3
40
P. Buchwalter et al. / Inorganica Chimica Acta 409 (2014) 330–341
Table 14
Relative proportion of crystalline phases detected for cluster 2 under He.
Table 15
Relative proportion of crystalline phases detected for cluster 3 under He.
Cluster 2
(°C)
00
00
00
00
50
00
00
Co
Co
4
(CO)10
(PO
(l
4
-PPh)
2
Cluster 3
% (°C)
Co
Co
4
(CO)
8 2
(l-dppa)
%
3
4
)
2
Co
Co
2
P
2 2
(P O
7
)
2
Co P
2
4
5
6
6
7
8
–
–
–
–
32
37
47
–
–
–
–
–
52
38
29
9
200
400
500
600
650
700
800
–
–
–
–
–
–
49
–
–
–
–
100
100
51
100
48
30
34
44
Mapping of the composition by SEM showed that this sample
originating from cluster 1 was rather homogeneous (Fig. 12).
Moreover, as expected, the EDX measurements give a molar ratio
Co/P about 2:1 (Table 4). When it comes to cluster 2 with the same
Co/P ratio as 1, XRD (Fig. 7) confirms the formation of crystalline
phosphate is obtained. Then, from 650 to 800 °C, only the propor-
tions vary (Table 14). At 650 °C, each phase accounts for ca. 1/3 of
the crystalline powder. When the temperature is increased, the
amounts of cobalt phosphate and cubic cobalt increase too, while
that of cobalt phosphide decreases. At 800 °C, cobalt and cobalt
2
Co P. The use of the Scherrer equation leads to an average crystal-
phosphate account for 47% and 44% respectively, while Co
2
P is only
lite size of 28 nm at 650 °C and 35 nm at 800 °C. Compared to 1, the
crystallite size varies not much between 650 and 800 °C. Actually,
cluster 2 decomposed at much lower temperature than 1 and 3
(Fig. 3) and the product exhibits a composition close to that of
9
%. So unlike in the case of cluster 1, the main crystalline phase is
3 4 2
not cobalt phosphide, but Co (PO ) despite an identical Co/P ratio
in the precursor cluster.
SEM observations in composition mode confirmed that the
sample is rather homogeneous (Fig. 10A). The EDX spectroscopy
indicates the presence of ca. 72% of Co for 14% of P. This analysis
is in quite good agreement with that found by XRD. The important
part of cubic cobalt in the crystalline part of the sample, and the
ratios Co to P in cobalt phosphate and phosphide (3 to 2 and 2 to
2
Co P (Fig. 13, Table 5 and Table 6) which is then already well crys-
tallized at 650 °C. One reason why cluster 2 decomposes more rap-
idly than clusters 1 and 3 might be attributed to the coordination
mode of the P atom of the
4
l -PPh ligand. Indeed, the Co-P bond en-
ergy in cluster 2 might be weaker than in the case of the single
bonds found in clusters 1 and 3. Also the chemical nature of the
phosphine might be important.
1
respectively) contribute to the high proportion of Co.
The behavior of cluster 3 is easier to understand under He
Cluster 3 shows once again a different behavior, owing to its
atmosphere. Indeed, only two crystalline phases were observed.
Co P is formed around 650 °C. It remains the only crystalline phase
at 700 °C. At 800 °C however, the cobalt phosphate Co (P ) ap-
pears. The stoichiometry of the starting cluster corresponds to a
Co/P ratio of 1:1 like in Co (P ). At 800 °C, the XRD peak intensi-
composition. Indeed, the phases obtained both at 650 and 800 °C
under H
intensities suggest that CoP is the main phase, accounting for 79%
at 650 °C (21% for Co P) and for 68% at 800 °C (32% for Co P). How-
ever, one should notice that its quantity decreases when the tem-
perature increases, while the amount of Co P increases. This is
2 2 2
/N are cobalt phosphides Co P and CoP (Fig. 8). XRD peak
2
2
2 7
O
2
2
2
2 7
O
ties suggest a mixture of about 50% of cobalt phosphate and 50% of
2
cobalt phosphide (Table 15). Hence, any other phosphate phase is
undoubtedly connected to the thermal stability of CoP [64]. Indeed,
amorphous. Actually, Co
and at high temperatures (600 °C from salt precursors) [63] while
Co P crystallizes around 650 °C, according to our observations.
