Synthesis and flame retardant testing of new boronated and phosphonated aromatic compounds

Article abstract of DOI:10.1039/c1jm14682c

The present report describes the preparation and use of some dimethyl terephthalate derivatives in transition metal-catalyzed coupling reactions to produce new reactive flame retardants. Dimethyl iodoterephthalate and dimethyl 2,5-diiodoterephthalate were

Full text of DOI:10.1039/c1jm14682c

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 1180
www.rsc.org/materials
PAPER
Synthesis and flame retardant testing of new boronated and phosphonated
aromatic compounds
a
Vladimir Benin, Sravanthi Durganala and Alexander B. Morgan
b
c
*
*
Received 20th September 2011, Accepted 19th October 2011
DOI: 10.1039/c1jm14682c
The present report describes the preparation and use of some dimethyl terephthalate derivatives in
transition metal-catalyzed coupling reactions to produce new reactive flame retardants. Dimethyl
iodoterephthalate and dimethyl 2,5-diiodoterephthalate were successfully employed in the preparation
of phosphonic and boronic esters and acids. The latter were tested for heat release with
a microcombustion calorimeter (ASTM D7309) to determine the potential for heat release reduction of
these flame retardant molecules. The results showed that the addition of boronic or phosphonic acids
greatly lowered the heat release, due to a condensed phase (char formation) mechanism. Adding ester
groups to the boronic acids or phosphonic acids could either completely remove all flame retardant
effects or make the molecule act more like a vapor phase flame retardant. Finally, the various potential
flame retardants were solvent blended with a thermoplastic polyurethane to assess their flammability
reduction effects by microcombustion calorimetry. The results of these experiments found that the
molecules that reduced heat release the most by themselves showed the greatest reduction in heat
release in a polyurethane as well, with the boronic acids yielding the greatest reduction in heat release.
rather than the polymer mass being pyrolyzed as high heat
Introduction
release decomposition products.3,6 To that end, we have
proposed the synthesis of boron and phosphorus functionalized
reactive molecules which could co-polymerize with thermoset-
type polymers such as epoxy and polyurethane. Boron has shown
some interesting condensed phase activity when available in
a boronic acid/boroxine structure,14,15 and phosphorus-based
structures are well known for creating char in polymers con-
taining heteroatoms such as nitrogen and oxygen in the polymer
structure.6,16 Due to the efficacy of boron and phosphorus in
flame retardancy, we have focused on the use of catalytic
processes to add boron and phosphorus based groups to
aromatic compounds such as terephthalates and phenols. This
has been done since these compounds could in turn act as reac-
tive flame retardants, which could address some of the environ-
mental/leaching issues that affect flame retardant materials.
Establishing the chemical structure–property relationships
prior to scale-up of a new flame retardant is very important, and
goes back to the case of considering reactive vs. additive flame
retardants. Making reactive flame retardants is not a trivial
activity, since those same reactive groups may lead to short shelf
lives for the flame retardants or high cost of synthesis due to
protection/deprotection schemes needed in the synthesis of those
molecules. Therefore, knowing what to scale up is of as much
importance as discovering a possible compound for flame
retardant evaluation. If funding is limited for discovery or scale-
up, but not both, one needs good measurement tools to ensure
the right molecule is made for evaluation. This becomes difficult
While modern polymeric materials bring many benefits by their
use in society, they typically suffer from flammability/fire risk
issues, and so they may be flame retarded to provide fire
protection in a variety of fire risk scenarios. Wood is certainly
a flammable material, but there are commodity plastics in use
today in furniture, consumer goods, and clothing that have heat
releases and fire risk potentials far greater than those of cellulosic
matter.1–5 When a polymeric material requires a flame retardant
to be used/sold in a particular application, one can use a non-
reactive flame retardant molecule/polymer, which is blended into
the polymer, or one can use a reactive flame retardant molecule,
which bonds directly to the polymer during polymerization or via
side-chain/grafting reactions.6,7 With concerns about environ-
mental persistence of some flame retardant additives that are not
bound to the polymer and may leach out over time,8–13 the use of
reactive flame retardants becomes more attractive. Further, there
is a desire to develop condensed phase (char forming) reactive
flame retardants, so more of the polymer fuel can be converted
into low-flammability carbon char (graphite/glassy carbon)
aDepartment of Chemistry, University of Dayton, Dayton, OH,
45469-2357, USA. E-mail: vladimir.benin@notes.udayton.edu
bChemtura Corporation, 1801 US Highway 52 W, West Lafayette, IN,
47906
cAdvanced Polymers Group, University of Dayton Research Institute,
Dayton, OH, 45469-0160, USA. E-mail: alexander.morgan@udri.
udayton.edu
1180 | J. Mater. Chem., 2012, 22, 1180–1190
This journal is ª The Royal Society of Chemistry 2012

View Online
Scheme 1
when the success of flame retardancy is determined by larger
scale regulatory pass/fail tests, and so a much smaller scale test is
needed to quantify the effectiveness of the flame retardant prior
to scale-up. There is a wealth of organic chemistry available
which can identify how to make the target flame retardant
molecule, but without confirmation from a small-scale test, one
may have to expend large amounts of resources to make enough
material for the regulatory pass–fail tests if the chemistry is
difficult or low-yielding. Therefore we believe that studying
flammability of small molecules themselves, followed by simple
experiments of mixing the potential flame retardants into
a polymer, is a reasonable approach to take for identifying and
screening flame retardant potential prior to scale-up, and that is
the key concept behind this paper.
combustion flow calorimeter (PCFC), a small scale standard
method which mimics flammability phenomena of pyrolysis and
oxidation quite well.17–21 In order to compare the charring and
lowered heat release potential of these new structures, we also
measured the heat release of terephthalic acid, dimethyl tere-
phthalate, and hydroquinone, which served as non-flame retar-
dant standards, thus allowing us to study structure/property
relationships prior to scale-up and incorporation of these mole-
cules into polymers for larger scale flammability testing. We then
combined these flame retardants into a thermoplastic poly-
urethane (TPU) which could serve as a small-scale mimic for
a polymer with a high need for new flame retardant chemistry,
polyurethane. Because polyurethane foams can be difficult to
synthesize at a small scale, trying a small scale mimic test makes
them an excellent candidate for this study. More importantly
though, polyurethanes, especially flexible polyurethane foams,
represent a notable fire threat in the home which requires new
flame retardants that lower heat release while enhancing char
In the current report, we wish to provide some details of their
use in the generation of phosphonic and boronic esters and acids.
