+ +

Macromolecules 1996, 29, 4397-4405 4397

Crystal Structures of Optically Active Polyamides Derived from

Di-O-methyl-L-tartaric Acid and 1,n-Alkanediamines: A Study
Combining Energy Calculations, Diffraction Analysis, and
Modeling Simulations

I. Iribarren, C. Alemán, J. J. Bou, and S. Muñoz-Guerra*

Departament d’Enginyeria Quı́mica, Universitat Politècnica de Catalunya ETSEIB,
Diagonal 647, 08028 Barcelona, Spain
Received September 15, 1995; Revised Manuscript Received February 2, 1996X

ABSTRACT: The crystal structures of a series of optically active polyamides derived from L-tartaric acid
and 1,n-alkanediamines (n ) 2, 4, 6, and 8) have been investigated. Experimental data provided by
X-ray diffraction of powders and fibers as well as by electron diffraction of single crystals were used to
determine the lattice parameters for each polymer. A triclinic unit cell with space group P1 was found
to be shared by the whole series. Semiempirical quantum mechanical calculations revealed that the
preferred conformation for these polytartaramides entailed the tartaric acid moiety in a gauche
arrangement with the amide groups rotated out of the plane containing the all-trans polymethylene
segment. Crystal models compatible with the crystallographic data were built and refined against X-ray
diffraction intensities with the linked-atom least-squares (LALS) methodology. The favored structure
appeared to consist of hydrogen-bonded pleated sheets packed with a stagger similar to that found in the
R-form of nylon 66. Modeling performed with CERIUS, a program enabling on-line observation of the
simulated X-ray and electron scattering as the molecular model is adjusted, led to substantially similar
results.

Introduction 185-312 °C, with Tm describing a zigzag curve with

Carbohydrate-based polyamides, i.e., polysacchar- local maxima and minima corresponding to polymers
amides, are receiving both increasing academic and with even and odd values of n, respectively. By analogy
industrial attention because of the natural abundance with other series of polycondensation polymers consist-
of their feedstocks and because they are promising ing of flexible polymethylene segments of varying
materials with novel technical possibilities.1 Varying length,15,16 such a correlation implies the existence of
properties, such as hydrophilicity/hydrophobicity, bio- different crystalline structures in each case.
degradability, and potential biocompatibility with other The crystal structures of polyaldaramides have not
naturals, make these polyamides interesting for a been investigated thus far, even though molecular
number of currently demanding applications. Poly- arrangements departing from those typically adopted
aldaramides are linear polysaccharamides produced by conventional nylons may be expected for these highly
from aldaric acids and aliphatic diamines. Although substituted chiral polyamides. As with simple sugar
polyaldaramides began to be explored in the earlier derivatives,17 repulsive effects between pendent oxy
1960s2 and were extensively investigated by Ogata et groups would be expected to prevent the backbone of
al. along the 1970s,3-6 a renewed interest in them has the aldaric acid moiety to be in an extended conforma-
emerged in these last few years.7-10 Recently, we have tion. Furthermore, possible intramolecular hydrogen
reported on the synthesis and properties of poly- bond interactions taking place between side-chain oxy-
tartaramides, which are polyaldaramides deriving from gen atoms and amide groups may contribute also to
the naturally occurring tartaric acid.11-14 Among the modify the conformation of the polymer. In this regard,
different types of polytartaramides that have been several studies carried out on N-alkyl-D-gluconamides
investigated by us, those based on 2,3-di-O-methyl-l- showed that these molecules crystallized in a sickle
tartaric acid and 1,n-alkanediamines, namely PnDMLT, conformation arising from a change in the torsion angle
around the C1-C2 bond in the alkyl chain and that the
crystal lattice is stabilized by both inter- and intramo-
lecular hydrogen bonds.18,19 More recently, computer
simulation studies based on experimental NMR data of
polyaldaramides have been reported.20 They suggest
that certain hydroxyl unprotected polyaldaramides tend
to form helical structures stabilized by intramolecular
have been the object of preferential attention. They hydrogen bonds; the symmetry displayed by the helix
stand out by combining an overal desirable pattern of seems to be determined largely by the stereochemical
properties with relatively easy accessibility. configuration of the carbohydrate moiety.21
PnDMLT’s are stereoregular polyamides that display As a first step in the development of a long-term
high crystallinity and show large optical activity in project intended to investigate the conformations and
solution. They exhibit a pronounced affinity for water crystal structures of polytartaramides in relation to
and hydrolyze faster than conventional polyamides.12 their chemical constitution and properties, the present
Melting points of the PnDMLT’s fall within the interval study has focused on PnDMLT’s having even values of
n. Powders, fibers, and single crystals of members with
* To whom all correspondence should be addressed. n ) 2, 4, 6, and 8 were analyzed by X-ray diffraction
X Abstract published in Advance ACS Abstracts, May 1, 1996. and electron microscopy and the crystal structures
+ +

4398 Iribarren et al. Macromolecules, Vol. 29, No. 12, 1996

Table 1. Characteristics of Polyamides PnDMLT Studied

in this Worka
polyamide Mwb Mnb Tm,c °C Tg,c °C [R]23Dd
P2DMLT 10 800 7 700 311.5 175.2
P4DMLT 25 100 15 200 266.1 100 121.0
P6DMLT 35 300 15 700 230.2 106 104.4
P8DMLT 69 300 44 600 208.1 89 92.4
a Data taken from ref 12. b Measured by GPC on trifluoroacety-

lated samples and calibrated against polystyrene standards.

