1
Electroreduction of tantalum fluoride in a room temperature ionic liquid at variable temperatures

2
The present paper deals with the electroreduction of TaF5 in the room temperature ionic liquid 1-butyl-1-methyl-pyrrolidinium bis(tri-fluoromethylsulfonyl)imide ([BMP]Tf2N) at different temperatures for the sake of electrodeposition of tantalum.

3
The study was carried out using cyclic voltammetry and chronoamperometry measurements complemented by SEM-EDAX and XRD investigations.

4
In situ scanning tunneling microscopy and IU tunneling spectroscopy were also utilized for characterization of the electrodeposits.

5
The results show that, in addition to the formation of insoluble compounds, Ta can be electrodeposited in the ionic liquid ([BMP]Tf2N) containing 0.5 M TaF5 at 200 °C on polycrystalline Pt and Au(111) electrodes.

6
By addition of LiF to the electrolyte, the quality and the adherence of the electrodeposit were found to be improved.

7
An in situIU tunneling spectrum with about 300 nm thickness of the electrodeposit shows metallic behaviour indicating the formation of elemental tantalum.

8
Moreover, the XRD patterns of the electrodeposit, obtained potentiostatically at −1.8 V (vs. Pt) in ([BMP]Tf2N) containing 0.25 M TaF5 and 0.25 M LiF on Pt electrode at 200 °C, show the characteristic patterns of crystalline tantalum.

Introduction

9
Tantalum has unique properties that make it useful for many applications, from electronics to mechanical and chemical systems.1

10
Its high melting point, ductility, toughness, and excellent corrosion resistance make it an attractive coating material for components exposed to high temperature, wear, and severe chemical environments.

11
Because of its good thermal stability at elevated temperatures, thin layers of tantalum are applied as diffusion barriers on silicon in ultralarge-scale integrated (ULSI) circuits to prevent copper atoms from diffusion into dielectrics or silicon substrate.2

12
In addition, as tantalum is almost completely resistant to body fluids and non-irritating for human tissue, it has been widely used for making appliance and implants.

13
The corrosion resistance of tantalum is attributed to a thin protective oxide film that forms spontaneously in air and that exhibits a high stability in most mineral acids, even in concentrated and hot ones, except in hydrofluoric acid or fuming sulfuric acid.3–5

14
Many efforts have been made to develop an electroplating process for the electrodeposition of Ta.

15
High temperature molten salts were found to be efficient baths for the electrodeposition of refractory metals.

16
Senderoff and Mellors have reported the first results on the electrodeposition of coherent coatings of Nb, Ta, Zr and Mo on steel from molten fluorides.6–12

17
For tantalum deposition, they have used the ternary eutectic mixture LiF–NaF–KF as a solvent and K2TaF7 as a source of Ta at temperature between 650 and 850 °C.7,8

18
Using the same eutectic mixture and K2TaF7 as a source of Ta, Dutra et al13. have reported that compact Ta deposits, free of dendrites, can be obtained using pulse currents.

19
Chamelot et al14,15. have shown the optimum conditions for tantalum electroplating in the electrolyte LiF–NaF–K2TaF7 at a temperature of 800 °C.

20
Lantelme et al16. have stated that Ta can be electrodeposited from NaCl–KCl–K2TaF7 melts at 720 °C.

21
They have found that the quality of the deposit improves on addition of NaF to the melt.

22
As noticed, the use of high temperature molten salts of alkali metal halides as electrolytes has been the only possible way to electrodeposit Ta.

23
However, these baths have many technical and economic problems, such as the loss in the current efficiency of the electrolysis process, due to the dissolution of metal after its deposition,17 and the expected corrosion problems at high temperatures.

24
To the best of our knowledge, until now no successful attempts have been made for Ta electrodeposition at room temperature or even at low temperature in ionic liquids.

