Chemical Physics Letters 517 (2011) 103–107

Contents lists available at SciVerse ScienceDirect

Chemical Physics Letters

journal homepage: www.elsevier .com/ locate /cplet t

In situ electrochemical-X-ray Photoelectron Spectroscopy: Rubidium metaldeposition from an ionic liquid in competition with solvent breakdown

Rahmat Wibowo a, Leigh Aldous a, Robert M.J. Jacobs b, Ninie S.A. Manan c,d, Richard G. Compton a,⇑a Department of Chemistry, Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QZ, United Kingdomb Department of Chemistry, Chemistry Research Laboratory, University of Oxford, Mansfield Road, Oxford OX1 3TA, United Kingdomc School of Chemistry and Chemical Engineering, The QUILL Centre, Queen’s University Belfast, Belfast BT9 5AG, United Kingdomd Department of Chemistry, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history:Received 17 August 2011In final form 6 October 2011Available online 14 October 2011

0009-2614/$ - see front matter � 2011 Elsevier B.V. Adoi:10.1016/j.cplett.2011.10.017

⇑ Corresponding author. Fax: +44 (0) 1865 275 410E-mail address: richard.compton@chem.ox.ac.uk (R

The electrodeposition of rubidium from an ionic liquid (IL) N-butyl-N-methylpyrrolidium bis(trifluoro-methylsulfonyl)imide ([C4mpyrr][NTf2]) has been performed and monitored at a Nickel mesh electrodeby using in situ electrochemical-X-ray Photoelectron Spectroscopy (XPS) measurements. At extremelyhigh current values during the deposition of the metal, the solvent breakdown was also observed. Bychoosing suitable low current values, electrodeposition of Rb can be promoted over the IL degradation.IL degradation was characterised by carbonisation of the electrode-IL-vacuum interface, with the lossof fluorine being relatively pronounced, consistent with reduction of the [NTf2]� anion.

� 2011 Elsevier B.V. All rights reserved.

1. Introduction

Rubidium is one of the Group I alkali metals, which are charac-teristically strong reducing agents. The formal potential (E0

f ) of theRb/Rb+ couple in aqueous solution is �2.94 V (vs. SHE) [1], whilethe E0

f for Rb/Rb+ in the ionic liquid [C4mPyrr][NTf2] has recentlybeen measured as �3.37 V (vs. Fc/Fc+) [2]. This property makesRb a potential candidate for use in solid batteries [3] as well as en-ergy storage devices [4]. Rubidium cannot be electrodepositedfrom aqueous solution or many organic solvents as their potentialwindows are not wide enough.

Room temperature ionic liquids (RTILs or ILs), which are typi-cally comprised entirely of ions, are promising media for the depo-sition for active metals such as Rubidium [2,5]. Their intrinsic ionicconductivity, chemical stability as well as wide electrochemicalwindows makes the electrodeposition of many active metals possi-ble [6]. Moreover, another important property of ILs is that theypossess near zero vapour pressure. The combination of theseproperties has facilitated novel spectroelectrochemical techniquesunder high vacuum conditions, including in situ electrochemical-X-ray spectroscopy. A number of publications are availableregarding the use of in situ electrochemical-XPS measurements inconjunction with ILs, probing the oxidation of Cu to Cu(I) [7],reduction of Fe(III) to Fe(II) [8], the intercalation of [C2mim][BF4]into HOPG [9], and the XPS induced charging [10] and electro-chemical shift in binding energy [11] of ionic liquids. The diverse

ll rights reserved.

..G. Compton).

range of other (non-electrochemical) studies relating to combinedIL-XPS studies has recently been reviewed [12].

We have reported the electrodeposition of Group I alkali metalsfrom the ionic liquid N-butyl-N-methylpyrrolidinium bis(trifluoro-methylsulfonyl)imide ([C4mpyrr][NTf2]) on Pt and Ni electrodes[2,13–15]. The fundamental kinetic and thermodynamic propertieshave been determined by means of simulation [2,13–15]. In addi-tion, recently we have reported the in situ electrochemical-XPSmeasurements to monitor the electrodeposition of potassium inthe ionic liquid of [C4mpyrr][NTf2] using in situ electrochemical-XPS measurements [16]. Unlike potassium, rubidium electrodepos-ition is close to the reductive breakdown of [C4mpyrr][NTf2]. In thisLetter, we report the use of in situ electrochemical-XPS to explorethe breakdown of the solvent in competition with the electrode-position of rubidium.

