CHAPTER3

MOLECULAR STRUCTURE AS A DETERMINANT OF HYDROTROPIC ACTION: A STUDY OF

POLYHYDROXYBENZENES

CHAPTER3

MOLECULAR STRUCTURE AS A DETERMINANT OF HYDROTROPIC ACTION

A STUDY OF POLY HYDROXYBENZENES

3.0. INTRODUCTION:

One of the remarkable features displayed by hydro tropes is the ability of some of them

to solubilize hydrophobic substances selectively. In other words, not all hydrophobes are

solubilized by a given hydrotrope to the same extent. Also, not all hydrotropes solubilize a given

solubilizate to the same extent. Generally, the solubilizing efficiency of an amphiphile, be it a

surfactant or a hydrotrope, is dependent on the hydrophile-lipophile _balance that the molecule

possesses. Usually conventional surfactants are considered to be more hydrophobic by virtue of

the long hydrocarbon chains they have, in comparison to the aromatic hydrotropes, where the

planar benzene ring, which is equivalent to 3-4 carbons in a straight chain, constitutes the

hydrophobic part. Yet some hydrotrope aggregates differ from surfactant micelles in displaying

a higher and somewhat more selective ability to solubilize guest molecules. While this selectivity

in solubilization is not exhibited by all hydrotropes, this ability of some of these agents to do so

suggests that hydrotrope aggregates display self-association and host-guest interactions that are

somewhat different from those seen in aqueous micelles.

The importance of such structural features (with respect to their amphiphilic character)

in terms of solubilizing efficiencies was brought about by looking at the efficiency of a series of

hydrotropes towards the same solute (Yamamoto et al., 1955). The substance that was needed

to be solubilized in a stable and effective form was vitamin B2 (riboflavin) and a variety of

hydrotropes were tried. The results obtained from these studies are shown in the foll<;>wing table.

25

Table 1: Solubility of riboflavin in aqueous hydrotropes (Yamamoto et al., 1955)

CONCEN1RATION OF THE HYDROTROPE

HYDROTROPE 1% 2% 3% 5% 10%

--------------------------------------------------------------------------------------------------------------------------------------------------------

Na benzoate 0.38 0.48 0.64 1.10 3.10

Na salicylate 0.53 0.96 1.30 2.90 7.70

(2-hydroxy benzoate)

Na gallate 0.67 0.92 1.46 2.85 7.34

(3,4,5-trihydroxy benzoate)

Na 3-hydroxy-2-naphthaoate 5.60 12.3 19.6 34.3 97.8

N a 4-picolinate 0.37 0.50 0.66 1.10 2.40

(pyridine-4-carboxylate)

The solubility of riboflavin ·in water is negligible, whereas in 10% Na benzoate solution,

the value increases to 3.1 g I litre. The table also suggests that introduction of polar groups to Na

benzoate does not dramatically change the hydrotropic efficiency, while changing from the

benzene ring to the more hydrophobic naphthalene moiety leads to a remarkably enhanced

hydrotropic behaviour. Introduction of the more polarizable hetero atom nitrogen, which results

in the loss of planarity of the molecule, reduces the hydrotropy of 4-picolinate. In a similar vein

follows the studies on the solubility of the cytotoxic agent chartreusin using hydroxybenzoate

class of hydrotropes (Poochikian and Craddock, 1979). Again Na benzoate increases the

solubility of chartreusin in water. Amongst the isomeric monohydroxy benzoates, the m-isomer

appears to have maximum hydrotrope efficiency, compared to the o-, and p-isomers. Even

isomeric dihydroxybenzoates show differing hydrotropy towards chartreusin. The important point

is that such structure-based differences are not commonly encountered with micellar systems.

Thus, it would be important to understand the self-aggregation features of hydrotropes and the

26

role of the molecular structure in modulating the hydrotropy.

