- Email: [email protected]
Ionisation and solvation of d -glucose
Carbohydrate Research, 140 (1985) 169-183 Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherfands
IONISATION JOHN A.
W. M.
Laboratory Eindhoven
AND SOLVATION
OF D-GLUCOSE
BEENACKERS*, BEN F. M. KWXER, AND HESSEL S. VAN DER BAAN
of Chemical Technology, (The Netherlands)
Eindhoven
University
of Technology,
P.O.
Box 513, 5600 MB
(Received September Z&t, 1984; accepted for publication, January 28th, 1985)
ABSTRACT
For a quantitative description of chemical reactions of carbohydrates in concentrated solutions, a detailed knowledge of the ionisation equilibria is a prerequisite. In a series of experiments involving a wide range of concentrations of D-glucose, the ideal or non-ideal solution models did not accord with the observations unless hydration was taken into account. A hydration number of 3.5 was found for molecular n-glucose, but no hydration for dissociated D-@UCOSC. This result is discussed in terms of intramolecular hydrogen-bonding within the Dglucose ion. INTRODUCTION
This study is part of a project on the heterogeneous alkaline isomerisation of carbohydrates using ion-exchange resins l. In considering these reactions, the ionisation and solvation of the carbohydrates is of great importance. The ionisation of carbohydrates in alkaline aqueous solutions causes mutarotation, and, via carbanions (enolate ion), isomerisation and degradation. Reducing mono- and oligo-saccharides are weak acids and the ionisation of the anomeric hydroxyl group is an essential step in the isome~sation and epimerisation reactions. As ionisation is much faster than mutarotation2,3, the ionisation constants of a and p forms can be distinguished. Los and Simpson4 found a value of 0.29 for Apf& (= pKm_,,, - pKP_olc)for the pyranose forms, whereas De Wit et aI. found a value of 0.19. When only one pK, value is reported, it must be considered to be an overall ionisation constant and these have been determined for many carbohydrates at various temperatures6-25, using a variety of methods. Table I lists the pK, values at 298 K for D-glucose, D-fructose, and D-mannose. All concentrations are expressed in mol.mv3. The concentrations used for pH and pKs are expressed in kmol.mw3, so that pK and pH values can be compared with literature data. From the data in Table I, it can be concluded that, at 298 K, pKo,, = 12.4 40.25, plc,, = 12.1 f0.3, and pkl,, = 12.1 t0.1. *Present address: DSM Research and Patents, P.O. Box l&6160 MD Geleen, The Netherlands. OOOS-6215/85/$03.30
@ 1985 Elsevier Science Publishers B.V.
J. A. W. M BEENACKERS, B F M. KUSTER, H S VAN DER BAAN
170 TABLE
I
IVNfSATlON
CONSTANTS
fpf(,)
FOR AQUEOUS
SOLUTIONS
~-Glucose
D-Fructose
12.23 12.107 12.09 12.96 12.34 12.49” 12.2@ 12.87 12.35 12.38 12.46 12.51 12.72’ 12.35 12.78O.* 12.60b,d 13.9d,’
11.99 11.693 11.68
298 K D-Mannose
*p-Anomer.
Ref. 11 12 13 14 15 16
12.67 12.21
%-Anomer. TABLE
AT
17 18 19 20 21 22 72 5 5 5
12.13
12.27 12.31 12.553’
12.08
14.2d,’
14.0”’
cAt 283 K. dAt 376278
K.
3
II
LITERATURE
pl(,,,
VALUES
AS
A FUNC-I-ION
OF THE
CO~C~~TRA~O~
OF D-GLUCOSE.
AND
DETERMINED
PO~NTiOMETRlCALLY
Concentration (kmoI.m-3)
0.05 0.10 0.20 0.50 1.0
of
D-&cose
Michaefis and RonaLS
Thamse#
290-292 K
291 K
27.7 K
12.46 12.44 12.40
12.97 12.93 12.88
12.38 12.28 12.26 12.05
When the results of De Wit et aL5 are omitted, D-fructose seems to be more acidic than D-glUCOSe (Apfc, = 0.27 +O,lO). Izatt et ~1.~~ascribed the lack of agreement between the results of the various studies to the differences of the ionic strength of the solutions used. At 273 K, ThamsexP found a slight increase of pK, with increase in ionic strength. Degani 23, however, was unable to find any influence. Only three authors have described a dependence of the pK, on the concentration of hexose. Michaelis and Rona and ThamsenX4. using a potentiometric method, found that pKs decreased with increase in the concentration of D-glucose (Table II). The data of De Wit and co-workers5,2h, obtained using n.m.r. and U.V. techniques, reflected an increase of pKolC with increase in concentration of hexose. Also, in an alkaline ion-exchange resin, proton abstraction will take place
171
IONISATION AND SOLVATION OF D-GLUCOSE TABLE III EXPERIMENTAL
RESULTS
AND
CALCULATIONS
OF KGtal;
c,,
IS NOT IN THE TABLE
BECAUSE
Expt. C, x 1O-3 C, x 1O-3 pH
c,
-
c,
PKGW
PKG,,
(8)
(9)
(10)
0.447 3.301 8.774
0.632 0.571 0.584
12.200 12.243 12.226 12.234
0.117 1.000 3.162
0.454 3.462 8.980
0.612 0.602 0.603
12.213 12.221 12.218 12.220
10.04 1.234 11.00 9.986 11.20 14.331
0.110 1.000 1.58
1.124 8.986 12.746
0.604 0.611 0.595
12.219 12.214 12.220 12.226
0.158
10.03 1.221 11.00 9.782 11.20 14.231
0.107 1.000 1.585
1.114 8.782 12.646
0.612 0.597 0.590
12.213 12.224 12.222 12.210
0.371 0.352 0.329
0.351
10.00 2.828 10.50 7.757 10.80 13.772
0.100 0.316 0.631
2.728 7.441 13.141
0.741 0.683 0.659
12.130 12.166 12.159 12.181
6
0.371 0.353 0.329
0.351
10.01 2.737 10.50 7.584 10.80 13.830
0.102 0.316 0.631
2.635 7.268 13.199
0.699 0.665 0.662
12.156 12.177 12.171 12.179
7
0.690 0.673 0.609
0.657
9.81 3.945 10.11 7.573 10.50 16.442
0.065 0.129 0.316
3.880 7.444 16.126
0.876 0.868 0.860
12.057 12.061 12.061 12.065
8
0.711 0.701 0.626
0.679
9.90 5.037 10.01 6.314 10.50 16.361
0.079 0.102 0.306
4.958 6.212 16.045
0.884 0.874 0.932
12.054 12.059 12.064 12.080
9
1.194 1.106 1.007
1.102
9.53 4.479 10.00 11.448 10.30 19.366
0.034 0.100 0.199
4.445 11.348 19.167
10
1.144 1.108 1.047
1.100
9.53 4.403 10.00 11.281 10.20 16.175
0.034 0.100 0.168
4.369 11.181 16.017
1.131 1.019 0.980
11.946 11.992 11.982 12.009
11
1.526 1.496 1.384
1.469
9.30 4.040 9.50 5.966 9.90 12.986
0.020 0.032 0.079
4.020 5.934 12.907
1.324 1.259 1.185
11.878 11.900 11.901 11.926
12
1.524 1.493 1.384
1.467
9.33 4.207 9.50 6.162 9.90 12.998
0.021 0.032 0.079
4.186 6.130 12.919
no.
