Thermally induced structural transformations of linear coordination polymers based on aluminum tris(diorganophosphates)

Article abstract of DOI:10.1039/c8dt02672f

The thermal transitions of inorganic-organic hybrid polymers composed of linear aluminum tris(diorganophosphate) chains with a general formula of catena-Al[O₂P(OR)₂]₃ (where R = C₁-C₈ alkyl group or phenyl moiety) have been studied by means of DSC, powder XRD, TGA and TG-QMS, as well as optical spectroscopy. DSC and XRD reveal that most of them undergo reversible structural transformations in the solid state between ?100 and 200 °C caused by the changes in conformation of their organic substituents; however, a translational displacement of the rigid polymeric chains occurs only in the case of the derivative bearing long 2-ethylhexyl groups, which becomes liquid at about 140 °C. The thermal decomposition of the studied polymers begins between 200 and 265 °C depending on the type of organic substituent R decorating their aluminophospate core. TGA combined with mass spectrometry of the evolved gaseous products shows that the pyrolytic decomposition of Al[O₂P(OR)₂]₃ proceeds either through β-elimination of olefin (for compounds with C₂-C₈ aliphatic ligands), or a homolytic cleavage of the P-OR bond (for methyl and phenyl derivatives); both processes are accompanied by condensation of the newly formed POH groups and liberation of water. Powder XRD, FTIR and SEM analyses of the solid residues indicate that thermolysis of Al[O₂P(OR)₂]₃ accompanied by olefin elimination leads to the formation of condensed aluminum phosphates, mainly aluminum cyclohexaphosphate, exhibiting porous morphology. On the other hand, thermal degradation of methyl or phenyl derivatives results in amorphous aluminophosphate residues, and the latter contains conducting carbonaceous phases.

Full text of DOI:10.1039/c8dt02672f

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Thermally induced structural transformations of
linear coordination polymers based on aluminum
tris(diorganophosphates)†
Cite this: Dalton Trans., 2018, 47,
16480
Maciej Dębowski, * Krzysztof Łokaj, Andrzej Ostrowski,
Janusz Zachara,
Paulina Wiecińska,
Paweł Falkowski,
Anna Krztoń-Maziopa and
Zbigniew Florjańczyk
The thermal transitions of inorganic–organic hybrid polymers composed of linear aluminum tris(diorgano-
phosphate) chains with a general formula of catena-Al[O2P(OR)2]3 (where R = C1–C8 alkyl group or
phenyl moiety) have been studied by means of DSC, powder XRD, TGA and TG-QMS, as well as optical
spectroscopy. DSC and XRD reveal that most of them undergo reversible structural transformations in the
solid state between −100 and 200 °C caused by the changes in conformation of their organic substitu-
ents; however, a translational displacement of the rigid polymeric chains occurs only in the case of the
derivative bearing long 2-ethylhexyl groups, which becomes liquid at about 140 °C. The thermal
decomposition of the studied polymers begins between 200 and 265 °C depending on the type of
organic substituent R decorating their aluminophospate core. TGA combined with mass spectrometry of
the evolved gaseous products shows that the pyrolytic decomposition of Al[O2P(OR)2]3 proceeds either
through β-elimination of olefin (for compounds with C2–C8 aliphatic ligands), or a homolytic cleavage of
the P–OR bond (for methyl and phenyl derivatives); both processes are accompanied by condensation of
the newly formed POH groups and liberation of water. Powder XRD, FTIR and SEM analyses of the solid
residues indicate that thermolysis of Al[O2P(OR)2]3 accompanied by olefin elimination leads to the for-
mation of condensed aluminum phosphates, mainly aluminum cyclohexaphosphate, exhibiting porous
morphology. On the other hand, thermal degradation of methyl or phenyl derivatives results in
amorphous aluminophosphate residues, and the latter contains conducting carbonaceous phases.
Received 30th June 2018,
Accepted 20th October 2018
DOI: 10.1039/c8dt02672f
rsc.li/dalton
these polymers is self-organization of solvated aluminum
cations and organophosphorous anions, which may lead to
Introduction
In recent years, coordination polymers based on organically the formation of various supramolecular structures such as
modified aluminum phosphates and phosphonates have spherical nanoparticles, fibres, rods, two-dimensional plates
received increased attention due to their potential application or metal–organic frameworks, depending on the starting
as adsorbents,1 processable precursors of mesoporous zeo- materials and synthetic procedure applied.1,3,5,6,11–20
lites,2 flame retardants,3,4 reinforcing and orientation enhan-
We have previously reported some structural and morpho-
cing fillers in polymer composites,3–8 and organogelators,9,10 logical properties of a large family of aluminum tris(diorgano-
as well as luminophores.10,11 The key step in the synthesis of phosphates) (DOPAls) obtained from phosphoric acid diesters
and different sources of aluminum.21 We have shown that this
class of hybrid polymers forms catena-Al[O2P(OR)2]3 chains
Faculty of Chemistry, Warsaw University of Technology, Noakowskiego 3,
propagating along the c-axis of their unit cell, whose parallel
location induces the formation of closely packed hexagonal
00-664 Warsaw, Poland. E-mail: debowski@ch.pw.edu.pl
†Electronic supplementary information (ESI) available: Values of DnPPAl unit
crystalline domains. Single-crystal X-ray diffraction studies
cell constants measured between room temperature and 205 °C, FTIR spectra
carried out at −173 °C for the sample of aluminum tris(di-
and tabulated PXRD data for the solid residues from DOPAl thermal degradation
ethylphosphate) (DEPAl) precipitated during a hydrothermal
(130 °C, 1.5 MPa) reaction between triethyl phosphate and
aluminum L-lactate revealed that this compound crystallizes in
at 600 °C, DSC traces of DnPPAl, DiPPAl and DPhPAl, PXRD diffractograms of
DMPAl, DEPAl, DnPPAl and DBPAl measured at different temperatures, PXRD
patterns of BEHPAl pyrolyzates, 27Al and 31P MAS NMR spectra of DiPPAl and
DPhPAl thermo- and pyrolyzates, TGA and TG-QMS profiles of DOPAls, SEM
images of the solid residues from DnPPAl, DiPPAl, DBPAl and DPhPAl thermally
degraded at 600 °C, Raman spectra of DPhPAl pyrolyzates. See DOI: 10.1039/
c8dt02672f
ˉ
a trigonal crystallographic system (P3 space group) and com-
prises two crystallographically non-equivalent chains, denoted
as A- and B-type chains (Fig. 1). The main structural difference
16480 | Dalton Trans., 2018, 47, 16480–16491
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
peratures, and the mechanism of aluminum tris(diorganophos-
phate) (DOPAl) thermal decomposition, as well as the structure
of products formed during its calcination up to 600 °C.
Experimental
Materials and sample preparation methods
All chemicals for syntheses were purchased from commercial
sources and were used without further purification: trimethyl
phosphate (TMP) (97%, Aldrich), triethyl phosphate (TEP)
(99%, Aldrich), tri-n-propyl phosphate (TnPP) (99%, Aldrich),
triisopropyl phosphate (TiPP) (99%, Aldrich), triphenyl phos-
phate (TPhP) (99%, Merck Chemicals), di-n-butyl phosphate
(DBP) (97%, Aldrich), bis(2-ethylhexyl) phosphate (BEHP)
(97%, Aldrich), aluminum L-lactate (AlLact) (97%, Aldrich), tri-
ethylaluminum (TEA) (1.9 M solution in toluene, Aldrich), and
sodium bicarbonate (pure, POCh Gliwice). Toluene used in the
reactions involving TEA was distilled from sodium-benzophe-
none ketyl solution and stored over dry 4 Å molecular sieves
under a nitrogen atmosphere.
Fig. 1 Crystal structure of DEPAl: (A) the view down the crystallo-
graphic b-axis, (B) the view down the crystallographic c-axis. The edges
of the DEPAl unit cell are indicated by purple, dashed lines.