c Determined by DSC. d Measured in formic acid.

examined in light of the collected diffraction data. Due

to the structural similitudes found for the different
PnDMLT’s, computational and modeling calculations
were carried out basically on poly(hexamethylene-2,3-
di-O-methyl-L-tartaramide) (P6DMLT) and the results
obtained therefrom were extended to the other members
of the series. The strategy adopted in this paper
combines semiempirical quantum mechanical calcula- Figure 1. Chemical formulas of model compound N,N′-
tions with diffraction results to unravel the preferred dimethyl-2,3-di-O-methyl-L-tartaramide (1) and the repeating
conformations for the polytartaramide chain that are unit of P6DMLT (2) with the dihedral angle and atom
notations used in this work indicated.
compatible with the crystallographic data. The ener-
getically favored conformers were used to construct
Quantum Mechanical Calculations. N,N′-Dimethyl-2,3-
crystal models which were refined against X-ray data
di-O-methyl-L-tartaramide (1) was selected as the model
using the linked-atom least-squares (LALS) methodol- molecule for calculating the conformational preferences of the
ogy.22 A parallel modeling analysis has been performed tartaric unit in polyamides PnDMLT. Energy calculations
by means of CERIUS,23 a program enabling experimen- were performed using the semiempirical AM1 Hamiltonian24
tal diffraction patterns to be compared with those which has been proved to be useful in the structural analysis
obtained from crystal models by computer simulation. of related polyamides.25,26 Two specific cases were considered
in these calculations: (i) Ψ ) Ψ ′ and (ii) Ψ ) -Ψ ′. In both
Materials and Methods cases, conformational maps for combinations of the dihedral
angles Ψ and υ were generated using the standard parameters
Experimental Procedures. The polytartaramides
implemented in the MOPAC package.27 Potential energy
(PnDMLT, n ) 2, 4, 6, 8) used in this work were prepared by
surfaces were calculated using a grid of 20°. Torsional angles
polycondensation of bis(pentachlorophenyl) 2,3-di-O-methyl-
of the amide groups were held fixed at 180°, whereas all other
L-tartarate with the corresponding 1,n-alkanediamines in a
chloroform solution as described in detail elsewhere.12 Some geometrical parameters were relaxed. Then, the effect of
data of these polyamides relevant to the present study are internal rotations of the tartaric unit on an infinite polymer
given in Table 1. chain was analyzed. The repeating unit of P6DMLT (2) was
used for these calculations, where contour conditions were
Densities were measured at 25 °C on pieces of polymer films
applied in order to generate regular conformations. The
by the flotation method in a KBr 25% (w/w) aqueous solution.
chemical formulas of 1 and 2 with indication of the notations
X-ray diffraction diagrams were recorded from both powders
and fibers. Fibers of P6DMLT and P8DMLT were obtained used to specify the conformational angles and atoms are
by stretching the polymer from the melt and subsequent represented in Figure 1. All the calculations were performed
annealing at a temperature near the melting point. In order on an IBM/3090 computer at the Centre de Supercomputació
to avoid decomposition at heating, fibers of P4DMLT were de Catalunya (CESCA).
prepared at room temperature by pulling out a concentrated Crystal Model Building and Refinement. Those con-
solution of the polymer in formic acid. In contrast to the other formations of the tartaric unit favored by AM1 calculations
members of the series, P2DMLT could not be oriented when were used for building the asymmetric unit of the P6DMLT
the same methods were applied. X-ray diffraction diagrams chain by means of LALS. Bond lengths and angles were kept
were registered on flat films in a modified Statton-type camera fixed at their standard values, and the amide group was held
using nickel-filtered Cu KR radiation of wavelength 1.542 Å, in a trans conformation. Standard hydrogen bond parameters
and they were calibrated with molybdenum sulfide (d002 ) were introduced and forced to be mantained within reasonable
6.147 Å). Reflection maximum intensities were measured with ranges of variations. The resulting models were then adjusted
a unidimensional Joyce Loebl MK III CS microdensitometer by taking into account both packing and geometric restrictions
and corrected for polarization and Lorentz factors. In the case imposed by experimental diffraction data. Finally, the crystal
of fiber patterns, corrections taking into account the orienta- structure was refined against X-ray intensities measured on
tion of the diffracting planes were additionally made for every fiber patterns.
spot. However, no absorption correction was needed since the
films used for diffraction were thin. The repeating unit of P6DMLT (2) in its energetically
Electron microscopy observations were carried out on a favored conformations was also used to model the crystal
Philips EM-301 instrument operating at 80 and 100 kV for lattice of this polymer by CERIUS. The unit cell parameters
bright field and electron diffraction, respectively. Crystalliza- and symmetry determined initially by X-ray and electron
tions were accomplished isothermally from diluted solutions diffraction were used for generating the structure. Adjust-
of polymers in either glycerol or triethylene glycol. Single ments were made to satisfy contour conditions, namely,
crystals were recovered by centrifugation and, after repeated coincidence in the space at both ends of 2. Forbidden close
washing with 1-butanol, deposited on electron microscope contacts between neighboring chains were removed, and
carbon-coated grids. Crystals to be observed in the bright field hydrogen bonds were made between chains separated by
were shadowed with Pt carbon at an angle of approximately appropriate distances. X-ray and electron diffractograms were
15°, whereas unshadowed samples were used for electron simulated on-line and compared with those available from
diffraction examination. Electron diffraction diagrams were experiments. At this stage, the model was refined until
recorded in the selected area mode under a minimum flux of optimum agreement between calculated and experimental data
radiation and were internally calibrated with gold (d111 ) 2.35 was attained. All CERIUS calculations and simulations were
Å). performed on a Silicon Graphics RI-4000.
+ +

Macromolecules, Vol. 29, No. 12, 1996 Crystal Structures of Optically Active Polyamides 4399

Figure 2. X-ray powder diffraction patterns of polytartar-

amides. (a) P2DMLT; (b) P4DMLT; (c) P6DMLT; (d) P8DMLT.

