Development of an Ionic-liquid Ion Beam Source for Secondary Ion Mass Spectrometry (SIMS)∗

Beam characteristics generated by vacuum electrospray of ionic liquids were investigated at gas pressures of 10−5 Pa. An imidazolium ionic liquid, 1-ethyl-3-methyl imidazolium bis(trifluoromethanesulfonyl) amide (EMI-TFSA), and an ammonium ionic liquid, N,N-diethyl-N-methyl-N-(2-methoxyethyl)ammonium bis(trifluoromethanesulfonyl) amide (DEME-TFSA), were tested. Transient response measurements showed that m/z values of EMI-TFSA charged droplets were smaller than those of DEME-TFSA charged droplets under the same experimental conditions. The m/z values of EMI-TFSA charged droplets decreased with decreasing flow rate. A target current exceeding 30 nA was produced at a capillary voltage of 1.6 kV even with a low flow rate of 3 nL/min, at which the peak value of m/z distribution was 7.9×10. The average mass, charge number and diameter of the charged droplets at m/z = 7.9×10 were estimated to be about 6.1×10 u, 77 and 23 nm, respectively. [DOI: 10.1380/ejssnt.2014.119]

The glycerol and aqueous charged droplets are generated by electrospraying glycerol and aqueous solutions, respectively [17,18].Mahoney et al. developed a novel ionization method termed massive cluster impact (MCI) using vacuum electrospray of glycerol solutions.They reported that the bombardment of highly charged glycerol droplets effectively desorbed biomolecular ions from surfaces [17,19,20].Unfortunately, however, the ion signals resulting from MCI become unstable after several hours of operation because of the charging of lens electrodes contaminated by involatile glycerol [21,22].
To avoid this issue, Hiraoka et al. developed a new contamination-free ionization method called electrospray droplet impact ionization (EDI), which employs atmospheric pressure electrospraying of volatile aqueous solu-tions [18].They observed intense signals for protonated drug molecules, amino acids, peptides, and proteins [22][23][24].Additionally, EDI allows for shallow-surface etching with little sample damage, so that EDI is expected to be applicable to the characterization of organic and inorganic multilayer systems with high-depth resolution [25].Unfortunately, however, its beam is so broad that it is of limited usefulness for imaging experiments [18,22].In EDI, charged particles emitted at atmospheric pressure are sampled through an aperture into a vacuum chamber and then accelerated toward a target.At atmospheric pressure or in a low-vacuum environment, charged particles and gas molecules undergo many collisions, resulting in the scattering of ion beams; consequently, it is difficult to produce a focused ion beam.
In contrast, vacuum electrospray would likely produce a focused beam, since charged droplets emitted in a highvacuum environment can be focused using an electrostatic lens.For volatile liquids such as water and ethanol, however, it is not easy to achieve stable electrospraying under high-vacuum conditions, since evaporation and freezing are inevitable [26,27].To avoid freezing, Ninomiya et al. demonstrated that continuous infrared laser irradiation is useful in vacuum electrospray of volatile liquids [28,29].
Another approach to avoid evaporation and freezing is use of a liquid with very low volatility.Such a liquid must also have high conductivity to prevent detrimental charging resulting from its accumulation on electrodes.For these reasons, we started research and development of a vacuum-electrospray beam source using an ionic liquid that has negligible vapor pressure and high ionic conductivity [30,31].
Ionic liquids [C + A − ] are room-temperature molten salts consisting of a polyatomic cation [C + ] and a polyatomic anion [A − ] [32].Since ionic liquids have low vapor pressures and high ionic conductivities, it is possible to continuously electrospray them in vacuum.With regard to vacuum electrospray of ionic liquids, Gamero-Castano et al. reported that sputtering and amorphization of silicon were successfully performed with an ionic-liquid beam [33][34][35].Also, ionic liquids have been shown to be effective as matrices in both SIMS and matrix-assisted laser desorption ionization (MALDI) [36][37][38].
As reported in previous papers [39][40][41][42], we had investigated the electrospray characteristics of a quaternary ammonium ionic liquid, N,N-diethyl-N-methyl-N-(2methoxyethyl)ammonium bis(trifluoromethanesulfonyl) amide (DEME-TFSA), thereby demonstrating that stable currents can be continuously generated.For a primary ion beam in SIMS, the chemical compositions of ionic liquids should be considered.This is because the chemical compositions of an ionic-liquid beam can affect secondary ion yields, which is of particular importance in SIMS.Hence, in the present study, we tested an imidazolium ionic liquid in addition to the ammonium ionic liquid (DEME-TFSA).As an imidazolium ionic liquid, we used 1-ethyl-3-methyl imidazolium bis(trifluoromethanesulfonyl) amide (EMI-TFSA).Using EMI-TFSA, we investigated the transient response of an ionic-liquid beam current, thereby evaluating the m/z distributions of charged droplets in the beam.We also examined the mass, charge number and diameter of the charged droplets.

