Macromolecules 2005, 38, 9333-9340 9333

Helical and Coil Conformations of Poly(ethylene glycol) in Isobutyric

Acid and Water

Michael L. Alessi,† Alexander I. Norman,†,‡ Sasha E. Knowlton,‡ Derek L. Ho,§ and

Sandra C. Greer*,†,‡
Department of Chemical and Biomolecular Engineering and Department of Chemistry and
Biochemistry, The University of Maryland College Park, College Park, Maryland 20742, and Center for
Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899
Received June 23, 2005; Revised Manuscript Received August 19, 2005

ABSTRACT: We show that when poly(ethylene glycol) (PEG) is dissolved in isobutyric acid at
temperatures below about 55 °C, the polymer molecule can form helices. Small-angle neutron scattering
indicates that in pure isobutyric acid and in isobutyric acid-rich aqueous solutions the polymer chains
form stiff rods that coexist with polymer coils when the polymer molecular weight is 2.38 × 104, 2.13 ×
105, and 2.87 × 105 g/mol, but that at the lower molecular weight of 1.73 × 102, only the polymer rods
form. The addition of chiral dopants causes a net optical rotation in the solution, indicating that the rods
are actually helices. Above about 60 °C in deuterated isobutyric acid (and above about 70 °C in
hydrogenated isobutyric acid), the helices convert to coils. In water, the PEG molecules form coils which
persist over the entire temperature range studied (25-60 °C).

Introduction A study of the molecular weight distributions of

polydisperse PEG in coexisting liquid phases has shown
Poly(ethylene glycol) (PEG) is a simple synthetic
a remarkable fractionation of the PEG in a particular
polymer: a chain of [-O-CH2-CH2-] units, usually
solvent system.12 Water and isobutyric acid are mutu-
terminated on each end by an -OH group. Other
ally soluble above about 26 °C but separate into two
terminations are possible, e.g., -OCH3, in which case
liquid phases below the upper critical solution temper-
the polymer is referred to as poly(ethylene oxide) (PEO).
PEG has both hydrophobic segments (the carbon atoms) ature at a mass fraction of about 39% isobutyric acid.13,14
and hydrophilic segments (the oxygen atoms) and is The molecular weight distributions of the PEG in the
soluble in both aqueous and organic solvents. The water-rich phase and in the isobutyric acid-rich phase
scientific literature on PEG is vast, amounting to about were determined by size exclusion chromatography. It
90 000 articles. PEG has many practical applications, was found that the majority of the polymer mass was
from detergents and paints to drug delivery and protein in the upper, isobutyric acid-rich phase. However, the
purification, and is arguably the most important water- majority of the higher molecular weight polymers were
soluble synthetic polymer.1,2 found to equilibrate into the lower, water-rich phase,
In the crystal,3,4 the PEG unit cell is monoclinic, and a significant reduction in polydispersity was ob-
containing four chains. Each chain adopts a distorted served in the water-rich phase. This observation led us
structure with seven monomer units forming two helical to examine the conformations of the PEG in the coexist-
turns. If the sequence -O-CH2-CH2-O- has a gauche ing phases.
(g) conformation about the -C-C- bond and a trans We report here experimental evidence of a helical
(t) conformation about the -C-O- bond, then the molecular conformation of PEG in this particular sol-
conformation of the sequence is “ttg”. The helical form vent, isobutyric acid. We study the PEG conformation
of PEG consists of such ttg sequences. The reported in mixtures of water and isobutyric acid. In water-rich
X-ray diffraction data3 show a wide distribution of the solutions, the PEG polymer chains form coils with the
C-O torsional angles, suggesting that some C-O bonds fractal dimension expected when excluded-volume in-
are not at the minimum potential energy and implying teractions are present. In isobutyric acid-rich solutions,
a distorted helix. Scanning tunneling microscopy of PEO the polymer chains form stiff rods that coexist with coils
adsorbed on graphite shows single-, double-, and mul- at molecular weights of 2.38 × 104, 2.13 × 105, and 2.87
tistranded helices; the average width of the single helix × 105, but only rods exist at the molecular weight of
is 7.6 Å.4 1.73 × 102. At temperatures above about 60 °C in
In water, the PEG molecule forms a loose coil.5 There deuterated isobutyric acid (and above about 70 °C in
is some experimental evidence of short-range ttg se- hydrogenated isobutyric acid), the rods revert to coils.
quences in PEG molecules in water,6-9 but simulations We show, by using a chiral dopant to induce optical
show no such helical sequences.10,11 The overall PEG activity, that the rods are, in fact, helices. There are
conformation in water is that of a coil. reports in the literature of polymers taking a helical
form when chiral monomers are in the polymer back-
† Department of Chemical and Biomolecular Engineering, bone, or chiral side chains are attached to the polymer,
The University of Maryland College Park.
‡ Department of Chemistry and Biochemistry, The University
or the polymer is in a chiral solvent.15 Like the poly-
of Maryland College Park.
isocyanates in chloroform or hexane,16 PEG in isobutyric
§ National Institute of Standards and Technology. acid is a polymer of achiral monomers in an achiral
* Corresponding author. E-mail: sgreer@umd.edu. solvent, which takes a helical conformation. The PEG
10.1021/ma051339e CCC: $30.25 © 2005 American Chemical Society
Published on Web 10/04/2005
9334 Alessi et al. Macromolecules, Vol. 38, No. 22, 2005

Table 1. Commercial PEG Samples Used, Showing the Termination Groups, Lot Numbers, Weight Average Molecular
Weight (Mw), Number Average Molecular Weight (Mn) and Polydispersity (Mw/Mn)

