Z. Phys. Chem. 218 (2004) 255–283 by Oldenbourg Wissenschaftsverlag, München
Ionic Liquids: Promising Solvents for Electrochemistry By Frank Endres ∗ Technische Universität Clausthal, Abteilung für Extraktive Metallurgie und Elektrochemie, Robert-Koch-Str. 42, D-38678 Clausthal-Zellerfeld (Received November 4, 2003; accepted November 27, 2003)
Ionic Liquids / Scanning Tunneling Microscopy / Semiconductors / Electrochemistry / Nanocrystals Ionic liquids are solvents that are solely composed of ions. By definition their melting points are below 100 ◦ C. Typical cations are substituted imidazolium ions, like 1-butyl3-methylimidazolium, or tetraalkylammonium ions, like e.g. trioctyl-methyl-ammonium. Some important anions are hexafluorophosphate, trifluoromethylsulfonate, bis(trifluoromethylsulfonyl)imide. Many ionic liquids have negligible vapour pressures even at temperatures of 300 ◦ C and more, they can have viscosities similar to water, ionic conductivities of up to 0.1 (Ω cm) −1 , and, which makes them interesting for electrochemistry, wide electrochemical windows of more than 6 Volt. In this review article recent results of the author are summarized. It is shown that with the scanning tunneling microscope the processes during phase formation can be probed in situ with high quality. An important result is that semiconductors, shown at the example of germanium, can be made electrochemically on the nanoscale and that the electronic properties (band gap) can be measured in situ with current/voltage tunneling spectroscopy. Ionic liquids will gain a rising interest in electrochemistry as elements and compounds can be made electrochemically which are not accessible by conventional aqueous or organic electrochemistry.
1. Introduction The properties of the electrode/electrolyte interface are of basic interest for electrochemical processes. Classical electrochemical techniques like cyclic voltammetry (CV) or impedance spectroscopy (EIS) give an integral insight into the macroscopic processes [1]. And even spectroscopic techniques [2–12] mainly yield an integral information on electrochemical processes. The invention of the scanning tunneling microscope (STM) by Binnig and Rohrer in 1982 has revolutionized modern surface science [13]. It was the first tool * E-mail:
[email protected]
256
F. Endres
that allowed to investigate processes on an electronically conducting surface in real space with nanometer resolution. In the following 20 years the STM was applied in numerous investigations of surfaces and interfaces under different conditions. Among them variable temperature measurements under vacuum conditions [14] as well as at the interface electrode/electrolyte under electrochemical conditions have to be mentioned. In 1986 it was shown by Sonnenfeld and Hansma [15] as well as by Bard [16] that the STM is well suited for in situ investigations under electrochemical conditions. Since then numerous studies on the nanoscale electrochemical phase formation on a variety of substrates have been published and mainly aqueous solutions have been employed for those studies. This has surely to do with the fact that aqueous solutions are technically important, furthermore organic impurities, that strongly alter the quality of in situ STM experiments, can be easily removed. Nevertheless, a shortcoming of aqueous solutions is that the electrochemical phase formation is limited to potential ranges where water is neither reduced nor oxidized. Depending on the composition of the solution and depending on the substrate an electrochemical window of about 1.2 V is obtained. At the cathodic limit hydrogen evolves, at the anodic limit either oxygen forms or the material of the electrode is oxidized. Less noble elements like aluminium and magnesium, refractory metals like tantalum and niobium but also semiconductors like silicon, germanium, gallium arsenide or gallium antimonide can as a consequence not be electrodeposited in aqueous solutions. With respect to the electrochemical windows organic solvents are already an improvement, and, for example, there are numerous studies on the electrodeposition of aluminium from organic solvents ([17] and references therein). Apart from a problematic purity of the deposits because of decomposition products of the solutions and because most of the organic solvents have high vapor pressures at elevated temperatures, variable temperature experiments are difficult to perform. In this context molten salts are already an interesting alternative, because they combine wide electrochemical windows with low vapor pressures so that a variety of elements can be electrodeposited from them [18–21]. For about 20 years now molten salts that are liquid at room temperature have been of great interest in basic research [22, 23]. The first systems that were reported in literature were liquids based on mixtures of AlCl3 and N-butylpyridinium chloride or 1-ethyl-3-methylimidazolium chloride, and the Lewis acidity can be adjusted by variation of the composition. With a molar excess of AlCl3 the liquid is Lewis acidic, with a molar excess of the organic salt the liquid is Lewis basic. Most of the metal salts are more or less well soluble in both types of liquids. The electrochemical windows of these room temperature molten salts are dependent on the composition and they can vary between 2 and 4 Volt [24, 25]. Such AlCl3 based liquids are well suited to investigate the electrodeposition of many metals and alloys and much literature is devoted to Al-alloys that are interesting for corrosion protection and light-weight construction. An overview on the electrodeposition of metals and semiconductors
Ionic Liquids: Promising Solvents for Electrochemistry
257
from low melting ionic liquids was recently given by the author [26, 27]. In 1992 the concept of low melting ionic liquids was significantly extended by the synthesis of air and water stable ionic liquids based on different organic cations and tetrafluoroborate [28]. Later, systems were developed that are based on organic cations like those mentioned above and organic anions like e.g. bis(trifluoromethylsulfonyl)imide. Especially liquids based on the latter anion are stable against hydrolysis, hardly miscible with water and they can be heated, even under air, without decomposition to temperatures as high as 300 ◦ C [29]. The electrochemical windows of such air and water stable ionic liquids can reach values of more than 4 Volt on platinum and gold [29, 30], and therefore they are particularly suited to investigate the electrodeposition of elements like germanium, silicon, titanium and other ones. In this review a short introduction in the importance of ionic liquids for the (nanoscale) electrodeposition of less noble elements will be given.
2. Low melting ionic liquids Molten salts have interesting physical and chemical properties: many chemical substances can be dissolved in them and especially the high temperature molten salts have a high ionic conductivity of more than 1 (Ω cm) −1 [31]. This high ionic conductivity as well as wide electrochemical windows have lead to a great importance in industry, for example as solvents for electrowinning of metals and semiconductors [18]. The melting points of the different systems vary from about −50 ◦ C for modern synthetic ionic liquids to more than 1000 ◦ C for the classical inorganic oxides and halides. In general, systems with melting points of 200 ◦ C and more are regarded as high temperature molten salts [32]. These classical molten salts require demanding experimental conditions, and in most cases for reproducible measurements water and oxygen free conditions have to be fulfilled. Especially the choice of materials is often limited as at high temperatures strong corrosion can set in. For about 20 years now “room temperature molten salts” based on AlCl 3 have been extensively studied. Due to the low temperatures these liquids were discussed for a while as promising solvents for secondary batteries. For about 10 years now the interest focuses more and more on air and water stable liquids with the aim, to handle them even under ambient conditions. By definition an ionic liquid is a low melting ionic system with a melting point below 100 ◦ C, [33, 34]. Especially for Green Chemistry purposes ionic liquids have gained a rising interest during the last few years [35]. Although this temperature limit is arbitrary, it represents the fact, that meanwhile ionic liquids are employed for various chemical and electrochemical studies at low temperatures, and the number of publications in literature with the topic “ionic liquid” rises strongly. Furthermore, the expression “ionic liquid” shall prevent the association with high temperatures as in the case of molten salts. In the beginning, the author of this article was often asked, how an STM can work
258
F. Endres
in molten salts at high temperatures, although it was always pointed out, that these molten salts are liquid at room temperature. After having changed the expression to “ionic liquids” in later talks these questions were no longer asked. Meanwhile modern ionic liquids are widely employed in basic research in organic and technical chemistry, furthermore they are interesting for polymer synthesis and enzyme reactions [36]. Historically, ionic liquids, i.e. systems with melting points below 100 ◦ C, can be divided into 3 groups: a) systems based on AlCl3 and different organic salts like for example N-butylpyridinium chloride and 1,3-dialkylimidazolium halides [23–25] b) systems based on organic cations like for example the 1,3-dialkylimidazolium-ions and perfluorinated anions like BF 4 − , PF 6 − and SbF 6 − [28, 29, 37, 38] and c) systems based on organic cations like for example the 1,3-dialkylimidazolium-ions and different hydrolysis resistant ions like for example tosylate, bis-(trifluoromethylsulfonyl)imide and others [29, 39]. In the following sections some properties of these classes of ionic liquids will be summarized
2.1 AlCl 3 based ionic liquids The first ionic liquids based on aluminium halides were presented in 1948 by Hurley and Wier at the Rice Institute in Texas. Their aim was to use these liquids for the electroplating of aluminium [40]. The extreme hygroscopic properties of these liquids and the lack of inert gas systems might have been the reason that these liquids were almost forgotten for about 30 years. It was around 1980 when the groups of Osteryoung and Wilkes started systematic investigations with these liquids [22, 23]. Most of the studies in literature deal with a liquid based on 1-ethyl-3-methylimidazolium chloride and AlCl 3 . Fig. 1 shows the phase diagram of this system [41]: it is remarkable that in a wide range of compositions melting points at room temperature or below are obtained. Qualitatively similar phase diagrams are obtained for mixtures of AlCl 3 and other organic halides [42]. Fig. 2 shows the dependence of the mole fraction of the various known chloroaluminate complexes as a function of the molar ratio of AlCl 3 and therefore as a function of the Lewis acidity: the result is that in Lewis basic systems only chloride and tetrachloroaluminate exist, whereas in the Lewis acid regime polynuclear chloroaluminate complexes are found. The high Lewis acidity of these complexes is one reason for the excellent solubility of many metal halides in these liquids. Of importance for electrochemical investigations are the good ionic conductivity and the adjustable electrochemical windows of chloroaluminate ionic liquids. The most important system 1-ethyl-3-methylimidazolium
Ionic Liquids: Promising Solvents for Electrochemistry
259
Fig. 1. The phase diagram of 1-ethyl-3-methylimidazolium chloride and aluminium chloride shows a compound formation for equimolar mixtures. Defined eutectica in the Lewis basic or Lewis acid regime are not observed because of glassy transitions (see [39]).
