1426
D. A. I. GORINGAND T. E. TIMELL
Vol. 64
MOLECULAR PROPERTIES OF SIX d!-O-METHYLGLUCUItOlOXYLANS' B Y D. A. I. GORING AND T. E. TIMELL Pulp m d Poper Research Institute of Canada and the Department of Chemistry, McGill University, Montreal, Canada Recezved March 18,I960
Six 4-0-methylglucuronoxylans have been isolated from various woody angiosperms in a maximum yield and with a minimurn of depolymerization. Osmotic pressure measurements gave number-average degrees of polymerization ranging from 185 to 234. Light-scattering measurements were carried out with dimethyl sulfoxide solutions of the polysaccharides. Reliable results were obtained only after colloidal material had been eliminated by ultracentrifugation a t 140,000 g, followed by ultracentrifugation a t 35,000 g in light-scattering cells especially designed for removal of both sedimenting and floating debris. The weight-average degrees of polymerization obtained varied between 440 and 500. A comparison between degrees of polymerization and intrinsic viscosities indicated that the acidic side chains of the 4-0-methylglucuronoxylans increased the effective volume of the molecules in cupriethylenediamine but that this volume was only half of that of cellulose in the seme solvent.
While the structural details of many of the xylans side chain and 3.6 0-acetyl groups per 10 xylose occurring in nature are now fairly well known12 residues, was also isolated from the wood of white little information is presently available concerning birch. The bark hemicellulose was obtained in almost their niolecular properties. It is evident from a recent review3 that in the past many of the poly- quantitative yield by extraction of the bark with saccharides ,studied had been severely degraded alkali after pretreatment with acid chlorite. The during their isolation, especially when acid chlorite 0-acetylated polysaccharide was isolated in a yield was sed.^,^ Low yield in the isolation often re- of 50% by extracting a chlorine holocellulose with sulted in products hardly representative of the dimethyl sulfoxide. The remaining hemicelluloses polysaccharide as it originally occurred in the living represented 70430% of the amount present in the plant. In most cases, organic esters of the hemi- wood and were obtained by direct extraction with celluloses have been investigated, many of them alkali of the latter. A portion of each hemicellulose was converted to constituting only a portion of the original material because of the poor solubility properties of these the fully substituted acetate derivative, a reaction derivatives. Lately, this difficulty has been partly which has been shown'l to involve no depolymerizaoverco~ne,~ but so far only number-average molec- tion of the polysaccharide. The acetates were soluble in chloroform containing some ethanol, ular weights have been reported. The present paper is concerned with the number- and osmotic pressure measurements were carred out and weight-average molecular weights of some hard- in this mixed solvent. The osmometers used12 wood xylans as determined from osmotic pressure contained large membranes, supported on both and light-seat'tering measurements, care being sides, thus making rapid and accurate measuretaken to avoid degradation in isolating the samples. ments possible. Linear relationships between Intrinsic viscosities were also measured from which reduced osmotic pressure and concentration were an indication of the molecular configuration was observed throughout, and results obtained with different membranes and osmometers were in excelobt'ained. Four of iihe hemicelluloses were 4-0-methyl- lent agreement (Fig. 1). The number-average glucuronoxylans containing a linear framework of molecular weights obtained are given in Table I. (1 + 4)-linked ,i3-D-xylopyranose residues, every TABLE I tenth of wh.ich, on the average, carried a 4-0OSMOMETRY DATA methyl-a-D-glucuronic acid residue attached as a single side chain through Cz. These polysacchaPolysaccharide (h/ZL)W-" Y n x 10-4 rides were obtained from the wood of white birch White birch 0 50 51 (Betula papyrifera),5 yellow birch (Betula lutea)6 White birch, 0-acetyl .60 43 and sugar maple (Acer sacch~rum),~ as well as White birch, bark 46 56 from the inner bark of white birch.