Polymer 42 (2001) 3489±3502

www.elsevier.nl/locate/polymer

Compatibilization and properties of PBT/PU polymeric alloys

P.S. Archondouli, N.K. Kalfoglou*
Department of Chemistry, University of Patra, 26500 Patra, Greece
Received 18 July 2000; received in revised form 25 September 2000; accepted 16 October 2000

Abstract
Melt mixed blends of poly(butylene terephthalate) (PBT) with a polyester type polyurethane (PU) were characterized by tensile testing,
dynamic mechanical analysis (DMA), differential scanning calorimetry (DSC), optical and electronic microscopy (SEM) and solution 1H
NMR. Blends prepared covered the complete composition range. Large deformation behavior, thermal properties and DMA indicated partial
component mixing and mechanical compatibility typical of a polymeric alloy. Morphology examination revealed good component dispersion
and strong interface adhesion. 1H NMR of blends showed the formation of a PBT±PU copolymer by ester±amide interchange reactions
whose extent depended on the PBT±PU ratio and melt mixing conditions. Evidence for the formation of the copolymer was also provided by
FT-IR. This in situ formed copolymer is responsible for the compatibilization achieved. q 2001 Elsevier Science Ltd. All rights reserved.
Keywords: PBT/PU alloys; In situ compatibilization; Blend characterization

1. Introduction between the ±NH of the PU and the ±CyO of the polyester.
Analogous ®ndings were reported by Ahn et al. [4] on simi-
In a recent work [1] the in situ compatibilization of poly- lar blends. The PU component was based on MDI, BD-1,4
(ethylene terephthalate) (PET) with a polyester type poly- as the hard segment and poly(ethylene adipate) or poly-
urethane (PU) was studied. It was established by 1H NMR (butylene adipate) as the soft segment. Solvent cast ®lms
that at the melt mixing conditions applied, an in situ PET± annealed in the differential scanning calorimeter (DSC)
PU copolymer was formed via ester±amide interchange between 230 and 2708C for various periods of time showed
reactions. This reactive compatibilization accounts for the convergence of the main components' Tgs. This was taken
mechanical, thermal and morphological properties observed as evidence for reactive compatibilization attributed to
in these polymeric alloys. ester±amide interchange reactions.
The present study is an extension of the above work and Very little work on aliphatic polyester/PU blends was
examines the compatibilization of poly(butylene terephtha- found in the literature. Poly(methyl methacrylate)/PU
late) (PBT)/PU blends. PBT is an easily crystallizable main blends consisting of linear homopolymers or prepolymers
chain aromatic polyester and inclusion of PU is expected to capable of forming interpenetrating networks were exam-
improve its impact strength. It was also of interest to exam- ined by Kim et al. [5,6]. In solvent cast ®lms distinct phase
ine whether a reactive compatibilization via a similar separation and component Tg invariance were reported for
mechanism as before would be obtained during melt the linear polymer blends while there was limited Tg conver-
mixing. gence for the interpenetrating system and results from elec-
Related past work on aromatic polyesters/PU blends dealt tron microscopy also indicated incompatibility but with a
primarily with the polycarbonate (PC)/PU pair. In solution ®ner distribution of the dispersed phases.
with cyclohexanone, polyester or polyether type PU was Analogous ®ndings were reported by Kwei, Frisch and
stated to be incompatible [2]. Melt mixing may lead to coworkers [7], who examined the morphology and dynamic
miscible or compatible polymer blends as the work of viscoelastic properties of interpenetrating networks based
Fabri et al. [3] demonstrated for PU elastomers based on on polyacrylate/urethane±urea prepolymers.
4,4 0 -diphenylmethane diisocyanate (MDI) and aliphatic In the present work compatibility characterization was
polycarbonate diols and/or butanediol-1,4 (BD-1,4). Misci- carried out on blends covering the complete composition
bility was attributed to hydrogen bonding interactions range. Techniques applied were tensile testing, DSC,
dynamic mechanical analysis (DMA), optical microscopy,
* Corresponding author. Tel.: 130-61-997-102; fax: 130-61-997-122. FT-IR, electron microscopy (SEM) of cryofractured and
E-mail address: n.kalfog@chemistry.upatras.gr (N.K. Kalfoglou). etched specimens, and 1H NMR.
0032-3861/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0032-386 1(00)00758-8
3490 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

