Solvothermal syntheses of Cu3P via reactions of amorphous red phosphorus with a variety of copper sources

Article abstract of DOI:10.1016/j.jssc.2004.10.004

Polycrystalline Cu₃P was successfully prepared under a wide variety of solvothermal conditions. The reaction of red phosphorus with several copper sources (copper metal, copper (I) iodide, copper (I) chloride and copper (II) chloride) at 150-200 °C for 1-2 days in water produced Cu₃P. Products were examined with powder X-ray diffraction, scanning electron microscopy and energy dispersive spectroscopy. Thermal analyses and optical spectroscopy were also performed. A detailed list of reaction conditions, products and impurity phases (where applicable) are reported.

Full text of DOI:10.1016/j.jssc.2004.10.004

ARTICLE IN PRESS
Journal of Solid State Chemistry 178 (2005) 970–975
www.elsevier.com/locate/jssc
Solvothermal syntheses of Cu3P via reactions of amorphous
red phosphorus with a variety of copper sources
Ã
Jennifer Ann Aitkena, Valentina Ganzha-Hazenb, Stephanie L. Brockb,
aDepartment of Chemistry and Biochemistry, Duquesne University, Pittsburgh, PA 15282, USA
bDepartment of Chemistry, Wayne State University, 5101 Cass Avenue, Detroit, MI 48202-3929, USA
Received 15 July 2004; accepted 3 October 2004
Abstract
Polycrystalline Cu3P was successfully prepared under a wide variety of solvothermal conditions. The reaction of red phosphorus
with several copper sources (copper metal, copper (I) iodide, copper (I) chloride and copper (II) chloride) at 150–200 1C for 1–2 days
in water produced Cu3P. Products were examined with powder X-ray diffraction, scanning electron microscopy and energy
dispersive spectroscopy. Thermal analyses and optical spectroscopy were also performed. A detailed list of reaction conditions,
products and impurity phases (where applicable) are reported.
r 2004 Elsevier Inc. All rights reserved.
Keywords: Solvothermal; Hydrothermal; Copper phosphide; Transition metal phosphide; Phosphide; Pnictide; Low temperature synthesis;
Microcrystalline
1. Introduction
group and transition metal pnictide (pnicogen=Group
15 element) materials [6–8]. Nanocrystalline Cu3P has
been previously prepared by the reaction of yellow
phosphorus and copper (II) chloride dihydrate in
ethylenediamine (en) solvothermally at 80–140 1C for
12 h [7] and in a dilute ammonia liquor at 140 1C for 10 h
[8]. In both cases it was proposed that the chelating
solvent and the yellow phosphorus play a very
important role in the formation of Cu3P. Specifically,
it was suggested that the yellow phosphorus reacts with
hydroxide groups, generated in the basicen or ammonia
solution from the hydrated metal salts, to produce
phosphate ions and the reducing equivalents necessary
to reduce the Cu2+ to Cu1+. The resultant Cu1+ ions
react directly with the yellow phosphorus as they are
generated to form the metal phosphide. The role of the
N-donor ligand is to control this reaction by coordinat-
ing copper ions, resulting in their slow release, and the
subsequent generation of a nanocrystalline material. It
was also reported that reactions performed with red
phosphorus or in aqueous solution did not yield
complete reactions [8].
Copper (I) phosphide has attracted attention due to
its existent and potential technological properties. This
complex material, see Fig. 1, enhances the sintering
behavior of 316 L stainless steel [1] and acts as a
reinforcing agent in high speed steel (HSS) composite
materials [2]. More recently, copper (I) phosphide has
gathered attention as a potential negative electrode
material in lithium batteries because of its good
cyclability and reversible capacity [3].
Traditionally, transition metal phosphides such as
Cu3P have been prepared via inconvenient routes that
are not practical for industrial scale preparation, such as
high temperature solid-state synthesis in sealed am-
poules [4] and metathesis reactions using Na3P, a
hazardous reagent [5]. Within the past decade, sol-
vothermal synthesis has been demonstrated to be a
successful method for the facile preparation of main
Ã
Corresponding author. Fax: +1 313 577 8822.
E-mail address: sbrock@chem.wayne.edu (S.L. Brock).