In an effort to understand the relation between hydrotropic behaviour and molecular·

structure, polyhydroxy benzene class of hydrotropes have been chosen. These compounds suit

the present study because a) many of them are efficient hydrotropes, b) several isomers of the

same formula are available for comparison enabling an assessment of molecular structural

features, and c) they are uncharged molecules, so that complications due to columbic factors are

avoided.·

3.1. MATERIALS AND METHODS:

The hydrotropes used in this study, namely catechol, resorcinol, hydroquinone (1,2-, 1,3-

1,4-dihydroxybenzenes respectively), pyrogallol, phloroglucinol (1 ,2,3-, 1 ,3,5-

trihydroxybenzenes respectively), and 4-methylcatechol, orcinol (1-methyl-3,5-

dihydroxybenzene), and 2-methylresorcinol were obtained from commercial sources in the

highest available pure form. Some of these compounds were purified by recrystallization,

wherever necessary. Since many of these hydrotropes are photosensitive, they were stored away

from light.

3.1.1. Solubilization Experiments: Fluorescein diacetate (FDA) was chosen again as the

representative solubilizate and followed the experimental protocol which was described in

chapter 2 under the materials and methods section.

3.1.2. Surface Tension Measurements: The procedure used for these measurements was the

same as described in chapter 2.

3.1.3. NMR Studies: NMR spectra were recorded on the Broker AM 300 and JEOL 400 MHZ

spectrometers. Proton chemical shifts of hydrotropes were measured in D20 solutions, using

HDO signal fixed at 4.80 ppm. Spin lattice relaxation times were measured using the inversion

recovery pulse sequence. For proton, the 90 degree pulse was of 4.3 JlS duration, while for 13C

27

it was 16 f.lS. T1 values were obtained after a nonlinear curve fit, and the accuracy of the values

so obtained was about 5%.

3.1.4. Electron Spin Resonance Studies: The spin probe in these studies is 5-doxyl stearate and

the procedure adopted for estimating the nitrogen hyperfine coupling constant and the

reorientational correlation time ('tc) of the spin probe in various systems was already described

in chapter 2 under the materials and methods section.

3.2. RESULTS AND DISCUSSION:

3.2.1.Hydroxybenzenes: The solubility of FDA in several hydro trope solutions as a function of

concentration is shown in Figure 3.1. The point that needs to be reemphasized here is that

hydrotropy is sigmoidal, and is operative in each case only beyond a characteristic concentration,

that is minimal hydrotrope concentration or MHC. From Figure 3.1 it is evident that the MHC

values of various dihydroxybenzenes differ as also their solubilization efficiencies. Considering

the amphiphilic character alone, catechol, resorcinol and hydroquinone should have displayed

comparable solubilizing efficiencies. The solubility of hydroquinone is very low (0.8 M)

compared to that of catechol which is 4M and to that of resorcinol which is 9 M. This low

solubility of hydroquinone is expected because of the negligible net dipole moment the molecule

has. Another difficulty that is associated with hydroquinone is that it is prone to easy oxidation

in air. Due to these difficulties not much work could be carried out with hydroquinone but the

preliminary experiments using FDA as the solubilizate revealed very little hydrotropy. Between

the two other dihydroxybenzenes, catechol is a better hydrotrope than resorcinol. Though their

polarities are about the same, catechol is expected to be more acidic than resorcinol owing to

intramolecular hydrogen bonding. It is evident from Figure 3.1, catechol is a more efficient

hydrotrope than resorcinol which in tum is better than hydroquinone.

Surface tension experiments of these hydroxybenzenes were carried out as a function of

28

4•5

4-0

3·!5

E c:: 2·!5 0

= ! 2·0

0 0

0+1

©t0+1 OH

2 ·~ OH

©ro~ 3

CH

OH

4 ~0+1 CH3

HO't9) 5 0

CH3

0 2 0.4 0·6 0·8 1-Q 1·2 1·4 1·6

(Hydrotrope] , !!! -1·8 2·0

Figure 3.1. Solubilization of fluorescein diacetate (FDA) by various hydrotropes at room temperature, as monitored by the absorbance of FDA at 480 nm as a function of increasing molarity of the hydrotrope in water.

their concentration and the results are shown in Figure 3.2. The curves for catechol and resorcinol

suggest that the surface active character that these molecular aggregates possess is about the

same. The break points in these surface tension curves indicate that both catechol and resorcinol

self-aggregate and form loose non-covalent assemblies beyond a given concentration. Thus the