* A fifth poly56 46 Yellow birch saccharide with the same constitution but with only 56 46 White elm 7 xylose residues per acid side chain was obtained Sugar maple .50 51 from the wood of white elm (Ulmus umeric~na).~ An O-:tcet~yl-4-O-methylglucuronoxylan containing The 4-0-methylglucuronoxylan from birch bark one (1 .-t 2)-,linked 4-O-methyl-a-~-glucuronicacid had evidently the highest molecular weight, the 0-acetylated polysaccharide from birch wood the (1) Paper preeented at the Symposium on Wood Hemicelluloses lowest. The high value observed for the bark xylan before the Divisian of Cellulose Chemistry a t the 136th Meeting of the American Chemical Society in Atlantic City, N. J., September, 1959. was rather surprising in view of the fact that this (2) G . 0. Aspinall, Advances in Carbohydrate Chem., 14, 429 (1959). was the only hemicellulose which had been sub(3) C. P. J . G1:utdemans and T. E. Timell, Suensk Pnpperstidn., 61, jected to treatment with acid chlorite, a reagent 1 (1958). known t o degrade polysaccharides consider(4) C. P. .J. Glzudemans and T. E. Timell, ibid., 60, 8G9 (1057). (5) C. P. J . Glaudemans and T. E. Timell, J . Am. Chem. SOL.. 80, a b l ~ . ~I4 , ' ~It was also the only hemicellulose 941, 1209 (1958). ( 6 ) T. E. Time11, ibid., 81, 4989 (1959). (7) T. E. Timell, Can. J . Chem., 37, 893 (1959). (8) A. Jabhar Mian and T. E. Timell, T a p p i , in press. (9) J. :K. Gillham and T. E. Timell, Can. J . Chem., 36, 410, 1467 (1058).
(10) T. E. Timell. J . Am. Chem. Soc., 82, 5211 (1960). (11) J. 0. Thompson, J. J. Becker and L. E. Wise, T n p p z , 36, 310 (1953). (12) J . V. Stabin and E. H. Immergut, J. Polymer Scz., 1 4 , 209 (1954).
isolated in tin almost quantitative yield, a fact which WHITE BIRCH - " u u - appeared t o preclude the possibility of a fractiona0.5 ti( n during; its isolation. ;W BIRCH, ACETYL ,. Greater difficulties than usual were encountered 0 5-in clarifying the chloroform-ethanol solutions for i W BIRCH, BARK light scattering. Clarification by ultracentrifuga- " tion was ineffective because of the high density of the solvent and filtration through glass, paper or u IYELLOW BIRCH " cellophane membranes of varying porosity resulted < 0.5-in fractioniition of the polymers and loss of up to 8 IMAPLE " 50% of the hemicellulose acetates. 0.5-Better results were obtained with solutions of the polysaccharides themselves in dimethyl sulfI oxide containing 2% of water. When dry, this 03 I 1 I I I I I 0 I O 2 0 3 0 solvent is extremely hygroscopic and the inclusion CONCENTRATION, w i n g / kg. of the water was intended to obviate a t least a part of this difficulty. Intrinsic viscosities measured in Fig. 1.-Osmometry data for six 40-methylglucuronoxylans. this solvent mixture and in dry dimethyl sulfoxide cm. differed by less tlhan lo%, indicating that the small quantity of water present had little effect on the configuration of the polymer in solution. In preliminary experiments, the previously described striations'j n-ere very pronounced and rendered accurate nieasurements impossible. A modified Dandliker and Kraut16 cell was therefore adopted, designed to eliminate both sedimenting and floating material (Fig. 2 ) . These striations in the scattering light-beam were due to colloidal debris, A B as suggested by tin experiment in which a 1% polymer soluticm was subjected to a preliminary ultra- Fig. 2.-Light-scattering cells according to (-4)Dandliker and Kraut and (B) present modification. centrifugation folr various lengths of time at 140,000 g, followed by ultmcentrifugation a t 35,000 g in the TABLE I1 modified light-sc,zttering cell. The results are sumLIGHT-SCATTERING DATA marized in Fig. 3. Both the dissymmetries (2)and -. MW (C/ the reduced turbidities (?