(mol.m-3) (mol.m-3)
(1)
(2)
(3)
(4)
1
0.0635 0.0611 0.0563
0.0602
10.05 0.559 11.00 4.301 11.50 11.936
0.112 1.000 3.162
2
0.0635 0.0610 0.0561
0.0602
10.97 0.571 11.00 4.462 11.50 12.142
3
0.171 0.156 0.148
0.158
4
0.171 0.156 0.148
5
(5)
(6)
(7)
11.958 11.984 11.985 12.012
11.890 11.885 11.900 11.926
172
.I.A. W. M. BEENACKERS. B. F. M KUSTER. H. S. VAN DER BAAN
KOH-burette u
_----------_
magnetic
pH control
ler
recorder reactor
magnetic stirrer
Fig 1. Reactor for potentiometric
titrations.
before isomerisation. Inside such a catalyst, the concentration of the components is relatively high. In view of the discrepancies associated with the literature data, the degree of ionisation at high concentrations of hexose has been determined. EXPERIMENTAL
The ionisation measurements were carried out by potentiometric titration using analytical grade chemicals, doubly distilled, CO,-free water, and a thermostated reactor (150 mL) provided with a magnetic stirrer (Fig. 1). The pH was measured with a glass electrode (Radiometer, type GK 2401 B) in combination with a pH controller (Radiometer, type TIT lc), and corrections were applied for the temperature and the concentration of the alkali. The electrode was calibrated with buffer solutions (Merck Titrisol). The titration was controlled by titrating pure water with CO,-free M KOH. When >5 mL of KOH solution was added, the measured and the calculated concentration of HO- differed by <5%. The isomerisation of the carbohydrate under study was neglected. A solution (50 mL) of D-glucose of known concentration under nitrogen was titrated to three different pH values at 298 K. Columns (l)-(5) of Table III show the final composition of the D-glucose solutions after adding the appropriate amount of alkali. RESULTS AND DISCUSSION
In solution, the chemical potential of each component27 is p, =
~7 + RT ln (Y, . C,),
and for the solvent.
(1)
IONISATION
pH,O
AND SOLVATION
=
&,O
+
RT
173
OF D-GLUCOSE
In bH,O
(4
. ch2O)y
where p&O is the chemical potential of pure water, and consequently the concentration of water is expressed as a fraction: cil,
=
(3)
~,dCH,O,pure.
As the experiments were carried out at constant pressure, PT is a function only of the temperature. For non-ideal solutions, y, + 1. When a solution is in equilibrium, the chemical potential (G) will be a minimum and AG will be zero. Hence, XV, . pi eq = .Z{v1 * ~7) + RT S{ V, * In (yi . CJ}eq = 0,
(4)
AG* = S{ q * /.~..f}= - RT ~{ Vi. In (yi * CJ}eq, and
(-5)
AGE = - RT~{v~ . In yi}.
(6)
As p: is a function only of the temperature, AG* will be also a function only of the temperature. The equilibrium constant is defined as In K, = Z{vi . In (y, . C,)} = - AG*/RT.
(7)
Because the pressure remains constant, the equilibrium constant should be only a function of the temperature. Applied to a solution of D-glucose (G) and including hydration of all species:
GH * (H20)h,,
+ q H,O%
G-
(HzO),,,_ + H+ . WA%,+
(8)
K
;H+
~40
- (H,O) ,,“+ + HO-
(HzO)hHO
(9)
where h = hydration number, q = ho- i-h,+ - hoH, and
(10)
p = h,+ + h,,-
(11)
KG = yG- ’ %
K H,O
=
YH+ . cH+
+ 1. ’
YH+ * cH+b3H
’ YHO-
’ %@I,0
’ %H
* h!,O
’ %,O>
’ ci20>
(12)
(13)
174
J. A. W M. BEENACKERS,
B. F. M. KUSTER,
H S. VAN DER BAAN
In the following sections, the equilibrium constant will be calculated on the basis of three assumptions: (a) ideal solution, no hydration, (b) non-ideal solution, no hydration, (c) non-ideal solution, hydration. In an ideal solution, the dissociation constants for n-glucose and water (Eqs. 22 and 13) are simplified to: K ?G) = Co- . Cn+/Con, and
(14)
K H,O(d= c,+ . q-I,-,
(15)
where the index (a) refers to the assumption (a). During titration, centration of o-glucose decreases and the following relations hold: C H+(aj= 103-pH CHO-(a)
CG-(a)
C GH(a)
=
mol.mm3
KHzO(&H+(~)
=
CK+(a)
-
CG(a)
the total con-
+
cH+(,)
-
CG-(a)
(16)
(PKH&l,298
-
C”O-(a)
=
13.9965)
(17)
(electro-neutrality)
(18)
(o-glucose balance)
(19)
In Table III, columns (6) and (7) show the results of the calculations of C,,and Co-. Furthermore, the values of Kc0 and pKo,,, according to Eq. 13 are
t pKG
12.5-
24
Q
Thamsen
m
MichaelIs
0
Present
and Rona
25
work
12 0\
I
0
0
500
1 1500
I
1000 c,---
Fig. 2. pK, as a function of the concentration the data of Thamsen” were used to recalculate
of D-glucose at 298 K. The temperature dependence them to the reference temperature of 298 K.
of
175
IONISATION AND SOLVATION OF D-GLUCOSE
presented in columns (8)-(10). 0 ur results, in combination with the potentiometric data of Thamsen24 and of Michaelis and Rona25, are given in Fig. 2. Our data give roughly the same concentration-p& relation as found by the other authors. However, it is clear that this approach does not lead to an equilibrium constant that is independent of the concentration. For a non-ideal solution without hydration, Eqs. 12 and 13 are simplified to: K G(b)
=
K H,O(b)
YG- . CG-. YH+ . &+/(YGH =
YH+ . cH+
* G.J, and
* YHO~ ’ cHO-/(YH20
’ CH,O),
(20)
(21)
with aHO 2 2 = yHO. 1 CHo. It is difficult to calculate the thermodynamic activity of water in a multi-component system29. For our experimental conditions (Table III), the concentration of Dglucose Co, is much higher than the ionic concentrations. For this reason, the activity of water was provisionally assumed to be equal to aHsOin a pure solution of D-glucose. The latter can be calculated from the measurements of Bonner and Breazeale30, who gave the activity coefficient of o-glucose (eon) and the osmotic coefficient (+oH) as a function of the molality (moH) of D-glucose in a neutral solution: yGH
=
1 + 0.022mbg
+on = 1 + O.O12m&$
(O <
IllGH <
2,
(23)
(O <
IllGH <
2,
(24)
The activity of water can be calculated from ref. 31 as In aHI = - mGH . Mqo * +GH . 10W3
(0 c mGH c 2).
(25)
The activity coefficients yGmand -yuo- have been calculated with the Debye-Htickel expression, as corrected by Robinson and Stokes31 for solvation and the activity of the solvent: log y = - Ar - I”*/(1 + B, * d, - I’“) - log (1 - 0.018 hi . mi) - h, . log aHZO,(26) with A,, K = 0.5115 kg’“.mol-ln (Debye-Hiickel constant), B 2ssK = 3.291 lo9 kgtn.molvz. m-l (Debye-Hiickel constant), d,- = diameter G- - 8 x lo-i0 m, d,,- = diameter HO- - 2 x lo-lo m, hi = solvation number: mol H,O per mol i, I = mxf + mH+ = mG- + mHO- = ionic strength (mol.kg-i).
J
176
A. W. M. BEENACKERS.
B. F M. KUSTER.
H. S VAN DER BAAN
The activity of water is calculated from yen, yo- , and yno- with the Gibbs-Duhem equation 26. It appeared that aHzOcalculated with Eq. 25 is a good approximation. The thermodynamic quantities are between the following limits. 1.000 < Yen < 1.045
(27)
0.882 < -yo- < 1.000
(28)
0.853 < -yHo- < 1.000
(29)
0.967 < aHto < 1.000
(30)
To calculate the ionisation applied, namely
constant,
a molality-molarity
m, = C,l@ - MO, . Co, - Mom .
C,-
- Mu,-
conversion
. C,,m),
has to be
(34
and for the density of the solution, p = 1000 + 0.067co.
(32)
In Fig. 3, the equilibrium constant pKo,,, is given as a function of the concentration of D-glucose. It is seen that the concentration dependence of p&,,, is almost unchanged with respect to that of pKo(,). Apparently, it is impossible to 12.3 7
PKG
q,,g-
pKGfbl
=
12 20-
PKG(C)
=
12.23
022
x
10 -3 cG
‘0.02
C G-
Fig. 3. pKG as a function
of the concentration
of
D-ghCOSe
The
points
belong
to pK,,,,.
