Determination of Proton Affinities and Acidity Constants of Sugars

Article pubs.acs.org/JPCA

Determination of Proton Affinities and Acidity Constants of Sugars Shuting Feng, Christina Bagia, and Giannis Mpourmpakis* Catalysis Center for Energy Innovation (CCEI) and Department of Chemical and Biomolecular Engineering, University of Delaware, Newark, Delaware 19716, United States S Supporting Information *

ABSTRACT: Proton transfer reactions play a key role in the conversion of biomass derived sugars to chemicals. In this study, we employ high level ab initio theoretical methods, in tandem with solvation effects to calculate the proton affinities (PA) and acidity constants (pKa) of various D-glucose and D-fructose tautomers (protonation−deprotonation processes). In addition, we compare the theoretically derived pH values of sugar solutions against experimentally measured pH values in our lab. Our results demonstrate that the protonation of any of the O atoms of the sugars is thermodynamically preferred without any significant variation in the PA values. Intramolecular hydrogen transfers, dehydration reactions, and ring-opening processes were observed, resulting from the protonation of specific hydroxyl groups on the sugars. Regarding the deprotonation processes (pKa), we found that the sugars’ anomeric hydroxyls exhibit the highest acidity. The theoretically calculated pH values of sugar solutions are in excellent agreement with experimental pH measurements at low sugar concentrations. At higher sugar concentrations the calculations predict less acidic solutions than the experiments. In this case, we expect the sugars to act as solvents increasing the proton solvation energy and the acidity of the solutions. We demonstrated through linear relationships that the pKa values are correlated with the relative stability of the conjugate bases. The latter is related to hydrogen bonding and polarization of the C−O− bond. A plausible explanation for the good performance of the direct method in calculating the pKa values of sugars can be the presence of intramolecular hydrogen bonds on the conjugate base. Both theory and experiments manifest that fructose is a stronger acid than glucose, which is of significant importance in self-catalyzed biomass-relevant dehydration reactions.

■

INTRODUCTION The continuous increase in energy demand associated with the decreasing fossil fuel reserves necessitates the adoption of alternative energy production and utilization methodologies. Furthermore, heavy dependence on imported traditional sources of energy, such as oil and natural gas, puts the energy and economic security of most developed countries at risk.1 Biomass appears to be one of the most significant, promising, and renewable sources of energy, since it can be converted to fuels and chemicals.2 One of the recently proposed routes is to produce hydrocarbon biofuels from woody-biomass sugars, such as glucose and fructose.3 According to U.S. Department of Agriculture statistics, the total annual production of sugar-based biomass in the United States, including plants, crops, and other forest products, is estimated to be about 800 million dry tons.4 The abundance of biomass in nature establishes it as an alternative supply of energy and chemicals of industrial importance.5 Glucose and fructose are representative biomass compounds of interest because they are polyols, abundant and structurally simpler than cellulose. Glucose, directly obtained from cellulosic biomass, can be reduced to sorbitol, which, in turn, can be converted to hexane through a series of catalyzed dehydration and hydrogenation reaction steps.6 As a result, hexane produced in this way can be used as a renewable © XXXX American Chemical Society

transportation fuel. In alternative chemical strategies, fructose can be converted to 2,5-dimethylfuran (DMF)7 and 5(hydroxymethyl)furfural (HMF).8 DMF has a higher energy density and boiling point than ethanol,7,9 which makes it a promising liquid biofuel. For the production of HMF,10 a “target molecule” with marked industrial importance, Brønsted-catalyzed dehydration of glucose or fructose is often employed. The dehydration mechanism has attracted significant attention and evolves through the protonation of one of the hydroxyl groups of the sugars.11 This is a very complicated process due to the polyfunctionality of carbohydrates. A Brønsted acid catalyst protonates a hydroxyl group (−OH is a poor leaving group) to a −H2O+ (better leaving group). The proton transfer process plays an important role in dehydration reactions. As a result, understanding proton transfer reactions in sugars is of marked significance in controlling yield and selectivity in biomass conversion reactions. Sugars are polyols, and similar to any other alcohol, they can be deprotonated and introduce acidity to the solution. The acidity constant (pKa) calculations have been extensively Received: April 4, 2013 Revised: May 22, 2013

A

dx.doi.org/10.1021/jp403355e | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

studied in the literature with electronic structure calculations.12 The applied strategies to calculate pKa values include (i) the direct or absolute method, (ii) the proton exchange or relative method, (iii) the hybrid cluster−continuum method, and (iv) the implicit−explicit method. The direct method involves a thermodynamic cycle13 (Scheme 1) which combines gas-phase

this, to validate our pKa results, we experimentally measure the pH values of the sugars under different solution concentrations. Our results demonstrate that the direct method can accurately calculate the pKa values of sugars and unravel the role of Hbonds in their acidities.