PEG sample lot no. source Mw g/mol Mn g/mol Mw/Mn

2kOH PEG2OH-2K Polymer Source 1.01 × 103 9.19 × 102 1.10
2kOCH3 PEG2OCH3-2K Polymer Source 1.73 × 103 1.56 × 103 1.11
8kOH 71K0114 Sigma 8.64 × 103 6.09 × 103 1.42
10kOH PEG2OH-10K Polymer Source 1.14 × 104 7.10 × 103 1.61
10kOCH3 P2963-2OCH3 Polymer Source 1.29 × 104 5.84 × 103 2.21
20kOH 425182/1 Fluka 2.38 × 104 2.10 × 104 1.14
200kOCH3 062725JO Aldrich 1.51 × 105 7.29 × 104 2.08
252k* P3624-EOOCH3 Polymer Source 2.87 × 105 2.48 × 105 1.16
337kOCH3 P1590-EO Polymer Source 2.13 × 105 1.53 × 105 1.40
1000kOCH3 N/A Polymer Source 9.33 × 106 4.57 × 106 2.04

helices occur as a racemic mixture, but an enantiomeric Neutron Research (NCNR) in Gaithersburg, MD,18 and were
excess is induced by a chiral dopant. probed over the q range19 from 0.0029 to 0.4014 Å-1, where q
) (4π/λ) sin(θ/2), θ is the scattering angle, the wavelength λ
Experimental Methods is 8 Å, and ∆λ/λ is 0.11. Samples were loaded in either 1 or 2
mm path length quartz cells and closed with Teflon plugs. The
Materials. Commercial PEG samples were used without raw data were corrected for background and parasitic scat-
further purification. The samples are listed in Table 1, where tering, placed on an absolute scale using a calibrated secondary
Mn is the number-average molecular weight, Mw is the weight- standard, and circularly averaged to yield the scattered
average molecular weight, and Mw/Mn is the polydispersity intensity I(q). The incoherent background from the pure
index. The molecular weights were all determined by size solvents was measured, corrected by the volume fraction
exclusion chromatography in our laboratory, as described displaced by the dissolved polymer, and subtracted from the
below. reduced SANS data. The data points in the q range from ca.
The solvents used were hydrogenated isobutyric acid (IBA, 0.3 to 0.4 Å-1 were then averaged to yield the estimated
Aldrich Chemical Co., 99.9% purity), fully deuterated iso- incoherent background from polymer in the sample, which was
butyric acid (d-isobutyric acid or d-IBA, Isotec, Inc., 98% subtracted from the data as well.
deuterated), D2O (Isotec, Inc., 99.999% D), deuterated acetic SANS data can be analyzed in three ways: (1) analyzing
acid (Aldrich, 99.9% D), DCl (Sigma-Aldrich, 99% D), and the slopes of I(q) in various q ranges; (2) modeling of the data
NaOD (Sigma-Aldrich, 99.9% D). Freshly distilled, deionized using specific models for the shapes and interactions of the
H2O was obtained from a Nanopure system (Barnstead, 18 MΩ scattering particles; (3) taking the Fourier transform of I(q)
cm). The chiral dopants used in the polarimetry experiments into real space. In the analysis of the various q ranges of I(q),20
were (S)-(+)-1,2-propanediol and (R)-(-)-1,2-propanediol (Lan- the Guinier regime at low q values (qRg e 1, where Rg is the
caster Research Chemicals, 98% pure enantiomers). radius of gyration of the polymer chain) gives information
The PEG concentrations studied by small-angle neutron about Rg:
scattering and by polarimetry were 12 ( 1 mg/mL, which are
well below the estimated overlap concentrations for the above-
listed PEG polymers in coil form. For PEG polymers in rod q2Rg2
ln[I(q)] ) ln[I(0)] - (1)
forms, the overlap concentrations are lower than for the coil 3
forms17 and approach 12 mg/mL or less for molecular weights
greater than about 20 000. where I(0) is the scattering at q ) 0. The fractal regime at
Size Exclusion Chromatography (SEC). For the SEC intermediate q values gives information about particle shape;
analyses, all PEG samples were diluted to ∼2 mg/mL in freshly for example, the slope in the fractal regime is -1 for a rod
distilled, deionized water. Each sample was analyzed using and -5/3 for a coil with excluded volume. The Porod regime at
an autosampler on a Waters SEC apparatus. The Waters SEC high q values gives information about the interface between
consists of a 1525 binary HPLC pump, a 2414 refractive index the particle and the solvent. The q value at which I(q) changes
detector, a 2487 dual wavelength absorbance detector, and a from the Guinier regime to the fractal regime corresponds to
717 autosampler. A series of three Ultrahydrogel columns (120, 2π/L, where L is the length of the particle. The transition from
250, and 2000) of dimension 7.8 × 300 mm were placed in the fractal regime to the Porod downturn corresponds to 2π/
series in a column heater and set to 45 °C. A mobile phase of D, where D is the diameter of the particle.21 However, the
0.01 M K2HPO4(aq) at pH ) 7.065 was used. The apparatus analysis of the q values at which I(q) changes slope “can only
was calibrated with known PEG standards (American Polymer be considered to be semi-quantitative, and ... does not usually
Standards Corp., Mentor, OH) of molecular weight 2500, provide accurate results.”22
10 225, 14 500, 30 225, 50 000, 111 000, 250 000, and 510 000 The second method of analysis is that using models,
g mol-1. Each sample was then injected for a run time of 60 especially those provided by NCNR.23 The analysis of SANS
min at a flow rate of 0.6 mL/min. An injection volume of 100 data for PEG solutions by Fourier transform (see below and
µL was used throughout. The subsequent chromatograms were Table 2, last column) indicated that under many conditions of
analyzed and converted to full molecular weight distributions molecular weight, solvent composition, and temperature the
using the Waters software package Breeze, version 3.30. PEG molecules formed a mixture of coil and rod conformations.
Small-Angle Neutron Scattering (SANS). For the SANS The presence of a mixture made the use of models problematic.
measurements, the PEG was dissolved in d-isobutyric acid and In one case, (second sample in Table 2), the PEG formed only
D2O, the deuteration of the solvent serving to reduce the rodlike species, and in that case we applied the core-shell
incoherent background scattering and provide contrast with cylindrical model.21 This model provides estimates of the cross-
the hydrogenated polymer. sectional radius and length of the core cylinder, assuming
Small-angle neutron scattering (SANS) measurements were circular cross sections and monodisperse polymer chains. The
carried out using the NG7 30-m SANS instrument at the volume fractions of the polymer and the scattering length
National Institute of Standards and Technology Center for densities of the solvents were fixed in that analysis.
Macromolecules, Vol. 38, No. 22, 2005 PEG in Isobutyric Acid and Water 9335