Fig. 1. Phase contrast micrographs of PBT/PU blends: (a) 100/0; (b) 90/10; (c) 75/25; (d) 50/50; (e) 25/75; (f) 10/90.

2. Experimental donated by Bayer A.G. It is a polyester type PU with hard

segments formed by the addition of BD-1,4 to MDI. The
2.1. Materials and specimens preparation soft segment, with Mn of ca. 2000 g mol 21 consists of poly-
ester chains formed by polycondesation of adipic acid and
PBT used was obtained from Nevicolor SpA, Italy. It is BD-1,4. Density was 1.23 g cm 23. PU was dried at 1008C
an injection molding grade with density 1.31 g cm 23 and for 24 h and PBT at 1108C for 5 h, in dynamic vacuo. The
MFI 35 (2508C, g/10 min). PU (Desmopan 359) was dried materials were blended at ca. 2508C under a blanket
 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502 3491

Fig. 2. Optical micrographs with crossed polars of PBT/PU blends: (a) 100/0; (b) 90/10; (c) 75/25; (d) 50/50; (e) 25/75; (f) 10/90; (g) quenched 75/25;
(h) annealed 75/25.
3492 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

of inert gas (Ar) in a home-made stainless steel bob- lar composite morphology was also obtained in PET/PU
and-cup type of mixer previously described [8]. PBT/PU blends [1]. Polarizing microscopy con®rms these ®ndings;
compositions prepared were 90/10, 75/25, 50/50, 25/75, see Fig. 2. The effect of annealing is also shown in Fig. 2(g)
10/90 (w/w). and (h) for one composition, 75/25. The main effect is the
Films were obtained by compression molding between increase of the spherulite size of PBT. The effect of tmix is
Te¯on sheets at ca. 2508C and 5 MPa, followed by quench- revealed in Fig. 3. Increasing tmix results in improved disper-
ing to 08C. sion (Fig. 3(a) and (b)). At increased PU levels a tmix
increase leads to the progressive development of composite
2.2. Apparatus and procedures PBT globules with a concomitant decrease of uniformity in
dispersion; Fig. 3(c)±(e).
Tensile tests were performed at 238C, according to ASTM
D882 at a crosshead speed of 10 cm min 21 using a J.J.
Tensile Tester Type 5001 and ®lm strips measuring 3:00 £ 3.1.2. SEM
0:65 £ 0:025 cm3 :
Additional information on morphology and phase adhe-
DSC measurements were carried out using the DSC
sion is provided by SEM. Cryofractured specimens show a
(SP 1 ) equipped with the Autocool accessory from Rheo-
mixed ductile±glassy fracture at high and medium PBT
metric Scienti®c Co. Sample weight was ca. 10 mg and the
levels; see Fig. 4. This characterizes a high degree of
thermal cycling applied for the crystallinity determination
phase adhesion in a heterogeneous blend. For phase sepa-
of blends was 25 ! 2508C with 208C heating rate, quench-
rated and poorly adhering phases after fracture one would
ing to 2608C and heating up to 2508C with 108C min 21. The
obtain smooth circular craters where the minor component
second heating run was recorded. All the experiments were
was occluded into the matrix. Fracture at ambient tempera-
performed under a constant ¯ow of dry nitrogen.
ture yields a ®brillar texture since both strongly attached
DMA data were obtained at 10 Hz with the RSA II
components may easily elongate: the PU by elastic defor-
mechanical spectrometer from Rheometric Scienti®c Co.