■

COMPUTATIONAL AND EXPERIMENTAL METHODS Ab Initio Calculations. We employed electronic structure calculations at various levels of theory, namely, Hartree−Fock (HF), density functional theory (DFT),22 MP2, CBS-QB3,23 and G424 using the Gaussian 09 computational package.25 All structures were fully optimized and the ground states were verified by the absence of any imaginary frequencies. For each different sugar tautomer, we accounted for structures with multiple orientations of hydrogen bonds. For the reaction where a proton is added to a hydroxyl group of an alcohol

Scheme 1. pKa Calculation Using the Direct Method

acidity calculations at a high level of theory with solvation calculations at a lower level of theory. Successful examples of this methodology have been demonstrated for pKa calculations on carboxylic acids14 and uracils.13 Despite the successes of the direct method, there are limitations originating from uncertainties in solvation calculations of ionic species, which are significantly larger than that of neutral species.15 On the other hand, the proton exchange method utilizes a reference acid in the thermodynamic cycle16 (Scheme 1S in Supporting Information), leading to the same number of ionic species on both sides of the reaction equation. As a result, this methodology allows for partial cancellation of errors associated with the solvation calculations of the ionic species.12 A drawback of this method, however, is that the reference acid needs to be chosen carefully to resemble the structure of the acid of interest and minimize errors in the solvation calculations. In addition, accurate pKa values of the chosen reference acid may not be always available, which further limits the generalization of this method.12 Alternative approaches are the cluster−continuum method (Scheme 2S in Supporting Information) and the implicit−explicit method (Scheme 3S in Supporting Information). Pliego and Riveros have used the cluster−continuum method in conjunction with the IPCM solvent model.17 The implicit−explicit method utilized by Kelly, Cramer, and Truhlar18 delivers considerably improved pKa results compared to the direct method. In this method, the pKa values are computed in an identical way as in the direct method, but an additional explicit water molecule interacts with the conjugate base forming an interactive complex. The utilization of the implicit−explicit method has been also recommended by Keith and Carter in the solvation energy calculations of radical species, since the explicit water molecule was found to stabilize the radical anion valence.19 Sugars appear in nature in different tautomers. On the basis of 1H and 13C NMR spectroscopic observations of D-fructose, this ketose consists of a mixture of β-pyranose, β-furanose, αfuranose, and α-pyranose tautomers (6:3:1:trace mixture) in water.20 On the other hand, D-glucose in water is a mixture of α-pyranose, β-pyranose, α-furanose, β-furanose, aldehyde form, and hydrated form tautomers in 37.63, 61.96, 0.108, 0.28, 0.0040, and 0.0059%, respectively, at 30 °C.21 In this paper, we employ highly accurate ab initio theoretical calculations of proton affinity (PA) and pKa values of D-glucose/fructose tautomers of interest. Since the amount of α-fructopyranose tautomer is negligibly small, the acyclic form (open chain) and the remaining three tautomers of D-fructose are accounted in this study. For glucose, we study the acyclic form, α-pyranose tautomer, and β-pyranose tautomer of D-glucose. On top of

ΔG °

ROH + H+ ⎯⎯⎯→ ROH 2+

(1)

the PA value is given by E PA = −ΔG°

(2)

On the other hand, for the deprotonation reaction, the change in standard Gibbs free energy is related to the acid dissociation constant, Ka: ΔGa °

HA ⎯⎯⎯⎯→ A− + H+

(3)

ΔGa° = −RT ln K a ≈ −2.303RT log K a

(4)

where Ka is defined by Ka =

[A−][H+] [HA]

(5)

and pKa is given by

pK a = −log K a

(6)

Therefore

pK a =

ΔGa° 2.303RT

(7)

In order to calculate ΔGa° in eq 7, the thermodynamic cycle13 shown in Scheme 1 is used. In Scheme 1 ΔGliq = ΔGa°; therefore, eq 7 can be written as

pK a =

ΔG liq 2.303RT

(8)

Since the direct calculation of ΔGliq is less accurate, we use the aforementioned thermodynamic cycle and ΔGliq can be expressed as ΔG liq = ΔGgas + ΔΔGsol = ΔGgas + ΔGs(H+) + ΔGs(A−) − ΔGs(HA)

(9)

It is worth noting that the gas-phase free energy calculations (Gibbs) use a reference state of 1 atm, whereas the solvation energy calculations use a reference state of 1 M. Therefore, we apply the following conversion to attain correct energy units: ΔGgas(1 M) = ΔGgas(1 atm) + RT ln(24.46)

(10)

From eqs 9 and 10, we derive the final formula for ΔGliq. B



■

ΔG liq = G(A−gas) + G(H+gas) − G(HA gas)

(11)

