in an attempt to determine the positions of the potassium hydroxide residues. His method waa based on the likely assumption that dimethyl sulfate reacts only with those hydroxyl groups to which OHe ion b attached. The results of his studies of product isolation employing simple carbohydrates (D-glucose,” sucrose,11ocellobiose,”JW maltose,”>” and lactosem) indicated that, for D-glucose, the potassium hydroxide is located preferentially, if not exclusively, a t the anomeric hydroxyl group; for sucrose, the potassium hydroxide is attached mainly to the primary hydroxyl groups; for cellobiose, the attachments are at the free anomeric hydroxyl group and at the hydroxyl groups on C-2 and (2-3; for lactose, the attachments are at the anomeria hydroxyl group of the ,D-glucopyranoseresidue and at the hydroxyl groups at C-2 and C-4 of the D-galactopyranose residue; and for maltose, the attachments are at the anomeric hydroxyl group and at the hydroxyl groups at C-2 and C-6 of one of the D-glucopyranose residues. the 1:1 potassium hydroxide adducts According to Heddle and Percival,gB of amylose and cellulose are methylated, by dimethyl sulfate, preferentially at the hydroxyl group on (2-2; no methylation of the hydroxyl group at C-6 was detected. Croon and coworkers,1m however, using more advanced analytical techniQues, later found that sodium hydroxide adducts of amylose and cellulose (which should not differ significantly from the potassium hydroxide adducts) can be methylated with dimethyl sulfate at all three of these hydroxyl groups. The relative rates for the methylation of cellulose at the hydroxyl groups on C-2, C-3, and C-6 are 3.5,1, and 2, respectively; for amylose, they are 6, 1, and 7, respectively. It was suggestedlm that intrachain hydrogen-bonds play an important role in the reactivity of the hydroxyl groups in both celluloBe and amylose.
(160)

I. Croon and B. Lindberg, Suenek hpperstid., 6B, 794 (1958); 60, 843 (1957); I. Croon, ibkd., 61, 919 (1968); I. Croon and E. Flsrnm, ibid., 61, 863 (1958); I. Cronu, Ada Chem. Scad., la, 1235 (1969).

ALKALI AND ALKALINE-EARTH METAL COMPLEXEB

269

Htwch can he vinylated with acetylene in the presence of potassium hydroxide in an W~UCOUH tctrahydrofuran medium.16' The mechanism poecJibly involven the addition of the potassio derivative of starch across the carbon-carbon triple bond of acetylene, with subsequent hydrolysis of the organometallic intermediate to give the vinyl ether. Such a mechanism has been postulated for the formation of vinyl ethers from monohydric alcohols and acetylene, in the presence of an alkali metal base as catalyst.162 The vinylation of amylose is very similar to the vinylation of amylopectin, except for the relative ratio of mono- to di-substitution.'6" With amylopectin, the proportion of disubstitution is greater. In both starches, the hydroxyl group on C-2 is slightly more reactive than the hydroxyl group on (2-6; there is little substitution at the hydroxyl group on C-3. The fact that a carbohydrate can be vinylated in the presence of a base catalyst in ,an aqueous system is a good argument for the existepce of carbohydrate alcoholates in aqueous media. However, successful vinylation does not imply that the concentration of the alcoholate is necessarily high. Because the interaction between the carbohydrate and the hydroxide ion is rapidly reversible, the concentration of the alcoholate may not have to be large for vinylation to proceed at a moderate rate. There have been no studies of adducts of alkalineearth metal hydroxides with a view to determining the position, or positions, of attachment of alkaline-earth metal hydroxide on carbohydrates.
OF

IV. ALCOHOLATES FROM REACTIONB, IN LIQUID AMMONIA, CARBOHYDRATES WITH ALKALI METALS, ALKALINE-EARTH METALS, AND ALKALI METAL AMIDES

Chablay16a* treated D-mannitol with sodium and with potassium in anhydrous liquid ammonia, and obtained the respective monoalkali metal alcoholates. Later, Schmid and coworkers1" reported the preparation of monosodium alcoholates of ethylene glycol, glycarol, wmannitol, methyl a-wglycopyranoeidq, starch, inulin, lichenin, glycogen, chitin, ~-glucose, and *fructose by treatment of the respective carbohydrate with sodium in liquid ammonia; similarly, starch, inulin, lichenin, D-mannitol, and
(161) J. W. Berry, H. Tucker, and A. J. Deutsahm, Jr., Z n d . Eng. Chum., Procese Dcsign Develop., I, 318 (1983). (162) W.Reppe, Ann., 601, 81 (1966). (163) J. W. Berry, A. J. Deutmhmen, Jr., and J. P. Evans, J . Otg. Chsm., I S , 2819 (1964). (1638) E.Chtlthy, Compt. Rend., 140, 1396 (1905). (164) L. Bnhrnid and U. Becker, Ber., 58, 1966 (1925); L. Sahmid, A. Wsschlcau, and E.Ludwig, Monubh., 4S, 107 (1928).

