ISSN 1061-9348, Journal of Analytical Chemistry, 2016, Vol. 71, No. 13, pp. 1221–1227. © Pleiades Publishing, Ltd., 2016. Original Russian Text © D.S. Kosyakov, E.A. Sorokina, N.V. Ul’yanovskii, E.A. Varakin, D.G. Chukhchin, N.S. Gorbova, 2015, published in Mass-spektrometriya, 2015, Vol. 12, No. 3, pp. 169–176.
ARTICLES
Carbon Nanocoatings: A New Approach to Recording Mass Spectra of Low-Molecular Compounds Using Surface-Assisted Laser Desorption/Ionization Mass Spectrometry D. S. Kosyakov*, E. A. Sorokina, N. V. Ul’yanovskii, E. A. Varakin, D. G. Chukhchin, and N. S. Gorbova Northern (Arctic) Federal University named after M.V. Lomonosov, nab. Severnoi Dviny 17, Arkhangelsk, 163002 Russia *e-mail:
[email protected] Received April 1, 2015; in final form, April 16, 2015
Abstract—A simple and rapid approach to obtaining target plates for the investigation of low-molecularweight compounds by surface-assisted laser desorption/ionization (SALDI) mass spectrometry is proposed. It consists in the vacuum sputtering of a carbon layer with a thickness of about 50 nm onto a metal surface. The resulting coatings are characterized by homogeneity, hydrophobicity, and high mechanical strength, which eliminates a possibility of mass spectrometer contamination. A comparison of the SALDI mass spectra of test compounds recorded using conventional carbon materials and carbon nanocoatings demonstrates advantages of the last named materials, such as high spectral resolution and the absence of spectral interferences at low m/z values. Keywords: mass spectrometry of small molecules, carbon, graphite-assisted laser desorption/ionization, surface-assisted laser desorption/ionization, sample preparation, nanocoatings DOI: 10.1134/S1061934816130086
INTRODUCTION Matrix-assisted laser desorption ionization (MALDI) mass spectrometry is an efficient method for the study of natural and synthetic polymers because of mild ionization conditions, generation of mainly singly charged ions, and, in using time-offlight analyzers, exclusively wide ranges of measured masses. Important advantages of the method, such as rapidity and tolerance to impurities, make it also highly attractive for work with low-molecular compounds. In spite of this, MALDI mass spectrometry of small molecules has still remained a rare method for certain reasons [1]. The most important one is the presence of matrix interferences in the recorded mass spectra, usually in the region m/z below 1000. To solve the above problem, analysts use approaches such as laser desorption/ionization of the studied substances directly from metal target plates (it can be applied to a very limited range of compounds), search for and application of new matrixes with the minimum numbers of peaks in the mass spectra [2], use of matrixes with high molecular weights, and also the inclusion of traditional matrixes into complexes, for example, with cyclodextrine [3]. The direction assuming the use of surface assisted laser desorption/ionization (SALDI) [1, 4] seems to be more fruit-
ful. In the latter approach, target plates are usually made of materials with nanostructured surfaces, capable of efficiently absorbing the energy of laser radiation and transferring it to analyte molecules. Graphite-assisted laser desorption/ionization (GALDI) is one of the most successful and promising SALDI versions [1]. Along with the low cost and availability of high-purity preparations, graphite possesses some other important advantages. These are intense absorbance in a wide wavelength range and the simplicity of sample preparation. In early works, suspensions of graphite microparticles (2–150 μm) in liquids mixed with the studied compounds [5] were used for this purpose and the fabrication of targets directly from graphite plates was proposed [6]. Later it was found that the best results in the study of small molecules can be obtained by applying a graphite coating onto a metal target plate using a graphite pencil [7]. This method also proved itself well in combinations of MALDI mass spectrometry with thin-layer chromatography. Putting graphite directly onto the surface of chromatographic zones allowed Borisov et al. to identify individual components without preliminary extraction and to avoid the smearing of spots [8, 9]. The efficiency of carbon in the laser desorption/ionization of analytes is substantially determined
1221
1222
KOSYAKOV et al.