0022-4596/$ - see front matter r 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2004.10.004

ARTICLE IN PRESS
J. Ann Aitken et al. / Journal of Solid State Chemistry 178 (2005) 970–975
971
The initial goal of our solvothermal reactions was to
find a set of conditions under which to produce copper
pnictide halides via a soft chemical route. We thought
that under conditions different than those utilized for
Cu3P we might be able to form CuaPbXc (X=halide)
type materials, which show promise as copper ion-
conductors [9]. To our surprise, we obtained Cu3P in
each case, regardless of the conditions employed. This
finding compelled us to re-examine the conditions
necessary to synthesize Cu3P solvothermally.
Here we report a wide variety of solvothermal
conditions under which Cu3P can be produced, includ-
ing the optimal conditions for preparing pure-phase
material. Our work shows that red phosphorus is
sufficiently active to be employed in lieu of yellow
phosphorus in the solvothermal preparation of Cu3P.
Furthermore, ethylenediamine or ammonia can be
substituted by H2O as a cleaner and cheaper solvent to
successfully synthesize Cu3P. We also report the optical
spectra and thermal analyses of several samples of Cu3P
that we obtained.
2. Experimental
2.1. Reagents
The chemicals in this work were used as obtained:
(i) copper metal powder, À625 mesh, 99.9% Alfa-Aesar,
Ward Hill, MA; (ii) copper (I) iodide, 99.999%, Aldrich,
Milwaukee, WI; (iii) copper (I) chloride 99% Acros,
Morris Plains, NJ; (iv) copper (II) chloride 99%
anhydrous, Acros, Morris Plains, NJ and (v) red
phosphorus amorphous powder, 100 mesh, EM Science,
Gibbstown, NJ.
2.2. Synthesis
2.2.1. General synthetic procedure
Red phosphorus powder and a copper source were
weighed in appropriate amounts and loaded into a
23 mL Teflon cup. No more than 13 mL of solvent or
mixture of solvents was added to the cup. The Teflon
cup was capped and then placed inside a stainless steel
Parr acid digestion bomb. The bomb was inserted into
an oven already at the reaction temperature,
150–200 1C, allowed to stay there for a time period of
1–2 days and subsequently removed from the oven and
allowed to cool to room temperature. The resulting
product was filtered by vacuum suction and washed with
distilled H2O and ether. In one example of an ideal
synthesis, 0.15 g red phosphorus (4.8 mmol) was com-
bined with 0.479 g CuCl (4.8 mmol) in 13 mL of
deionized H2O and heated at 200 1C for 24 h.
Fig. 1. Crystal structure of Cu3P projected down the a-axis showing
layers of phosphorus trigonal prisms stacked in the c-direction (a) and
the structure of one such layer viewed down the c-axis (b). Phosphorus
atoms are shown as gray spheres while the black spheres represent
copper atoms.

ARTICLE IN PRESS
J. Ann Aitken et al. / Journal of Solid State Chemistry 178 (2005) 970–975
972
2.3. Physical measurements
in the formation of the Cu3P material obtained; second,
yellow phosphorus is very active under solvothermal
conditions, acting both as a reducing agent and a
reactant; and third, an enormous amount of excess
phosphorus is needed to ensure a complete reaction and
generate pure phase product. Our investigation shows
that none of these ‘‘key factors’’ pointed out by Qian
and coworkers are necessary to synthesize Cu3P.
2.3.1. Powder X-ray diffraction
Powder X-ray diffraction patterns were collected with
a Rigaku RU 200 B rotating anode X-ray diffractometer
operating at 40 kV and 150 mA and using graphite
monochromatized CuKa radiation. Samples were
ground with mortar and pestle and deposited onto a
glass slide coated with a thin layer of Vaseline. Data
were collected from 101 to 701 2y: Phases were identified
using the search match capabilities of the JADE
program along with the ICDD (International Center
for Diffraction Data) powder diffraction file (PDF)
database.
Our initial experiments were conducted using yellow
phosphorus, the most volatile and reactive form of
phosphorus; however, we quickly determined that
yellow phosphorus was not practical to work with and
resulted in a number of phosphate byproducts. Amor-
phous red phosphorus is much less reactive than yellow
phosphorus under normal conditions, but this makes it
safer and easier to handle in air. It is well known that
many starting materials that are unreactive or insoluble
under normal conditions can react or become solubi-
lized under solvothermal conditions. Indeed, we have
demonstrated that red phosphorus can react with a
number of copper sources solvothermally to form Cu3P.