The target phosphonic and boronic esters and acids were tested
for heat release/char formation potential with a pyrolysis
Scheme 2
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 1180–1190 | 1181

View Online
Fig. 1 Structures of several flame retardants, studied by pyrolysis combustion flow calorimetry (PCFC).
formation to slow down flame spread and time to flashover (total
fire loss) conditions.22–24
in ethanol as a solvent, in the presence of base (triethylamine or
dicyclohexylmethylamine) and Pd(OAc)2/Ph3P.
Iodoterephthalate 1 proved to be also an excellent starting
material for the introduction of a boronic ester moiety. In
particular, the cyclic boronic ester 3 was prepared in good yield,
using (dppp)2NiCl2 as a catalyst (Scheme 1).29 The reaction was
carried out in the presence of triethylamine or dicyclohexyl-
methylamine as a base, leading to the product in similar yields.
Finally, pinacol boronate 3 was hydrolyzed in acidic conditions
(3 M HCl) to yield the boronoterephthalic acid 4. Earlier
attempts to prepare 4 from 2-bromo-p-xylene, using Grignard
chemistry followed by permanganate oxidation, were not
Results and discussion
The research was composed of three parts. The first was the
synthesis of potential reactive flame retardants. The second part
was studying these potential flame retardants by themselves via
PCFC to see what their potential for heat release and char
formation would be. Finally, these substances were combined
with a thermoplastic polyurethane (TPU) and tested again via
PCFC to see if the findings from the 2nd part justified the
synthesis of the materials in the 1st part. If successful, this
approach would begin to help solidify a testing scheme and
screening criteria for new potential flame retardants, thus telling
the chemist what to focus on from a resource perspective.
Table 1 PCFC summary data for compounds 2–4, and 9–14
HRR HRR
Char peak
yield value/ temp./ HR/ % THR
Compound (wt%) W gꢀ1 ꢁC kJ gꢀ1 Reduction description
peak
Total
Char
Part I. A. Preparation of monophosphonic and monoboronic
esters/acids
10
4
0.0 450
0.0 441
0.0 501
41.2 181, 89 355, 481 9.6
41.5 160, 91 355, 483 9.4
35.3 180, 83 355, 481 9.6
378
382
380
19.3
19.4
19.4
0
0
0
None visible
Several structures containing a single phosphonic or boronic
ester (or acid) functionality were prepared using dimethyl iodo-
terephthalate 1. Although the latter has been mentioned in the
scientific literature,25 no particular reference provides an explicit
protocol for its preparation and characterization. We prepared it
in a single step, from commercially available dimethyl amino-
terephthalate (Scheme 1). The latter was diazotized in a H2O/
H2SO4 mixture, followed by treatment with aqueous KI solu-
tion.26 Compound 1 was subjected to transition metal-catalyzed
coupling reactions in order to introduce phosphonic or boronic
ester functionality.
50.5
51.5
50.5
96.4
96.4
96.4
77.8
78.9
78.9
0
1/2 pan of
black foamy
residue
Dark brown
powder
9
66.1 5, 21
66.0 5, 20
65.9 5, 21
43.2 69
44.6 65
45.1 59
0.0 613
0.0 611
0.0 619
2.0 710
3.2 838
2.6 782
0.0 402
0.0 390
0.0 399
0.0 44, 345 175, 312 18.1
0.0 44, 333 183, 315 18.4
0.0 41, 341 184, 314 18.3
0.1 652
0.1 712
0.0 684
42.19 817
48.11 875
38.17 837
365, 407 0.7
365, 406 0.7
365, 407 0.7
11
12
3
475
477
475
247
249
248
295
292
294
326
323
324
4.3
4.1
4.1
16.8
22.7
19.5
Pan full of
black shiny
foamy residue
None visible
0
0
23.3 ꢀ18.9
22.9 ꢀ16.8
22.6 ꢀ15.3
Black shiny
oil film
on bottom
Very thin
gray film
The preparation of phosphonic esters was successfully
accomplished following one of two general strategies, utilizing
a trialkyl- or a dialkyl phosphite correspondingly (Route 1 and
Route 2, Scheme 1). Thus trimethylphosphite and triethylphos-
phite were successfully employed in the preparation of
compounds 2a,b (Route 1). The protocol was adapted from
a patented work of Hashiba et al. and involves reaction of 1 with
a large excess of the trialkyl phosphite (also used as a solvent in
these conditions), at elevated temperature, in the presence of
PdCl2 as a catalyst.27
2a
2b
13
14
17.9
17.5
17.6
8.7
10.7
10.2
7.7
6.1
6.6
0
0
0
63.3
66.5
64.1
on bottom
None visible
261
259
259
321
319
320
24.4
24.5
24.6
9.0
None visible
Pan full
8.2
8.8
of black
shiny foamy
residue
The generation of phosphonate esters with dialkylphosphites
as starting materials was accomplished following the procedure
of Dezfuli and Goosen (Route 2).28 The reactions were conducted
1182 | J. Mater. Chem., 2012, 22, 1180–1190
This journal is ª The Royal Society of Chemistry 2012

View Online
successful. We demonstrated that under basic conditions a novel
dimeric sodium salt of 4 was isolated, instead of the free acid.30
although in both cases significant amounts of byproducts were
generated. Method (1) yielded consistently mixtures of the target
8, the monoboronic ester 3, and dimethyl terephthalate, i.e.
byproducts of hydrodeboronation. Following Method (2) the
main impurity was a highly insoluble, likely polymeric material,
whose formation could speculatively be based on a Suzuki
coupling reaction between the starting diiodide 7 and the
product 8.