+

In eq 11, G(H gas) is calculated at the CBS-QB3 level to be −6.28 kcal/mol, and an experimental value of −264.61 kcal/ mol is used for ΔGs(H+).14 In order to calculate ΔGliq, we still need to compute the gas-phase Gibbs energies and the solvation energies of HA and A−. The structures are fully optimized and the Gibbs free energies of HA and A− in the gas phase are calculated at the CBS-QB3 level. The solvation effects are calculated at the HF/6-31+G(d) level using the CPCM solvation model.26 The solvation energy of HA and A− is the energy difference between their Gliq and Ggas values (both calculated at the HF/6-31+G(d) level). This methodology has been successfully applied by Liptak et al.14 to accurately calculate the pKa values of carboxylic acids. The deprotonation of the sugars introduces acidity in solution (increase of hydrogen cations). As a result we can assign pH values: pH = −log[H+]

(12)

Equation 13 is the well-known Henderson−Hasselbalch equation which describes the relationship between pH and pKa. pH = pK a + log

[A−] [HA]

(13)

Assuming that the concentrations of the dissociated A− and H+ ions in eq 5 are approximately the same, we attain the expression pH =

pK a − log[HA] 2

(14)

Experimental pH Determination. β-D-Fructopyranose (99%) and α-D-glucopyranose (99%) were purchased from Sigma Aldrich and used without further purification. The solutions were prepared by weighing different amounts of Dfructose and D-glucose and dissolving in volumetric flasks with distilled water. The concentration details of D-fructose and Dglucose solutions are presented in Table 1. Table 1. D-Fructose and D-Glucose Solution Details D-fructose

D-glucose

concn (g/L)

wt (g)

vol (mL)

wt (g)

vol (mL)

10.0 50.0 100.0 200.0 300.0 400.0

0.5 2.5 2.5 5.0 7.5 10.0

50 50 25 25 25 25

0.5 1.25 2.5 5.0 7.5 10.0

50 25 25 25 25 25

RESULTS AND DISCUSSION

First, we assess the performance of different ab initio methods in calculating the proton affinity (PA) values of simple alcohols, such as ethanol, 1-propanol, 2-propanol, and tert-butanol. These results are presented in Figure 1. Both CBS-QB3 and G4 levels of theory accurately reproduce the NIST Chemistry Webbook PA values.27 Since the CBS-QB3 method is computationally less expensive than the G4 method, it will be used from now on in the PA and pKa (protonation− deprotonation) calculations of the sugars. Previous work has employed a range of different theoretical methods, including CBS-QB3, G3, DFT, and CCSD(T), to calculate changes in enthalpy and free energy for a number of gas-phase deprotonation reactions, and the results were compared to experimental data.28 It was found that the CCSD(T) method was computationally most expensive and provided the most accurate results, whereas the CBS-QB3 method was considered the best when both accuracy and computational cost were taken into account. This finding reinforces the selection of the CBS-QB3 method in the PA and pKa calculations of the sugars. An important observation is that the PA values of the alcohols increase with their substitution degree, that is, PAtertiary > PAsecondary > PAprimary irrespective of the theory level. This is because the protonated alcohols can be considered as carbocations stabilized by the presence of water (CnH2n+1+(H2O)) and the carbocation stability increases with their substitution degree. A similar dependence has been experimentally observed by Gorte on Brønsted catalysts, such as zeolites,29 suggesting that the stability of the carbenium ion depends heavily on the substitution degree of the cation. In addition to Brønsted catalysts, the carbenium ion stability has been shown to play an important role in the Lewis-catalyzed alcohol dehydration reaction on metal oxides, such as γAl2O3.30 In Table 2 we present the calculated PA values for different tautomers of D-glucose and D-fructose using the CBS-QB3 method. Each tautomer consists of five hydroxyl groups and one carbonyl group. All six oxygen atoms of the molecules can be protonated. As a result, six different PA values are calculated and numbered. The detailed numbering of the oxygencontaining groups is presented in Figure 2. The calculated PA values of all different hydroxyl groups are presented in Table 2. We notice that there is a small variation in the PA values of the different hydroxyls, which is less than 10% of their total value. This observation agrees with the finding of Caratzoulas,31 who proposed that the protonation of any of the ring-hydroxyl groups of fructose is equally likely, using a QM/ MM type of calculations. The highest PA value for each tautomer is highlighted with an asterisk in Table 2. We were not able to calculate the PA values of all different hydroxyls, due to hydrogen-bond-mediated proton transfers. An example is the #1 hydroxyl of fructose. During geometry optimization the proton is transferred from the #1 hydroxyl to hydroxyls with higher PAs. This behavior is observed for every tautomer. All the observed intramolecular hydrogen transfers are shown in Table 2. We found that protonation of the #1 hydroxyl of αglucopyranose results in dehydration of the molecule (formation of H2O and significant elongation of C−O bond). Dehydration is also observed when protonating the tertiary hydroxyls of α-fructofuranose and β-fructopyranose. The tertiary hydroxyl (#2 hydroxyl in Figure 2d,e) for both fructopyranose and fructofuranose exhibits the highest PA,

+ RT ln(24.46) + ΔGs(H+) + ΔGs(A−) − ΔGs(HA)

Article

The pH values of the solutions were measured at ambient temperature using an epoxy body electrode (Cole-Parmer) attached to a Fisher Scientific Accumet Basic/AB15 pH meter. Before taking any measurements, the pH meter was calibrated using three standard buffers from Fisher Scientific at pH 4.00 (certified pH 3.99−4.01 at 25 °C), 7.00 (certified pH 6.99− 7.01 at 25 °C), and 10.00 (certified pH 9.98−10.02 at 25 °C). C



Article

Figure 1. Proton affinities (PAs) of alcohols calculated at various theoretical levels.

