Metal-Assisted Secondary Ion Mass Spectrometry: Influence of Ag and

Oct 15, 2004 - Sadia Sheraz née Rabbani , Irma Berrueta Razo , Taylor Kohn , Nicholas P. Lockyer , and John C. Vickerman. Analytical Chemistry 2015 8...
0 downloads 0 Views 773KB Size
Download PDF
Anal. Chem. 2004, 76, 6777-6785

Metal-Assisted Secondary Ion Mass Spectrometry: Influence of Ag and Au Deposition on Molecular Ion Yields L. Adriaensen,*,† F. Vangaever,‡ and R. Gijbels†

Department of Chemistry (MiTAC), University of Antwerp, Universiteitsplein 1 B-2610 Wilrijk, Belgium, and Agfa-Gevaert N.V., B-2640 Mortsel, Belgium

A series of organic dyes and a pharmaceutical are used to study secondary ion yield enhancement by metal deposition. The molecules were dissolved in methanol and spin-casted on silicon substrates. Subsequently, silver or gold was evaporated onto the samples to produce a very thin coating. The coated samples, when measured with TOF-SIMS, showed a considerable increase in characteristic secondary ion intensity. Gold evaporated samples appeared to exhibit the highest signal enhancement. A major advantage of the metallization technique is that it can be used on real world samples. In particular, additive containing organic crystals could be studied; however, the observed signal increase does not occur at any given moment. The time between metal deposition on the sample surface and the measurement of the sample with TOf-SIMS appears to have an important influence on the enhancement of the secondary ion intensities. Therefore, the metal-coated samples were measured at different times after sample preparation. The results show that, depending on the sample and the metal deposited, the secondary ion signals reach their maximum at different times. Further study will be necessary to reveal the mechanism responsible for the observed enhancement effect. Since its development in the early 1970s, static SIMS (S-SIMS) has become a well-established analytical technique for the characterization of organic and biological materials.1,2 Apart from its capability to render molecular information from the uppermost surface layers, it is also able to provide ion images of the molecular distribution at the surface. Model S-SIMS studies typically involve analyte deposition on noble metal substrates. It is preferable to have monolayers, because under these circumstances, maximum ion yields are reached for S-SIMS.3,4 Furthermore, specifically tailored model systems are also being examined. The self* To whom correspondence should be addressed. Phone: +32-3-820-23-89. Fax: +32-3-820-23-76. E-mail: [email protected]. † University of Antwerp. ‡ Agfa-Gevaert N.V. (1) Hagenhoff, B. SIMS Analysis of Organic Materials. Secondary Ion Mass Spectrometry SIMS IX; Benninghoven, A., et al., Eds.; John Wiley: Chichester, 1993, pp 753-756. (2) Gillen, G. J. Am. Soc. Mass Spectrom. 1993, 4, 419-423. (3) Van Vaeck, L.; Adriaens, A.; Gijbels, R. Mass. Spectrom. Rev. 1999, 18, 1-47. 10.1021/ac049108d CCC: $27.50 Published on Web 10/15/2004