MHC values for catechol (0.8 M) and resorcinol (0.7 M) are comparable. The difference in

solubilizing efficiency of catechol and resorcinol cannot be because of either the MHC values

or the amphiphilicity displayed by these molecules. Thus, the clue to their different hydrotropic

abilities should lie in their molecular structure, which would determine the packing and the

organization of the functional unit in each case, namely the hydrotropic aggregate. As discussed

in chapter 2, hydrotropy is a collective molecular effect (Balasubramanian and Friberg, 1993;

Balasubramanian et al., 1989). Saleh and El-Khordagui have proposed a model wherein they

suggest that aromatic hydrotropes might self-aggregate in a stack type fashion, where the planar

aromatic moiety is thought to aid in stacking while the hydrophilic part aids in high solubility in

water (Saleh and El-Khordagui, 1985). The results on hydroquinone, catechol and resorcinol

suggest that subtle differences in the structure might be reflected as differences in the packing

and organizational geometry·of the self-aggregates, which might account for the hydrotropic

differences amongst the three dihyhydroxybenzenes.

Support for this point of view comes from the results with pyrogallol. This molecule is

even more hydrophilic and somewhat less surface active (Figure 3.2) than catechol or resorcinol

(soluble upto 5M in ·water). Yet it solubilizes FDA with greater efficiency (see Figure 3.1). In this

context, it should be noted that the solubility of phloroglucinol (1,3,5-trihydroxybenzene) in

water is only 0.8 M and it is a poor hydrotrope. Molecular models reveal that it is possible to

organize a more compact arrangement of stacks of pyrogallol as compared with that of

phloroglucinol, and of catechol compared to that of resorcinol or hydroquinone.

3.2.2. Methylated hydroxybenzenes: Looking at the different hydrotropic efficiencies displayed

by various dihydroxybenzenes, it would be interesting to see the effect of introducing a relatively

29

E u ' Ill .. c: >- 58

0

c: 0 Ill c: ., ., u 0 -.. ~

rJ)

I_ 4- Mtlhyl Catechol

2 Ruorcinol

3. Catechol

4_ Pyrogallol

5. Orcinol

6. 2-Mtlhyl Ruorcmol

2

08 1-0 1·2 1·4 16

(Hydrotrop~], M-

4

6

20

Figure 3.2. Variation in the air-water surface tension with hyd.rotrope concentration at room temperature.

hydrophobic methyl group at different positions on the benzene ring of the hydroxybenzenes. To

address this point, a series of methylated hydroxybenzenes were chosen. The molecules in the

present study are 4-methylcatechol (4MC), 2-methylresorcinol (2MR), and orcinol (3,5-

dihydroxytoluene). The structures of these molecules are shown in Figure 3.1. Figure 3.2 reveals

that 2MR and orcinol are slightly more surface active than catechol, resorcinol and pyrogallol.

Interestingly, 4MC displays maximum surface activity and is in fact comparable in this feature

to Na cumenesulfonate (Balasubramanian et al., 1989) and some classical micellar surfactants.

Thus, introduction of methyl group enhances the amphiphilicity of the molecules and leads to a

reduction in surface tension.

Figure 3.1 shows that the introduction of the methyl group improves the hydrotropy of

orcinol vis-a-vis resorcinol, and in 4MC vis-a-vis catechol. Replacing one hydroxyl group of

phloroglucinol (1,3,5-trihydroxybenzene) by a methyl group leads to the far superior hydrotrope

orcinol (1,3-dihydroxy-5-methyl benzene). It is interesting to note that the solubility of orcinol

in water is several-fold higher than that of phloroglucinol, probably because of a larger dipole

moment. Similarly, 2MR appears to be a better hydrotrope than pyrogallol. The most important

observation that is shown in Figure 3.1 is that 4MC is the best hydrotrope in this family. The

effect of shifting the position of hydroxyl group, i.e. orcinol vis-a-vis 4MC, has led to I)

increased surface activity, ii) slight increase in MHC, and iii) substantially improved hydrotropic

behaviour. Thus amongst the compounds studied here, the hydrotropic efficiency appears to be

in the order: . .