Table 2. Proton Affinities of D-Glucose and D-Fructose Using CBS-QB3 (Values in kJ/mol)a oxygen no. glucose

fructose

acyclic β-pyranose α-pyranose acyclic β-pyranose β-furanose α-furanose

1

2

3

4

5

6

ΔPA

782.60 734.14 762.69 1→2 1→6 1→5 1→6

746.10 754.71 756.74 817.17* 818.62* 816.29* 816.69*

794.45 760.46 750.53 3→5 776.09 3→2 3→2

797.31 741.21 802.78 4→2 787.04 731.96 736.81

809.73* 786.73* 740.57 776.68 763.26 803.88 785.94c

805.08 6→4 810.58* 775.33 792.14b 6→2 807.33

63.62 52.59 70.01 41.84 55.36 84.33 79.88

a

Asterisk denotes the O atom of the sugar with the highest affinity for a proton. bC(2)−O(6) bond breaks during the protonation process. cC(2)− O(5) bond elongates during the protonation process.

Figure 2. Numbering of the different functional groups of (a) acyclic D-glucose, (b) acyclic D-fructose, (c) β-glucopyranose, (d) β-fructopyranose, and (e) β-fructofuranose. For simplicity, the structures of α-glucopyranose and α-fructofuranose are not shown here.

PA values of the sugars are even higher than the PA of the tertiary butanol (alcohol with the highest PA in Figure 1). This

which agrees with the theoretically computed results of Assary et al. using the G4MP2 levels of theory.32 Note that the highest D



Article

hydroxyls gained a proton from an adjacent, more acidic hydroxyl. Our calculated pKa value for cyclic fructose is very close to the experimental value reported in the literature35 (e.g., experimental pKa of cyclic fructose = 12.27, calculated = 11.96). We found that the anomeric hydroxyls of both cyclic tautomers of D-glucose exhibit considerably higher acidity than other hydroxyls. This result agrees with Salpin and Tortajada’s theoretical study on D-glucose, in which they proposed that the anomeric hydroxyl is significantly more acidic in both α- and βglucopyranoses.36 We also observe a C(1)−O(5) bond elongation when deprotonating the anomeric hydroxyls of both D-glucose tautomers (i.e., for α-glucopyranose the C(1)− O(5) bond elongates from 1.41 to 1.55 Å) which could potentially result in a ring opening. Similar observation has been made by Mulroney et al. using the AM1 (semiempirical) and HF levels of theory.37 We observed the same behavior for cyclic D-fructose. The anomeric hydroxyl exhibits the lowest pK a in all different fructose tautomers, suggesting a thermodynamic preference of deprotonation over other sites. It has been suggested by Assary et al. that the deprotonation of the anomeric hydroxyl is also kinetically preferred.32 In Figures 3a and 4a, we present the calculated pH values for glucose and fructose in aqueous solution at different concentrations and compare them to the experimentally measured pH values obtained in our lab. Our calculated pH values are in excellent agreement with the experimental ones at low sugar concentration. We calculated the pH values of sugars using their lowest pKa values. Figures 3b and 4b show the pH curves that are produced by deprotonating the different hydroxyl groups of the most acidic tautomers of glucose and fructose (β-glucopyranose and β-fructopyranose). We found that the pH values of the most and least acidic hydroxyl groups on these tautomers differ by as much as three pH units. Most importantly, both theory and experiments manifest that fructose solutions are more acidic than glucose solutions. This observation is of marked importance in Brønsted acid selfcatalyzed dehydration reactions of sugars: Fructose is a stronger acid than glucose. Using the experimental pH values at different sugar concentrations and eqs 6, 8, 11, and 13, we can theoretically back calculate the solvation energy of proton, ΔGs(H+), in sugar solution. Although we make the coarse assumption that the proton is solvated in water (infinite solvation), even at concentrated solutions (40 g/100 mL), the ΔGs(H+) values presented in Table 5 show a trend with a physical meaning. The ΔGs(H+) increases (absolute number) with increasing sugar concentration, and it is always higher than the ΔGs(H+) in water (−264.61 kcal/mol from ref 14). This makes a great deal