(C)
IONISATION
TABLE
AND SOLVATION
177
OF D-GLUCOSE
IV
HYDRATIONOF D-OLUC~SE(LITERATUREDATA)
---
Experimental method
h CH 267 K
278K
2.3 2.2 6 5
298 K 3.5 3.5 1.8
3.5 >lO 3.7 2.7 --
2.0
Ultrasonic46 Ultrasoniti* Dielectric relaxation53 170-N.m.r. relaxation53 Dielectric relaxationjl “0-N.m.r. relaxationSI Compressibility5’ “0-N.m.r. relaxation” Dielectric relaxation54 Freezing process5’ Activity method%
eliminate this concentration dependency in this way, even when using the best thermodynamic data from the literature. For a non-ideal solution with hydration, the literature data on the hydration of molecular D-ghCOSe are given in Table IV. At 298 K, it is seen that most authors have reported a hydration number of 3.5. For the hydration number of G-, no literature data are available. From the entropy change during ionisation, conclusions have been drawn28,32T33 about the hydration of GH and G-. The entropy change during ionisation in water is a result of (a) the change of the number of particles (from the point of view of statistical thermodynamics, an increase of the number of particles causes an increase in the entropy of the system; when the hydration of the species formed differs from that of the non-dissociated compound, hydration will have an influence on the total entropy change), (b) the increased ionic strength (ions give an increase of the electrostatic field in the solution; the solvent water is strongly polar, so that the water molecules will be hindered in their rotation”135, and this effect causes a decrease in entropy upon ionisation), (c) the intramolecular hydrogen-bonding (an increase will lead to a decrease of the entropy of the D-glucose molecule32,“). For ~-glucose in solution, an entropy change upon ionisation of - 110 J.mol-*.K-i is calculated14a20,22,49. Allen and Wright33 ascribed this negative entropy effect to a decrease in the number of particles by an increase in the hydration of D-glucose during ionisation. However, in our opinion, the entropy change upon ionisation cannot be explained only by assuming an increase of the hydration of D-glucose. The electrostatic field, combined with intramolecular hydrogen-bonding, must have a dominating effect. The stoichiometric coefficient p, as defined in Eqs. 9 and 11, is generally For the hydration of H+ and HO-, mostly 1 given as 2 in the literature 10-13.37,38.
J. A.
1’78
W. M. BEENACKERS,
B. F. M. KUSTER, 13 S VAN DER BAAN
and 0 are assumed37%X.Inside the ion-exchange resin, the concentration of SH, S-, and HO- can be very high. According to Schwabez9, it is impossible to determine activity coefficients at high concentrations of electrolyte. In seeking to describe the ionisation inside the resin, the literature information on the activity coefficients of the various components in our system was replaced for the simple assumption that the excess free energy AGE (Eq. 5) is zero and that further effects must be ascribed to hydration. For the water, the relative concentration (Ci;,O.free)of hydration water is not taken into account. This approach was also used by others19-@. Eqs. 12 and 13 then are transformed into:
with Cr!ro,, : .
= 1 - (6.28 Co + ho, - C,,
in Eq. 35, the totaf relative water concentration
cNzO,total
=
+ ho- + C,-)/55508. was used:
1 - 6.28 C&55508 (0 c Co G 2000)
as calculated directly from literature dataJ*. To obtain a concentration-independent defined (Eq. 37):
(36)
Koc,), an optimisation criterion 6 was
to be calculated from all 36 experiments of Table III. Minimising this criterion, with the requirement that h, > 0, yields as optimal data:
hGH=
3.5 and ho.- = 0 for q = 1 - ho, = -2.5.
The hydration number of 3.5 for molecular D-glucose agrees well with most of the literature data for this temperature (see Table IV). With these best estimates of the hydration numbers, the plc,,,, value can be plotted as a function of the concentration of o-glucose. In Fig. 3, it is seen that, in this way, a concentration-independent ionisation constant is obtained. It was noted above that hydration decreases with increase in temperature. Shiio46 described the “adsorption” of water per hydroxyl group with a Langmuir adsorption equation: In [hon/no,l(l
- honfnon)] = AH/CRT) +
cl,
(38)
IONISATION
TABLE
AND SOLVATION
179
OF D-GLUCOSE
V
HYDRATION
OF D-GLUCOSE
Temperature (W
293 298 308 318
AS A FUNCIION
OF THE TEhfPERATlJRE8
Experimental data of ShiioM
Calculated data
h G”,exo
h FH
4.2 3.5 2.8 2.2
4.05 3.12 2.92 2.07
Langmuir adsorption
LinearlFreundlich s hvd 1.28 3.93 1.74 3.47
h GH
s hvd
4.12 3.60 2.79 2.19
0.40 0.80 0.02 0.01 1.23
10.42 AH (kJ.mol-I) “Including
- 20
- 55
the data of Shiio&.
where non = number of hydroxyl groups (for o-glucose, non = 5) and c1 = a constant. For o-glucose, a heat of hydration of -55 kJ.mol-i was fourid. According to this model, the hydration of o-glucose cannot exceed the number of hydroxyl groups. At low temperature, however, Harvey et aL4’ found a hydration number of at least 10. For this reason, the experimental data of Shiio46 were recalculated with a linear and with a Freundlich adsorption model. The temperature dependence of those two models can be described with Eq. 39. In bon = - AH/(RT) + c2
(Linear/Freundlich)
(39)
The results are given in Table V. In columns (3) and (5), the hydration numbers calculated using Eqs. 38 and 39, respectively, using best estimates for AH and Ciare given: S,,,, is defined as
hyd= (h,, - ho~,exp)~~m,exp x loam-
(40)
It is clear that the linear/Freundlich adsorption models give a better fit for the description of the hydration of D-glucose. The corresponding heat of “adsorption” is then calculated to be -20 kJ.mol-‘. When this temperature dependence was applied to our experimental results, the hydration numbers given in Table VI were TABLE
VI
HYDRATION
T WI
h GH
OF GH AS A FUNCTION
267 9.2
218 6.4
OF THE TEMPERATURE
298 3.6
303 3.2
313 2.4
323 1.9
333 1.5
J A. W. M BEENACKERS.
180
B. F. M. KUSTER, H S VAN DER BAAN
found. The low-temperature hydration number ho,,,, x is about the same as that found by Harvey et a1.47. The temperature dependency of the ionisation constant is given by De Wilt*’ as: Eion.G
=
-16.8 kJ.mol-l.
(41)
The best formulae for the ionisation function of the hydration number: K G(c) =
c,-~* C,,,+/[C,,
of D-glucose can thus be presented
as a
. C;I,O.free(‘-hq
KG(c)‘KHzo(c) = c,- *cEi,o.free (‘+hG”‘G3,. c,,-)
(43)
with ho, = 3.6 exp [2OOOO/R. (l/T
(44)
- l/298)], and
K occj = 0.583 exp [16800/R . (l/T
- l/298)].
(45)
DISCUSSION
It has been found that, during the ionisation of D-glucose, the hydration disappears. This can be understood from the difference in conformation between the ionic and the molecular forms of D-glucose. In the D-glucose molecule, there are few if any intramolecular hydrogenbonds48. Proton abstraction from the anomeric hydroxyl group gives a negative charge on O-l. Delocalisation of this negative charge to C-2/6 by intramolecular hydrogen-bonding was suggested by Rendleman @. The conformation of D-glucose can then be considered as shown in Fig. 4 for the (Yform. All ions shown in Fig. 4 would exist in equilibrium with each other, with the C-l-anion preponderating. Due to intramolecular hydrogen-bonding, the free rotation of the hydroxyl groups decreases and they will be oriented preferentially in certain directions for optimal hydrogen-bonding. This effect contributes to the decrease in entropy of the Dglucose molecule on ionisation. The difference in entropy decrease between LY-and /3-D-glucopyranose during ionisation, measured by Los and Simpson4’, can be ascribed to a difference of the strength of ax,eq and eq,eq hydrogen bonds (see Fig. 5). In an ax,eq sequence of hydroxyl groups, e.g., HO-l ,2 in a-D-glucose, the system can easily adopt the geometry for efficient hydrogen bonding. The reverse is true for an eq,eq sequence of hydroxyl groups”. The hydration of sugars can be considered as hydration of the hydroxyl groups46,st ) involving hydrogen bonding as well as further hydration of hydrate water molecules. After ionisation of D-glucose, the hydroxyl groups are oriented and stabilised by intramolecular hydrogen-bonding. Consequently, no hydroxyl groups of D-glucose are then available for bonding to solvent water molecules.
IONISATlON AND SOLVATION OF D-GLUCOSE
H-O
/,’
'CH2
?’