/e) decreased markedly Polysaccharide T)CA z x 10for the first three hours of high-speed ultracentriWhite birch 29 4 1.55 76 fugation after which they remained constant. In 29.4 1.63 contrast, the decrease in concentration over the 34 3 1.28 entire period was less than 27,. Obviously, a small 24.4 1.4s portion of the polysaccharide was of colloidal diWhite birch, 0-acetyl 24.4 1 08 78 mensions. Whether this material was a non-xylan Whit e birch, bark 27.0 1.06 70 impurity or an aggregated xylan, its presence proYellow birch 30 0 1.33 75 duced results which were not representative of the White elm 29.4 1.18 70 bulk of the polymer. h preliminary ultracentrifugation a t 140,000g for 3 hr. was therefore adopted as value was accordingly used throughout. As shown a standard procedure prior to the usual centrifuga- in Table 11, the values agree fairly well a_nd tion a t 35,000 6' in the cell. Longer times were are consistent with the number-average values (Mn) avoided because of the possibility of molecular in Table I. fractionation of 1.he polysaccharides. Kumber- and weight-average degrees of polyTypical Light-scattering data are given in Fig. 4. merization are summarized in Table 111, together The variation of the dissymmetries with concentra- with intrinsic viscosities determined in cupriethyltion exhibited no clear trend and a mean value was enediamiEe and dimethyl sulfoxide containing 2% therefore adopted (Table 11). The linear extra- water. Pnand P, represent the number of xylose polation to zero concentration applied for obtaining residues in the linear framework of the polysaccha( c / T ) ~ -is~ d5o s~!own. The weight-average molecride, exclusive of the acid side chains. There is no ular weights ( M - ) in Table I1 were calculated great difference in the Pn and P, values obtained from these values and from the refractive index in- for thg various 4-0-methylglucuronoxylans. The crements (dnldc) in the usual way. The values for ratio P,/Pn is 2.4 += 0.2 which is near the value of dn/dc varied solnewhat for the different polysac- 2 expected for a polymer with a Flory distribution. charides, alheit r i t h no apparent trend. A mean The number-average degrees of polymerization found here are considerably higher than those re( 1 3 ) T L: rum11 a n d E C Jahn. Svensk Papperstzdn., 6 4 , 831 ported by other investigators for 4-0-methylglu(1951). (14) J. D. Wethern, T a p p z , 36, 1067 (1952). curoxylans from similar sources. In two case^^^,^* (15) M. M Huque, J . Jaiiorzyn and D. A . I. Goring, J . Polymer degradation undoubtedly occurred during the isoScz , 39,9 (1959) u
-
1
i
w
I
U
I
U
v
4
"
-
-
U
a,
(16) W. B. 3andliker and J Kraut, J . A m . Chem. SOC.,78, 2380 (1956).
(17) E. Husemann, J . prakt. Chem., lS6, 13 (1940). (18) M.A. Millett and A . J. Stsmm, Tnxs JOURNAL. 61, 134 (1947).
D. A. I. GORING AND T. E. TIMELL
1428
I
I
I
I
2
4 6 8 TOTAL 1IME ( h r ) PRECENTRIFUGATION AT 14qOOO g,
Fig. 3.-Effect of preliminary ultracentrifugation on dissymmetry (Z), reduced turbidity (.r/c), and concentration (c).
Vol. 6-1.
appears quite conceivable that the insoluble part had a higher average molecular weight than the soluble portion. The only previous attempt to determine the weight-average degree of polymerization of a xylan seems to be a light-scattering investigation of a hemicellulose from beech wood.*O This 4-O-methylglucuronoxylan21was converted - to a mixed acetate-benzoate derivative, the P, of which was found to be 175 f 10. This low value was probably due to several factors, including the degrading effect of the acid chlorite used for delignifying the wood, and the fact that fractionations were carried out for purification of the polysaccharide. Light-scattering data recently obtained with nitrated, native celluloses from white birch wood22 indicate a weight-average degree of polymerization for these polysaccharides of 9,000. Other hardwood celluloses appear to have approximately the same molecular ~ v e i g h t . ~When ~ this value is compared with the P, values found here for hardwood xylans, 450-500, the considerable difference in molecular size between these two wood components becomes evident. The angular dependence of light-scattering has been used for evaluation of the configuration of the beech xylan referred to above.20 It is clear, however, that the present dissymmetries were so near to unity and so markedly affected by small amounts of aggregates that this parameter cannot be used to elucidate the molecular shape. A somewhat more reliable indication of the configuration can be obtained by comparing the intrinsic viscosity with the degree of polymerization. If, as with cellulose in c~priethylenediamine,~~ the coefficient a in the Mark-Houwinck equation [SI =
0 0
I
u
I 6 8 IO 1000 c ( g c m - 9 , Fig. 4.-Light-scattering data for three 4-0-methylglucuronoxylans. 2
2
4
lation of the hemicelluloses. An aspen xylanl9 gave esters which were only partly soluble and it (19) 5. 0. Thompaon and L.