II. EXPERIMENTAL
Figure 1 shows a schematic illustration of the experimental apparatus for measuring beam characteristics.Details of the experimental apparatus have been reported previously [42].Briefly, the apparatus consists of a stainless-steel capillary, a three-dimensional positioning stage, three electrodes, a beam target, a vacuum chamber, current detectors, and high-voltage power supplies.The capillary (Eisyo Metal, MTS-70) has an inner diameter of 30 µm.One end of the capillary was installed inside the vacuum chamber.A positive voltage was applied to the capillary with a DC power supply.The electrodes are stainless-steel circular disks of 90 mm in outer diameter.The first electrode has a 3-mm-diameter central hole (i.e., aperture) so that charged droplets emitted from the capillary may pass through it.The gap length between the tip of the capillary and the first electrode was about 5 mm.The second electrode is a beam deflector.It consists of parallel plates of 20 mm length and width, with a 10-mm gap between the plates.One of the plates was electrically grounded.The other was connected to a positive

III. RESULTS AND DISCUSSION
Figure 3 shows the transient responses of the target currents, I(t), before and after applying a pulse voltage of +1 kV to the second electrode.The EMI-TFSA and DEME-TFSA were electrosprayed continuously at a capillary voltage of +1.6 kV and a flow rate of 50 nL/min.The pulse voltage was periodically applied at 100 Hz for a duration of 2000 µs.The periodic I(t) signals were averaged through 128 cycles in the oscilloscope.The I(t) curves provide information about the velocity distribution of charged species in the electrosprayed beam.After a pulse voltage of +1 kV was applied at time t = −2000 µs, the target current decreased with time, becoming stable at I = 0.The decrease at EMI-TFSA was faster than that at DEME-TFSA.At time t = 0, the second electrode was rapidly grounded, and the target current subsequently recovered.Also, the recovery at EMI-TFSA was faster than that at DEME-TFSA.The faster decrease and recovery indicate that the EMI-TFSA beam consisted of charged droplets with smaller m/z values than the DEME-TFSA beam under the same electrospray conditions.were faster than those at 5 and 50 nL/min.This means that the m/z values of charged droplets decreased with decreasing flow rate.
The transient response of the target current was analyzed by a time of flight (TOF) technique.In general, the mass-to-charge ratio (m/z) is calculated from the following equation: where e is the elementary charge (1.6022×10 −19 C), V the acceleration voltage, t the time of flight, u the unified atomic mass (1.66054×10 −27 kg), and L the flight length.
On the basis of Eq. ( 1), we converted the time after t = 0 to an m/z value, thus roughly evaluating m/z values of charged species.The transient response after t = 0 corresponds to the recovery of the target current.In the present study, the decrease of the target current after t = −2000 µs was not analyzed.This is because the decrease of the target current would be somewhat distorted by the pulse voltage of +1 kV applied to the second electrode; that is to say, the electric field caused by the pulse voltage would push charged species in the beam to the target.The acceleration voltage V was assumed to be 1.6 kV, which corresponds to the voltage applied to the capillary.The flight length L was assumed to be 0.35 m, which corresponds to the distance between the first electrode and the target.Figure 5 shows difference values of the transient target current, ∆I, as a function of m/z.The ∆I was evaluated at a time interval of 50 µs.It can be seen that the m/z values obtained at 3 nL/min were smaller than those at 5 and 50 nL/min.At 3 nL/min, the highest peak was observed at m/z = 7.9×10 4 , which corresponds to a gas cluster of Ar + 1975 .The maximum stable charge of a droplet is given by the Rayleigh equation [43,44]; the charge number z can be represented as [18,23]  where γ is the surface tension of the liquid, ε 0 the dielectric permittivity of the vacuum (8.85×10 −12 C 2 J −1 m −1 ), and R the radius of the droplet.It is known that initial droplets emitted from a capillary tip are charged to around 70% of the Rayleigh limit on average [45,46].Hence, Eq. ( 2) can be rewritten as The surface tension and density of EMI-TFSA are 34.9×10−3 Jm −2 and 1.52 gcm −3 , respectively [47,48].On the basis of Eq. ( 3), we calculated the relationship between the mass and charge number of a droplet.In the calculation, the density of the droplet was assumed to be the same as that of the ionic liquid.
Figure 6 shows the relationship between the mass and charge number of the charged droplets at m/z = 7.9×10 4  The curved line is the calculated result that corresponds to 70% of the Rayleigh limit.The straight line corresponds to the experimental result, which is the peak m/z value at 3 nL/min.The intersection of the curved line with the straight line provides an estimated value of mass and charge number of the droplets at m/z = 7.9×10 4 .The mass and charge number were estimated to be about 6.1×10 6 u and 77, respectively.The corresponding droplet may be expressed as [(CA) 15,500 + 77C] 77+ , where C and A denote a cation and an anion of EMI-TFSA.The diameter of the droplet was estimated to be about 23 nm.
Note that actual values of the masses and charge numbers of generated droplets will have distributions.Since the above values of the mass and charge number were roughly estimated on the basis of the peak m/z value and the average charge state of droplets (Eq.( 3)), these estimated values should serve just as guidelines for true distribution of mass and charge number.
Figure 7 shows a typical time dependence of the target current when EMI-TFSA was electrosprayed at 1.6 kV at a constant flow rate of 3 nL/min.A relatively stable FIG.7: Time dependence of the target current in a continuous operation when EMI-TFSA was electrosprayed at 1.6 kV and a constant flow rate of 3 nL/min.A stable current exceeding 30 nA was generated continuously.
target current exceeding 30 nA was generated for 30 min.In the present ion source, the duration of continuous beam generation depends on both the volume of an ionic liquid inside the syringe and the flow rate of the ionic liquid.In the present experiment, a 100 µL syringe was used and the flow rate was 3 nL/min.Consequently, the maximum duration of beam generation was about 550 hours, which seems enough for SIMS analysis.