mation, the PBT by cold drawing; see Fig. 5(a)±(c). A simi-
Specimen dimensions were 3:0 £ 0:5 £ 0:01 cm3 :
lar morphology was obtained for PET/PU blends only at
Examination of the effects of possible ester±amide inter-
increased PU levels [1]. This is attributed to the higher Tg,
change reactions was performed by means of high resolu-
PET. Leaching out PU reveals the PBT matrix formed at
tion solution 1H NMR spectroscopy. Spectra were measured
increased PBT levels Ð Fig. 6(a), the interpenetrating
with an Avance DPX 400 MHz with a magnetic ®eld
component distribution at the median composition Ð Fig.
strength of 9.4 T. The samples were dissolved in a 2/1 (v/
6(b) and the spherical PBT composite inclusions at higher
v) mixture of tri¯uoroacetic acid and deuterated chloroform,
PU contents Ð Fig. 6(c).
to obtain 5 wt% solutions.
Optical micrographs with phase contrast and crossed
polar arrangements were obtained with an Olympus BH-2
3.2. Tensile properties
microscope.
SEM was carried out on a JEOL model JSM-500 instru-
The results and their standard deviation of tensile testing
ment. Cryofractured or etched surfaces were examined at a
are summarized in Table 1 in terms of strength s b, yield
tilt angle of 08.
stress s y, elongation % at break e b, and energy to tensile
Selective leaching of PU was carried out using dimethyl-
failure Eb; the last quantity is related to impact tensile
formamide (DMF) (6 days at ambient temperature).
strength.
FT-IR spectra were obtained using a Perkin±Elmer 1600
These ultimate tensile behavior results, especially e b,
spectrometer.
indicate good interphase component adhesion typical of a
polymeric alloy. Signi®cant PBT impact improvement is
3. Results achieved with the addition of 25±50 wt% PU. Table 1 indi-
cates that tmix $ 5 min is adequate though optical micro-
3.1. Morphology scopy indicated improved dispersion at tmix ù 10 min.
Further increase of tmix may be detrimental to tensile proper-
3.1.1. Optical microscopy ties Ð possibly the result of PU decomposition as TGA data
Phase contrast may reveal blend heterogeneity since have shown. Aging affects the properties in PBT-rich
hDPU ˆ 1:550 [9] and hDPBT ˆ 1:587; the latter was calculated compositions. This could be the result of PBT crystallinity
[10] using formulas suitable for semicrystalline polymers. increase that may proceed at room temperature [11].
At positive phase contrast dark areas should correspond to Annealing affects mechanical properties adversely not
PBT. The effect of composition variation is shown in Fig. 1. because of PBT crystallinity increase (see Table 2), but
Increasing PU content leads to a coarser phase distribution because of enhanced PBT spherulitic growth, as optical
and at compositions $50/50, PBT is dispersed as spherical microscopy indicated and because of an increased level of
globules (containing PU inclusions) in a PU matrix. A simi- phase separation; see above.
 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502 3493

Fig. 3. Phase contrast micrographs of PBT/PU blends. Effect of mixing time: 75/25 blend, (a) tmix ˆ 10 min; (b) tmix ˆ 15 min. 25/75 blend, (c) tmix ˆ 5 min;
(d) tmix ˆ 10 min (e) tmix ˆ 15 min.