E. Wiae, T a p p i .
35, 331 (1952).
IC@
where [77] denotes the intrinsic vizcosity and K is a constant, is unity, the ratio [q]/Pnshould be constant. Even if n departed somewhat from unity, this ratio would be almost coilstant in the present case because of the narrow range of molecular weights involved. As is shown in Table 111, the ratio of the intrinsic viscosity in dimethyl sulfoxide to the gumber-aver/ P ~virtu, age degree of polymerization, [ q ] ~ ~ l swas ally constant for the four 4-0-methylglucuronoxylans from wood. The larger ratio noted for the native, 0-acetylated polysaccharide was probably due to the greater solvation of the 0-acetyl groups by the dimethyl sulfoxide, which would cause the macromolecule- to expand. Contrary to this, the ratio [ ~ ] c E D /was P ~somewhat variable, ranging to 6.00 X The 4-0-methylfrom 3.69 X glucuronoxylans all carried acid side chains with carboxyl groups which would ionize in a solvent such as cupriethylenediamine. This would cause polyelectrolyte expansion which would csusc an increase in the intrinsic viscosity over that expected for an uncharged chain. Thus, the exceptionnlly Rt. Horio, R. Imamura and H. Inagahi, ?bid., 38, 210 (1950). (21) G. A. Adams. Can. J. Cfiem., 35,556 (1957). (22) D.A. .I Goring . and T. E. Timell, Suensk Papperstrdn , 111 preir. (23) T. E. Timell. ibid., 60, 830 (1957). (24) E. II. Immergut, B. G. Rsnby and 11. F. :?ark. J n d . En8 Chsm.,48, 2483 (1953). (20)
Oct., 1960
high value for the elm xylan is in accord with the higher acidit;y of this p~lysaccharide.~It is probable that in vismcometric determinations of the molecular weight of 4-O-methylglucuronoxylans, dimethyl sulfoxide is a more suitable solvent than cupriethyleriediamine. O F ~'Ol.YMERIZhTION AND INTRINSIC \rISCOSITIES
O F TIIE Polysnccharide
White birch White birch, 0-acetyl White birch, bark Yellow birch Whiteelm Sugar maple
POLYSACCHARIDES 1000
1000
-
.-
P3
1 ' ~
1,w
[9lzM8/
[.I
:~~CED
Table IV. All preparations were homogeneous on electrophoresis on glass fiber paper in a borate buffer.
TABLE IV YIELDAND GENERALPROPERTIES OF THE POLYSACCHARIDES ObWood, tained,
Polysaccharide
TABLE I11 1) EGREES
1429
hfOLECUL.4R P R O P E R T I E S OF S I X 4-0-METHYLGLUCURONOXYLANS
I',
1 . 1 1 ~ ~ s Pn
215 500
0.89
4.13
0.67
3.11
180 470
.87
4.82
.98
5.44
'234 464 192 495 185 440 215 . .
.84 .93 1.11 0.97
3.59 4.84 6.00 4.50
..
..