Table 2. Values of Radius, R (Coil or Helix), Radius of Gyration, Rg (for a Coil, Rg ≈ R), and Dmax (See Text) for Various
Molecular Weights of Linear PEG in Different Solvents at Different Temperaturesa
mol wt, Mw g/mol solvent temp /°C R/Å Rg/Å Dmax/Å helix or coil?
1.73 × 103 D2O 60 12.0 16* 62 coil
1.73 × 103 dIBA/D2O (0.39/0.61) 60 3.5 (3.3) 40 (34) helix
1.73 × 103 AcOD 43 16.5 16.1* 75 coil
1.73 × 103 AcOD 55 16.5 17.5* 80 coil
2.38 × 104 D2O 30 35.0 43* 185 coil
2.38 × 104 D2O 55 41.3 46* 215 coil
2.38 × 104 D2O 60 40 46* 190 coil
2.38 × 104 dIBA/D2O (0.46/0.54)-2 30 6.0 43, 44* coil + helix
2.38 × 104 dIBA/D2O (0.46/0.54)-2 40 6.5 49, 60* coil + helix
2.38 × 104 dIBA/D2O (0.46/0.54) 45 5.5 57.6, 51* coil + helix
2.38 × 104 dIBA/D2O (0.46/0.54) 55 6.5 50.6, 49.2* coil + helix
2.38 × 104 dIBA/D2O (0.39/0.61) 60 3.5 28, 34.5* coil + helix
2.38 × 104 dIBA 30 5.0 45, 42.4* coil + helix
2.38 × 104 dIBA 43 5.0 41, 42.3* coil + helix
2.38 × 104 dIBA 55 6.0 43.5, 38* coil + helix
2.38 × 104 dIBA 60 40.0 46* 190 coil
2.38 × 104 AcOD 25 42.3 44* 250 coil
2.38 × 104 AcOD 55 44.0 45* 226 coil
2.38 × 104 AcOD/D2O (0.39/0.61) 25 46.0 44* 200 coil
2.38 × 104 AcOD/D2O (0.39/0.61) 55 46.0 42* 200 coil
2.38 × 104 DCl (0.1M) 25 39.4 42* 150 coil
2.38 × 104 DCl (0.001M) 25 41.0 40* 147 coil
2.38 × 104 NaOD (0.001M) 25 42.5 48* 195 coil
2.38 × 104 D2O/NaCl 25 44.0 45* 195 coil
2.13 × 105 D2O 60 130.0 200* 400 coil
2.13 × 105 dIBA/D2O (0.46/0.54)-2 30 20.0 (3.3) 101* aggregated helix
2.13 × 105 dIBA/D2O (0.46/0.54) 55 50.0 (2.9) 77* aggregated helix
2.13 × 105 dIBA 30 5 101 coil + helix
2.13 × 105 dIBA 43 6 120 coil + helix
2.13 × 105 dIBA 55 5.5 102 coil + helix
2.13 × 105 dIBA 60 130.0 161 320 coil
2.87 × 105 D2O 43 87.0 90* 445 coil
2.87 × 105 D2O 60 100.0 148* 300 coil
2.87 × 105 AcOD 43 56.0 62* 450 coil
2.87 × 105 AcOD 55 59.0 63* 450 coil
2.87 × 105 dIBA/D2O (0.39/0.61) 43 16.0 aggregated helix
2.87 × 105 dIBA/D2O (0.39/0.61) 55 24.0 71 aggregated helix + coil
2.87 × 105 dIBA 60 90.0 118* 234 coil
a Parameters are obtained from the p(r) function, from Guinier plots (*), and from modeling to a core cylinder model (parentheses). The
estimated uncertainties are (0.6 Å for R and ( 5 Å for Dmax, both from p(r); (1 Å for Rg from Guinier plots; (0.1 Å for R from modeling.
Abbreviations: dIBA ) deuterated isobutyric acid; AcOD ) deuterated acetic acid. Concentrations of solvents are given in mass fractions;
“-2” means that the solvent was in two phases. The last column indicates the conformations of the PEG molecules. In mixtures of coil and
helix, the Rg from p(r) refers to the coils.