© 2004 American Chemical Society

assembly process5 and the Langmuir-Blodgett technique6 are the most widely applied methods to construct such ordered systems. The analysis of “real world” samples, originating either from the life sciences or from the material sciences, started only in the 1990s.7,8 Charging effects, irregular shapes, and the composition of these samples often pose difficulties during S-SIMS analysis; therefore, dedicated equipment and sample preparation methods are often needed. Effective high spatial resolution surface analysis of molecular materials by TOF-S-SIMS requires the optimization of the secondary ion yield, since a decrease in the area of analysis results in the reduction of the number of molecules within each pixel. Development of yield-improving methods is one of the major goals in SIMS research. On one hand, new primary ion bombardment conditions are being tested for this purpose. The use of polyatomic or cluster primary ions, such as SF5+,9 Cn+ 10, or Aun+ 11 have shown promising results. On the other hand, several sample preparation techniques, such as the use of MALDI matrixes12 or halide additives,13 were tested in this regard. In the present work, small amounts of Ag or Au are deposited on the sample surface to increase the secondary ion intensities. The sister technique in which organic molecules are cast on metal substrates has been frequently used and described in the literature over the last two decades.14,15 Although this sample preparation route has a positive effect on the secondary ion yields of organic (4) Lenaerts, J.; Verlinden, G.; Van Vaeck, L.; Gijbels, R.; Geuens, I. TOF-SIMS Analysis of Carbocyanine Dyes Adsorbed on Silver Subsrates. Secondary Ion Mass Spectrometry SIMS XII, Benninghoven, A., et al., Eds.; Elsevier: Amsterdam, 1999, pp 115-118. (5) Gillen, G.; Bennet, J.; Tarlov, M. J.; Burgess, D., Jr. Anal. Chem. 1994, 66, 2170-2174. (6) Nowak, R. W.; Gardella, J. A.; Wood, T. D.; Zimmerman, P. A.; Hercules, D. M. Anal. Chem. 2000, 72, 4585-4590. (7) Benninghoven, A.; Hagenhoff, B.; Niehuis, E. Anal. Chem. 1993, 65, 14, 630-639. (8) Belu, A. M.; Davies, M. C.; Newton, J. M.; Patel, N. Anal. Chem. 2000, 72, 5625-5638. (9) Gillen, G.; Roberson, S. Rapid Commun. Mass Spectrom. 1998, 12, 13031312. (10) Wong, S. C. C.; Hill, R.; Blenkinsopp, P.; Lockyer, N. P.; Weibel, D. E.; Vickerman, J. C. Appl. Surf. Sci. 2003, 204, 219-222. (11) Davies, N.; Weibel, D. E.; Blenkinsopp, P.; Lockyer, N.; Hill, R.; Vickerman, J. C. App. Surf. Sci. 2003, 204, 223-227. (12) Wu, K. J.; Odom, R. W. Anal. Chem. 1996, 68, 5, 873-882. (13) Gusev, A. I.; Choi, B. K.; Hercules, D. M. J. Mass Spectrom. 1998, 33, 480485. (14) Zimmerman, P. A.; Hercules, D. M.; Benninghoven, A. Anal. Chem. 1993, 65, 983.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004 6777

(sub)monolayers, it has the major disadvantage of not being applicable on materials that cannot be prepared as thin films, such as compounds with a very poor solubility or just real world samples. Under these circumstances, the evaporation of metals on the sample surface, called metal-assisted secondary ion mass spectrometry (MetA-SIMS), offers an interesting alternative. Literature concerning this subject is relatively scarce. As a consequence, the possibilities this technique has to offer have not yet been fully explored. Linton et al. were the first to report the use of metallization.16 By means of a TEM grid, a 150-nmthick, patterned Ag layer was deposited on polymer samples to enhance the Ag-cationized ion signals. Although promising results were obtained, it is questionable whether significant amounts of secondary ions were emitted from the large Ag covered areas. As such, maximal ion yields may not have been attained. More recent publications have shown that coverage of polymeric materials with a very thin and continuous, as opposed to a patterned, metal overlayer also causes an increase in the ionization efficiency for S-SIMS.17,18 Most of the MetA-SIMS19 experiments described in the literature were conducted on polymers and polymer additives. In this study, it is demonstrated that MetA-SIMS can also be applied to other types of materials. Small quantities of Ag or Au are deposited on cationic and anionic dyes and on the pharmaceutical Risperidone. In the first part of the discussion, the enhancement effect on the spectral information, obtained from spin-coated samples, is studied. A comparative study is made between Ag and Au overlayers. The results from metal-covered samples are also compared to those obtained from samples spin-casted on Ag and Au substrates. Several TOF-S-SIMS measurements show that metallized samples are not stable over time. These instabilities, which are likely caused by diffusion processes of the metals in the organic material or vice versa, give rise to quite strong variations of the secondary ion signals.20 As a result, the enhancement effect of deposited metals is very time-dependent and, thus, has to be interpreted with care. In the last part of the discussion, MetA-SIMS is applied on additive-containing organic crystals. In this section, our major goal is to demonstrate the possibilities of the metal deposition sample preparation technique for solving problems concerning real world samples. It is indicated that, apart from a reduction of the sample (15) Keller, B. A.; Hug, P. A new type of support based on ion-coating polymers for selective generation of cationized species in TOF-SIMS spectra. Secondary Ion Mass Spectrometry, SIMS XII Proceedings; Benninghoven, A., Bertrand, P., Migeon, H. N., Werner, H. W., Eds.; Elsevier: Amsterdam, 2000, pp 749-885. (16) Linton, R. W.; Mawn, M. P.; Belu, A. M.; DeSimone, J. M.; Hunt, M. O.; Menceloglu, Y. Z.; Cramer, H. G.; Benninghoven, A. Surf. Interface Anal. 1993, 20, 991-999. (17) Delcorte, A.; Me´dard, N.; Bertrand, P. Anal. Chem. 2002, 74, 4955-4968. (18) Delcorte, A.; Bour, J.; Aubriet, F.; Muller, J. F.; Bertrand, P. Anal. Chem. 2003, 75, 6875-6885. (19) Delcorte, A.; Bertrand, P. Interest of Silver and Gold Metallization for Molecular SIMS and SIMS Imaging. Applied Surface Science: Proceedings of the 14th International Conference on Secondary Ion Mass Spectrometry and Related Topics; Benninghoven, A., et al., Eds.; Elsevier: Amsterdam, 2004, p 231, in press. (20) Adriaensen, L.; Vangaever, F.; Gijbels, R. Organic SIMS: the Influence of Time on the Ion Yield Enhancement by Silver and Gold Deposition. Applied Surface Science: Proceedings of the 14th International Conference on Secondary Ion Mass Spectrometry and Related Topics; Benninghoven, A., et al.; Eds.; Elsevier: Amsterdam, 2004, 231, pp 256-260.