Part I. B. Preparation of diboronic esters/acids
p-Xylene was diiodinated by using periodic acid/iodine, in
a mixture of acetic acid, sulfuric acid and water, to yield 2,5-
diiodo-p-xylene 5 (Scheme 2).31 Compound 5 was subjected to
oxidation, using KMnO4 in a t-butanol/water solvent mixture,
followed by acidification, to yield 2,5-diiodoterephthalic acid 6.
Fischer esterification of the acid in excess methanol, using
sulfuric acid as a catalyst, generated the dimethyl ester 7.
Compound 7 was then subjected to transition metal catalyzed
coupling, utilizing two different methods: (1) use of pina-
colborane, with (Ph3P)2PdCl2 as a catalyst and Et3N as a base, in
toluene,29 or (2) use of bis(pinacolato)diboron, with
(dppf)2PdCl2 as a catalyst and KOAc as a base in DMF. The
cyclic diboronic ester 8 was generated following either route,
Diboronic ester 8 was successfully hydrolyzed to 2,5-dibor-
onoterephthalic acid 9, using conditions similar to those
employed in the preparation of 4 (3 M HCl).
Part II. Thermal analysis and micro combustion calorimeter
studies
To understand the ability of these molecules to serve as flame
retardants, we chose to study their inherent heat release via the
use of a microscale combustion calorimetry technique. This
Fig. 2 HRR curves for terephthalic acid (a), boronoterephthalic acid 4 (b), 2,5-diboronoterephthalic acid 9 (c), and phosphonoterephthalic acid 11 (d).
This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 1180–1190 | 1183

View Online
technique, known as pyrolysis combustion flow calorimetry
(PCFC) is an ASTM standard method for measuring the heat
release of organic materials via oxygen consumption calorimetry.
The method is a standard technique under ASTM D7309, and
has been successfully used to measure the heat release (inherent
flammability) of a wide range of polymers, flammable solids and
liquids, and small molecules.17–19,21,29 In this study we investi-
gated the target molecules described above (structures 2–4 and 9,
Fig. 1), using the PCFC method. Also included in the study were
2-phosphonoterephtahlic acid 11, whose preparation we recently
reported,32 and hydroquinone 14, with a cyclic phosphonate ester
functionality.33 It was found that a significant portion of the
studied structures did show reductions in their inherent flam-
mability (heat release) but the amount of reduction is very
chemical structure-dependent.
Typical results from the PCFC focus on heat release
measurements and the results that were recorded from each of
the materials are shown in Table 1. The data in the table cover
the following measurements:
Fig. 4 DSC curve of 2,5-diboronoterephthalic acid 9.
structure (compare compounds 4 and 9 to structure 10) we can
see that the addition of boronic acid greatly increases char yield
and results in significant reductions in heat release. However, we
see a two peak heat release behavior for these materials when
compared to the control sample (compare Fig. 2b and c to
Fig. 2a) and the heat release rate profile for the monoboronic
acid 4 is very different from that seen for the diboronic acid 9
(compare Fig. 2b and c).
ꢂ Char yield: obtained by measuring the sample mass before
and after pyrolysis. The higher the char yield, the more carbon/
inorganic material left behind. As more carbon is left behind, the
total heat release should decrease.
ꢂ HRR peak(s): the recorded peak maximum of heat release
rate (HRR), found during each experiment. The higher the HRR
value, the more heat given off at that event. This value roughly
correlates to the peak heat release rate that would be obtained by
a cone calorimeter. The temperatures at which the peak HRR
was detected are listed as well.
The likely reason for this different heat release behavior is the
formation of boroxine structures as the boronic acids are heated.
Boronic acids are well known to form boroxine structures
around 180–200 ꢁC, and once this network is in place, it has
a strong tendency to char rather than burn.14 Very likely, the
monoboronic acid 4 forms a single boroxine, and then, when this
is heated further, the carboxylic acid groups become involved,
which in turn creates a complex boron oxycarbide structure
(Fig. 3). The first peak of HRR is probably from some initial
combustion of the monomeric boroxine, and the 2nd peak of
HRR is probably from thermal decomposition of the boron
oxycarbide as it forms.
ꢂ Total HR: this is the total heat release (HR) for the sample,
which is the area under the curve(s) for each sample analysis. The
% reduction in total heat release is also given.
ꢂ Char notes: description of the sample residues collected from
each test.
In Table 1, compounds 10, 12 and 13 are control samples as
they represent the molecular cores, without the flame retardant
(boron- or phosphorus-containing) functionalities. From the
data we can clearly see that the addition of boron or phosphorus
lowers the total HR of the molecule, sometimes as much as 96%.
Reductions in HRR peak values are noted as well. If we focus on
the addition of boronic acid moiety to the terephthalic acid
In the case of the diboronic acid 9, a boroxine network must be
formed right away, which could then form faster a thermally
stable boron oxycarbide char, thus resulting in a much lower
peak HRR, and smaller secondary peak of heat release at lower
temperature. While the formation of the boroxine is pretty clear,
the mechanism of condensation between the boroxine and
carboxylic acid groups (or thermal decomposition of the
carboxylic acids into additional cross-linked structures) is not
clear and is likely a complex thermal condensation reaction.
DSC of compound 9 shows two endothermic events for this
material, one probably for loꢁss of water, hydrated/complexed
with the boronic acid, at 112 C, and then a conversion of the
boronic acid to boroxine at 166 ꢁC (Fig. 4).
The phosphonic acid functionalized molecule32 (Compound
11, Fig. 2d) also shows high char yields and low heat release
rates, as would be expected, since it is well known that
phosphonic acids can condense with carbon–oxygen bonds to
form polyphosphate cross-links which rapidly lead to char.6
The results become very interesting when studying the pinacol
boronate (Compound 3, Fig. 5b) and dimethyl phosphonate
(Compound 2a, Fig. 5c) when compared to the control sample
Fig. 3 Proposed mechanisms for the thermal decomposition of bor-
onoterephthalic acid (4) and 2,5-diboronoterephthalic acid (9).