3.3. Dynamic mechanical properties As seen in Fig. 7, addition of 10 wt% PU leads to stiffen-
ing, possibly the result of intermolecular interaction of the
DMA data are reported at various compositions on two components. On the basis of modulus variation the
quenched blends at tmix ˆ 10 min; see Figs. 7 and 8 for blends are separated into two groups Ð the result of matrix
storage E 0 and loss modulus E 00 , respectively, and for the inversion at the median composition. Fig. 7, inset shows the
75/25 and 25/75 compositions at various mixing times, Figs. modulus variation of blends at two representative composi-
9 and 10, respectively. tions, which were leached to remove PU (75/25 and 25/75).
3494 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

the two components at the interface is evident from the

fact that both primary Tg relaxations, of PBT (at ca. 508C)
and of PU (at ca. 2208C) are present, though converging to
each other. At PU contents $50 wt % a single broad relaxa-
tion is observed Ð the result of merging of two peaks. A
routine deconvolution procedure of these relaxations gave
the component peaks shifted to each other. These are also
reported in Table 2 in parentheses. Fig. 8, inset shows the
DMA spectra of the above leached blends. The spectra show
primary relaxations intermediate between those of pure
components. These ®ndings are further discussed in the
last section. The effect of tmix on DMA spectra is shown in
Figs. 9 and 10 for the 75/25 and 25/75 blends, respectively.
Increasing tmix results in a progressive ¯exibilization of PBT
(see Fig. 9), with a concomitant decrease of Tg Ð see inset
of Fig. 9 on the E 00 variation. Of interest also is the fact that
at low tmix (5 min) the two separate component peaks are
observed and these merge into one at high tmix (.10 min).
This is evidence for an increased degree of compatibiliza-
tion with tmix. These effects are at variance with those
obtained for the 25/75 blend; see Fig. 10. Indeed it is
shown (inset of Fig. 10), that at increased tmix the blend
shows a lower degree of dispersion since the intermediate
single peak moves to lower temperatures (ca. 2208C at
t mix ˆ 10 min†: These ®ndings are generally in line with
the phase-contrast microscopy, see Fig. 3, in that when
two relaxations appear the phase distribution becomes
coarser. Data on the main relaxations obtained for these
blends as well as for annealed samples are summarized in
Table 2, where thermal data have also been included. For
the PU-rich blend annealing leads to considerable E 0 rein-
forcement; spectra are reported elsewhere [12]. As E 00 varia-
tion and Tg component shifts indicate (see Table 2), annealing
promotes phase separation.

3.4. Thermal properties

Thermal data obtained during the second heating scan are

reported in Table 2. Depending on heating rate and/or ther-
mal history pure PBT gives two close-lying melting
endotherms [13]: a very small one at ca. 2108C preceding
the main endotherm that is obtained by recrystallization
during the second heating scan. This main Tm and the asso-
ciated degree of crystallinity XC as well as the crystallization
temperature Tc are reported in Table 2. These endotherms
are also obtained in annealed blends. Since crystal reorga-
nization is hindered by PU the two endotherms observed are
further apart. PU in quenched blends crystallizes slowly;
therefore, during the second heating only the Tm,PBT is
Fig. 4. SEM of cryofractured surfaces of PBT/PU blends: (a) 75/25; (b) 50/ observed. Since in 25/75 and 10/90 blends the Tm observed
50; (c) 25/75. was close to that of pure PU, thermal measurements were
also made on specimens leached to remove PU. The results
The stiffness is quite different from that of PBT and re¯ects indicated that the endotherms originally observed belong to
the formation of a modi®ed polymer via compatibilization. PBT. In general, crystallization of PBT, as bulk crystallinity
This is also evident in the corresponding spectra of E 00 XC indicates, is retarded with increasing PU content Ð a
reported in Fig. 8, inset; see below. In Fig. 8 mixing of result expected in partially miscible polymer alloys. An
 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502 3495

Fig. 5. SEM of PBT/PU blends fractured at ambient temperature: (a) 75/25;

(b) 50/50; (c) 25/75. Fig. 6. SEM of etched PBT/PU blends: (a)75/25; (b) 50/50; (c) 25/75.

exception is noted for the PU-rich blends where PBT is component. Table 2 also shows the effect of annealing of
phase separated and its crystallization is less hindered. a PU-rich blend and of tmix in lowering Tm,PBT. Increasing tmix
Except for pure PU and the annealed PU-rich blend reduces Tg,PBT while Tg,PU cannot be clearly de®ned (see
(25/75), all Tm reported in Table 2 refer to the PBT Fig. 10, inset). A signi®cant Tm,PBT and Tc,PBT reduction is
3496 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