Fig. 2. Molar ratio of the different anionic species in chloroaluminate-liquids. (1): Cl− , (2): AlCl4 − , (3): Al2 Cl7 − , (4): Al3 Cl10 − , (5): Al4 Cl13 − , (6): Al2 Cl6 (see [43]).
chloride/AlCl 3 has an ionic conductivity of 1.17 × 10−2 (Ω cm)−1 at 30 ◦ C (44 mol-% AlCl3 ), with 50 mol-% and 67 mol-% of AlCl3 values of 2.27× 10 −2 (Ω cm)−1 and 1.54 × 10−2 (Ω cm)−1 at room temperature were reported in literature [23]. These values are well below those of high temperature molten
260
F. Endres
Fig. 3. Simplified electrochemical windows of N-butylpyridinium chloride (BPC) and 1-ethyl-3-methylimidazolium chloride (EMIC) with AlCl3 on inert substrates like for example glassy carbon (see [24]).
salts but, nevertheless, they reach values of 0.1–0.3 molar aqueous KCl solutions [44]. At 100 ◦ C the mentioned ionic liquid has a specific conductivity of 7 × 10−2 (Ω cm)−1 for equimolar mixtures [41]. For the liquid made of 1-butyl3-methylimidazolium chloride and AlCl 3 the reported conductivities are about 30% lower [41]. The strongly simplified electrochemical windows of 2 well investigated systems on inert substrates are presented in Fig. 3. In Lewis acid systems chlorine forms from chloroaluminate complexes at the anodic limit, at the cathodic limit aluminium is electrodeposited. In basic systems chlorine from chloride forms at the anodic limit, at the cathodic limit a less defined oligomerization of the organic cations sets in. Equimolar mixtures can have electrochemical windows of up to 4 Volt. Practically neutral systems have to be buffered by NaCl or CaCl2 , as small variations in the solvent composition – for example during electrodeposition or electrooxidation of metals – can strongly alter the acidity [45]. For in situ STM measurements the viscosity of ionic liquids is of importance, as for a good resolution in the STM pictures the movement of the tip should not be damped too much. For equimolar 1,3-dialkylimidazolium chloride/AlCl 3 systems absolute viscosities between 2 × 10 −2 and 3 × 10−2 Pa s at 25 ◦ C were reported, in the Lewis basic regime they can rise to 10 × higher values [41]. These values for the viscosity are more than one order of magnitude higher than that of water at room temperature, but in the experience of the author there is no disadvantageous effect on the resolution in in situ STM experiments on the nanometer scale. The density of the 1-alkyl-3-methylimidazolium chloride/AlCl 3 liquids is almost linearly dependent on the length of the side chain, and the density
Ionic Liquids: Promising Solvents for Electrochemistry
261
decreases with increasing length. Equimolar mixtures of 1,3-dimethylimidazolium chloride and 1,3-dibutylimidazolium chloride, respectively, with AlCl 3 have at 60 ◦ C densities of 1.31 and 1.13 g/cm 3 [41]. It has to be mentioned that the organic salts and AlCl3 are strongly or even extremely hygroscopic. Especially AlCl 3 absorbs water rapidly and quantitatively, as a consequence HCl and different oxochloroaluminates form in the liquid [24]. Although there are physical and chemical strategies to remove these impurities [38], it is strongly recommended to prevent water as well as possible from the beginning. AlCl 3 should be of 99.999% quality, 1-butyl-3-methylimidazolium chloride should be purified by recrystallization up to 3 times followed by vacuum drying at elevated temperatures. AlCl 3 based ionic liquids are not suited for the electrodeposition of silicon, germanium, titanium and other less noble elements: in Lewis acid media at least a codeposition of Al sets in, in Lewis basic media kinetic and thermodynamic reasons involved in the complexation of the respective halides seem to prevent an electrodeposition.
2.2 Air and water stable ionic liquids Both basic and applied studies in chloroaluminate systems are limited by their extreme hygroscopic behavior. Therefore it was a big step forward when in 1992 the synthesis of the air and water stable ionic liquid 1-ethyl3-methylimidazolium tetrafluoroborate was reported [28]. In the beginning the purity of this interesting liquid was not yet satisfying. One problem with the AgBF 4 based synthesis was a remarkable and difficult to remove amount of the silver salt that remained in the liquid. Apart from that the use of AgBF 4 is rather cost intensive. Furthermore, this liquid shows a complete miscibility with water [48], and it is not straightforward to remove water because at higher temperatures a strong hydrolysis of BF 4 − sets in liberating HF. Water contents below 200 ppm can hardly be realized. 1-ethyl3-methylimidazolium hexafluorophosphate, which can be made comparably easily in an aqueous routine, is an improvement because it is much less soluble in water [47]. In 1996 the synthesis of 1-butyl-3-methylimidazolium hexafluorophosphate was first reported [37]. At room temperature it takes up a maximum of 2 weight% of water [48], furthermore it is relatively easy to dry this liquid at elevated temperatures and thus make it practically water free. Water contents below 10 ppm can be easily achieved, and thoroughly dried liquids do not give any hint for water in the cyclic voltammogram [30]. It can be stated that in recent years remarkable progress was made in the synthesis of air and water stable ionic liquids. There are many possible applications meanwhile and in the opinion of the author it is difficult to predict where ionic liquids will be employed in the future. Apart from media for the extraction of metal ions from aqueous or organic solutions [49] chiral ionic liquids, that can be employed in the asymmetric or-
262
F. Endres
ganic synthesis, have been reported [50]. In general the trend is to make chemically and thermally stable ionic liquids that can be handled under ambient conditions, and for these purposes the already mentioned anion bis(trifluoromethylsulfonyl)imide is very promising [29]. For an overview on the most important synthesis routes as well as the most important physical properties, see reference [38]. The electrochemical windows of these air and water stable ionic liquids are remarkably high. For 1-butyl-3-methylimidazolium hexafluorophosphate a value of 6.35 Volt for glassy carbon has been reported, for microcrystalline/polycrystalline gold a value of 5.95 Volt has been published [51]. For Au(111) a more realistic value of a little more than 4 Volt was measured by the author. If this value is compared to that one of water (ca. 1.2 Volt), it is evident that ionic liquids have a great potential for electrochemistry. The melting point of [BMIm]PF 6 is −61 ◦ C [37, 55], the density at 30 ◦ C is 1.37 g/cm 3 , the absolute viscosity is 0.3 Pa s [55], and for the specific ionic conductivity at 25 ◦ C and 22 ◦ C values of 1.46 × 10−3 (Ω cm)−1 [55] and 1.3 × 10−3 (Ω cm)−1 [56] were reported in literature. The ionic conductivity and the viscosity are one order of magnitude worse than those of Lewis acid AlCl3 based ionic liquids, with the consequence, that distinct Ohmic drops and a reduced electrochemical reversibility are observed. This will hardly alter electrochemical processes in the STM experiments since they are performed potentiostatically near equilibrium conditions in most cases. But, as an example, it shall be mentioned here that even in such apparently simple processes like the electrodeposition or electrooxidation of silver or copper, which are reversible in chloroaluminate liquids, irreversible phenomena and remarkable charge transfer resistances are observed [52, 53]. Nevertheless, these liquids are well suited for the elctrodeposition of intrinsic and semiconducting germanium nanocrystallites. Hitherto, such Ge nanocrystallites have only been reported to be made by molecular beam methods under ultra high vacuum conditions. In the opinion of the author, ionic liquids will presumably gain a great interest in electrochemical nanotechnology as also silicon and many compound semiconductors will be accessible by electrochemical means. Currently, the interest in the new air and water stable ionic liquids is, unfortunately, still focussed on studies in organic synthesis as well as in homogeneous catalysis. But these liquids are also pretty interesting for industrial applications as they are chemically extraordinarily stable, furthermore they have low vapor pressures so that organic synthesis could also be performed at higher temperatures, without the usual environmental problems caused by organic solvents. Furthermore ionic liquids can easily be purified, surplus educts can often be simply removed under vacuum or by extraction with water. For such applications the review articles of Wasserscheid [39], Welton [54] and Seddon [35] as well as a recent book [36] are recommended here.