Results and Discussion

X-ray Diffraction. The powder diagrams recorded
from PnDMLT for n ) 2, 4, 6, and 8 are shown in Figure
2. The general appearance displayed by these patterns
indicates that a notable crystallinity is present in all
cases. When these diagrams are compared in detail,
the possibility that a similar type of crystal structure
is shared by the four polymers is immediately suggested.
This was definitely confirmed by the results obtained
in the diffraction analysis of oriented samples. The
pattern shown in Figure 3a was taken from a fiber of
P6DMLT stretched from the melt. Similar patterns
were obtained from P4DMLT and P8DMLT. The domi-
nant features common to the fiber diagrams of PnDMLT
are (a) the existence on the equator of two strong
reflections spaced at about 5.6 and 4.5 Å, respectively,
Figure 3. X-ray fiber diffraction patterns of P6DMLT. The
and (b) the presence of an intense first layer line with fiber axis is vertical. (a) Experimentally observed pattern. (b)
a spacing steadily increasing with n, the latter obviously Pattern simulated by CERIUS from the triclinic model P1.
corresponding to the axial repeat of the structure. The
similitude observed in the scattering along the equator with respect to the all-trans conformation adopted in
demonstrates that the same side-by-side packing of the R-form of nylons.28 The puckering is still consider-
chains is adopted in all cases. On the other hand, the ably deeper than that taking place in the γ-form of
distribution of the reflections along the strata reveals nylons.15 The plot of c against n results in a straight
that the 00l planes are inclined about 60° with respect line (Figure 4) with a slope of about 1.20 Å, which is a
to the polymer chains, which are aligned parallel to the distance slightly shorter than the calculated mean rise
fiber axis. of one methylene in the trans conformation (1.26 Å).
A triclinic lattice containing one chain per unit cell Such a small but significant difference could be ac-
was found to be consistent with the diffraction data counted for if the polymethylene segment is assumed
available for the four polymers. According to what is to be tilted about 20° with respect to the c axis. On the
usual in nylons,28 the structure is interpreted as a other hand, when n is made to be equal to 0, a value
stacking of hydrogen-bonded sheets regularly staggered. close to 6 Å is inferred for c, which represents the height
The interchain distance within the sheets is about 5 Å, of the sequence -HNCOCH(OCH3)CH(OCH3)CONH-
which coincides with the a dimension of the crystal; the in the three polytartaramides. In order to be accom-
intersheet distance is given by the 010 spacing and modated in such a space, the tartaric moiety must
ranges from 5 to 6 Å. Indexing of the four polymers on assume a skewed conformation provided that the
the basis of this structure together with their respective -HNCO- groups retain their usual planar arrange-
observed and calculated spacings is listed in Table 2. ment.
The c dimension of the structure, i.e., the axial repeat A rather different situation is thought to occur in the
length, was found to vary from 10.8 to 15.7 Å as n case of P2DMLT, where the shortening of the axial
increased from 4 to 8. This implies that polytartar- repeat turns out to be less than 1 Å per residue.
amides PnDMLT with n ) 4, 6, and 8 crystallize with Although diffraction data obtained from this compound
the chain contracted about 1.5-2 Å per repeating unit are comparatively poor, the deviation observed for the
+ +

4400 Iribarren et al. Macromolecules, Vol. 29, No. 12, 1996

Table 2. Interplanar Spacings (dhld) and Crystal Lattice Parameters (Å) of Polyamides PnDMLT
P2DMLT P4DMLT P6DMLT P8DMLT

d(obsd) d(obsd) d(obsd) d(obsd)

h k l X-rays d(calcd) X-rays ED d(calcd) X-rays ED d(calcd) X-rays ED d(calcd)
0 1 0 5.60 5.55 5.60 5.65 5.58 5.53 5.45 5.46 5.18 5.15
1 0 0 4.65 4.67 4.60 4.60 4.59 4.57 4.60 4.54 4.54 4.60 4.51
-2 1 0 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50
0 0 1 7.65 7.67 9.03 9.08 11.20 11.33 12.60 12.74
0 1 1 6.43 6.43 6.62 6.64 6.40 6.35 6.35 6.18 6.20 6.20
1 0 1 4.32 4.35 4.53 4.50 4.60 4.57 4.65 4.63
0 -1 1 3.89 3.90 4.17 4.15 3.97 4.03
1 1 1 3.30 3.31 3.20 3.19 3.20 3.10 3.15 3.11
0 1 2 5.63 6.10 6.12
1 0 2 4.48 4.48 3.60 3.65 4.00 3.98 4.23 4.22
0 2 2 3.15 3.18 3.05 3.10
1 0 3 2.83 2.82 3.26 3.27 3.58 3.57
0 1 4 2.68 2.67 3.30 3.30 3.93 3.92
1 0 4 2.69 2.67 2.99 2.98
0 1 5 2.63 2.62 3.12 3.12
0 1 6 2.18 2.16 2.54 2.56
a, b, c 4.95, 6.87, 9.10 5.00, 7.06, 10.80 5.00, 6.84, 13.20 5.00, 6.82, 15.7
R, â, γ 59.0, 90.0, 106.6 59.4, 90.0, 110.0 61.5, 90.0, 111.6 57.0, 90.0, 111.3
F, g mL-1
obsd 1.25 1.21 1.13 1.18
calcd 1.34 1.26 1.19 1.18