2. Experimental

N-butyl-N-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide, [C4mpyrr][NTf2], was prepared using standard literaturemethods [17]. The Rb[NTf2] salt was prepared by the reaction ofaqueous solutions of rubidium hydroxide with a 5% molar excessof H[NTf2], followed by repeated dissolution in water then dryingunder vacuum, until pH neutral.

XPS measurements were performed using a VG ESCALAB MkIIspectrometer equipped with a monochromatic Al Ka X-ray source(photon energy of 1486.6 eV). Survey scans and detailed scanswere recorded with a constant analyser pass energy of 100 and20 eV, respectively. Prior to measurement, the cell was put under

104 R. Wibowo et al. / Chemical Physics Letters 517 (2011) 103–107

vacuum in a desiccator overnight to remove dissolved water andgases. In order to prevent ionic liquid loss during degassing, a spe-cially designed lid was used. The cell was then rapidly transferredto the XPS entry chamber and evacuated for a further 2 h beforebeing moved into the sample chamber.

All electrochemical measurements were carried out using acomputer-controlled PGSTAT-12 Autolab potentiostat (Eco-Che-mie, Netherlands).

3. Results and discussion

3.1. Experimental set-up

The in situ electrochemical-XPS cell (Figure 1a) and holder (Fig-ure 1b) used in this study were fabricated in-house and are modi-fications of a previously reported electrochemical setup [16].Briefly, the cell holder was a previously described mobile cellholder [16] that had been fixed permanently onto the end of anextendable and retractable manipulator arm which was insertedinto the XPS chamber through a high vacuum flange. The sam-ple-holding cell consisted of two parts, namely two hollow cylin-ders of PEEK with the bottom of one sealed by a modified copperXPS stub, which were manufactured and joined as previously de-tailed elsewhere [16]. An arm was placed on the side of the cellin order to allow movement of the cell with a standard manipula-tor arm. The cell was lowered from above onto two pogo spring

Figure 1. In situ electrochemical-XPS cell (a) and associated holder inside the XPSchamber (b).

pins emerging from the holder which entered tapered holes inthe cell and made electrical contact with the working and refer-ence electrodes, respectively, via copper foil. Much of the counterelectrode passed through a hole in the holder before coming to reston a copper o-ring set in the holder. Three shielded wires ran fromthe holder to the potentiostat via an electrical feed-through in theXPS housing. Once the cell was secured on top of the holder, thestandard XPS holder was retracted and the entire assembly ex-tended until the sample was in line with the X-ray source.

3.2. XPS measurement of the ionic liquid

The XPS survey scan of 0.1 M Rb[NTf2] in [C4mpyrr][NTf2] re-corded in the XPS-electrochemistry cell prior to any electrochem-ical experiments is depicted in ‘0 h’ spectrum in Figure 2. Veryweak Rb signals were observed, although this is easily explainedby the relatively dilute nature of the solution (in terms of atomic% values). Signals were observed corresponding to C, N, O, F andS. Analysis of these signals were similar to those expected basedupon the stoichiometric ratio for the pure ionic liquid (deviationless than 5%), using the widely available and frequently used stan-dard atomic sensitivity factors by Wagner et al. [18]. Minor differ-ences between measured and known atomic ratios during XPSmeasurements of ILs have been reported before, and have beenattributed to system-dependant deviations which require thedevelopment of system-specific atomic sensitivity factors using‘ultrapure’ ILs [12,19]. In this work the sensitivity factors of Wag-ner et al. [18] were found to provide a sufficient degree of accuracy.No impurities were detected in the [C4mpyrr][NTf2], either by XPSor cyclic voltammetry.

Weak signals were also observed corresponding to the Ni meshelectrode (Ni 2p3/2 at 855 eV). This relatively weak signal (com-pared to the quantity of Ni present) was due to the take-off angleof 75� employed on the curved Ni wires of the mesh, e.g. if the sam-ple was rotated to give a take-off angle of 90� a significantly stron-ger Ni peak was observed, although measurements could not beperformed at this angle when the cell contained a liquid sample.