6778

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

charging, a considerable increase in secondary ion yields of ion species, originating from the crystal material itself as well as from the additives, is recorded. EXPERIMENTAL SECTION Risperidone was purchased from Johnson & Johnson (Beerse, Belgium). The organic carbocyanine dyes and crystals, analyzed in this work, were synthesized at Agfa Gevaert (Mortsel, Belgium). The dyes were dissolved in methanol to a concentration of 10-3 M. Subsequently, aliquots of 10 µL were spin-coated (8000 rpm for 120 s) onto Si substrates. Prior to spin-coating, the substrates were cleaned with methanol in an ultrasonic bath. The deposition of Ag and Au was done by means of a BAL-TEC sputter-coater SCD005 using Argon as operating gas, a working pressure of 0.05 mbar, a sputter current of 60 mA, and a working distance of 50 mm. Ag and Au substrates were prepared by depositing Ag and Au for 250 s on a cleaned Si substrate. The acquisition of SIMS spectra was performed on a TOFSIMS IV instrument (Cameca/Ion-Tof) using a pulsed Ga+ liquid metal ion source (25 keV) or a pulsed SF5+ electron impact ionization source (9 keV) with a current of 0.2 pA. The same primary ion flux of 1012 ions/cm2, well below the so-called static limit, was used for the acquisition of all the SIMS spectra. A detailed description of the apparatus and its principles of operation are given by Briggs et al.21 and Vickerman et al.22 The primary ion pulses, typically 1 ns, gave rise to mass spectra with a mass resolution of 7400 with Ga+ ions and 4000 with SF5+ ions (full width half-maximum) for [C2H5]+. On each sample, at least three different positions were analyzed. All spectra were obtained from 100 × 100 µm2 areas with an acquisition time of 150 s. Secondary ion images were taken from 500 × 500 µm2 areas, using Ga+ primary ions (no. scans, 5; no. shots/pixel, 32). A flood gun with 30 eV electrons was applied to those samples that showed signs of charging effects. Images of Si substrates, covered with a thin Ag and Au layer, were taken with a FEI Sirion ultrahigh-resolution scanning electron microscope (SEM). 3. RESULTS AND DISCUSSION Signal Enhancement Effect. Initial SIMS measurements on carbocyanine dyes and Risperidone samples treated with Ag and Au showed that metal deposition on the sample surface prior to TOF-S-SIMS analysis results in an enhancement of the secondary ion yields in comparison to the untreated samples. The molecular structures of the analyzed molecules are depicted in Figure 1. In the first instance, some experiments were carried out to establish the best conditions for metal deposition. For this purpose, several amounts of Au and Ag were deposited onto the surface of the samples and measured with S-SIMS. The measured secondary ion intensities were monitored as a function of the evaporated metal quantity. The results obtained for various ions of Ag-metallized Risperidone and dye C1 samples are displayed in Figure 2. The monitored signals at m/z 411 ([R + H]+) and 233 originate from Risperidone, whereas the ions at m/z 649 ([C1]+), 577, 359, and 299 come from dye C1. The Ag+ intensity (21) Briggs, D.; Seah, M. Practical Surface Analysis: Ion and Neutral Spectroscopy; John Wiley: Chichester, 1990; Vol 2; pp 303-420. (22) Vickerman, J. C.; Briggs, D. Tof-SIMS: Surface Analysis by Mass Spectrometry; IM Publications and Surface Spectra Ltd.: Chichester, 2001, p 789.