Preparation of DOPAls. The DOPAl samples were obtained
according to one of the synthetic methods described in our
between them is the conformation of the ethyl groups around previous paper, in which a detailed discussion of their
physicochemical, morphological and spectral properties is
the phosphorus centres. In A-type DEPAl chains, the ethyl
groups have a gauche–gauche (G+G+) conformation, whereas in also given.21 Table 1 contains selected literature data concern-
B-type chains they adopt a gauche–trans (G−T and G+T) arrange- ing the preparation and composition of the samples investi-
gated in the present work.
ment. Both the type of organic group and the method of prepa-
ration strongly affect the morphology of the final product.
Graphitization of DPhPAl samples. DPhPAl samples were
Therefore, one can obtain highly crystalline fibers with hexag- thermally degraded in a horizontal tube furnace (Carbolite
STF 16/75/450) at 600 °C in air for 6 h, or under an inert gas
onal cross-sections, irregular agglomerates or semicrystalline
wax-like phases. It was also demonstrated that DOPAls are atmosphere (7 L h−1 flow of argon) for 6 and 24 h.
soluble in primary amines and this feature can be utilized for
the synthesis of epoxy nanocomposites.21
Analytical methods
In order to examine and possibly expand the range of prac-
Elemental analysis. The hydrogen and carbon contents in
tical applications of DOPAls, we have decided to study in detail the investigated materials were determined using
their thermal properties. In this paper we present and discuss PerkinElmer CHNS/O II 2400 instrument.
a
the data concerning the reversible structural transitions of
Powder X-ray diffraction (PXRD) studies. PXRD patterns at
DOPAls occurring in the solid state at low or moderate tem- room temperature were recorded with a Bruker D8 Advance
Table 1 Literature data on the preparation and elemental analysis of DOPAl samples21
Results of elemental analysis
DOPAla
DOPAl formula (theoretical values)b
Method of synthesis
C/wt%
H/wt%
DMPAl
DEPAl
DnPPAl
DiPPAl
DBPAl
BEHPAl
DPhPAl
C6H18AlO12P3 (C, 17.92; H, 4.51%)
C12H30AlO12P3 (C, 29.64; H, 6.22%)
C18H42AlO12P3 (C, 37.90; H, 7.42%)
C18H42AlO12P3 (C, 37.90; H, 7.42%)
TOP-M1c
TOP-M1c
TOP-M2d
TOP-M2d
DOP-TEAe
DOP-M2f
DOP-M3g
18.18
29.89
37.92
37.95
43.87
56.53
55.42
4.65
6.13
7.48
7.44
8.49
10.11
4.07
C
24H54AlO12P3 (C, 44.04; H, 8.32%)
C48H102AlO12P3 (C, 58.16; H, 10.37%)
C36H30AlO12P3 (C, 55.83; H, 3.90%)
a Aluminum tris(diorganophosphate) containing methyl (DMPAl), nPr (DnPPAl), iPr (DiPPAl), nBu (DBPAl), 2-ethylhexyl (BEHPAl) or phenyl
(DPhPAl) groups. b Values of carbon and hydrogen content (in wt%) calculated for the assumed DOPAl formula. c Reaction of AlLact with a water-
miscible phosphoric acid triester carried out in water under reflux.21 d Reaction of AlLact with a hydrophobic phosphoric acid triester carried out
in water, requiring the use of a pressure vessel and temperatures in the range of 130–150 °C.21 e Reaction of phosphoric acid diester with TEA.21
f Reaction of AlLact with phosphoric acid diester carried out in water under reflux.21 g Reaction of AlLact with a hydrophobic phosphoric acid
diester carried out in water for 24 h, requiring the use of a pressure vessel and temperatures in the range of 130–150 °C.21
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16480–16491 | 16481

View Article Online
Paper
Dalton Transactions
automated diffractometer equipped with a LynxEye position
Optical microscopy. Visual observations of DOPAl samples
sensitive detector using Cu-Kα radiation (λ = 0.15406 nm). heated from room temperature to 200 °C were made on a
Diffraction data were collected in the Bragg–Brentano (θ/2θ) Boetius apparatus (VEB Kombinat NAGEMA).
horizontal geometry (flat reflection mode) between 3° and 60°
SEM analysis. SEM images were obtained on a Zeiss Ultra
(2θ) in 0.03° steps at 10 s per step.
Plus field emission scanning electron microscope equipped
Variable temperature powder XRD (VT-PXRD) studies. with a GEMINI column. Before the measurements, all samples
VT-PXRD patterns were recorded on a Bruker D8 Discover were coated with an electron conductive layer of carbon utiliz-
instrument equipped with a VANTEC-1 position sensitive detec- ing a high-vacuum carbon sputter coater.
tor and an Anton Paar DCS-350 heating stage (temperature
Specific surface area. The total specific surface area (TSSA)
stability 1 K) using Cu-Kα radiation and a step size of 0.0183°. of the samples was determined by the nitrogen sorption
The data were collected between 4° and 60° (2θ). In the case of method on a Micromeritics ASAP 2020 apparatus at the tem-
DnPPAl, the TOPAS software22 (Bruker AXS) was utilized for perature of liquid nitrogen. Before the measurement, each
indexing the VT-PXRD patterns using 22 reflections (2θ < 38°), sample was degassed in vacuo first at 90 °C for 1 h, and then
as well as Pawley refinement of the obtained cell parameters.
for additional 4 hours at 120 °C (DOPAls) or 300 °C (solid resi-
Differential scanning calorimetry. DSC measurements were dues from DOPAl thermolysis in air). The results from the
conducted on a TA Instruments DSC Q200 V24.2 Build 107 5-point measurements were approximated with the Brunauer–
apparatus. The first heating run from −100 to 200 °C was per- Emmett–Teller (BET) adsorption isotherm.
formed at a heating rate of 10 °C min−1, and then, cooling to
Electrochemical impedance spectroscopy (EIS). Impedance
−100 °C at a rate of 10 °C min−1 was applied. The second spectra in the frequency range from 107 down to 1 Hz and
heating run was measured at a heating rate of 10 °C min−1 up temperature range 20–70 °C were recorded by using a compu-
to 200 °C.
ter-controlled setup based on a Solartron 1296 Dielectric
Thermogravimetric analysis. TGA studies were performed Interface and Schlumberger SI 1255 HF Frequency Response
using a Derivatograph Q 1500-D instrument operating at a Analyzer. Impedance spectra were analysed with EIS Spectrum
heating rate of 5–10 °C min−1 from room temperature to Analyser 1.0 software by fitting suitable equivalent circuits col-
800 °C. Measurements were carried out in a stream of air or lected in its internal library of models.23 As the graphitized
nitrogen gas.
DPhPAl materials were impossible to compress into the form
The volatile thermal decomposition products were analysed of pellets, for the EIS measurements the powder samples were
using a Netzsch STA 449 C Jupiter instrument coupled with a placed in a swagelok-type cell and pressed between parallel
Netzsch QMS 403C Aeolos quadrupole mass spectrometer. The stainless steel electrodes (13 mm in diameter) until the impe-
samples were heated from 35 up to 1000 °C with a heating rate dance, measured at room temperature, remained constant.
of 5 °C min−1 under an inert gas atmosphere (argon, flow rate
90 mL min−1).
FTIR spectroscopy. FT-IR spectra were recorded on
Thermo Scientific Nicolet iS5 spectrometer equipped with an
iD7 diamond ATR accessory.
a
Results and discussion
Temperature-dependent phase transitions and
transformations in the DOPAl structure below 200 °C
Raman spectroscopy. Raman spectra were collected on a
Thermo Nicolet Almega XR spectrometer equipped with a DSC studies (Fig. 2–3 and S1–S2, ESI†) reveal that DOPAls,
532 nm laser. with exception of derivatives containing either the isopropyl
MAS NMR spectroscopy. NMR experiments were conducted (Fig. S3A, ESI†) or phenyl group (Fig. S3B, ESI†), undergo
on a Bruker Avance III 400 spectrometer, with an Ultra-Shielded endothermic phase transitions when heated from −100 to
Plus Wide Bore superconducting magnet, operating at 400.13, 200 °C. These transitions are reversible, as indicated by the
161.91 and 104.26 MHz for 1H, 31P and 27Al, respectively. All presence of respective exothermic DSC peaks on the cooling
measurements were conducted using a 4 mm broadband MAS curves. They are typically located at temperatures 10–20 °C
probe head and 4 mm zirconia rotors with Kel-F caps. For all lower than the positions of the respective signals on the
measurements, the MAS spinning rate was set at 8 kHz.
heating curves. Both the number and temperatures of these
For 27Al spectra, a simple one-pulse sequence (without 1H transitions depend significantly on the type of organic substi-
decoupling) was applied with a 4 μs excitation pulse (27Al π/2) tuent present in diorganophosphate ligands.
and 0.2 s relaxation delay. 2000 data points and 512 scans FIDs
The methyl derivative exhibits two endothermic phase tran-
were accumulated; the spectral width was 83 kHz (800 ppm).
sitions, as indicated by the presence of two sharp signals with
31P spectra were run using one pulse sequence with 1H high maxima at 144 and 187 °C (Fig. 2A); their enthalpies are equal
power decoupling during acquisition (High Power decoupling to 25.7 and 6.4 kJ mol−1, respectively. DEPAl (Fig. 2B) and
sequence). The spectral width was 131.5 kHz (812 ppm); 4000 DnPPAl (Fig. S1, ESI†) show two endotherms at temperatures
data points and 200 scans FIDs with 10 s relaxation delay were exceeding 100 °C, but undergo also some additional energy-
collected for each spectrum. The 31P excitation pulse (π/2) was requiring processes, which occur below 0 °C. These low-temp-
set at 4.5 μs and 22.7 kHz pulse with a spinal 16 sequence was erature transitions are characterized by broad signals on the
applied during acquisition for 1H decoupling.
DSC curve located at approximately −72 °C for the n-propyl
16482 | Dalton Trans., 2018, 47, 16480–16491
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
XRD pattern of BEHPAl21 exhibits a broad halo at 2θ angle
around 20° that can be attributed to the presence of the amor-
phous phase, as well as some sharp signals characteristic of
crystalline regions.
To gain a greater understanding of the nature of solid state
thermal transitions of DOPAls, we carried out powder X-ray
diffraction studies at elevated temperatures. These PXRD
measurements were performed at temperatures of a few
Celsius degrees below and above the area of the respective
endothermic DSC peak (Fig. S4–S6, ESI†). The analysis of diffr-
action data collected for DMPAl, DEPAl, DnPPAl and DBPAl
allows for two important conclusions. First of all, the presence
of the relatively well-formed X-ray diffraction peaks above the
temperature of the endothermic process is a clear and final
proof that this transformation is not related to the melting of
DOPAl crystalline domains and the resulting solid phase exhi-
bits a highly ordered structure. Secondly, the evident simi-
larities between the shapes of the high-temperature PXRD pat-
terns and their analogues recorded at 25 °C suggest that at
elevated temperature DOPAls probably retain a hexagonal
close-packing of their polymeric chains. Taking this into
account, the visible shift of the highest-intensity diffraction
peak to the lower values of 2θ indicates an increase in the dis-
tance between the parallel DOPAl chains during heating. The
systematic studies of the temperature effect on the unit cell
parameters were performed for DnPPAl samples, based on
their VT-PXRD patterns (Fig. S7 and S8, ESI†) measured
between 25 and 205 °C in two different temperature modes
(heating or cooling). The results of those studies are graphi-
cally presented in Fig. 4, whereas the numerical data are
enclosed in Tables S1 and S2 (ESI†).
Fig. 2 DSC thermograms of DMPAl (A) and DEPAl (B). The 1st heating
(red) and cooling (blue) curves are presented. Enthalpies (in kJ mol−1) of
the first-order transitions are given in parentheses.
As can be seen in Fig. 4, there are three different regions of
the variation of DnPPAl unit cell parameters within the investi-
gated temperature range:
region A from room temperature to ca. 175 °C,
region B between approximately 180 and 190 °C,
region C above ca. 190 °C.
In the first region, the heating of the DnPPAl sample
Fig. 3 DSC trace of BEHPAl. The 1st heating (red) and cooling (blue) resulted in a gradual expansion of its unit cell volume (V), as
curves are presented. Tinfl denotes the temperature of the inflection point.
well as the lattice parameters a and c. The almost linear and
continuous character of those changes, especially in the
cooling mode, is indicative of the absence of any polymorphic
derivative, or −62 and −32 °C in the case of DEPAl. transitions. The eqn (S1)–(S6) (ESI†) obtained on the basis of
Furthermore, their enthalpies are rather small and do not the experimental data from the A region can be utilized for the
exceed 3.2 kJ mol−1. In the case of DBPAl, two endothermic calculation of both the linear thermal expansion coefficient
processes were observed at 13 and 77 °C (Fig. S2, ESI†). Optical along a specified crystallographic axis (αL) and the volumetric
microscopy observations indicate that all of these thermal thermal expansion coefficient (βV). More specifically, the
transitions proceed without changing the physical state of the slopes of the straight lines described by those equations can
solid DOPAl samples. In contrast, the sample of BEHPAl give the values of αL or βV since they are defined as24,25
becomes an isotropic liquid at about 140 °C and its DSC trace
βV ¼ ðVref Þꢀ1 ꢁ ðdV=dTÞ
αL ¼ ðLref Þꢀ1 ꢁ ðdL=dTÞ
ð1Þ
shows a set of endothermic transitions in the range of
60–140 °C (total enthalpy of 16.6 kJ mol−1) (Fig. 3). One can
also observe a second-order phase transition at −78 °C, which
ð2Þ
has its equivalent on the cooling curve (shifted to the lower where Vref and Lref are a given unit cell volume (reference
temperatures by ca. 10 °C) and is most likely related to the volume) and a given lattice parameter (reference length) in
change in heat capacity of the amorphous phase. Actually, the some particular crystallographic direction, respectively. It
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16480–16491 | 16483

View Article Online
Paper
Dalton Transactions
Table 2 Coefficients of the volumetric and linear thermal expansion of
the low-temperature DnPPAl phase existing over the temperature range
of 298.15–448.15 K a
Thermal coefficient αa × 105/Å K−1 αc × 105/Å K−1 βV × 105/Å3 K−1
Heating mode
Cooling mode
12.6(5)
12.3(4)
14.4(11)
13.9(5)
40.4(10)
39.4(8)
a 298.15 < T < 448.15 K for the VT-PXRD measurements in a heating
mode and 298.15 < T < 458.15 K for the same measurements in a
cooling mode.
compound. The temperature window in which such a phenom-
enon occurs is rather small and one can already observe only
the high-temperature phase of aluminum tris(di-n-propyl-
phosphate) (HT-DnPPAl) at a temperature of around 195 °C or
higher (region C in Fig. 4).
In comparison with LT-DnPPAl, the values of both the
a-parameter and unit cell volume of HT-DnPPAl are increased
by approximately 2.4 and 4.5%, respectively, whereas the
c-parameter is slightly lower and reaches the values observed
at room temperature (see the set of data for T = 185 °C in a
cooling mode, Table S2 in the ESI†). Based on this, one can
state that above 185 °C the DnPPAl unit cell is further
expanded along the a-axis, while getting more flat. However, in
our opinion the degree of those changes does not imply a
large rearrangement of atoms in the DnPPAl crystal lattice and
the formation of a new crystalline phase (polymorphic tran-
sition). A more plausible explanation is formulated if one
assumes that both the molecular architecture of a single
DOPAl chain and its interactions with neighbouring chains are
effectively altered by the conformational transformations of
organic groups R in the (RO)2PO2 ligands involving rotations
around the C–C, O–C and P–O bonds. One can expect that
those rotations can be frozen at a sufficiently low temperature
and such a phenomenon occurred in the case of the DEPAl
crystal cooled to −173 °C. At this temperature, in one popu-
lation of DEPAl chains (A-type chains) all alkyl substituents are
in a low-energy gauche–gauche conformation characterized by
the smallest strains in the phosphate tetrahedron, whereas the
second population of DEPAl chains (B-type chains) exhibits
less thermodynamically favourable gauche–trans arrangement
of ethyl groups.21 Since the C–C and C–O rotations should be
characterized by lower energy constraints (activation energy),
the low-temperature transitions observed in the DSC thermo-
Fig. 4 The effect of temperature on the DnPPAl unit cell parameters a,
c and volume V. The areas of occurrence of the low-temperature (A) and
high-temperature DnPPAl phase (C) are indicated; B is a transition
●
region. Different phases of DnPPAl are marked with ✦ and . The modes
of the VT-PXRD measurements are indicated by colours: red for a
heating (25 → 205 °C) and blue for a cooling (205 → 25 °C) mode.
should be noted that Vref and Lref are estimated at some gram of DnPPAl or several other DOPAls are probably related
chosen reference temperature (Tref) and in our studies those to the overcoming of energy barriers associated with coopera-
were the values of the DnPPAl unit cell parameters (a, c and V) tive movements of carbon and oxygen atoms. However, one
measured at Tref = 25 °C (298.15 K). The results of such calcu- cannot completely exclude the possibility that rotations
lations are presented in Table 2. Their analysis shows that up around the P–O bonds can also be involved.
to 175 °C some anisotropy in the DnPPAl thermal expansion
occurs, with the c-parameter being the softest against heating groups in the DOPAl chain will not only alter its dimensions,
(αc > αa). especially the diameter, but also modify the force of its van
It is logical to assume that a different orientation of organic
Above 175 °C, a transition region (denoted as B in Fig. 4) der Waals interactions with neighbouring chains. All of this
begins in which the low-temperature DnPPAl phase should result in a non-continuous change of the a-parameter,
(LT-DnPPAl) coexists with another crystalline phase of this which is strictly dependent on the distance between DOPAl
16484 | Dalton Trans., 2018, 47, 16480–16491
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Table 3 TGA/DTG data of DOPAl samples measured in air
chains. These phenomena are most likely responsible for the
differences in the a-parameter values occurring between
LT-DnPPAl and HT-DnPPAl phases.
Sample
T5 a/°C
T10 a/°C
Tmax b/°C
ΔmOX,600 c/wt%
One can also expect that the changes in arrangement of
organic ligands will affect the otherwise rigid inorganic core of
the DOPAl chain formed by AlO6 octahedra and PO4 tetrahe-
dra. This may be performed either by inflicting strains in the
phosphate tetrahedron and altering its geometry, or by chan-
ging the angle at which the O–P–O bridges are inclined with
DMPAl
DEPAl
DnPPAl
DiPPAl
DBPAl
BEHPAl
DPhPAl
295.6
251.8
270.7
274.0
269.0
220.6
340.0
304.0
257.5
272.6
278.0
269.6
230.9
377.2
325.2
55.23
54.95
59.60
56.15
64.79
74.65
73.65
274.4d, 287.8
272.4d, 284.6
286.0
270.0
258.3
411.6d, 417.1, 431.3
respect to the c-axis of the DOPAl unit cell. Regardless of the a T5 and T10 are temperatures at which 5% and 10% of weight loss
occurred, respectively. b Temperature at the maximum decomposition
exact mechanism of this process, it will only slightly modify
rate derived from the DTG curve (see Fig. S9 and S10 in the ESI).
c Weight loss of the sample heated to 600 °C in air (as % of the initial
the locations of aluminum ions within DOPAl chains. Since
the Al⋯Al distance equals 0.5 c,21 every change in the confor- sample mass). d Refers to the main weight loss.
mation of the organic group will influence not only the a-para-
meter of the DOPAl unit cell, but also its repeatability along
the c-axis. From this point of view, the behaviour of DnPPAl at
elevated temperatures can be explained as follows: at about
175 °C the diminishing van der Waals interactions between
DnPPAl chains allow for another endothermic rearrangement
of their organic ligands and moving them apart (expansion
along the a-axis), which is accompanied by a relaxation of
strains within the AlO6 and PO4 polyhedra and return of alumi-
num centers to closer positions within a single DnPPAl chain.
However, this is only a hypothesis that needs further experi-
mental verification.
In comparison with n-alkyl derivatives, DPhPAl and DiPPAl
containing sterically bulky phenyl and isopropyl groups prob-
ably need more space and energy for conformational move-
ments and do not reveal thermal transitions below 200 °C (see
Fig. S3, ESI†).
worth noting that DOPAls with 2-ethylhexyl or methyl groups
exhibited different behaviour and after decomposition began
to stabilize their weight at much lower (BEHPAl) or higher
(DMPAl) temperatures (Fig. S9, ESI†). Among all of the investi-
gated compounds DPhPAl seems to be the most resistant to
the thermo-oxidative decomposition since both its T5 and
Tonset temperatures exceeded 380 °C, whereas its sample
ceased to change weight at approximately 520 °C. It is interest-
ing that DPhPAl started losing its weight at temperature as low
as 238 °C, although the volatilization of the first 5 wt% of its
sample proceeded slower and was more energy-consuming
than for the other DOPAls (Fig. S10, ESI†). The measured total
weight loss during the heating of DOPAls in air up to 600 °C
depended on the type of organic ligand and reached approxi-
mately 55–75% (Table 3).
The collected data let us conclude that most of the investi-
gated DOPAls show the positional and orientational order
while heating up to 200 °C. Only chains bearing long 2-ethyl-
hexyl side groups can gain sufficient translational mobility to
generate a liquid phase.
For the purpose of further studies on thermal properties of
DOPAls, we carried out their pyrolysis in argon. It allowed us
not only to establish their resistance to heating under an inert
atmosphere, but also to investigate and identify the volatile
products formed during their decomposition – the latter due
to a simultaneous analysis of the evolved gas in a quadrupole
mass spectrometer. The results of such measurements are
summarized in Table 4 and Fig. S11–S17 in the ESI.†
Thermal degradation of DOPAls above 200 °C
The results of thermogravimetric analysis show that under an
oxidative atmosphere BEHPAl exhibited the lowest thermal
stability, since its sample began to decompose around 200 °C
(a bend point on the TG curve) and stabilized its weight at ca.
269 °C (Fig. S9, ESI†). The same conclusion can be derived if
other typical criteria for thermal stability of compounds, such
as the temperature corresponding to 5% of weight loss (T5,
Table 3) or the extrapolated onset temperature of degradation
(Tonset, Fig. S9, ESI†), are considered. The susceptibility of
BEHPAl for thermolysis is most likely due to the large number
of easily oxidized C–H bonds present in the compound’s struc-
ture, as well as the fact that it underwent melting before
decomposition. The other DOPAls were more stable as their
weight loss started between 235 and 265 °C, with Tonset tem-
perature being usually from 5 to 35 °C higher than the respect-
ive bend point on the TG curve (Fig. S9, ESI†). For the majority
of samples containing alkyl substituents, the main weight loss
process ended around 300 °C, with the extrapolated endset
temperature (Tendset) within the range of 283.2–291.0 °C. It is
Table 4 TGA/DTG data of DOPAl samples measured under an argon
atmosphere
Sample
T5 a/°C
T10 a/°C
Tmax b/°C
ΔmAr,600 c/wt%
DMPAl
DEPAl
DnPPAl
DiPPAl
DBPAl
BEHPAl
DPhPAl
339.4
320.0
308.7
258.1
289.5
218.0
384.8
349.7
324.0
312.7
258.8
301.0
226.2
394.3
368.8
334.0
318.7
259.1
313.5
256.4
413.1
59.83
60.93
66.96
54.65
65.28
73.10
78.07d
a T5 and T10 are temperatures at which 5% and 10% of weight loss
occurred, respectively. b Temperature at the maximum decomposition
rate derived from the DTG curve (see Fig. S11–S17 in the ESI). c Weight
loss of the sample heated to 600 °C under an argon atmosphere (as %
of the initial sample mass). d Weight loss of the sample heated to
800 °C under an argon atmosphere (as % of the initial sample mass).
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16480–16491 | 16485

View Article Online
Paper
Dalton Transactions
Based on the difference in a response of DOPAls to heating through a 4-center transition state, in which the hydrogen
under an oxidative atmosphere and in argon one can divide atom of the POH group resulting from olefin elimination
them into the following categories:
forms a bond with an oxygen atom of the remaining ethoxy
(a) DOPAls susceptible to thermo-oxidation, for which elim- group. The subsequent rupture and rearrangement of the
ination of oxygen from the atmosphere results in the increase PO–H and P–OEt linkages lead to elimination of ethanol.
in thermal stability, as indicated by higher values of both the
The TG-QMS data suggest that the pyrolytic decomposition
T5 and Tonset parameters – those are DMPAl, DEPAl, DnPPAl of other DOPAls containing aliphatic C2–C8 groups follows the
and DBPAl. It is also evident that under an inert atmosphere pathways established for DEPAl, although their main volatile
those compounds undergo a slower decomposition since their products usually are terminal olefins with only some traces of
characteristic temperatures related to the rate of the process alcohols (Fig. S12–S15, ESI†).26 It should be noted that in each
(T5, T10 and Tmax) are few tens of °C higher than the respective case water is also evolved and its QMS signal is detected at
parameters estimated in air (see Tables 3 and 4);
almost the same temperatures as analogous peaks characteriz-
(b) Compounds, for which oxidation seems to be practically ing alkenes. This suggests that most probably the conden-
irrelevant to the initiation of thermal degradation, causing sation of POH groups occurs immediately after their formation
almost no changes in the Tonset temperatures – this group in the system.
includes both BEHPAl and DPhPAl. It should be noted,
In the case of DMPAl, the gaseous products of thermal
however, that among the members of this group oxygen can decomposition contained both methanol (m/z = 15, 31 and
have a different impact on the rate of weight loss and the 32)26 and dimethyl ether (m/z = 29, 45 and 46)26 (Fig. S16,
temperature at which the sample weight reaches a constant ESI†). This result strongly indicates that, during heating, the
value. For example, under an inert atmosphere a DPhPAl P–OCH3 bonds may have undergone a homolytic cleavage.29,30
sample required heating to a higher temperature to lose 5% of Methoxyl radicals formed in such a reaction are susceptible to
its weight than during the measurement in air (384.8 °C versus further transformations that lead to cleavage of hydrogen or
340.0 °C, see Tables 3 and 4), and stopped changing its weight methyl groups.29 One can expect that this process requires
well above the temperature of 520 °C observed during thermo- more energy than β-elimination of alkene resulting in a better
oxidation (see Fig. S10 and S17, ESI†). On the other hand, thermal stability of DMPAl. As far as the pyrolysis process is
for BEHPAl the T5, T10 or Tmax parameters did not change concerned, another feature distinguishes DMPAl from other
significantly with lowering oxygen concentration (see DOPAls subjected to our studies – an apparent presence of
Tables 3 and 4);
QMS peaks assigned to TMP (molecular ion with m/z = 140)26
(c) DOPAls, for which degradation is hindered under an oxi- and some products of its thermal decomposition, e.g. di-
dative atmosphere. This situation applies to DiPPAl, for which methylphosphite (molecular ion with m/z = 110).26 There are
the range of decomposition (an area between Tonset and also mass spectrometry signals that can be attributed to vola-
Tendset) in argon is narrower and shifted to the lower tempera- tile oxophosphorus ionic species: PO (m/z = 47), PO2 (m/z = 63)
ture region compared to the measurements in air. Such and HOPO2 (m/z = 80).30 Based on this, one can assume that at
unusual behaviour is difficult to explain, but it may be related least a part of phosphorus content of DMPAl is volatilized and
to the physicochemical properties (e.g. thermal stability and removed from the sample during heating under an inert atmo-
volatility) of the products formed during thermo-oxidation.
sphere. The reaction pathway leading to the formation of TMP
Thermogravimetry coupled with quadrupole mass spec- is not clear. One of the plausible explanations might be the
trometry (TG-QMS) showed the presence of the peaks charac- homolytic substitution at the phosphorus center in the di-
teristic of both ethylene (m/z = 26, 27 and 28)26 and ethanol methylphosphate ligand.
(m/z = 31, 45 and 46)26 in the mass spectra of the volatile pro-
The mass spectrum of the volatile products of DPhPAl ther-
ducts of DEPAl pyrolysis (Fig. S11, ESI†). Based on this infor- molysis (Fig. S17, ESI†) contains only peaks assigned to OH
mation, one can assume that under such conditions thermal ions (m/z = 17) as well as phenol (m/z = 94).26 The intensity of
degradation of DEPAl begins similarly to the thermolysis of those signals is rather low compared to similar spectra of
any alkyl phosphate containing hydrogen atoms directly other DOPAls, suggesting lower concentration of ions and the
bonded to a β-carbon of an organic substituent, i.e., via pyrol- respective molecules, from which they originated. Furthermore,
ysis of the ester linkages and olefin elimination, as indicated it should be noted that even after prolonged heating at 600 °C, a
in eqn (3).27–29 This step is accompanied by the condensation solid residue from DPhPAl thermolysis contains a fraction of
of the newly formed POH groups, leading to water formation elemental carbon: 7.47 and 6.54 wt% after 6 and 24 h in argon,
(m/z = 18)26,28
respectively. Interestingly, the carbon deposit (5.73 wt%) was
detected even in the case of the DPhPAl sample thermally
ð3Þ degraded for 6 h at 600 °C under oxidative conditions.
Therefore, one can expect that the degradation of DPhPAl may
ꢂO2PðOCH2CH3Þ2 ! ꢂO2PðOHÞ2 þ 2CH2vCH2
The presence of ethanol in gaseous products of DEPAl occur via polyaromatic structures, which are not detected by the
pyrolysis can be explained by the occurrence of a process con- TG-QMS apparatus.
current to β-scission of olefin. As in the case of thermal degra-
The measured total weight loss during the heating of the
dation of ethyl esters of phosphoric acid,28,29 it may proceed analyzed DOPAls up to 600 °C depended on the type of organic
16486 | Dalton Trans., 2018, 47, 16480–16491
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
ligand present in their structure and reached approximately
55–75%. Changing of the atmosphere during DOPAl heating
from oxidative to inert either contributed significantly to the
decrease in the weight of the solid residue (DMPAl, DEPAl,
DnPPAl), or practically had little or no effect on it (DiPPAl,
DBPAl, BEHPAl).
Composition of the solid residues from DOPAl thermolysis
The solid products of DOPAl thermal decomposition were
studied by means of infrared and MAS NMR spectroscopy, as
well as the PXRD technique. We found that the FTIR spectra of
these products did not contain signals characteristic of the
C–H, C–C or O–H bond vibrations (Fig. S18, ESI†). Instead,
over the spectral range of 400–1300 cm−1, one can observe very
strong and broad absorption bands that originate from
vibrations of the Al–O or P–O linkages in the inorganic alumi-
num phosphates.31,32 As can be seen in Fig. S18 (ESI†), the
infrared spectra of thermolyzates obtained from DOPAls con-
taining C2–C8 aliphatic groups are very similar to each other
both in the number and location of absorption bands,
although some of those signals can differ in intensity and
resolution (compare spectra of the solid residues from DiPPAl
and DEPAl). This suggests that all of those samples may be
characterized by the same, or very similar composition, but
probably differ in terms of crystallinity or amorphous phase
content. Based on the literature spectroscopic data32 one can
identify in the analysed samples one or more types of con-
densed aluminum phosphates, for example aluminum meta-
phosphates. This hypothesis is supported by the occurrence of
absorption signals located between 700 and 900 cm−1 (sym-
metric stretches of the P–O–P bridge), as well as the existence
of a strong IR band at 1279 cm−1 resulting from the νas mode
of the “external” (non-bridging) P–O bonds.32 Furthermore, in
the 1200–1010 cm−1 region some additional bands are
detected which may be attributed to the P–O stretching
vibrations of a mixed symmetric–antisymmetric origin occur-
ring in an “external” PO2 group of the metaphosphate
ligand.33
Fig. 5 PXRD patterns of the solid residues obtained in the course of
DOPAl thermal degradation carried out for 6 h in air at 600 °C (black
lines), and aluminum cyclohexaphosphate standard (Powder Diffraction
File No 39-0159,35 red bars and stars). Reflexes characteristic of alumi-
num metaphosphate type A (Powder Diffraction File No 13-0430)35 are
indicated with a ( ) symbol.
Regarding the possible structure of DOPAl pyrolyzates, a phous phase beside crystalline domains. In each DOPAl pyroly-
very interesting feature of their FTIR spectra is the simul- zate the amount of vitreous phase is different: the smallest for
taneous appearance of strong bands located in the DiPPAl and DBPAl, whereas samples obtained from DMPAl
950–970 cm−1 region and at ca. 1000 cm−1. These signals arise and DPhPAl are almost entirely amorphous. These data are in
from antisymmetric stretching vibrations of bonds in the P–O–P good agreement with the results of FTIR analysis, i.e. explain
bridge of the metaphosphate ligand – the one with higher the differences in development of some particular infrared
frequency is characteristic of ring structures, whereas the other absorption bands – a broadening of FTIR signals is a clear
may indicate the presence of chain-like moieties.32,33 Such a indication of increasing amorphousness of the sample.32
phenomenon suggests that the solid residues from DOPAl Furthermore, the PXRD measurements also confirm FTIR find-
thermo-oxidation are in fact a mixture of aluminum metapho- ings regarding the composition of the crystalline fraction of
sphates. To investigate further this possibility we carried out DOPAl pyrolyzates. The majority of PXRD reflections observed
the PXRD measurements – their results are presented in Fig. 5 in Fig. 5 can be attributed to aluminum cyclohexaphosphate,
and Tables S4–S11 (ESI†).
Al2P6O18, consisting of cyclic, hexameric anions [P6O18]6− iso-
As can be seen in Fig. 5, almost all of the obtained PXRD lated from each other and sharing corners with AlO6
diffractograms contain multiple reflexes which suggests a high octahedra.34,35 The other diffraction signals occurring on the
degree of structural order within the samples. However, a PXRD diffractograms can be also assigned to crystalline forms
characteristic broad halo on the PXRD baseline with of aluminum metaphosphate,35 for example its polymorph A
maximum around 2θ 24° indicates the formation of an amor- (Powder Diffraction File No 13-0430).36 In accordance with
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16480–16491 | 16487

View Article Online
Paper
Dalton Transactions
those data are the results of the MAS NMR measurements con-
ducted for such a mixture. As can be seen for DiPPAl thermoly-
zate, its 27Al MAS NMR spectrum (Fig. S19 in the ESI†) shows
two rather sharp resonance signals at −23.65 ppm and
30.09 ppm. The former peak is located within the spectral
region characteristic of six-coordinated aluminum sites of
octahedral symmetry37–43 and can be assigned to the Al2P6O18
phase, whereas the latter is characteristic of the AlO4
tetrahedra37–43 and results from other phases of aluminum
metaphosphate. The 31P MAS NMR spectrum of the same
sample (Fig. S19 in the ESI†) contains several signals indicat-
ing the presence of at least 4 populations of inequivalent phos-
phorus nuclei; however, their chemical shifts (δP between −40
and −50 ppm) are within the range characteristic of tetrahedral
phosphorus in aluminophosphate glasses40 or polycrystalline
aluminum metaphosphates.44,45 A relatively small difference
in the location of those peaks can be explained by a bonding
of the phosphate tetrahedron with different types of aluminate
polyhedra, as well as a variation in the number of neighbour-
ing tetrahedral phosphate species connected to a single PO4
tetrahedron.37,39,40,43
Fig. 6 SEM images of the solid residues from DOPAl thermal degra-
dation carried out in air at 600 °C. Sample obtained from: (A, B) DMPAl,
(C, D) DEPAl, (E, F) BEHPAl.
It should be noted that changing of the atmosphere during
pyrolysis from oxidative to inert seems to have no influence on
the composition of crystalline DOPAl pyrolyzates, as evidenced
by the diffraction patterns of BEHPAl degraded in argon (see
Tables S10–S11 and Fig. S20 in the ESI†).
shows that the particles present in DOPAl thermolyzates are
generally of the same type (for example rod-like) as in the start-
ing material, as far as the shapes are considered. The most
important morphological differences occur in the small scale
in the form of nanometric pores, most probably resulting from
the liberation of gaseous products of thermal degradation – it
is especially evident in the case of DEPAl (Fig. 6C and D) or
DnPPAl (Fig. S21A and S21B, ESI†), and can be also observed
for DMPAl (Fig. 6B). The formation of a porous structure is
additionally confirmed by the TSSA measurements. For
example, in the course of thermolysis the initial TSSA value of
DMPAl (0.8 m2 g−1) increased to 10.7 m2 g−1, whereas in the
case of DEPAl those changes were relatively smaller (ΔTSSA =
500%, from 1.7 to 10.2 m2 g−1).
It should be noted that in the case of some of the DOPAl
thermolyzates one can observe not only the rod-like particles
but also a population of those exhibiting a continuous surface
– a good example can be DiPPAl (Fig. S21C and S21D, ESI†).
Another remarkable thing concerns the apparent discrepancy
between a high degree of morphological order observed in the
case of SEM images of DMPAl (Fig. 6A and B) or DPhPAl
(Fig. S22, ESI†) thermolyzates and their obvious amorphous-
ness indicated by the PXRD studies (Fig. 5): one could expect
that the lack of crystallinity should prevent the formation of
particles with well-defined shapes.
On the other hand, the 27Al and 31P MAS NMR spectra of
the amorphous DPhPAl thermolyzates (Fig. S19 in the ESI†)
show that their structure depends on the pyrolysis conditions.
In the oxidative environment DPhPAl decomposes to a mixture
of aluminum metaphosphates, in which octahedral Al sites
prevail resulting in a broad resonance observed at −23.47 ppm.
The occurrence of some additional signals located around or
above 0 ppm can be attributed to the presence of some other
aluminate moieties, e.g. the AlO5 and AlO4 polyhedra.37,40 These
Al-containing species are in part responsible for a large number
of signals observed in the 31P MAS NMR spectrum of that
sample (a middle chart on the left, in Fig. S19, ESI†). The
removal of oxygen from the atmosphere results in the increase
in the amount of tetrahedral aluminate species (a peak at δAl
around 30 ppm, see a top chart on the right in Fig. S19, ESI†),
and a shift of the 31P MAS NMR signals to lower frequencies
(see a top chart on the left in Fig. S19, ESI†).
At this point, we cannot give a satisfactory explanation for
the role of oxygen in the mechanism of thermal degradation of
DOPAls. One can assume that oxygen has no effect on the
β-elimination process followed by the formation of crystalline
domains of aluminum metaphosphate. However, oxygen may
take part in radical reactions and thereby affect the formation
of the amorphous phase of DPhPAl thermolyzates.
Morphology of DOPAl thermolyzates
Properties of carbon deposits present in DPhPAl pyrolyzates
Scanning electron microscopy revealed that the thermal treat- A physical state (black or grey solids) and the previously men-
ment at 600 °C only partially influenced the morphology of tioned results of elemental analysis indicated the presence of
DOPAl samples. A comparison of their SEM images obtained elemental carbon in DPhPAl thermo- and pyrolyzates. This
before21 and after thermolysis (Fig. 6 and S21–S22, ESI†) prompted us to investigate some properties that might be con-
16488 | Dalton Trans., 2018, 47, 16480–16491
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
nected with such a carbon deposit. Since Raman spectroscopy traces of humidity. It should be noted that these were only pre-
is a valuable analytical tool enabling structural analysis of liminary studies that indicate the possibility of enhancement
carbon-based materials containing carbon atoms in sp2 of DPhPAl pyrolyzates’ conductivities by a proper optimization
hybridization,46–50 we applied this technique in our studies. As of their synthesis.
can be seen in Fig. S23 (ESI†) the Raman spectra of the solid
residues from DPhPAl thermolysis contain two distinctive
absorption bands located at approximately 1596 and
Conclusions
1340 cm−1. The former signal is characteristic of the G band of
a highly ordered graphene or graphite (the in-plane Csp2–Csp2 We have shown that the mode of thermally induced structural
stretching mode), whereas the latter one can be ascribed to the transformation and decomposition of inorganic polymers
D band originating from defects and structural disorder in the based on aluminum tris(diorganophosphates) depends signifi-
graphene layers.46–49 The complete lack of the band positioned cantly on the type of organic side group. DOPAl bearing long
at 2700 cm−1 (namely the 2D or G′ band)50 suggests that the 2-ethylhexyl groups behaves in a similar way to semicrystalline
obtained carbon deposits did not contain pristine graphene or organic polymers: it becomes liquid at about 140 °C and exhi-
graphite domains.47 In fact, our experimental Raman spectra bits a second-order phase transition resembling the glass tran-
of DPhPAl thermolyzates closely resemble those of graphene sition in organic polymers. The other DOPAls preserve hexa-
oxide (GO) or reduced graphene oxide (rGO) indicating that gonally closely packed crystalline domains at least up to
one or both of those materials may be present in the 180–200 °C. Most of them, however, are conformationally
samples.47,48
mobile and the structural disorder is induced by the rotation
It is well known that electrical properties of GO strictly around the P–O, O–C and C–C bonds in organic ligands occur-
depend on the number of oxygen-containing functional ring at a sufficiently high temperature. This results in an
groups causing disorder in graphene layers: pure GO is increase in the DOPAl interchain distance and thermal expan-
thought to be an insulator and requires transformation to rGO sion of its unit cell dimensions. It may even lead to some
to gain electrical conductivity.51,52 We have examined conduc- rearrangement of the rigid inorganic core of DOPAl chains at
tivity of DPhPAl pyrolyzates and the results of such measure- temperature well above 100 °C due to a reversible deformation
ments are presented in Fig. 7. As can be seen these materials of the PO4 and AlO6 polyhedra and modification of the way
exhibit rather low electrical conductivity (values between 10−6 they are connected: for DnPPAl such a phenomenon has been
and 10−8 S cm−1), although it increases with temperature. detected at ca. 185 °C.
Taking into account a very low content of pyrolytic carbon in
The pyrolytic decomposition of DOPAls containing aliphatic
the sample (below 7.5 wt%) one can assume that it is formed C2–C8 substituents proceeds mainly by β-elimination of olefins
by a well-conducting rGO phase that was deposited on the and is followed by condensation of the newly formed P–OH
surface of insulating aluminum metaphosphate grains. groups. These chemical transformations lead to the formation
Although not studied in detail, we think that electronic con- of condensed aluminum phosphates, with the aluminum
ductivity dominates in the prepared samples due to their cyclohexaphosphate (Al2P6O18) crystalline phase being particu-
chemical nature; however, some ionic contribution may also larly common. Therefore, this type of DOPAl might serve as
arise from the presence of polar functional groups and/or the precursors of porous ceramic materials, which to some extent
preserve the original shape of particles present in the sub-
strate. However, their practical application as flame retardants
might be limited due to the insufficient thermal stability and
generation of olefins upon heating. The methyl and phenyl
derivatives exhibit significantly higher thermal stability and
their pyrolysis leads to elimination of phosphorus-containing
moieties, which may improve the flame retardancy.53 The
mechanism of such transformations is not elucidated yet. One
can only suggest that it involves a homolytic scission of the
P–OC bonds. Furthermore, the pyrolysis of DPhPAl results in
the formation of a conducting carbon deposit on the alumi-
num phosphate core. Although the conductivity of those
materials, being in the range of 10−6–10−8 S cm−1, is not very
impressive one can consider them as potential antistatic
agents to be used in polymer composites. We have found that
both the structure and conducting properties of carbonaceous
materials obtained by pyrolysis of hybrid polymers based on
metal organophosphates decorated by aromatic substituents
Fig. 7 Temperature dependence of the conductivity of DPhPAl pyroly-
may vary significantly depending on both the type of metal
center and organic ligand.
zates obtained after: 6 h (black line and squares) or 24 h (red line and
circles) of pyrolysis at 600 °C in argon.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16480–16491 | 16489

View Article Online
Paper
Dalton Transactions
21 M. Dębowski, K. Łokaj, A. Wolak, K. Żurawski, A. Plichta,
J. Zachara, A. Ostrowski and Z. Florjańczyk, Dalton Trans.,
2016, 45, 8008.
Conflicts of interest
There are no conflicts to declare.
22 TOPAS V.5, Bruker AXS, Karlsruhe, Germany, 2014.
23 A. S. Bondarenko and G. A. Ragoisha, in Progress in
Chemometrics Research, ed. A. L. Pomerantsev, Nova Science
Publishers, New York, 2005, pp. 89–102 (the program is avail-
able online at http://www.abc.chemistry.bsu.by/vi/analyser/).
Acknowledgements
This research was partially funded by National Science Centre,
Poland within the OPUS-11 research scheme, project number 24 R. Hammond, K. Pencheva, K. J. Roberts, P. Moughin and
2016/21/B/ST5/00126. D. Wilkinson, J. Appl. Crystallogr., 2005, 38, 1038.
The authors would like to thank Dr Eng. Magdalena Zybert 25 S. I. Sadovnikov, A. I. Gusev, A. V. Chukin and
from Faculty of Chemistry, Warsaw University of Technology,
for conducting the TSSA measurements.
A. A. Rempel, Phys. Chem. Chem. Phys., 2016, 18, 4617.
26 NIST Mass Spec Data Center, Stein, S. E. director, Mass
Spectra, in NIST Chemistry WebBook, NIST Standard Reference
Database Number 69 [Online], ed. P. J. Linstrom and
W. G. Mallard, National Institute of Standards and
Technology, Gaithersburg MD, can be found under http://
webbook.nist.gov/chemistry (accessed March 19, 2018). NIST
MS number: 133176 (1-butanol), 18918 (1-butene), 18911 (iso-
butylene), 78 (dimethyl ether), 18814 (ethylene), 118507
(ethanol), 2838 (2-ethyl-1-hexanol), 227612 (2-ethyl-1-hexene),
229809 (methanol), 133909 (phenol), 289584 (1-propanol),
50 (propene), 125515 (TMP), 230474 (dimethylphosphite).
27 E. J. P. Zegers and E. M. Fisher, Combust. Sci. Technol.,
1998, 138, 85.
Notes and references
1 K. Maeda, Y. Kiyozumi and F. Mizukami, J. Phys. Chem. B,
1997, 101, 4402.
2 T. Kimura, Chem. Mater., 2005, 17, 337.
3 J. J. Ma, J. X. Yang, Y. W. Huang and K. Cao, J. Mater.
Chem., 2012, 22, 2007.
4 H. Galina, M. Oleksy, M. Heneczkowski, Z. Florjańczyk,
M. Dębowski and K. Łokaj, EP Pat, 2628766B1, 2014.
5 Z. Florjańczyk, A. Wolak, M. Dębowski, A. Plichta,
J. Ryszkowska, J. Zachara, A. Ostrowski, E. Zawadzak and 28 P. A. Glaude, H. J. Curran, W. J. Pitz and C. K. Westbrook,
M. Jurczyk-Kowalska, Chem. Mater., 2007, 19, 5584. Proc. Combust. Inst., 2000, 28, 1749.
6 Z. Florjańczyk, M. Dębowski, A. Wolak, M. Malesa and 29 V. Lhomme, C. Bruneau, N. Soyer and A. Brault, Ind. Eng.
J. Płecha, J. Appl. Polym. Sci., 2007, 105, 80. Chem. Prod. Res. Dev., 1984, 23, 98.
7 B. Mossety-Leszczak, B. Strachota, A. Strachota, 30 O. P. Korobeinichev, V. M. Shvartsberg and S. B. Il’in,
M. Steinhart and M. Šlouf, Eur. Polym. J., 2015, 72, 238. Combust., Explos. Shock Waves, 1997, 33, 270.
8 B. Mossety-Leszczak and M. Włodarska, Polym. Compos., 31 S. Sambasivan and K. A. Steiner, Investigation of a New
2017, 38, 277.
Class of Disordered Inorganic Materials; Technical Report for
Air Force Office of Scientific Research, Arlington, VA, 2001,
and references therein.
9 G. P. Funkhouser, N. Tonmukayakul and F. Liang,
Langmuir, 2009, 25, 8672.
10 Z. Ding, B. Chen, J. Ding, L. Wang and Y. Han, J. Colloid 32 A. Rulmont, R. Cahay, M. Liegeois-Duyckaerts and P. Tarte,
Interface Sci., 2014, 425, 102. Eur. J. Solid State Inorg. Chem., 1991, 28, 207.
11 Y. W. Huang, J. J. Ma, J. X. Yang, K. Cao and Z. Lu, J. Phys. 33 J. M. Rojo, J. L. Mesa, L. Leyama and T. Rojo, J. Mater.
Chem. C, 2012, 116, 22518. Chem., 1997, 7, 2243.
12 K. Maeda, J. Akimoto, Y. Kiyozumi and F. Mizukami, 34 A. Oudahmane, A. Mbarek, M. El-Ghozzi and D. Avignant,
Angew. Chem., Int. Ed. Engl., 1995, 34, 1199. Acta Crystallogr., Sect. E: Struct. Rep. Online, 2010, 66, i17.
13 Z. Florjańczyk, A. Lasota, A. Wolak and J. Zachara, Chem. 35 Powder Diffraction File Database PDF-2, International
Mater., 2006, 18, 1995.
Centre for Diffraction Data, Newtown Square, Pennsylvania,
14 H. Tanaka and M. Chikazawa, J. Mater. Chem., 1999, 9, 2923.
USA, 1998. Cards 39-0159, 13-0430, 13-0266.
15 M. Schulz, M. Tiemann, M. Fröba and C. Jäger, J. Phys. 36 T.-H. Yoon, L.-Y. Hong and D.-P. Kim, in Advanced
Chem. B, 2000, 104, 10473.
Functional Materials, ed. H.-G. Woo and H. Li, Zhejiang
University Press, Hangzhou and Springer-Verlag Berlin
Heidelberg, 2011, ch. 4, p. 128.
16 Y. Huang and Y. Zhimin, J. Am. Chem. Soc., 2005, 127, 2731.
17 A. Vioux, J. Le Bideau, P. H. Mutin and D. Leclercq, Top.
Curr. Chem., 2004, 232, 145.
37 G. Tricot, B. Doumert, B. Ravel, M. Bria, J. Trebosc and
H. Vezin, Solid State Nucl. Magn. Reson., 2017, 84, 137.
18 Metal Phosphonate Chemistry: From Synthesis to Applications,
ed. A. Clearfield and K. Demadis, Royal Society of 38 M. Haouas, F. Taulelle and C. Martineau, Prog. Nucl. Magn.
Chemistry, Cambridge, 2011. Reson. Spectrosc., 2016, 94–95, 11.
19 K. J. Gagnon, H. P. Perry and A. Clearfield, Chem. Rev., 39 L. van Wüllen, S. Wegner and G. Tricot, J. Phys. Chem. B,
2012, 112, 1034. 2007, 111, 7529.
20 R. Murugavel, A. Choudhury, M. G. Walawalkar, 40 L. van Wüllen, G. Tricot and S. Wegner, Solid State Nucl.
R. Pothiraja and C. N. R. Rao, Chem. Rev., 2008, 108, 3549.
Magn. Reson., 2007, 32, 44.
16490 | Dalton Trans., 2018, 47, 16480–16491
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
41 G. Tricot, L. Delevoye, G. Palavit and L. Montage, Chem.
Commun., 2005, 5289.
ed. J. I. Jang, Nova Science Publishers, Inc., New York,
2013, ch. 19, p. 403.
42 J. Schneider, S. L. Oliveira, L. A. O. Nunes, 48 H. J. Kim, S.-M. Lee, Y.-S. Oh, Y.-H. Yang, Y. S. Lim,
F. Bonk and H. Panepucci, Inorg. Chem., 2005, 44,
423.
D. H. Yoon, C. Lee, J.-Y. Kim and R. S. Ruoff, Sci. Rep.,
2014, 4, 5176, DOI: 10.1038/srep05176.
43 J. Tsuchida, J. Schneider, M. T. Rinke and H. Eckert, 49 M. Molenda, A. Chojnacka, P. Natkański, E. Podstawka-
J. Phys. Chem. C, 2011, 115, 21927.
44 J. Schneider, S. L. Oliveira, L. A. O. Nunes and
H. Panepucci, J. Am. Ceram. Soc., 2003, 86, 317.
45 T.-H. Yoon, L.-Y. Hong and D.-P. Kim, in Advanced
Functional Materials, ed. H.-G. Woo and H. Li, Zhejiang
Proniewicz, P. Kuśtrowski and R. Dziembaj, J. Therm. Anal.
Calorim., 2013, 113, 329.
50 Thermo Fisher Scientific Application Note 51901, https://
tools.thermofisher.com/content/sfs/brochures/D19504~.pdf,
(accessed May 2018).
University Press, Hangzhou and Springer-Verlag Berlin 51 X. Huang, F. Liu, P. Jiang and T. Tanaka, presented in part
Heidelberg, 2011, ch. 4, p. 103.
46 A. A. Dubale, W.-N. Su, A. G. Tamirat, C.-J. Pan,
at 2013 IEEE International Conference on Solid Dielectrics
(ICSD), Bologna, Italy, July, 2013.
B. A. Aragaw, H.-M. Chen, C.-H. Chen and B.-J. Hwang, 52 V. B. Mohan, R. Brown, K. Jayaraman and D. Bhattacharyya,
J. Mater. Chem. A, 2014, 2, 18383. Mater. Sci. Eng., B, 2015, 193, 49.
47 I. Childres, L. A. Jauregui, W. Park, H. Cao and Y. P. Chen, 53 L. Liu, Y. Huang, Y. Yang, J. Ma, J. Yang and Q. Yin, J. Appl.
in New Developments in Photon and Materials Research,
Polym. Sci., 2017, 134, 45445.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16480–16491 | 16491