PnDMLT’s and imply that the hydrogen-bonded sheets

run parallel to the long dimension of the crystals. Due
to the oblique geometry of the triclinic lattice, diffraction
patterns composed of hkl reflections with l * 0 were
registered more frequently than those corresponding to
the [001] axis zone. In addition, some reflections may
appear occasionally doubled, in particular those arising
from 0kl planes. The few number of repeating units
accommodated within the lamellar height entails a
distortion of the reciprocal lattice, meaning that more
than one plane may be intercepted by the Ewald sphere.
These features become well illustrated in the diagram
depicted in Figure 6a, which was recorded from a crystal
of P6DMLT lying on its basal plane, i.e., with the chains
leaning about 30° with respect to the electron beam
Figure 4. Repeating unit length of polyamides n,4 plotted direction. The spacings measured for each polymer by
against n for different conformations. O, fully extended electron diffraction, regardless of which was the dif-
conformation characteristic of the R-form; 9, 21 helical con-
formation characteristic of the γ-form; 2, conformation ob-
fracting zone recorded, are compiled in Table 2 where
served for the polytartaramides studied in this work. they can be compared with those estimated by X-ray
diffraction.
value of c is out of the acceptable margin of experimental Conformational Analysis of the Tartaric Acid
error. Accordingly, the conformation of the chain in Unit. The Ψ-υ conformational energy maps resulting
P2DMLT is thought to be more relaxed than in other for model compound 1 under the constraints stated for
PnDMLT’s in spite of the fact that the same type of cases i and ii are shown in Figure 7. Contour lines are
crystal lattice seems to be adopted by this polymer. drawn for energy increments of 1 kcal mol-1, revealing
Electron Microscopy and Electron Diffraction. that several low-energy conformational regions (below
Crystallization of polyamides PnDMLT for n ) 4, 6, and 5 kcal mol-1) are possible for both cases. Whereas low-
8 from diluted solutions in polyols yielded lamellar energy regions appear clearly separated in case i,
crystals of the type shown in Figure 5. Attempts made contour lines from 3 kcal mol-1 upwards enveloping a
to crystallize P2DMLT under similar conditions were continuous range of favored conformations are found for
unsuccessful. The morphological features displayed by case ii. In the latter case, the map becomes symmetrical
the crystals are those typical of polymers crystallizing with respect to the line Ψ ) 180°, as expected from the
in a lattice comprised of chains interlinked by a uni- conformational analogy resulting for opposite values of
directional array of hydrogen bonds.29 They have in Ψ and Ψ ′. Relative energies and dihedral angles
common a thickness of about 60 Å, which is within the calculated for the minimum-energy conformations exist-
40-80-Å range usually observed for polyamide lamellae ing in the low-energy regions found in each case are
with chains arranged in a layered structure.30 Further- listed in Table 3.
more, their width-to-length ratio tends to increase with The low-energy conformations computed in the gas-
the value of n, as reasonably expected for a gradual phase for model compound 1 should be considered only
decreasing in hydrogen-bonding density. as a first approximation to initiate the polymer struc-
The electron diffraction patterns recorded from these ture simulation. Since the crystal field may modify the
crystals were consistent with the triclinic structure relative energy order resulting from such calculations,
established on the sole basis of X-ray data. Moreover, all the conformations included in Table 3 were initially
their great similarities corroborate the occurrence of a taken into account for the construction of the polymer
common side-by-side mode of packing for all the chain.
+ +

Macromolecules, Vol. 29, No. 12, 1996 Crystal Structures of Optically Active Polyamides 4401

Figure 5. Lamellar crystals of PnDMLT isothermally grown

from solution and their corresponding electron diffraction Figure 6. Single-crystal electron diffraction patterns of
patterns. (a) P4DMLT; (b) P6DMLT; (c) P8DMLT. Arrows P6DMLT. (a) Experimentally observed pattern. (b) Electron
indicate the orientation of the crystal relative to the diffraction diffraction pattern simulated for the [012] axis zone of the
pattern. triclinic structure (P1 model).
ø3, ø4, ø2 ′, and ø3 ′) were frozen in an all-trans confor-
Conformation of the Polytartaramide Chain. mation. This approach takes after the γ-form of nylons
The conformational analysis and subsequent model where the C-N bond is known to be rotated, causing
building and simulations were carried out primarily for the polymethylene zigzag to abandon the plane of the
polyamide P6DMLT. In view of the structural analogies amide group.15 By this means, an additional contraction
evidenced by diffraction experiments, the conclusions of around 0.35 Å per amide group is attained. Our
obtained therein may be reasonably extended to P4DMLT results revealed that only the a and g low-energy
and P8DMLT. conformations of the tartaric acid unit were able to
In the first step, the values of the dihedral angles provide a chain repeat compatible with the experimental
given for the seven favored conformations of 1 were used data. The values resulting for the torsion angles about
to build the repeating unit of P6DMLT, and the unit the N-C bonds in these two conformations were re-
heights resulting for each case were compared with the spectively ø1 ) ø1 ′ ) 85° and ø1 ) -ø1 ′ ) 101°. The
experimental value of c ) 13.20 Å determined by X-ray initial tartaric dihedrals were slightly modified also in
diffraction. In order to attain a satisfactory agreement order to achieve an optimum adjustment. The repeating
between calculated and observed data, the dihedral unit of the two conformers are schematically repre-
angles ø1 and ø1 ′ were varied, whereas all other sented in Figure 8, showing that the only remarkable
dihedrals involved in the polymethylene segment (ø2, difference between them is concerned with the relative
+ +