is because the presence of multiple OH groups in sugars contributes to stabilizing the H+ through hydrogen bonds. Interestingly, a ring-opening process is observed when the ring oxygen of β-fructopyranose is protonated (C−O bond breaks between the #2 carbon and the ring oxygen). This ring-opening process has also been observed by Wlodarczyk et al. at the B3LYP/6-31+G* level.33 At this point, we turn to pKa calculations. The pKa is a measure of a sugar’s acidity, or equivalently, its ability to lose H+. In order to calculate the pKa values of the sugars, first, we test our approach by calculating the pKa of acetic acid. The Ggas in eq 11 is computed at the CBS-QB3 level, the same level which we calculated the PA values. The solvation effects can be calculated using various methods. To calculate the solvation energy, methods 1 and 3, as presented in Table 3, optimize the Table 3. pKa Calculation of Acetic Acid Using Different Methodologies and Comparison with Experimental Value acetic acid pKa = 4.75 (exptl) methodology 1. 2. 3. 4.

gas-HF opt; liquid-HF SPC gas-HF SPC; liquid-HF opt gas and liquid-HF opt gas and liquid-HF SPC

pKa

diff

7.09 7.19 6.16 7.21

2.34 2.44 1.41 2.46

gas-phase structure at the HF level, whereas methods 2 and 4 perform single point HF calculations (SPC) at the CBS-QB3 optimized structures. In the CPCM calculations, methods 2 and 3 optimize the structures at the HF level, while methods 1 and 4 perform single point HF calculations at the CBS-QB3 optimized structures. All the results are compared to an experimental pKa value of 4.75.34 As shown in Table 3, method 3 reproduces the most accurate results among all the different methods. As a result, for the pKa calculations of the sugars we applied method 3. It should be noted that the difference of 1.41 between the experimental value of acetic acid pKa and our calculated one corresponds to an energy difference of only 1.92 kcal/mol in the ΔG values of eq 11. In general, it should be noted that the pKa calculations are very sensitive to solvation effects as an energy change of approximately 1.36 kcal/mol will affect the pKa values by one unit. The calculated pKa values of all the hydroxyl groups of the sugars are presented in Table 4. The lowest pKa for each tautomer is noted with an asterisk. Some hydroxyl acidity values could not be calculated because the conjugate bases were not stable. When these hydroxyls were deprotonated, intramolecular hydrogen transfer occurred. As a result, these Table 4. Calculated pKa Values of D-Glucose/Fructosea,b

oxygen no. glucose

fructose

acyclic β-pyranose α-pyranose acyclic β-pyranose β-furanose α-furanose

1

2

3

4

5

6

ΔpKa

CO 14.08* 14.30* 16.14* H-trans H-trans 16.53

H-trans 18.92 18.12 CO 11.96* 12.43* 13.71*

14.11* 18.11 18.26 17.16 15.87 H-trans 15.86

19.22 18.06 15.87 19.19 17.26 18.73 18.15

H-trans ring O ring O 17.16 16.49 ring O ring O

H-trans 20.65 15.86 H-trans ring O H-trans H-trans

5.11 6.58 9.54 3.05 2.85 3.11 9.88

a

(ΔpKa is the difference in the pKa values between the various OH groups on each sugar). bH-trans denotes proton transfer from a neighboring OH group to the deprotonated site. Carbonyl and ring oxygens do not have any hydrogens to contribute to acidity. E



Article

Figure 3. (a) Theoretical and experimental pH values of D-glucose solution and (b) calculated pH values from deprotonation of different OH groups of D-β-glucopyranose solutions.

Figure 4. (a) Theoretical and experimental pH values of D-fructose solution and (b) calculated pH values from deprotonation of different OH groups of D-β-fructopyranose solutions.

Table 5. Calculated Proton Solvation Energies in Sugar Solution (Values in kcal/mol) from Experimentally Measured pH Values Making the Assumption That the [H+] Is Solvated in H2O concn (g/100 mL) D-glucose D-fructose

1.0

5.0

10.0

20.0

30.0

40.0

−265.09 −265.22

−265.23 −267.85

−265.99 −268.53

−267.14 −269.32

−267.75 −269.66

−268.67 −269.35

becomes more stable. A similar linear relationship between the pKa values of the alcohols and their calculated deprotonation enthalpies (ΔH°acid) has been reported by Tian et al.38 The stability of the conjugate bases, in turn, depends partially on the intramolecular hydrogen bond network. We calculated the stability of each conjugate base by generating structures with different hydrogen bond orientations. We found that the most acidic conjugate base tends to maximize the intramolecular hydrogen bonds. From the geometries of the conjugate bases of β-glucopyranose (Figure 5b) and β-fructopyranose (Figure 5c), we observe multiple hydrogen bonds directed to the negatively charged, deprotonated centers. This hydrogen bonding partially decreases the localization of the charge and increases the stability of the conjugate base, resulting in greater acidity (lowest pKa values of both sugars in Figure 5a). Similarly, an inverse proportionality between the number of hydrogen bonds and the pKa values of a series of alcohols has been reported in the literature.38 An additional reason that affects the stability of