Tables 1–3 list 23 of approximately 60 different
reactions that we performed, which yielded Cu3P as
the major product or as a pure phase material. We
successfully determined the right conditions under
which copper metal, copper (I) iodide, copper (I)
chloride and copper (II) chloride react with red
2.3.2. Scanning electron microscopy and energy dispersive
spectroscopy
Samples were mounted onto an adhesive carbon tab
that was attached to an aluminum specimen stub and
examined using a Hitachi S-2400 Scanning Electron
Microscope operating at 25 keV. Energy dispersive
spectroscopy was performed using an EDAX unit with
an active area of 10 mm2, and data were analyzed using
the Genesis Spectrum V. 2.5 software package by
EDAX INC. Data were collected with an accumulation
time of 60 s.
2.3.3. Differential thermal analysis (DTA)
Table 1
Summary of solvothermal reactions performed at 150 1C for 24 h using
en or en/H2O mixtures
Thermal analyses were performed using a Shimadzu
DTA50 Thermal Analyzer. Samples were sealed in small
carbon coated fused silica ampoules. Alumina was used
as a reference material. The difference in mV between the
sample and the reference was measured as a function of
temperature up to 1100 1C. The sample and reference
were heated and cooled at a rate of 10 1C/min. Multiple
heating and cooling cycles were performed.
Copper Ratio of
Ratio
copper source of red
phosphorus
% en Result of
powder
X-ray
source
diffraction
CuCl
3
1
100
Cu3P+Cu
CuCl
CuCl
CuCl
CuCl
3
3
3
3
1
1
1
1
50
2.5
2
Cu3P+Cu
Cu3P+Cu
2.3.4. Optical spectroscopy
Optical diffuse reflectance measurements were mea-
sured on a Cary 5000 double-beam double-monochro-
mator spectrophotometer equipped with the praying
mantis attachment and operating from 2500 to 200 nm.
The samples and a BaSO4 100% reflectance standard
were prepared by grinding the material to a fine powder
and loading it into the sample holder. The reflectance
versus wavelength data were used to estimate the band
gap by converting reflectance to absorption data [10].
Cu3P+Cu+CuCl
Cu3P+CuCl
1
CuCl2
3
1
100
Cu3P+Cu
Table 2
Comparison of solvothermal reactions performed at 150 1C and 200 1C
in H2O for 24 h
Copper Ratio of
source
Ratio of red Reaction
copper source phosphorus temperature of powder
Result
3. Results and discussion
(1C)
X-ray
diffraction
3.1. Synthesis
CuI
CuI
1
1
5
5
150
200
Cu3P+CuI
Cu3P
Three key factors have been previously proposed in
the synthesis of Cu3P [7,8]. First, an N-donor ligand
such as ethylenediamine or ammonia plays a crucial role
CuCl
CuCl
1
1
1
1
150
200
Cu3P+CuCl
Cu3P

ARTICLE IN PRESS
J. Ann Aitken et al. / Journal of Solid State Chemistry 178 (2005) 970–975
973
Table 3
Summary of solvothermal reactions performed at 200 1C in H2O for
24 h using several copper sources (a designates a prominent amorphous
background)
Copper
source
Ratio of
copper source
Ratio
of red
phosphorus
Result of
powder
X-ray
diffraction
Cu metal
Cu metal
Cu metal
Cu metal
1
1
Cu3P+Cu
Cu3P+Cu
Cu3P+Cu
Cu3P
1
8
1
1a
10
10a
CuI
CuI
CuI
1
1
1
1
2
5
Cu3P+CuI
Cu3P+CuI
Cu3P
Fig. 2. Powder X-ray diffraction pattern of Cu3P (CuCl+red P,
1:1 molar ratio, 200 1C, 24 h) versus that obtained from the powder
diffraction file (line diagram: PDF # 71-2261).
CuCl
CuCl
CuCl
CuCl
2.8
2
1
Cu3P+CuCl
Cu3P+CuCl
Cu3P
1
1
1
1.5
1
Cu3P+a
grinding of the precursors to create an active surface
area. The EDS spectrum showed only copper and
phosphorus in the product. If less than ten equivalents
of phosphorus is used, copper metal is found in the
powder X-ray diffraction pattern. For example, if we
performed a similar reaction with a 1:8 molar ratio of
starting materials, we obtained peaks for both Cu3P and
copper metal in the powder X-ray diffraction pattern.
