Studies of new volatile copper(I) complexes with triphenylphosphite and perfluorinated carboxylates

Article abstract of DOI:10.1016/S0277-5387(99)00199-0

Copper(I) complexes with triphenylphosphite and aliphatic perfluorinated carboxylates of the type [Cu(μ-RCOO){P(OC₆H₅)₃}]₂, where R=CF₃, C₂F₅, C₃F₇, C₆F₁₃, C₇F₁₅, C₈F₁₇ and C₉F₁₉ have been prepared as viscous liquids and characterised by ¹³C, ¹⁹F, ³¹P, ⁶³Cu NMR, IR and mass spectra. In the liquid state, a dimeric structure with Cu in trigonal symmetry linked by the bridging carboxylates is proposed. Temperature-dependent ³¹P and ⁶³Cu NMR spectra were determined, in the 333-233 K range and single lines from both nuclei were detected, suggesting the fast exchange of triphenylphosphite in acetonitrile solution. Copper relaxation is sufficiently slow to record the signal, which is the first time that this has been reported for complexes with P(OC₆H₅)₃. Examination of ¹³C and ¹⁹F resonances confirm coordination of carboxylates in solution where few dimeric species with a geometry lower than T_d symmetry can exist. Thermal decomposition of complexes proceeds as a multistage process, yielding a mixture of Cu, Cu₂O and Cu₂P₂O₇. Elsevier Science Ltd.

Full text of DOI:10.1016/S0277-5387(99)00199-0

Polyhedron 18 (1999) 2941–2948
www.elsevier.nl/locate/poly
Studies of new volatile copper(I) complexes with triphenylphosphite and
perfluorinated carboxylates
*
´
E. Szl«yk , I. Szymanska
´
Nicholas Copernicus University, Department of Chemistry, 87-100 Torun, Poland
Received 9 March 1999; accepted 9 July 1999
Abstract
Copper(I) complexes with triphenylphosphite and aliphatic perfluorinated carboxylates of the type [Cu(m-RCOO)hP(OC6H5)3j]2,
where R5CF3, C2F5, C3F7, C6F13, C7F15, C8F17 and C9F19 have been prepared as viscous liquids and characterised by 13C, 19F, 31P,
63Cu NMR, IR and mass spectra. In the liquid state, a dimeric structure with Cu in trigonal symmetry linked by the bridging carboxylates
is proposed. Temperature-dependent 31P and 63Cu NMR spectra were determined, in the 333–233 K range and single lines from both
nuclei were detected, suggesting the fast exchange of triphenylphosphite in acetonitrile solution. Copper relaxation is sufficiently slow to
record the signal, which is the first time that this has been reported for complexes with P(OC6H5)3. Examination of 13C and 19F
resonances confirm coordination of carboxylates in solution where few dimeric species with a geometry lower than Td symmetry can
exist. Thermal decomposition of complexes proceeds as a multistage process, yielding a mixture of Cu, Cu2O and Cu2P2O7.
Elsevier Science Ltd. All rights reserved.
1999
Keywords: Copper(I); Triphenylphosphite; NMR spectra
1. Introduction
phenylphosphine and smaller steric hindrances because
phenyl rings are more flexible along the C–O–P bond [6].
We assumed that these features would be effective in
complex stabilisation and in the preparation of Cu(I)
complexes. The main purpose of the presented work was
isolation of new Cu(I) complexes and to elucidate the
coordination sphere composition and symmetry using 13C,
19F, 31P, 63Cu NMR, mass spectra and infrared results. Due
to fast ligand exchange at the labile Cu(I) centre, 31P and
63Cu NMR studies have been used relatively infrequently
for determining the structural characteristics of copper(I)
complexes. Besides the good NMR sensitivities of 63Cu
and 65Cu, the significant quadrupole moments imply that
only a limited number of copper resonances have been
observed in solution [7–10] and, in this paper, we report
on further examples. We have found that gold(I) and
silver(I) complexes with perfluorinated carboxylates and
tertiary phosphines presented sufficient volatility for
chemical vapour deposition (CVD) [11–15], hence, the
other purpose of this work was to prepare volatile pre-
cursors for CVD of copper oxide or copper metal thin
films. Thermal studies were undertaken to determine the
mechanism of thermal decomposition and to ascertain the
Copper(I) complexes with oxygen donor ligands such as
carboxylates have mono- or multinuclear structures de-
pending on the nature of the carboxylate and other ligands
in the coordination sphere [1–4]. These compounds have
been studied less than Cu(II) analogues due to the in-
stability of copper(I)–oxygen bonds. One can explain this
as a weak bond between hard oxygen and soft Cu(I) using
the theory of hard–soft acids–bases [5]. Our interest was
focused on perfluorinated carboxylates in which oxygen
atoms could perform as a harder donor, rather than in the
nonfluorinated analogues, hence, one may expect that they
could be bonded in Cu(I) complexes in different ways that
may result in interesting new structures. Tertiary phos-
phines, as p-electroacceptor ligands, are likely to stabilise
Cu(I) complexes, which may also have effects upon the
strength of the Cu–O bond. Therefore, we have chosen
triphenylphosphite [P(OC6H5)3, tpp] as the ligand, as it
should exhibit stronger p-acceptor properties than tri-
*Corresponding author. Tel.: 148-56-611-4304; fax: 148-56-654-
2477.