Table 1
Tensile properties of quenched blends

PBT/PU blends Mixing temperature (8C) s y (MPa) s b (MPa) e b (%) DL/L0 e b (%) ln(L/L0) Eb (J cm 23)

tmix ˆ 10 min
100/0 ± 43 ^ 5 28 ^ 3 102 ^ 43 69 ^ 19 33 ^ 9
90/10 250 33 ^ 3 23 ^ 2 171 ^ 58 98 ^ 24 43 ^ 6
75/25 250 27 ^ 3 36 ^ 4 670 ^ 108 204 ^ 14 161 ^ 45
75/25 a 250 ± 35 ^ 2 24 ^ 6 21 ^ 5 7^1
75/25 b 250 32 ^ 3 24 ^ 5 363 ^ 69 153 ^ 16 76 ^ 15
50/50 250 ± 36 ^ 3 559 ^ 61 188 ^ 10 172 ^ 44
25/75 250 ± 29 ^ 2 534 ^ 58 184 ^ 9 89 ^ 10
25/75 a 250 ± 25 ^ 2 235 ^ 48 120 ^ 14 47 ^ 11
25/75 b 250 ± 29 ^ 2 447 ^ 26 170 ^ 5 88 ^ 9
10/90 250 ± 28 ^ 3 511 ^ 21 181 ^ 3 81 ^ 8
0/100 ± ± 56 ^ 5 807 ^ 130 220 ^ 14 222 ^ 22
tmix ˆ 5 min
75/25 250 27 ^ 2 39 ^ 4 701 ^ 90 207 ^ 12 167 ^ 22
25/75 250 ± 33 ^ 4 522 ^ 64 182 ^ 11 113 ^ 9
25/75 a 250 ± 27 ^ 5 297 ^ 80 137 ^ 19 53 ^ 15

tmix ˆ 15 min
75/25 250 19 ^ 3 30 ^ 3 557 ^ 65 188 ^ 10 92 ^ 21
25/75 250 ± 30 ^ 4 492 ^ 63 177 ^ 11 96 ^ 24
25/75 a 250 ± 29 ^ 3 449 ^ 32 170 ^ 6 121 ^ 16
tmix ˆ 20 min
75/25 250 11 ^ 1 23 ^ 3 440 ^ 60 168 ^ 11 57 ^ 15
a
Annealed at 1408C for 30 min.
b
After 20 days physical aging at 258C.