The third method of data analysis involves the Fourier helices in equal number (a racemic mixture) and thus give no
transformation of the scattering data. The Fourier transforma- optical rotation. However, it is known that chiral impurities
tion of I(q) involves the integral24 can lead to an “enantiomeric excess”.31-33 Therefore, we tested
the effect of a chiral dopant on the optical rotation of the PEG
∫ I(q)qr sin (qr) dq
1 ∞ in isobutyric acid to test for the presence of helices. The dopant
p(r) ) (2)
2π2 0 was 1,2-propanediol, in either (+) or (-) enantiomer, at a mass
fraction of 9 × 10-4.
which yields the pair distance distribution function p(r), where The Jasco P1010 polarimeter consists of a 10 cm path length
r is the distance in real space. The p(r) is a histogram of the glass cell with an interior volume of 8 mL. The P1010 uses a
distances between pairs of points within the scattering object. halogen lamp with a wavelength of 589 nm selected by a filter
The p(r) can be calculated from SANS data using the general- and has an accuracy of 0.002° with a reproducibility of 0.002°
ized indirect Fourier transform (GIFT) method of Glatter and or better for measurements less than 1°. To measure the
co-workers.25 In our analysis, p(r) data have been calculated optical rotation, R, as a function of temperature, the polarim-
using the GIFT software, PCG version 1.01.02.26 The GIFT eter was configured with a VWR 1160 bath to circulate
analysis used here assumes that the particles are so dilute as temperature-controlled water through a jacketed polarimeter
to be noninteracting, and thus it is not dependent on any model cell (Jasco, 10 cm path length, 3.5 mL volume). Connecting
for the structure factor. Tygon tubing was insulated tightly with foam to minimize heat
Thus, p(r) provides additional information regarding loss between the water bath and the sample. The temperature
the dimension, shape, and flexibility of the scattering at the sample was measured with a thermistor connected to a
species.24,25,27-30 For example, spherical particles give rise to Sper Scientific digital thermometer. The temperature was
a symmetric p(r) about a central maximum. A perfectly controlled and measured to an accuracy and precision of about
cylindrical particle will have a p(r) with a very sharp peak at 0.1 °C.
low values of r, which will then decrease (linearly if a rigid We first measured the optical rotation of solvent (hydroge-
cylinder or approximately linearly with some oscillation if nated isobutyric acid or H2O) with added chiral dopant. We
nonrigid) to a point where p(r) ) 0. This p(r) ) 0 point defines then dissolved PEG (12 mg/mL) into the solvent/dopant
the longest distance in the scattering object, Dmax, which is mixture and measured the optical rotation of that mixture.
the same as the length, L, for a cylinder. We subtracted the optical rotation of the solvent + dopant from
Polarimetry. If the PEG molecules form rods that are the optical rotation of the solvent + dopant + PEG to obtain
actually helices, then they will form left- and right-handed the optical rotation of the PEG. Each measurement was carried
9336 Alessi et al. Macromolecules, Vol. 38, No. 22, 2005

Figure 1. SANS intensity, I, as a function of wave vector, q,

of PEG 20kOH (Mw ) 2.38 × 104), in D2O and in d-isobutyric
acid, at 55 °C. Slopes were calculated over the range 0.03 < q
< 0.19 Å-1 for the d-isobutyric acid solution and over the range
0.03 < q < 0.15 Å-1 for the D2O solution.

out by repeated integrations (10 s integration times) of the

signal and corrected for concentration and background effects.