4402 Iribarren et al. Macromolecules, Vol. 29, No. 12, 1996

Figure 8. Schematic representation of the two conformations

favored for the asymmetric unit of P6DMLT corresponding
respectively to conformations a and g of the tartaric unit.
Model a: Ψ ) Ψ′ ) -60°; ø1 ) ø1′ ) 85°. Model g: Ψ ) -Ψ′ )
-60°; ø1 ) -ø1′ ) -101°.

determine the degree of polarity displayed by the crystal

structure.
A further interesting feature of these conformations
which is worthy of comment concerns the gauche effect
between lateral methoxy oxygens. It is known that
glycols with the threo configuration tend to be arranged
with the dihedral angle O-C-C-O near (60° in order
to relieve electronic interactions between neighboring
oxygens.31 The dihedral values calculated for such
sequence in P6DMLT were -43.4° and -63.3° for
models a and g, respectively. In both cases, the interac-
tion between the vicinal OR groups becomes alleviated,
in particular in model g, where the dihedral attains a
value very near to that considered to be a standard
gauche conformation. The same effect has been recently
reported to occur in other related compounds, as is the
case for certain cyclic acetals of L-tartaramides.32
Crystal Structure Modeling and Refinement by
LALS. Crystal models were constructed with LALS by
placing the repeating unit of P6DMLT (2) with the
tartaric moiety in a or g conformation and the remaining
part of the molecule in an arbitrary arrangement, into
the experimentally determined unit cell. The overall
Figure 7. Conformational maps of model compound N,N′- conformation had to be slightly modified in order to
dimethyl-di-O-methyl-L-tartaramide for cases (i) Ψ ) Ψ′ and bring the ends of the repeating unit into coincidence in
(ii) Ψ ) -Ψ′. the space. Space groups C2 and P1 were adopted for
Table 3. Torsional Angles (deg) and Relative Energies conformations a and g, respectively, in agreement with
(kcal mol-1) of the AM1 Low-Energy Conformations for their respective molecular symmetries. In the former
Model Compound 1 case, the unit cell was monoclinic and contained two
case model Ψ υ Ψ′ ∆E chains related by 2-fold rotation and screw axes, whereas
in the second case, the unit cell was triclinic. Although
i a -60 -160 -60 0.0 both conformations appeared to be feasible on stereo-
b 60 40 60 1.3
c -80 0 -80 1.4 chemical grounds, they showed notable differences in
ii d 180 -140 180 0.0 their ability to form hydrogen bonds. The best hydrogen-
e -80 -80 80 0.5 bond geometry attained for the model with chains in
f -20 60 20 1.2 conformation a involved a N-H‚‚‚O angle smaller than
g -60 180 60 2.3 150° and a N‚‚‚O distance larger than 3.3 Å. Such
parameter values turn out to be quite unusual for
orientation of the two amide groups contained in the hydrogen bonding in polyamides.33 Furthermore, they
asymmetric unit. In model a, the carbonyls are pointing are in clear disagreement with infrared data reported
in directions approximately 60° to each other, whereas for polytartaramides,12 which indicated that hydrogen-
they are arranged nearly in a trans conformation in the bonding interactions in these systems are not very
g model. The spatial arrangement adopted by the different in strength to those taking place in conven-
HNCO groups along the chain will play an important tional nylons. On the contrary, all hydrogen bonds
role in the formation of the hydrogen bonds and will could be comfortably made in the model made up of
+ +

Macromolecules, Vol. 29, No. 12, 1996 Crystal Structures of Optically Active Polyamides 4403