also obtained. At PU-rich compositions and increased tmix 3.5.1.2. Molecular interactions. Of the various absorption
(15 min) no Tm endotherms are observed. In general both Tm bands associated with inter- or intramolecular interactions,
and Tc data obtained with increasing tmix indicate a progres- the hydrogen bonding associated with the ±NH stretching
sively less perfect crystal structure due to higher degrees of vibration at 3332 cm 21 is of relevance to the present study
compatibilization and are in agreement with FT-IR spectral (see Fig. 12). Hydrogen bonding also involves the
changes at increased tmix; see below. group from the soft segment and the group from
the hard segment of PU. The non-associated ±NH group
3.5. Spectroscopy absorbs at ca. 3450 cm 21. Inspection of Fig. 12 shows that
in PU practically all of the ±NH± groups are hydrogen
3.5.1. FT-IR bonded. Addition of PBT disrupts the hydrogen bonded
Spectra were obtained for various compositions and for structure yielding free ±NH± groups. The net result is a
the 75/25 blend at various tmix. shift of the maximum to a higher frequency 3332 !
The results are analyzed in terms of PBT crystal transfor- 3352 cm21 †: The width of the absorption band in blends is
mation and intermolecular interactions. attributed to the variety of the associated species involved,
with the
3.5.1.1. Effect on crystal structure. PBT crystallizes in two ester group of PBT. Finally the reduction in strength of
forms [14]: the a-crystal with a g±t±g conformation and the this band is a result of dilution caused by the addition of
b-crystal (formed under stress) with a t±t±t conformation. polyester.
Most characteristic absorptions for these crystal forms were Ester±amide interchange reactions that could conceiva-
observed and are reported elsewhere [12]; e.g. Fig. 11 shows bly take place cannot be studied in these blends since in pure
the spectrum for the a-crystal (combination of skeleton and PU the characteristic bands of ±CyO and ±NH± are broad:
rocking vibrations, at 916 cm 21) and of the b-crystal at 1675±1760, 1500±1560 and 1180±1220 cm 21 for the
965 cm 21. In all cases [12] increasing tmix reduces the Amide I, II and III conformations, respectively, causing
intensity of the a-crystal absorptions. This may be the coupled vibrations [16]. The PBT ±CyO absorption and
result of destruction of the packing in the a-crystals and/or the ±C±O± stretching vibration at 1700±1750 and 1000±
the elimination of trans and high energy (A) bond rotations 1120 cm 21, respectively, are also broad. Thus in blends any
[15]. shift due to molecular interactions would not be identi®able.
 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502 3497

Table 2
Viscoelastic and thermal properties of blends

PBT/PU blends High Tg (8C) Low Tg (8C) Tm (PBT) (8C) Tc (PBT) (8C) Xc (PBT) (%) a

tmix ˆ 10 min
100/0 50 ± 224 191 22
90/10 47 ± 221 182 19
75/25 45 ± 221 178 20
75/25 c 48 ± 213, 221 179 21
75/25 d e e
219 178 25
50/50 30 f ± 200 168 g

(28) (13)
25/75 ± 27 f 217 165 5
(19) (9)
25/75 c 46 218 196 (PU) 177 12
210, 220
10/90 ± 213 f 219 177 11
(14) (11)
0/100 ± 220 201, 219 137, 114 26 b

tmix ˆ 5 min
75/25 47 29 216 160 21
25/75 ± 22 f 219 161 13
tmix ˆ 15 min
75/25 38 ± 194 142 18
25/75 ± 0f ± ± ±
25/75 c ± 2f ± ± ±
tmix ˆ 20 min
75/25 36 ± 187 149 18
Etched samples
75/25 33 f ± 217 175 28
25/75 50 19 219 158 6
a
DHf0 PBT† ˆ 33:44 cal g21 [11].
b
DHf0 PU† ˆ 5:83 cal g21 [19].
c
After annealing at 1408C for 30 min.
d
After 20 days physical aging at 258C.
e
Not examined.
f
Broad relaxation.
g
Merging of the two endotherms.

To obtain unequivocal information on possible intercompo- progressively higher mixing times, one detects an additional
nent interaction, compositions 75/25 and 25/75 were exam- peak at 7.78 ppm when tmix $ 10 min. The amount of copo-
ined after prolonged extraction of PU. The spectra, Figs. 13 lymer formed was estimated to be 3 and 5 wt% for (c) and
and 14, showed the characteristic absorptions at 1535 and (d), respectively. This was not observed for the 25/75 blend
1600 cm 21 present in PU and attributed to Amide II confor- at similar tmix (Fig. 15(v) and (vi)).
mation and CyC stretching vibration of the aromatic ring, The new peak has been attributed to the change of envir-
respectively. These results are in line with the DMA and onment of the aromatic H due to the formation of a PU±PBT
NMR measurements indicating mixing at the interphase and copolymer Ð the result of isocyanate groups from PU
possible copolymer formation; see below. chemically reacting with terminal carboxyl groups of
PBT; see below. Similar ®ndings were reported by Pillon
3.5.2. 1H NMR and Utracki [17] who studied melt mixed PET/PA-66 blends
In the previous study [1] NMR gave evidence for the in the presence of a catalyst. As to why the new additional
formation of a copolymer and a possible mechanism was peak is not observed in PU rich blends a possible explana-
proposed based on ester±amide interchange reactions. Simi- tion is given below.
lar results were obtained in the present work. In Fig. 15(i) of
the pure PBT spectrum the single peaks at 7.74 and 4.12,
1.64 ppm, correspond to the aromatic (a) and aliphatic (b, c) 4. Discussion
protons, respectively. Focusing on the aromatic H in the
spectrum of the 75/25 blend, (Fig. 15(ii)±(iv)), taken at Static and dynamic mechanical properties as well as
3498 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