Ionic Liquids: Promising Solvents for Electrochemistry
263
3. The scanning tunneling microscope (STM) [57] The scanning tunnelling microscope was first introduced in 1982 by Gerd Binnig and Heinrich Rohrer [12] who were then awarded in 1986 with the Nobel Prize for physics. The principle of operation is remarkably simple: a metallic sharp tip is approached to an electronically conductive substrate at a distance of about 1 nm and a voltage is applied between tip and sample. Due to the quantum mechanical tunneling effect electrons will “tunnel” from the tip to the surface or vice versa, depending on the polarity. This tunneling current is exponentially dependent on the tip/substrate distance and it can be used as a feedback signal to control precisely the distance between tip and sample. Height differences can be probed and directly represented in real space. Typical tunneling currents in the experiment are between 0.5 and 50 nA.
3.1 In situ STM measurements under electrochemical conditions When the STM was first employed under electrochemical conditions the following problem occurred: the STM tip acts as an electrode in the electrochemical cell and leak currents of several hundreds of nanoamperes were easily obtained. If it is taken into account that typical tunneling currents are only a few nanoamperes it is clear that an in situ STM imaging is not straightforward under such conditions. Therefore the STM tip has to be insulated by a paint or a varnish, with the exception of the very end of the tip. The electrochemically active surface is then reduced such that the leak currents are well below the tunneling currents. There are several approaches in literature: besides Apiezon wax [58], silicon-rubber [59] and nail varnish [60], coatings with epoxy resins [61, 62] and glasses [63] were suggested. For experiments in ionic liquids hitherto only an anodic electrophoretic paint from BASF has had sufficient chemical stability: leak currents of 50 pA and below can be obtained with a well insulated tip at room temperature.
3.2 Requirements for operation in ionic liquids AlCl3 based ionic liquids are extremely hygroscopic, as mentioned before. Therefore, for in situ STM measurements inert gas conditions with water and oxygen contents below 1 ppm are necessary which requires an inert gas glove box with a gas purification system. Even water that is physically adsorbed to the steel parts of the STM head is rapidly and quantitatively absorbed by that type of ionic liquids. The liberated HCl together with the oxygen traces leads to a remarkable corrosion. As commercial STM heads hardly resist such conditions, in house designed STM heads have been employed for the in situ STM measurements. These heads involve a parallel approach of the sample plate to the tip so that the coarse and fine approach are done automatically by the software, which allows that the whole experiment can be set up in the glove box.
264
F. Endres
These heads have several advantages over the commercial ones: (1) the construction is held as simple as possible so that all parts of the head can be easily cleaned and polished. (2) the STM-scanners are well insulated against the environment and well resistant sealing materials are employed. Where possible all parts are fixed by screws so that in case of damage a low-priced repair of the heads is possible. For some parts like e.g. the sample plate molybdenum was used as material because it does not show the normal corrosion, furthermore it can be easily cleaned by heating under vacuum. Furthermore, its thermal expansion coefficient is 4–5 × lower than that one of stainless steel which leads to a low thermal drift. (3) With three step motors at an angle of 120 ◦ C to each other three fine grade screws are driven so that a parallel approach results. (4) Improved STM preamplifiers were constructed so that basic noise is reduced to values below 50 pA, as a result current/voltage tunneling spectra can be acquired reproducibly in high quality. Furthermore, as a rule, all parts of the STM head, that can be heated under vacuum, were treated such in order to remove traces of water as well as possible. The assembled STM head was put in the glove box into a vacuum-tight vessel made of stainless steel and transferred to a vibration damped table outside the glove box. Thus, STM measurements under inert gas conditions can be performed at high quality, according to the experience of the author for at least one week. Although the air and water stable ionic liquids can be handled in air, hygroscopic precursors like the germanium halides and other hygroscopic compounds require inert gas conditions.
4. Nanoscale electrodeposition in ionic liquids Electrodeposition of metals from AlCl3 based ionic liquids is well known and many publications on this topic exist. Furthermore, the electrodeposition of compound semiconductors like GaAs or InSb was reported in literature. In general, all metals that can be electrodeposited in aqueous media can also be obtained from ionic liquids. In part, even a better quality is obtained, e.g. for palladium [64], which in aqueous solutions is conventionally obtained as a sponge due to the coevolution of hydrogen and its dissolution in the metal during its electrodeposition. From ionic liquids shining and even nanocrystalline thin layers can be obtained. An overview on metal and semiconductor electrodeposition from ionic liquids was recently given by the author in [26]. The most important metals shall therefore only be mentioned here: copper [65], silver [66], gold [67], tin [68], zinc [69], nickel, cobalt [70, 71], cadmium [72], tellurium [73], antimony [74] and indium [75] belong to those metals that can also be electrodeposited from aqueous solutions. Especially in the case of less noble elements like Sb, In and Zn there is an advantage to aqueous media as no metal oxides are formed when the potentiostatic control is interrupted. In Lewis acid AlCl 3 based ionic liquids Al forms at the cathodic limit,
Ionic Liquids: Promising Solvents for Electrochemistry
265
which means that those metals can be electrodeposited that are more noble than Al. Those metals, that are less noble than Al can only be codeposited or Al-alloys form. Among those metals, that can not be electrodeposited in aqueous media or only under a massive hydrogen evolution, aluminium [17], gallium [76], iron [77, 78] sodium [79] and lithium [80] shall be mentioned here. In the form of Al-alloys or solid solutions of the respective metals with Al, niobium [81, 82], titanium [83], chromium [84] and lanthanum [85] can be electrodeposited. Especially for the latter one it was reported that with organic additives it could be obtained in thin layers in elemental form. Furthermore, alloys of Al with iron, copper, nickel, cobalt and silver were investigated [86]. There are only a few studies on semiconductor electrodeposition for example of GaAs [87] and InSb [88]. The latter one was electrodeposited from a liquid based on InCl3 and SbCl3 . In both cases the stoichiometry of bulk layers was not satisfactory and besides the desired semiconductor the respective metals were codeposited. As already mentioned, the elemental semiconductors silicon and germanium cannot be electrodeposited without Al from AlCl3 based ionic liquids. In the opinion of the author the new water and air stable ionic liquids can have a promising future for such purposes.