Figure 1. Overview of the molecular structures and the short notations of the cationic (C) and anionic (A) dyes and Risperidone analyzed in this work. (Et, ethyl; TMS, trimethylsilyl).

Figure 3. High-resolution SEM images of Si substrates after Au and Ag deposition: (a) Au deposited on Si for 20 s and (b) Ag deposited on Si for 20 s.

Figure 2. Secondary ion intensities of signals at m/z 411 and 233, originating from spin-coated Risperidone samples, and of signals at m/z 649, 577, 359, and 299, originating from spin-coated dye C1 samples, as a function of the amount of Ag deposited on the sample surface. Inset: Secondary ion intensity of Ag as a function of the amount of deposited Ag.

(depicted in the inset) increases almost linearly with the amount of Ag up to 120 s of deposition time. After that, it still continues to increase, but with a less steep slope. The highest signal intensities for (protonated) molecular ions are recorded from samples that have been metal-coated for 10-20 s, equivalent to the deposition of a 3-5-nm-thick metal layer on the surface. Longer deposition times bring about a decrease of the ion intensities. Deposition times of 60 s or more result in an especially rapid deterioration of the secondary ion intensities. This is probably due to the complete coverage of the organic material by the metal layer, so the organic molecules are no longer accessible to TOF-SIMS. Similar conclusions can be drawn when the experiment is repeated with Au layers instead of with Ag layers (not shown). On the basis of these results, samples were all coated for 20 s unless stated otherwise. There are essentially three crystal growth modes by vapor deposition: layer-by-layer growth, island growth, and layer-plusisland growth.23 Due to the fact that numerous parameters, such as surface dangling bonds, lattice mismatch, defect concentration,

and substrate temperature, etc., play an important role during the growth, limited understanding has been achieved for predicting the actual growth mode of a specific system. SEM images of Si substrates coated with Au and Ag during 20 s are shown in Figure 3. As expected, the deposited Ag and Au metals do not completely cover the substrate, but tend to form nanometer-sized metal islands on the surface. When the metal layers cover the entire samples, detection of any underlying molecules would be highly unlikely with TOF-S-SIMS, since the primary ions would not actually hit the organic samples. Moreover, Au and Ag give rise to somewhat different patterns on the Si surface. Au forms bigger islands than Ag. When the metals were deposited on an organic layer on Si, similar observations were made. Spin-coated samples of C1, C2, C3, A1 and Risperidone, metallized with Ag and Au, were measured with TOF-S-SIMS in positive mode. The intensities recorded for the (protonated) molecular ions on untreated samples and on Ag and Au bearing samples are summarized in Figure 4. The resulting mass spectra of an untreated, a Ag-coated, and a Au-coated C1 sample are shown in Figure 5. For the sake of clarity, the number of ion counts recorded for the molecular ion and for the adduct ions are shown between brackets. The other measured samples give rise to mass spectra with the same characteristics (not shown). The data in Figures 4 and 5 demonstrate that the deposition of Ag and Au on the sample surface can lead to enhancement of the secondary ion yields. In the case of the anionic dye A1 (spectra not shown), it even leads to the detection of the protonated molecular ion at (23) Zhang, X. B.; Vasiliev, A. L.; Van Tendeloo, G.; Yan, H.; Yu, L. M.; Thiry, P. A. Surf. Sci. 1995, 340, 317-327.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6779

Figure 4. Secondary ion intensities of the (protonated) molecular ions of dyes C1, C2, C3, A1, and Risperidone measured on nonmetallized and on Ag- and Au-metallized samples.