Table 4. Observed and Calculated Spacings (dhkl, in Å) Table 5. Unit Cell Parameters, Conformational Angles,
and Structure Factors (F) for the P1 Model of P6DMLT and Hydrogen-Bond Geometriesa for the P1 and C2
Crystal Models of P6DMLT
Fobsd -
spot h k l dobsd Fobsd dcalcd Fcalcd Fcalcd LALS CERIUS
Equator P1 C2 P1 C2
1 0 1 0 5.53 11.48 5.46 10.89 0.59
unit cell parameters
2 1 0 0 4.54
a 5.00 5.00 5.00 5.00
1 0 4.57 19.37 4.55 20.24
b 6.84 12.65 6.84 12.6
-1 -0.87
3 1 1 0 2.95
c 13.20 13.20 13.20 13.20
2 0 4.66 2.94 6.80
61.5 60.0 61.5 60.0
-1 -2.14
4 0 2 0 3.48 2.73 3.85
R
â 90.0 90.0 90.0 90.0
-0.37
5 1 0 2.50 0.89 2.50 1.76
γ 111.6 90.0 111.6 90.0
-2 -0.87
First Layer Line torsional angles
6 0 0 1 11.20 37.07 11.33 42.54 -5.47 Ψ -76.2 -86.5 -50.2 -80.0
7 0 1 1 6.40 10.60 6.35 13.76 -3.16 υ -179.0 177.0 180.0 180.0
8 1 0 1 4.57 Ψ′ 76.2 -86.5 50.2 -80.0
-1 1 1 4.60 29.30 4.57 25.93 3.37 ø1 -131.7 110.3 -101.0 85.0
9 0 -1 1 4.17 19.49 4.15 21.43 -1.94 ø2 ) ø3 ) ø4 ) ø2′ ) ø3′ 180.0 180.0 180.0 180.0
10 1 1 1 3.20 c1′ 131.7 110.3 101.0 85.0
-1 2 1 3.20 14.28 3.19 13.02 1.26 H-bond geometry
11 1 -1 1 2.61 d(H‚‚‚O) 1.92 3.42 1.89 2.52
-1 -2 1 16.47 2.60 12.90 3.57 d(N‚‚‚O) 2.93 3.77 2.90 3.29
Second Layer Line ∠N-H‚‚‚O 173.8 102.5 170.0 146.6
12 0 0 2 5.66 R factor, % 14.5 >50
0 1 2 5.67 13.98 5.63 18.56 -4.58 a Distances in Å and angles in deg.
13 1 0 2 3.98
-1 1 2 4.00 18.46 3.98 20.64 -2.18 Table 6. Cartesian coordinates of one residue of model
14 0 2 2 3.18 P1 of P6DMLT
2 0 2 3.18
0 -1 2 3.20 atom x y z
1 1 2 3.21 N1 -6.895 -7.851 -10.478
1 -1 2 3.22 H-N1 -6.536 -7.619 -11.393
-1 2 2 3.15 9.17 3.22 8.83 0.34 C1 -6.594 -7.210 -9.190
Third Layer Line H1-C1 -7.125 -6.261 -9.123
15 0 1 3 3.27 H2-C1 -6.912 -7.862 -8.376
-1 1 3 3.26 11.45 3.27 12.15 -0.70 C2 -5.078 -6.959 -9.082
H1-C2 -4.758 -6.306 -9.895
Fourth Layer Line H2-C2 -4.546 -7.908 -9.148
16 1 0 4 3.30 6.67 3.30 10.50 -3.83 C3 -4.762 -6.288 -7.732
H1-C3 -5.081 -6.938 -6.918
chains in g conformation. These findings enabled us to H2-C3 -5.294 -5.337 -7.665
discard the model based on conformation a (C2 model) C4 -3.246 -6.035 -7.624
H1-C4
and to consider the triclinic lattice in space group P1
-2.927 -5.383 -8.437
H2-C4 -2.714 -6.984 -7.690
as the preferred model (P1 model) for the crystal C5 -2.931 -5.363 -6.274
structure of P6DMLT. H1-C5 -3.249 -6.015 -5.460
The P1 model was refined against X-ray intensities H2-C5 -3.462 -4.413 -6.207
C6
measured on the fiber diffraction pattern shown in H1-C6
-1.414 -5.111 -6.166
Figure 3a. Sixteen independent spots were used for
-1.095 -4.460 -6.979
H2-C6 -0.882 -6.060 -6.232
refinement, leading to a satisfactory R factor of 14.5%. N2 -1.007 -4.470 -4.907
Observed and calculated spacings and structure factors H-N2 -1.437 -4.843 -4.083
for all the reflections used in the refinement process are CdO2 -0.132 -3.468 -4.877
compared in Table 4. When the C2 model was subjected O2 0.391 -3.018 -5.908
C7 0.153
to similar refinement, an R factor well out of acceptable
-2.947 -3.484
O-C7 0.272 -3.985 -2.630
limits was obtained. A comparative description of H-C7 1.081 -2.374 -3.492
models C2 and P1, including conformational and crystal C8 -1.005 -2.039 -3.027
parameters as well as the geometry of the optimized O-C8 -2.147 -2.758 -2.973
hydrogen bonds, is given in Table 5. The Cartesian H-C8 -1.126 -1.219 -3.735
coordinates for all the atoms included in the asymmetric C9 0.527 -3.521 -1.388
H1-C9 0.625
unit of P6DMLT in the P1 model are given in Table 6.
-4.360 -0.699
H2-C9 1.455 -2.949 -1.396
The three projections (a, b, and c) of model P1 drawn H3-C9 -0.292 -2.879 -1.065
by means of the ORTEP program34 are depicted in C10 -3.162 -1.963 -2.573
Figure 9. H1-C10 -4.083 -2.543 -2.530
H-C10
Features of the crystal structure of P6DMLT which
-2.940 -1.560 -1.585
H-C10 -3.282 -1.143 -3.281
are worth noting are as follows: CdO1 -7.758 -8.842 -10.268
(1) The conformation of the molecule is certainly O1 -8.189 -9.112 -9.137
complex, with the tartaric unit arranged as gauche, the array of 5.0 × 12.6 Å; this accounts for the orthogonality
diamine unit twisted at the HN-CH2 bond, and the observed in the electron diffraction patterns.
polymethylene segment in the all-trans conformation; (3) Hydrogen bonding takes place along the a direc-
the result is a compressed chain describing a wavy tion of the crystal, giving rise to nonpolar sheets
trajectory which follows the c axis of the structure. composed of chains spaced out 5.0 Å; this distance is
(2) Although the formal unit cell is triclinic because significantly larger than that normally found in hydro-
of the lack of symmetry of the molecule in the g gen-bonded sheets of polyamides30 and polypeptides,35
conformation, the chains are positioned in a rectangular which used to be in the range 4.6-4.9 Å.
+ +

4404 Iribarren et al. Macromolecules, Vol. 29, No. 12, 1996

Figure 9. Views of the a, b, and c crystallographic planes of

the triclinic structure (model P1) of polytartaramide P6DMLT.