However, if this resulting material is then placed in a
bomb with additional phosphorus (0.064 g of product
(Cu3P+Cu)+0.09 g of phosphorus) and heated again to
200 1C for 1 day, a pure material can be obtained.
Several ratios of Cu:P were attempted, see Table 3, with
the 1:1 reaction having the most copper metal impurity.
The reaction of copper iodide and red phosphorus in
a 1:5 molar ratio heated at 200 1C for 1 day in H2O also
produced Cu3P in pure form (Table 3). The powder
X-ray diffraction pattern of the product showed peaks
for Cu3P and no residual CuI was observed. EDS
analysis gave a semiquantitative composition of Cu2.87P.
If the ratio of CuI:P is changed to 1:1 or 1:2 at 200 1C,
the powder X-ray diffraction pattern shows Cu3P plus a
large amount of unreacted CuI.
Although one might expect similar reactivity for CuCl
as that found for CuI, CuCl has proved far more
reactive. The reaction of copper (I) chloride and red
phosphorus in a 1:1 ratio heated in a bomb to 200 1C for
1 day in H2O produced pure Cu3P (Fig. 2, Table 3). If
the ratio of CuCl is increased, large peaks indicative of
CuCl are observed in the powder X-ray diffraction
pattern. However, this product can be washed with
dimethylsulfoxide (50 mL, three times with stirring) to
obtain a phase pure powder diffraction pattern for
Cu3P. If the ratio of P is increased we see an amorphous
presence in the powder X-ray diffraction pattern and an
excess of phosphorus in the EDS spectrum (Table 3).
Thus, the optimal amount of phosphorus needed to
produce phase-pure Cu3P is considerably less for CuCl
CuCl2
CuCl2
CuCl2
CuCl2
1
1
1
1
4
Cu3P+CuCl
Cu3P
Cu3P
4.5
5
8
Cu3P+a
aIngredients were ground together before being placed in the bomb.
phosphorus in H2O to produce pure phase Cu3P.
Therefore, en or ammonia does not play a crucial role
in the synthesis of Cu3P. Ironically, although we
performed some reactions in en and en/H2O mixtures
at 150 1C, we never obtained pure Cu3P under those
conditions (Table 1). The fact that H2O is a suitable
solvent to make Cu3P is a significant finding since it is
much more industrially practical to produce Cu3P in
H2O because it is easier and safer to work with, as well
as significantly less costly.
The critical factors in the synthesis of phase-pure
Cu3P in water are temperature, stoichiometry and
precursor identity. Byproducts are obtained for reac-
tions conducted at 150 1C, whereas pure phase Cu3P c an
be obtained at 200 1C (Table 2) once the appropriate
stoichiometries have been determined for the various
precursors. This is consistent with a lower reactivity for
red phosphorus relative to yellow.
Reactions in which en makes up X2% of the solvent
mixture always produce metallic copper as a byproduct
(Table 1). Indeed, en is used as a reducing agent in
electrolysis copper plating baths, so this is not a surprise
[11]. This led us to test whether copper metal could be
used directly as the copper source in these solvothermal
reactions.
The reaction of copper metal with red phosphorus in
a 1:10 molar ratio heated at 200 1C for 1 day in H2O
produced pure Cu3P (Table 3). The powder X-ray
diffraction pattern did not show the presence of any
other product. However, complete reaction requires

ARTICLE IN PRESS
J. Ann Aitken et al. / Journal of Solid State Chemistry 178 (2005) 970–975
974
(3 equivalents of phosphorus) than for CuI (15
equivalents of phosphorus).
It has been reported that Cu3P has a moderate range
of homogeneity that can depend on the synthetic
conditions used. In a study undertaken by Olofsson it
was demonstrated that at 700 1C the homogeneity range
lies approximately between Cu2.82P and Cu2.73P. It was
also found that by changing the reaction temperature
from 400 to 800 to 850 1C a change in lattice parameters
can be observed, where at higher temperature smaller
lattice parameters were obtained, suggesting a larger
extent of Cu vacancies. Our experiments also show that
Cu3P displays a homogeneity range, since we obtained
slightly different ratios of Cu:P for nearly every
sample obtained, while the powder diffraction patterns
looked nearly identical. Under all solvothermal condi-
tions essayed, a slightly copper deficient material was
obtained, as determined by EDS, which could be
attributed to either intrinsicdefetc formation in the
copper phosphide materials as postulated by Olofsson,
or to the presence of amorphous phosphorus.
