Laser Desorption/Ionization Using Surface of dispersed Single Wall Carbon Nanotubes∗

We showed that our new single wall carbon nanotubes substrate (SWCNTs substrate) with a nano-sized network structure on its all surface areas was very efficient for high performance ionization of analyte molecules. A preparation procedure of our SWCNTs substrate requires mainly two processes that are a well dispersion of SWCNTs in preparing solution and a very thin laying-up on a surface of base plate. Moreover, we found that base plates for laying-up of SWCNTs determines its analytical performance and a SUS plate is a good base plate. We expect that further investigation of base plates in our SALDI-MS technique using the SWCNTs leads to more efficient ionization. [DOI: 10.1380/ejssnt.2009.546]


I. INTRODUCTION
Carbon nanotubes (CNTs) have attracted considerable attention since the discovery in 1991 [1].Although CNTs have been investigated their unique physical properties in many researches, all of their properties have not been revealed yet.The CNTs have been applied in many fields due to their novel structural, electronic, and mechanical properties.
Matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) is a soft LDI-MS method which can detect sample molecules with no fragmentation, and very effective for structural characterization of organic molecules such as synthetic polymers [2][3][4] and biological samples [5].This technique is a powerful analytical tool and has been used in many fields.In the MALDI-MS measurements, analyte molecules must be mixed with organic matrix molecules which have large absorbance at a wavelength of laser irradiation (typically for a 337 nm nitrogen laser).Although the MALDI-MS has been remarkably successful as a soft ionization technique for these molecules, this ionization technique has several weak points due to this mixing procedure.One is unsuitability for a detection of low molecular compounds because of interference background peaks caused from matrix reagents in the low-mass range.Other is that selection of organic matrixes and sample preparation methods influence results of the LDI measurements.
To avoid these problems, several soft laser desorption/ionization techniques without using organic matrix molecules have been developed.One of these techniques, which utilizes surface effects generated from various substrates, is generally classified surface assisted LDI-MS (SALDI-MS).In the SALDI-MS technique, analyte ions are desorbed directly from the SALDI substrate surface which has a huge specific surface area.Therefore, the SALDI technique does not require the entrainment action of added matrix molecules for desorption.Recently, various inorganic materials such as nanoparticles [6][7][8], and porous silicon [9][10][11] have been proposed as these SALDI-MS substrates.
In this study, we focused on single wall carbon nanotubes (SWCNTs).Considering their size, structure and huge specific surface areas, it is presumed easily that their surface of the SWCNTs also has distinct characteristics.In our previous study, we demonstrated that the SALDI-MS technique employing SWCNTs was effective for irradiating laser pulses with several kinds of wavelengths of 532 nm and 1064 nm, due to their efficient absorption of laser energy in wide wavelength region from ultraviolet (UV) to near-infrared (IR) [12].This SALDI-MS technique has a possibility to enable soft desorption/ionization of various kinds of analyte species by eliminating overlaps between their absorption wavelength region and wavelength of irradiating laser.Although our SWCNTs substrate indicated some advantages for SALDI-MS, however, it was not been discussed sufficiently regarding this ionization technique.In order to improve our SALDI-MS substrate, we attempted to compare preparation procedure and kinds of base plates for SALDI-MS using SWCNTs.

Preparation and characterization of SALDI-MS substrate
In order to disperse SWCNTs, three kinds of solvents, ethanol, cyclohexane, THF, were used.The SWCNTs were suspended in each solution and sonicated until they were dispersed.Each SWCNTs suspension was deposited on base plates, SUS and silicon, respectively.A surface of SWCNTs substrate was observed by scanning electronic microscopy (SEM, Keyence, VE-9800).

Instruments
The mass spectrometric measurements were performed using two time-of-flight (TOF) mass spectrometer.One is a MALDI-TOF mass spectrometer (Shimadzu; Kratos Axima CFR) equipped with a nitrogen laser operating at a wavelength of 337 nm.The other is a laboratory-built linear TOF mass spectrometer equipped with Nd:YAG laser (second harmonic wavelength of 532 nm).In this laboratory-built linear TOF-MS, the pulse width of the laser was 3 ns, and the acceleration voltage for ions was 4 kV and mass spectra were obtained by the positive ion mode.The mass spectra were acquired by accumulating 128 single-shot mass spectra.
For mass spectrometric measurements, the PEG samples (1.0 mg) were dissolved in 1.0 ml methanol.Sample solution (0.50 µl) was deposited on the target plates and dried in the air.

III. RESULTS AND DISCUSSION
The SWCNTs were dispersed in three kinds of solvents and then these suspensions were deposited on each SUS plates.SALDI mass spectra of PEG 1000 on these substrates were observed by irradiating of UV laser (337 nm), as shown in Fig. 1.As shown in Fig. 1(a), the SALDI-MS substrate utilizing the SWCNTs dispersed in ethanol showed highest analytical performance among three substrates.This mass spectrum of PEG 1000 was observed with good intensity despite absence of matrix molecules, and the mass distribution was clearly observed similar to standard distribution obtained by conventional MALDI-MS technique [4].Each peak of this mass distribution were separated by 44 m/z which was the mass of PEG repeating unit (-CH 2 CH 2 O-) and a series of main peaks was the same showed that the surface morphology of the SWCNTs-Si-substrate was similar to that of Although the SWCNTs-seat is composed of thick SWCNTs bundle structure, the SWCNTs-seat e on its all surface (Fig. 2 .When the SWCNTs were dispersed in these solvents under the same condition, the ethanol solution showed fine dispersion of SWCNTs, whereas the cyclohexane solution was not capable of dispersing them well.This difference between solvents might be due to their polarity though the actual factor was not clear at this stage.These experimental results suggest that the dispersion of SWCNTs is associated with the observed SALDI mass spectra.Thus, we decided to select ethanol as an appropriate solvent to the SWCNTs substrate. Our laboratory-built linear TOF mass spectrometer is capable of varying target plates readily in LDI-MS measurements.We attempted to perform SALDI-MS analysis with three kinds of target plates; SUS plate and silicon plate with dispersed SWCNTs on its surface, SWCNTs-SUS-substrate and SWCNTs-Si-substrate, and SWCNTsseat. Figure 2 shows SEM images of surfaces of the SWCNTs-Si-substrate and the SWCNTs-seat.As shown in Fig. 2(a), the same showed that the surface morphology of the SWCNTs-Si-substrate was similar to that of the SWCNTs-SUS-substrate [12].Although the SWCNTsseat is composed of thick SWCNTs bundle structure, the SWCNTs-seat has a nano-sized network structure on its all surface (Fig. 2(b)).That is, the SWCNTs-seat has a larger surface area occupied by the SWCNTs bundles than that of each SWCNTs substrate (SWCNTs-SUS- substrate and SWCNTs-Si-substrate).
Figure 3 shows the SALDI mass spectra of PEG 1000 deposited on each substrate by irradiating of visible laser (532 nm).Both spectra of PEG 1000 on the SWCNTs-SUS-substrate (Fig. 3(a)) and the SWCNTs-Si-substrate (Fig. 3(b)) were obtained with sufficient sensitivity and a series of ion peaks separated by 44 units was observed.For comparison, although the sample solution was deposited onto flat silicon and SUS plates respectively, any ion peaks of PEG 1000 could not be detected in the LDI-MS measurements with same condition (Figs.3(d)(e)).This observation indicates that amount of the ion caused by single-or multi-photon absorption process was below detection limit and therefore contribution of single-and multi-photon absorption process can be excluded in this experimental condition.The SWCNTs-Si-substrate, thus, would have an effect as a SALDI-MS substrate similar to the SWCNTs-SUS-substrate.However, the observed SALDI mass distribution of PEG 1000 by the SWCNTs-Si-substrate showed that its shape was dissymmetric and the maximal point appeared at the low mass region.This difference of the observed mass distribution by each SWC-NTs substrate shows that the observed SALDI mass spectra would be affected by base plates (SUS and silicon plate).From this experimental result, we speculated that the absorbed energy by the SWCNTs was diffused and then transmitted to the base plate, thus, the difference of the observed mass spectra was occurred by base plates.
In contrast, the SALDI-MS with the SWCNTs-seat could not generate the ion peaks of PEG 1000 in this experimental condition, despite its wide superficial area including nano-sized roughness (Fig. 3(c)).The network structure of the SWCNTs-seat has not only wide superficial area including nano-sized roughness, but also large thickness.We deduced that the defect of the SWCNTsseat was attributed to diffusion of absorbed energy from an irradiation source internally similar to bulk metal.It is reported that bulk metal can not take up the energy internally due to dispersion of absorbed energy, by Tanaka et al. who have pioneered the SALDI-MS research using cobalt particles [13].Thus, since the energy diffusion prevents to transfer absorbed energy to analyte molecules efficiently, the ion peaks of PEG 1000 would not be obtained in the SALDI-MS measurement with the SWCNTs-seat.
These results suggest that it is necessary for an effective SALDI substrate to be capable of staying the energy from laser irradiation at its surface without the diffusion and loss of the energy.Therefore, the SWCNTs substrate has a possibility of becoming more effective SALDI substrate by dispersing on a surface with very thinly owing to avoiding the diffusion and loss of the energy.

IV. CONCLUSION
We showed that the SWCNTs substrate requires mainly two processes that are a well dispersion of SWCNTs in preparing solution and a very thin laying-up on a surface of plate.SWCNTs-SUS-substrate showed higher analytical performance as a SALDI-MS substrate than SWCNTs-seat composed of thick SWCNTs bundles and a nano-sized network structure on its all surface areas.We concluded that it is necessary for an effective SALDI substrate to be capable of staying the energy at its surface without the diffusion and loss of the energy.We expect that the SWCNTs substrate would become more efficiently by dispersing on a surface of plate as possible as thinly owing to avoiding the diffusion and loss of energy.

Figure 3
Figure 3 shows the SALDI mass spectra of PEG 1000 rradiating of visible laser (532 nm).Both spectra of PEG 1000 on the SWCNTs-SUS-substrate substrate (Fig. 3(b)) were obtained with sufficient sensitivity and a series of ion peaks separated by 44 unit was observed.For comparison, although the sample solution was deposited onto flat silicon and SUS plates respectively, any ion peaks of PEG 1000 could not be FIG.2: SEM images of the surface of (a) SWCNTs-Sisubstrate at magnification of 30 000, and (b) SWCNTs-seat at magnification of 15 000.