.57 .56 .65
2.96 3.02 3.02
The mean Value of [ q ] C E D / p n was 4.7 x low3, which is considerably smaller than the value reported for cellulose24within this molecular range, 8.1 X Evidently, a 4-0-methylglucuronoxylan with. the same number of residues in the main chain as the cellulose molecule will, nevertheless, occupy only one-half of the effective volume of the latter in cupriethylenediamine, and this in spite of the additional volume and polyelectrolyte expansion caused by the presence of the acid side chains. The xylose residue is of the same length as the glucose residue and both are probably present in the C 1 conformation in xylans and cellulose. The reason for t,kiis difference in hydrodynamic volume is therefore not clear a t the present. Further physicochemical studies with fractions over a wide range of molecular weight's should reveal additional configumtional details. Experimental Isolation of the Hemicelluloses.-Extractive-free wood nicnl (800 g.. 40-60 mesh) from a freshly cut specimen of white birch wm extracted for 10 hr. in a nitrogen atmosphere with 247, (w./w.) potassium hydroxide ( 4 1.). After filtration through s.interec1glass and washing with water, the filtrate and wasliings ( 6 l .) were poured into ethanol (15 l.) and hydrochloric :Acid W:LS added unt,il a pH of 2.5 had been reached. Thc: precipkate was recovered on the cent,rifuge and washed five timas with three times its volume of 70% ethanol, followed by anhydrous ethanol and petroleum ether. The m Iterinl NBS dried for a week in mcuo over potassium hydroxide to give a white, fluffy powder ( l i 6 g., 22% of the mood). T h t hemic*ellulose was ash-free and, on hydrolysis, g_.nvccrnly xylose and uronic acids on the paper chromutogmn-i. 'rhc other 4-O-methylglucuronoxylans were isolated in :b simihr v:ay. The O-ace~;yl-~-O-mstliylglucuronoxylanwas obtained from a whit,e birch chlorine holocellulose13 by exhnustive extraction n-ibh dirrtethd sulfoxide25 as described previo ~ i s l y '(yield, ~ 50%j The hemicellulose was ash-free and w:ts easily sc'lublc in water. Extractive-free inner bark of white birch was treated with hot. water and hot aqueous ammoiiium oxalate f ' x removnl of pectic material, followed by treat,inent with acid chlorite as dcscribed elsewhere.* Extraction of the rca:tirie n;it,h aqueous potassium hydroxide gave a 4-0-methylglucuronoxylan containing only traces of glucose residues and repreventing almost all the xyloseyiolding mnterin.1 in tJhe bark. The yields obtained and the general proper t k s of the polysaccharides are summarized in (25) E. Higglund. R . Lindbcrg and 3. McPherson, Acta Chem., Scand., 10, 1160 (13.56).
%
%
Methoxyl,
%
Uronic anhydr.,
%
Xylose 5qr acld group
ia]Da
Whitebirch 29 22 2 . 0 10.9 10 -83" White birch, 9.8 10 -860b 0-acetyl 33 17.5 2 . 1 White birch, 9.9 10 -68" bark 30 28 2.2 23 17.5 2.0 11.2 10 -81" Yellowbirch Whiteelm 14 10 2 . 6 15.2 7 -70" Sugar maple 17 13 2 . 1 12.0 10 -80" Specific rotation at 20" in 10% sodium hydroxide. I n water. Acetylation of the Hemicelluloses.-Dry hemicellulose (2.0 9.) was dissolved or swollen in freshly redistilled and dry formamide (40 ml.).aE After 24 hr., dry pyridine was added (80 ml.) and, with intervals of 12 hr., three portions of redistilled acetic anhydride (20 ml. each). The reaction mixture was shaken vigorously and soon settled t o a transparent gel. After an additional 24 hr. a t room temperature, the mixture was added with stirring t o ice-water (3 1.) containing 2% hydrochloric acid. The product was washed with ice-water until neutral and then with ethanol and petroleum ether. After drying in uucuo a t room temperature for 3 days, a hard, grayish powder was obtained (3.1 g.). Acetyl analysesa7 indicated complete substitution throughout. The acetates were soluble in chloroform only after addition of 10% (v./v.) of ethanol. Osmotic Pressure Measurements.-The instrument used wa8 the Stabin-Immergutlz modification of the Zimm-Myerson osmometer.$* The membranes were freshly prepared cellophane films which had never been allowed t o dry.lo Syringes with long needles were employed for filling the OSmometers and mercury was used for closing the instruments completely during measurements. The osmotic height was measured with a cathetometer by the static method, equilibrium being established within 3-5 hr. The temperature was 30 0.01'. No diffusion through the membranes was noticed. The solvent was a 9: 1 (v./v.) mixture of reagent grade chloroform and anhydrous ethanol. Concentrations were determined by weight, thus eliminating the need for density measurements. The osmotic height was determined a t 5-6 different concentrations. In the case of the 4-0-methylglucuronoxylan from white birch wood, two series of measurements were made with different membranes and with different osmometers. The results obtained agreed closely (Fig. 1). Molecular weights were calculated from the general relationship. ir/c = R T / a n [ l Bc]
*
+
where B is the second virial coefficient and the other symbols have their usual significance. Linear extrapolation to zero concentration yielded the value (h/w),o where h was the osmotic height in cm. solution and w was the concentration in g./kg. solution." Number-average molecular weights were obtained from the simplified relationship SIn= 25,70O/(h/w),a The corresponding degree of polymerization, 8, (number of (26) J. F. Carson and W. (1946); 70, 293 (1948).