Fig. 7. Temperature dependence of storage modulus E 0 of PBT/PU blends: (B) 100/0; ( ) 90/10; (K) 75/25; (X) 50/50; (L) 25/75; (O) 10/90; (S) 0/100.
Inset Ð storage modulus of leached blends: (Ð) 75/25; (± ±) 25/75.

Fig. 8. Temperature dependence of loss modulus E 00 of PBT/PU blends:(B) 100/0; ( ) 90/10; (K) 75/25; (X) 50/50; (L) 25/75; (O) 10/90; (S) 0/100.
Inset Ð storage modulus of leached blends: (Ð) 75/25; (± ±) 25/75.
 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502 3499

Fig. 9. Effect of mixing time on storage modulus of 75/25 blends: (-O-) t ˆ 5 min; (-W-) t ˆ 10 min; (-A-) t ˆ 15 min; (-X-) t ˆ 20 min. Inset Ð loss modulus.

Fig. 10. Effect of mixing time on storage modulus of 25/75 blends: (-W-) t ˆ 5 min; (-A-) t ˆ 10 min; (-O-) t ˆ 15 min.
3500 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

Fig. 11. FT-IR spectrum of 75/25 PBT/PU blend. Effect of mixing time.

thermal data combined with morphological features support

Fig. 13. FT-IR spectrum of leached 25/75 PBT/PU blend.
the view that at the mixing conditions employed polymeric
alloys of PBT/PU are obtained. Additional DMA data in
blends where PU was selectively removed, see Fig. 8, samples and NMR give evidence for the formation of a
inset, give evidence for the formation of a blend with relaxa- PBT±PU copolymer. Its presence facilitates component
tions intermediate to those of PBT and PU; 338C for the 75/ dispersion and subsequent reactive compatibilization when
25 and 19 and 508C (shoulder) for the 25/75 blend. If physi- adequate mixing time is employed. In addition, interactions
cal intermolecular interactions were responsible for the between the carbonyl groups of PBT and the imidic
relaxations shifts observed in blends, in etched blends one hydrogen of PU may contribute since during melt mixing
would recover PBT with a main relaxation at 508C. Thus
some chemical bonding between partners is involved. More-
over, spectroscopic evidence based on FT-IR of leached

Fig. 12. FT-IR spectra of PBT/PU blends. Effect of composition. Fig. 14. FT-IR spectrum of leached 75/25 PBT/PU blend.
 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502 3501

Fig. 15. NMR spectra of PBT/PU blends. Effect of composition and mixing time: (i) pure PBT; 75/25 blend, (ii) tmix ˆ 5 min; (iii) tmix ˆ 10 min; (iv) tmix ˆ 15 min;
25/75 blend, (v) tmix ˆ 10 min; (vi) tmix ˆ 15 min.