.+--0,
181
CH:!
4===
,’
,,’
H-O,
CHz
Fig. 4. Intramolecular hydrogen-bonding in the cu-D-glucopyranose ion.
I’
a-~-Glucose dSI0,
=
non -110
P-o-Glucose
J mol-‘OK-’
dSI0”
=
,,+-0,
CHZ
ton -83
J
rid-
@ K-’
Fig. 5. Intramolecular hydrogen-bonding in the ions of a- and ED-glucopyranose.
LIST OF SYMBOLS
AT ; c; C&O C, di
E F, FH, Fru G, GH, Glc AG AG” AGE
Debye-Hiickel constant activity of component i Debye-Hiickel constant concentration of component i relative water concentration constant diameter of component i activation energy
m kJ.mol-’
D-fNCtOSe
D-glucose change of chemical potential difference of free energy of pure components excess free energy
kJ.mol-1 kJ.mol-1 kJ.mol-’
182
AH h, I K, M, MH, Man M, m, nGII
P PH PKS q R S SSH T y, 3:
J. A
W
M. BEENACKERS,
B F. M. KUSTER,
H S VAN DER BAAN
kJ.mol-i change of enthalpy hydration number of sugar i mol.kg-r ionic strength equilibrium constant D-mannose kg.moI1 mol weight of component i molkg-’ molality of component i number of hydroxyl groups of GH stoichiometric coefficient acidity: pH = 3 - log C,+ pK,:pK, = 3 - log KS stoi~hiometric coefficient J.mol-r.K-i gas constant sugar (S = SH + S-): G, F, M ionised sugar molecular sugar temperature activity coefficient of component i on molarity scale activity coefficient of component i on molality scale optimisation criterion chemical potential of component i in the solution J.mol-1 chemical potential of the pure component J.mol-r chemical potential of pure water J.mol-* stoichiometric coefficient liquid density kg.m-3 osmotic coefficient of component i
REFERENCES I J. A. W. M. BEENACKERS. Thesis, Elndhoven University of Technology, 2 C. S. HUDSON AND
3 4 5 6 7 8 9 10 11
12 13
J. K. DALE. 1. Am. Chem. Sot.,
1980.
39 (1917) 32@3-328. V. S. R. RAO AND J. F. FOSIXR. J. Phys. Chem., 69 (1965) 636-640. J. M. Los AND L. B. SIMPSON,Reck Trav. Chim. Pays-&~, 73 (19.54) 941-958. G. DE WIT, A. P. G. KIEBOOM. AND H. VAN BEKKUM, Reel. Trav. Chim. Pays-Bus, 98 (1979) 355361. K. SMOLENSKIAND W. KOZI.OWSKI, Rocr. Chem., 16 (1936) 270. K. SMOLENSKIAND W. KOZLOWSKI, Rocz. Chem.. 16 (1936) 281-287. B CAPON AND W. G. OVEREND, Adv. C~rbo~ydr. Chem., 15 (1960) 32. B. CASU, M. REGGIANI, G. G. GALLO, AND A. VIGERANI, ~e~ruhedron, 22 (1966) 3061-3083. V. YA LVGINA AND E. S. LYGIN. Izv. Vyssh. Uchebn. .&wed., Ashch. Tekhnol., 5 (1970) 34-36; Chem. Abstr., 74 (1971) 65,906. T MADSEN, 2. Phys. Chem., 36 (1901) 290. P. HIRSCH AND R. SCHLAGS, 2. Phys. Chem., Aba. A, 141 (1929) 387-412. F. URBAN AND P. A. SHAFFER, J. Bioi. Chem., 94 (1931-1932) 697-715.
IONISATION AND SOLVATION OF D-GLUCOSE 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
183
P. S~UCHAYAND R. SCHAAL, Bull. SOC. Chim. Fr., (1950) 819-824. G. C. TILDE AND W. F. K. Wmxn-JohEs, Trans. Faraday Sot., 49 (1953) 243-251. J. M. Los AND L. B. SIMPSON,Reel. Trav. Cfiim. Pays-Bas, 73 (1954) 941-958. N. A. RAMAIAH AND S. S. KATIVAR, Proc. Annu. Conv. Sugar Technol. Assoc. India, 29 (1961) 77-84. G. GUILLOT AND P. RUMPF, C. R. Acad. Sci., 259 (1964) 4064. C. A. BL~TON AND H. CHAIMORICH,J. Am. Chem. Sot., 88 (1966) 4082-4089. R. M. IzA~, J. H. RY~~ING, L. D. HANSEN.AND J. J. CHRISTENSEN,J. Am. Chem. Sot., 88 (1966) 2641-2645. H. G. J. DE WILT, Thesis, Eindhoven University of Technology, 1969. J. J. CHRISTENSEN, J. H. RWING. AND R. M. IZATF, J. Chem. Sot., B, (1970) 1646-1648. CH. DEGANI, Carbohydr. Res., IS (1971) 329-332. J. ‘D~AMSEN,Acta Chem. Stand., 6 (1952) 270-284. L. MICHAELISAND P. RONA, Biochem. Z., 49 (1913) 232-248. G. DE WIT, A. P. G. -BOOM, AND H. VAN BEKKUM, Carbohydr. Res., 74 (1979) 157-175. K. DENBIGH,The Principles of Chemical Equilibrium, Cambridge University Press, 1971. G. DE WIT, Thesis, Delft University of Technology, 1979. K. SCHWABE,Croat. Chem. Acta, 44 (1972) 127-137. 0. D. BONNERAND W. H. BREAZEALE, 1. Chern. Eng. Data, 10 (1965) 32S327. R. A, ROB~SON AND R. H. STOKES, Electrolyte Sol~o~* 2nd edn., Butte~orths, London, 1959. J. A. RENDLEMAN,JR., Adv. Chem. Ser., 117 (1973) 51-69. C. R. ALLEN AND P. C. WRIGHT, J. Chem. Educ., 41 (1964) 251-253. R. W. GURNEY, Ionic Processes in Solution, McGraw-Hill, New York, 1953. E. J. KING, International Enyclopedia of Physical Chemistry and Chemical Phystcs, Topic 15. Equilibrium Properties of Electrolyte Solutions, Vol. 4, Pergamon, Oxford, 1965, p. 4. H. HONIGAND H. WEIDMANN,Carbohydr. Res., 73 (1979) 260-266. H. SCHMID, ~onatsh. Chem., 100 (1969) 1654-1661. H. LISCHKA,Theor. Chim. Acta, 31(1973) 39-48. G. YAGIL AND M. ANBAR, J. Am. Chem. Sot., 85 (1963) 2376-2380. M. ANBAR, M. BOBTELSKY,D. SAMNEL, B. SILVER,AND G. YAGIL, J. Am. Chem. Sot., 85 (1963) 2380-2384. G. YAGIL, J. Phys. Chem., 71 (1967) 1045-1052. G. YAGIL, J. Phys. Chem., 73 (1968) 1610-1612. G. YAGIL, IF. J. Chem., 9 (1971) 329-350. G. YAGU, J. Chem. Sot., Chem. Commun., (1977) 672-673. Tabellenboekje ten dienste van laboratoria, K.N.C.V., Centen, Netherlands, 1962. H. Swrro, J. Am. Chem. Sot., 80 (1958) 70-73. J. M. HARVEY, M. C. R. SYMONS,AND J. RICHARD,Nature (London), 261 (1976) 435-436. J. A. RENDLEMAN,JR., Adv. Chem. Ser., 117 (1973) 51-69. J. M. Los AND L. B. SIMPSON,Reel. Trav. Chim. Pays-Bas, 76 (1957) 267-285. M. HANACK, Confo~afio~ Theory, Academic Press, New York, 1%5. F. FRANKS, D. S. REID, AND A. SUGGET~,J. So&ion Chem., 2 (1973) 99-118. T. YASJNAGA, I. Usur, K. IWATO, AND M. MIURA, Bull. Chem. Sot. Jpn., 37 (1964) 1658-1669. M. J. TAIT, A. SUGGETT, F. FRANKS,S. ABLE~?~,AND P. A. QUINKENDEN,J. Solution Chem., 1 (1972) 131-151. A. SUGGE-IT,J. Solution Chem., 5 (1976) 33-47. Y. MYAHARA, Kagaku, 29 (1959) 365.