SEM observations support these results. The bright NPs (spot 1
on Fig. 11B) are made of Co and P, with a ratio around 2:1, which is
consistent with the crystalline Co P detected by XRD. The phase at
spot 2 is oxygen-rich and has a lower Co:P ratio close to 1, there-
fore it appears less bright. This can be ascribed to Co (P ), as de-
tected by XRD above 800 °C. The SEM observations indicate that
the main phase is Co (P ) although it is not fully crystalline at
00 °C. According to XRD, the amount of crystalline phase is similar
2
(P
2
O
7
) is known to crystallize very slowly
the CoP phase is stable at temperatures as high as ca. 1000 °C,
while Co
lites determined using the Scherrer equation are about 14 nm for
CoP at 650 °C, and around 30 nm at 800 °C. The size of the Co
2
P can still exist at 1380 °C. The average size of the crystal-
2
2
P
crystallites could not be estimated in this sample, because the
reflection peak used in the case of compound 1 does not appear
on the diffractogram, Co P being the minor component.
2
The SEM observations corroborate these first results. At 650 °C,
two different regions were observed on the pictures (Fig. 14) using
the composition mapping mode: the relative brightness of the re-
gions of spots 1 and 2 suggests that the average atomic number is
higher in spot 2 than in spot 1. The ‘‘grey’’ part seems to be the
main phase. The EDX spectroscopy (Table 7) suggests a Co/P ratio
close to 1:1 for spot 1, which is consistent with CoP. On the other
2
2
2
O
7
2
2 7
O
8
2
to that of Co P and the rest is amorphous.
As expected, the He gas protects the sample better against oxi-
dation than N , since we could identify cobalt phosphide in all
2
cases. However, it is obvious that a neutral atmosphere is not suf-
ficient to yield cobalt phosphide without forming phosphates.
Therefore we used a slightly reducing atmosphere consisting of
2
hand, spot 2 gives a ratio close to 2:1, which points toward Co P. At
800 °C, the phenomenon is quite similar, that is a ‘‘grey’’ phase
(predominant), and ‘‘white’’ and faceted crystals (Fig. 15). Here
again, the elemental analysis from EDX is clear: ‘‘white’’ crystals
have a ratio Co to P of 2:1, corresponding to Co P, and the ‘‘grey’’
2
part is CoP (Co to P ꢂ 1:1).
H
2
/N
phide from our clusters. Indeed, according to XRD (Fig. 6), cluster
forms Co P both at 650 and 800 °C. Heating to 800 °C improves
2
(5/95%) which appears to be the key to yield cobalt phos-
1
2
the crystallinity of the product. The crystallite size was evaluated
Kꢃk
by using the Scherrer equation S ¼
, where K is a shape factor
bꢃcos h
5. Conclusions
equal to 0.9 for spherical crystallite, k is the wavelength of the X-
ray used, b is the difference of the full width at half maximum
The study of the thermal behavior of three tetranuclear Co-P
organometallic carbonyl clusters under different atmospheres re-
vealed that a slightly reducing atmosphere is necessary to obtain
cobalt phosphide only, and completely avoid the formation of co-
balt phosphate. The combined use of TGA, (temperature-resolved)
XRD and SEM allowed us to understand the influence of the precur-
(
FWHM) between peaks of the studied material and a standard
which gives the resolution of the instrument, and h the corre-
sponding Bragg angle of the diffraction peak analyzed. Applied
on the d020 reflections it leads to an average size of 17 nm at
6
50 °C, and 46 nm at 800 °C.