of sense. By comparing the PA values of sugars (see Table 2, range 734.14−818.62 kJ/mol) with that of water, which is calculated to be 657.78 kJ/mol, we observe that the sugars have a greater affinity for H+. As a result, in high sugar concentrations, these polyols act as solvents, increasing the solvation energy of the protons. We believe that this is the reason why our calculated theoretical pH values are in excellent agreement with the experimental results at low concentrations, but deviate at higher sugar concentrations. Finally, in Figure 5a we plot the theoretically calculated pKa values of the different hydroxyls of β-glucopyranose and βfructopyranose, against the relative energies of the deprotonated species (conjugate base). The conjugate-base structures of β-glucopyranose and β-fructopyranose with the lowest energies are set to zero in Figure 5a. As expected, the pKa values of the sugars are linearly proportional to the ΔE of the conjugate bases. This trend demonstrates that the sugar pKa decreases (the sugar becomes more acidic) as its conjugate base F



Article

Figure 5. (a) Calculated pKa values for β-glucopyranose and β-fructopyranose as a function of the relative stability of the conjugate bases and optimized geometries of (b) β-glucopyranose and (c) β-fructopyranose conjugate bases. Hydrogen bonds are indicated with dotted lines, whereas the elongated C−C bonds are indicated in green. Charge density distribution on two different conjugate bases of β-fructopyranose, exhibiting the (d) lowest and the (e) highest pKa values. The most negatively charged atom is indicated in red, whereas the most positively charged is in green (NBO analysis).

bonds are expected to change the solvation energy of a proton. This effect can only be captured by molecular dynamics simulations. Both theory and experiments demonstrated that fructose solutions are more acidic than glucose solutions. This observation is very important in Brønsted acid self-catalyzed dehydration reactions, where fructose can act as a stronger acid than glucose. Finally, we found that the pKa values derived by deprotonating the different hydroxyls of the sugars are linearly proportional to the relative energies of their conjugate bases, which in turn, are related to the number of hydrogen bonds formed and the polarization of the C−O− bond. From a theoretical point of view, we believe that part of the success of the direct method in calculating the pKa values of the sugars can be attributed to the intramolecular hydrogen bonds from −OH groups that stabilize the conjugate base. These hydroxyl groups intrinsically act as explicit solvent molecular groups and result in decreasing the errors in the solvation energy (similar to the implicit−explicit method).

the conjugate bases is the polarizability of the molecule and specifically of the C−O− bond. In Figure 5d,e we plot the charge density distribution (natural bond orbital analysis) on the conjugate bases that are related to the lowest and highest pKa values for β-fructopyranose. As can be noted, the βfructopyranose conjugate base of Figure 5d (lowest pKa) exhibits higher polarization on the C−O− bond due to the more electropositive character of the C atom, when compared to the C−O− bond of the structure in Figure 5e (highest pKa). This results to increased electrostatic interactions in the vicinity of the deprotonated center (O−).

■

CONCLUSIONS Using a highly accurate ab initio level of theory, we calculated the proton affinity (PA) and pKa values of different tautomers of D-glucose (acyclic, β-pyranose, and α-pyranose) and Dfructose (acyclic, β-pyranose, β-furanose, and α-furanose), considering every different protonation−deprotonation site. Our results show that there is a variation of less than 10% in the PA values of the different sites on the sugars. This suggests that protonation of any of the oxygen atoms of the sugars is equally likely, even though hydrogen-bond-mediated proton transfers can happen from one protonation site to another. The existence of multiple intramolecular hydrogen bonds on the sugars contributes to higher PA values, when compared to the PA values of simple alcohols. Interestingly, protonation of the #1 hydroxyl of α-glucopyranose, as well as of the tertiary hydroxyls of α-fructofuranose and β-fructopyranose, leads to dehydration of the sugars. Moreover, when the ring oxygen of β-fructopyranose is protonated, the ring opens. By combining highly accurate ab initio calculations with solvation effects, we investigated the pKa (deprotonation) of the sugars. The anomeric hydroxyls of the sugars were found to exhibit the highest acidities. Using the lowest calculated pKa values, we calculated theoretical pH values of the sugars in water solutions. On top of this, we experimentally measured the pH values of sugar solutions in water. The theoretical results are in excellent agreement with the experimental measurements at low sugar concentrations, showing that our calculated pKa values are very accurate. At higher sugar concentrations, our experiments show more acidic solutions (lower pH) than the theoretically calculated values based on pKa. A plausible explanation is that, at high sugar concentrations, intermolecular hydrogen

■

ASSOCIATED CONTENT

S Supporting Information *

pKa calculation schemes with the proton exchange, hybrid cluster-continuum, and implicit-explicit methods. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

This work was supported as part of the Catalysis Center for Energy Innovation, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Award No. DE-SC0001004. The authors would like to thank Dr. Stavros Caratzoulas from the Catalysis Center for Energy Innovation at the University of Delaware for fruitful discussions. G