IONISATION JOHN A.
W. M.
Laboratory Eindhoven
AND SOLVATION
OF D-GLUCOSE
BEENACKERS*, BEN F. M. KWXER, AND HESSEL S. VAN DER BAAN
of Chemical Technology, (The Netherlands)
Eindhoven
University
of Technology,
P.O.
Box 513, 5600 MB
(Received September Z&t, 1984; accepted for publication, January 28th, 1985)
ABSTRACT
For a quantitative description of chemical reactions of carbohydrates in concentrated solutions, a detailed knowledge of the ionisation equilibria is a prerequisite. In a series of experiments involving a wide range of concentrations of D-glucose, the ideal or non-ideal solution models did not accord with the observations unless hydration was taken into account. A hydration number of 3.5 was found for molecular n-glucose, but no hydration for dissociated D-@UCOSC. This result is discussed in terms of intramolecular hydrogen-bonding within the Dglucose ion. INTRODUCTION
This study is part of a project on the heterogeneous alkaline isomerisation of carbohydrates using ion-exchange resins l. In considering these reactions, the ionisation and solvation of the carbohydrates is of great importance. The ionisation of carbohydrates in alkaline aqueous solutions causes mutarotation, and, via carbanions (enolate ion), isomerisation and degradation. Reducing mono- and oligo-saccharides are weak acids and the ionisation of the anomeric hydroxyl group is an essential step in the isome~sation and epimerisation reactions. As ionisation is much faster than mutarotation2,3, the ionisation constants of a and p forms can be distinguished. Los and Simpson4 found a value of 0.29 for Apf& (= pKm_,,, - pKP_olc)for the pyranose forms, whereas De Wit et aI. found a value of 0.19. When only one pK, value is reported, it must be considered to be an overall ionisation constant and these have been determined for many carbohydrates at various temperatures6-25, using a variety of methods. Table I lists the pK, values at 298 K for D-glucose, D-fructose, and D-mannose. All concentrations are expressed in mol.mv3. The concentrations used for pH and pKs are expressed in kmol.mw3, so that pK and pH values can be compared with literature data. From the data in Table I, it can be concluded that, at 298 K, pKo,, = 12.4 40.25, plc,, = 12.1 f0.3, and pkl,, = 12.1 t0.1. *Present address: DSM Research and Patents, P.O. Box l&6160 MD Geleen, The Netherlands. OOOS-6215/85/$03.30
@ 1985 Elsevier Science Publishers B.V.
J. A. W. M BEENACKERS, B F M. KUSTER, H S VAN DER BAAN
170 TABLE
I
IVNfSATlON
CONSTANTS
fpf(,)
FOR AQUEOUS
SOLUTIONS
~-Glucose
D-Fructose
12.23 12.107 12.09 12.96 12.34 12.49” 12.2@ 12.87 12.35 12.38 12.46 12.51 12.72’ 12.35 12.78O.* 12.60b,d 13.9d,’
11.99 11.693 11.68
298 K D-Mannose
*p-Anomer.
Ref. 11 12 13 14 15 16
12.67 12.21
%-Anomer. TABLE
AT
17 18 19 20 21 22 72 5 5 5
12.13
12.27 12.31 12.553’
12.08
14.2d,’
14.0”’
cAt 283 K. dAt 376278
K.
3
II
LITERATURE
pl(,,,
VALUES
AS
A FUNC-I-ION
OF THE
CO~C~~TRA~O~
OF D-GLUCOSE.
AND
DETERMINED
PO~NTiOMETRlCALLY
Concentration (kmoI.m-3)
0.05 0.10 0.20 0.50 1.0
of
D-&cose
Michaefis and RonaLS
Thamse#
290-292 K
291 K
27.7 K
12.46 12.44 12.40
12.97 12.93 12.88
12.38 12.28 12.26 12.05
When the results of De Wit et aL5 are omitted, D-fructose seems to be more acidic than D-glUCOSe (Apfc, = 0.27 +O,lO). Izatt et ~1.~~ascribed the lack of agreement between the results of the various studies to the differences of the ionic strength of the solutions used. At 273 K, ThamsexP found a slight increase of pK, with increase in ionic strength. Degani 23, however, was unable to find any influence. Only three authors have described a dependence of the pK, on the concentration of hexose. Michaelis and Rona and ThamsenX4. using a potentiometric method, found that pKs decreased with increase in the concentration of D-glucose (Table II). The data of De Wit and co-workers5,2h, obtained using n.m.r. and U.V. techniques, reflected an increase of pKolC with increase in concentration of hexose. Also, in an alkaline ion-exchange resin, proton abstraction will take place
171
IONISATION AND SOLVATION OF D-GLUCOSE TABLE III EXPERIMENTAL
RESULTS
AND
CALCULATIONS
OF KGtal;
c,,
IS NOT IN THE TABLE
BECAUSE
Expt. C, x 1O-3 C, x 1O-3 pH
c,
-
c,
PKGW
PKG,,
(8)
(9)
(10)
0.447 3.301 8.774
0.632 0.571 0.584
12.200 12.243 12.226 12.234
0.117 1.000 3.162
0.454 3.462 8.980
0.612 0.602 0.603
12.213 12.221 12.218 12.220
10.04 1.234 11.00 9.986 11.20 14.331
0.110 1.000 1.58
1.124 8.986 12.746
0.604 0.611 0.595
12.219 12.214 12.220 12.226
0.158
10.03 1.221 11.00 9.782 11.20 14.231
0.107 1.000 1.585
1.114 8.782 12.646
0.612 0.597 0.590
12.213 12.224 12.222 12.210
0.371 0.352 0.329
0.351
10.00 2.828 10.50 7.757 10.80 13.772
0.100 0.316 0.631
2.728 7.441 13.141
0.741 0.683 0.659
12.130 12.166 12.159 12.181
6
0.371 0.353 0.329
0.351
10.01 2.737 10.50 7.584 10.80 13.830
0.102 0.316 0.631
2.635 7.268 13.199
0.699 0.665 0.662
12.156 12.177 12.171 12.179
7
0.690 0.673 0.609
0.657
9.81 3.945 10.11 7.573 10.50 16.442
0.065 0.129 0.316
3.880 7.444 16.126
0.876 0.868 0.860
12.057 12.061 12.061 12.065
8
0.711 0.701 0.626
0.679
9.90 5.037 10.01 6.314 10.50 16.361
0.079 0.102 0.306
4.958 6.212 16.045
0.884 0.874 0.932
12.054 12.059 12.064 12.080
9
1.194 1.106 1.007
1.102
9.53 4.479 10.00 11.448 10.30 19.366
0.034 0.100 0.199
4.445 11.348 19.167
10
1.144 1.108 1.047
1.100
9.53 4.403 10.00 11.281 10.20 16.175
0.034 0.100 0.168
4.369 11.181 16.017
1.131 1.019 0.980
11.946 11.992 11.982 12.009
11
1.526 1.496 1.384
1.469
9.30 4.040 9.50 5.966 9.90 12.986
0.020 0.032 0.079
4.020 5.934 12.907
1.324 1.259 1.185
11.878 11.900 11.901 11.926
12
1.524 1.493 1.384
1.467
9.33 4.207 9.50 6.162 9.90 12.998
0.021 0.032 0.079
4.186 6.130 12.919
no.