m/z 807 and several characteristic fragment ions at higher m/z values, which could barely be distinguished from the background noise in spectra from untreated samples. Although the intensity enhancement is observed for all of the recorded ions, it appears to be very dependent on the ion species. In general, the (protonated) molecular ions undergo the largest intensity improvement. Enhancement factors ranging from 1 to 20 for Agcoated samples and from 3 to 48 for Au-coated samples are recorded. The fragment ions do not necessarily exhibit the same enhancement effect as the (protonated) molecular ions. Several experiments showed that while the (protonated) molecular ions reach their optimal intensity after 20 s of Ag/Au deposition, some molecular fragments attain a maximum yield at higher deposition times. In Figure 2, the evolution of fragment ions at m/z 299, 359, 577, and 621, originating from dye C1, illustrates this. It is clear that the signal at m/z 359 reaches a maximal intensity value after 60 s of Ag deposition instead of after 20 s, as was the case for the C1 molecular ion at m/z 649. Usually, Au causes a stronger enhancement effect than Ag. A possible reason might be the difference in atomic mass between the two metals. Since Pt and Au do not differ much in atomic mass, new samples were coated with several amounts of Pt; however, none of these gave evidence of increased secondary ion yields. A remarkable finding is that the measured samples hardly give rise to Ag and Au adduct ions under primary ion bombardment. Moreover, the recorded spectra are qualitatively very similar to the spectra taken from nonmetallized samples, as can be seen from Figure 5. The detected secondary ions are the same for nonmetallized and metallized samples, but the ion intensities can be different. This is a substantial difference from the work of Delcorte et al.,17 in which it was reported that the deposition of a thin Au layer on polymer samples resulted in a series of new Au cationized ion signals. Secondary ion signals obtained from metal-coated samples after measurement with SF5+ primary ions also showed an intensity rise; however, the enhancement factor is generally lower as compared to Ga+ measurements. When measured with Ga+ primary ions, the C1 signals are enhanced by a factor 20 after Ag deposition, whereas measurement with SF5+ results in a secondary ion yield increase of a factor of 2. However, it has to be noted that bombardment of nonmetallized samples with SF5+ primary ions already results in higher ion intensities as compared to that with Ga+ primary ions. 6780 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

The samples were also studied in the negative mode. None of the untreated samples produced intense molecular secondary ion signals. Only elements such as fluorine or chlorine, incorporated in the studied molecules, produced signals in the spectra with significant intensities. In contrast to the positive mode, no enhancement effect was observed after the deposition of a thin Au or Ag layer on the sample surface. In fact, the signals showed a decrease, rather than an increase. Only the Cl- signal was detected with higher intensities. This is probably due to contamination introduced in the sample during the metal coating procedure, because dye 2, which does not contain any chlorine, also produced a quite intense Cl- peak in the negative TOF-S-SIMS spectra. Ag and Au Substrates. In this section, the use of metal deposition is evaluated in comparison to the use of Ag and Au substrates. Spin-coated samples of the organic dyes C1 and A1 and Risperidone on Ag and Au substrates were prepared as described in the Experimental Section and measured with TOF-S-SIMS (Ga+). In Figure 6, the results (positive mode) of a comparison between the use of metal substrates and the metal-coated samples are summarized for the molecular ion and for some characteristic fragments of dye C1. Similar results were found for the other studied compounds. Metal substrates appear to give rise to secondary intensities up to a factor of 10 higher than the metallization method. In general, the use of Au substrates offers higher ion yields than Ag substrates do, just as Au deposited on the sample surface causes a higher enhancement effect as compared to deposited Ag. Measurements in the negative mode lead to a similar trend concerning the differences between the Au and Ag substrates. In contrast to metallization, the use of Au and Ag substrates produces an increase in the negative secondary ion intensities. However, characteristic ion signals at higher m/z values are still not recorded in the negative mass spectra. In summary, the use of Ag and Au substrates tends to provide higher secondary ion yields as compared to deposition of Ag and Au, but this approach is not generally applicable. The metallization method, on the other hand, still outperforms substrates such as Si, while, more importantly, it can be applied to a larger set of samples, even to real world samples. Time Dependence. To test the stability of the metallized samples, spin-coated carbocyanine dyes on Si were covered with a 3-5 nm Ag or Au layer. These samples were measured 5, 14, and 21 days after metal deposition. In Figure 7 the recorded intensities for the molecular ions and some characteristic fragments of dyes C2 (Figure 8a, b) and A1 (Figure 8c, d) are presented. It should be noted that the intensity of the (protonated) molecular ion signals measured on the untreated dye samples (no metal coating) decreases as the samples get older. Most of the characteristic fragment ions show just the opposite behavior, namely, an increase in ion yield as a function of time. The phenomenon can be attributed to the slow degradation of the spincoated samples, which starts directly after the sample preparation. However, in general, these intensity variations remain within acceptable limits. The results in Figure 7 show that the evolution of metal-coated samples is very unpredictable. For example, the Ag-coated dye C2 sample gives optimal [C2]+ intensities 14 days after metal deposition, whereas the samples covered with Au have already

Figure 5. Positive S-SIMS and MetA-SIMS spectra of spin-coated dye C1 samples. Top spectrum, without metal deposition; middle spectrum, after 20-s deposition of Ag on the sample surface; bottom spectrum, after 20-s deposition of Au on the sample surface.