Results and Discussion

Uncertainties and error bars on graphs are given as
one standard deviation.
Small-Angle Neutron Scattering (SANS). Mea- Figure 2. (a) The pair distance distribution function, p(r),
surements were made on PEG solutions with various for PEG 20kOH, Mw ) 2.38 × 104, at 55 °C, for different mass
solvent compositions and polymers of various molecular fractions, w, of d-isobutyric acid in D2O. (b) The radius of
gyration (Rg) of PEG coils and the radius (R) of PEG rods
weights, at several temperatures. Plots of all the I(q) (helices) for the system in (a), as obtained from GIFT analysis
data are available.34 Table 2 gives the parameters from and Guinier plots, as a function of mass fraction of d-isobutyric
the SANS analysis. acid in D2O.
For the system deuterated isobutyric acid + D2O, the
critical composition is 39% by mass of d-isobutyric acid dimensions from the transition points are are subject
and the critical temperature is about 42 °C,35 so samples to errors in determining the point of the slope change
at this composition would be in the two-phase region and are known not to be accurate,22 we report only the
for temperatures less than 42 °C; this condition did not values from model analyses, Guinier plots, and GIFT
occur for the 39% by mass samples. Samples with analyses in Table 2. Guinier plots were not possible for
compositions that differ from 39% d-isobutyric acid will samples where the concentration exceeded the overlap
break into two liquid phases at temperatures less than concentration.
42 °C. Three samples that were in the two-phase region Solvent Concentration Dependence. Figure 2a
are indicated in Table 2 by “-2” on the solvent composi- shows p(r) for the same data as Figure 1, for PEG
tion. For these samples, the SANS beam spanned the 20kOH, Mw ) 2.38 × 104, at three compositions of
two phases and averaged over them. d-isobutyric acid + D2O. It is obvious that on changing
Figure 1 shows a typical plot of I(q) with the features solvent composition there is a dramatic change in the
typically seen in the SANS data for these systems. The p(r). The p(r) for PEG in pure D2O (w ) 0) has one peak
data are for PEG 20kOH (see Table 1), Mw ) 2.38 × but is not symmetrical, indicating the PEG conformation
104, in pure D2O and in pure d-isobutyric acid at 55 °C. to be an elongated coil, so that the maximum in p(r)
In D2O, the slope at q ) 0.03-0.19 Å-1 is close to -5/3, gives the radius of gyration, Rg, of 41.3 ( 1.2 Å, in
which is characteristic of a polymer coil with excluded agreement with the Rg determined from a Guinier plot
volume.20 In d-isobutyric acid, the slope at q ) 0.03- (46.0 ( 0.6 Å). The coil shape is not spherical and has
0.15 Å-1 is close to -1, which is characteristic of a rod. a maximum particle dimension of 215 Å.
The increase in SANS intensity at q < 0.008 Å-1 for The addition of d-isobutyric acid to a weight fraction
both samples indicates that aggregates of polymer of 0.46 brings about an overall increase in intensity and
chains were also present, a well-known phenomenon in the appearance of a second peak at lower values of r.
aqueous PEG solutions (see below). A high-q decrease The inflection point of this intense, sharp peak yields
in I(q) occurs in d-isobutyric acid but not in D2O; this is the cross-sectional diameter of a rodlike particle.25 The
the Porod36 regime with a slope of about -4, which second, broader peak is due to the radius of gyration of
indicates scattering from the sharp nonfractal interfaces coiled PEG. In other words, there is coexistence of coiled
of the rods, supporting the conclusion of a rod conforma- PEG and rodlike PEG under these conditions. The
tion in d-isobutyric acid and a coil conformation in D2O. dimensions from p(r) are Rg ) 50.6 ( 1.2 Å (cf. Guinier
In Figure 1 for polymer 20kOH in pure d-isobutyric plot, 49.2 ( 0.6 Å) ) radius of coil and R ) 6.5 ( 0.1 Å
acid, the transition points in I(q) correspond to q values ) radius of rod. The meaning of Dmax, the point at which
of about 0.030 Å-1 for 2π/L and about 0.20 Å-1 for 2π/d. p(r) goes to zero as r increases, becomes hard to
These would correspond to a rod of length L ≈ 209 Å interpret when both rods and coils are present and is,
with a diameter d ≈ 31 Å. Given the fact that the in fact, not well-determined by the GIFT procedure, so
Macromolecules, Vol. 38, No. 22, 2005 PEG in Isobutyric Acid and Water 9337

Figure 4. P(r) for PEG 20kOH, Mw ) 2.38 × 104, in pure D2O

at 30 °C (open circles) and 55 °C (filled diamonds).

do not show any significant temperature dependence

(see Table 2).
The relative invariant, Q, calculated from Q )
∫qq12I(q) dq and normalized, is plotted in Figure 3b for
PEG 20kOH in pure d-isobutyric acid. If we assume that
helical rods scatter more than do random coils,37 then
the data in Figure 3b are consistent with this assump-
tion, since Q decreases as the temperature increases.
Figure 3b makes clear, from the gradual decrease in Q,
that, in fact, coils and rods coexist at temperatures
below 60 °C in pure d-isobutyric acid. Above 60 °C, only
PEG coils exist.
The p(r) for PEG 20kOH in pure D2O at two temper-
atures are shown in Figure 4. The PEG has coil
conformations at both temperatures. An elongated
Figure 3. (a) Pair distance distribution function for PEG spherical scattering particle with Rg ∼ 45 Å persists as
20kOH, Mw ) 2.38 × 104, in pure d-isobutyric acid, for four the temperature is raised from 30 °C through to 55 °C.
different temperatures. (b) Relative invariant, Q, as a function Clearly PEG forms coils in D2O at all temperatures.
of temperature for PEG 20kOH in pure d-isobutyric acid.
Molecular Weight Dependence. We have shown
above that in pure d-isobutyric acid at temperatures
we do not list values of Dmax except when only rods or above 60 °C the rod conformation of PEG 20kOH, Mw
only coils are present. ) 2.38 × 104, changes to a coil conformation. Figure 5a
The differences between these two p(r) profiles in shows I(q) for PEG of a lower molecular weight, PEG
Figure 2a are magnified when the solvent is pure 2kOCH3 (Mw ) 1.73 × 103), in pure d-isobutyric acid at
d-isobutyric acid (w ) 1). The peak corresponding to the 60 °C. The core-shell cylinder model with L ) 40.0 (
cross section of a rod increases in intensity and resolu- 1.2 Å and R ) 3.3 ( 0.1 Å fits the SANS data very well.
tion. The broad peak, corresponding to the radius of Figure 5b shows the p(r) profile for this sample, which
gyration of the coil, is reduced in intensity. Hence, the indicates a rigid rod with L ) 40.0 ( 1.3 Å and R ) 3.5
ratio of helical-to-coil conformations is increased with ( 0.6 Å. This is the smallest molecular weight that was
d-isobutyric acid content. Here Rg(coil) ) 43.5 ( 1.2 Å studied and the only sample that indicated the presence
(cf. Guinier plot, 38.0 ( 0.6 Å) and R ) 6.0 ( 0.1 Å. of only rods/helices, even at 60 °C. This sample indicates
Figure 2b plots the coil and rod radii as a function of that as the molecular weight of the polymer is reduced,
mass fraction of d-isobutyric acid. The error bars the tendency for a helical structure is greater.
indicate one standard deviation, so the dimensions from We note from Table 2 that Rg is larger in D2O than
GIFT and Guinier analysis are in agreement, and there in d-isobutyric acid at this temperature, indicating that
is no significant dependence of the dimensions on D2O is the better solvent for PEG. Figure 6 shows Rg of
solvent composition. the polymer coils as obtained from p(r) as a function of
Temperature Dependence. The p(r) profiles for molecular weight in D2O. The exponent of the molecular
PEG 20kOH, Mw ) 2.38 × 104, in pure d-isobutyric acid weight dependence is consistent (within error) with that
at four different temperatures are given in Figure 3a. expected (0.60) for a polymer coil in a good solvent.38
At low temperatures, a sharp peak is observed at low Polymer Aggregation. The formation of aggregates
values of r, with a broad peak at higher values. As in or clusters of PEG molecules in water has been well-
Figure 2a, these peaks correspond to the cross-sectional known39 but controversial, with arguments both that
diameter of a rod and the radius of gyration of a coil, aggregation is the result of impurities40-42 and that
respectively. The peak due to the coil increases in aggregation is an inherent property of PEG in water.43-45
intensity (with respect to the rod peak) and resolution In our SANS measurements, the aggregation of the
with temperature. This is evidence of a coexistence of PEG causes an increase in I(q) at low q (q < 0.003 Å-1).
coil and rod conformations of PEG in d-isobutyric acid In the course of our studies, we observed samples with
that favors a rod at low temperatures and favors a aggregation and samples without aggregation. All the
random coil (forming an elongated spherical scattering polymer samples showed aggregation in D2O and in
particle) at temperatures of 60 °C and higher. R and Rg mixtures of D2O and d-isobutyric acid. In pure d-
9338 Alessi et al. Macromolecules, Vol. 38, No. 22, 2005