(mol.m-3) (mol.m-3)
(1)
(2)
(3)
(4)
1
0.0635 0.0611 0.0563
0.0602
10.05 0.559 11.00 4.301 11.50 11.936
0.112 1.000 3.162
2
0.0635 0.0610 0.0561
0.0602
10.97 0.571 11.00 4.462 11.50 12.142
3
0.171 0.156 0.148
0.158
4
0.171 0.156 0.148
5
(5)
(6)
(7)
11.958 11.984 11.985 12.012
11.890 11.885 11.900 11.926
172
.I.A. W. M. BEENACKERS. B. F. M KUSTER. H. S. VAN DER BAAN
KOH-burette u
_----------_
magnetic
pH control
ler
recorder reactor
magnetic stirrer
Fig 1. Reactor for potentiometric
titrations.
before isomerisation. Inside such a catalyst, the concentration of the components is relatively high. In view of the discrepancies associated with the literature data, the degree of ionisation at high concentrations of hexose has been determined. EXPERIMENTAL
The ionisation measurements were carried out by potentiometric titration using analytical grade chemicals, doubly distilled, CO,-free water, and a thermostated reactor (150 mL) provided with a magnetic stirrer (Fig. 1). The pH was measured with a glass electrode (Radiometer, type GK 2401 B) in combination with a pH controller (Radiometer, type TIT lc), and corrections were applied for the temperature and the concentration of the alkali. The electrode was calibrated with buffer solutions (Merck Titrisol). The titration was controlled by titrating pure water with CO,-free M KOH. When >5 mL of KOH solution was added, the measured and the calculated concentration of HO- differed by <5%. The isomerisation of the carbohydrate under study was neglected. A solution (50 mL) of D-glucose of known concentration under nitrogen was titrated to three different pH values at 298 K. Columns (l)-(5) of Table III show the final composition of the D-glucose solutions after adding the appropriate amount of alkali. RESULTS AND DISCUSSION
In solution, the chemical potential of each component27 is p, =
~7 + RT ln (Y, . C,),
and for the solvent.
(1)
IONISATION
pH,O
AND SOLVATION
=
&,O
+
RT
173
OF D-GLUCOSE
In bH,O
(4
. ch2O)y
where p&O is the chemical potential of pure water, and consequently the concentration of water is expressed as a fraction: cil,
=
(3)
~,dCH,O,pure.
As the experiments were carried out at constant pressure, PT is a function only of the temperature. For non-ideal solutions, y, + 1. When a solution is in equilibrium, the chemical potential (G) will be a minimum and AG will be zero. Hence, XV, . pi eq = .Z{v1 * ~7) + RT S{ V, * In (yi . CJ}eq = 0,
(4)
AG* = S{ q * /.~..f}= - RT ~{ Vi. In (yi * CJ}eq, and
(-5)
AGE = - RT~{v~ . In yi}.
(6)
As p: is a function only of the temperature, AG* will be also a function only of the temperature. The equilibrium constant is defined as In K, = Z{vi . In (y, . C,)} = - AG*/RT.
(7)
Because the pressure remains constant, the equilibrium constant should be only a function of the temperature. Applied to a solution of D-glucose (G) and including hydration of all species:
GH * (H20)h,,
+ q H,O%
G-
(HzO),,,_ + H+ . WA%,+
(8)
K
;H+
~40
- (H,O) ,,“+ + HO-
(HzO)hHO
(9)
where h = hydration number, q = ho- i-h,+ - hoH, and
(10)
p = h,+ + h,,-
(11)
KG = yG- ’ %
K H,O
=
YH+ . cH+
+ 1. ’
YH+ * cH+b3H
’ YHO-
’ %@I,0
’ %H
* h!,O
’ %,O>
’ ci20>
(12)
(13)
174
J. A. W M. BEENACKERS,
B. F. M. KUSTER,
H S. VAN DER BAAN
In the following sections, the equilibrium constant will be calculated on the basis of three assumptions: (a) ideal solution, no hydration, (b) non-ideal solution, no hydration, (c) non-ideal solution, hydration. In an ideal solution, the dissociation constants for n-glucose and water (Eqs. 22 and 13) are simplified to: K ?G) = Co- . Cn+/Con, and
(14)
K H,O(d= c,+ . q-I,-,
(15)
where the index (a) refers to the assumption (a). During titration, centration of o-glucose decreases and the following relations hold: C H+(aj= 103-pH CHO-(a)
CG-(a)
C GH(a)
=
mol.mm3
KHzO(&H+(~)
=
CK+(a)
-
CG(a)
the total con-
+
cH+(,)
-
CG-(a)
(16)
(PKH&l,298
-
C”O-(a)
=
13.9965)
(17)
(electro-neutrality)
(18)
(o-glucose balance)
(19)
In Table III, columns (6) and (7) show the results of the calculations of C,,and Co-. Furthermore, the values of Kc0 and pKo,,, according to Eq. 13 are
t pKG
12.5-
24
Q
Thamsen
m
MichaelIs
0
Present
and Rona
25
work
12 0\
I
0
0
500
1 1500
I
1000 c,---
Fig. 2. pK, as a function of the concentration the data of Thamsen” were used to recalculate
of D-glucose at 298 K. The temperature dependence them to the reference temperature of 298 K.
of
175
IONISATION AND SOLVATION OF D-GLUCOSE
presented in columns (8)-(10). 0 ur results, in combination with the potentiometric data of Thamsen24 and of Michaelis and Rona25, are given in Fig. 2. Our data give roughly the same concentration-p& relation as found by the other authors. However, it is clear that this approach does not lead to an equilibrium constant that is independent of the concentration. For a non-ideal solution without hydration, Eqs. 12 and 13 are simplified to: K G(b)
=
K H,O(b)
YG- . CG-. YH+ . &+/(YGH =
YH+ . cH+
* G.J, and
* YHO~ ’ cHO-/(YH20
’ CH,O),
(20)
(21)
with aHO 2 2 = yHO. 1 CHo. It is difficult to calculate the thermodynamic activity of water in a multi-component system29. For our experimental conditions (Table III), the concentration of Dglucose Co, is much higher than the ionic concentrations. For this reason, the activity of water was provisionally assumed to be equal to aHsOin a pure solution of D-glucose. The latter can be calculated from the measurements of Bonner and Breazeale30, who gave the activity coefficient of o-glucose (eon) and the osmotic coefficient (+oH) as a function of the molality (moH) of D-glucose in a neutral solution: yGH
=
1 + 0.022mbg
+on = 1 + O.O12m&$
(O <
IllGH <
2,
(23)
(O <
IllGH <
2,
(24)
The activity of water can be calculated from ref. 31 as In aHI = - mGH . Mqo * +GH . 10W3
(0 c mGH c 2).
(25)
The activity coefficients yGmand -yuo- have been calculated with the Debye-Htickel expression, as corrected by Robinson and Stokes31 for solvation and the activity of the solvent: log y = - Ar - I”*/(1 + B, * d, - I’“) - log (1 - 0.018 hi . mi) - h, . log aHZO,(26) with A,, K = 0.5115 kg’“.mol-ln (Debye-Hiickel constant), B 2ssK = 3.291 lo9 kgtn.molvz. m-l (Debye-Hiickel constant), d,- = diameter G- - 8 x lo-i0 m, d,,- = diameter HO- - 2 x lo-lo m, hi = solvation number: mol H,O per mol i, I = mxf + mH+ = mG- + mHO- = ionic strength (mol.kg-i).
J
176
A. W. M. BEENACKERS.
B. F M. KUSTER.
H. S VAN DER BAAN
The activity of water is calculated from yen, yo- , and yno- with the Gibbs-Duhem equation 26. It appeared that aHzOcalculated with Eq. 25 is a good approximation. The thermodynamic quantities are between the following limits. 1.000 < Yen < 1.045
(27)
0.882 < -yo- < 1.000
(28)
0.853 < -yHo- < 1.000
(29)
0.967 < aHto < 1.000
(30)
To calculate the ionisation applied, namely
constant,
a molality-molarity
m, = C,l@ - MO, . Co, - Mom .
C,-
- Mu,-
conversion
. C,,m),
has to be
(34
and for the density of the solution, p = 1000 + 0.067co.
(32)
In Fig. 3, the equilibrium constant pKo,,, is given as a function of the concentration of D-glucose. It is seen that the concentration dependence of p&,,, is almost unchanged with respect to that of pKo(,). Apparently, it is impossible to 12.3 7
PKG
q,,g-
pKGfbl
=
12 20-
PKG(C)
=
12.23
022
x
10 -3 cG
‘0.02
C G-
Fig. 3. pKG as a function
of the concentration
of
D-ghCOSe
The
points
belong
to pK,,,,.