For the reaction of copper (II) chloride plus red
phosphorus in H2O at 200 1C the best ratio is 1:5, similar
to the case of CuI. Increased copper (II) chloride
amounts produced CuCl as a byproduct, whereas
increased amounts of phosphorous resulted in a
significant amorphous background in the powder
diffraction pattern, which we attribute to red phos-
phorus.
While our results show that an excess of phosphorus
is always necessary to form pure Cu3P, the excess is not
nearly as much as was reported by Qian and coworkers.
Their reactions required 22 mol of phosphorus (66
equivalents) for every 1 mol of CuCl2 Á 2H2O while our
greatest excess of phosphorus was 10 mol (30 equiva-
lents) for every 1 mol of copper metal. The reaction of
copper (I) chloride with red phosphorus requires the
least amount of excess phosphorus, only two additional
equivalents (i.e. Cu:P 1:1). The finding that less
phosphorus can be used than originally proposed makes
the formation of Cu3P under solvothermal conditions
much more atom efficient. Additionally, despite being
less reactive, red phosphorus can be employed success-
fully in lieu of yellow phosphorus.
3.3. Physical properties
Optical diffuse reflectance measurements were per-
formed for powdered samples of Cu3P. Analyses of
several Cu3P products obtained from different reaction
conditions all showed an optical absorption edge of
approximately 0.8 eV, indicating that Cu3P is a relatively
narrow gap semiconductor, consistent with the black
color of the material (Fig. 3).
Differential thermal analyses were performed for
several samples of Cu3P obtained under different
reaction conditions. The expected thermal behavior
was obtained for a sample obtained from the reaction
of CuI and P in a 1:5 molar ratio heated to 200 1C for 1
3.2. Structure and X-ray diffraction
Cu3P is a very interesting material since both its
composition and structure have been the center of some
controversy. The structure of Cu3P was originally
proposed to be a member of the anti-LaF3 structure-
type, along with many alkali metal pnictide 3:1 phases,
¯
crystallizing in space group P3C [12,13]. The structure
¯
was later solved in space group P3m [14]. Finally, in
day in H2O to yield a material of stoichiometry Cu2.87P
(Fig. 4). The DTA diagram shows an endothermicpeak
1972 Olofsson proposed that the space group P63cm was
correct [15]. Indeed, we calculated the powder X-ray
¯
diffraction pattern based on the P3C model and that of
the P63cm model to find that there were no differences in
the number of observed peaks and little difference in
intensities. Our diffraction data fits with the Olofsson
model and therefore that is the structure we based our
indexing upon (Fig. 2).
The structure of Cu3P contains both P and Cu with
high coordination numbers. Each P atom is surrounded
by eleven Cu atoms. The coordination sphere is quite
distorted but can be best described as a trigonal prism
with additional capping atoms on each of the three
rectangular faces and the two triangular faces. The
trigonal prisms are connected in the ab-plane to form
‘‘layers’’ that stack along the c-axis (Fig. 1). The
coordination of Cu is also irregular. Each Cu has either
three or four phosphorus atoms as neighbors with
distorted trigonal pyramidal and tetrahedral coordina-
tion spheres, respectively. In addition, each Cu atom has
nine Cu neighbors.
Fig. 3. Optical diffuse reflectance spectrum converted to absorption
for a powdered sample of Cu3P (CuCl+red P, 1:1 molar ratio, 200 1C,
24 h).

ARTICLE IN PRESS
J. Ann Aitken et al. / Journal of Solid State Chemistry 178 (2005) 970–975
975
stoichiometric proportions, we have shown the sol-
vothermal synthesis of Cu3P to be much easier, safer
and more efficient than originally proposed.
Acknowledgments
Support from the Wayne State University-Institute of
Manufacturing Research and S. L. Brock’s National
Science Foundation Career Award (DMR-0094273) is
gratefully acknowledged. The Cary 5000 and Shimadzu
DTA50 were purchased with funds provided by
Duquesne University start-up funding.