by the particle size of the inorganic matrix. In this connection, a number of works on the use of carbon nanomaterials as MALDI matrixes is known [10]. Xu with coauthors first proposed the use of carbon nanotubes (CNT), capable of efficiently capturing analyte molecules and ensuring their desorption under laser irradiation without significant fragmentation for recording mass spectra of small molecules [11]. Carbon nanotubes have demonstrated advantages, such as rather low thresholds of the intensity of laser radiation for the desorption/ionization of low-molecular analytes, good reproducibility of the recorded mass spectra, and better resolution in comparison to organic matrixes. Various approaches to the optimization of the process of sample preparation using CNT, including also the immobilization of nanotubes on the surface [12] and their chemical modification (carboxylation) [13] have also been developed by the present time. Among the disadvantages of graphite and other carbon matrixes are problems with the uniform application of coatings onto the target plate surface, presence of signals from carbon clusters in a wide range of m/z, and, which is very important because of the technical features of the mass spectrometer, possibility of the deposition of carbon in the ion source with its rapid contamination [1]. In our opinion, the specified problems can be solved by applying ultrathin uniform graphite coatings, strongly bound to the surface of a GALDI target plate and containing extremely small amounts of carbon, instead of individual particles. The results presented in [14] and demonstrating the high efficiency of graphene as a substrate for GALDI favor this approach, in spite of a decrease in the sensitivity of the method because of the formation of extremely thin carbon layers on the target plate surface. Nevertheless, we still found no data in the literature on the application of technologies of the creation of durable carbon coatings for use in MALDI mass spectrometry. The only exception is provided by publication [15], the authors of which proposed the formation of nanolayers of tetrahedral amorphous carbon, several hundreds of nanometers in thickness, on the surface of a disposable polymeric target plates using a laser. The samples obtained were used for recording spectra of peptides and proteins using an organic matrix. In this connection, of great interest is the development of a simple and available laboratory method for preparing target plates with durable carbon coatings suitable for use in the mass spectrometry of small molecules. This method can be based on the technology of vacuum deposition and widely used to prepare nanodimensional conducting sample coatings in scanning electron microscopy and X-ray microanalysis [16]. Its optimization for use in MALDI mass spectrometry, assessment of possibilities for recording mass spectra of compounds of different classes, and comparison
with the known approaches were the goals of this study. EXPERIMENTAL Test compounds. To test and compare the efficiencies of different types of carbon matrixes (coatings), we selected a number of natural compounds from different classes with molecular weights in the range 170– 450 Da, and also polyethyleneglycol with an average molecular weight of about 600 Da (Table 1). All analytes were used without additional purification. Reagents and materials. To obtain carbon matrixes (coatings), we used rods of spectrally pure graphite (99.995%, Aldrich, Germany), soft graphite pencils (8B, Koh-i-Noor Hardtmuth, Czech Republic), and also multilayered carbon nanotubes (6–13 nm × 2.5– 20 μm, >98%, Aldrich, Germany). Solutions were prepared using acetonitrile (LC-MS grade, Merck, Germany); methanol (HPLC grade Lichrosolv, Merck, Germany); trifluoroacetic acid (≥99%, Sigma-Aldrich, United States); and also deionized water with a specific resistance of 18.2 MΩ cm, prepared with a Simplicity UV system (Millipore, France). Equipment. Mass spectra were recorded in the reflectron mode using an Axima Performance tandem time-of-flight (TOF-TOF) MALDI mass spectrometer (Shimadzu-Biotech, Great Britain). A nitrogen UV laser (λ = 337 nm) with pulse duration of 3 ns and energy of 130 μJ was the source of radiation. Spectra were recorded by accumulating data of 200 laser shots (100 profiles each of 2 shots) collected from different points in the target plate. The maximum frequency of shots was 50 Hz. In all experiments, we used time-lag focusing, the delay time of the extraction of ions from the source was optimized for m/z values characteristic for the studied compounds. To reduce possible interferences with the matrix and also impurities with low m/z values, we used the cutoff of ions with small masses (up to m/z = 300 and 100 for arginine). All studies were performed in the positive ion mode. The control of the mass spectrometer and data collection and processing were performed using the Launchpad 2.9. software package (Shimadzu–Biotech, Great Britain). Mass spectra were processes using the mMass 5.5.0 software (©M. Strohalm) [17]. Carbon nanocoatings (CNC) were obtained using a Q150T ES deep vacuum sputtering system (Quorum Technologies, Great Britain), equipped with a rotating sample holder, and also accessories for the thermal evaporation of carbon from filaments and graphite rods. Microscopic studies of target plates with applied samples were conducted with an Axio Imager M2m optical microscope, a Zeiss Sigma VP scanning electron microscope (Carl Zeiss, Germany), and a MultiMode 8 atomic-force microscope (Bruker, Germany).