0277-5387/99/$ – see front matter
PII: S0277-5387(99)00199-0
1999 Elsevier Science Ltd. All rights reserved.

´
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
2942
suitability of complexes as precursors for CVD, which
should be competitive to organometallics used at present.
2.3. Synthesis
Copper(II) carboxylates of the general formula
Cu(RCOO)2 where R5CF3, C2F5, C3F7, C6F13, C7F15,
C8F17 and C9F19 were prepared by reaction of a slight
excess of copper(II) basic carbonate with the respective
carboxylic acid in a water or water–ethanol solution. The
copper content, thermal analysis, mass and IR spectra
confirmed the compositions of the carboxylates obtained.
Complexes of Cu(I) were synthesised in Schlenk glass-
ware in an argon atmosphere. In the general procedure,
copper(II) carboxylate (0.531023 mol) of the above
formula was placed in the Schlenk tube, either dissolved or
suspended in 20 cm3 of freshly distilled acetonitrile, and
copper powder (2.531023 mol) was added. The suspen-
sion obtained was stirred until the solution was pale yellow
in colour. Next, P(OC6H5)3 (1.031023 mol) was added
and the reaction mixture was stirred for about 12 h at
ambient temperature, filtered and the solvent was evapo-
rated in a vacuum line, leaving a pale yellow viscous
liquid. The complexes were unstable in air, turning green.
The results of Cu determination were as follows (%)
(calc./found): (1) C40H30Cu2F6O10P2, Cu (13.0/12.9);
2. Experimental
2.1. Materials
Triphenylphosphite (99%) and perfluorinated carboxylic
acids (97–99%) were purchased from Aldrich and were
used without further purification. Acetonitrile, of analytical
grade (Fluka), was dried over molecular sieves 4A, then
over CaH2 and, finally, was distilled under dry nitrogen
from P4O10 prior to use. Copper powder for organic
synthesis (Aldrich) and copper(II) basic carbonate (ana-
lytical grade) from POCh (Poland) were used as received.
Sodium carboxylates, for NMR and IR experiments, were
obtained from sodium bicarbonate (POCh) and perfluori-
nated carboxylic acids from the same batch as above.
2.2. Methods
(2)
(3)
(4)
(5)
(6)
C42H30Cu2F10O10P2,
C44H30Cu2F14O10P2,
C50H30Cu2F26O10P2,
C52H30Cu2F30O10P2,
C54H30Cu2F34O10P2,
Cu
Cu
Cu
Cu
Cu
(11.8/11.6);
(10.8/10.8);
(8.6/8.2);
(8.1/7.9);
(7.6/7.3);
NMR spectra in acetonitrile were recorded using a
Varian Gem. 200 MHz spectrometer; 13C was recorded at a
frequency 50 MHz and using TMS as the standard, 19F was
recorded at 188 MHz, with CFCl3 as a standard, 31P was
recorded at 80.95 MHz and the standard used was 80%
H3PO4, 63Cu was recorded at 56.79 MHz with
[Cu(CH3CN)4]ClO4 as the standard. Temperature-depen-
dent 31P and 63Cu NMR spectra (333–223 K) were
obtained using a Bruker MSL 300 MHz spectrometer in
the same solvent and using the same standards. To measure
the linewidth of the peaks accurately, no baseline correc-
tion or apodization function was used in the data process-
ing. Linewidths were measured using manual optimisation
within the peakfinder module. The error in linewidth
measurements was assumed to be equal to the digital
resolution, which was 1.95 Hz. IR spectra were determined
using a Perkin-Elmer 2000 FT IR spectrophotometer in the
range 4000–400 cm21 using film on KBr plates, whereas
400–100 cm21 spectra were run using film on poly-
ethylene plates. Thermal studies were performed on a
MOM OD-102 Derivatograph (Paulik and Paulik, Hun-
gary). The atmosphere over the sample was nitrogen, the
heating range was 20–5008C, at a heating rate of 2.5
8/min, the sample mass was 20–90 mg, the TG range was
50 or 100 mg and the reference material was Al2O3.
Powder diffractograms were recorded on an HZ64/A-2
DRON-1 (Russian Federation) diffractometer using CuKa
radiation, with l50.1542 nm. Copper was determined
using a Carl Zeiss Jena AAS spectrometer with Cu(NO3)2
solution as the standard. Mass spectra were recorded on an
AMD-640 mass spectrometer using the electron ionization
method.
(7) C56H30Cu2F38O10P2, Cu (7.2/7.0). The EI mass
spectra were recorded for the complexes in an attempt to
confirm the formulations. The fragments confirming stoi-
chiometric
composition
were
[Cu2(RCOO)2hP(OC6H5)3j]1 and [Cu2(RCOO)2]1. They
were detected at (1) 662, 352; (2) 762, 452; (3) 862, 552;
(4) 1162, 852; (5) 1262, 952; (6) 1362, 1052, respectively,
whereas for (7), only [Cu2(RCOO)2]1 at 1152 was found.
The mass spectra results suggest a dimeric structure for the
compounds and proved that the formulas proposed above
were correct.
3. Results and discussion
3.1. IR spectral analysis
Mode of carboxylic acids residues coordination can be
proposed from COO asymmetrical and symmetrical vi-
brations bands frequency analysis. Absorption bands of
COO asymmetrical and symmetrical modes appeared in
the ranges 1672–1690 cm21 and 1395–1432 cm21, respec-
tively (Table 1). As a spectroscopic criterion of car-
boxylates binding mode we have applied the parameter
defined as DnCOO5nasym2nsym [1,16–18]. Estimation of
the carboxylate linkage with a metal ion is based on the
relation between the DnCOO value calculated for the
considered complex and the DnCOO value found in the

´
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
2943
Table 1
a
Characteristic IR spectral frequencies [cm21
]
Compound
na(COO)
ns(COO)
D
D1
na(P–O)
na(Cu–O)
ns(Cu–O)
n(Cu–P)
(1) [Cu(m-CF3COO)hP(OC6H5)3j]2
(2) [Cu(m-C2F5COO)hP(OC6H5)3j]2
(3) [Cu(m-C3F7COO)hP(OC6H5)3j]2
(4) [Cu(m-C6F13COO)hP(OC6H5)3j]2
(5) [Cu(m-C7F15COO)hP(OC6H5)3j]2
(6) [Cu(m-C8F17COO)hP(OC6H5)3j]2
(7) [Cu(m-C9F19COO)hP(OC6H5)3j]2
1676
1690
1690
1691
1672
1694
1681
1432
1404
1395
1401
1404
1401
1410
244
286
295
290
268
293
271
223
268
272
272
272
272
276
890
890
885
890
890
888
895
273
296
293
294
293
295
292
210
226
218
220
206
218
206
160
149
149
150
148
145
152
a D5na(COO)2ns(COO) for complex; D15na(COO)2ns(COO) for RCOONa; na(P–O)5863 cm21 for P(OC6H5)3.
identical sodium carboxylate. The parameter DnCOO, calcu-
lated for the examined complexes, was compared to DnCOO
for the appropriate sodium carboxylate and the data from
mass spectra favoured a bridging coordination (Table 1).
Absorption bands of n(P–O) in the examined spectra
appeared in the region 885–895 cm21 (Table 1), whereas
in the spectrum of free tpp, they were at 863 cm21 [19].
Such a significant change can be taken as evidence of
triphenylphosphite coordination [20,21].
Spectra in the range of metal–ligand vibrations revealed
bands that could be assigned to Cu–O(RCOO) and Cu–P
stretching vibrations. From group theory calculations, for
C2v coordination sphere geometry, stretching vibrations of
Cu–O(RCOO) should be of the type A1 and B1, whereas
O–Cu–O angle deformation modes should be of the type
B1 and all of them were active in IR spectrum. The
Cu–O(RCOO) stretching vibration bands were found in
the range 273–296 cm21 (B1) and 200–226 cm21 (A1),
corresponding to the data reported [2]. The absorption
band observed in the range 148–160 cm21 most likely can
be assigned to Cu–P vibrations [22–24]. These results and
fragments detected in mass spectra are in favour of a
dinuclear structure of complexes with bridging carbox-
ylates and a monodentate-coordinated molecule of tri-
phenylphosphite.
One might expect that the 13C resonance of the COO
carbon would give rise to the most significant changes
upon coordination, but spectra of the complexes studied
exhibit a very weak COO signal that was shifted slightly in
relation to free acids (Table 2).
The fluorine resonances have been affected by Cu–O
bond formation in a more pronounced way than carbon and
the most distinct changes were observed for CaF2 signals
(Table 2). The resonance lines of the CaF2 groups upon
coordination were shifted downfield (D253.7–4.8 ppm) in
relation to the spectra for free acids. The coordination shift
observed for CaF2 can be taken as evidence of carboxylate
complexation in acetonitrile solution. Analogous effects
were observed in the spectra of Ag(I) and Au(I) complex-
es with perfluorinated carboxylates and tertiary phosphines
[11–13]. Uncoordinated carboxylates can be detected in
solution of sodium carboxylates of the analogous acids.
Therefore, we have recorded 19F NMR of the equivalent
sodium salts in acetonitrile solution, under the same
conditions as used for complexes being studied (Table 2,
D3 and D5). Resonances of the CaF2 fluorine in complexes
2–7, and of CF3 for 1, are shifted downfield (D350.4–1.7
ppm) in relation to those for sodium carboxylates. The
coordination shift of CaF2resonances can be taken as
evidence of a Cu–OOCR bond, although it is difficult to
correlate it with the mode of carboxylate binding.
3.2. NMR spectral analysis
Single, broad 31P (Dn1 / 2 ranging from 230–1040 Hz)
resonances in the spectra of complexes appeared in the
range 108.0–110.1 ppm, which is 20.6–22.7 ppm upfield
in relation to uncoordinated P(OPh)3 (Table 2). The
The chemical shifts for 13C, 19F, 31P and 63Cu NMR
resonances of the complexes are listed in Table 2.
Table 2
13C, 19F, 31P and 63Cu NMR spectral dataa
Compound
13C [ppm]
19F [ppm]
31P [ppm]
dCP D6
0.4 109.8 220.9
1.8 20.2 108.2 222.5 105.0
0.6 20.4 108.0 222.7 64.2
63Cu
dCCOO2 D1
dCCaF2 D2
D3
dCCF3 D4
3.2 1.9
D5
dC [ppm] Dn1 / 2 [Hz]
(1) [Cu(m-CF3COO)hP(OPh)3j]2
(2) [Cu(m-C2F5COO)hP(OPh)3j]2
(3) [Cu(m-C3F7COO)hP(OPh)3j]2
(4) [Cu(m-C6F13COO)hP(OPh)3j]2 159.0
(5) [Cu(m-C7F15COO)hP(OPh)3j]2 159.0
(6) [Cu(m-C8F17COO)hP(OPh)3j]2 159.1
(7) [Cu(m-C9F19COO)hP(OPh)3j]2 159.3
159.0
159.3
159.4
0.6
2
2
2
97.0
8750
22520
11400
9280
13250
8150
0.1 241.4
4.2
0.7 24.9
0.6 24.0
1.7 23.3
1.2 23.2
1.0 23.1
23.6
20.2 239.7
20.3 237.2
20.4 237.7
20.2 237.8
20.1 238.0
3.7
4.8
4.2
4.0
3.9
0.5
0.6
0.9
0.2
0.0 108.9 221.8 110.0
0.2 110.1 220.6 125.0
0.3 109.0 221.7 105.0
109.1 221.6 150.0
b
b
20140
a D15dCCOO22dACOO2; D25dCCaF22dACaF2; D35dCCaF22dSCaF2; D45dCCF32dACF3; D55dCCF32dSCF3; D65dC2d P(OC6H5)3;
P(OC6H5)35130.7 ppm; C, complex; A, carboxylic acid; S, sodium carboxylate.
d
b Due to the low solubility of C9F19COONa, its spectrum could not be recorded.

´
2944
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
upfield shift in 31P spectra favours the p-back donation
from Cu(I) to phosphorus in tpp. The signal is not split,
indicating that, in solution, the CuP4 moiety does not exist
as a prevailing form or that a fast exchange reaction causes
the lack of 31P–63Cu coupling. Exchange reactions present
in solution can be studied using temperature-dependent 31P
NMR spectra. We have recorded spectra in the 333–233 K
range, expecting to find the coalescence temperature
between different forms in solution. However, in the
spectra of 2, 5 and 7, splitting of the phosphorus signal
was not found.
The detection of a singlet suggests the appearance of at
least two pseudotetrahedral species in equilibrium that
averaged and broadened the signal [25–27]. Considering
the above results, one can suggest coordination of tri-
phenylphosphite along with bridging carboxylates and
acetonitrile. The prevailing existing species can be de-
scribed as isomers of [Cu(CH3CN)(m-RCOO)hP(OPh)3j]
(Fig. 1). This assumption may be explained by the fast
exchange of P(OPh)3, and the lack of a favoured configu-
ration with one phosphorus in the coordination sphere,
which is typical for the large phosphine ligands such as
triphenylphosphite [28] and triphenylphosphine (PPh3).
Table 3
31P and 63Cu NMR data of [Cu(CH3CN)4]ClO4 (c50.1 m):nP(OC6H5)3
system in acetonitrile solution
n
63Cu NMR
31P NMR
d [ppm]
Dn1 / 2 [Hz]
d [ppm]
Dn1 / 2 [Hz]
0.0
0.5
1.0
1.5
2.0
3.0
4.0
5.0
0.2
19.0
62.7
440
2560
4680
5300
7600
6360
6710
5080
2
2
1370
1240
680
460
280
108.9
106.5
109.7
104.7
117.6
118.1
118.3
138.2
141.0
152.0
145.1
149.2
360
160
We have checked the possibility of [CuhP(OPh)3j4]1
detection in the acetonitrile solution recording 31P NMR
spectra for [Cu(CH3CN)4]ClO4: P(OPh)3 system. The 31P
single resonance lines for the systems Cu:nP(OPh)3 from
1:0.5 to 1:5 were observed (Table 3). Even at the ratio 1:5,
the resonance line from the uncoordinated P(OPh)3 (130.7
ppm) was not found. Temperature-dependent spectra of
solutions with M:L ratios of 1:1 and 1:4 were also
investigated. In the 1:1 system, only an upfield shift of
Fig. 1. Possible species existing in equilibrium after dissolving [Cu(m-RCOO)hP(OPh)3j]2 in acetonitrile.

´
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
2945
phosphite signals, with decreasing temperature, was ob-
served. However, when the ratio of M:L was 1:4, the
single phosphorus resonances that had been shifted down-
field were detected, and these were split at 233 K into two
lines (127.4 and 109.5 ppm) (Fig. 2). The latter was
positioned at the same frequency as in the 1:1 system and
was in a similar range as for complexes 1, 4, 5, 6 and 7.
We can presume that this signal is from the species where
one molecule of P(OPh)3 is coordinated.
The main factors that influenced the 63Cu NMR signals
were the quadrupolar relaxation of 63,65Cu, and the nature
and rate of dynamic processes occurring in solution. These
in turn were influenced by the temperature, distortions
from the regular tetrahedral symmetry, and the electronic
and steric properties of ligands. Studies of Cu(I) halide
complexes [29–31] with phosphine ligands revealed that
the dissociation of the complex in solution increased with
ligand size, which, for tertiary phosphines, can be ex-
pressed by the Tolman cone angle. Furthermore, [CuL4]X
is regarded as a dominant species in solution [29,32–35],
although in cases where there is an aromatic substituent on
phosphines, several mononuclear and polynuclear com-
plexes can also exist, i.e., CuL3X, CuL2X or Cu2L3X2,
depending on the coordinative properties of X [29]. Ligand
exchange rates are usually too fast on the NMR time scale
to identify individual species, except for very low tempera-
tures. As reported, the extent of ligand dissociation in
solution and the corresponding fast quadrupolar relaxation
of the Cu nuclei in the low symmetry coordination resulted
in a lack of 63Cu resonances [25].
63Cu spectra of the complexes revealed a broad singlet
(Dn1 / 2 from 8150 to 22 520 Hz) in the range 64.2–150.0
ppm. In relation to complexes with PR3, where R5Me or
Et, they are shifted upfield and their frequency is closer to
that of PR32x(OR)x ligands [36]. The 63Cu resonances
reported for tetrahedral [CuhP(OR)3j4]Cl where R5Me or
Et were found to be in the range 70–90 ppm [7,36] in the
form of a quintet from Cu–P coupling, except for
[CuhP(OC6H5)3j4]ClO4, for which the authors did not
observe a 63Cu signal. To the best of our knowledge, the
63Cu resonances reported here are the first reported for
Cu(I) complexes with triphenylphosphite. The expected
spin–spin coupling 1J(31P–63Cu) in 63Cu and 31P NMR
spectra of the complexes under discussion was not de-
tected. The lack of line-splitting can be caused by the fast
exchange of P(OC6H5)3 and the low symmetry of the
coordination centre. Considering the 13C, 19F, 31P NMR
and IR spectra results, one can assume a lower symmetry
(C3v or C2v) for the complexes studied in acetonitrile
solution.
We have registered temperature-dependent 63Cu NMR
spectra of 5 and 7 in the range 333–233 K (Table 4).
Initially, the signals are shifted upfield, but, from 253 K,
the reverse tendency is observed. The linewidth for 5 with
a decrease in temperature expands followed by a signifi-
cant decrease and a subsequent increase. In the case of 7,
Fig. 2. Temperature-variable 31P NMR spectra of the [Cu(CH3CN)4]-
ClO4:4P(OC6H5)3 system in acetonitrile solution.

´
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
2946
Table 4
maximum, but at 233 K, it was impossible to determine the
signal position due to the very strong line broadening that
occurred. It should be added that resonance linewidths
between 333–273 K are close to those obtained for M:L5
1:1, but further cooling broadened the signals very rapidly.
Bearing in mind that the 63Cu line appears only when the
cation [CuhP(OR)3j4]1, which undergoes fast chemical
exchange is present in the solution, than with temperature
decrease the reduction of the linewidth should be observed.
In our system, the opposite effect was detected, that is, it
favoured species with a symmetry lower than tetrahedral.
We presume that our solution cannot be classified as a
system with very fast quadruple relaxation, otherwise,
simultaneously with the increase in the linewidth in 63Cu
NMR, a rapid narrowing of the signal in 31P NMR should
be observed.
Temperature-variable 63Cu NMR spectra in acetonitrile solution
T [K]
(5) [Cu(m-C7F15COO)-
hP(OPh)3j]2
(7) [Cu(m-C9F19COO)-
hP(OPh)3j]2
d [ppm]
Dn1 / 2 [Hz]
d [ppm]
Dn1 / 2 [Hz]
333
313
297
273
253
233
2
2
94.9
111.2
110.4
79.1
22.9
50.7
6060
6130
7240
8150
25 220
18 880
2
2
112.1
111.2
146.2
123.1
13190
14230
5430
7500
Dn1 / 2 increases up to 253 K and, afterwards, a reduction is
observed. It should be added that, for 7 in the 333–297 K
range, the chemical shift and linewidth changes are similar
to those for the systems presented below without car-
boxylates. Below 297 K for 7, and for 5 over the entire
range of temperatures, the overall pattern of changes is
different, most probably being caused by coordinated
carboxylates in the studied system.
In the case of [Cu(CH3CN)4]ClO4:nP(OC6H5)3, where
M:L51:1, the relation of the 63Cu chemical shift to
temperature was examined and a linear relationship (corre-
lation coefficient R250.9241; straight line slope coeffi-
cient, 20.46) was observed in the 333–273 K range,
however, below 273 K, a slight deviation was detectable.
The magnitude of the straight line slope coefficient reached
a mean value between the systems with nitrogen donor
ligands (20.7) and the complexes mentioned above of the
type [CuhP(OR)3j4]ClO4 (ca. 20.1), where R is an ali-
phatic substituent. An increase in the linewidth with a
reduction in temperature (3880 Hz at 333 K to 17 200 Hz
at 233 K) was also observed. For the system where
M:L51:4, the function d5f(T) at 273 K reached a
3.3. Thermal analysis
Results of thermal analysis are listed in Table 5.
Examination of the thermoanalytical curves suggests a
two-stage, exothermic decomposition process of complex-
es 1, 2 and 3 (Table 5). Analysis of the thermogravimetric
(TG) curve indicates that, in the first step, decomposition
of a carboxylic acid residue and one molecule of phosphite
occurs, which is followed by dissociation of the second
molecule of tpp. Decomposition reactions of 1, 2 and 3
were completed in the range 618–653 K and the final
product appeared to be a mixture of Cu, Cu2O and
Cu2P2O7 (3:1:1 for 1; 1:1:1 for 2 and 3). Powder diffrac-
tograms of the final products revealed lines for Cu, 0.181,
0.208 nm, and Cu2O, 0.249, 0.215, 0.175, 0.151 nm, that
correspond to those reported in the Powder Diffraction File
[37]. IR spectra of the final product revealed strong
Table 5
Results of thermal analysis (in nitrogen)
Compound
Heat
effect
Temperature range [K]
Weight loss on TG [%]
Detached
group
Final product
xCu:yCu2O:zCu2P2O7
a
a
a
Ti
Tm
Tf
Found
Calc.
(1) [Cu(m-CF3COO)hP(OC6H5)3j]2
(2) [Cu(m-C2F5COO)hP(OC6H5)3j]2
(3) [Cu(m-C3F7COO)hP(OC6H5)3j]2
(4) [Cu(m-C6F13COO)hP(OC6H5)3j]2
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
exo
298
478
298
493
298
488
298
413
498
313
428
508
298
418
503
303
423
568
453
533
458
478
618
493
653
488
653
413
498
643
428
508
643
418
503
638
708
49.3
32.2
51.8
29.1
56.0
26.7
17.5
48.4
20.6
19.0
47.9
19.7
20.6
46.7
20.6
88.1
49.5
31.9
52.2
28.9
56.3
26.4
P(OC6H5)3
P(OC6H5)3
P(OC6H5)3
3:1:1
1:1:1
1:1:1
548
b
473
65.2
21.0
553
P(OC6H5)3
P(OC6H5)3
1:1:1
1:1:1
b
(5) [Cu(m-C7F15COO)hP(OC6H5)3j]2
(6) [Cu(m-C8F17COO)hP(OC6H5)3j]2
(7) [Cu(m-C9F19COO)hP(OC6H5)3j]2
478
67.4
19.7
538
b
473
548
473
87.9
88.6
1:1:1
1:1:1
a Ti, initial temperature; Tm, maximum temperature; Tf, final temperature.
b Unable to determine.

´
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
2947
vibrations in the range 1065–1071 cm21 that were charac-
teristic for Cu2P2O7 [38]. Similar final products were
reported for other copper(I) phosphine complexes, e.g.
[CuhP(C6H5)3j4]BF4 [39]. Complexes 4, 5 and 6 de-
compose in three exothermic processes and the final
product is the same mixture as for 2 and 3. For 1–5, the
results of analysis of TG curves are in favour of tpp
detachment in the second stage. The onset temperatures of
the consecutive stages for 6 are very close; therefore, the
total mass loss on the TG curve was taken as evidence of
tpp and carboxylate dissociation. Complex 7 decomposed
in a single stage to the same mixture as for 2–6. The onset
temperature of the first exotherm is in the range 298–313
K, suggesting that the perfluorinated chain had a small
influence on the thermal stability of the complexes studied.
The temperatures of the final product formation are in the
range acceptable for CVD purposes with the hot wall
reactor (625 K). The final products of complex decomposi-
tion are discouraging for use in the CVD of copper or
copper oxide films, however, the complexes are volatile
viscous liquids and can decompose in different ways under
conditions applied for CVD or other deposition techniques.
Ciesielski (Center for Macromolecular Research, L« odz) for
recording temperature-dependent NMR spectra.
´ ´
References
[1] D.J. Darensbourg, M.W. Holtcamp, E.M. Longridge, K. Klausmeyer,
J.H. Reibenspies, Inorg. Chim. Acta 227 (1994) 223.
[2] D.A. Edwards, R. Richards, J. Chem. Soc., Dalton Trans. 7 (1975)
637.
[3] R.D. Hart, P.C. Healy, G.A. Hope, D.W. Turner, A.H. White, J.
Chem. Soc., Dalton Trans. 5 (1994) 773.
[4] R.L. Beddoes, J.A. Connor, D. Dubowski, A.C. Jones, O.S. Milles,
R. Price, J. Chem. Soc., Dalton Trans. 10 (1981) 2119.
[5] D.F. Shriver, P.W. Atkins, C.H. Langford (Eds.), Inorganic Chemis-
try, Oxford University Press, Oxford, 1994, pp. 212–216.
[6] C.A. Tolman, Coord. Chem. Rev. 77 (1977) 313.
[7] P. Granger, in: P.S. Pregosin (Ed.), Transition Metal Nuclear
Magnetic Resonance, Elsevier, Amsterdam, 1991, pp. 265–273.
[8] B. Mohr, E.E. Brooks, N. Rath, E. Deutsch, Inorg. Chem. 30 (1991)
4541.
[9] S.J. Berners-Price, C. Brevard, A. Pagelot, P.J. Sadler, Inorg. Chem.
25 (1986) 596.
[10] S.J. Berners-Price, L.A. Colquhoun, P.C. Healy, K.A. Byriel, J.V.
Hanna, J. Chem. Soc., Dalton Trans. (1992) 3357.
[11] E. Szl«yk, I. L« akomska, A. Grodzicki, Polish J. Chem. 68 (1994)
1529.
4. Conclusions
[12] E. Szl«yk, I. L« akomska, A. Grodzicki, Polish J. Chem. 69 (1995)
1103.
[13] E. Szl«yk, I. L« akomska, A. Grodzicki (Eds.), Trends in Coordination
Chemistry, Slovak Technical University Press, 1995, p. 151.
[14] D.L. Schultz, T.J. Marks, Adv. Mater. 6 (1994) 719.
[15] V.G. Minkina, Russ. Chem. Bull. 42 (1993) 1461.
[16] K. Nakamoto (Ed.), Infrared, Raman Spectra of Inorganic, Coordi-
nation Compounds, Wiley & Sons, New York, 1986, p. 233.
[17] G.B. Deacon, R.J. Phillips, Coord. Chem. Rev. 33 (1980) 227.
[18] N.F. Curtis, J. Chem. Soc. (A) (1968) 1579.
[19] Atlas IK-Spectrov fosforoorganichieskih soiedinienij, Moskva,
1977, p. 237.
[20] V.I. Kurashov, D.G. Pobiedimskij, P.A. Kirpichnikov, Doklady
Akad. SSSR 238 (1978) 1407.
[21] V.I. Kurashov, D.G. Pobiedimskij, Koord. Khimia 5 (1979) 1162.
[22] D.A. Edwards, R. Richards, Spectrochim. Acta 34A (1978) 167.
[23] R.L. Keiter, J.G. Verkade, Inorg. Chem. 9 (1970) 404.
[24] G.A. Bowmaker, R.D. Hart, B.E. Jones, B.W. Skeleton, A.H. White,
J. Chem. Soc., Dalton Trans. 24 (1995) 3063.
[25] A. Marker, M.J. Gunter, J. Magn. Reson. 47 (1982) 118.
[26] S. Kitagawa, M. Munakata, M. Sasaki, Inorg. Chim. Acta 120 (1)
(1986) 77.
[27] D.G. Pobiedimskij, D. Szeller, T.N. Kondratieva,V.I. Kurashov, P.A.
Kirpichnikov, Doklady Akad. SSSR 250 (1980) 661.
[28] J.R. Black, W. Levason, M.D. Spicer, M. Webster, J. Chem. Soc.,
Dalton Trans. 20 (1993) 3129.
These results show that neat isolated complexes can be
regarded as trigonal dimeric species with bridging car-
boxylates and triphenylphosphite of general formula
[Cu(m-RCOO)h(P(OPh)3j]2. This type of complexation is
different from tetrameric [40] or polymeric [41] structures
found for Cu(I) carboxylates in the solid state [42].
Trigonal geometry was found for dimeric Cu(I) halide
complexes with triphenylphosphine, but, to the best of our
knowledge, not for triphenylphosphite. In our complexes,
the bridging carboxylates favour dimer formation. Analysis
of 63Cu NMR spectra suggests the fast exchange of tpp in
acetonitrile solution in the range 333–233 K. Copper
relaxation is sufficiently slow to record the signal, which
can most probably be related to the bridging carboxylates.
The complexes examined are the first systems with tpp
where 63Cu resonances were observed. This is a rare
example of complexes with a geometry lower than Td, for
which 63Cu signals have been registered. In Cu(I) com-
plexes with carboxylates and monodentate phosphine, only
the chelating mode of carboxylate was noticed [3], but the
bridging mode seems to be characteristic for tertiary
phosphites.
[29] S.J. Lippard, J.J. Mayerle, Inorg. Chem. 11 (1972) 753.
[30] D.J. Fife, W.M. Moore, K.M. Morse, Inorg. Chem. 21 (1984) 1684.
[31] E.L. Muetterties, C.W. Alegranti, J. Am. Chem. Soc. 92 (1970)
4114.
[32] P.S. Pregosin, R.W. Kunz, 31P, 13C NMR of Transition Metal
Complexes, Springer-Verlag, Berlin, 1979.
Acknowledgements
[33] J.T. Gill, J.J. Mayerle, P.S. Weleker, D.F. Lewis, D.A. Ucko, D.J.
Barton, D. Stowens, S.J. Lippard, Inorg. Chem. 15 (1975) 1155.
[34] A.P. Gaughan, Z. Dori, J.A. Ibers, Inorg. Chem. 13 (1974) 1657.
[35] M.B. Dines, Inorg. Chem. 11 (1972) 2949.
[36] P. Kroneck, O. Lutz, A. Nolle, H. Oehler, Z. Naturforsch. 35a
(1979) 221.
Financial support from the Polish Committee for Sci-
entific Research (KBN) is gratefully acknowledged, grant
PB 619/T09/97/13. The authors wish to thank Z. Hup-
penthal (N.C.U.) for 31P, 63Cu NMR measurements and W.

´
E. Szl«yk, I. Szymanska / Polyhedron 18 (1999) 2941–2948
2948
[37] Powder Diffraction File, Sets 4-836, 5-667, 22-1090; Joint Commit-
tee on Diffraction Standards, 1977.
[40] M.G.B. Drew, D.A. Edwards, D.A. Richards, J. Chem. Soc., Dalton
Trans. 3 (1977) 299.
[38] R.A. Nyquist, R.O. Kagel (Eds.), Infrared Spectra of Inorganic
Compounds, Academic Press, New York, 1971, pp. 186–187.
[39] V.A. de Lucca Neto, A.E. Mauro, V. Sargentelli, M. Ionashitro,
Thermochim. Acta 260 (1995) 235.
[41] M.G.B. Drew, D.A. Edwards, D.A. Richards, J. Chem. Soc., Chem.
Commun. 4 (1973) 124.
[42] T.P. Lockhart, D.A. Haitko, Polyhedron 4 (1985) 1745.