2024 Volume 65 Issue 8 Pages 946-953
Copper-based conductive ink has received attention to fabricate thin, flexible, and lightweight devices through printing techniques. In particular, the copper formate based conductive inks, called metal organic decomposition (MOD) inks, which are fabricated by using amine ligand coordinated copper formate have been well studied because of higher oxidative resistant against air than that of metallic copper before thermal annealing. The copper formate moiety of amine ligand coordinated copper formate complexes varies by changing the coordinated amine species, however, for the thermal decomposition temperature of such ink is not well understand. Here, we analyzed the influence of the amine ligand on the thermal decomposition temperature drop of copper formate moiety. Amine–copper formate complexes bearing several amine ligands containing primary amine, pyridine, and imidazole groups were fabricated. The relationship between the evaluated decomposition temperature and the differences between the wavenumbers of symmetric and antisymmetric vibration mode peaks in Fourier transform infrared (FTIR) spectra was found to be nearly linear. This finding demonstrates that the thermal decomposition temperature is governed by the structural modification of formate ions by the coordination of amine groups. Based on this finding, the order of thermal decomposition temperature of the copper moiety is predictable by using FTIR measurement.
Recently, flexible, thin, and lightweight devices have been manufactured using a printing technique known as the printed electronics technology [1–6]. Compared to traditional lithography-based technology, this printed electronics technology reduces material consumption and the number of process steps such as making the mask, and etching.
In the fabrication of electronics devices using the printed electronics technology, conductive inks are widely used to print conductive circuits and electrodes on a flexible substrate, such as plastics and paper [7, 8]. Among conductive materials, metallic copper has the potential to be used for fabricating conductive ink owing to the second highest conductivity (ρ = 1.72 × 10−8 Ωm), high migration resistivity, and greater abundance than that of silver (which has been the most used element to fabricate conductive ink)-indicating potential for use in fabricating conductive ink [9]. Accordingly, many researchers have studied and reported copper conductive inks [10–13]. However, metallic copper is inherently unstable against oxidation in air. Consequently, nanosized Cu crystals rapidly oxidize under atmospheric conditions to generate copper oxide (Cu2O). More recently, copper formate (CuF)-based conductive ink known as the metal organic decomposition (MOD) ink, has been reported [14–26]. CuF is one of the self-reducible copper salts with infinite coordination network [27]. Additionally, the thermal decomposition of the formate anion provides a reducing atmosphere to synthesize the Cu metal [28, 29]. Consequently, metallic copper circuits and electrodes can be fabricated by the thermal annealing followed by the printing with the MOD inks.
Coordination of amine molecules with Cu formate is essential for the fabrication of MOD ink. This is because of two reasons. First, the solubility of the complex can be modified by changing the amine ligand. Because of the infinite coordination structure of CuF, it is difficult to dissolve CuF in any solvent while keeping the coordination bonding between the copper ion and the formate ion [30]. In contrast, the coordination of the amine ligand to CuF breaks the infinite coordination structure of CuF and forms a CuF-based the mononuclear coordination compound (L-CuF). Second, coordination of the amine ligands modifies the thermal decomposition temperature of the CuF moiety of L-CuF. This change in thermal decomposition temperature is important for the fabrication of CuF-based MOD inks. To print the conductive pattern on a substrate with low thermal resistivity, the thermal decomposition temperature of L-CuF must be low. Conversely, a lower decomposition temperature of L-CuF will results in a Cu conductive pattern with highly resistivity [31]. Therefore, precise tuning of the thermal decomposition temperature of L-CuF is required for the fabrication of L-CuF based MOD ink with desired characteristics. In this context, the reason for the decrease in the thermal decomposition temperature of the CuF moiety depending on the amine ligands remains unknown. Paquet et al. hypothesized that the breaking of the infinite CuF coordination networks owing to the coordination of an amine ligand would result in a decreasing the thermal decomposition temperature [32]. According to our previous findings, the formation of an amine coordinated mononuclear metal oxalate complex lowered the thermal decomposition temperature of the metal oxalate moiety [33–36]. Therefore, the breaking of the coordination network with additive thermal energy is key to decreasing the thermal decomposition temperature of self-reducible coordination polymer. Paquet et al. have also hypothesized that the possible number of the hydrogen bonding between the formate ion and the coordination group surrounding the central copper ion influences the decomposition temperature of CuF moiety [31, 32]. Consequently, the decomposition temperature appears to depend on the coordination group surrounding the central copper ion. In contrast, an amine-coordinated CuF complex with the same coordination group such as, amine and pyridine, but different substituent group shows various decomposition temperature. Therefore, these two hypotheses are not sufficient to fully explain the reason for the change in thermal decomposition temperature.