FT-IR results indicated disruption of the hydrogen bonded leached blends gives evidence for a copolymer formation
PU. Ð the result of reactive compatibilization by both routes I
The sequence of reactions that may take place is summar- and II.
ized in Scheme 1. Reaction path (a) is reversible and takes A referee suggested that transesteri®cation reactions
place during the initial stage of melt mixing. The R , NCO between PBT and the polyester soft segment phase of the
chains that result from thermal dissociation may also parti- PU may be possible, if a residual catalyst (used for the manu-
cipate in reaction paths I and II involving the terminal facture of PBT) is present. Though this possibility cannot be
active groups of PBT. Reaction path I is slow and irrever- ruled out we do not have spectroscopic evidence to support it.
sible, involving the terminal carboxyl of PBT. This reaction
leads to the modi®cation of the NMR spectrum of the
aromatic protons of PBT observed and is expected to 5. Conclusions
yield measurable concentrations of copolymer I only at
PBT-rich compositions, v.a. Reaction II involving the term- 1. Polymeric alloys with improved mechanical properties
inal hydroxyl of PBT is much faster [18] and produces are obtained by reactive melt blending of PBT with
copolymer II, which does not lead to modi®cation of the thermoplastic PU.
NMR spectrum. Given the preference of the R , NCO to 2. DMA and spectroscopic evidence support the view that
react via route II, it is reasonable to expect that at low PBT during melt blending ester±amide interchange reactions
ratios in blends, most of it is consumed to form type II take place, forming a PBT±PU copolymer. The latter
copolymer. Thus the concentration (if any) of copolymer I promotes good component dispersion and interphase
is too low to detect by NMR. However the FT-IR of the bonding via intermolecular reactions.
3502 P.S. Archondouli, N.K. Kalfoglou / Polymer 42 (2001) 3489±3502

Scheme 1.

[4] Ahn TO, Jung S, Lee J, Jeong HM. J Appl Polym Sci 1997;64:2363.
[5] Kim SC, Klempner D, Frisch KC, Radigan W, Frisch HL. Macro-
Acknowledgements molecules 1976;9:258.
[6] Kim SC, Klempner D, Frisch KC, Frisch HL. Macromolecules
The authors thank Profs J. Kallitsis, K. Gravalos and Dr 1976;9:263.
[7] Matsuo M, Kwei TK, Klempner D, Frisch HL. Polym Engng Sci
K. Samios for useful discussions. Thanks are also due to
1970;10:327.
Nevicolor SpA and Bayer A.G. for providing materials [8] Papadopoulou CP, Kalfoglou NK. Polymer 1997;38:4207.
and information. This work was supported in part by the [9] Seymour RB, Carraher CE. Structure±property relationships in poly-
Operational Programme for Education and Initial Voca- mers. New York: Plenum Press, 1984 (chap. 4).
tional Training, Programme on Polymer Science and Tech- [10] van Krevelen DW. Properties of polymers. New York: Elsevier, 1972
(p. 200).
nology-3.2a, 33H6, administered jointly by the Ministry of
[11] Illers KH. Colloid Polym Sci 1980;258:117.
Education in Greece and the European Community [12] Archondouli P. MSc thesis, in preparation.
(Programme JOULE 1998-2000, PL-970014). [13] Hobbs SY, Pratt CF. Polymer 1975;16:462.
[14] Hopfe I, Pompe G, Eichorn K-J. Polymer 1997;38:2321.
[15] Stambaugh BD, Koenig JL, Lando JB. Polym Lett 1977;15:299.
References [16] Bellamy LJ. The infra-red spectra of complex molecules. 2nd ed.
London: Methuen, 1958 (chap. 120).
[1] Samios CK, Gravalos KG, Kalfoglou NK. Eur Polym J 2000;36:937. [17] Pillon LZ, Utracki LA. Polym Engng Sci 1984;24:1300.
[2] Peterson RJ, Corneliussen RD, Rozelle LT. Polym Prepr (Am Chem [18] Ravve A. Organic chemistry of macromolecules. New York: Marcel
Soc, Div Polym Chem) 1969;10:385. Dekker, 1967 (p. 295).
[3] Fambri F, Penati A, Kolarik J. Polymer 1997;38:835. [19] Papadopoulou CP, Kalfoglou NK. Polymer 1998;39:7015.