(4) The tartaric units are located at the same level

within the sheets, but they are regularly shifted along
the b direction of the crystal. This enables the structure
to save some room to accommodate the lateral methoxy
groups in the intersheet space.
The conclusions drawn for P6DMLT may be readily
extended to P4DMLT and P8DMLT provided that the
Figure 10. Compared profiles of experimental (a) and simu-
corresponding enlargement of the axial repeat of the lated (b) powder diffraction pattern of model P1 for poly-
structure is duly taking into account. On the other tartaramide P6DMLT.
hand, the discordant c value recorded for P2DMLT
prevents us from doing the same in this case, although in space group C2 or P1 were applied accordingly.
a similar structure is likely to be adopted by the Stereochemical conclusions similar to those derived from
polymer. Unfortunately, insufficient experimental data the LALS analysis were drawn when contour conditions
could be obtained for P2DMLT for a detailed analysis were imposed. The triclinic unit cell with the chain in
of its structure to be undertaken with acceptable reli- the g conformation appeared to be the favored model
ability. also, and therefore, this was the structure primarily
Modeling and Diffraction Simulation by considered henceforth. The orientation and position of
CERIUS. CERIUS has been reported to be a useful the chain within the lattice was then adjusted in order
tool in the structural study of polymers for which to obtain an optimum correspondence between calcu-
sufficient diffraction data are available and the confor- lated and experimentally observed diffraction patterns.
mation of the asymmetric unit in the gas-phase state The conformation of the chain was only slightly modified
has been previously established.36,37 Since both condi- at this stage. Diffraction patterns with three different
tions are fulfilled in the case of P6DMLT, it was textures, powder, fiber, and single crystal, were used
considered of interest to see how crystal models of this for this purpose.
polymer constructed by CERIUS compare to those (i) X-ray Powder Patterns. In Figure 10, the
generated by LALS. Furthermore, since CERIUS af- experimental X-ray powder diffraction pattern is com-
fords the possibility of simulating the diffraction pat- pared with that simulated for the P1 model. The
terns from a given crystal model, the program may be numerical intensities estimated for the six rings which
used to check the validity of the proposed structures. appear clearly differentiated in the powder diagram of
Firstly, the asymmetric unit of polyamide P6DMLT P6DMLT (shown in Figure 2c) are given in Table 7. A
in either conformation a or g was placed at the origin very good agreement was obtained for both positions
of the unit cell, and the symmetry operations implicit and intensities of the maxima with a satisfactory R
+ +

Macromolecules, Vol. 29, No. 12, 1996 Crystal Structures of Optically Active Polyamides 4405