3.3. In situ electrodeposition-XPS measurements

Galvanostatic electrolysis was performed by passing a constantcurrent at a Ni mesh electrode in the electrochemistry-XPS cell

Figure 2. XPS survey scans (pass energy 100 eV, 5 scans) recorded duringelectrolysis at �5 � 10�4 A at a Ni mesh electrode floating on a [C4mpyrr][NTf2]containing 0.1 M Rb[NTf2]. The scan were recorded after 0, 1 and 6 h of electolysis.

Figure 3. (a) XPS high resolution scans (pass energy 20 eV, 30 scans) for the Rb 3dregion taken during electrolysis at �1 � 10�5 A at a Ni mesh electrode floating on[C4mpyrr][NTf2] containing 0.1 M Rb[NTf2]. The scan were recorded after 1, 3, 5, 16,22 and 26 h of electrolysis and (b) Plot of Rb peak area detected at the Ni-IL-vacuuminterface as a function of electrolysis time, as measured by XPS.

R. Wibowo et al. / Chemical Physics Letters 517 (2011) 103–107 105

containing 0.1 M Rb[NTf2] in [C4mpyrr][NTf2].1 The progress of theelectrolysis was monitored periodically by XPS, using both wide sur-vey scans as well as detailed scans in the Rb 3d region.

Figure 2 displays the survey scans recorded before and duringelectrolysis at a fixed current of �5 � 10�4 A. It can be clearlyobserved that after 1 h of electrolysis a significant Rb 3d signalwas present, due to accumulation of Rb by its electrodepositionat the Ni-IL-vacuum three-phase boundary. The signal increasedwhen the electrolysis was extended to 6 h, indicating increasingquantities of Rb metal. In addition to the growth of the Rb features,it was also observed that the stoichiometric ratio of the constituentIL elements (C, N, O, F and S) gradually changed with the atomicpercentage of C increasing from ca. 44% to ca. 69% after 1 h electrol-ysis, indicating significant carbonisation of the ionic liquid wasoccurring in conjunction to Rb electrodeposition. After 6 h the Nfeatures were also observed to merge to form one single feature,rather than the two features expected for the different N presentin the anion and cation [12]. Weingarth have previously monitoredthe reductive decomposition of [C2mim][BF4] at a Pt electrode byXPS, where as an increasingly negative potential was applied aminor N feature (shifted by 2.5 eV relative to the imidazolium N)was observed and was attributed to a decomposition product ofthe cation [11].

A further experiment was performed by decreasing the currentby a factor of 50, to �1 � 10�5 A. Figure 3a shows detailed scansrecorded in the Rb 3d region during electrolysis. From the figureit can clearly be seen that a Rb signal was observed after 1 h ofelectrolysis, and that the Rb signal increased as the electrolysistime was extended. The quantity of Rb detected was found to in-crease in an approximately linear manner with time (Figure 3b),and after 26 h a total of 0.94 C was passed, corresponding toapproximately 0.8 mg of Rb metal deposited (assuming 100% col-umbic efficiency).

After 26 h the electrolysis was ceased and all connections run-ning from the potentiostat to the XPS unplugged. The systemwas allowed to remain like that for 1 h before another XPS mea-surement recorded. There were no significant differences beforeand after the system was allowed to rest for 1 h, indicating thatthe Rb electrodeposit was stable. After this the cell was recon-nected, and the Rb deposit oxidised at an applied potential of 0 Vvs. Pt wire quasi-reference for 30 min. Figure 4 displays the de-tailed XPS scans for Rb 3d taken (a) after 26 h of electrodepositionand (b) after the subsequent application of 0 V for 30 min. The de-crease in the size of the Rb peak corresponds to the oxidation of Rbto Rb[NTf2], some of which diffused away on the timescale of theoxidation.

Detailed analysis of the composition of the IL at the Ni-IL-vac-uum interface was performed by examining the IL signals from arange of survey scans (Figure 5). The corresponding atomic % ofeach IL element obtained from the detailed scans as a function ofelectrolysis time has been displayed in Figure 6, as well as solidlines representing the expected (stoichiometric) values for the ILbased upon. Minor discrepancies between measured and expectedelemental compositions at short times are associated with non-optimised atomic sensitivity factors, as described by Kolbecket al. [20]. The IL composition was observed to stay essentially con-stant within the first 3 h of electrolysis. However, from 5 h on-wards deviation consistent with the carbonisation of the ionic

1 The potentiostatic electrolysis of K from [C4mpyrr][NTf2] with concurrent XPSmeasurement has been previously reported, with the onset of K electrodepositionfrom [C4mpyrr][NTf2] being clearly resolved from the solvent breakdown features [2].In the case of Rb, potentiostatic electrolysis could not be achieved in [C4mpyrr][NTf2]as the difficulties introduced by the similar potentials for Rb nucleation and[C4mpyrr][NTf2] reduction are compounded by the limitations of quasi-referenceelectrodes and the volatility of the ferrocene/ferrocenium reference couple [25].

liquid was observed, likely due to the exhaustion of Rb+ near theelectrode surface after extended electrolysis. The C signal was ob-served in increase from ca. 44% to ca. 58% over 22 h, while the Nsignal was observed to remain constant, S and O decreased slightlyand the F signal decreased significantly. This could be consistentwith the reduction of the pyrrolidinium cation at the Ni electrode(likely through cleavage of an N–C bond to form methylpyrrolidineand a butyl radical or ring opening to a dibutylmethylamine radi-cal, based upon quantum chemical calculations [21]. followed bysubsequent reduction and/or polymerisation of the reduced ter-tiary amine), and concurrent lost of the [NTf2]� anion at the coun-ter electrode. Alternatively, findings by other researchers havepointed towards the possibility of electrochemical reduction ofthe [NTf2]� anion [22–24], with the major gaseous product havingbeen identified as trifluoromethane [23]. The reductive loss of tri-fluoromethane appears to agree with the XPS monitored

Figure 4. XPS high resolution scans for Rb 3d recorded (a) after 26 h ofelectrodeposition and (b) after oxidation of the deposit for 30 min at 0 V.

Figure 5. XPS survey scans (pass energy 100 eV, 5 scans) recorded during electrol-ysis at �1 � 10�5 A at a Ni mesh electrode floating on [C4mpyrr][NTf2] containing0.1 M Rb[NTf2]. The scans were recorded after 0, 1, 3, 5, 16, 22 and 26 h of electrolysis.

Figure 6. The atomic percentages of C, N, O, F and S calculated from the XPS highresolution scan (excluding Rb and Ni signals) recorded during electrolysis at�1 � 10�5 A. The solid line represents the theoretical value of the atomicpercentage. Error bars are associated with the peak fitting and instrumental error.

Figure 7. The atomic percentages of C (s) and Rb (d) calculated from the XPSSpectra recorded during electrolysis at (a) �5 � 10�4 A and (b) �1 � 10�5 Ashowing the relative decreases in IL decomposition at lower current.

106 R. Wibowo et al. / Chemical Physics Letters 517 (2011) 103–107

degradation of the [C4mpyrr][NTf2], namely the rapid loss of F incomparison with the other elements.

In order to show the effect of different currents in the electrode-position of rubidium and the decomposition of the IL, the change inatomic percentage of Rb and C during the electrolysis is compared(Figure 7a and b). At relatively high current (Figure 7a), the trendsshow both deposition and decomposition are prevalent. At lowercurrent (Figure 7b) almost constant C is observed although Rb issignificantly deposited. Thus the lower current promotes deposi-tion over decomposition.

4. Conclusions

The electrodeposition of rubidium has been performed by in situelectrodeposition-XPS employing a novel cell design. The in situelectrochemical-XPS technique allows monitoring of the progressof Rb electrodeposition as well as the competing decompositionof the ionic liquid during the electrodeposition. Passing a highcathodic current for the electrodeposition resulted in rapid decom-position of the ionic liquid in addition to Rb electrodeposition. Onthe other hand, lowering the current for electrodeposition slowsdown the decomposition during the electrodeposition process.

Acknowledgements

Professor Christopher Hardacre is thanked for the kind donationof the ionic liquid, which was synthesised by N.S.A.M. N.S.A.M.