D. Maclay, J . Am. Chem. Soc., 68, 1015
(27) L. B. Genung and E. C. Mallatt, Znd. Eng. Chrm., Anal. Ed., 13, 369 (1941). (28) B. H. Zimm and I. Myerson, J. A m . Chem. Sac., 68, 911 (1946).
(26) Provided by American Viscose Corporation through the courtesy of Dr. C. P. J. Glaudemans. (30) P. M. Doty and H. M. Spurlin in E. Ott, H. hl Spurlin and AI. W. Grafflin, Editors, "Cellulose and Cellulose Derivatives," Interscience Publishers, New York, N. Y., 1955, pp. 1133-1163.
1430
D. A. I. GORING AND T . E. TIMELL
Vol. 04
xylose residues in the main xylan chain), was calculated from L u ~ o x . ~The ~ , refractive ~~ index correction of Carr and the relationship P . = an(n/Mr),where n was the number Zimm41was applied. The dissymmetry, 2 , was c:tlculatcd as of xylose residues per acid side chain and M , was the molec- the ratio of the intensities a t 45 and 135'. hfeasurements made a t six different concentrations, e, ranging from ular weight of the repeating unit of the p ~ l y s a c c h a r i d e . ~were ~ Viscosity Measurements.-The instrument used was a zero to 0.01 g./ml. and solvent scatter was subtracted from Craig-Henderson viscometer,32 designed to eliminate effects that of the solution. Some polysaccharides, and especially of surface tension and to minimize kinetic energy corrections. the 0-acetylated xylan, gave strongly fluorescent solutions. Its capillary radius was 0 01778 em., capillary length 18.59 This difficulty was overcome as described in a previous cm., hydrostatic head 25.0 cm., and efflux volume 0.852 ml. publication.'2 The original concentration of the polymer The temperature was 30". Samples of dry hemicellulose was determined by adding a carefully measured aliquot to a (225 mg.) were dissolved in M-cupriethylenediamine or in 1:l mixture of ethanol and ethyl acetate. The precipitate freshly distilled dimethyl sulfoxide containing 2y0 (v./v.) of formed was collected on a sintered glass filter cruciblr water (15.0 ml ). For measurements, 10 ml. was used, and (coarse), provided with a mat of arid-washed asbestos. dilutionH were made directly in the viscometer, the concen- After washing successively with ethanol and petroleum trations being varied from 1.5 to 0.2 g./dl. Reduced viscosi- ether, the crucible with its content was dried a t 60" i? ties were plotted against concentrations according to Hug- vacuo for 3 hr. and weighed. The reproducibility of thls gins33when linear relationships were obtained throughout. method was excellent. Refractive index increments were measured in a BriceExtrapolation to zero concentration gave the intrinsic visPhoenix differential refractometer." Both solutions and cosity. Light-scattering Measurements.-Light-scattering meas- reference solvent had to be protected from osmospheric moisture to ensure reproducible results. In addition to the urements were made with a B r i c e P h ~ e n i x ~photometer ~-" usual precautions, equal quantities of solution and solvent a t a wave 1eng;h of 5460 A . The instrument was modified were carried through identical manipulations prior to measas described previously86~37to permit the use of light-scatter- urements. Excellent agreement (&1.2%) was found being cells amenable to ultracentrifugation as described by tween duplicate values of the increment on any one sample. Dandliker and Kraut.3* Even in