Fig. 4. DTA diagram (10 1C/min heating and cooling) obtained for
Cu3P (CuI+red P, 1:5 molar ratio, 200 1C, 24 h).
at ꢀ1021 1C upon heating, corresponding to a melting
point, and an exothermicpeak at ꢀ963 1C upon cooling,
indicative of a recrystallization. These results are in close
agreement with the melting point of 1022 1C reported by
Ugai and coworkers [16]. This suggests that the low
copper content is a function of intrinsic defects, rather
than the presence of amorphous phosphorus, since the
latter would be expected to reduce the melting point.
However, several other samples which appeared to be
pure from the analysis of powder X-ray diffraction data,
energy dispersive spectra and optical diffuse reflectance
measurements gave slightly different thermal analysis
diagrams, resulting in a shift of up to 10 1C in the
melting/recrystallization curve. Furthermore, the ther-
mal analysis diagrams show a subtle exothermicpeak
occurring slightly above 800 1C that we cannot assign
to a specific thermal event. The differences between
samples and extra thermal events may be due to small
impurities, the formation of eutectics and/or reactions
such as Cu3P+P-CuP2.
References
[1] H. Preusse, J.D. Bolton, Powder Metall. 42 (1999) 51–62.
[2] (a) W.C. Zapata, C.E. DaCosta, J.M. Torralba, J. Mat. Sci. 33
(1998) 3219–3225;
(b) M.M. Oliveira, J.D. Bolton, Powder Metall. 38 (1995)
131–140.
[3] H. Pfeiffer, F. Tancret, M.-P. Bichat, L. Monconduit, F. Favier,
T. Brousse, Electrochem. Commun. 6 (2004) 263–267.
[4] O. Olofsson, Acta Chem. Scand. 19 (1965) 229–241.
[5] A.L. Hector, I.P. Parkin, J. Mater. Chem. 4 (1994) 279–283.
[6] Y.T. Qian, Adv. Mater. 11 (1999) 1101–1102 and references
therein.
[7] Y. Xie, H.L. Su, X.F. Qian, X.M. Liu, Y.T. Qian, J. Solid State
Chem. 149 (2000) 88–91.
[8] H.L. Su, Y. Xie, B. Li, X.M. Liu, Y.T. Qian, Solid State Ionics
122 (1999) 157–160.
[9] (a) E. Freudenthaler, A. Pfitzner, Solid State Ionics 101–103
(1997) 1053–1059;
(b) E. Freudenthaler, A. Pfitzner, D.C. Sinclair, Mater. Res. Bull.
31 (1996) 171–176.
[10] (a) W.W. Wendlandt, H.G. Hecht, Reflectance Spectroscopy,
Interscience Publishers, New York, 1996;
(b) G. Kotum, Reflectance Spectroscopy, Springer, New York,
1969;
(c) S.P. Tandon, J.P. Gupta, Phys. Status Solidi 38 (1970)
363–367.
4. Conclusion
[11] Metal Finishing Guidebook & Directory, 65th ed., Metals and
Plastics Publications, 1997, p. 419.
We have shown that a variety of starting materials
can be used to prepare Cu3P as a bulk, microcrystalline
pure phase material. Most importantly, we have
removed the prejudice that highly reactive yellow
phosphorus must be used for the solvothermal prepara-
tion of Cu3P and that an N-donor coordinating solvent
plays a key role in this synthesis. By making Cu3P using
a copper source, red phosphorus and H2O in closer to
[12] B. Steenberg, Arkiv. Kemi., Mineralog. Geol. 12A (1938) 1–15.
[13] M. Mansmann, Z. Kristallogr 122 (1965) 399–406.
[14] H. Schlenger, H. Jacobs, R. Juza, Z. Anorg. Allg. Chem. 385
(1971) 177–201.
[15] O. Olofsson, Acta Chem. Scand. 26 (1972) 2777–2787.
[16] (a) Y.A. Ugai, O.Y. Gukov, Izv. Akad. Nauk Neorg. Mat. 5
(1969) 598–599;
(b) O.Y. Gukov, Y.A. Ugai, V.R. Pshestanchik, V.Z. Anokhin,
Izv. Akad. Nauk Neorg. Mat. 8 (1972) 167–168.