1184 | J. Mater. Chem., 2012, 22, 1180–1190
This journal is ª The Royal Society of Chemistry 2012

View Online
Fig. 5 HRR curves for dimethyl terephthalate (a), and its pinacol boronate (b) and dimethyl phosphonate derivatives (c).
of dimethyl terephthalate 12 (Fig. 5a). By capping the boronic
acid with a pinacol functionality, no boroxine network can be
formed and so heat release actually increases for this molecule
relative to the control. This makes sense in that there is more
carbon present on the boronate to be combusted, and since
the boroxine cannot form, the molecule is free to pyrolyze
completely and burn to completion. Indeed, the HRR profile
for this molecule is sharper, indicating a rapid pyrolysis and
combustion.
The dimethyl phosphonate functionalized terephthalate shows
no char, indicating that the molecule was fully pyrolyzed, but it
does show a reduction in heat release. This suggests that the
molecule may be a vapor phase flame retardant such that the
phosphonate breaks off from the terephthalate after the molecule
is pyrolyzed and pushed into the instrument’s 900 ꢁC combustion
furnace. Phosphorus in some forms is known to be a vapor phase
flame retardant which inhibits combustion,6 and so it seems
likely that while the reduction in total HR is modest for this
Fig. 6 HRR curves for hydroquinone (a), and its cyclic phosphonate ester 14 (b).
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 1180–1190 | 1185

View Online
Table 2 PCFC heat release data for PU foam and Texin 990R (TPU)
functionality will serve as a condensed phase flame retardant as
well. The boronate ester functionality appears to have no effect
on flammability at all, and would likely make flammability worse
in the presence of a polymer, while the organophosphonates may
have some vapor phase activity depending upon the polymer they
are compounded into. In a polymer with no oxygen in the
backbone, this molecule is likely to be a vapor phase flame
retardant, while in a polymer in the presence of oxygen, the
phosphonate could work in both vapor phase and condensed
phase,6 unless it pyrolyzes out of the polymer before getting
a chance to crosslink/char. Obviously the flame retardant will
have to be tested in an actual polymer matrix to determine its
flame retardant mechanism in that particular polymer, but from
the data collected here, we can infer how the molecules are likely
to behave if particular chemical groups are present for interac-
tion between the flame retardant molecule and polymer during
polymer thermal decomposition.
Char HRR
yield
peak(s)
(wt%) value/W gꢀ1
Total
HR/kJ gꢀ1 Char notes
Sample
Polyurethane foam-1 0.91
Polyurethane foam-2 0.82
Polyurethane foam-3 0.93
217, 366
213, 367
215, 415
25.5
25.6
25.8
Light black
residue all
over pan, 1
small dot
on bottom
Black thin
fine film all
over inside
of pan and
top part of
outside
Texin 990R-1
Texin 990R-2
Texin 990R-3
1.14
0.53
0.95
169, 176, 332 27.5
203, 346
184, 361
27.2
27.5
molecule, the phosphonate functionality has the strong potential
to be a vapor phase flame retardant while the pinacol boronate
has no vapor phase activity. However, one must look to the rest
of the structure in this analysis. Since phosphorus reacts with
other oxygen atoms to form polyphosphate linkages which lead
to char, we can state that compound 2a volatilizes before any
phosphorus and oxygen can react. Structure 14 (Fig. 6b), which
is also a phosphonate ester34 but has free hydroxyl groups, shows
an effect opposite to compound 2a. Specifically, it shows high
char yields and lowered heat release, indicating that it is
a condensed phase/char forming flame retardant despite the alkyl
phosphonate. Therefore, in the case of compound 14, the phos-
phorus and oxygen react first, before volatilizing, while
compound 2a volatilizes first and then reacts. Sample 2b (HRR
curve not shown) behaves similarly to 2a. Changing the chemical
structure of the phosphonate to a diethyl phosphonate results in
a decrease in total HR reduction and a small heat release event at
lower temperature, but otherwise the heat release behavior is the
same, further indicating that this type of functionality leads to
more vapor phase behavior.
Part III. TPU + flame retardant blends and PCFC testing
As mentioned in the introduction, polyurethane foam is a target
for the development of new flame retardant chemistry, but as
also mentioned, making a small scale polyurethane foam is not
a trivial exercise. Therefore we sought to see if we could use
a thermoplastic polyurethane (TPU) as a screening matrix to mix
the flame retardants into for PCFC testing. The advantage of
a TPU over a PU foam is that solvent blending of flame retar-
dants and TPU is possible at 1 gram batch scales, something not
possible if one were to try and make a homogeneous flexible PU
foam. However, we first had to determine if the TPU was
a reasonable mimic for small scale polymer + flame retardant
blend tests, and we conducted PCFC tests on PU foam and
a TPU (Texin 990R). From the heat release rate data collected
(Table 2, Fig. 7), we can conclude that the TPU chosen is not
a perfect mimic for a polyurethane foam, but it is still a reason-
able choice when one compares the HRR curves (Fig. 7).
So with this information we studied the effect of the most
promising flame retardants from section II in the TPU matrix.
Fig. 8 shows the relevant data for compounds 11 and 14.
From these results we can conclude that the boronic acids are
high char forming materials, and when combined with a polymer
where they can form boroxine networks, they will serve as
condensed phase flame retardants. Likewise, the phosphonic acid
Fig. 7 HRR curves for PU foam and Texin 990R.
This journal is ª The Royal Society of Chemistry 2012
1186 | J. Mater. Chem., 2012, 22, 1180–1190

View Online
Fig. 8 HRR curves for TPU + compound 11 (left), and TPU + compound 14 (right).
The phosphonate ester (Compound 14) showed good flame
retardancy when tested by itself, but when mixed into the TPU, it
was not as effective and appears instead to make flammability
worse. It still helps with char yield but does raise total heat
release. Only the free phosphonic acid (11) increases the char
yield and lowers heat release.