4.1 Electrodeposition of copper The electrodeposition of copper from aqueous solutions, especially the underpotential deposition, has been well investigated with in situ STM. In sulfuric acid media one finds in the UPD regime 2 well defined redox couples at +230 and +50 mV vs. Cu/Cu 2+ . In the UPD regime a Cu monolayer forms in three steps as also ex situ measurements showed [89–92]: at E > +230 mV first a mobile Cu adlayer is deposited; between +100 and +200 mV vs. Cu/Cu 2+ an ordered adsorption sets in, at +5 mV, finally, a completely closed Cu monolayer forms. Furthermore, the anion of the electrolyte has an important influence on the formation of the superstructures that are observed in the UPD regime [93], and the peak positions in the UPD regime of the cyclic voltammogram change as soon as chloride is added to the sulfuric acid solution. The electrodeposition of copper from AlCl3 based ionic liquids shows some interesting differences to aqueous media. For Lewis acid liquids Cu is always electrodeposited from Cu(I) species [94]. Furthermore, Cu + in ionic melts is not solvated in a classical manner it is rather surrounded by an ion shell. In Fig. 4 two cyclic voltammograms are merged which show the electrochemistry in the the UPD and the OPD regime: at electrode potentials of E > +1.1 V vs. Cu/Cu + gold oxidation sets in, at E < 0 V the bulk electrodeposition of copper begins, and the overvoltage for this process is only 5–10 mV on Au(111). In the UPD regime 3 redox processes are found at +450 mV, at +400 mV and at +80 mV vs. Cu/Cu + . Between +800 mV and +500 mV vs. Cu/Cu + the Au(111) terraces can be probed in high quality and the integral charge for electrodeposition at +500 mV would correspond to a copper cover-
266
F. Endres
Fig. 4. Overview on the electrochemical behavior of BMIC/AlCl3 (34/66 mol-%, ca. 5 mmol/l CuCl2 ) on Au(111): at +1100 mV gold oxidation begins at the step edges, at +1200 mV on the whole of the surface. Three UPD processes, followed by overpotential deposition at E < 0 V are observed. The inlet shows the UPD regime: at 450 mV, at 400 mV and at 80 mV redox couples are obtained. The first process at 450 mV corresponds to step decoration, between 400 and 350 mV an (8 × 8) superstructure forms that closes to a monolayer. The third process at 80 mV is correlated to the growth of clusters as well as to the growth of a second monolayer (v = 100 mV/s) [115]. – Reproduced by permission of the PCCP Owner Societies.
age of 20%, provided, an anion coadsorption does not correspond to the charge. If the electrode potential is set to +450 mV vs. Cu/Cu + , the gold step edges are decorated by copper. In Fig. 5 the processes at +400 mV and +350 mV vs. Cu/Cu + are shown, furthermore one can see the decoration of the steps. It is evident that at +400 mV an ordered structure forms on the surface, which can be well probed at +350 mV in feedback mode. An analysis of the distances gives a hexagonal periodic structure of 2.4 ± 0.2 nm which, referred to Au(111), corresponds to an (8 × 8) superstructure. The integral charge gives as a result 0.7 Cu monolayers, under the assumption that anion adsorption does not contribute to the measured charge. The superstructure can be the result of both anion adsorption similar to the observation in aqueous media, but it is also possible that an ordered but not completely closed Cu superstructure is deposited forming a Moir´e pattern with the underlying Au(111) structure. Concerning the Moir´e effect ref. [95] shall be mentioned here. At +250 mV vs. Cu/Cu + the superstructure disappears and a completely closed Cu monolayer with a coverage of Θ = 1.0 ± 0.1 is obtained. Upon further reducing of the electrode potential clusters with heights of more than one nanometer start growing on the terraces and preferentially at the step edges, at +50 mV a sec-
Ionic Liquids: Promising Solvents for Electrochemistry
267
Fig. 5. On the left in situ STM picture a potential step from +450 to +400 mV was performed at the arrow, and after about 1 min an ordered hexagonal structure forms. At +350 mV (right picture) the hexagonal structure can be well identified as the inlet (Fourier filtered, 13.3 × 13.3 nm2 ) shows. The (8 × 8) superstructure would correspond to a coverage of 0.7 Cu monolayers [115]. – Reproduced by permission of the PCCP Owner Societies.
ond monolayer with a height of 200 ± 20 pm forms (Fig. 6). Both the second monolayer and the growth of clusters in the UPD regime have not yet been reported for aqueous solutions. Maybe aluminium is codeposited to a certain extent which might stabilize the clusters and the second monolayer. It is interesting that upon reoxidation even at electrode potentials as high as +300 mV vs. Cu/Cu + the dissolution of those clusters is slow and only at E > +500 mV vs. Cu/Cu + the original gold surface is recovered. If a bulk layer of Cu is suddenly electrooxidized many in the average 250 pm deep holes can be found in the surface, and during one hour these defects heal completely which means that Cu and Au formed a surface alloy. On Highly Oriented Pyrolytic Graphite copper is only electrodeposited in the OPD regime and the overvoltage for the electrodeposition is about 100 mV. The deposition starts preferentially on the HOPG basal planes and the kinetics of phase formation is determined by 3-dimensional nucleation. At −100 mV vs. Cu/Cu + nucleation is instantaneous, at −150 mV the current transients are best described by the model of a progressive nucleation at a limited number of active sites, and at −200 mV the model of progressive nucleation can be best fitted to the experimental data. At E < −250 mV vs. Cu/Cu + none of the simple models can be applied, probably because of the codeposition of aluminium.
4.2 Electrodeposition of germanium from air and water stable ionic liquids There are only a few studies on the electrodeposition of germanium in literature. Winkler, who discovered germanium, reported in 2 articles, that with
268
F. Endres
Fig. 6. At +50 mV vs. Cu/Cu+ a second Cu monolayer with a height of 200 ± 20 pm, forms, furthermore a distinct cluster growth on the first Cu monolayer and at the step edges occurs (x, y = 120 nm, unity of z-axis in height profile: nm) [115]. – Reproduced by permission of the PCCP Owner Societies.
electroreduction in an ammoniacal solution containing tartrates he could get on platinum a loosely adhering germanium film [96, 97]. Later the electrodeposition of germanium was investigated from molten salts, aqueous alkaline media and organic solvents. Rius and coworkers reported that germanium can be obtained in a quality of 99.9% from GeO 2 in silica based melts at temperatures of about 1000–1100 ◦ C [98]. Furthermore borax melts [99] as well as mixtures of NaF and KF [100] were employed for the electrodeposition of germanium. Its electrodeposition from aqueous media has not been very successful up to now. Hall and Koenig observed in 1934, that in a 3 molar aqueous KOH solution thin Ge layers can be obtained from GeO 2 but the deposition is strongly disturbed by hydrogen evolution [101]. Rather recent studies did not show appreciably improved results [102]. Organic solvents were first applied by Fink and Dokras in 1949. With GeI 4 at 140–150 ◦ C they could obtain in glycol based
Ionic Liquids: Promising Solvents for Electrochemistry
269
solutions and with a current efficiency of 45% grayish, a few micrometer thick brittle germanium films [103]. Szkeley reported in 1951 about the Ge electrodeposition from GeCl4 in different glycolic solutions [104]. He reported that with propylene glycol and 7% GeCl 4 at 59 ◦ C and a constant current density of 0.4 mA/cm 2 roughly 130 µm thick, shining films could be obtained. But the current efficiency in this system was only about 1%. In the following years only a few studies were done on the electrodeposition of germanium. Thus, it can be concluded that there are only a few literature data on the electrodeposition of germanium, to the knowledge of the author systematic electrochemical investigations like cyclic voltammetry or impedance spectroscopy have hitherto not yet been reported. The interest in germanium rised enormously in the recent years as it was found that germanium quantum dots made by Molecular Beam Techniques have interesting optical properties: as one example it shall be mentioned here that such Ge quantum dots on Si(111) show a photoluminescence around 1 eV [105]. Especially with respect to size dependent quantum effects (quantum confinement) such studies are of great interest. The motivation was therefore to find a way to make germanium in a defined manner by electrochemical means and to investigate the processes with the in situ STM in detail on the nanometer scale. The new air and water stable ionic liquids are solvents of the choice for several reasons: 1) they have wide electrochemical windows of more than 4 Volt; 2) they can easily be dried so that a hydrolysis of the germanium halides does not occur; 3) at room temperature they have a negligible vapor pressure; 4) they have a good ionic conductivity; 5) GeCl 4 and GeBr 4 can be dissolved at room temperature up to a concentration of about 0.5 mol/l. In the following, the results of germanium deposition from GeX 4 (X = I, Br, Cl) on Au(111) in the ionic liquid 1-butyl-3-methylimidazolium hexafluorophosphate are briefly summarized. 4.2.1 Germanium on Au(111) from GeI 4 [106] Fig. 7a shows the Cyclic Voltammogram of dry 1-butyl-3-methylimidazolium hexafluorophosphate on Au(111), in Fig. 7b with GeI 4 added in a concentration of ca. 0.1–1 mmol/l. The electrode potentials are referred to the bulk deposition of germanium. It is obvious that the electrochemical window of the dry ionic liquid is a little more than 4 Volt. At the cathodic limit the organic cation is irreversibly reduced and with the STM a less defined deposit can be probed that dissolves well in the liquid when the potentiostat is switched off. At the anodic limit gold oxidation occurs as can be probed with the in situ STM. If germaniumtetraiodide is added in a concentration of about 0.1–1 mmol/l, several redox processes occur: E 1 most likely is correlated to the I 2 /I − redox couple, as this is independent of the substrate, at E 2 gold step oxidation starts, at E 3 250 ± 20 pm high islands form, at E 5 finally the bulk deposition of germanium sets in. E 4 could not be correlated to a definite process on the
270
F. Endres
Fig. 7. (a) Cyclic voltammogram of 1-butyl-3-methylimidazolium hexafluorophosphate on Au(111). The electrochemical window is a little more than 4 Volt, at the cathodic limit it is determined by a less defined reduction of the cation at the anodic limit by the oxidation of gold; (b) If GeI4 is added, 5 redox processes show up. E 1 is most likely correlated to the I2 /I − redox couple, E 2 is correlated to gold step oxidation, at E 3 250 pm high islands grow, at E 5 germanium bulk deposition sets in. E 4 could not be correlated without doubt with a surface process. With the exception of E 3 all peak currents show a linear dependence on the square root of scan rate, the processes are mainly diffusion controlled. The electrode potential for the electrodeposition of germanium in this system was determined from the rising cathodic current of E 5 .