Figure 7. (a) SANS profiles for PEG 20kOH at 25 °C in

different solvents: deuterated acetic acid, 0.1 wt % DCl in D2O,
Figure 5. (a) SANS intensity profile for PEG 2kOCH3 and 0.1 wt % NaOD in D2O. (b) The corresponding p(r) profiles.
(Mw ) 1.73 × 103) in d-isobutyric acid + D2O (0.39/61) at 60
°C and (b) the corresponding p(r) profile. that the coil-rod transition in PEG is not a function of
pH. Figure 7 shows I(q) and p(r) for these solvents. The
p(r) functions are typical of elongated spheres, corre-
sponding to elongated polymer coils, for which dimen-
sions are given in Table 2.
Polarimetry. PEG exhibited no optical rotation in
H2O containing a chiral dopant. PEG exhibited no
optical rotation in (hydrogenated) isobutyric acid with-
out a chiral dopant. However, PEG did indeed exhibit
a net optical rotation in doped isobutyric acid. The
observation of the induced optical rotation is evidence
that the PEG “rods” seen by SANS are, in fact, helical
in conformation. The optical rotation in doped isobutyric
acid was larger, and the measurements were more
reproducible when the sample was first heated above
Figure 6. Radius of gyration, Rg, for PEG coils in D2O as a 60 °C and then cooled to room temperature. This is
function of weight-average molecular weight, at 60 °C. consistent with the observation from SANS that the
PEG rod in d-isobutyric acid reverts to a coil at a
isobutyric acid, the extent of aggregation depended on temperature between 55 and 60 °C: Presumably the
the molecular mass of the polymer: PEG of lower heating converts the helix to a coil, and the resulting
molecular mass showed less aggregation, such as PEG coil is affected by the dopant when it refolds. All the
2kOCH3 as shown in Figure 5a. Thus, aggregation samples on which we report here were first heated to
always occurred in the presence of D2O, and aggregation 60 °C before the polarimetry measurements were made.
occurred in d-isobutyric acid except at Mw less than or The first polarimetry studies were made at room
equal to 1.73 × 103. temperature. Figure 8a shows that the sign of the
Table 2 shows four samples at the highest molecular optical rotation follows the chirality of the dopant: (+)-
weights in which the values of R were considerably 1,2-propanediol gives a (+) rotation for the PEG, and
larger than those of the other samples. We interpret this (-)-1,2-propanediol gives a (-) rotation for the PEG.
as due to the aggregation of the helices into larger Figure 8a also shows that the magnitude of the optical
species. rotation of the PEG increases as the concentration of
Other Solvents. Our initial hypothesis was that the the dopant increases. Figure 8b shows the optical
formation of the PEG rods was related to the pH of the rotation as a function of the molecular weight of the
solvent. We examined the SANS from PEG 20kOH in PEG, for the PEG samples described above under
DCl (0.1 wt % in D2O), in pure deuterated acetic acid “Experimental Methods, Materials”. There is no clear
and in NaOD (0.1 wt % in D2O), all at 25 °C. In all three dependence of the optical rotation on molecular weight,
cases, the PEG assumed a coil conformation, indicating but there is a symmetry around 0° rotation in that the
Macromolecules, Vol. 38, No. 22, 2005 PEG in Isobutyric Acid and Water 9339

Figure 9. Specific rotation of PEG 20kOH (Mw ) 2.38 × 104)

in hydrogenated isobutyric acid with chiral dopants, (+)- and
(-)-1,2-propanediol, as a function of temperature.

unfolding of the PEG helices into coils. This result in

hydrogenated isobutyric acid is to be compared with the
SANS results (see Figure 3) in deuterated isobutyric
acid that show a helix-coil transition at 60 °C. That
the optical rotation does not change abruptly indicates
that the helix-to-coil transition occurs slowly over a
temperature range and that a coexistence of helices and
coil conformations persists, consistent with the observa-
tions made by SANS (Figure 3). When the samples are
cooled, the optical rotation reappears, indicating the
conversion of the coils back to helices.
We observe the helix-coil transition at about 60 °C
in deuterated isobutyric acid and at about 68 °C in
hydrogenated isobutyric acid. This suggests that the
helix is more stable in hydrogenated isobutyric acid than
in deuterated isobutyric acid. We do not understand this
difference.