The monoboronic acid 4, however, lowers heat release greatly
in the TPU (Fig. 9, Table 3) and was noted to keep the char
mostly intact (it did not flow into the corners of the crucible,
Fig. 10). The HRR shape for the TPU is greatly changed in the
presence of 4 (Fig. 9) which suggests that the monoboronic acid is
chemically reacting with the TPU during decomposition and
likely is assisting in char formation through potential cross-
linking reactions between polyurethane and boronic acid groups.
While we have not isolated chemical species to assign a mecha-
nism of flammability reduction, it is likely that the boronic acid
first forms a boroxine network around 180–200 ꢁC, as verified by
DSC (Fig. 3 and 4) and then this boroxine begins to complex with
urethane linkages. Boron is pre-disposed to form complexes with
the free electron pair on nitrogen and therefore it is likely that the
boroxine intermediate is complexing with decomposing urethane
groups which in turn serves as crosslinking sites for char
formation. Overall, the higher char yield, reduction in total HR,
and change in HRR behavior suggest that the monoboronic acid,
Fig. 9 HRR for TPU + boronoterephthalic acid (4).
Table 3 HRR data for TPU and TPU + FR blends
Char
yield (wt%)
HRR peak(s)
value/W gꢀ1
Total
HR/kJ gꢀ1
% THR
reduction
Sample
Char notes
Texin 990R-1
Texin 990R-2
Texin 990R-3
1.14
0.53
0.95
11.78
11.12
10.27
4.81
4.91
5.51
169, 176, 332
203, 346
184, 361
22, 281
251, 268
254, 275
274, 262, 19
265, 269, 32
253, 253, 19
256, 230
244, 67, 48
195, 72, 47
251, 63, 49
27.5
27.2
27.5
24.9
25.1
25.5
27.5
27.5
27.3
27.4
21.9
21.9
22.2
0.00
0.00
0.00
9.12
8.39
6.93
ꢀ0.36
ꢀ0.36
0.36
Black thin fine film all over inside of pan
and top part of outside
TPU 990 + 10% FR #11-1
TPU 990 + 10% FR #11-2
TPU 990 + 10% FR #11-3
TPU 990 + 10% FR #14-1
TPU 990 + 10% FR #14-2
TPU 990 + 10% FR #14-3
TPU 990 + 10% FR #14-4
TPU 990 + 10% FR4-1
TPU 990 + 10% FR4-2
TPU 990 + 10% FR4-3
Black all over inside of pan with some
flakey residue
Black all over inside of pan, no flakes,
small residue around edge
5.61
0.00
15.09
15.11
14.98
20.07
20.07
18.98
Small ball of ash in center of pan, no film
on rest of pan
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 1180–1190 | 1187

View Online
Experimental section
Pyrolysis combustion flow calorimetry (PCFC) was accom-
plished with an MCC-1 (Govmark Inc, USA) at 1 ꢁC sꢀ1 heating
rate under nitrogen from 150 to 900 ꢁC using method A of
ASTM D7309 (pyrolysis under nitrogen). Each sample was run
in triplicate to evaluate reproducibility of the flammability
measurements. Differential scanning calorimetry was accom-
plished with a Q1000 DSC (TA Instruments, USA), under
nitrogen from 25 to 300 ꢁC at 5 ꢁC minꢀ1
.
1H and 13C spectra were recorded at 300 MHz and 75 MHz
respectively and referenced to the solvent (CDCl3: 7.27 ppm and
77.0 ppm). Elemental analysis was provided by Atlantic Micro-
lab, Norcross, GA.
Fig. 10 Final chars for TPU + compound 11 (left) and TPU +
compound 4 (right).
with its inherently high char yield, is a suitable candidate for PU
foam flammability studies along with the phosphonic acid.
Additionally, the monoboronic acid 4 is likely to bring some
enhanced char formation and possibly even prevent dripping
during burning since the (TPU + compound 4) did not flow
during thermal decomposition. Obviously this must be verified
with larger scale experiments, but still, the change in char
structure may be a promising one and would be an additional
benefit of the small scale PCFC testing if such a char structure
was found to correlate to larger scale flammability behavior.
Dimethyl iodoterephthalate 1 has been reported previously,
but we could not find an actual synthetic procedure in any of the
available references. Thus, the full details of its preparation are
reported here, along with pertinent NMR data. Literature search
yields a single reference for each of the phosphonates 2a and 2b,35
and details on the preparation, or NMR data, of 2a are not
available. Taking into account that a different synthetic protocol
was used in the current work, we have found it justified to list
preparation details and NMR data for both phosphonates.
Dimethyl aminoterephthalate was purchased from Acros
Organics. Phosphonoterephthalic acid (11),32 2,5-diiodotereph-
thalic acid (6),31 and 2-(5,5-dimethyl-2-oxo-1,3,2-dioxaphos-
phane-2-yl)hydroquinone (14)33 were prepared using previously
published synthetic protocols.
Conclusions
This report has outlined the preparation and use of some
dimethyl terephthalate derivatives in various transition metal-
catalyzed coupling reactions. Dimethyl iodoterephthalate and
dimethyl 2,5-diiodoterephthalate have been successfully used in
the generation of phosphonic and boronic esters, using Pd- or
Ni-containing catalysts. The boronic esters have been utilized
further to generate the previously unknown mono- and dibor-
onoterephthalic acids.
Dimethyl iodoterephthalate (1)
A mixture of dimethyl aminoterephthalate (10.50 g, 50.0 mmol),
water (10 mL), ice (40 g) and 98% sulfuric acid (15 mL) was cooled
in an ice-water bath. Solution of NaNO2 (3.65 g, 50.0 mmol) in
water (10 mL) was added dropwise at the same temperature, over
a 0.5 h period. The reaction mixture was stirred for additional 30
min at 0–5 ꢁC, then poured slowly into a solution of KI (9.96 g,
60.0 mmol) in water (10 mL) at ambient temperature. The resul-
tant mixture was stirred for additional 20 min at 60 ꢁC, then
cooled and the excess iodine quenched with aqueous NaHSO3.