4MC> pyrogallol> catechol> orcinol> resorcinol> phloroglucinol- hydroquinone.

4MC appears to be better than sodium butyl monoglycolsulfate (NaBMGS), the short

chain aliphatic anionic which is considered to be an excellent broad spectrum hydrotrope. To

date, the only hydrotrope that appears to be better than 4MC is Na xylenesulfonate, which is able

to solubilize almost twice the amount of FDA as 4MC is able to do. NaXS has been used by

several groups and found to be very good in terms of its solubilizing ability (Mckee, 1946; Booth

30

and Everson, 1948).

3.2.3. NMR Studies on 4-methyl catechol: In light of the excellent hydrotropy displayed by

4MC, it would be worthwhile to study some features of its self-aggregation in a more detailed

fashion.

Figure 3.3 shows the concentration dependant changes in the 1H NMR chemical shifts of

the ring protons as well as those of the methyl group of protons of 4MC in aqueous solution. It

is evident that sharp changes are obtained for each proton at about 0.4 M, which is. the

concentration beyond which hydrotropy is exhibited (that is its MHC value, see Figure 3.1 ), and

the surface activity of the molecule levels off (see Figure 3.2). This suggests that beyond 0.4 M

in water 4MC self-aggregates into a non-covalent assembly.

The microenvironmental features of the 4MC aggregate were monitored using both

intrinsic and extrinsic spectral probe methods. The fact that the chemical shift of the methyl

protons, shown in Figure 3.3, is upfield shifted beyond 0.4 M indicates a ring current induced

shielding of these protons, which could happen when the molecules aggregate one on top of

another. If it is so, then the downfield shift that is seen in the ring proton signals is also to be

expected (see Figure 3.3)

In order to look at the molecular motion of a given self-aggregate one could measure the

spin-lattice relaxation time of a designated proton. Figure 3.4 shows the spin-lattice relaxation

time values of the methyl protons as a function of concentration of 4MC as well as that of the

conventional surfactant, sodium octanoate. Sodium octanoate was chosen because of the

following reasons, namely a) its erne value is about 0.4 M, which is close to the MHC value of

4MC .• b) among the alkanoates, it appears to mark the transition from hydrotropes to micelle

forming agents (Danielsson and Stenius, 1971 ), and c) the micellar aggregates of octanoate are

thought to be small and not as densely packed as in the higher alkanoates.

Figure 3.4 shows that there is a two-fold reduction in the spin-lattice relaxation time (T1)

of the methyl protons of 4MC beyond its MHC value, reflecting on the fact that the molecular '

31

O OH

H*OH ""I CH " HA

aCH3

HD

0 A

J: -2 64

• HA

• -2 68 a:l

0 I

e a. a.

co

0 HB

0 0

01 02 o3 o• 05 o-6 01 oe og 10

[ 4- Mrthyl Catrchol], ~

Figure 3.3. Concentration-dependent changes in the chemical shift values of the various protons of 4- methyl catechol ( 4MC).

Figure 3.4. Concentration-dependent variation in the spin-lattice relaxation times {T1) of the methyl protons of the micellar surfactant sodium ·octanoate and of the hydrotrope 4MC in water, at room temperature, at an operating NMR frequency of 300 Mhz.

motion is restricted in the self-aggregate. Na octanoate also exhibits a similar behaviour but a

closer inspection of both the curves of 4MC and Na octanoate reveals some differences in their

T 1 behaviour in terms of their aggregation behaviour. With Na octanoate, the T 1 value drops from

about 2.7 s at 0.1 M to a levelled off value of about 1.5 s upon formation of the micellar

assembly. This behaviour is expected of a micellar amphiphile where aggregation is sharp and

thought of as pseudo-phase separation (Tanford, 1973; Lindmann and Wennerstorm, 1980). With

the hydrotrope 4MC, on the other hand, the T1 behaviour is different; it stays roughly the same

until the MHC is reached, and progressively decreases thereafter. Such behaviour is usually

attributed to a more gradual association, or even a step-wise aggregation process, which is likely

when intermolecular stacking occurs. The T 1 values of the methyl protons in 4MC are also lower

than those in the octanoate micelles, perhaps indicative of a greater degree of molecular packing

or restriction. Surfactant micelles are thought to possess a liquid-like interior (Lindmann and

Wennerstorm, 1980; Gruen, 1985), while the microenvironment in the molecular aggregates of

aromatic hydrotropes is still to be described with certainity.