Table 7. Simulation of the X-ray Powder Diagram of References and Notes

P6DMLT: Comparison between Observedb and
Calculated Intensities for the P1 and C2 Crystal Models (1) Thiem, J.; Bachmann, F. Trends Polym. Sci. 1994, 12, 425.
Icalcd (2) Bird, T. P.; Black, W. A. P.; Deward, E. T.; Hare, J. B. J.
Chem. Soc. 1963, 1208.
ring h k la dobsd, Å Iobsd P1 C2
(3) Ogata, N.; Sanui, K.; Iijima, K. J. Polym. Sci., Polym. Chem.
1 0 0 1 11.50 53 63 100 Ed. 1973, 11, 1095.
2 0 1 1 6.35 68 62 5 (4) Ogata, N.; Hosoda, Y. J. Polym. Sci., Polym. Chem. Ed. 1975,
3 0 1 0 5.40 19 20 44 13, 1973.
4 1 0 0
1 0 1 4.60 100 114 56 (5) Ogata, N.; Sanui, K. J. Polym. Sci., Polym. Chem. Ed. 1977,
5 0 -1 1 4.17 15, 1523.
1 0 2 3.95 92 101 9 (6) Ogata, N.; Sanui, K.; Nakamura, H.; Kuwahara, M. J. Polym.
6 -2 1 0 2.50 7 9 2 Sci., Polym. Chem. Ed. 1980, 18, 939.
a Indexes taken from Table 2. b Spacings measured on the (7) Bachmann, F.; Thiem, J. Makromol. Chem. 1993, 194, 1035.
powder diagram shown in Figure 2c. (8) Hashimoto, K.; Wibullucksanakul, S.; Matsuura, M.; Okada,
M. J. Polym. Sci., Polym. Chem. Ed. 1993, 31, 3141.
factor of 13.4%. A good concordance between calculated (9) Kiely, D. E.; Chen, L.; Lin, T.-H. J. Am. Chem. Soc. 1994,
and observed diffraction powder diagrams is considered 116, 571.
to be prerequisite for further consideration of the (10) Bueno, M.; Galbis, J. A.; Garcı́a-Martı́n, M. G.; De Paz, M.
simulated structure. The agreement attained for the V.; Zamora, F.; Muñoz-Guerra, S. J. Polym. Sci., Part A:
Polym. Chem. 1995, 33, 299.
simulated diffraction pattern generated from the C2
model was much poorer. This was mainly due to the (11) Rodrı́guez-Galán, A.; Bou, J. J.; Muñoz-Guerra, S. J. Polym.
Sci., Polym. Chem. Ed. 1992, 30, 713.
symmetry restraints imposed by space group C2, which
(12) Bou, J. J.; Rodrı́guez-Galán, A.; Muñoz-Guerra, S. Macro-
limited severely the flexibility of the refinement proce- molecules 1993, 26, 5664.
dure.
(13) Bou, J. J.; Iribarren, I.; Muñoz-Guerra, S. Macromolecules
(ii) X-ray Fiber Patterns. The X-ray diffraction 1994, 27, 5263.
pattern of a fiber P6DMLT in the triclinic structure (14) Bou, J. J.; Muñoz-Guerra, S. Polymer 1995, 36, 181.
generated by CERIUS is shown in Figure 3b. The
(15) Kinoshita, Y. Makromol. Chem. 1959, 33, 1.
analogy achieved in both positions and intensities of the
(16) Doak, K.; Campbell, H. N. J. Polym. Sci. 1955, 18, 213.
reflections is excellent. All the equatorial reflections
and layer lines are satisfactorily reproduced. It should (17) Angyal, S. J. In The Carbohydrates; Pigman, W., Horton, D.,
Eds.; Academic: London, 1972; Vol. IA, p 195.
be noted that the appearance of the calculated fiber
diagram appears to be highly sensitive to orientational (18) Zabel, V.; Müller-Fahrnow, A.; Hilgenfeld, R.; Saenger, W.;
Pfannemüller, B.; Enkelmann, V.; Welte, W. Chem. Phys.
effects. Therefore, much care must be taken in defining Lipids 1986, 39, 313.
the orientation factor of the fiber in the simulation (19) Müller-Fahrnow, A.; Hilgenfeld, R.; Heese, H.; Saenger, W.
proccess in order to achieve reliable results. Carbohydr. Res. 1988, 176, 165.
(iii) Electron Diffraction Patterns. Electron dif- (20) Chen, L.; Haraden, B.; Kane, R. W.; Kiely, D. E.; Rowland,
fraction patterns recorded from single crystals of R. S. In Computer Modelling of Carbohydrate Molecules;
P6DMLT contain reflections corresponding to more than French, A. D., Brady, J. W., Eds.; American Chemical
one reciprocal plane. The diagram shown in Figure 6a Society: Washington, DC, 1990; p 141.
is interpreted as arising from a crystal lying on the 001 (21) Chen, L.; Kiely, D. L. Polym. Prepr. (Am. Chem. Soc., Div.
crystal plane. In order to reproduce properly such a Polym. Chem.) 1993, 34, 550.
diagram, electron diffraction patterns were calculated (22) Campbell-Smith, P. J.; Arnott, S. Acta Crystallogr., Sect. A
1978, 34, 3.
by tilting the crystal around the a-axis. The simulated
pattern arising from the [012] axis zone of the crystal (23) CERIUS2 1.6, Molecular Simulations Inc., Burlington, MA.
(tilting conditions: θ ≈ 30°, æ ) 0°) is shown in Figure (24) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J.
6b; the concordance with the experimental pattern is P. J. Am. Chem. Soc. 1985, 107, 3902.
satisfactory for both distribution and intensities of (25) Alemán, C.; Franco, L.; Puiggalı́, J. Macromolecules 1994, 27,
4298.
diffraction maxima.
The conformational parameters and hydrogen-bond (26) Alemán, C.; Bella, J. Biopolymers 1995, 35, 257.
geometry calculated with CERIUS for the P1 model (27) Stewart, J. J. P. MOPAC, QCPE Bull. 1983, 3, 1.
after adjustment and for the crude C2 model have been (28) Bunn, C. W.; Garner, E. V. Proc. R. Soc. London, Ser. A 1947,
included in Table 5 to illustrate how they compare with 189, 39.
those coming out from modeling by LALS. A reasonable (29) Muñoz-Guerra, S.; Prieto, A.; Montserrat, J. M.; Sekiguchi,
H. J. Mat. Sci. 1992, 27, 89.
agreement is found between the results obtained by the
two methods, in particular when differences between (30) Wunderlich, B. Macromolecular Physics; Academic: New
York, 1973; Vol. 1.
models P1 and C2 are considered. It should be noted
that the torsional angles resulting for the model refined (31) (a) Olson, W. K.; Sussman, J. L. J. Am. Chem. Soc. 1982,
104, 270. (b) Olson, W. K. J. Am. Chem. Soc. 1982, 104, 278.
by CERIUS are very close to those introduced in the
(32) Alemán, C.; Martı́nez de Ilarduya, A.; Giralt, E.; Muñoz-
starting conformations. This is mainly due to the fact Guerra, S. submitted.
that small changes in the dihedral angles hardly affect
(33) Cannon, C. G. Spectrochim. Acta 1960, 16, 302.
the appearance of the simulated diffraction patterns.
(34) Jonhson, C. R. ORTEP Report ORNL-3794; Oak Ridge
This is one of the main reasons requiring that an energy National Laboratory: Oak Ridge, TN, 1965.
conformational analysis has to be carried out prior to
(35) Fraser, R. D. B.; MacRae, T. P. Conformation in Fibrous
the modeling with CERIUS. Proteins; Academic: New York, 1973.
Acknowledgment. Financial support to carry out (36) Voigt-Martin, I. G.; Garbella, R.; Schumacher, M. Macromol-
this study given by the CICYT (Comisión Interminis- ecules 1992, 25, 961.
terial de Ciencia y Tecnologı́a, Grant No. MAT93-0555- (37) Voigt-Martin, I. G.; Simon, P.; Yan, D.; Yakimansky, A.;
Bauer S.; Ringsdorf, H. Macromolecules 1995, 28, 243.
CO2-02) is gratefully acknowledged. We are indebted
to CESCA for computational facilities. MA951394V