surface. At E > 2 V vs. Ge gold oxidation is such strong that the thin gold film is destroyed within a short time. With the exception of E 3 all peak currents show a linear dependence on the square root of the scan rate indicating mainly diffusion controlled reactions. The electrode potential for the beginning electrodeposition of germanium (in this system) was determined from the rising cathodic current E 5 . Fig. 8 shows an STM picture, where the electrode potential was set at the indicated position from the open circuit potential to +500 mV vs. Ge. On the upper part of the picture the typical gold steps can be identified, their average height is 250 ± 20 pm. On the lower part of the picture, i.e. after the potential step was performed, islands with an average height of 250 ± 20 pm grow, and after a short period of oscillation (because of the sensitively adjusted feedback circuit) the surface is probed in high quality. It was observed that these islands close within a few minutes to a defect rich monolayer whose topography hardly changes when the electrode potential approaches 0 V vs. Ge. In a wider sense both the islands and the defect rich monolayer are an underpo-
Ionic Liquids: Promising Solvents for Electrochemistry
271
Fig. 8. Upon a potential step from the open circuit potential to +500 mV vs. Ge (arrow, downscan) islands with an average height of 250 ± 20 pm grow, which close within a few minutes to a defect rich monolayer [30]. – Reproduced by permission of the PCCP Owner Societies.
tential deposition of germanium. The electrooxidation of the islands with GeI4 as precursor was difficult to investigate as in the presence of iodide this process is very close to gold oxidation so that the processes can hardly be distinguished. With GeCl 4 , nevertheless, these investigations were successful [107]. Furthermore it was very difficult to investigate the growth of the bulk phase. If the electrode potential of the tip was held close to that of the sample, the growth in the OPD regime could hardly be investigated because the tip obviously touched the growing clusters. When higher electrode potentials were applied to the tip in part no deposition at all was observed. For the investigations of the deposition therefore the tip was retracted by about 50 µm, the deposition was performed and then the sample was reapproached to the tip. The sequence in Fig. 9 shows a measurement where the tip was reapproached 10 minutes after the deposition in the OPD regime had been started. First a layer starts growing that consists of small crystallites with heights of a few nanometers. Within one hour the surface transforms into a layered structure with height differences of the terraces of 330 ± 30 pm, the literature value for Ge(111) bilayers is 328 pm. Obviously a complex ripening of the surface sets in resulting in well defined bilayers. In order to exclude that this annealing is induced by the tip the experiment was reproduced under the same conditions, with the exception that the reapproach was done after 1 hour. The result is shown in the upper 2/3 of Fig. 10. Under these conditions a terrace like structure with average heights of 330 ± 30 pm is observed, too. Interestingly there are many defects in the surface (arrows in Fig. 10) whose depth can be up to 1 nm.
272
F. Endres
Fig. 9. A Ge layer that was deposited with retracted tip at −1000 mV vs. Ge shows a complicated annealing: within one hour the surface transforms into a terrace structure with heights of 330 ± 30 pm. This height difference corresponds to Ge(111) bilayers.
Fig. 10. In the upper part of the in situ STM picture at −1000 mV vs. Ge 330 ± 30 pm high Ge(111) bilayers are probed. The white arrows mark up to 1 nm deep defects in the surface. At the (black) arrow the electrode potential was increased to E = −500 mV vs. Ge: spontaneously wormlike nanostructures with maximum heights of 1 nm form in the surface [30]. – Reproduced by permission of the PCCP Owner Societies.
Ionic Liquids: Promising Solvents for Electrochemistry
273
At the marked position in Fig. 10 (arrow) the electrode potential was set to −500 mV vs. Ge, and one can see that wormlike nanostructures form on the surface. The height of these worms does not extend 1 nm. It is interesting that this partial oxidation occurred on the whole of the surface and not preferentially at step edges where it starts in the case of many metals. If the electrode potential is reduced to −1000 mV vs. Ge, these nanostructures heal completely and there are several processes on the surface: both new clusters can grow and islands dissolve, furthermore pinch off phenomena occur. Within about 2 hours all these processes lead to a transformation of the wormlike nanostructures into a terrace like surface structure. Under the reported conditions the maximum thickness of the Ge layer was in the range of 20 nm. Apparently the electrodeposition is irreversible as no stripping peak was observed in the cyclic voltammogram. Nevertheless the thin germanium film disappears completely as soon as the potentiostat is switched off. It could be found that a chemical attack of GeI 4 on the deposited Ge was the reason whereby Ge x I y species seem to form. Such observations were also made in electrodeposition studies of germanium from organic solvents [103]. The disadvantage of this system is that micrometer thick layers that are well suited to ex situ analysis could not be obtained. Furthermore no stable nanoclusters could be obtained, they rather transform into Ge(111) bilayers. Furthermore the oxidation of the islands in the UPD regime could hardly be investigated because it only occurred at electrode potentials where gold oxidation started at the steps, too. Therefore, for further studies GeBr 4 and GeCl4 were employed, both substances are liquid at room temperature and they can be dissolved up to about 0.5 mol/l in 1-butyl-3-methylimidazolium hexafluorophosphate. 4.2.2 Germanium on Au(111) from GeBr 4 [108] Fig. 11 shows the cyclic voltammogram of 1-butyl-3-methylimidazolium hexafluorophosphate, which was saturated with GeBr 4 . The scan rate was 1 mV/s. The CV is dominated by 2 irreversible mainly diffusion controlled reduction processes: the first one with a minimum at about +700 mV vs. Ge is mainly correlated to the reduction of Ge(IV) to Ge(II), at 0 V the deposition of the bulk phase starts as can also be seen with the naked eye as a black deposit starts to form. In some measurements (as in Fig. 11) a double peak was observed, which disappears in the second cycle. Maybe deposition at defects or a surface alloying play a role here. The peak currents in both reduction processes show a linear dependence on the square root of the scan rate indicating that they are mainly diffusion controlled. At about −1 V vs. Ge the reduction of the organic cation begins. Typical stripping peaks are not observed and, as described above, GeBr 4 also chemically attacks the deposited Ge. At electrode potentials of E > +1 V vs. Ge the oxidation of the gold step edges starts. The in situ STM measurements on gold under open circuit conditions showed repro-
274
F. Endres
Fig. 11. Cyclic voltammogram of GeBr4 , saturated in 1-butyl-3-methylimidazolium hexafluorophosphate, v = 1 mV/s: the first reduction process at E ≈ +700 mV is mainly correlated with the reduction of Ge(IV) to Ge(II), at 0 V Ge bulk deposition begins. At E > +1 V gold is strongly oxidized [108]. – Reproduced by permission of the PCCP Owner Societies.