these cells, however, The values obtained for the various fractions differed sompstriations were observed in the light beam due to the pres- what, however, and a mean value of 0.064 0.006 was ence of micellar debris, and a new type of cell was accord- used. ingly devised and used in all subsequent experiments (Fig. The light-scattering data are summarized in Table I1 2). The narrow tubings a t the base and top of the new cell where ( c / ~ ) ~ - - ovalues, dissyEmetries ( 2 ) and weight,effectively captured all sedimenting or floating material. average molecular weights (M-) are given. Molecdar The solvent was dimethyl sulfoxide, freshly redistilled in weights were calculated from the relationship vacuo, and containing 2% of water (v./v.). A 1%solution 3 X 4 N - ~of the polymer in this solvent was subjected to a preliminary ~@ . " = 32r3n02(dn/dc)2(c/7),,o ultracentrifugation a t 140,000 g (40,000 r.p.m.) for 3 hr. in the No. 40 rotor of a Spinco Model L ultracentrifuge.% The where X is the wave length of the light, 'V is Avogadro's clear solution vias carefully decanted and transferred to the number and no is the refractive index of the solvent. Dislight-scattering cell which was floated with carbon tetra- symmetry corrections were applied. The rather high values chloride-ethanol mixtures in the cups of a SW 25.1 swinging for 2 for the first two measurements with the 4-0-methylbucket rotor of the same ultracentrifuge. Centrifugation glucuronoxylan from white birch wood were probably due to was a t 35,000 g (20,000 r.p.m.) for 1 hr. The cells were a shorter time (2 hr .) of preliminary ultracentrifugation. transferred directly to the photometer and scattering inten- The molecular weight of this polysaccharide was accordingly sities were measured at 45, 90 and 135" relative t o the inci- computed from the mean of the four c / r values, corrected by dent beam. The turbidity, 7 , was calculated from the 90" a factor based on the two lower dissymmetries. The values scatter by means of a calibration factor determined with obtained for the polysaccharide from sugar maple were not deemed reliable and have therefore not been included in (31) F. W. Baith and T. E. Timell, J . Am. Chsm. Soc.. 80, 6320 Tables I1 and 111. ~~
(1958). (32) A. W. Craig and D. A . Henderson, J . Polymer Sci., 19, 215 ( 1958). (33) hf. 1,. Huggins, J . Am. Chem. Soc., 64,2716 (1942). (34) B. A. Brice, M. Halaer and R. Speiser, J . O p t . SOC.Amer.. 40, 768 (1950). (35) Manufactured by the Phoenix Precision lnstrument Co., Philadelphia, Pa. (36) M. M. Huque, D. A. I. GoringandS. G. Mason, Can. J . Chsm., 36, 1962 (1908). (37) M.IM. H u p e , J Jaworzyn and D .4. I. Goring, J . Polymer Sce., 39, 9 (1959). (38) hlanufactnred by Beckman Instruments, Inc , South Pasadena, Cal forma.
Acknowledgment.-The authors wish to express their gratitude to Mr. W. L. St'eyn, who carried out the osmotic pressure and light'-ecattering measurements. They are also thankful to Mr. J. Jaworzyn for valuable assistance. (39) A colloidal silica. manufactured by E. I. d u Pont de Nemours and C r , Wilmington, Del. (40) D. A. I. Goring, M. Senez, B. Melanson and hI. M. Huque. J . Colloid Sci., 12, 412 (1957). (41) C. I. Carr and B. H. Zimm, J . Chem. Phys., 18, 1616 (1950). (42) P. R. Gupta and D. A . I. Goring, Can. J. Chem.. 38, 270 (1960).