Figure 6. Secondary ion intensities of the molecular ion (m/z 649) and several fragment ions (m/z 577, 361, 335, 299, and 277) of dye C1 measured on Si, on Ag- and Au-covered samples, and on Ag and Au substrates.

reached a maximum for [C2]+ after 5 days. The fragment ion at m/z 243 gives a high yield after 14 and 21 days for Au- as well as for Ag-coated samples. The behavior of the ion yields for dye A1 samples are again totally different. For instance, the [M + 2H]+

signal at m/z 807 and the characteristic fragment signal at m/z 299 are most intense after 14 days on both Ag- and Au-covered samples. In conclusion, these measurements show that Au and Ag metallization induces an enhancement of the secondary ion yield of the studied samples; however, due to the influence of time, it is unclear at which time the maximal enhancement will occur. This not only makes it difficult to know when the best time is to analyze the samples but also poses problems regarding the reproducibility of the results. Taking this time-dependence into account and the fact that samples covered with too much metal do not show any enhancement of the secondary ion yields, the yield enhancement might be explained as follows. After the deposition of a thin metal layer on the sample surface, diffusion of the organic species on top of the metal islands or penetration of the metal particles through the organic material24 seems to be a prerequisite for ion yield enhancement. The occurrence of diffusion processes can explain the observed time effect in MetA-SIMS experiments. When samples are measured too soon after metallization, that is, before sufficient diffusion has taken place, no enhancement effect is recorded. The analysis of samples in which the organic molecules have diffused too much also gives no yield enhancement. It is possible that, as a result of the diffusion, one arrives at a situation in which the metal islands act in a way similar to metal substrates. Molecular dynamics experiments, previously conducted by (24) Whitten, J. E.; Gomer, R. J. Phys. Chem. 1996, 100, 2255-2259.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6781

Figure 7. The influence of time on MetA-SIMS results. Secondary ion intensities for the molecular ions and some characteristic fragments of dyes C2 (see Figure 8a, b) and A1 (see Figure 8c, d) measured 5, 14, and 21 days after metal deposition.

Delcorte et al.,25 suggest that heavy metal substrates facilitate molecular desorption via a very efficient upward reflection of the projectile momentum. Furthermore, the ejected molecules appear to have only a limited amount of internal energy, thus lowering the possibilities of fragmentation processes to occur. These metal substrate qualities can clearly result in the enhancement of secondary ion intensities. Real World Samples. The possible use of MetA-SIMS for real world samples was tested on millimeter-sized organic crystals, which contained some additives in addition to the main crystal molecules. For the engineering of organic crystals, data regarding the distribution of molecules on and in the crystals is very important, since it holds information about the crystal growth mechanisms and the behavior of the additives present in the crystals. However, this information is often difficult to obtain by SIMS, since the secondary ion intensities of organic molecules, embedded in a matrix, are rather low. In addition, organic crystals are nonconductive, so charging effects also negatively influence the measurements. In the following experiments, MetA-SIMS is tested on millimeter-sized organic crystals after coating the samples with a thin layer of Au. Attention is paid to the effect of Au metallization on the recording of spectra as well as on the acquisition of secondary ion images. Figure 8 provides an overview of the studied crystal molecules and additives. To compensate for charging of the crystals, an electron flood gun was used. Gilmore et al.26 have shown that although neutralization electrons are of low energy (typically 10-30 eV) and have small cross sections for damage, they can cause a significant amount of (25) Delcorte, A.; Vanden Eynde, X.; Bertrand, P.; Vickerman, J. C.; Garrison, B. J. J. Phys. Chem. B 2000, 104, 2673-2691. (26) Gilmore, I. S.; Seah, M. P. Appl. Surf. Sci. 2002, 187, 89-100.

6782 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

Figure 8. Schematic representation of the structures and short notations of the crystal molecules and additives.