IV. SUMMARY
To develop a vacuum-electrospray beam source for secondary ion mass spectrometry (SIMS), we studied the beam characteristics generated by vacuum electrospray of an imidazolium ionic liquid, 1-ethyl-3-methyl imidazolium bis(trifluoromethanesulfonyl) amide (EMI-TFSA), as well as an ammonium ionic liquid, N,N-diethyl-N-methyl-N-(2methoxyethyl)ammonium bis(trifluoromethanesulfonyl) amide (DEME-TFSA).We electrosprayed the ionic liquids at gas pressures of 10 −5 Pa and measured the transient response of the target current.Transient response analysis enabled us to study the m/z distribution of charged species in the beam, thereby showing that m/z values of EMI-TFSA charged droplets were smaller than those of DEME-TFSA.We also confirmed that the m/z values of EMI-TFSA charged droplets diminished with decreasing flow rate.Also, we demonstrated that a stable target current exceeding 30 nA were continuously generated even at a low flow rate of 3 nL/min.
On the basis of the obtained m/z value of 7.9×10 4 at 3 nL/min, we estimated the average mass and charge number of the charged droplets.The average mass and charge number were estimated to be about 6.1×10 6 u and 77, respectively.The diameter of the corresponding droplet was estimated to be about 23 nm.
In terms of application as a primary ion beam source for SIMS, higher beam energy and smaller beam size will be required in addition to mass filtering of generated charged particles.To accelerate and focus mass-separated ion beams, we plan to fabricate a vacuum-electrospray ion source equipped with a mass filter, sufficient insulating parts to maintain at least 10 kV, and an Einzel lens for focusing.The performance of the beam source will be the

FIG. 1 :
FIG.1: Schematic illustration of the experimental apparatus for measuring the beam characteristics generated by vacuum electrospray of the ionic liquid.

Figure 2 FIG. 3 :
FIG.3: Transient responses of the target currents before and after applying a pulse voltage of +1 kV to the second electrode.Electrospraying was continuously performed at a capillary voltage of +1.6 kV with a flow rate of 50 nL/min.

Figure 4
Figure3shows the transient responses of the target currents, I(t), before and after applying a pulse voltage of +1 kV to the second electrode.The EMI-TFSA and DEME-TFSA were electrosprayed continuously at a capillary voltage of +1.6 kV and a flow rate of 50 nL/min.The pulse voltage was periodically applied at 100 Hz for a duration of 2000 µs.The periodic I(t) signals were averaged through 128 cycles in the oscilloscope.The I(t) curves provide information about the velocity distribution of charged species in the electrosprayed beam.After a pulse voltage of +1 kV was applied at time t = −2000 µs, the target current decreased with time, becoming stable at I = 0.The decrease at EMI-TFSA was faster than that at DEME-TFSA.At time t = 0, the second electrode was rapidly grounded, and the target current subsequently recovered.Also, the recovery at EMI-TFSA was faster than that at DEME-TFSA.The faster decrease and recovery indicate that the EMI-TFSA beam consisted of charged droplets with smaller m/z values than the DEME-TFSA beam under the same electrospray conditions.Figure4shows transient responses of I(t) when EMI-TFSA was electrosprayed continuously at a capillary voltage of +1.6 kV and flow rates of 3, 5 and 50 nL/min.The figure shows that the decrease and recovery at 3 nL/min

FIG. 6 :
FIG.6: Relationship between the mass and charge number of the charged droplets.The curved line corresponds to 70% of the Rayleigh limit.The straight line corresponds to the peak m/z value of 7.9×10 4 at 3 nL/min.The intersection of the curved line with the straight line provides estimated value of the mass and charge number of the droplets.