Summary and Conclusions

We find that in isobutyric acid PEG molecules can
take the form of helices at temperatures below about
68 °C. There is a coexistence of helices and coils for
higher molecular weights, but at low molecular weight
Figure 8. (a) Specific rotation of PEG 10kOH (Mw ) 1.14 × the polymers form only helices. The PEG helices revert
104) in isobutyric acid with chiral dopants, (+)- and (-)-1,2-
propanediol, as a function of mass fraction of dopant. (b)
to coils above about 68 °C in hydrogenated isobutyric
Specific rotation of PEG samples in isobutyric acid with chiral acid (and above about about 60 °C in deuterated
dopants, (+)- and (-)-1,2-propandiol at 0.1 mass %, as a isobutyric acid). We know of no other solvent in which
function of PEG molecular weight. (c) Specific rotation of PEG PEG has the helical conformation. We confirm that PEG
samples in isobutyric acid with chiral dopants, (+)- and (-)- molecules form coils in D2O.
1,2-propanediol at 0.1 wt %, where the PEG has different
polymer terminating groups. All these measurements were
The coexistence of PEG coils and PEG helices can
made at room temperature. have two forms: Each PEG molecule can form either a
coil or a helix, or each PEG molecule can itself be part
(-) and (+) dopants acted in equal and opposite direc- coil and part helix. It is difficult to distinguish between
tions on the induced PEG rotation, consistent with the two from the neutron scattering data. We would
Figure 8a. Figure 8c shows the rotation of PEG samples expect p(r) to have a more complex form in the case of
with nearly the same molecular weights but with polymers that are mixed coils and helices, and thus we
different terminating groups on the PEG molecules. tend to favor the case of polymers as either coils or
There is no clear pattern of end effects. helices, but we cannot firmly prove that at this time.
Next we measured the optical rotation of PEG in Why do the PEG molecules form helices in isobutyric
doped isobutyric acid as a function of temperature. acid? We can imagine three mechanisms: (1) there is a
Figure 9 shows the magnitude of the optical rotation steric effect of the isobutyric acid with the PEG that
as a function of temperature for 20kOH PEG doped in induces the helical conformation; (2) the water present
both (+)- and (-)-1,2-propanediol enantiomers. For each as a hydration layer on the PEG (even in pure isobutyric
measurement, the sample was allowed to reach equi- acid) serves as a solvophilic layer and the PEG forms a
librium (∼15 min) before the next measurement was helix to best present that layer to the solvent; (3) the
taken. For both enantiomers, the induced optical rota- water present in the hydration layer is released by the
tion decreases as the temperature increases and be- formation of the helix, thus increasing the entropy and
comes zero at about 70 °C. We propose that the driving the conformational change. We are designing
disappearance of the optical rotation indicates the new experiments to test these mechanisms.
9340 Alessi et al. Macromolecules, Vol. 38, No. 22, 2005