ducibly a surface with striped structures. Upon rising the electrode potential the surface becomes smoother, but also the oxidation of gold starts, preferentially at step edges and defects [108]. Therefore the in situ STM experiments were always started at the open circuit potential and the electrode potential was then decreased to more cathodic values. In Fig. 12 the STM pictures of the surface at +200, +100, +50 and 0 mV vs. Ge are represented. Fig. 12a shows the surface at +200 mV vs. Ge. At this electrode potential the surface is completely covered with a thin germanium layer with a rather metallic behavior: in the current voltage tunneling spectrum there is no band gap. From GeCl4 quite a similar surface is obtained. At +100 mV vs. Ge striped structures begin to grow on the surface, their heights are between 300 and 500 pm (Fig. 12b), at +50 mV by lateral merging islands with lengths between 50 and 80 nm and a height of about 1 nm form (Fig. 12c and d): if an epitactic growth of Ge was the case this height would correspond to about 3 Ge(111) bilayers. If the electrode potential is further reduced, the height of these islands rises and at E = −250 mV vs. Ge micrometer thick germanium films can be obtained. The tunneling spectrum of 200 nm thick Ge layers shows intrinsic semiconducting behavior with a symmetrical band gap of about 0.7 ± 0.1 eV. This value is in good agreement with the literature value of 0.67 eV for microcrystalline germanium at 300 K [109]. In contrast, the gold surface under open circuit conditions exhibits a typical metallic behavior. The “UPD” layer of Fig. 12 shows in the tunneling spectrum a characteristics that is between that one of the gold surface and that one of the
Ionic Liquids: Promising Solvents for Electrochemistry
275
Fig. 12. (a) shows a closed “UPD” layer of Ge on Au(111) at +200 mV vs. Ge, (b) at +100 mV: 300–500 pm high striped structures (arrows) grow on the surface. (c) shows the surface at +50 mV vs. Ge: the striped structures seem to merge laterally and at 0 V vs. Ge islands with lengths between 50 and 80 nm and heights in the nanometer regime grow; (d) the Ge islands have heights of about 1 nm [108]. – Reproduced by permission of the PCCP Owner Societies.
semiconducting germanium film. It can be seen in the tunneling spectrum that beginning with positive gap voltages the current in the backward scan is higher than in the forward scan. In literature such effects are explained on the basis of surface states that are formed in the electrical field between tip and sample [110]. Furthermore, a strong electrical field can lead to a band bending in the sample, which could also result in a hysteresis. Ex situ XPS measurements have shown without doubt that the topmost 5 nm of thick germanium layers consist of elemental germanium [111]. Together with the tunneling spectra this proves that indeed elemental germanium was electrodeposited. 4.2.3 Germanium on Au(111) from GeCl 4 [107] Both with GeI 4 and GeBr 4 in a wider sense underpotential deposition of germanium was observed on Au(111) but in both cases the reoxidation was difficult to investigate. In the presence of iodide and bromide, respectively, gold oxidation is such favored by complexation of gold ions with halide that germanium
276
F. Endres
Fig. 13. Starting at the open circuit potential the cyclic voltammogram of GeCl4 , saturated in 1-butyl-3-methylimidazolium hexafluorophosphate shows 2 irreversible mainly diffusion controlled redox processes on Au(111). The first process with a minimum at +500 mV vs. Ge is mainly correlated to the reduction of Ge(IV) to Ge(II), the second process corresponds to the bulk deposition of Ge. At about −1 V vs. Ge the irreversible reduction of the organic cation starts. The oxidation current between about +0.7 and +1.0 V is correlated to the electrooxidation of Ge, at E > 1.2 V gold oxidation begins at the step edges. With the exception of the step oxidation at +1400 mV these processes could not be correlated to defined processes on the surface [107]. – Reproduced by permission of the PCCP Owner Societies.
oxidation and beginning gold decomposition are close together. Fortunately, with GeCl 4 in the already mentioned ionic liquid the electrodeposition and oxidation of germanium can be investigated in the UPD and in the OPD regime. Fig. 13 shows the cyclic voltammogram of GeCl 4 , that is saturated in 1-butyl3-methylimidazolium hexafluorophosphate: as in the case of GeBr 4 the CV is mainly determined by the Ge(IV)/Ge(II) process as well as by the deposition of germanium at E < 0 V vs. Ge. At E > 1.2 V vs. Ge gold oxidation slowly starts at the step edges. Fig. 14 gives an overview on the electrode processes observed in this system: for this purpose together with the STM measurement at +1200 mV vs. Ge a slow Linear Sweep Voltammetry (LSV) scan (1 mV/s) towards more negative potentials was started. The respective electrode potentials are given in the top and in the bottom of the pictures, the values in between can be linearly interpolated. Fig. 14a–d shows the result: in the upper part of Fig. 14a the typical gold terraces with average heights of 250 pm can be identified. At about +900 mV vs. Ge under these conditions the first island formation is seen (arrow in Fig. 14a). Further measurements showed (see [107]) that at +1000 mV vs. Ge first the steps are decorated, upon subsequent decreasing of
Ionic Liquids: Promising Solvents for Electrochemistry
277
Fig. 14. These in situ STM pictures show an experiment, where together with a slow STM scan (450 sec per picture) a Linear Sweep Voltammetry scan with a scan rate of 1 mV/s was performed: beginning at +1200 mV vs. Ge islands form on the surface (arrows in (a)) at about +900 mV vs. Ge. These islands close on further reducing of the electrode potential to a rough monolayer (b and c). At −50 mV the growth of Ge nanocrystallites begins, a typical height profile shows their shape. At the arrow in (c) the LSV scan was stopped, and the electrode potential was held constant at −50 mV vs. Ge, resulting in a slow growth of the crystallites (d) [107]. – Reproduced by permission of the PCCP Owner Societies.
the electrode potential to +950 mV and +750 mV vs. Ge 150 and 250 pm high islands grow, and when these islands are stripped holes remain in the surface. If the electrode potential is further reduced, at +300 mV vs. Ge (Fig. 14b, bottom) a closed but rather rough layer forms on the electrode surface. Between +300 and 0 mV the structure of this layer hardly changes and at −50 mV vs.
278
F. Endres
Fig. 15. In (a) the electrode potential was set from +1200 mV to +750 mV. Spontaneously islands with an average height of 250 ± 20 pm grow, and at these electrode potentials hardly any further growth is observed. If the islands are stripped from the surface, lots of difficult to probe holes remain in the surface (arrows and inlet in (b)). Their measurable depths are between 30 and 100 pm. A surface alloy between Ge and Au is very likely [107]. – Reproduced by permission of the PCCP Owner Societies.
Ge (Fig. 14c) the growth of germanium nanoclusters with widths between 10 and 40 nm and heights of a few nanometers begins. The height to width ratio of these clusters is around 1/10 in the beginning. At −50 mV the LSV scan was stopped, and the growth of the clusters was observed for a while. Fig. 14d shows the same site 8 min later, and the surface is now completely covered. If the electrode potential is increased to values up to 1.2 V vs. Ge, the main parts of the deposits dissolve rapidly. But it is interesting that first some clusters with heights of up to a few nanometers remain on the surface, and they are only slowly dissolved. Furthermore, many islands of 250 pm height remain, and they disappear only within a few hours. In Fig. 15 an experiment is shown, where at the arrow the electrode potential was set from +1200 mV vs. Ge to +750 mV: islands with an average height of 250 pm grow (Fig. 15a). If then suddenly the electrode potential is set to +1200 mV vs. Ge the islands disappear rapidly leaving difficult to probe holes in the surface with apparent depths between 30 and 100 pm. These holes heal within a few minutes. The phase diagram Au/Ge shows an eutecticum at x(Ge) = 28 mol-% and at a temperature of 361 ◦ C [112], and a solid solution of germanium in gold with a maximum solubility of 3 mol-% at about 200 ◦ C is known. Furthermore, gold/germanium alloys were reported [113], and a grain boundary alloy of germanium in gold is known [114]. At +300 mV vs. Ge a completely closed about 300 pm thick rough layer with a rather metallic behavior forms, similar to that one observed for the deposition from GeBr 4 . If the electrode potential is further reduced to E < 0 V with the tip at +800 mV the already described cluster growth sets in.
Ionic Liquids: Promising Solvents for Electrochemistry
279
If the electrode potential is held at +200 mV vs. Ge and that of the sample at +100 mV, exclusively under the STM tip cluster agglomerates grow, that seem to consist of clusters that are only a few nanometers wide. Current/voltage tunneling spectra on those clusters show a band gap of 0.3–0.4 eV which is not symmetrical to zero voltage. One possible explanation for the growth of these clusters might be that a jump to contact transfer from the tip to the sample occurs, another explanation might be that in the electrical field between tip and sample a preferred growth sets in. 4.2.4 A comparative view on germanium electrodeposition from an ionic liquid The electrodeposition of germanium on Au(111) from the ionic liquid 1-butyl3-methylimidazolium hexafluorophosphate and the three germanium halides GeI 4 , GeBr 4 and GeCl4 shows a few common characteristics but also a few differences. With all three substances there is in a wider sense an underpotential deposition, and before the bulk growth sets in a complete monolayer forms. Its maximum height is 300 pm. Only with GeCl 4 electrodeposition in the UPD and OPD regime could be investigated reversibly, in the other cases the oxidation of the UPD layer occurs in a potential range where also the oxidation of gold begins. The initial states of deposition in the OPD regime could hardly be probed with GeI 4 . It seems that the initial deposition leads to nanoclusters that on the time scale of about one hour transform into Ge(111) bilayers. With GeBr 4 such a transformation was only observed in part, with GeCl 4 stable nanoclusters can be obtained that are stable for days. Both with GeBr 4 and GeCl4 at room temperature micrometer thick layers can be electrodeposited from saturated solutions, whereas with GeI 4 the thickness seems to be limited to 100 nm. Qualitative in situ current/voltage tunneling spectroscopy was only performed at layers that were made from GeBr 4 and GeCl4 . The result is that the 300 pm thick UPD layer shows metallic behavior, but the work functions seem to be higher than for pure gold. A band gap could not be observed. Beginning with a thickness of about 20 nm for a Ge layer on Au(111) a band gap of about 0.7 ± 0.1 eV is observed, within the limits of errors the value that has to be expected.