■

Article

Resonance. D-Fructose and D-Turanose. J. Am. Chem. Soc. 1971, 93 (11), 2779−2781. (21) Roslund, M. U.; Tahtinen, P.; Niemitz, M.; Sjoholm, R. Complete Assignments of the (1)H and (13)C Chemical Shifts and J(H,H) Coupling Constants in NMR Spectra of D-Glucopyranose and All D-Glucopyranosyl-D-glucopyranosides. Carbohydr. Res. 2008, 343 (1), 101−112. (22) (a) Becke, A. D. Density-Functional Thermochemistry. 3. The Role of Exact Exchange. J. Chem. Phys. 1993, 98 (7), 5648−5652. (b) Lee, C. T.; Yang, W. T.; Parr, R. G. Development of the ColleSalvetti Correlation-Energy Formula into a Functional of the ElectronDensity. Phys. Rev. B 1988, 37 (2), 785−789. (23) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. A Complete Basis set Model Chemistry. VI. Use of Density Functional Geometries and Frequencies. J. Chem. Phys. 1999, 110 (6), 2822−2827. (24) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. Gaussian-4 Theory. J. Chem. Phys. 2007, 126 (8), No. 084108. (25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009. (26) Barone, V.; Cossi, M. Quantum Calculation of Molecular Energies and Energy Gradients in Solution by a Conductor Solvent Model. J. Phys. Chem. A 1998, 102 (11), 1995−2001. (27) Linstrom, P. J.; Mallard, W. G. NIST Chemistry WebBook, NIST Standard Reference Database Number 69; National Institute of Standards and Technology: Gaithersburg, MD, June 2005. (28) (a) Liptak, M. D.; Shields, G. C. Comparison of Density Functional Theory Predictions of Gas-Phase Deprotonation Data. Int. J. Quantum Chem. 2005, 105 (6), 580−587. (b) Pickard, F. C.; Griffith, D. R.; Ferrara, S. J.; Liptak, M. D.; Kirschner, K. N.; Shields, G. C. CCSD(T), W1, and Other Model Chemistry Predictions for GasPhase Deprotonation Reactions. Int. J. Quantum Chem. 2006, 106 (15), 3122−3128. (c) Pokon, E. K.; Liptak, M. D.; Feldgus, S.; Shields, G. C. Comparison of CBS-QB3, CBS-APNO, and G3 Predictions of Gas Phase Deprotonation Data. J. Phys. Chem. A 2001, 105 (45), 10483−10487. (29) Gorte, R. J. What Do We Know about the Acidity of Solid Acids ? Catal. Lett. 1999, 62 (1), 1−13. (30) Roy, S.; Mpourmpakis, G.; Hong, D.-Y.; Vlachos, D. G.; Bhan, A.; Gorte, R. J. Mechanistic Study of Alcohol Dehydration on γ-Al2O3. ACS Catal. 2012, 2, 1846−1853. (31) Caratzoulas, S.; Vlachos, D. G. Converting Fructose to 5Hydroxymethylfurfural: A Quantum Mechanics/Molecular Mechanics Study of the Mechanism and Energetics. Carbohydr. Res. 2011, 346 (5), 664−672. (32) Assary, R. S.; Kim, T.; Low, J. J.; Greeley, J.; Curtiss, L. A. Glucose and Fructose to Platform Chemicals: Understanding the Thermodynamic Landscapes of Acid-Catalysed Reactions Using HighLevel ab Initio Methods. Phys. Chem. Chem. Phys. 2012, 14, 16603− 16611. (33) Wlodarczyk, P.; Paluch, M.; Hawelek, L.; Kaminski, K.; Pionteck, J. Studies on Mechanism of Reaction and Density Behavior during Anhydrous D-Fructose Mutarotation in the Supercooled Liquid State. J. Chem. Phys. 2011, 134 (17), No. 175102.