(C)
IONISATION
TABLE
AND SOLVATION
177
OF D-GLUCOSE
IV
HYDRATIONOF D-OLUC~SE(LITERATUREDATA)
---
Experimental method
h CH 267 K
278K
2.3 2.2 6 5
298 K 3.5 3.5 1.8
3.5 >lO 3.7 2.7 --
2.0
Ultrasonic46 Ultrasoniti* Dielectric relaxation53 170-N.m.r. relaxation53 Dielectric relaxationjl “0-N.m.r. relaxationSI Compressibility5’ “0-N.m.r. relaxation” Dielectric relaxation54 Freezing process5’ Activity method%
eliminate this concentration dependency in this way, even when using the best thermodynamic data from the literature. For a non-ideal solution with hydration, the literature data on the hydration of molecular D-ghCOSe are given in Table IV. At 298 K, it is seen that most authors have reported a hydration number of 3.5. For the hydration number of G-, no literature data are available. From the entropy change during ionisation, conclusions have been drawn28,32T33 about the hydration of GH and G-. The entropy change during ionisation in water is a result of (a) the change of the number of particles (from the point of view of statistical thermodynamics, an increase of the number of particles causes an increase in the entropy of the system; when the hydration of the species formed differs from that of the non-dissociated compound, hydration will have an influence on the total entropy change), (b) the increased ionic strength (ions give an increase of the electrostatic field in the solution; the solvent water is strongly polar, so that the water molecules will be hindered in their rotation”135, and this effect causes a decrease in entropy upon ionisation), (c) the intramolecular hydrogen-bonding (an increase will lead to a decrease of the entropy of the D-glucose molecule32,“). For ~-glucose in solution, an entropy change upon ionisation of - 110 J.mol-*.K-i is calculated14a20,22,49. Allen and Wright33 ascribed this negative entropy effect to a decrease in the number of particles by an increase in the hydration of D-glucose during ionisation. However, in our opinion, the entropy change upon ionisation cannot be explained only by assuming an increase of the hydration of D-glucose. The electrostatic field, combined with intramolecular hydrogen-bonding, must have a dominating effect. The stoichiometric coefficient p, as defined in Eqs. 9 and 11, is generally For the hydration of H+ and HO-, mostly 1 given as 2 in the literature 10-13.37,38.
J. A.
1’78
W. M. BEENACKERS,
B. F. M. KUSTER, 13 S VAN DER BAAN
and 0 are assumed37%X.Inside the ion-exchange resin, the concentration of SH, S-, and HO- can be very high. According to Schwabez9, it is impossible to determine activity coefficients at high concentrations of electrolyte. In seeking to describe the ionisation inside the resin, the literature information on the activity coefficients of the various components in our system was replaced for the simple assumption that the excess free energy AGE (Eq. 5) is zero and that further effects must be ascribed to hydration. For the water, the relative concentration (Ci;,O.free)of hydration water is not taken into account. This approach was also used by others19-@. Eqs. 12 and 13 then are transformed into:
with Cr!ro,, : .
= 1 - (6.28 Co + ho, - C,,
in Eq. 35, the totaf relative water concentration
cNzO,total
=
+ ho- + C,-)/55508. was used:
1 - 6.28 C&55508 (0 c Co G 2000)
as calculated directly from literature dataJ*. To obtain a concentration-independent defined (Eq. 37):
(36)
Koc,), an optimisation criterion 6 was
to be calculated from all 36 experiments of Table III. Minimising this criterion, with the requirement that h, > 0, yields as optimal data:
hGH=
3.5 and ho.- = 0 for q = 1 - ho, = -2.5.
The hydration number of 3.5 for molecular D-glucose agrees well with most of the literature data for this temperature (see Table IV). With these best estimates of the hydration numbers, the plc,,,, value can be plotted as a function of the concentration of o-glucose. In Fig. 3, it is seen that, in this way, a concentration-independent ionisation constant is obtained. It was noted above that hydration decreases with increase in temperature. Shiio46 described the “adsorption” of water per hydroxyl group with a Langmuir adsorption equation: In [hon/no,l(l
- honfnon)] = AH/CRT) +
cl,
(38)
IONISATION
TABLE
AND SOLVATION
179
OF D-GLUCOSE
V
HYDRATION
OF D-GLUCOSE
Temperature (W
293 298 308 318
AS A FUNCIION
OF THE TEhfPERATlJRE8
Experimental data of ShiioM
Calculated data
h G”,exo
h FH
4.2 3.5 2.8 2.2
4.05 3.12 2.92 2.07
Langmuir adsorption
LinearlFreundlich s hvd 1.28 3.93 1.74 3.47
h GH
s hvd
4.12 3.60 2.79 2.19
0.40 0.80 0.02 0.01 1.23
10.42 AH (kJ.mol-I) “Including
- 20
- 55
the data of Shiio&.
where non = number of hydroxyl groups (for o-glucose, non = 5) and c1 = a constant. For o-glucose, a heat of hydration of -55 kJ.mol-i was fourid. According to this model, the hydration of o-glucose cannot exceed the number of hydroxyl groups. At low temperature, however, Harvey et aL4’ found a hydration number of at least 10. For this reason, the experimental data of Shiio46 were recalculated with a linear and with a Freundlich adsorption model. The temperature dependence of those two models can be described with Eq. 39. In bon = - AH/(RT) + c2
(Linear/Freundlich)
(39)
The results are given in Table V. In columns (3) and (5), the hydration numbers calculated using Eqs. 38 and 39, respectively, using best estimates for AH and Ciare given: S,,,, is defined as
hyd= (h,, - ho~,exp)~~m,exp x loam-
(40)
It is clear that the linear/Freundlich adsorption models give a better fit for the description of the hydration of D-glucose. The corresponding heat of “adsorption” is then calculated to be -20 kJ.mol-‘. When this temperature dependence was applied to our experimental results, the hydration numbers given in Table VI were TABLE
VI
HYDRATION
T WI
h GH
OF GH AS A FUNCTION
267 9.2
218 6.4
OF THE TEMPERATURE
298 3.6
303 3.2
313 2.4
323 1.9
333 1.5
J A. W. M BEENACKERS.
180
B. F. M. KUSTER, H S VAN DER BAAN
found. The low-temperature hydration number ho,,,, x is about the same as that found by Harvey et a1.47. The temperature dependency of the ionisation constant is given by De Wilt*’ as: Eion.G
=
-16.8 kJ.mol-l.
(41)
The best formulae for the ionisation function of the hydration number: K G(c) =
c,-~* C,,,+/[C,,
of D-glucose can thus be presented
as a
. C;I,O.free(‘-hq
KG(c)‘KHzo(c) = c,- *cEi,o.free (‘+hG”‘G3,. c,,-)
(43)
with ho, = 3.6 exp [2OOOO/R. (l/T
(44)
- l/298)], and
K occj = 0.583 exp [16800/R . (l/T
- l/298)].
(45)
DISCUSSION
It has been found that, during the ionisation of D-glucose, the hydration disappears. This can be understood from the difference in conformation between the ionic and the molecular forms of D-glucose. In the D-glucose molecule, there are few if any intramolecular hydrogenbonds48. Proton abstraction from the anomeric hydroxyl group gives a negative charge on O-l. Delocalisation of this negative charge to C-2/6 by intramolecular hydrogen-bonding was suggested by Rendleman @. The conformation of D-glucose can then be considered as shown in Fig. 4 for the (Yform. All ions shown in Fig. 4 would exist in equilibrium with each other, with the C-l-anion preponderating. Due to intramolecular hydrogen-bonding, the free rotation of the hydroxyl groups decreases and they will be oriented preferentially in certain directions for optimal hydrogen-bonding. This effect contributes to the decrease in entropy of the Dglucose molecule on ionisation. The difference in entropy decrease between LY-and /3-D-glucopyranose during ionisation, measured by Los and Simpson4’, can be ascribed to a difference of the strength of ax,eq and eq,eq hydrogen bonds (see Fig. 5). In an ax,eq sequence of hydroxyl groups, e.g., HO-l ,2 in a-D-glucose, the system can easily adopt the geometry for efficient hydrogen bonding. The reverse is true for an eq,eq sequence of hydroxyl groups”. The hydration of sugars can be considered as hydration of the hydroxyl groups46,st ) involving hydrogen bonding as well as further hydration of hydrate water molecules. After ionisation of D-glucose, the hydroxyl groups are oriented and stabilised by intramolecular hydrogen-bonding. Consequently, no hydroxyl groups of D-glucose are then available for bonding to solvent water molecules.
IONISATlON AND SOLVATION OF D-GLUCOSE
H-O
/,’
'CH2
?’
.+--0,
181
CH:!