5. Summary and outlook In the present review a few results of the nanoscale electrodeposition of metals and semiconductors from ionic liquids were summarized. It is shown that in situ STM measurements can be performed in high quality under inert gas in these new liquids. Due to their wide electrochemical windows of up to 6 Volt, their low vapor pressures and their high chemical stability and due to the fact that they can nowadays be tailor-made ionic liquids have a great chance in electrochemistry and electrochemical nanotechnology. The trend today is to design
280
F. Endres
hydrophobic well ion- conducting liquids that can be handled in air. The aim of this article was to show that even semiconductors like germanium, that hitherto with nanometer resolution were deposited only under UHV conditions by physical methods, can be made by electrochemical means. In the opinion of the author maybe in a few years ionic liquids will be employed routinely in electrochemistry like water and organic solvents are today. In any case they deserve to be employed more widely in electrochemistry.
References 1. A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley & Sons, New York, Chichester, Weinheim, Brisbane, Singapore, Toronto (2001). 2. J. D. E. McIntyre and D. M. Kolb, Symp. Faraday Soc. 4 (1970) 99. 3. D. M. Kolb, in Spectroelectrochemistry: Theory and Practice, R. J. Gale, Ed., Plenum, New York (1988) p. 87. 4. D. L. Jeanmaire and R. P. VanDuyne, J. Electroanl. Chem. 84 (1977) 1. 5. A. Bewick and K. Kunimatsu, Surf. Sci. 101 (1980) 131. 6. L. Blum, H. D. Abruna, J. White, J. G. Gordon, G. L. Borges, M. G. Samant, and O. R. Melroy, J. Chem. Phys. 85 (1986) 6732. 7. M. F. Toney, O. R. Melroy, in Electrochemical Interfaces, H. Abruna, Ed., VCH, Weinheim (1991), p. 55. 8. R. Schumacher, Angew. Chem. Int. Edn. 29 (1990) 329. 9. V. L. Shannon, D. A. Koos, and G. L. Richmond, J. Chem. Phys. 87 (1987) 1440. 10. P. M. Saville, M. Gonsalves, A. R. Hillmann, and R. Cubitt, J. Phys. Chem. B 101 (1997) 1. 11. Y. Y. Tong, E. Oldfield, and A. Wieckowski, Anal. Chem. A 70 (1998) 518. 12. J. C. Vickerman, Ed., Surface Analysis: The Principal Techniques, John Wiley & Sons, Chichester, New York, Weinheim, Brisbane, Singapore, Toronto (1997). 13. G. Binnig and H. Rohrer, Helv. Phys. Acta 55 (1982) 726. 14. L. Kuipers, R. W. M. Loos, H. Neerings, J. Ter Horst, G. J. Ruwiel, A. P. De Jongh, and J. W. M. Frenken, Rev. Sci. Instrum. 66 (1995) 4557. 15. R. Sonnenfeld and P. K. Hansma, Science 232 (1986) 211. 16. H.-Y. Liu, F. R. F. Fan, C. W. Lin, and A. J. Bard, J. Am. Chem. Soc. 108 (1986) 3838. 17. Y. Zhao and T. J. VanderNoot, Electrochim. Acta 42 (1997) 3. 18. D. Pletcher and F. C. Walsh, Industrial Electrochemistry, 2nd ed., Blackie Academic & Professional, London, Glasgow, New York, Tokyo, Melbourne, Madras (1993). 19. P. Taxil, B. Lafage, O. Boiko, P. Chamelot, and K. Serrano, NATO ASI Ser., Ser. 3(53) (Refractory Metals in Molten Salts) (1998) 131. 20. D. Wei and M. Okido, Curr. Top. Electrochem. 5 (1997) 21. 21. I. Galasiu, R. Galasiu, and J. Thonstad, Nonaqueous Electrochem. (1999) 461. 22. J. Robinson and R. A. Osteryoung, J. Am. Chem. Soc. 101 (1979) 323. 23. J. S. Wilkes, J. A. Levisky, R. A. Wilson, and C. L. Hussey, Inorg. Chem. 21 (1982) 1263. 24. C. L. Hussey, in Chemistry of Nonaqueous Solutions Current Progress, G. Mamantov and A. I. Popov, Eds., VCH-Publishers (1994), p. 227. 25. R. T. Carlin and J. S. Wilkes, in Chemistry of Nonaqueous Solutions Current Progress, G. Mamantov and A. I. Popov, Eds., VCH-Publishers (1994), p. 277.
Ionic Liquids: Promising Solvents for Electrochemistry
281
26. F. Endres, Chem. Phys. Chem. 3 (2002) 144. 27. F. Endres, in Ionic Liquids in Synthesis, P. Wasserscheid and T. Welton, Eds., Wiley-VCH (2002). 28. J. S. Wilkes and M. J. Zaworotko, J. Chem. Soc. Chem. Comm. (1992) 965. 29. P. Bonhôte, A.-P. Dias, N. Papageorgiou, K. Kalyanasundaram, and M. Grätzel, Inorg. Chem. 35 (1996) 1168. 30. F. Endres and C. Schrodt, Phys. Chem. Chem. Phys. 2 (2000) 5517. 31. G. J. Janz, Ed., Molten Salts Handbook, Academic Press, New York, London (1967). 32. Y. Ito and T. Nohira, Electrochim. Acta 45 (2000) 2611. 33. K. R. Seddon, J. Chem. Tech. Biotechnol. 68 (1997) 351. 34. K. R. Seddon, Kinet. Catal. Engl. Transl. 37 (1996) 693. 35. M. J. Earle and K. R. Seddon, Pure. Appl. Chem. 72 (2000) 1391. 36. P. Wasserscheid and T. Welton, Eds., Ionic Liquids in Synthesis, VCH-Wiley (2002). 37. P. A. Z. Suarez, J. E. L. Dullius, S. Einloft, and R. F. de Souza, J. Dupont, Polyhedron 15 (1996) 1217. 38. R. Hagiwara and Y. Ito, J. Fluorine Chem. 105 (2000) 221. 39. P. Wasserscheid and W. Keim, Angew. Chem. Int. Ed. 39 (2000) 3772. 40. F. H. Hurley and T. P. Wier, J. Electrochem. Soc. 98 (1951) 207. 41. A. A. Fannin, D. A. Floreani, L. A. King, J. S. Landers, B. J. Piersma, D. J. Stech, R. L. Vaughn, J. S. Wilkes, and J. L. Williams, J. Phys. Chem. 88 (1984) 2614. 42. P. R. Gifford and J. B. Palmisano, J. Electrochem. Soc. 134 (1987) 610. 43. H. A. Øye, M. Jagtoyen, T. Oksefjell, and J. S. Wilkes, Mater. Sci. Forum 73–75 (1991) 183. 44. Y. C. Wu, K. W. Pratt, and W. F. Koch, J. Sol. Chem. 18 (1989) 515. 45. P. Koronaios, D. King, and R. A. Osteryoung, Inorg. Chem. 37 (1998) 2028. 46. J. D. Holbrey and K. R. Seddon, J. Chem. Soc., Dalton Trans. (1999) 2133. 47. J. Fuller, R. T. Carlin, H. C. De Long, and D. Haworth, J. Chem. Soc. Chem. Comm. 3 (1994) 299. 48. J. L. Anthony, E. J. Maginn, and J. F. Brennecke, J. Phys. Chem. B 105 (2001) 10 942. 49. A. E. Visser, R. P. Swatloski, W. M. Reichert, R. Mayton, S. Sheff, A. Wierzbicki, J. H. Davis, and R. D. Rogers, Chem. Comm. (2001) 135. 50. P. Wasserscheid, W. Keim, C. Bolm, and A. Bösmann, PCT Int. Appl. (2001) WO 0155060. 51. P. A. Z. Suarez, V. M. Selbach, J. E. L. Dullius, S. Einloft, C. M. S. Platnicki, D. S. Azambuja, R. F. de Souza, and J. Dupont, Electrochim. Acta 42 (1997) 2533. 52. P.-Y. Chen and I.-W. Sun, Proc. Electrochem. Soc. 98(11) (1998) 55. 53. Y. Katayama, S. Dan, T. Miura, and T. Kishi, J. Electrochem. Soc. 148 (2001) C102. 54. T. Welton, Chem. Rev. 99 (1999) 2071. 55. P. A. Z. Suarez, S. Einloft, J. E. L. Dullius, R. F. de Souza, and J. Dupont, J. Chim. Phys. Phys.-Chim. Biol. 95 (1998) 1626. 56. C. Nanjundiah, F. Mc Devitt, and V. R. Koch, J. Electrochem. Soc. 144 (1997) 3392. 57. D. A. Bonnell and B. D. Huey, in Scanning Probe Microscopy and Spectroscopy: Theory, Techniques and Applications, D. A. Bonnell, Ed., 2nd ed., Wiley-VCH, New York, Chichester, Weinheim, Brisbane, Singapore, Toronto (2001), Chap. 2. 58. L. A. Nagahara, T. Thundat, and S. M. Lindsay, Rev. Sci. Instrum. 60 (1989) 3128. 59. D. J. Trevor, C. E. D. Chidsey, and D. N. Loiacono, Phys. Rev. Lett. 62 (1989) 929. 60. S.-L. Yau, C. M. Vitus, and B. C. Schardt, J. Am. Chem. Soc. 112 (1990) 3577.