JOURNAL OF ANALYTICAL CHEMISTRY
Vol. 71
No. 13
2016
CARBON NANOCOATINGS
1223
Table 1. Studied compounds. No.
Compound
Monoisotopic mass, Da
Structural formula
Characteristics of preparation
NH2 1
L-arginine
O
NH
H2N
174.11
>95%, Sigma-Aldrich (Japan)
302.04
>95%, Sigma-Aldrich (India)
339.15
>98%, Sigma (Germany)
342.12
>98%, Sigma (Germany)
442.38
>98%, Aldrich (China)
OH
NH
OH OH 2
Quercetin
O
HO
OH OH O O O 3
Papaverine
O N
O OH
OH O 4
O
Cellobiose
HO
O OH
OH HO OH
OH
H 5
OH
Betulin
H H HO
6
Polyethylene glycol (PEG-600)
H H
O
44.03n + 18.01
n OH
Preparation of target plates and application of samples. In all experiments, we used standard target plates made of unpolished stainless steel as microtiter plates, each containing 384 sample wells (Kratos, Great Britain). Graphite coatings were applied using pure graphite rods (G) and soft graphite pencils (GP), by painting over a well to obtain a completely uniform coating of the surface, avoiding the formation of loosing particles. Carbon nanotubes were predispersed in a water– methanol mixture (1 : 1) under ultrasonic treatment for 5 min, thus obtaining a stable colloidal solution with the concentration 2 mg/mL. Then we put 1 μL of JOURNAL OF ANALYTICAL CHEMISTRY
Vol. 71
Mn = 570–630, Bioultra, Sigma (Germany)
the obtained solution onto the well surface and dried in air. Test compound solutions with concentrations of 1 mg/mL were prepared in a mixture of water with acetonitrile (1 : 1), to which 0.1% of trifluoroacetic acid was added. A 0.5-μL portion of the obtained solution was applied onto the prepared target plate and dried in air. Preparation of carbon nanocoatings. To obtain CNC, we used the technology of the electrothermal atomization of carbon in vacuum (at the temperature ~3000 K and residual pressure <10–4 Pa), widely used for the formation of conducting coatings in electron
No. 13
2016
1224
KOSYAKOV et al.
(b)
nm 173.2 124.0 0.2 0.2 1 µm
(a)
0.4
0.6
0.8
µm
0.2
0.4
0.6
0.8 µm
Fig. 1. (a) SEM microphotograph and (b) AFM-image of a carbon nanocoating on the surface of a MALDI target plate.
microscopy. A comparison of two types of carbon sources, carbon filaments and graphite rods, showed that the last source offers substantial advantages and gives a more uniform and reproducible coating of the whole surface of a MALDI target with the optimum thickness of the applied layer. In this regard, the next experiments were performed with graphite rod electrodes thinned to a diameter of 1.4 mm. The optimum conditions of the formation of coatings based on such criteria as the uniformity of the application of carbon, strength of the formed layer (absence of separating particles capable of polluting the ion source of the mass spectrometer), and the maximum intensity of analyte mass spectra were selected experimentally as follows: current during the pulse evaporation of carbon, 75 A; number of pulses, 3; pulse duration, 1.5 s; and pause between pulses, 30 s. The total time for the preparation of one target by the described method was 10–15 min, including all stages of the preparation of the experimental installation. RESULTS AND DISCUSSION Study of carbon nanocoatings. The results of study of the prepared targets by scanning electron microscopy (Fig. 1a) demonstrate the exclusive smoothness and uniformity of the coating formed under these conditions. We observed on its surface only single individual carbon particles, from 10 to 100 nm in size; rather large particles (to 1 μm) were met seldom. Pictures of end faces are indicative of a good reproducibility of the thickness of the formed carbon layers; this parameter fluctuated in a narrow range of 50–60 nm. A similar result was obtained in the study of artificially prepared CNC chips on the target surface by atomic-force microscopy (Fig. 1b). The proposed approach allows the researcher to obtain targets for SALDI mass spectrometry, characterized by high hydrophobicity, favoring the high-
quality application of samples from organic solvents, and also with the minimum amount of applied carbon (the mass of coating on a well is several times smaller than typical amounts of graphite or carbon nanotubes applied by conventional methods). The high strength of carbon layer prevents the dispersion of the coating material in the course of laser desorption, considerably reducing the risk of contamination of the mass spectrometer. The uniformity of the coating on the target surface, ensuring the recording of high-quality mass spectra of analytes, is illustrated by optical pictures of the surface of a steel target with CNC and other types of carbon materials used (Fig. 2). Mass-spectrometric characteristics of carbon nanocoatings. The study of the dependences of peak intensities in the mass spectra of the studied carbon matrixes on the intensity of laser radiation showed that CNC differ by the highest threshold of carbon ionization: peaks corresponding to carbon clusters appeared at a pulse energy of 52 μJ. For graphite and graphite pencil, the threshold value was 40 μJ and for CNT, 33 μJ, which is 1.5 times lower than that for CNC. Mass spectra of the obtained carbon nanocoating and also of spectrally pure graphite, graphite pencil, and carbon nanotubes applied onto a steel target plate are presented in Fig. 3. It is evident that nanocoatings are characterized not only by the lowest total ion current at different energies of the laser pulse, but also by the minimum number of peaks in the region m/z 100– 600, which is very important for the study of lowmolecular compounds. Surface-assisted laser desorption/ionization of lowmolecular compounds. Using CNC, we could record mass spectra of all test compounds. The spectra exhibited peaks of both protonated molecules and molecules cationized by alkali metals (Table 2). This effect is characteristic for any carbon matrix and was many times described in the literature [1, 8, 9]. A comparison of the results obtained in the application of a car-
JOURNAL OF ANALYTICAL CHEMISTRY
Vol. 71
No. 13
2016
CARBON NANOCOATINGS
50 µm
(a)
50 µm
(b)
50 µm
(c)
1225
50 µm
(d)
Fig. 2. Optical micrographs of the surface of a steel target plate with applied (a) graphite, (b) pencil lead, (c) carbon nanotubes, and (d) carbon nanocoating.
bon nanocoating, carbon nanotubes, graphite, and lead of a graphite pencil indicated a smaller relative intensity of peaks of [M + Li]+, [M + Na]+, and [M + K]+ ions for CNC, which is, probably, due to a higher purity of the target material prepared by the proposed method. The majority of the studied compounds were characterized by close values of threshold energies of the laser pulse for desorption/ionization from target plates with different types of carbon coatings (Table 3). Exceptions are provided by betulin and cellobiose, whose mass spectra could not been recorded in the studied energy range of laser radiation (up to 70 μJ) using a graphite pencil and carbon nanotubes (compound 5). The carbon nanocoating differs by a possibility of recording sufficiently intense spectra of all of the studied analytes at the intensity of UV radiation much lower than the threshold of carbon ionization. A comparison of the intensities of the main peaks in the mass spectra of the studied compounds JOURNAL OF ANALYTICAL CHEMISTRY
Vol. 71
(Table 4) shows that the carbon nanocoating is inferior to carbon nanotubes and, for some of analytes, to graphite. This is due to a small amount of the applied carbon and also due to a rather small surface area of the smooth coating. Nevertheless, for the same reasons, carbon nanocoatings demonstrated exclusively high resolution compared to the other matrixes, twoto eightfold exceeding the corresponding values for CNT. As a whole, SALDI mass spectra recorded using CNC in quality significantly exceed the results obtained with other carbon materials. As an illustration, Fig. 4 presents a comparison of mass spectra of polyethyleneglycol applied onto CNC and carbon nanotubes. ACKNOWLEDGMENTS This work was performed in the Arktika Core Facility Center of the Northern (Arctic) Federal University named after M.V. Lomonosov and supported by the Ministry of Education and Science of the Rus-
No. 13
2016
1226
KOSYAKOV et al.
100
100
(a)
(a) 1
0
Relative intensity, %
Relative intensity, %
50
50 2
0
50 3
0
80 60 40 20
50 0
100
200
300 m/z
400
500
4 600
100
Relative intensity, %
Relative intensity, %
1
0 50
2
0 50
3
0
700 600 m/z
800
900 (b)
80 60 40 20
50 0
500
400
100
(b)
50
0
100
200
300 m/z
400
500
4 600
Fig. 3. Mass spectra of different carbon materials recorded at the energy of laser radiation (a) 35 and (b) 50 μJ: 1, carbon nanocoating; 2, slate pencil; 3, graphite; 4, carbon nanotubes.
0
500
400
700 600 m/z
800
900
Fig. 4. Mass spectra of PEG-600 recorded using (a) carbon nanocoating and (b) carbon nanotubes. Energy of laser pulse 40 μJ.
Table 2. Relative peak intensities (Irel) in mass spectra of analytes recorded using different carbon materials Matrix
Irel for the compound
Ion 1
2
3
4
5
6
Carbon nanocoating
Н]+
[М + [M + Li]+ [M + Na]+ [М + К]+
100 0 7 9
100 0 54 34
100 0 0 0
0 0 100 67
0 100 0 0
0 4 100 49
Graphite
[М + Н]+ [M + Li]+ [M + Na]+ [М + К]+
100 25 88 95
100 22 79 27
100 7 24 17
0 16 100 35
0 78 28 100
0 40 97 100
Graphite pencil
[М + Н]+ [M + Li]+ [M + Na]+ [М + К]+
100 0 15 43
100 0 25 23
100 0 0 8
0 0 0 0
0 0 0 0
0 26 76 100
Carbon nanotubes
[М + Н]+ [M + Li]+ [M + Na]+ [М + К]+
100 33 69 77
100 47 49 13
100 30 15 0
0 85 100 43
0 0 0 0
0 51 100 66
JOURNAL OF ANALYTICAL CHEMISTRY
Vol. 71
No. 13
2016
CARBON NANOCOATINGS
1227
Table 3. Comparison of threshold energies of laser pulse for the desorption/ionization of the studied compounds using different carbon matrixes Ionization threshold, μJ, for carbon matrixes
Compound carbon nanocoating
graphite
carbon nanotubes
graphite pencil
35 35 20 35 45 35
35 30 30 45 45 30
30 30 15 35 – 30
35 35 30 – – 35
1 2 3 4 5 6
Table 4. Comparison of characteristics of the main peaks in mass spectra of analytes recorded using carbon matrixes Relative intensity of the main peak (resolution) Compound 1 2 3 4 5 6
Laser energy, μJ 29 25 22 29 29 25
carbon nanocoating
graphite
1549 (1824) 762 (3411) 1687 (3678) 1097 (2117) 90 (4881) 255 (6491)
1683 (942) 1896 (890) 930 (1476) 1051 (1890) 256 (2062) 1708 (1551)
carbon nanotubes 1717 (821) 1948 (852) 1848 (466) 1858 (951) 0 1358 (1439)
graphite pencil 640 (462) 879 (436) 741 (427) 0 0 440 (335)
sian Federation (Contract no. 14.594.21.0004, unique identifier of works RFMEFI59414X0004) and Russian Foundation for Basic Research (project no. 1303-12238 ofi-m).
10. Ugarov, M.V., Egan, T., Khabashesku, D.V., Schultz, J.A., Peng, H., Khabashesku, V.N., Furutani, H., Prather, K.S., Wang, H-W.J., Jackson, S.N., and Woods, A.S., Anal. Chem., 2004, vol. 76, no. 22, p. 6734.
REFERENCES
11. Xu, S., Li, Y., Zou, H., Qiu, J., Guo, Z., and Guo, B., Anal. Chem., 2003, vol. 75, no. 22, p. 6191.
1. Hillenkamp, F. and Peter-Katalinic, J., MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Weinheim: Wiley, 2014. 2. Shanta, S.R., Kim, T.Y., Hong, J.H., Lee, J.H., Shin, C.Y., Kim, K.-H., Kim, Y.H., Kim, S.K., and Kim, K.P., Analyst, 2012, vol. 137, no. 24, p. 5757. 3. Yamaguchi, S., Fujita, T., Fujino, T., and Korenaga, T., Anal Sci., 2008, vol. 24, no. 11, p. 1497. 4. Peterson, D.S., Mass. Spectrom. Rev., 2007, vol. 26, no. 1, p. 19. 5. Sunner, J., Dratz, E., and Chen, Y.-C., Anal. Chem., 1995, vol. 67, no. 23, p. 4335. 6. Kim, J. and Kang, W., Bull. Korean Chem. Soc., 2000, vol. 21, no. 4, p. 401. 7. Berger-Nicoletti, E., Wurm, F., Kilbinger, A.F.M., and Frey, H., Macromolecules, 2007, vol. 40, no. 3, p. 746. 8. Borisov, R.S., Polovkov, N.Yu., Zhilyaev, D.I., Esparsa, S.A., and Zaikin, V.G., Mass-Spektrometriya, 2014, vol. 11, no. 2, p. 107. 9. Borisov, R.S., Zhilyaev, D.I., Esparsa, Sandoval’S.A., Polovkov, N.Yu., and Zaikin, V.G., Mass-Spektrometriya, 2014, vol. 11, no. 4. JOURNAL OF ANALYTICAL CHEMISTRY
Vol. 71
12. Ren, S., Zhang, L., Cheng, Z., and Guo, Y., J. Am. Soc. Mass. Spectrom., 2005, vol. 16, no. 3, p. 333. 13. Pan, C., Xu, S., Hu, L., Su, X., Ou, J., Zou, H., Guo, Z., Zhang, Y., and Guo, B., J. Am. Soc. Mass. Spectrom., 2005, vol. 16, no. 6, p. 883. 14. Dong, X., Cheng, J., Li, J., and Wang, Y., Anal. Chem., 2010, vol. 82, no. 14, p. 6208. 15. Winkler, W., Balika, W., Hausberger, P., Kraushaarb, H., and Allmaiera, G., J. Mass. Spectrom., 2010, vol. 45, no. 5, p. 566. 16. Echlin, P., Handbook of Sample Preparation for Scanning Electron Microscopy and X-Ray Microanalysis, New York: Springer, 2009. 17. Strohalm, M., Kavan, D., Novak, P., Volny, M., and Havlicek, V., Anal. Chem., 2010, vol. 82, no. 11, p. 4648.
No. 13
Translated by E. Rykova 2016