25
Only one attempt to electrodeposit Ta at 100 °C in a mixture of TaCl5, LiF and 1-ethyl-3-methyl imidazolium chloride [(EMIm) Cl] has been reported.18

26
The authors claim that they could make Ta films of thickness up to 100 μm in the former bath with a composition of 30 mol% TaCl5, 10 mol% LiF and 60 mol% (EMIm) Cl.

27
It is worth noting that we could not electrodeposit tantalum using the same bath composition.

28
Therefore, it seemed of interest to study the electroreduction of TaF5 in the water and air stable ionic liquid 1-butyl-1-methyl pyrrolidinium bis(tri-fluoromethylsulfonyl)imide for the sake of Ta electrodeposition.

29
In comparison to the imidazolium based ionic liquids, depending on the substrate the cathodic limit for irreversible cation reduction is about 500 mV wider.

Experimental

30
The electroreduction of tantalum fluoride was studied in the water and air stable ionic liquid 1-butyl-1-methyl-pyrrolidinium bis(tri-fluoromethylsulfonyl)imide ([BMP]Tf2N) which was purchased from Merck KGaA (EMD).

31
The liquid was dried under vacuum for 12 h at a temperature of 100 °C to water contents below 3 ppm (by Karl–Fischer titration) and stored in an argon filled glove box with water and oxygen below 1 ppm (OMNI-LAB from Vacuum-Atmospheres).

32
In cyclic voltammograms on platinum there is no evidence for electrochemically active water.

33
TaF5 (Alfa, 99,99%) and LiF (Alfa, 98,5%) were used without further purification.

34
The cyclic voltammetry measurements were performed in the glove box using a VersaStat™ II Potentiostat/Galvanostat (Princeton Applied Research) controlled by PowerCV and PowerStep software.

35
Gold substrates of Arrandee (gold films of 200–300 nm thickness deposited on chromium-covered borosilicate glass), Au(111) on mica purchased from Molecular Imaging and platinum sheet of thickness 0.5 mm (Alfa, 99,99%) were used as working electrodes, respectively.

36
Directly before use, the gold substrates were very carefully heated in a hydrogen flame to red glow, Pt-substrates were cleaned for 10 min in an ultrasonic bath in acetone, then heated in a hydrogen flame to red for a few minutes.

37
Pt-wires (Alfa, 99,99%) were used as reference and counter electrodes.

38
For the measurements at room temperature, the electrochemical cell was made of polytetrafluoroethylene (Teflon) and clamped over a Teflon covered Viton o-ring onto the substrate, thus yielding a geometric surface area of 0.3 cm2.

39
At elevated temperature, a quartz round flask was used as an electrochemical cell.

40
There is no evidence that the quartz flask is etched by the fluoride of TaF5.

41
Prior to use, all parts in contact with the solution were thoroughly cleaned in a mixture of 50/50 vol% H2SO4/H2O2 followed by refluxing in pyrogene free water ((aqua destillata ad iniectabilia).

42
The STM experiments were performed with in-house built STM heads and scanners under inert gas conditions (H2O and O2 < 1 ppm) with a Molecular Imaging Pico Scan 2500 STM controller in feedback mode.

43
The STM experiments were performed in an air conditioned laboratory with ΔT < ±1 °C.

44
Usually the approach is done overnight so that during the STM measurement a thermal equilibrium is obtained giving rise to a minimum thermal drift.

45
STM tips were prepared by electrochemical etching of tungsten wires (0.25 mm diameter) and electrophoretically coated with an electropaint (BASF ZQ 84-3225 0201).

46
During the STM experiments the electrode potential was controlled by the PicoStat from Molecular Imaging.

47
For the current/voltage tunnelling spectroscopy the tip was positioned on the site of interest and the tip voltage was scanned between an upper and a lower limit.

48
During this procedure the feedback is switched off.

49
A high resolution field emission scanning electron microscope (Carl Zeiss DSM 982 Gemini) was utilized to investigate the surface morphology of the deposited film and energy dispersive X-ray analysis was used to determine the film composition.

50
The XRD diffractograms of the deposited aluminium were acquired by Siemens D-5000 diffractometer with Co Kα radiation.

Results and discussions

Electrodeposition of Ta on polycrystalline platinum

51
Fig. 1 shows the cyclic voltammogram of 1-butyl-1-methyl-pyrrolidinium bis(tri-fluoromethylsulfonyl)imide ([BMP]Tf2N) ionic liquid containing 0.5 M TaF5 on platinum electrode at room temperature.

52
The electrode potential was scanned from the open circuit potential, −0.2 V vs. Pt, in the negative direction at a rate of 10 mV s−1.

53
As seen, the cyclic voltammogram exhibits only one cathodic process in the forward scan at −1.6 V vs. Pt with a current peak of ca. −2.1 mA at −2.2 V. The reduction peak may be attributed to the electrodeposition of Ta or to other insoluble tantalum compounds at the electrode surface.

54
At such electrode potential, the formation of a black deposit on the electrode surface is clearly visible.

55
The peak current of the reduction peak shows a square root dependence on the scan rate revealing that the reduction process is mainly controlled by diffusion of the electroactive species to the electrode surface.

56
At a potential of about −2.4 V, the reduction of the organic cation of the ionic liquid starts to occur.

57
In the reverse scan, the cathodic current continues to flow and crosses the forward scan at negative currents.

58
This current loop is typical for nucleation processes.

59
The anodic scan crosses the voltage axis at a potential of about −1.3 V producing a cathodic overpotential of about −0.3 V for the deposition process.

60
The anodic peak observed on the anodic scan at about 0.25 V is attributed to the incomplete stripping of the electrodeposit.

61
This can be confirmed by the almost complete removing of the black deposit.

62
However, the ratio of anodic to cathodic charge is lower than one, revealing some irreversibility of this system.

63
Stripping seems to be kinetically hindered, which is a common phenomenon in air and water stable ionic liquids.

64
The potential was set at −2.3 V for 1 h to form a thick layer of the electrodeposit on the electrode surface.

65
Visually, the deposit is black and appears to be thick and less adhering to the surface since it can easily be removed by washing with acetone.

66
As we assumed some kinetic hindrance in the formation of elemental tantalum, we performed the electrodeposition of Ta in the same ionic liquid at different temperatures, 100, 150 and 200 °C.

67
We would like to mention here that air and water stable ionic liquids are well suited to variable temperature studies.

68
In our experience, ([BMP]Tf2N) ionic liquid remains stable up to at least 300 °C under inert gas conditions.

69
Fig. 2 shows the effect of increasing the temperature on the cyclic voltammograms of ([BMP]Tf2N) containing 0.5 M TaF5 on platinum electrode at a scan rate of 10 mV s−1.

70
As can be seen the cyclic voltammograms exhibit a somewhat different behaviour compared with the cyclic voltammogram at room temperature.

71
Here, the cyclic voltammograms show two reduction peaks on the cathodic branch and two oxidation peaks on the anodic branch of the cyclic voltammograms.

72
The first cathodic peak is assumed to be correlated to the electrolytic reduction of Ta(v) to Ta(iii), since TaF3 is known to be a stable compound.

73
The second reduction peak might be attributed to the reduction of Ta(iii) to Ta(0), as a black deposit is clearly seen even by the naked eye.

74
In the backward scan, the first anodic peak might be attributed to the oxidation of Ta to Ta(iii) and the second peak is correlated to oxidation Ta(iii) to Ta(v).

75
Here it is also evident that the reoxidation of the deposit is not fully reversible.

76
The peak currents of both the reduction and oxidation peaks also depend linearly on the square root of the scan rate, indicating diffusion control.

77
As shown in Fig. 2, the peak potentials of the cathodic peaks slightly shift to less negative values and the potential of the oxidation peaks slightly move to the positive direction with increase in temperature.

78
The peak currents of the reduction and oxidation peaks remarkably increase with rising temperature.

79
This is due to the increased mobility of the electroactive species towards the electrode surface which, in turn, leads to increasing the reaction rate, either of the reduction or oxidation.

80
Both conductivity and viscosity of such types of ionic liquids are strongly dependent on temperature.

81
The mechanical quality and adherence of the deposit were found to be superior to room temperature deposits.

82
At 200 °C, the deposit appears to be dendritic and it is better adherent than those which were obtained at 100 or 150 °C.

83
Nevertheless, the deposits can be removed by ultrasonic cleaning with acetone in all cases.

84
The SEM micrograph of such a deposit made at 200 °C at −1.3 V for 1 h is presented in Fig. 3a.

85
As seen, the electrodeposit contains mainly granules and was analysed as a tantalum species containing C, O, F and S as revealed from the accompanying EDAX profile, Fig. 3b.

86
A typical ratio of the surface concentration (wt.%) of Ta/F is 4.2/1, as determined by EDAX analysis.

87
The presence of C, O, F and S is due to the remaining ionic liquid at the electrode surface.

88
Nevertheless, this layer can be removed by washing with isopropanol followed by boiling in water.

89
After such treatment only crystalline and elemental Ta can be detected at the electrode surface using XRD, EDAX and AES.

90
Further studies are currently in progress and will be published in a future paper.

91
In order to know whether or not the electrodeposit contains crystalline tantalum, X-ray diffraction patterns (XRD) were acquired.

92
Surprisingly, the XRD patterns of the initial electrodeposits showed only the characteristic patterns of the Pt substrate, and there was no evidence for crystalline tantalum.

93
This means that the electrodeposits are either amorphous or so small in their particle size that within the resolution of our device no XRD patterns could be obtained.

94
In order to improve the crystallinity of the electrodeposits, the sample was annealed at 800 °C under vacuum (10−4 mbar) for 5 h, then the sample was reinvestigated by XRD.

95
Now XRD patterns show clearly the characteristic signals of crystalline Ta (JCPDS 25-1280, 19-1290 and 04-0788) and Ta2O5 (JCPDS 21-1198 and 19-1298), Fig. 4.

96
These results confirm that the electrodeposits contain crystalline Ta.

97
It is not too surprising that Ta2O5 forms, as at such vacuum there is enough trace oxygen that can oxidize our tantalum.

98
Furthermore, an EDAX analysis showed that there remained no fluoride in the deposit after thermal annealing.

99
This might indicate that the low valent Ta–F compounds have a high vapour pressure.

100
The results presented here imply that first a thin crystalline tantalum layer is deposited (about 200–300 nm) on top of which a non-stoichiometric tantalum subfluoride layer with some trapped ionic liquid grows with thicknesses of several micrometers.

Electrodeposition of Ta on Au(111)

101
Fig. 5 shows the cyclic voltammogram of ([BMP]Tf2N) containing 0.5 M TaF5 on Au(111) at room temperature.

102
As shown, two reduction processes are recorded in the forward scan.

103
The first one starts at a potential of −0.5 V with a peak at −0.75 V, it might again be correlated to the electrolytic reduction of Ta(v) to Ta(iii).

104
The second process starts at a potential of −1.5 V and is accompanied by the formation of a black deposit on the electrode surface.

105
This can be attributed to the reduction of Ta(iii) to Ta metal simultaneously with the formation of insoluble tantalum compounds on the electrode surface.

106
The anodic peak recorded on the backward scan is due to the dissolution of the electrodeposit which, however, does not seem to be complete.

107
Then the anodic current increases as a result of gold dissolution at E > 1.5 V. Unlike the behaviour of this system on polycrystalline Pt at room temperature, the electrochemical reduction of Ta(v) obviously occurs in two steps on Au(111).

108
However, as in the case of Pt, the deposit obtained on gold is also less adherent and can easily be removed by washing with acetone.

109
Furthermore, the bulk phase of this deposit does not show XRD patterns.

110
The cyclic voltammogram of ([BMP]Tf2N) containing 0.5 M TaF5 on gold substrate at a temperature of 200 °C is shown in Fig. 6.

111
The electrochemical reduction of Ta(v) also occurs in two steps as revealed by the presence of two cathodic peaks in the forward scan.

112
The first cathodic process starts at about −0.5 V which is presumably the result of the reduction of Ta(v) to Ta(iii) and the second process starts at a potential of about −1.1 V, accompanied by the deposition of a black layer on the electrode surface.

113
In the reverse scan, a shoulder was recorded at a potential of about −0.1 V which presumably represents the first oxidation step of Ta to Ta(iii).

114
Then, at about 0.3 V, dissolution of the electrodeposit sets in.

115
The large increase in anodic current at a potential of 1 V is due to electrochemical dissolution of gold.

116
It is worthy of mentioning that the mechanical quality and the adherence of the electrodeposits improve at 200 °C.

117
The SEM micrograph of Fig. 7a shows the surface morphology of a thick layer of the electrodeposit prepared potentiostatically at a potential of −1.2 V vs. Pt for 1 h at 200 °C.

118
The deposit appears to be thick and dense and contains a few cracks.

119
It is a common observation that during the electrodeposition process, internal or residual stress can occur causing fine cracks in the deposit.

120
We tried to remove the deposited layer (by ultrasonic cleaning in acetone) and to investigate the bare surface of the substrate.

121
Our aim was to investigate whether or not the surface is covered by a thin and compact film of Ta, before the formation of the thick deposit started.

122
Fig. 7b shows a SEM micrograph of the gold substrate after removing the thick deposit using ultrasonic treatment in acetone until the bare surface of gold is recovered.

123
As seen, there are many remaining islands of Ta which are tightly bound to the surface.

124
The deposit contains fine crystallites with sizes well below 100 nm, as shown in the SEM micrograph of Fig. 7c which represents the area shown in the white square in Fig. 7b.

125
EDAX profile taken within the marked area shown in the micrograph of Fig. 7b reveals the presence of Ta on the surface and very small amount of N and C which might originate from some trapped ionic liquid, Fig. 7d.

126
It is interesting to mention that within the resolution of the EDAX detector we did not find any hint for fluoride.

127
In the light of the aforementioned results, we may conclude that tantalum can be electrodeposited from ([BMP]Tf2N) containing TaF5.

128
However, the quality of thick layers of the electrodeposit needs to be improved since in addition to the metal deposition, also insoluble tantalum compounds might be formed owing to side reactions.

129
This is maybe the reason for the bad adherence of a thick layer of the deposit because it does not solely contain metallic Ta.

In situ STM measurements

130
In situ STM measurements under potentiostatic conditions can give valuable information on the electrodeposition of Ta in the employed ionic liquid ([BMP]Tf2N).

131
The STM picture of Fig. 8a shows a typical surface of gold on mica substrate (Au(111)) in the ionic liquid ([BMP]Tf2N) containing 0.5 M TaF5 at open circuit potential.

132
As seen, the surface is characterized by terraces with average step heights of about 250 pm, typical for Au(111).

133
By applying a potential of −1.25 V (vs. Pt) (see Fig. 5), the nature of the surface changes as seen in the STM picture of Fig. 8b.

134
A rough layer of Ta is formed rapidly and some triangularly shaped islands with heights of several nanometers grow above the deposited layer (Fig. 8b).

135
With ongoing time, these islands grow vertically and laterally and finally merge together to a thick layer.

136
The 3-D STM picture of Fig. 9a shows the topography of the electrodeposit, with a thickness of about 300 nm.

137
The thickness of the deposit was determined in situ from the z-position of piezo, which is a standard procedure in our laboratories.

138
In order to investigate if the in situ deposit is metallic or not, current/voltage tunneling spectroscopy was conducted.

139
It was already shown by us that the IU tunneling spectroscopy is a valuable technique for in situ characterization of electrodeposited semiconductors19–21 and metals.

140
We could show with the in situIU tunnelling spectroscopy that germanium with layer thicknesses of 20 nm and more is semiconducting with a symmetric band gap of 0.7 ± 0.1 eV.19,20

141
On the other hand, we have found that very thin layers of germanium with thicknesses of several monolayes exhibit clearly metallic behaviour.19,20

142
In the case of silicon, we could show for the first time that in an ionic liquid elemental silicon can be obtained which has at room temperature a band gap of 1.0 ± 0.2 eV for a layer of 100 nm thickness.21

143
A typical in situ tunneling spectrum of a 300 nm thick layer of the electrodeposit at different positions is shown in Fig. 9b.

144
As seen, the IU spectrum exhibits metallic behaviour with an exponential-like rise of the current (see ref. 22) indicating that the electrodeposited layer seems to be elemental Ta.

145
Together with the ex situ measurements we can conclude that the initial reduction of TaF5 in ([BMP]Tf2N) leads to a very thin layer of metallic tantalum.

146
However, detailed STM and tunneling spectroscopy studies are required and are currently running in our laboratories.

Effect of addition of LiF

147
We would like to report on some experiments, where we added LiF to the above mentioned ionic liquid containing TaF5.

148
From literature data in high temperature molten salts, it was found that addition of fluorides of alkali metals, such as LiF or NaF, to the fused salts facilitates the electrodeposition of tantalum.13–16

149
Therefore, it seemed of interest to examine the effect of adding LiF to the employed ionic liquid containing TaF5 on the electrodeposition of Ta.

150
We found that the mechanical quality and the adherence of the electrodeposited Ta in ([BMP]Tf2N) containing TaF5 can be improved by addition of LiF to the electrolyte.

151
Fig. 10 shows the cyclic voltammogram of ([BMP]Tf2N) containing 0.25 M TaF5 and 0.25 M LiF on Pt electrode at 200 °C.

152
Here, three cathodic peaks are recorded in the forward scan before reduction of the organic cation of the ionic liquid sets in at E = −2.2 V. This observation indicates that the reduction of Ta(v) to Ta metal obviously occurs in at least three steps and not in two steps as in the case of LiF absence.

153
The SEM micrograph of such a deposit (Fig. 11a) made at −1.8 V for 1 h at 200 °C shows a smooth, coherent and dense layer.

154
The electrodeposit was analysed as metallic tantalum, as revealed from the corresponding XRD patterns (Fig. 11b) which show four characteristic peaks of crystalline Ta (JCPDS 25-1280, 19-1290 and 04-0788).

155
The peaks are pretty sharp though of low intensity compared to the peaks of the substrate.

156
However, for micrometer thick layers we still find varying amounts of fluoride in the deposit.

157
The role of LiF in the deposition process is not clear at the moment.

158
Maybe due to the ionic polarizability of Li+ the Ta–F bonds are weakened leading to a facilitation of Ta deposition.

159
Further work is now under progress in our laboratory to shed more light on the influence of addition of lithium salts.

160
Moreover, detailed in situ STM experiments are currently performed.

Conclusions

161
We have presented the first results on the electroreduction of TaF5 in the ionic liquid 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide ([BMP]Tf2N) at different temperatures with the aim of tantalum electrodeposition.

162
It was found that the electrodeposition of tantalum at room temperature is not straightforward.

163
However, at 200 °C Ta can be electrodeposited in addition to the formation of insoluble tantalum compounds on the electrode surface.

164
The quality and the adherence of the electrodeposit were found to be improved upon addition of LiF to the electrolyte.

165
XRD patterns of the electrodeposit, obtained potentiostatically at −1.8 V in ([BMP]Tf2N) containing 0.25 M TaF5 and 0.25 M LiF on Pt electrode at 200 °C, clearly show the characteristic patterns of crystalline tantalum.

166
Furthermore, in situIU tunneling spectra of about 300 nm thick layers show metallic behaviour indicating the formation of elemental tantalum.

167
Our studies show that the electrodeposition of Ta in the ionic liquid ([BMP]Tf2N) at low temperature is possible.

168
However, to obtain bulk tantalum layers further work is needed.