REFERENCES

(1) Vlachos, D. G.; Chen, J. G.; Gorte, R. J.; Huber, G. W.; Tsapatsis, M. Catalysis Center for Energy Innovation for Biomass Processing: Research Strategies and Goals. Catal. Lett. 2010, 140 (3−4), 77−84. (2) Mettler, M. S.; Vlachos, D. G.; Dauenhauer, P. J. Top Ten Fundamental Challenges of Biomass Pyrolysis for Biofuels. Energy Environ. Sci. 2012, 5 (7), 7797−7809. (3) Regalbuto, J. R. Cellulosic BiofuelsGot Gasoline? Science 2009, 325 (5942), 822−824. (4) Wool, R. P.; Sun, X. S. Overview of Plant Polymers: Resources, Demands, and Sustainability. Bio-Based Polymers and Composites; Elsevier: New York, 2005; p 2. (5) Roman-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Phase Modifiers Promote Efficient Production of Hydroxymethylfurfural from Fructose. Science 2006, 312 (5782), 1933−1937. (6) Huber, G. W.; Chheda, J. N.; Barrett, C. J.; Dumesic, J. A. Production of Liquid Alkanes by Aqueous-Phase Processing of Biomass-Derived Carbohydrates. Science 2005, 308 (5727), 1446− 1450. (7) Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Production of Dimethylfuran for Liquid Fuels from Biomass-Derived Carbohydrates. Nature 2007, 447 (7147), 982−985. (8) Nikbin, N.; Caratzoulas, S.; Vlachos, D. G. A First PrinciplesBased Microkinetic Model for the Conversion of Fructose to 5Hydroxymethylfurfural. ChemCatChem 2012, 4 (4), 504−511. (9) Choudhary, V.; Pinar, A. B.; Sandler, S. I.; Vlachos, D. G.; Lobo, R. F. Xylose Isomerization to Xylulose and Its Dehydration to Furfural in Aqueous Media. ACS Catal. 2011, 1 (12), 1724−1728. (10) Choudhary, V.; Burnett, R. I.; Vlachos, D. G.; Sandler, S. I. Dehydration of Glucose to 5-(Hydroxymethyl)furfural and Anhydroglucose: Thermodynamic Insights. J. Phys. Chem. C 2012, 116 (8), 5116−5120. (11) Guo, N.; Caratzoulas, S.; Doren, D. J.; Sandler, S. I.; Vlachos, D. G. A Perspective on the Modeling of Biomass Processing. Energy Environ. Sci. 2012, 5 (5), 6703−6716. (12) Ho, J.; Coote, M. L. A Universal Approach for Continuum Solvent pK(a) Calculations: Are We There Yet? Theor. Chem. Acc. 2010, 125 (1−2), 3−21. (13) Jang, Y. H.; Sowers, L. C.; Cagin, T.; Goddard, W. A. First Principles Calculation of pKa values for 5-Substituted Uracils. J. Phys. Chem. A 2001, 105 (1), 274−280. (14) Liptak, M. D.; Shields, G. C. Accurate pKa Calculations for Carboxylic Acids Using Complete Basis Set and Gaussian-n Models Combined with CPCM Continuum Solvation Methods. J. Am. Chem. Soc. 2001, 123 (30), 7314−7319. (15) Cramer, C. J.; Truhlar, D. G. A Universal Approach to Solvation Modeling. Acc. Chem. Res. 2008, 41 (6), 760−768. (16) Namazian, M.; Kalantary-Fotooh, F.; Noorbala, M. R.; Searles, D. J.; Coote, M. L. Møller-Plesset Perturbation Theory Calculations of the pKa Values for a Range of Carboxylic Acids. J. Mol. Struct: THEOCHEM 2006, 758 (2−3), 275−278. (17) Pliego, J. R., Jr.; Riveros, J. M. Theoretical Calculation of pKa Using the Cluster-Continuum Model. J. Phys. Chem. A 2002, 106 (32), 7434−7439. (18) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. Adding Explicit Solvent Molecules to Continuum Solvent Calculations for the Calculation of Aqueous Acid Dissociation Constants. J. Phys. Chem. A 2006, 110 (7), 2493−2499. (19) Keith, J. A.; Carter, E. A. Quantum Chemical Benchmarking, Validation, and Prediction of Acidity Constants for Substituted Pyridinium Ions and Pyridinyl Radicals. J. Chem. Theory Comput. 2012, 8 (9), 3187−3206. (20) (a) Dais, P.; Perlin, A. S. Stabilization of the β-Furanose Form, and Kinetics of the Tautomerization of D-Fructose in Dimethyl Sulfoxide. Carbohydr. Res. 1985, 136 (Feb), 215−223. (b) Allerhand, A.; Doddrell, D. Carbon-13 Fourier Transform Nuclear Magnetic Resonance. VII. Study of Anomeric Equilibria of Ketoses in Water by Natural-Abundance Carbon-13 Fourier Transform Nuclear Magnetic H



Article

(34) Sargeant, E. P.; Dempsey, B. Ionisation Constants of Organic Acids in Aqueous Solution; Pergamon Press: New York, 1979. (35) CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 2012. (36) Salpin, J. Y.; Tortajada, J. Gas-Phase Acidity of D-Glucose. A Density Functional Theory Study. J. Mass Spectrom. 2004, 39 (8), 930−941. (37) Mulroney, B.; Peel, J. B.; Traeger, J. C. Theoretical Study of Deprotonated Glucopyranosyl Disaccharide Fragmentation. J. Mass Spectrom. 1999, 34 (8), 856−871. (38) Tian, Z.; Fattahi, A.; Lis, L.; Kass, S. R. Single-Centered Hydrogen-Bonded Enhanced Acidity (SHEA) Acids: A New Class of Brønsted Acids. J. Am. Chem. Soc. 2009, 131 (46), 16984−16988.

I


Determination of Proton Affinities and Acidity Constants of Sugars

Recommend Documents