282
F. Endres
61. A. A. Gewirth, D. H. Craston, and A. J. Bard, J. Electroanl. Chem. 261 (1989) 477. 62. R. Christoph, H. Siegenthaler, H. Rohrer, and H. Wiese, Electrochim. Acta 34 (1989) 1011. 63. M. J. Heben, M. M. Dovek, N. S. Lewis, R. M. Penner, and C. F. Quate, J. Microsc. (Oxford) 152(3) (1988) 651. 64. H. C. De Long, J. S. Wilkes, and R. T. Carlin, J. Electrochem. Soc. 141 (1994) 1000. 65. C. L. Hussey, L. A. King, and R. A. Carpio, J. Electrochem. Soc. 126 (1979) 1029. 66. X.-H. Xu and C. L. Hussey, J. Electrochem. Soc. 139 (1992) 1295. 67. X.-H. Xu and C. L. Hussey, Proc. Electrochem. Soc. 16 (1992) 445. 68. X.-H. Xu and C. L. Hussey, J. Electrochem. Soc. 140 (1993) 618. 69. W. R. Pitner and C. L. Hussey, J. Electrochem. Soc. 144 (1997) 3095. 70. J. A. Mitchell, W. R. Pitner, C. L. Hussey, and G. R. Stafford, J. Electrochem. Soc. 143 (1996) 3448. 71. W. R. Pitner, C. L. Hussey, and G. R. Stafford, J. Electrochem. Soc. 143 (1996) 130. 72. M. A. M. Noel and R. A. Osteryoung, J. Electroanl. Chem. 293 (1996) 139. 73. E. G.-S. Jeng and I.-W. Sun, J. Electrochem. Soc. 144 (1997) 2369. 74. M. Lipsztjan and R. A. Osteryoung, Inorg. Chem. 24 (1985) 3492. 75. J. S.-Y. Liu and I.-W. Sun, J. Electrochem. Soc. 144 (1997) 140. 76. P.-Y. Chen, Y.-F. Lin, and I.-W. Sun, J. Electrochem. Soc. 146 (1999) 3290. 77. C. Nanjundiah, K. Shimizu, and R. A. Osteryoung, J. Electrochem. Soc. 129 (1982) 2474. 78. M. Lipsztajn and R. A. Osteryoung, Inorg. Chem. 24 (1985) 716. 79. G. E. Gray, P. A. Kohl, and J. Winnick, J. Electrochem. Soc. 142 (1995) 3636. 80. B. J. Piersma, Proc. Electrochem. Soc. 94 (1994) 415. 81. N. Koura, T. Kato, and E. Yumoto, Hyomen Gijutsu 45 (1994) 805. 82. G. T. Cheek, H. C. De Long, and P. C. Trulove, Proc. Electrochem. Soc. 99 (2000) 527. 83. N. Guo, J. Guo, and S. Xiong, Fushi Kexue Yu Fanghu Jishu 10 (1998) 290. 84. M. R. Ali, A. Nishikata, and T. Tsuru, Electrochim. Acta 42 (1997) 2347. 85. T. Tsuda and Y. Ito, Proc. Electrochem. Soc. 99 (2000) 100. 86. R. T. Carlin, H. C. De Long, J. Fuller, and P. C. Trulove, J. Electrochem. Soc. 145 (1998) 1598. 87. M. W. Verbrugge and M. K. Carpenter, AlChE J. 36 (1990) 1097. 88. M. K. Carpenter and M. W. Verbrugge, J. Mater. Res. 9 (1994) 2584. 89. Y. Nakai, M. S. Zei, D. M. Kolb, and G. Lehmpfuhl, Ber. Bunsenges. Phys. Chem. 88 (1984) 340. 90. J. G. Gordon, O. R. Melroy, and M. F. Toney, Electrochim. Acta 40 (1995) 3. 91. L. Blum and D. A. Huckaby, J. Electroanl. Chem. 315 (1991) 255. 92. L. Blum and D. A. Huckaby, Proc. Electrochem. Soc. 93–5 (1993) 232. 93. M. S. Zei, G. Qiao, G. Lehmpfuhl, and D. M. Kolb, Ber. Bunsenges. Phys. Chem. 91 (1987) 349. 94. C. Nanjundiah and R. A. Osteryoung, J. Electrochem. Soc. 130 (1983) 1312. 95. I. Amidror, The Theory of the Moir´e Phenomenon, Kluwer Academic Publishers, Dordrecht, Boston, London (2000). 96. C. Winkler, J. Prakt. Chem. 34 (1886) 208. 97. C. Winkler, Z. Anal. Chem. 26 (1887) 359. 98. A. Rius, P. Colon, and A. Arbacho, Electrochim. Acta 11 (1966) 1497. 99. M. J. Barbieux-Andrieux, Ann. Chim. 10 (1944) 754. 100. R. Monnier and P. Tissot, Helv. Chim. Acta 47 (1964) 2203. 101. J. L. Hall and A. E. Koenig, Trans. Electrochem. Soc. 65 (1934) 215.
Ionic Liquids: Promising Solvents for Electrochemistry
283
102. M. Schüßler, T. Statzner, C. I. Lin, V. Krozer, J. Horn, and H. L. Hartnagel, J. Electrochem. Soc. 143 (1996) L73. 103. C. G. Fink and V. M. Dokras, J. Electrochem. Soc. 95 (1949) 80. 104. G. Szekely, J. Electrochem. Soc. 98 (1951) 318. 105. O. Leifeld, A. Beyer, E. Müller, D. Grützmacher, and K. Kern, Thin Solid Films 380 (2000) 176. 106. F. Endres, Phys. Chem. Chem. Phys. 3 (2001) 3165. 107. F. Endres and Sh. Zein El Abedin, Phys. Chem. Chem. Phys. 4 (2002) 1649. 108. F. Endres and Sh. Zein El Abedin, Phys. Chem. Chem. Phys. 4 (2002) 1640. 109. J. I. Pankove, Optical Processes in Semiconductors, Dover Publications Inc., New York (1970). 110. S. Gilbert and J. H. Kennedy, Surf. Sci. 225 (1990) L1. 111. F. Endres, Electrochem. Solid State Lett. 5 (2002) C38. 112. T. B. Massalski, Ed., Binary Alloy Phase Diagrams, ASM International Materials Park, Ohio (1990), 2nd ed. 113. A. M. Edwards, M. C. Fairbanks, and R. J. Newport, Philos. Mag. B 63(2) (1991) 457. 114. S. Ingrey and B. MacLaurin, J. Vac. Sci. Techn. A 2 (1984) 358. 115. F. Endres and A. Schweizer, Phys. Chem. Chem. Phys. 2 (2000) 5455.