270

J. A. RENDLEMAN, JR.

methyl a-Pglucopyranoside reacted with potassium to give the respective monopotassium alcoholates. Muskatla' improved the procedure, and prepared potaslilium alcoholat- of various highly substituted monosaccharides. The report on the preparation of the alcoholatea of D-glucose and D-fructose is probably incorrect. Liquid ammonia itself reacts with either a free or an acylated anomeric hydroxyl group,laa affording the corresponding glycosylamine, without affecting acyloxy groups attached to other carbon atoms. However, alkali metals remove acyl group of O-acylated glycosylamines to give alkali metal derivatives. Lithium, sodium, and potctssium in liquid ammonia have been used107 t o prepare the respective monoalkali metal derivatives of sucrose. Prey and Grundschobe+ used similar means to prepare the monosodio derivative. Di-, tri-, tetra-, and pentasodio derivatives have also been prepared. ~6~ ~ 7 0 Ammonia of solvation, which is strongly held by the alcoholates, can be almost entirely removed by extraction with toluene.'@ Amagasa and Onikura171 treated sucrose with potassium in liquid ammonia under preasure at room temperature, and obtained a dipotassio derivative. Further addition of potassium to the reaction mixture gave a pyrophoric substance that appeared to be a mixture of hexa- and hepta-potaseio derivatives. D-Mannitol, likewise capable of undergoing polysubstitution, can form a di- and a tri-lithium alcoholate when it is treated with lithium in liquid ammonia; sodium and potassium give disubstitution products; and calcium gives a monocalcium alcoholate.lS1Treatment of cellulose with aodium17* can lead to the trisodio derivative. The first atom of sodium enters rapidly; the second and third, slowly. Sodium in either pyridine or morpholine reacts with sucrose to give ti substance that, reportedly, does not have the properties of an al~oholate.~~7 Further work with these solvent media is necessary before any definite conclusions regarding the reaction products can be made. Sodamide in liquid ammoniala7reacts with sucroae much faster than does
(165) I. E. Muekat, J . Am. Chcm. Soc., 66, 693, 2449 (1934); P. A. Levene and I. E. Mudcat, J . BioZ. Chcm., 106, 431 (1934). (166)R.8.Tipeon, MeuIods Carbohvdde C h . ,9, 160 (1903). (167) P. C. Arni, W. A. P. Black, E. T.Dewar, J. C. Petenron, and D. Rutherford, J . A w l . Clam. (London), 8, 180 (1959). (168) V. Prey and F. Crundschober, Mondsh., 81, 1186 (1960). (169) W. A. P. Black, E.T.Dewar, and D. Ruthexford, J . C h . Soc., 3073 (1959). (170) W. A. P. Black, E. T. &war, J. C. Patereon, and D. Rutherford, J . Appl. C h . (London), 8, 286 (1959). (171) M. Amagaea and N. Onikurs, Kogyo Kagaku Zasshi, 69, 2 (1949). (172) P. C. Scherer and R. E. Hussey, J . Am. Chem.Soe., 68, 2344 (1931); P.Shorigin and N. Makarowva-SemljanRkaja, Rer., 88, 1713 (1936).

ALKALI AXD ALKALINE-EARTH METAL COMPLEXES

271

sodium to give sodium alcoholates. The use of sodamide haa permitted the isolation of a heptasodio derivative. An octasodio compound that has o t a s s i u m amide reacts with pyrophoric properties has also been reported. P o-msnnito1161 in liquid ammonia to produce a dipotado derivative. Tipson'" has described, in detail, two general procedures,both involving the use of liquid ammonia and alkali metal, for the methylstion of none offered procedures for reducing carbohydrcttes and their derivatives. H t4e preparation of 1,2: 5 ,6di-O-isopropylidene-3-O-potassio-cu-D-glucofurmose, methyl tetra-0-potassio-a-o-mannopyranoside, and tri-0-sodiocellulose; and for the conversion of these alkali metal derivatives into 1,2 :5 ,&li-Oisopropylidene-3-0-methyl-~~-~-glu~ofuranose, methyl tetraO-methyl-a-~-mannopyranoside, and tri-0-methylamylose, respectively. Factors that dictate the choice of the alkali metal in the preparation of methyl ethers are: (1) the solubility of the metal, (2) the solubility of the alcoholate. (3) the reactivity of the alcoholate, and (4) the solubility of the metal halide formed during reaction of the alcoholate with methyl
halide.


相关文章:
更多相关标签: