Resource-Efficient High-Yield Ionothermal Synthesis of Microcrystalline Cu3-xP

Article abstract of DOI:10.1021/acs.inorgchem.6b01397

Polycrystalline Cu_3-xP was successfully synthesized in different ionic liquids comprising imidazolium and phosphonium cations. The reaction of elemental copper and red phosphorus in trihexyltetradecylphosphonium chloride at 200 °C led to single-phase Cu_3-xP (x = 0.05) within 24 h with a quantitative yield (99%). Liquid-state nuclear magnetic resonance spectroscopy of the ionic liquids revealed degeneration of the imidazolium cations under the synthesis conditions, while phosphonium cations remain stable. The solid products were characterized with X-ray powder diffraction, scanning electron microscopy, energy-dispersive X-ray spectroscopy, solid-state nuclear magnetic resonance spectroscopy, and elemental analysis. A reinvestigation of the electronic transport properties of Cu_2.95(4)P showed metallic behavior for the bulk material. The formation of CuP₂ during the synthesis of phosphorus-rich Cu_3-xP (x ≥ 0.1) was observed.

Full text of DOI:10.1021/acs.inorgchem.6b01397

Article
pubs.acs.org/IC
Resource-Efficient High-Yield Ionothermal Synthesis of
Microcrystalline Cu3−xP
Alexander Wolff,† Julia Pallmann,† Richard Boucher,‡ Alexander Weiz,† Eike Brunner,† Thomas Doert,*,†
and Michael Ruck†,#
‡
†Department of Chemistry and Food Chemistry and Department of Material Science, Technische Universitat Dresden, 01062
Dresden, Germany
̈
#Max Planck Institute for Chemical Physics of Solids, 01187 Dresden, Germany
S
* Supporting Information
ABSTRACT: Polycrystalline Cu3−xP was successfully synthe-
sized in different ionic liquids comprising imidazolium and
phosphonium cations. The reaction of elemental copper and
red phosphorus in trihexyltetradecylphosphonium chloride at
200 °C led to single-phase Cu3−xP (x = 0.05) within 24 h with
a quantitative yield (99%). Liquid-state nuclear magnetic
resonance spectroscopy of the ionic liquids revealed
degeneration of the imidazolium cations under the synthesis
conditions, while phosphonium cations remain stable. The
solid products were characterized with X-ray powder
diffraction, scanning electron microscopy, energy-dispersive
X-ray spectroscopy, solid-state nuclear magnetic resonance spectroscopy, and elemental analysis. A reinvestigation of the
electronic transport properties of Cu2.95(4)P showed metallic behavior for the bulk material. The formation of CuP2 during the
synthesis of phosphorus-rich Cu3−xP (x ≥ 0.1) was observed.
between approximately Cu2.90P and Cu2.75P.27,28 The copper
INTRODUCTION
■
deficit is accompanied by vacancies at two of the four
In the past decade the scientific community has begun to
crystallographic copper positions in the structure.27 Recently,
consider phosphorus to be a limited resource.1−3 Phosphorus,
de Trizio et al. identified the 6c Wyckoff positions as being
as a key element for living cells, plays a major role in food
most favorable for vacancies based on density functional theory
production. Approximately 80% of the globally mined
(DFT) calculations.29 They concluded that vacancies should
phosphate rock is used for the production of fertilizers.4
form even in Cu3P, which agrees with the intrinsic sub-
While other critical resources, such as natural oil or gas, may
stoichiometric nature of this compound as proposed by
become replaceable, this is not the case for phosphorus.5,6
Schlenger et al. and Olofsson.27,28 On this basis it seems
Thus, the waste of phosphorus should rigorously be reduced,
appropriate to generally refer to this compound as Cu3−xP. It
even in the production of inorganic phosphorus-containing
has been suggested that these vacancies are the origin of the
materials.
specific plasmon dynamics of Cu3−xP.29,30
One class of phosphorus-containing compounds is the binary
metal phosphides, which have a wide variety of structures,
There are different reports about the band gap of Cu3−xP in
literature. According to optical31 and scanning tunneling
interesting properties, and technical uses.7−9 They have been
spectroscopy,30 the gap is found between 0.8 and 1.3 eV.
studied intensively, for instance, as catalytic materials.10−13 The
Recently, p-type semiconducting behavior was claimed for
optical properties of Zn3P2 can be used in photovoltaic
Cu3−xP nanoparticles.32 However, observations of compara-
applications.14−16 Even superconductivity has been discovered
tively low electrical resistance33 and a Knight shift as well as
in the metal-rich molybdenum phosphides Mo3P, Mo8P5, and
Mo4P3.17 The main-group compounds InP and GaP are well-
known semiconductors used in several applications, for
example, as light-emitting diodes.18,19
The compound commonly referred to as Cu3P is a deoxidizer
for copper alloys and a sintering additive in metallurgy.20−23 It
also has, along with other metal or semimetal phosphides,
attracted attention as a potential candidate for anodes in
lithium ion batteries.24−26 The crystal structure of Cu3P was
described in 1972 by Olofsson.27 His investigations as well as
later work by Schlenger et al. revealed a homogeneity range
Korringa-type behavior of the spin−lattice relaxation rate in
solid-state nuclear magnetic resonance (NMR) spectroscopy34
indicate that the bulk material is metallic in nature.
In previous works, crystalline Cu3−xP was synthesized by the
direct reaction of the elements copper and red phosphorus in
sealed silica tubes at 560 °C for at least 5 d.28 Typically
byproducts, such as Cu or CuP2, depending on the ratio of the
Received: June 9, 2016
Published: August 25, 2016
© 2016 American Chemical Society
8844
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
Article
starting materials, were also obtained. To overcome this
problem, surface solid-state reactions were developed. In this
type of reaction, red phosphorus is suspended, e.g. in ethanol,
and then sprayed on a copper foil.24,35 The foil is fired at
temperatures of T ≥ 250 °C, leading to the formation of
Cu3−xP on the surface.25 The powder can then be scratched off
in order to obtain the product, however, the yield is limited by
the surface area of the copper foil. Another production route
involves a mechanochemical approach based on ball milling.
However, in addition to the presence of impurities, e.g. WC,
from the milling balls,26 the crystallinity of the product is rather
poor and an additional heat treatment may be necessary.35
Subsequently, solvothermal syntheses were developed as a
more convenient route for obtaining high quality Cu3−xP
material. The respective starting materials and reaction
conditions are listed in Table 1. In this kind of synthesis, the
EXPERIMENTAL SECTION
■
Starting Materials. The ionic liquids (ILs) 1-butyl-3-methylimi-
dazolium chloride ([BMIm][Cl], IoLiTec, 99%) and trihexyltetrade-
cylphosphonium chloride ([P66614][Cl], CYPHOS, 95%) were dried at
110 °C for 12 h under vacuum. Copper (ABCR, 99.9%, spherical, 100
mesh) was treated with H2 at 400 °C before use. Following a
procedure described by Brauer et al., red phosphorus (ABCR,
99.999%) was washed with sodium hydroxide, rinsed with refluxing
water, and finally dried in vacuum.45
Synthesis. All compounds were handled in an argon-filled
glovebox (M. Braun; p(O2)/p0 < 1 ppm, p(H2O)/p0 < 1 ppm). In a
typical synthesis, 381 mg (6.00 mmol) of copper and 63 mg (2.03
mmol) of red phosphorus were carefully weighed directly into a glass
flask with a magnetic stirring bar (final ratio Cu/P = 2.95:1). Note that
any loss of phosphorus, which is a sticky powder, causes impurities of
unreacted copper in the product, whereas an excess of it leads to
impurities of CuP2. The starting materials were covered with either
1310 mg (7.5 mmol) of [BMIm][Cl] or 1298 mg (2.5 mmol) of
[P66614][Cl]. The flask was closed with a glass stopper and
subsequently placed in an oil bath. The flask was quickly heated to
200 °C and then stirred vigorously. After the mixture was stirred for 24
h, the flask was removed from the bath and allowed to cool to room
temperature. To isolate the product the mixture was dispersed in
ethanol (96%), centrifuged (4400 rpm), and then redispersed in fresh
ethanol with the help of ultrasound. This procedure was repeated three
times. Afterward the obtained powder was dried in air at 60 °C for 12
h. According to powder X-ray analysis and energy-dispersive X-ray
(EDX) spectroscopy, single-phase samples of Cu2.95(4)P are obtained
with yields of 372 mg (84%) for [BMIm][Cl] and 438 mg (99%) for
[P66614][Cl].
Elemental Analysis. For the elemental analysis of Cu, P, Cl, C, O,
and Na, combinations of different methods were used. The analysis of
Cu, P, and Na was made with inductively coupled plasma atomic
emission spectroscopy using a 5100 SVDV (Agilent). A C200 CHLH
(LECO) was used for the detection of organic carbon (combustion
method). The oxygen content was probed with the carrier gas hot
extraction method using a TCH 600 (LECO). The chloride analysis
was made using the combustion ion chromatography method (AQF-
2100, Mitsubishi).
Powder X-ray Diffraction. Powder X-ray diffraction (PXRD) was
performed at 296(1) K on an X’Pert Pro MPD diffractometer
(PANalytical) equipped with a curved Ge(111) monochromator by
using Cu Kα1 radiation (λ = 154.056 pm). Le Bail analysis was done
within the software package Jana200646 using an internal LaB6
standard (660b, National Institute of Standards and Technology,
USA).
Scanning Electron Microscopy. The dried as-prepared samples
were dispersed in ethanol, and a few drops were placed on polished
silicon wafers. After evaporation of the ethanol, the wafers were fixed
on a sample holder with a graphite pad. Scanning electron microscopy
(SEM) was performed by using an SU8020 instrument (Hitachi) with
a triple detector system for secondary and low-energy backscattered
electrons (Ue = 3 kV).
Table 1. Selected Solvothermal Syntheses of Crystalline
Cu3−xP
starting materials
conditions
solvent
ref
a
b
CuCl, TOP
350 °C, 0.8 h
230 °C, 0.25 h
220 °C, 2 h
140 °C, 10 h
180 °C, 12 h
140 °C, 12 h
190 °C, 10 h
200 °C, 24 h
200 °C, 24 h
200 °C, 24 h
OLA , ocytlamine
29,32
30
c
CuCl, PH3
OLA, TOPO
Cu, Pwhite
octadecene/toluene
aqueous NH3
glycol/ethanol/water
ethylenediamin
ethylenediamin
H2O
36
CuCl2·2H2O, Pwhite
CuSO4·5H2O,Pwhite
CuCl2·2H2O, Pwhite
Cu, Pred
37
39
38
35
Cu, Pred
31
CuCl, Pred
H2O
31
CuCl2, Pred
H2O
31
CuI, Pred
200 °C, 24 h
H2O
31
a
b
c
Trioctylphosphine. Oleylamine. Trioctylphosphine oxide.
choice of the starting material is essential. For example,
nucleation has to be controlled during the synthesis of
nanoparticles. Therefore, at least one precursor must be
sufficiently soluble in order to achieve this control. This was
accomplished by using toxic phosphorus reagents like
phosphane30 or white phosphorus36 as well as with different
copper precursors.29,30,32,36 During the synthesis of nano-
particles, a substitution of this toxic reagent with trioctylphos-
phine29,32 was feasible, but the temperature had to be increased
to achieve decomposition of the phosphorus precursor.
Especially during the synthesis of bulk material with white
phosphorus, a considerable excess had to be used to obtain
single-phase products.37−39
Cu3−xP powders were also successfully synthesized using
nontoxic red phosphorus.31,35 Aitken et al. showed that
replacement of ethylenediamine by pure H2O as the solvent
had beneficial effects on the synthesis. In their approach various
copper precursors were employed; however, a sizable excess of
red phosphorus was necessary in all cases. The smallest excess
(2 equivalents) was needed for the reaction with CuCl, and the
largest excess had to be used for the reaction with copper metal
(30 equiv.).
Energy-Dispersive X-ray Analysis. The samples were pressed
into pellets with a diameter of 6 mm under vacuum with a load of 2
tons. The pellets were then embedded under vacuum into an epoxy
polymer in which graphite was dispersed. Afterward, the pellets were
wet-ground with a MetaServ 250 (Buehler, silicon carbide grinding
paper) and subsequently polished with a VibroMet 2 (Buehler,
MasterPrep alumina suspension 0.05 μm). The polished pellets were
finally fixed on the sample holders with a graphite pad. The
compositions of the samples were determined by quantitative EDX
(Ue = 20 kV) analysis using an SU8020 instrument (Hitachi) equipped
with a Silicon Drift Detector X-MaxN (Oxford). In light of the
preparation method, the elements C and O (both part of the polymer)
and Al (from polishing suspension) were omitted in EDX
quantifications.
Here we report a convenient and resource-efficient route to
synthesize microcrystalline Cu3−xP from nontoxic red phos-
phorus at low temperatures with a very high yield, which should
also be up-scalable for industrial processes. This method
combines a relatively short reaction time and the pressureless
setup of ionothermal40−44 syntheses with the high atomic
efficiency of solid-state reactions.
Nuclear Magnetic Resonance Spectroscopy. All liquid-state
NMR experiments were performed using an Avance 400 (Bruker)
8845
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
Article
spectrometer at 348(2) K. The samples were transferred into tubes
with polytetrafluoroethylene valves (Deutero) in an argon-filled
glovebox. The pure ILs were dried before use as described above;
the reacted ILs were transferred into the glovebox without exposure to
air. The 31P NMR spectra were recorded at a resonance frequency of
1
161.98 MHz, and the H NMR spectra were recorded at a resonance
frequency of 400.13 MHz by using a 5 mm high-resolution probe. A
pulse length of 16 μs at 14 W (31P) and 10.15 μs at 20 W (1H) and
relaxation delays of 2 s (31P) and 1 s (1H) were used. The chemical
shifts were referenced relative to H3PO4 for 31P NMR measurements
1
and relative to tetramethylsilane for H NMR measurements. Magic
angle spinning (MAS) NMR experiments were performed on a Bruker
Avance 300 spectrometer using a commercial double-resonance 2.5
mm MAS NMR probe operating at a resonance frequency of 121.5
MHz and a MAS frequency of 16 kHz. For these measurements the
phosphorus chemical shifts were referenced relative to H3PO4. For the
determination of the chemical shift of signals in the 31P MAS NMR
spectra, the center of gravity of the signals was used.
Quantum Chemical Calculations. Scalar-relativistic DFT calcu-
lations were performed in the Elk code,47 which is based on the full-
potential linearized augmented plane-wave method within the local-
density approximation. A number of full-potential linearized
augmented-plane wave basis functions up to RMTKmax = 8 (where
RMT is the average radius of the muffin-tin spheres, and Kmax is the
maximum value of the wave vector K = k + G) were considered and
the maximum length of G for expanding the interstitial density and
potential was set to 12. A total number of 16 k-points was obtained in
the irreducible part of the Brillouin zone.
Figure 1. PXRD patterns of Cu3−xP from reaction of Cu and Pred
(2.95:1 molar ratio) in [P66614][Cl] and [BMIm][Cl] vs that calculated
on the basis of the Inorganic Crystal Structure Database No. 15056.
Electrical Resistivity Measurements. A sample with the
composition of Cu2.95(4)P was selected for electrical resistivity
measurements. The powder was checked for purity with PXRD and
SEM, and the composition was determined by EDX. The fine ground
powder was subsequently compacted with spark plasma sintering
(SPS) using an SPS-515 ET setup (Fuji Electronic Industrial) under a
pressure of 100 MPa with a heating rate of 50 K/min at a temperature
of 350 °C for 10 min. Afterward the pellet was manually broken into
pieces and checked for purity by PXRD, SEM, and EDX again. One
piece of the pellet was selected for the resistivity measurements. This
piece was contacted on a holder using Au wires in a four-point
measurement configuration and in a configuration with two pairs of
opposing contacts for Hall measurements. Finally, the dependence of
the resistance was measured with the sample in He exchange gas
between 4 and 300 K.
RESULTS AND DISCUSSION
■
Synthesis. In a previous work, we observed an unexpected
reactivity of red phosphorus with iodine in the IL [BMIm][Cl]·
2AlCl3.48 Subsequent experiments aiming at a targeted
synthesis of phosphides showed that a mixture of copper
metal and red phosphorus (element ratio 2.95:1) in ILs led to a
dark solid product. After the IL was removed with ethanol and
the product was dried, a fine greyish powder was obtained. The
powder is single-phase Cu3−xP according to PXRD (Figure 1).
EDX mapping showed a homogeneous distribution of Cu and
P inside the sample (Figure S1, Supporting Information)
making the presence of X-ray amorphous red phosphorus
improbable. The results of the elemental analysis (Table S1,
Supporting Information) exclude residuals of the ILs on the
particle surface. Only low amounts of oxygen and chloride
contaminations as well as sodium impurities, most probably
caused by the used glassware, were found. Closer investigations
of the powder with SEM revealed particles, which are
inhomogeneous in size and have irregular shapes (Figure 2).
The particle sizes range from ∼1 μm to 15 μm. Interestingly,
the cations of the ILs appear not to substantially affect the
particle morphology.
Figure 2. SEM images of the as-prepared Cu3−xP (after washing and
drying).
In the course of this work, a few conditions that yield an
incomplete reaction of the starting materials could be identified.
The substitution of the chloride anion of [BMIm][Cl] by
bromide or iodide led to a noticeable decrease of the reaction
rate, while [P66614][decanoate] or [P66614][bis(tri¯uoromethyl-
sulfonyl)imide] were found to lead to a significant amount of
impurities of copper when applying the same reaction
conditions. Therefore, it should be considered that chloride
anions might play a role during the activation of the starting
materials. There are already many examples for the activation of
P
white by nucleophilic reagents.49 Recently, nucleophilic loosen-
ing of P−P bonds has also been observed with Pred in different
8846
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
Article
1
Figure 3. Liquid-state H NMR and 31P NMR spectra of the ILs before (black lines) and after (red lines) reaction of copper and red phosphorus
inside the liquids. For reasons of clarity, only the ranges of the spectra with signals are shown.
organic solvents.50 Because a decreased reaction rate was
observed while lowering the nucleophilicity of the anion, it was
assumed that a comparable mechanism might play a role during
this synthesis. However, this must be verified in future
experiments. Remarkably, replacing metallic copper as the
starting material by CuCl, CuI, or CuCl2 failed and yielded only
unidentifiable products with poor crystallinity. Obviously, the
copper halides are not appropriate as redox partners for the
reduction of the red phosphorus in ILs.
The new synthesis route presented here, although related to
the one described by Aitken et al.,31 shows some major
improvements compared to the approaches presented
previously. First, the reaction is highly efficient. No excess of
phosphorus is needed, and the product yield is high. While the
reaction in [P66614][Cl] runs almost quantitatively (yield 99%),
side reactions of the phosphorus with the imidazolium cation
reduce the yield to 84% in [BMIm][Cl] (see next section). The
ILs could be removed easily from the product with a common
organic solvent in both cases. This is a great improvement
compared to the classic preparation of binary and ternary metal
phosphides in metal fluxes, for example, Sn, because a
treatment with hydrochloric acid is no longer needed.51 The
second advantage of this synthesis route arises from the fact
that ILs are generally not volatile.52 The experimental
procedure can be heavily simplified; no pressure control or
autoclave are needed. Instead, the reaction can be conducted in
a common glass flask, as neither solvent nor starting materials
generate autogenous pressure. This illustrates another advant-
age of red phosphorus. In addition to its nontoxicity, it does not
develop a significant vapor pressure below 400 °C (in contrast
to, for example, white phosphorus or PH3).53 Moreover, the
flask allows for vigorous stirring, which easily leads to
comparatively short reaction times. As shown in this work,
inert conditions are needed to suppress side reactions with
oxygen or moisture, which can be achieved by sealing the flask
with a simple glass stopper. An upscaling of the reactions seems
only limited by the size of the reaction flask and the strength of
the stirring mechanism.
Reaction Studies with Liquid-State Nuclear Magnetic
Resonance Spectroscopy. To obtain more detailed insights
into the reaction pathway, liquid-state NMR experiments with
1H and 31P nuclei were performed, Figure 3. In practice, two
different reaction paths should be considered. First, phosphorus
may react with copper to form a soluble complex that
decomposes at elevated temperatures to form the final product
Cu3−xP. In this case, additional signals of the dissolved species
should appear in the 31P NMR spectrum. Second, phosphorus
could simply diffuse into copper in a similar way as observed for
common solid-state reactions. The second scenario has already
been observed for solvothermal synthesis of nanocrystalline
Cu3−xP29,32 and other metal-rich phosphide nanoparticles.54
In our experiments, the ILs were measured before use as well
as after the reaction. No changes in the NMR spectra of
[P66614][Cl] (Figure 3, top) were detected, which supports the
8847
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
Article
latter pathway in accordance with literature results.29,32,36,54
However, the exact mechanism and the question of whether the
IL contributes to the activation of the reactive species must be
investigated in more detail in the future. The fact that no
degeneration of the [P66614] cation occurred is probably one
reason for the almost quantitative yield of Cu3−xP from this
reaction. In contrast, numerous additional signals are detected
1
in the H and 31P NMR spectra when conducting the same
reaction in [BMIm][Cl] (Figure 3, bottom). Differences in the
spectra highlight a degeneration of the imidazolium cation
through the reaction with phosphorus. Interestingly, the PXRD
analysis (Figure 1) does not show any copper impurities in the
Cu3−xP product. Most likely, the phosphorus species detected
in the 31P NMR are coordinated by copper ions, an assumption
that is also supported by the considerably lower product yield
as compared to the reaction in [P66614][Cl]. As the type of
organic cation does not seem to have any influence on the
particle morphology (Figure 2), degeneration of the cation is
not expected to affect the reaction path itself but lead to a loss
of starting materials instead. This is probably the reason for the
reduced product yield that we observed for reactions in
[BMIm][Cl]. Consequently, a precise control of the stoichio-
metric ratio of the starting materials can only be achieved with
[P66614][Cl], which becomes crucial when a single-phase
product is desired (see next section). Both ILs can easily be
recovered by distilling off the organic solvent used for product
separation. However, in terms of yield, sustainability, and
upscaling, [P66614][Cl] is the better choice, as no significant loss
occurs during reaction and recovery.
Figure 4. (top) The crystal structure of Cu3−xP according to
Olofsson.27 Bonds between Cu(1)/(2) and P are omitted.
Coordination of Cu(1) and Cu(2) is visualized with a blue and a
red polyhedron, respectively. (bottom) Projection of the honeycomb
net. The Cu(1) and Cu(2) atoms are omitted.
Structural Analysis of Cu3−xP. The structure of Cu3−x
P
was analyzed by Olofsson in 1972.27 It crystallizes in the
acentric hexagonal space group P63cm with one P atom and
four Cu atoms in the asymmetric unit. The coordination
polyhedron of the P atom is a trigonal prism with capping
atoms on all faces. The coordination of Cu(1) and Cu(2) by P
atoms is distorted tetrahedral; in the case of Cu(3) and Cu(4)
it is distorted trigonal.
≤ 2.95:1. EDX of respective samples synthesized in [P66614][Cl]
(Figure S2 and Table S2, Supporting Information) led to the
compositions Cu2.95(4)P, supporting the hypothesis of an
intrinsic sub-stoichiometry of Cu3−xP under standard con-
ditions, which is consistent with results of DFT calculations of
de Trizio et al.29
To get an overview of the homogeneity range, samples with
different nominal compositions of Cu3−xP (x = 0, 0.05, 0.1, 0.2,
[...], 0.7) were prepared. As already described, samples with
nominal composition Cu3P are known to contain residual
copper, whereas those with a composition of Cu2.95P are single-
phase according to the results of PXRD and SEM/EDX
measurements. EDX analyses of samples with higher
phosphorus contents show fluctuations in the chemical
composition, caused by the precipitation of a byproduct,
most likely CuP2, as revealed by EDX mappings and point
analyses (see Figures S3−S5 and Table S3, Supporting
Information). However, X-ray powder diagrams show no
CuP2 reflections either due to the small amount of the
byproduct or due to an inferior cystallinity of the precipitates.
CuP2 precipitate formation has already been observed for
samples with a composition of Cu2.85P.27 According to our
results the formation of a CuP2 precipitate already starts at a
nominal composition of Cu2.9P, so batches with x = 0.05 are the
only single-phase Cu3−xP samples obtained and characterized in
this work.
More interestingly, the structure can be subdivided into
layers. The copper atoms in trigonal coordination are the nodes
in a honeycomb net parallel to (001). The rings of this net are
buckled, but the Cu−P distances are very similar (235.1 to
236.7 pm). The tetrahedrally coordinated copper atoms are
connecting these layers along the c-axis, but it is worth
mentioning that some of the Cu−P distances are even shorter
(234.4 to 296.6 pm) than those found in the rings (Figure 4).
As mentioned previously, Olofsson described how the sub-
stoichiometry of Cu3−xP manifests itself in partially occupied
Cu(1) and Cu(2) positions.27 de Trizio et al. recently further
substantiated the experimental evidence by DFT calculations,
which showed a negative value of the free energy of vacancy
formation at the Cu(1) and Cu(2) positions.29 With decreasing
copper content, the number of tetrahedrally coordinated Cu
atoms between the honeycomb layers is reduced, giving the
compound a more layered character. This is consistent with the
results of Schlenger et al., who observed a contraction of the
unit cell while lowering the copper content.28 This contraction,
however, does not occur predominantly along the c-axis.
Although our results in general agree with those of Olofsson
and Schlenger et al., some minor additions concerning the
homogeneity range were noted. Single-phase samples of
stoichiometric Cu3P could not be obtained; all attempts to
achieve this composition resulted in elemental copper as a
byproduct instead. The Cu byproduct vanishes for ratios Cu/P
In the following, 31P MAS NMR spectroscopy and Le Bail
analyses were made on the same batches already analyzed by
SEM (Figure S6, Supporting Information). The NMR peaks
show an asymmetric broadening, which can be attributed to the
presence of different Cu3−xP phases within one batch (Figure
5). However, no signals from the CuP2 precipitation were
observed. More obvious is, however, the considerable peak shift
8848
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
Article
of the Cu3−xP structure at its copper-rich limit must be
investigated in the future by more elaborate experiments.
Electronic Behavior. The results of the solid-state NMR
experiments indicate a metallic character of the Cu3−xP
material. Metallic conductivity of Cu3−xP was proposed by
Robertson et al., who measured a low electrical resistance on a
single crystal.33 However, semiconducting behavior was claimed
by several other groups.29−31 To shed some light on the
electronic properties the band structure of idealized stoichio-
metric Cu3P was calculated (Figure 7). According to these
calculations, even the copper-rich boundary phase should be
metallic.
Figure 5. 31P MAS NMR spectra of Cu3−xP with different nominal
compositions synthesized in [P66614][Cl]. Compositions given in
quotation marks are nominal compositions.
going from copper-rich to copper-poor compositions. Compar-
ing the nominal compositions Cu3P and Cu2.95(4)P, the shift of
the main peak is negligible supporting the intrinsic sub-
stoichiometry of this compound again (Figure 5). The
phosphorus-rich edge of the homogeneity range was found at
an approximate composition of Cu2.6P, which is slightly less
than reported before (x = 0.3).27,28 The NMR signal of Cu3−xP,
shifts in total ∼32 ppm downfield between samples with
nominal compositions of Cu2.9P and Cu2.8P (Figure 5). As
already discussed in a previous work, Cu3−xP is known to
exhibit a Knight shift, which is considered as indicative for a
conducting sample.34 Thus, the considerable offset between the
31P signals of Cu2.9P and Cu2.8P indicates a strongly changed
Knight shift. This means phosphorus-rich samples are expected
to have enhanced electron mobility, that is, a higher
conductivity. The chemical shift of the 31P signal correlates
well with the change of the unit cell dimensions, which makes a
proportional relation between both parameters probable
(Figure 6; for detailed data see Table S4, Supporting
Information).
Figure 7. Band structure of Cu3P on the basis of the structure reported
by Olofsson.27
To corroborate the theoretical results, the temperature-
dependent resistance of a sample with the composition
Cu2.95(4)P was measured. The sample shows metallic behavior
as the resistance decreases monotonically with decreasing
temperature to 4 K (Figure S7, Supporting Information).
Likewise, Hall measurements at 300 K evidence a carrier
density of ∼1 × 1027 carriers·per cubic meter, which is in the
same range as those for the alkali metals, for example.55
CONCLUSION
■
It is worth mentioning that it was not possible to fit the
diffraction pattern of a sample with the nominal composition
Cu3P with the hexagonal structure model satisfactorily due to a
considerable broadening of the reflections. A possible distortion
Microcrystalline Cu3−xP can be synthesized by using an
ionothermal synthesis route most efficiently in the ionic liquid
[P66614][Cl]. The starting materials, copper metal and red
phosphorus, are readily available and practically nontoxic. The
IL-based approach significantly simplifies the experimental
procedure, because no autogenous pressure is generated. A
common glass flask with stopper is sufficient as reaction vessel,
making it easy to avoid contamination by air and moisture.
[P66614][Cl] promotes the reaction but is neither affected by the
starting material nor degenerates thermally within the reaction
cycle and is, thus, easily recyclable without influence on the
overall phosphorus efficiency. Yields of single-phase Cu2.95(4)P
close to 100% were obtained. Our results confirm that a phase
of the precise composition Cu3P probably does not exist. Bulk
Cu2.95(4)P is metallic. Solid-state NMR spectroscopy revealed
the possibility of enhanced electron mobility in phosphorus-
rich samples of Cu3−xP.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.6b01397.
■
S
Figure 6. Correlation between chemical shift (center of gravity of the
signal) and change of the cell dimension of samples with different
compositions synthesized in [P66614][Cl]. Compositions given in
quotation marks are nominal compositions.
8849
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
SEM images of Cu3−xP, EDX data for Cu2.95(4)P, data of
Article
(22) Kato, T.; Adachi, S.; Aoyagi, T.; Naito, T.; Yamamoto, H.;
Nojiri, T.; Yoshida, M. IEEE J. Photovoltaics 2012, 2, 499−505.
(23) Solovyov, S. E. Oxygen Scavengers. In Kirk-Othmer Encyclopedia
of Chemical Technology; Seidel, A., Bickford, M., Eds.; John Wiley and
Sons, Inc: Hoboken, NJ, 2014; pp 1−31.
(24) Pfeiffer, H.; Tancret, F.; Brousse, T. Electrochim. Acta 2005, 50,
4763−4770.
(25) Chandrasekar, M. S.; Mitra, S. Electrochim. Acta 2013, 92, 47−
54.
elemental analysis of Cu3−xP, solid-state NMR data,
results of lattice parameter fitting, temperature-depend-
ent resistance of Cu2.95(4)P (PDF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: thomas.doert@tu-dresden.de.
(26) Stan, M. C.; Klopsch, R.; Bhaskar, A.; Li, J.; Passerini, S.; Winter,
̈
M. Adv. Energy Mater. 2013, 3, 231−238.
Author Contributions
All authors have given approval to the final version of the
manuscript.
(27) Olofsson, O.; et al. Acta Chem. Scand. 1972, 26, 2777−2787.
(28) Schlenger, H.; Jacobs, H.; Juza, R. Z. Anorg. Allg. Chem. 1971,
385, 177−201.
Notes
(29) De Trizio, L.; Gaspari, R.; Bertoni, G.; Kriegel, I.; Moretti, L.;
Scotognella, F.; Maserati, L.; Zhang, Y.; Messina, G. C.; Prato, M.;
Marras, S.; Cavalli, A.; Manna, L. Chem. Mater. 2015, 27, 1120−1128.
(30) Manna, G.; Bose, R.; Pradhan, N. Angew. Chem., Int. Ed. 2013,
52, 6762−6766; Angew. Chem. 2013, 125, 6894−6898.
(31) Ann Aitken, J. A.; Ganzha-Hazen, V.; Brock, S. L. J. Solid State
Chem. 2005, 178, 970−975.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We would like to thank P. Nockemann (Queen’s Univ, Belfast)
for helpful discussions. We gratefully acknowledge G.
Auffermann (MPI CPfS, Dresden) for elemental analysis, I.
(32) De Trizio, L.; Figuerola, A.; Manna, L.; Genovese, A.; George,
C.; Brescia, R.; Saghi, Z.; Simonutti, R.; Van Huis, M.; Falqui, A. ACS
Nano 2012, 6, 32−41.
Veremchuk (MPI CPfS, Dresden) for SPS, A. Brunner (TU,
̈
Dresden) for help on SEM and EDX measurements, A. Isaeva
(TU, Dresden) for band structure calculations, and S. Yogendra
(TU, Dresden) for measuring NMR spectra. We thank the
Deutsche Forschungsgemeinschaft for financial support within
the Priority Program SPP1708.
(33) Robertson, D. S.; Snowball, G.; Webber, H. J. Mater. Sci. 1980,
15, 256−258.
́
́
(34) Furo, I.; Bakonyi, I.; Tompa, K.; Zsoldos, E.; Heinmaa, I.; Alla,
M.; Lippmaa, E. J. Phys.: Condens. Matter 1990, 2, 4217−4225.
(35) Bichat, M.-P.; Politova, T.; Pfeiffer, H.; Tancret, F.;
Monconduit, L.; Pascal, J.-L.; Brousse, T.; Favier, F. J. Power Sources
2004, 136, 80−87.
REFERENCES
■
(36) Carenco, S.; Hu, Y.; Florea, I.; Ersen, O.; Boissier
̀
e, C.;
(1) Steen, I. Phosphorus Potassium 1998, 227, 25−31.
(2) Smil, V. Annu. Rev. Energy Environ. 2000, 25, 53−88.
Mez
́
ailles, N.; Sanchez, C. Chem. Mater. 2012, 24, 4134−4145.
(37) Su, H. L.; Xie, Y.; Li, B.; Liu, X. M.; Qian, Y. T. Solid State Ionics
1999, 122, 157−160.
(3) Schroder, J. J.; Cordell, D.; Smit, A. L.; Rosemarin, A. Sustainable
̈
Use of Phosphorus, EU Tender ENV.B.1/ETU/2009/0025; Plant
Research International Wageningen UR: Wageningen, Netherlands,
2010.
(38) Xie, Y.; Su, H. L.; Qian, X. F.; Liu, X. M.; Qian, Y. T. J. Solid
State Chem. 2000, 149, 88−91.
(39) Liu, S.; Qian, Y.; Xu, L. Solid State Commun. 2009, 149, 438−
440.
(4) Van Vuuren, D. P.; Bouwman, A. F.; Beusen, A. H. W. Global
Environ. Chang. 2010, 20, 428−439.
(40) Cooper, E. R.; Andrews, C. D.; Wheatley, P. S.; Webb, P. B.;
Wormald, P.; Morris, R. E. Nature 2004, 430, 1012−1016.
(41) Parnham, E. R.; Drylie, E. A.; Wheatley, P. S.; Slawin, A. M. Z.;
Morris, R. E. Angew. Chem., Int. Ed. 2006, 45, 4962−4966; Angew.
Chem. 2006, 118, 5084−5088.
(42) Parnham, E. R.; Morris, R. E. J. Am. Chem. Soc. 2006, 128,
2204−2205.
(43) Parnham, E. R.; Morris, R. E. Acc. Chem. Res. 2007, 40, 1005−
1013.
(5) Cordell, D.; Drangert, J. O.; White, S. Global Environ. Chang.
2009, 19, 292−305.
(6) Cordell, D.; White, S. Sustainability 2011, 3, 2027−2049.
(7) Von Schnering, H. G.; Honle, W. Chem. Rev. 1988, 88, 243−273.
̈
(8) Pottgen, R.; Honle, W.; von Schnering, H. G. Phosphides: Solid-
̈
̈
State Chemistry. In Encyclopedia of Inorganic Chemistry, 2nd ed.; King,
R. B., Ed.; John Wiley & Sons, Ltd: Chichester, U.K., 2005; Vol. 8, pp
4255−4308.
(9) Bawohl, M.; Nilges, T. Z. Anorg. Allg. Chem. 2015, 641, 304−310.
(10) Muetterties, E. L.; Sauer, J. C. J. Am. Chem. Soc. 1974, 96, 3410−
3415.
(44) Morris, R. E. Chem. Commun. 2009, 2990−2998.
(45) Brauer, G. Handbuch Der Praparativen Anorganischen Chemie,
̈
3rd ed.; F. Enke: Stuttgart, 1962.
(11) Brock, S. L.; Perera, S. C.; Stamm, K. L. Chem. - Eur. J. 2004, 10,
3364−3371.
(12) Oyama, S. T.; Gott, T.; Zhao, H.; Lee, Y. K. Catal. Today 2009,
143, 94−107.
(13) Prins, R.; Bussell, M. E. Catal. Lett. 2012, 142, 1413−1436.
(14) Bhushan, M.; Catalano, A. Appl. Phys. Lett. 1981, 38, 39−41.
(15) Nayar, P. S.; Catalano, A. Appl. Phys. Lett. 1981, 39, 105−107.
(16) Luber, E. J.; Mobarok, M. H.; Buriak, J. M. ACS Nano 2013, 7,
8136−8146.
(17) Shirotani, I.; Takaya, M.; Kaneko, I.; Sekine, C.; Yagi, T. Phys. C
2001, 357−360, 329−332.
(18) Bachmann, K. J. Annu. Rev. Mater. Sci. 1981, 11, 441−484.
(19) Bera, D.; Qian, L.; Tseng, T. K.; Holloway, P. H. Materials 2010,
3, 2260−2345.
(20) Preusse, H.; Bolton, J. D. Powder Metall. 1999, 42, 51−62.
(21) Nikanorov, S. P.; Peller, V. V. Continuous Casting Design by
the Stepanov Method. In Handbook of Metallurgical Process Design;
Totten, G. E., Funatani, K., Xie, L., Eds.; Marcel Dekker, Inc: Basel,
Switzerland, 2004; pp 295−349.
(46) Jana2006, The Crystallographic Computing System; Institute of
Physics: Praha, Czech Republic, 2015 Petricek, V.; Dusek, M.;
Palatinus, L. Z. Kristallogr. - Cryst. Mater. 2014, 229, 345−352.
(47) The Elk FP-LAPW Code, version 1.4.22, 2009−2015; Online:
http://elk.sourceforge.net.
(48) Groh, M. F.; Paasch, S.; Weiz, A.; Ruck, M.; Brunner, E. Eur. J.
Inorg. Chem. 2015, 2015, 3991−3994.
(49) Scheer, M.; Balaz
́
s, G.; Seitz, A. Chem. Rev. 2010, 110, 4236−
4256.
(50) Dragulescu-Andrasi, A.; Miller, L. Z.; Chen, B.; McQuade, D. T.;
Shatruk, M. Angew. Chem., Int. Ed. 2016, 55, 3904−3908; Angew.
Chem. 2016, 128, 3972−3976.
(51) Kanatzidis, M. G.; Pottgen, R.; Jeitschko, W. Angew. Chem., Int.
̈
Ed. 2005, 44, 6996−7023; Angew. Chem. 2005, 117, 7156−7184.
(52) Ionic Liquids in Synthesis, 2nd ed.; Wasserscheid, P., Welton, T.,
Eds.; Wiley-VCH: Weinheim, Germany, 2008.
(53) Eckstein, N.; Hohmann, A.; Weihrich, R.; Nilges, T.; Schmidt, P.
Z. Anorg. Allg. Chem. 2013, 639, 2741−2743.
8850
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851

Inorganic Chemistry
Article
(54) Henkes, A. E.; Vasquez, Y.; Schaak, R. E. J. Am. Chem. Soc. 2007,
129, 1896−1897.
(55) Palenskis, V. World J. Condens. Matter Phys. 2013, 3, 73−81.
8851
DOI: 10.1021/acs.inorgchem.6b01397
Inorg. Chem. 2016, 55, 8844−8851