The product precipitate was filtered, washed with water and dried
to yield an orange-brown solid. Recrystallization from methanol
yielded a yellow solid (10.24 g, 64%). 1H NMR (CDCl3) d 3.86 (s,
3H), 3.87 (s, 3H), 7.80 (d, J ¼ 8.1 Hz, 1H), 8.02 (dd, J1 ¼ 1.8 Hz,
J2 ¼ 8.1 Hz, 1H), 8.56 (d, J ¼ 1.5 Hz, 1H).
When studying the heat release of these molecules and their
potential as flame retardants, we see that groups capable of
crosslinking and condensed phase char formation yield the
greatest reduction in heat release. The boronic and phosphonic
acid groups reduced heat release the most, followed by the
phosphonates. However, if the boronic acid group is present as
an ester (such as the pinacol ester in this paper), heat release is
increased, which indicates that this material would make a poor
flame retardant additive. When comparing this result to the
dimethyl phosphonate, which also does not form any char, some
heat release reduction is observed and so this molecule may have
some value as a vapor phase flame retardant. When the flame
retardants are combined with a TPU polymer, some similar
effects of heat release reduction are noted, but the boronic acid
and phosphonic acid containing molecules show the greatest
reduction in heat release which roughly correlates to their
effectiveness found when the FR additives were tested alone. Of
course the correlation is not perfect, but for screening purposes
to identify what materials are worth scaling up for additional
work, the use of PCFC to screen flame retardant performance
prior to scale-up appears to be justified. Obviously more work is
needed to verify the screening criteria that we outline in this
paper and further justify the use of this tool for flame retardant
material development. To that end, we intend to study this
metrology need further via studies of existing commercial flame
retardants and flame retardant polyurethanes.
Preparation of phosphonate esters 2 (Route 1)
A mixture of dimethyl iodoterephthalate (1 equiv.), trialkyl
phosphite (4 equiv.) and PdCl2 (0.1 equiv.) was purged with
nitrogen, and then stirred for 5 h at 150 ꢁC, under nitrogen
atmosphere. Water and 1,2-dichloroethane were added after
cooling to ambient temperature. The organic layer was sepa-
rated, washed two times with water, dried (MgSO4) and the
volatile components removed under reduced pressure. The
residue was purified on a silica gel column, using methylene
chloride, followed by CH2Cl2 : EtOAc ¼ 3 : 1, followed by pure
EtOAc. The last two fractions were combined, solvent was
removed under reduced pressure, to yield the product.
1188 | J. Mater. Chem., 2012, 22, 1180–1190
This journal is ª The Royal Society of Chemistry 2012

View Online
Preparation of phosphonate esters 2 (Route 2)
the same purification protocol, but with diminishing amounts of
water. Purity was monitored by NMR. Yield: 0.54 g (55%, off-
A mixture of dimethyl iodoterephthalate (1 equiv.), dialkyl
phosphite (1.2 equiv.), triethylamine (1.5 equiv), triphenylphos-
phine (0.06 equiv.) and Pd(OAc)2 (0.02 equiv.) in ethanol (4 mL
per 1 mmol of terephthalate) was purged with nitrogen, then
refluxed for 18 h under nitrogen atmosphere. After cooling to
ambient temperature the solvent was removed under reduced
pressure. The residue was dissolved in a small amount of CH2Cl2
and purified on a silica gel column, using methylene chloride,
followed by CH2Cl2 : EtOAc ¼ 3 : 1, followed by pure EtOAc.
The last two fractions were combined, solvent was removed
under reduced pressure, to yield the product.
ꢁ
1
white solid). Mp 305 C (dec). H NMR (D2O) d 7.57 (d, J ¼
7.2 Hz, 1H), 7.86 (d, J ¼ 7.9 Hz, 1H), 7.97 (s, 1H). 13C NMR
(D2O, indirectly referenced to acetone-d6 in D2O) d 127.2, 132.3,
135.9, 141.4, 173.1, 177.5. Anal. Calcd for C8H7BO6: C, 45.77; H,
3.36. Found: C, 45.34; H, 3.26%.
Dimethyl 2,5-diiodoterephthalate (7)
Although previously reported,31 compound 7 was prepared by us
using Fischer esterification, instead of intermediate conversion of
the starting acid to the corresponding acid chloride. Its synthesis
is therefore described in detail: 2,5-diiodoterephthalic acid 6
(3.00 g, 7.18 mmol) was dissolved in methanol (40 mL), conc.
H2SO4 (4 mL) was added, and the solution was refluxed for 5 h.
The reaction mixture was cooled in the refrigerator overnight,
the product was filtered and washed with small amount of cold
Dimethyl (dimethylphosphono)terephthalate (2a)
Yield: 81%. Colorless oil. 1H NMR (CDCl3) d 3.83 (d, J ¼
11.4 Hz, 6H), 3.96 (s, 3H), 3.97 (s, 3H), 7.79 (dd, J1 ¼ 4.9 Hz,
J2 ¼ 8.0 Hz, 1H), 8.25 (dt, J1 ¼ 1.5 Hz, J2 ¼ 8.0 Hz, 1H), 8.56
(dd, J1 ¼ 1.6 Hz, J2 ¼ 14.3 Hz, 1H).
1
methanol, to yield a white crystalline solid (2.01 g, 64%). H
NMR (CDCl3) d 3.84 (s, 6H), 8.34 (s, 2H).
Dimethyl (diethylphosphono)terephthalate (2b)
Dimethyl 2,5-bis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)
terephthalate (8)
Yield: 86%. Colorless oil. 1H NMR (CDCl3) d 1.34 (t, J ¼ 7.5 Hz,
6H), 3.96 (s, 3H), 4.08–4.22 (m, 4H), 7.76 (dd, J1 ¼ 4.8 Hz, J2 ¼
7.8 Hz, 1H), 8.24 (dt, J1 ¼ 1.5 Hz, J2 ¼ 7.8 Hz, 1H), 8.56 (dd,
J1 ¼ 1.5 Hz, J2 ¼ 14.1 Hz, 1H).
Method 1: a mixture of dimethyl 2,5-diiodoterephthalate 7 (1.00 g,
2.24 mmol), 4,4,5,5-tetramethyl-1,3,2-dioxaboralane (1.35 g,
10.54 mmol, 1.53 mL), triethylamine (0.93 g, 9.18 mmol, 1.28 mL)
and (Ph3P)2PdCl2 (0.11 g, 0.156 mmol) in toluene (25 mL) was
purged with nitrogen and refluxed for 12 h under inert atmo-
sphere. Water was added and the layers separated. The organic
layer was dried (MgSO4) and the solvent removed under reduced
pressure. The residue was purified on a silica gel column (CH2Cl2),
yielding the product as a white solid (0.51 g, 49%). Additional
purification was achieved via recrystallization from methanol. Mp
253–255 ꢁC. 1H NMR (CDCl3) d 1.43 (s, 24H), 3.93 (s, 6H), 8.05 (s,
2H). 13C NMR (CDCl3) d 25.1, 52.8, 84.5, 132.7, 136.4, 168.1.
Anal. Calcd for C22H32B2O8: C, 59.23; H, 7.23. Found: C, 59.51;
H, 7.34%. Method 2: dimethyl 2,5-diiodoterephthalate 7 (0.50 g,
1.12 mmol), bis(pinacolato)diboron (0.62 g, 2.44 mmol), KOAc
(0.33 g, 3.37 mmol) and (dppf)2PdCl2 (46 mg, 0.056 mmol) were
placed in a flask, which was evacuated and backfilled with dry
nitrogen. Anhydrous DMF (10 mL) was added and the reaction
mixture was stirred for 20 h at 80 ꢁC. The mixture was cooled and
DMF evaporated under reduced pressure. Benzene (40 mL) was
added and the suspension was vacuum filtered. The filtrate was
washed successively with aq. NaCl and water, the organic layer
was separated, dried (MgSO4) and the solvent removed under
reduced pressure. The residue was washed with cold pentane to
remove excess bis(pinacolato)diboron, followed by purification
on a silica gel column (CH2Cl2), leading to isolation of the product
as a white solid (0.19 g, 38%).
Dimethyl 2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)
terephthalate (3)
A mixture of dimethyl iodoterephthalate (3.53 g, 11.02 mmol),
4,4,5,5-tetramethyl-1,3,2-dioxaboralane (1.76 g, 13.78 mmol,
2.00 mL), dicyclohexylmethylamine (4.05 g, 20.75 mmol, 4.41
mL) and (dppp)2NiCl2 (0.17 g, 0.32 mmol) in toluene (20 mL)
was purged with nitrogen and refluxed for 12 h under inert
atmosphere. TLC analysis (CH2Cl2) indicated the presence of
two main components: the target material (slower running) and
dimethyl terephthalate. The contents were poured into dilute
HCl, stirred for 5 min and the layers separated. The organic layer
was dried (MgSO4) and the solvent removed under reduced
pressure. The residue was purified on a silica gel column
(hexane : CH2Cl2 ¼ 1 : 1), the solvent removed under reduced
pressure and the product was isolated as a white solid (2.51 g,
71%). Additional purification was via recrystallization from
pentane. Mp 69–71 ꢁC. 1H NMR (CDCl3) d 1.44 (s, 12H), 3.94 (s,
6H), 7.99 (dd, J1 ¼ 8.1 Hz, J2 ¼ 0.6 Hz, 1H), 8.08 (dd, J1 ¼ 8.1
Hz, J2 ¼ 1.8 Hz, 1H), 8.16 (dd, J1 ¼ 1.7 Hz, J2 ¼ 0.6 Hz, 1H). 13
C
NMR (CD3CN) d 24.9, 52.4, 52.6, 84.3, 128.7, 130.2, 132.7,
133.4, 137.3, 166.4, 167.8. Anal. Calcd for C16H21BO6: C, 60.03;
H, 6.61. Found: C, 59.93; H, 6.78%.
Boronoterephthalic acid (4)
2,5-Diboronoterephthalic acid (9)
Hydrochloric acid (3 M, 10 mL) was added to boronic ester 3
(1.50 g, 4.68 mmol) and the mixture was stirred for 4 h at 60 ꢁC,
then cooled to room temperature and the solvent was evapo-
rated. The residue was dissolved in water upon heating, the
solution was cooled and the crystallized impurity (terephthalic
acid) was vacuum filtered. The mother liquor was evaporated to
dryness, and the resultant solid was subjected two more times to
Diboronic ester 8 (0.25 g, 0.56 mmol) was placed in a flask and 3 M
HCl (4 mL) was added. The resultant mixture was stirred for 12 h
at 40 ꢁC. Then solvent was removed under reduced pressure, the
residue was suspended in ether, stirred for 30min, then filtered and
dried to yield an off-white solid (0.11 g, 77%). Mp 330 ꢁC (dec). 1H
NMR (D2O) d 7.79 (s, 2H). 13C NMR (D2O, indirectly referenced
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 1180–1190 | 1189

View Online
9 P. Blomqvist, L. Rosell and M. Simonson, Fire Technol., 2004, 40, 39–
58.
10 P. Blomqvist, L. Rosell and M. Simonson, Fire Technol., 2004, 40, 59–
73.
to acetone-d6 in D2O) d 127.0, 141.3, 179.0. Anal. Calcd for
C8H8B2O8: C, 37.86; H, 3.18. Found: C, 38.12; H, 3.33%.
11 H. M. Stapleton, N. G. Dodder, J. H. Offenberg, M. M. Schantz and
S. A. Wise, Environ. Sci. Technol., 2005, 39, 925–931.
12 L. Tange and D. Drohmann, Polym. Degrad. Stab., 2005, 88, 35–40.
13 I. Wadehra, Fire Mater., 2005, 29, 121–126.
14 A. B. Morgan and J. M. Tour, J. Appl. Polym. Sci., 2000, 76, 1257–
1268.
15 M. Sacristan, T. R. Hull, A. A. Stec, J. C. Ronda, M. Galia and
V. Cadiz, Polym. Degrad. Stab., 2010, 95, 1269–1274.
16 S. V. Levchik and E. D. Weil, J. Fire Sci., 2006, 24, 345–364.
17 M. Schuster, T. Rager, A. Noda, K. D. Kreuer and J. Maier, Fuel
Cells, 2005, 5, 355–365.
18 R. E. Lyon and R. N. J. Walters, J. Anal. Appl. Pyrolysis, 2004, 71,
27–46.
19 B. Schartel, K. H. Pawlowski and R. E. Lyon, Thermochim. Acta,
2007, 462, 1–14.
20 C. Q. Yang, Q. He, R. E. Lyon and Y. Hu, Polym. Degrad. Stab.,
2010, 95, 108–115.
21 A. B. Morgan and M. Galaska, Polym. Adv. Technol., 2008, 19, 530–
546.
Solvent blending of TPU + flame retardant general procedures
A 2 g batch of TPU (Texin 990R) + flame retardant was prepared
in the following manner. To 1.8 g of TPU, 0.2 g of flame retar-
dant was added. This pellet + powder blend was added into a 250
mL glass beaker equipped with a magnetic stir bar. To this
beaker was added 200 g of DMF (Aldrich, anhydrous, 99.8%
purity) and the beaker was placed on a heated stir plate. The
mixture was heated to 150 ꢁC, and stirred at 300 rpm. All of the
TPU dissolved in 4 hours, and the material was mixed for one
additional hour, after all of the TPU was noted to have dissolved.
After mixing, the material was allowed to cool, the mixture was
poured into a porous Teflon sꢁheet-lined aluminium tray, and
dried in a vacuum oven at 160 C for 12 hours. The final dried
TPU film was then removed from the Teflon sheet and tested for
residual solvent content via TGA. If solvent was found, the
material was dried again for another 12 hours at 160 ꢁC.
22 R. H. Kramer, N. Witten and J. W. Gilman, in Inerflam 2010
Proceedings, Interscience Communications Ltd., Nottingham, UK,
2010, pp. 499–508.
23 J. Lefebvre, M. Le Bras, B. Bastin, R. Paleja and R. Delobel, Journal
of Fire Sciences, 2003, 21, 343–367.
24 S. V. Levchik and E. D. Weil, Polym. Int., 2004, 53, 1585–1610.
25 Y. Wu, H. Guo, X. Zhang, T. D. James and J. Zgao, Chem.–Eur. J.,
2011, 17, 7632–7644.
26 O. Bolton and J. Kim, J. Mater. Chem., 2007, 17, 1981–1988.
27 S. Hideo, M. Nomura, K. Tokunaga and I. Hashiba, in
Phenylphosphonic acid derivative and production process therefor,
Eur. Pat. Appl., EP0855400 (A1), Nissan Chemical Ind. Ltd.,
Japan, 1998, p. 26.
28 M. K. Dezfuli and L. J. Goossen, Synlett, 2005, 445–448.
29 P. M. Iovine, M. A. Kellett, N. P. Redmore and M. J. Therien, J. Am.
Chem. Soc., 2000, 122, 8717–8727.
Acknowledgements
Funding for this project was provided by the National Institute
of Standards and Technology, Building and Fire Research
Laboratory (NIST-BFRL), Fire Research Grant
#
70NANB9H9183. The authors wish to thank Mary Galaska,
Kathy Schenck, and Kathleen Beljan for assistance with PCFC
sample preparation and testing.
References
30 S. Simmons, A. Fratini and V. Benin, Acta Crystallogr., Sect. C:
Cryst. Struct. Commun., 2011, 67, m123–m125.
31 R. J. Perry, B. D. Wilson, S. R. Turner and R. W. Blevins,
Macromolecules, 1995, 28, 3509–3515.
1 R. N. Walters and R. E. Lyon, J. Appl. Polym. Sci., 2002, 87, 548–563.
2 D. J. Irvine, J. A. McCluskey and I. M. Robinson, Polym. Degrad.
Stab., 2000, 67, 383–396.
32 S. Branion and V. Benin, Synth. Commun., 2006, 36, 2121–2127.
33 A. Worku, A. Bharadwaj, R. Thibault, M. Wilson and D. Potts, in
PCT Int. Appl., WO2010025165 (A1), Dow Global Technologies,
USA, 2010, p. 47.
34 A. Worku, A. Bharadwaj, R. Thibault, M. Wilson and D. Potts, in
Aromatic Phosphonates for Fireproofing Agents for Polymers, PCT
Int. Appl., WO2010025165 (A1), Dow Global Technologies Inc.,
USA, 2010, p. 47.
35 H. Suzuki, M. Nomura, K. Tokunaga and I. Hashiba, in
Phenylphosphonic acid derivative and production process thereof
Eur. Pat. Appl., EP0855400, Nissan Chemical Industries Ltd.,
Japan, 1998, p. 26.
3 S. Bourbigot and S. Duquesne, J. Mater. Chem., 2007, 17, 2283–2300.
4 H. Singh and A. K. Jain, J. Appl. Polym. Sci., 2009, 111, 1115–1143.
5 M. M. Hirschler, Polym. Adv. Technol., 2008, 19, 521–529.
6 Fire Retardancy of Polymeric Materials, ed. C. A. Wilkie and A. B.
Morgan, Taylor and Francis, Boca Raton, FL, 2010, ISBN 978-1-
4200-8399-6.
7 E. D. Weil and S. V. Levchik, Flame Retardants for Plastics and
Textiles, Hanser Publications, Cincinnati, OH, 2009, ISBN 978-3-
446-41652-9.
€
8 M. Simonson, P. Blomqvist, A. Boldizar, K. Moller, L. Rosell,
C. Tullin, H. Stripple and J. O. Sundqvist, in SP Report 2000:13,
2000, ISBN 91-7848-811-7.
1190 | J. Mater. Chem., 2012, 22, 1180–1190
This journal is ª The Royal Society of Chemistry 2012