3.2.4. ESR Studies on 4-methyl catechol: Electron spin resonance or ESR spectral method has

been used to monitor the microenvironmental polarity and microviscosity values of the

hydrotrope assemblies of 4MC.The spin probe, 5-doxylstearate, was solubilized in the 4MC

hydrotrope assemblies and the nitrogen hyperfine coupling constant (aN) was measured and the

reorientational correlation time ( 'tc) of the motion of the probe therein estimated. As mentioned

in the earlier chapter, the value of~ is sensitive to the polarity of the medium, while 't c increases

with the vicosity of the medium (Berliner, 1976). The aN value of 5-doxyl stearate in 1M 4MC

was found to be 17.2 G, which is comparable to the values of 17.25 G and 17.05 G measured in

hydrotropic assemblies of NaBMGS and Na cumenesulfonate respectively. These values were

seen to be higher at concentrations below the MHC(l8.0 G), indicating that the: hydrotrope·

aggregates offer a microenvironment of lowered polarity than the bulk phase. The 'tc values were

estimated for a few hydrotropes and also for Na octanoate and these are listed in Table 2.

32

Table 2. Reorientational correlation times of 5-doxylstearate in several systems:

--------------------------------------------------------------------------------------------------------------

SYSTEM 'tc VALUES (nanoseconds)

Water 0.20

1.25M Resorcinol 0.33

L25M NaBMGS 0.77

LOOM 4MC 0.76

LOOM Na octanoate 1.10

The above table suggests that the probe experiences greater hindrance in hydrotrope

assemblies of 4MC than in those of resorcinol, but has comparable mobility with that of in

NaBMGS.Its mobility appears somewhat greater in hydrotrope assemblies than in micelles,

which might indicate that the former are somewhat more loosely packed than the latter.

3.2.5. Crystal Structure Data on Hydroxybenzenes: Since hydrotropes self-aggregate at fairly

high concentrations, it might be possible to draw a correlation between the crystal structure and

the intermolecular interactions and organization that are displayed at high concentrations in

solution, where molecular aggregates are detected; for example would these be the incipient

structures leading ultimately to the crystal order? The crystal structure of catechol has been

investigated by Brown ( 1966) and by Wunderlitch and Mootz ( 1971 ). Pairs of catechol molecules

related by a centre of symmetry are linked together by hydrogen bonds in the crystal. Successive

pairs of molecules form thick layers parallel to the 00 I plane by means of further hydrogen bonds

which form a helical array around the crystallographic screw axis. The crystal structure of 4MC

is not known till date. Since 4MC appears to be the best hydrotrope it would be interesting to find

out its crystal structure and attempts are being made in this direction. Also, molecular modelling

might offer a clue its mode of packing in the aggregated state and such modelling studies are also

33

being attempted currently. In any event molecular layering or stacking similar to that of catechol

is speculated for 4-methyl catechol. The stacking in case of 4MC might even be better due to the

increased hydrophobic interaction by virtue of the presence of the methyl group.

The crystal structure of resorcinol has been determined by Robertson ( 1936) and is quite

similar. The crystal structure of pyrogallol (Becker et al., 1972) also reveals a very close-packed

network of intramolecular and intermolecular hydrogen bonds. It is likely that in concentrated

aqueous solutions, such intermolecular network and layering occur in these hydrotropes. Such

stacking has been suggested as a possibility in the case of several hydrotropic salts (Badwan et

al., 1983; Rath, 1965).

Thus it appears that in addition to the amphiphilic nature of the hydrotrope, its molecular

structure also plays a crucial role in determining its hydrotropic efficiency. The fact that

information obtained from crystal structure data supports the solution state behaviour of

hydrotropes strengthens this point further.

34