Herein, we propose and demonstrate that the structural change of the formate ion in L-CuF modifies the thermal decomposition temperature. First, we prepared two alkylamine, three pyridine and three imidazole derivatives coordinated L-CuF. Subsequently, the thermal decomposition temperature and structure of the formate in L-CuF were measured by thermogravimetry mass spectrum system (TGMS) and Fourier-transform infrared (FTIR) spectroscopy, respectively. Differences between the wavenumbers of symmetric- and antisymmetric-vibration mode peaks of carboxylic acid and thermal decomposition temperature show a liner relationship. This result suggests that the structural characteristics such as C-O bond length, M-O bond strength, and O-C-O angle control the thermal decomposition temperature of the CuF moiety in L-CuF.
Copper(II) formate tetrahydrate (CuF·4H2O, 98%) was purchased from Fujifilm Waco Chemicals. 2-isopropylimidazole (2-iProim, 98.0%) and methanol (99.8%) were purchased from Sigma-Aldrich Co. LTd. Hexylamine (HexAm, 99.0%), dodecylamine (DodAm, 97.0%), 1-propylimidazole (1Proim, 97.0%), 1-isopropylimidazole (1iProim, 97.0%), 2-benzylpyridine (2Benzpy, 98.0%), 4-methoxypyridine (4Metpy, 98.0%), 4-tert-butylpyridine (4tBuPy, 96.0%), and acetonitrile (99.5%) were purchased from Tokyo Chemical Industry Co. Ltd. Hexane (96.0%), 1-propanole (99.0%), and toluene (99.0%), Kanto chemical Co., Inc. All chemicals were used as received without further purification.
2.2 Preparation of alkylamine, imidazole, and pyridine coordinated Cu formate complexAlkylamine-, imidazole-, or pyridine coordinated-CuF complexes were prepared by modified method reported previously [37]. The procedure of the preparation of 1Proim-coordinated Cu(II) formate (1Proim-CuF) is described below.
CuF·4H2O (1.13 g, 5 mmol) and 1Proim (1.02 g, 9 mmol) were mixed with 30 mL of acetonitrile. Then, the mixture was stirred for 30 min in ambient conditions. The changing from a pale blue suspension to deep blue solution suggested the coordination of 1Proim to copper ion. The mixture was filtrated to remove the unreacted copper formate. Acetonitrile was removed by using rotary evaporator under reduced pressure. After removing the acetonitrile, a dark blue oil containing 1Proim-CuF was recovered. The recovered oil was dissolved in toluene to crystallize 1Proim-CuF for 3 days. The obtained crystals were filtrated and washed with toluene. Elemental analysis of 1Proim-CuF indicate that it consists of one CuF unit and two 1Proim molecules [calculated weight percent (mass%) for C14H22N4O4Cu·1.0H2O: C 42.65, H 5.82, N 14.23; found mass% C 42.65, H 5.82, N 14.23]. The procedures for preparation of other alkylamine-, imidazole-, and pyridine-coordinated Cu(II) formate complexes are described in the Supporting Information.
2.3 CharacterizationDiffuse reflectance UV-vis spectra were collected using SHIMADZU 2600i equipped with integrating sphere attachment (ISR-2600) ranging from 400 to 1400 nm. Powder X-ray diffraction (XRD) patterns were measured using a Rigaku Miniflex600 equipped with Dtex Ultra II in the range from 10 to 40°. Fourier-transformation infrared (FTIR) spectra were obtained using a Thermo Scientific Nicolet iS5 in the 525 to 4000 cm−1 range. Thermogravimetric-mass analysis (TGMS) was performed using a NETZSCH STA2500R connected with a JEOL JMS-Q1050 Master Quad GC/MS. The measurements were performed under an Ar flow of 50 mL/min. The temperature was increased to 200°C at a rate of 5°C/min.
2.4 DFT calculationOptimizations of the structure of the complexes were performed by density functional theory (DFT) with B3LYP/6-31G(d) level of theory in vacuo using with Gaussian 16 Rev A.03 [38]. The vibrational frequencies and strengths of the corresponding optimized structures were calculated from the optimized structures and no imaginary frequencies were found. The calculated wavenumbers were multiplied by the vibrational scaling factor of 0.960 for the correction [39].
To determine the factors controlling the decomposition temperature of the CuF moiety, amine ligand-coordinated Cu formate complexes were prepared using eight different types of ligands (L-CuF). (Fig. 1) Elemental analysis revealed that the ratio of the Cu ion, formate ion, and amino ligand was 1:2:2 in all the L-CuFs (see Appendix). Figure 2 shows the photograph of the CuF·4H2O (Fig. 1(a)) and L-CuFs (Fig. 2(b)–(f)). Each crystals exhibits a blue color. Figure A1 shows the diffuse reflectance UV-vis spectra of L-CuFs obtained after mixing CuF and amine ligand show the different shape of absorption bands with absorption maximum ranging from 560 to 690 nm (Fig. A1). The absorption band of copper complex in the visible region is attributed to the d-d transition of Cu(II) ion. Therefore, the absorption band is sensitive to the coordination geometry around the central Cu(II) ion [40]. Accordingly, this color change indicates that the coordination state of Cu(II) ion is changed by mixing and coordinating the mixed amine ligand, alkylamine, imidazole, and pyridine derivatives.
Illustration of the ligand used for the preparation of L-CuFs. (a) HexAm, (b) DodAm, (c) 1Proim, (d) 1iProim, (e) 2iProim, (f) 2Benzpy, (g) 2Metpy, and (h) 4tBupy.
Photograph of the (a) CuF, (b) HexAm-CuF (c) DodAm-CuF, (d) 1Proim-CuF, (e) 1iProim-CuF, (f) 2iProim-CuF, (g) 2Benzpy-CuF, (h) 2Metpy-CuF, and (i) 4tBupy-CuF.
The XRD patterns of the CuF·4H2O and L-CuFs are shown in Fig. 3. The diffraction patterns of the L-CuF were distinct from that of the CuF·4H2O. Additionally, the no diffraction peaks assigned as CuF·4H2O are observed in any L-CuF diffraction patterns. This indicates that the obtained L-CuFs were completely converted to a different crystal structure.
XRD patterns of (a) CuF, (b) HexAm-CuF (c) DodAm-CuF, (d) 1Proim-CuF, (e) 1iProim-CuF, (f) 2iProim-CuF, (g) 2Benzpy-CuF, (h) 2Metpy-CuF, and (i) 4tBupy-CuF.
The differences (Δ) between the wavenumbers of the symmetric (νs) and antisymmetric (νas) vibration mode peaks in the infrared spectra are attributed to the binding geometry of the carboxylate bond for metal containing species. The review by Deacon and Philips and Nakamoto’s book provide the guideline to determine the binding geometry of carboxylate bond for metal ion. Here, to study the binding geometry of CuF and L-CuF, FT-IR spectra were measured [41, 42].
Figure A2 shows the FTIR spectrum of CuF·4H2O. The Δ value was 195 cm−1, which is close to the bridging geometry (140 cm−1 < Δ < 190 cm−1). The single crystal structure of CuF·4H2O with the infinite coordination network shows that the formate ion in CuF·4H2O binds to Cu ions forming a bridging geometry [27]. Therefore, the binding geometry of the formate ion in CuF·4H2O is evaluated by FTIR spectrum. The FTIR spectra of L-CuFs are shown in Fig. 4(a)–(h). The peak positions of νas and νs of COO− were assigned with reference to the DFT calculation result of L-CuFs, respectively. The peak positions of νas and νs, as well as the value of the Δ values, are summarized in Table 1. The Δ value of the L-CuFs complexes were ranging from 224 cm−1 to 310 cm−1 for monodentate geometry (200 cm−1 ≦ Δ). In addition, two formate ions are coordinated to one copper ion by forming monodentate geometry in the single crystal structure of amine coordinated copper formate complexes such as 4tBuPy-CuF [37]. Accordingly, the FTIR spectra reveal that the infinite coordination network of CuF·4H2O was broken by the coordination of amine ligands, leading to the formation of mononuclear copper formate-amine complexes.
Experimental (found) and calculated (calc) FTIR spectra of (a) HexAm-CuF, (b) DodAm-CuF, (c) 1Proim-CuF, (d) 1iProim-CuF, (e) 2iProim-CuF, (f) 2Benzpy-CuF, (g) 4Metpy-CuF, and (h) 4tBupy-CuF.
The thermal decomposition temperatures of the L-CuF complexes were measured using TG-MS system. TG curves are shown in Fig. A3. The TG-MS spectra (m/z = 44) of the CuF complexes are shown in Fig. 5(a)–(d), respectively. Metallic copper is formed via thermal decomposition of CuF moiety according to the following equation,
\begin{equation*} \textit{Cu}(\textit{HCOO})_{2} \to \textit{Cu}_{(s)} + H_{2(g)} + 2\textit{CO}_{2(g)} \end{equation*} |
Therefore, CO2 gas release is important event for evaluation of thermal decomposition temperature of L-CuF complexes. As previously reported [43], the weight loss of CuF·4H2O was caused by the removal of H2O molecules at approximately 100°C and the release of CO2 gas generated by the decomposition of formate ion at approximately 190°C (Fig. A3(a) and Fig. 5(a)).
TGMS spectra of CuF (a), alkylamine coordinated CuF (b), pyridine group coordinated CuF (c), and imidazole group coordinated CuF (d).
In the case of L-CuFs, the CO2 gas was released first weight loss by thermal decomposition of copper formate moiety decreased lower than that of CuF·4H2O. For the primary amine series of L-CuF (Fig. 5(b)), the decomposition temperatures were decreased to 120°C for HexAm-CuF and 123°C for DodAm-CuF, respectively. In the series of imidazole group fused CuF complexes (Fig. 4(c)), the decomposition temperature of 2iProim-CuF (134°C) is slightly higher than that of 1Proim-CuF (108°C) and 1isoProim-CuF (107°C). Figure 5(d) shows the TG curve and mass spectra of the pyridine group coordinated CuF complex. The decomposition temperature of 4tBupy-CuF (98°C) and 4Metpy-CuF (104°C) was lower than that of alkylamine coordinated CuF. However, the decomposition temperature of 2Benzpy-CuF (120°C) was nearly identical to that of alylamine coordinated CuF.
In recent studies, the decomposition temperature of L-CuF was determined based on the coordinated functional group, which altered the number of hydrogen bonds between the amine ligand and the formate ion [30, 31]. In contrast, our study shows that the coordinated functional group cannot determine the decomposition temperature of the L-CuF. This suggests that the change in the decomposition temperature of the CuF moiety is due to not only coordination functional group but also due to other factors.
As shown in the TGMS results, the thermal decomposition of L-CuF corresponds to the decomposition of formic acid ions in L-CuF. As demonstrated above, Δ, which is the difference between νas and νs in the FTIR spectrum, can be utilized to estimate the coordination geometry, at the same time, it is also possible to estimate structural information such as bond length and angle of carboxylic acid. Therefore, we focused on the FTIR spectrum and investigated the relationship between Δ and the decomposition temperature. Figure 6 shows that the plot of the Δ vs. the decomposition temperature of L-CuFs evaluated by TG and GCMS. As seen in Fig. 6, the decomposition temperature dropped as the value of Δ increased. Correspondingly, the greater the Δ value indicates the shorter C-O bond length, stronger M-O bond, and smaller O-C-O angle [43]. This change in the structure of the formate ion in L-CuF directly determines the decomposition temperature, indicating the stability of the C-H bond in the formate ion.
The plots of the value of thermal decomposition temperature of L-CuFs and deference of the differences between the wavenumber of symmetric and antisymmetric vibration mode peaks in infrared spectra.
In this paper, we demonstrated that the thermal decomposition temperatures of the CuF moiety of L-CuFs can be estimated by analyzing the positions of peaks corresponding to COO− νas and νs stretching in the FTIR spectra. The decomposition temperature of the CuF moiety is an important parameter for controlling the performance of MOD ink. Prior to this study, the species of functional group coordinated to the copper ion of the amine ligand was regarded as one of the key to predict the decomposition temperature of L-CuF. Therefore, due to lack of the suitable theory and method to predict the decomposition temperature of CuF moiety, it is difficult to presume the decomposition temperature of L-CuF with the same functional group. In this paper, we prepared L-CuF coordinated with the same functional group, such as pyridine, imidazole, and primary amine, which had different decomposition temperatures. Specially, the decomposition temperature of L-CuF was linearly decreased with increasing value of Δ, which is the difference between νas and νs in the FTIR spectra. This corresponds to the structural change of formate ion. The FTIR technique entails short measurement time and easy sample fabrication. In addition, the structure of the formate ion in the mixture of mixied copper complex, solvents, and additives, such as binder which are component of conductive ink can be measured by FTIR technique. Therefore, FTIR technique will be powerful tool to expect the order of decomposition temperature of CuF-based MOD inks.
This work was supported by the Grant-in-Aid for Scientific Research (B) (No. 20H02499) of the Japan Society for the Promotion of Science (JSPS), Hosokawa Powder Technology Foundation (Grant Number HPTF21115), NIMS Joint Hub Program, and the Cooperative research program of “Network Joint Research Center for Materials and Devices”.
Cu(II) formate tetrahydrate (CuF·4H2O, 1.47 g, 6.5 mmol) and hexylamine (HexAm, 1.02 g, 13 mmol) was added in 30 mL of acetonitrile. Then the mixture was stirred at an ambient condition for 30 min. The mixture was changed from pale blue colored suspension to deep blue colored solution. The mixture was filtrated to remove unreacted CuF. The filtered solution was cooled on an ice bath to obtain blue colored crystal. Obtained purple colored crystal was filtrated and washed with cooled acetonitrile. The elemental analysis of HexAm-CuF show the following result: calcd weight percent (mass%) for C14H32N2O4Cu·0.8H2O, C 45.29, H 8.77, N 7.55; found mass% C 45.29, H 9.06, N 7.53.
Synthesis of Dodecylamine-Cu FormateCuF·4H2O (1.47 g, 6.5 mmol) and dodecylamine (DodAm, 2.24 g, 12 mmol) was added in the mixed solvent of 15 mL of acetonitrile, 5 mL of hexane, and 3 mL of toluene. Then the mixture was stirred at an ambient condition for 30 min. The mixture was changed from pale blue colored suspension to light blue colored solution. The mixture was filtrated to remove unreacted CuF. The filtered solution was cooled on an ice bath to obtain light blue colored crystal. Obtained purple colored crystal was filtrated and washed with cooled acetonitrile. The elemental analysis of DodAm-CuF show the following result: calcd mass% for C26H56N2O4Cu·0.9H2O, C 57.63, H 10.99, N 5.21; found mass% C 57.69, H 10.71, N 5.17.
Synthesis of 4-tert-Butylpyridine-Cu FormateCuF·4H2O (2.26 g, 5 mmol) and 4-tert-butylpyridine (4tBupy, 2.44 g, 9 mmol) was added in 50 mL of acetonitrile. Then the mixture was stirred at an ambient condition for 30 min. The mixture was changed from pale blue colored suspension to deep blue colored solution. The mixture was filtrated to remove unreacted CuF. The filtered solution was cooled on an ice bath to obtain purple colored crystal. Obtained purple colored crystal was filtrated and washed with acetonitrile. The elemental analysis of 4tBuPy-CuF show the following result: calcd mass% for C20H28N2O4Cu, C 56.53, H 6.70, N 6.69; found mass%, C 56.54, H 6.60, N 6.58.
Synthesis of 2-Benzylpyridine-Cu FormateCuF·4H2O (1.0 g, 4.5 mmol) and 2-benzylpyridine (2Benzpy, 2.18 g, 12.6 mmol) was added in 25 mL of hexane. Then the mixture was stirred on an ice plate maintained at 0°C for 1 day. The mixture was gradually changed from pale blue colored suspension to purple colored suspension. Purple colored crystal was filtrated and washed with acetonitrile. The elemental analysis of 2BnzPy-CuF show the following result: calcd mass% for C26H24N2O4Cu·4.5H2O mass%, C 54.40, H 5.73, N 5.21; found mass% C 57.70, H 4.49, N 4.75.
Synthesis of 4-Methoxypyrydine-Cu FormateAt first the mixture of 4-Methoxypyridine (MetPy, 1.85 g, 17 mmol) and 25 mL of acetonitrile was cooled on an ice bath for 30 min. Then, CuF·4H2O (1.34 g, 6 mmol) was added into the mixture. Then the mixture was stirred on an ice plate maintained at −7°C for 60 min. After stirring, blue colored powder was obtained. The obtained powder was collected by filtrate and washed cooled acetonitrile. The elemental analysis of 4MetPy-CuF show the following result: calcd mass% for C14H16N2O6Cu·1.0H2O, C 43.02, H 4.07, N 7.15; found C 42.95, H 4.61, N 7.15.
Synthesis of 1-Isopropylimidazole-Cu FormateCuF·4H2O (CuF, 1.13 g, 5 mmol) and 1-isopropylimidazole (1iProim, 0.99 g, 9 mmol) was added in 25 mL of acetonitrile. Then the mixture was stirred at an ambient condition for 30 min. The mixture was changed from pale blue colored suspension to deep blue colored solution. The mixture was filtrated to remove unreacted CuF. Acetonitrile was removed under reduced pressure by using rotary evaporator. After removing the acetonitrile, 1iProim fused CuF was collected as pale blue powder. The collected powder was washed by toluene. The elemental analysis of 1iProim-CuF show the following result: calcd mass% for C14H22N4O4Cu·0.6H2O, C 43.60, H 6.02, N 14.50; found C 43.42, H 5.78, N 14.39.
Synthesis of 2-Isopropylimidazole-Cu FormateCuF·4H2O (1.13 g, 5 mmol) and 2-isopropylimidazole (2iProim, 0.99 g, 9 mmol) was added in 25 mL of methanol. Then the mixture was stirred at an ambient condition for 30 min. The mixture was changed from pale blue colored suspension to deep blue colored solution. The mixture was filtrated to remove unreacted CuF. Methanol was removed under reduced pressure by using rotary evaporator. After removing the methanol, 2iProim-CuF was collected as deep blue colored oil. The collected oil was dissolved in toluene to crystalize 2iProim-CuF for 3 days. Obtained crystal was filtrated and washed with toluene. The elemental analysis of 2iProim-CuF show the following result: calcd mass% for C14H22N4O4Cu·0.8H2O, C 43.19, H 6.07, N 14.37; found C 43.24, H 5.80, N 14.38.
Diffuse reflectance UV-vis spectra of CuF (a), Alkylamine coordinated CuFs (b), Pyridine coordinated CuFs (c), Imidazole coordinated CuFs (d).
FTIR spectra of CuF.
TG curve of CuF (a), alkylamine-CuF (b), pyridine group coordinated CuF (c), and imidazole group coordinated CuF (d).