4===
,’
,,’
H-O,
CHz
Fig. 4. Intramolecular hydrogen-bonding in the cu-D-glucopyranose ion.
I’
a-~-Glucose dSI0,
=
non -110
P-o-Glucose
J mol-‘OK-’
dSI0”
=
,,+-0,
CHZ
ton -83
J
rid-
@ K-’
Fig. 5. Intramolecular hydrogen-bonding in the ions of a- and ED-glucopyranose.
LIST OF SYMBOLS
AT ; c; C&O C, di
E F, FH, Fru G, GH, Glc AG AG” AGE
Debye-Hiickel constant activity of component i Debye-Hiickel constant concentration of component i relative water concentration constant diameter of component i activation energy
m kJ.mol-’
D-fNCtOSe
D-glucose change of chemical potential difference of free energy of pure components excess free energy
kJ.mol-1 kJ.mol-1 kJ.mol-’
182
AH h, I K, M, MH, Man M, m, nGII
P PH PKS q R S SSH T y, 3:
J. A
W
M. BEENACKERS,
B F. M. KUSTER,
H S VAN DER BAAN
kJ.mol-i change of enthalpy hydration number of sugar i mol.kg-r ionic strength equilibrium constant D-mannose kg.moI1 mol weight of component i molkg-’ molality of component i number of hydroxyl groups of GH stoichiometric coefficient acidity: pH = 3 - log C,+ pK,:pK, = 3 - log KS stoi~hiometric coefficient J.mol-r.K-i gas constant sugar (S = SH + S-): G, F, M ionised sugar molecular sugar temperature activity coefficient of component i on molarity scale activity coefficient of component i on molality scale optimisation criterion chemical potential of component i in the solution J.mol-1 chemical potential of the pure component J.mol-r chemical potential of pure water J.mol-* stoichiometric coefficient liquid density kg.m-3 osmotic coefficient of component i
REFERENCES I J. A. W. M. BEENACKERS. Thesis, Elndhoven University of Technology, 2 C. S. HUDSON AND
3 4 5 6 7 8 9 10 11
12 13
J. K. DALE. 1. Am. Chem. Sot.,
1980.
39 (1917) 32@3-328. V. S. R. RAO AND J. F. FOSIXR. J. Phys. Chem., 69 (1965) 636-640. J. M. Los AND L. B. SIMPSON,Reck Trav. Chim. Pays-&~, 73 (19.54) 941-958. G. DE WIT, A. P. G. KIEBOOM. AND H. VAN BEKKUM, Reel. Trav. Chim. Pays-Bus, 98 (1979) 355361. K. SMOLENSKIAND W. KOZI.OWSKI, Rocr. Chem., 16 (1936) 270. K. SMOLENSKIAND W. KOZLOWSKI, Rocz. Chem.. 16 (1936) 281-287. B CAPON AND W. G. OVEREND, Adv. C~rbo~ydr. Chem., 15 (1960) 32. B. CASU, M. REGGIANI, G. G. GALLO, AND A. VIGERANI, ~e~ruhedron, 22 (1966) 3061-3083. V. YA LVGINA AND E. S. LYGIN. Izv. Vyssh. Uchebn. .&wed., Ashch. Tekhnol., 5 (1970) 34-36; Chem. Abstr., 74 (1971) 65,906. T MADSEN, 2. Phys. Chem., 36 (1901) 290. P. HIRSCH AND R. SCHLAGS, 2. Phys. Chem., Aba. A, 141 (1929) 387-412. F. URBAN AND P. A. SHAFFER, J. Bioi. Chem., 94 (1931-1932) 697-715.
IONISATION AND SOLVATION OF D-GLUCOSE 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
183
P. S~UCHAYAND R. SCHAAL, Bull. SOC. Chim. Fr., (1950) 819-824. G. C. TILDE AND W. F. K. Wmxn-JohEs, Trans. Faraday Sot., 49 (1953) 243-251. J. M. Los AND L. B. SIMPSON,Reel. Trav. Cfiim. Pays-Bas, 73 (1954) 941-958. N. A. RAMAIAH AND S. S. KATIVAR, Proc. Annu. Conv. Sugar Technol. Assoc. India, 29 (1961) 77-84. G. GUILLOT AND P. RUMPF, C. R. Acad. Sci., 259 (1964) 4064. C. A. BL~TON AND H. CHAIMORICH,J. Am. Chem. Sot., 88 (1966) 4082-4089. R. M. IzA~, J. H. RY~~ING, L. D. HANSEN.AND J. J. CHRISTENSEN,J. Am. Chem. Sot., 88 (1966) 2641-2645. H. G. J. DE WILT, Thesis, Eindhoven University of Technology, 1969. J. J. CHRISTENSEN, J. H. RWING. AND R. M. IZATF, J. Chem. Sot., B, (1970) 1646-1648. CH. DEGANI, Carbohydr. Res., IS (1971) 329-332. J. ‘D~AMSEN,Acta Chem. Stand., 6 (1952) 270-284. L. MICHAELISAND P. RONA, Biochem. Z., 49 (1913) 232-248. G. DE WIT, A. P. G. -BOOM, AND H. VAN BEKKUM, Carbohydr. Res., 74 (1979) 157-175. K. DENBIGH,The Principles of Chemical Equilibrium, Cambridge University Press, 1971. G. DE WIT, Thesis, Delft University of Technology, 1979. K. SCHWABE,Croat. Chem. Acta, 44 (1972) 127-137. 0. D. BONNERAND W. H. BREAZEALE, 1. Chern. Eng. Data, 10 (1965) 32S327. R. A, ROB~SON AND R. H. STOKES, Electrolyte Sol~o~* 2nd edn., Butte~orths, London, 1959. J. A. RENDLEMAN,JR., Adv. Chem. Ser., 117 (1973) 51-69. C. R. ALLEN AND P. C. WRIGHT, J. Chem. Educ., 41 (1964) 251-253. R. W. GURNEY, Ionic Processes in Solution, McGraw-Hill, New York, 1953. E. J. KING, International Enyclopedia of Physical Chemistry and Chemical Phystcs, Topic 15. Equilibrium Properties of Electrolyte Solutions, Vol. 4, Pergamon, Oxford, 1965, p. 4. H. HONIGAND H. WEIDMANN,Carbohydr. Res., 73 (1979) 260-266. H. SCHMID, ~onatsh. Chem., 100 (1969) 1654-1661. H. LISCHKA,Theor. Chim. Acta, 31(1973) 39-48. G. YAGIL AND M. ANBAR, J. Am. Chem. Sot., 85 (1963) 2376-2380. M. ANBAR, M. BOBTELSKY,D. SAMNEL, B. SILVER,AND G. YAGIL, J. Am. Chem. Sot., 85 (1963) 2380-2384. G. YAGIL, J. Phys. Chem., 71 (1967) 1045-1052. G. YAGIL, J. Phys. Chem., 73 (1968) 1610-1612. G. YAGIL, IF. J. Chem., 9 (1971) 329-350. G. YAGU, J. Chem. Sot., Chem. Commun., (1977) 672-673. Tabellenboekje ten dienste van laboratoria, K.N.C.V., Centen, Netherlands, 1962. H. Swrro, J. Am. Chem. Sot., 80 (1958) 70-73. J. M. HARVEY, M. C. R. SYMONS,AND J. RICHARD,Nature (London), 261 (1976) 435-436. J. A. RENDLEMAN,JR., Adv. Chem. Ser., 117 (1973) 51-69. J. M. Los AND L. B. SIMPSON,Reel. Trav. Chim. Pays-Bas, 76 (1957) 267-285. M. HANACK, Confo~afio~ Theory, Academic Press, New York, 1%5. F. FRANKS, D. S. REID, AND A. SUGGET~,J. So&ion Chem., 2 (1973) 99-118. T. YASJNAGA, I. Usur, K. IWATO, AND M. MIURA, Bull. Chem. Sot. Jpn., 37 (1964) 1658-1669. M. J. TAIT, A. SUGGETT, F. FRANKS,S. ABLE~?~,AND P. A. QUINKENDEN,J. Solution Chem., 1 (1972) 131-151. A. SUGGE-IT,J. Solution Chem., 5 (1976) 33-47. Y. MYAHARA, Kagaku, 29 (1959) 365.