R. Wibowo et al. / Chemical Physics Letters 517 (2011) 103–107 107

acknowledges the financial support of the Ministry of Higher Edu-cation Malaysia and University of Malaya via a SLAI fellowship.R.W. thanks the Directorate of Higher Education, The Ministry ofNational Education, Republic of Indonesia for funding. Mr J. Free-man is thanked for assistance with Figure 1, and Mr. C. Jones forthe construction of the cell and holder.

References

[1] Y. Marcus, Pure Appl. Chem. 57 (1985) 1129.[2] R. Wibowo, L. Aldous, S.E.W. Jones, R.G. Compton, Chem. Phys. Lett. 492 (2010)

276.[3] D.V. Pawar, T.S. Magdum, S.P. Rasal, S.H. Chavan, Bull. Electrochem. 11 (1995)

294.[4] G. Angelino, C. Invernizzi, Sol. Energy 82 (2008) 637.[5] F.P.D. Endres, D. MacFarlane, A. Abbott, Electrodeposition from Ionic Liquids,

Wiley-VCH, Weinheim; Chichester, 2008.[6] S.Z. El Abedin, F. Endres, ChemPhysChem 7 (2006) 58.[7] F.L. Qiu, A.W. Taylor, S. Men, I.J. Villar-Garcia, P. Licence, Phys. Chem. Chem.

Phys. 12 (2010) 1982.[8] A.W. Taylor, F.L. Qiu, I.J. Villar-Garcia, Chem. Commun. (2009) 5817.[9] A. Foelske-Schmitz, D. Weingarth, H. Kaiser, R. Kotz, Electrochem. Commun. 12

(2010) 1453.[10] I.J. Villar-Garcia, E.F. Smith, A.W. Taylor, F.L. Qiu, K.R.J. Lovelock, R.G. Jones, P.

Licence, Phys. Chem. Chem. Phys. 13 (2011) 2797.

[11] D. Weingarth, A. Foelske-Schmitz, A. Wokaun, R. Kotz, Electrochem. Commun.13 (2011) 619.

[12] K.R.J. Lovelock, I.J. Villar-Garcia, F. Maier, H.P. Steinruck, P. Licence, Chem. Rev.110 (2010) 5158.

[13] R. Wibowo, L. Aldous, E.I. Rogers, S.E.W. Jones, R.G. Compton, J. Phys. Chem. C114 (2010) 3618.

[14] R. Wibowo, S.E.W. Jones, R.G. Compton, J. Phys. Chem. B 113 (2009) 12293.[15] R. Wibowo, S.E.W. Jones, R.G. Compton, J. Chem. Eng. Data 55 (2010) 1374.[16] R. Wibowo, L. Aldous, R.M.J. Jacobs, N.S.A. Manan, R.G. Compton, Chem. Phys.

Lett. 509 (2011) 72.[17] D.R. MacFarlane, P. Meakin, J. Sun, N. Amini, M. Forsyth, J. Phys. Chem. B 103

(1999) 4164.[18] C.D. Wagner, L.E. Davis, M.V. Zeller, J.A. Taylor, R.H. Raymond, L.H. Gale, Surf.

Interface Anal. 3 (1981) 211.[19] F. Maier et al., Angew. Chem.-Int. Ed. 45 (2006) 7778.[20] C. Kolbeck, M. Killian, F. Maier, N. Paape, P. Wasserscheid, H.P. Steinruck,

Langmuir 24 (2008) 9500.[21] M.C. Kroon, W. Buijs, C.J. Peters, G.J. Witkamp, Green Chem. 8 (2006) 241.[22] P.C. Howlett, E.I. Izgorodina, M. Forsyth, D.R. MacFarlane, Z. Phys. Chem.-Int. J.

Res. Phys. Chem. Chem. Phys. 220 (2006) 1483.[23] E. Markevich et al., Electrochim. Acta 55 (2010) 2687.[24] S.P. Ong, O. Andreussi, Y.B. Wu, N. Marzari, G. Ceder, Chem. Mater. 23 (2011)

2979.[25] C.P. Fu, L. Aldous, E.J.F. Dickinson, N.S.A. Manan, R.G. Compton, Chem.

Commun. 47 (2011) 7083.