Our study was initially motivated by an effort to (15) Green, M. M.; Park, J.-W.; Sato, T.; Teramoto, A.; Lifson, S.;
understand the microscopic origins of the dramatic Selinger, R. L. B.; Selinger, J. V. Angew. Chem., Int. Ed. 1999,
38, 3138-3154.
fractionation of PEG in a solution of water + isobutyric (16) Lifson, S.; Andreola, C.; Peterson, N. C.; Green, M. M. J. Am.
acid.12 Water and isobutyric acid are mutually soluble Chem. Soc. 1989, 111, 8850-8858.
above about 26 °C but separate into two liquid phases (17) Teraoka, I. Polymer Solutions: An Introduction to Physical
below the upper critical solution temperature (UCST).13,14 Properties; Wiley-Interscience: New York, 2002.
When PEG is added to the two-phase water + isobutyric (18) Glinka, C. J.; Barker, J. G.; Hammouda, B.; Krueger, S.;
Moyer, J. J.; Orts, W. J. J. Appl. Crystallogr. 1998, 34, 430-
acid system, the longer polymer chains tend to go to the 445.
lower, aqueous phase and the shorter polymer chains (19) NG3 and NG7 30-Meter SANS Instruments Data Acquisition
tend to go to the upper, acidic phase. The average Manual; Cold Neutron Research Facility at the National
molecular weight in the lower phase is about twice that Institute of Standards and Technology, 2002.
in the upper phase. The greater mass of the polymer (20) Higgins, J. S.; Benoit, H. C. Polymers and Neutron Scattering;
Clarendon Press: Oxford, 1994.
(∼80%) is in the upper, acidic phase, even though PEG
(21) Schurtenberger, P. In Neutrons, X-rays, and Light Scattering
is more soluble in water than in isobutyric acid.12 This Methods Applied to Soft Condensed Matter; Lindner, P.,
new work has shown that the PEG molecules are coils Zemb, T., Eds.; Elsevier Science B.V.: New York, 2002; p 541.
in the water phase and helices in the isobutyric acid (22) Pedersen, J. S.; Schurtenberger, P. Macromolecules 1996, 29,
phase. We hope that the elucidation of the mechanism 7602-7612.
of the helix formation will elucidate the fractionation (23) http://www.ncnr.gov/resources/index.html.
(24) Glatter, O. In Neutrons, X-rays, and Light: Scattering
behavior as well. Methods Applied to Soft Condensed Matter; Lindner, P.,
Zemb, T., Eds.; Elsevier: Amsterdam, 2002; p 73.
Acknowledgment. This research was supported by (25) Glatter, O. In Neutron, X-ray, and Light Scattering: Intro-
the National Science Foundation, Chemistry Division. duction to an Investigative Tool for Colloidal and Polymeric
We also acknowledge support of the National Institute Systems; Lindner, P., Zemb, T., Eds.; North-Holland: New
York, 1990; pp 33-82.
of Standards and Technology (NIST), U.S. Department (26) GIFT software available from O. Glatter, U. o. G., Austria.
of Commerce, in providing the neutron research facili- Details can be found at the Web site http://physchem.kfuni-
ties used in this work. We thank Jack Douglas, Philip graz.ac.at/sm/.
R. Deshong, Lyle Isaacs, Yiwei Fei, and Andrew Lyon (27) Glatter, O. J. Appl. Crystallogr. 1977, 10, 415-421.
for helpful discussions. Certain products and equipment (28) Glatter, O. J. Appl. Crystallogr. 1980, 13, 577-584.
are mentioned by name only to clarify the experimental (29) Glatter, O. In Neutrons, X-rays, and Light: Scattering
Methods Applied to Soft Condensed Matter; Lindner, P.,
conditions used; this does not mean that they are the Zemb, T., Eds.; Elsevier: Amsterdam, 2002; p 103.
best for the purpose or that NIST endorses them. (30) Hansen, S. J. Appl. Crystallogr. 2003, 36, 1190-1196.
(31) Soai, K.; Shibata, T.; Sato, I. Acc. Chem. Res. 2000, 33, 382-
References and Notes 390.
(32) Singleton, D. L.; Vo, L. K. J. Am. Chem. Soc. 2002, 124,
(1) Bailey, F. E.; Koleske, J. V. Alkylene Oxides and Their 10010-10011.
Polymers; Marcel Dekker: New York, 1991; Vol. 35. (33) Yashima, E. Anal. Sci. 2002, 18, 3-6.
(2) Harris, J. M., Zalipsky, S., Eds. Poly(ethylene glycol): Chem- (34) Alessi, M. L. Ph.D. Dissertation; University of Maryland:
istry and Biological Applications; American Chemical Soci- College Park, MD, 2004; p 162.
ety: Washington, DC, 1997; Vol. 680. (35) Greer, S. C. Ber. Bunsen-Ges. Phys. Chem. 1977, 81, 1079-
(3) Takahashi, Y.; Tadokoro, H. Macromolecules 1973, 6, 672- 1081.
675.
(36) Porod, G. Kolloid-Z. 1951, 124, 83-114.
(4) Yang, R.; Yang, X. R.; Evans, D. F.; Hendrikson, W. A.; Baker,
J. J. Phys. Chem. 1990, 94, 6123-6125. (37) Koh, A. Y. C.; Heenan, R. K.; Saunders, B. R. Phys. Chem.
(5) Vennaman, N.; Lechner, M. D.; Oberthuer, R. C. Polymer Chem. Phys. 2003, 5, 2417-2423.
1987, 28, 1738-1748. (38) Grosberg, A. Y.; Khokhlov, A. R. Statistical Physics of
(6) Matsuura, H.; Fukuhara, K. J. Mol. Struct. 1985, 126, 251- Macromolecules; American Institute of Physics: New York,
260. 1994.
(7) Yang, X.; Su, Z.; Wu, D.; Hsu, S. L.; Stidham, H. D. (39) Polverari, M.; van de Ven, T. G. M. J. Phys. Chem. 1996, 100,
Macromolecules 1997, 30, 3796-3802. 13687-13695.
(8) Begun, R.; Matsuura, H. J. Chem. Soc., Faraday Trans. 1997, (40) Devanand, K.; Selser, J. C. Nature (London) 1990, 343, 739-
93, 3839-3848. 741.
(9) Oesterhelt, F.; Rief, M.; Gaub, H. E. New J. Phys. 1999, 1, (41) Porsch, B.; Sundelof, L.-O. Macromolecules 1995, 28, 7165-
6.1-6.11. 7170.
(10) Smith, G. D.; Bedrov, D.; Borodin, O. J. Am. Chem. Soc. 2000, (42) Faraone, A.; Magazu, S.; Maisano, G.; Migliardo, P.; Tetta-
122, 9548-9549. manti, E.; Villari, V. J. Chem. Phys. 1999, 110, 1801-1806.
(11) Bedrov, D.; Smith, G. D. J. Chem. Phys. 2003, 118, 6656- (43) Hammouda, B.; Ho, D.; Kline, S. Macromolecules 2002, 35,
6663. 8578-8585.
(12) Shresh, R. S.; MacDonald, R. C.; Greer, S. C. J. Chem. Phys. (44) Ho, D. L.; Hammouda, B.; Kline, S. R. J. Polym. Sci., Part B:
2002, 117, 9037-9049. Polym. Phys. 2003, 41, 135-138.
(13) Greer, S. C. Phys. Rev. A 1976, 14, 1770-1780. (45) Duval, M. Macromolecules 2000, 33, 7862-7867.
(14) Venkataraman, T. S.; Narducci, L. M. J. Phys. C: Solid State
Phys. 1977, 10, 2849-2861. MA051339E