Figure 9. Positive MetA-SIMS spectra of real world samples: (a) crystal Cr1 + Ad1 without metal, (b) crystal Cr1 + Ad1 after 20-s deposition of Au, (c) crystal Cr1 + Ad2 without metal, and (d) crystal Cr1 + Ad2 after 20-s deposition of Au.

damage in the sample when the delivered electron fluences are too high (∼2 × 1016 electrons/cm2). For some materials, signal intensity changes of a factor of 4 have been recorded. To take good spectra, it is therefore recommended to use electron fluences below 6 × 1014 electrons/cm2. Measurements on spin-coated Risperidone samples both with and without a thin metal layer have shown that the working conditions of the flood gun as used in this work do not result in substantial changes in signal intensities (results not shown). Therefore, it is assumed that neutralization electrons do not significantly influence the enhancement effect caused by deposited Ag and Au metals. The crystals were first measured without any Au deposited on the surface, using Ga+ primary ions. In Figure 9a and c, the positive mass spectra, obtained from the crystals Cr1 + Ad1 and Cr1 + Ad2, respectively, are shown. Under these measurement conditions, the signals originating from the main crystal molecules, such as the adduct ion [Cr1 + H]+ and the fragment ion at m/z

343, are present in the spectra at high intensities. The additives, however, appear more difficult to detect. To produce an increase in the secondary ion yields, Au was deposited on the crystal surface for 20 s. For comparison, the spectra of the crystals covered with a thin Au layer are shown in Figure 9b and d, respectively. These spectra were taken 4 days after the metal deposition. All of the secondary ions are recorded with higher intensities, as compared to the ion yields of the nonmetallized samples. However, the intensity enhancement highly depends on the type of ion. For the Cr1 + Ad1 crystal, the protonated molecular ion of the additive is recorded with an intensity increase of a factor of 250, whereas the signals originating from the Cr1 molecule undergo an enhancement of a factor ranging from 2 to 10. Measurements of the Au-bearing crystal Cr1 + Ad2 showed an increase in the protonated additive Ad2 yield by a factor of 10. Despite the observed enhancement effect, the [Ad2 + H]+ ion still has an ion count below 2000. However, the signal at m/z 471, Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6783

Figure 10. An example of MetA-SIMS applied to a real world sample: positive secondary ion images of the organic crystal Cr1 + Ad1 before and after Au (20 s) deposition. (Area: 500 × 500 µm2; no. scans, 5; no. shots per pixel, 32; an asterisk (*) denotes signals originating from the additive Ad1).

which was hardly detected in the nonmetallized crystal, is characteristic of the additive Ad2 and appears very intense in the spectrum. Furthermore, after Au deposition, the crystal molecules are also recorded as Au adducts [Cr1 + Au]+, which is not the case for the tested additives. The samples were also measured in negative mode. Only a few molecular signals with rather low intensities are present in the recorded spectra (not shown). The spectrum of crystal Cr1 + Ad2 also contains an elemental signal at m/z 19, due to the fluorine present in the Ad2 additive. Deposition of Au on the crystal surface results in slightly better ion yields for the molecular fragments for both crystal types. However, the F- ion signal is not enhanced as a result of metallization; in fact, its intensity even decreases slightly. Measurements show intense Cl- signals in the spectra of both crystal types after Au deposition. Since none of the studied molecules contain chlorine, the presence of the Cl- ion is probably due to contamination, introduced in the samples during Au deposition. As a result of the increased additive ion yields brought about by the Au metal at the crystal surface, it is possible to make positive secondary ion images of the crystal compounds. To illustrate this, a few ion images of crystal Cr1 + Ad1 before and 6784 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

after Au deposition are depicted in Figure 10. On the untreated samples, it is possible to make ion images of several crystal signals (e.g., m/z 431 and 343), but the additive signal (e.g., [Ad1 + H]+) intensities are too low to give any useful information concerning the molecular distribution on the sample surface. After the deposition of a thin Au layer, clear images can be recorded from main crystal molecules as well as from the additive, since the maximal number of counts per pixel for the [Ad1 + H]+ ion is increased from 2 to 167. The crystal signals also show a yield enhancement, although not so dramatic. For the crystal Cr1 + Ad2, similar observations could be made (not shown). Due to the low secondary ion yields in the negative mode, ion images of the F- distribution on the Cr1 + Ad2 crystal surface were the only ones achievable (not shown). Although the deposition of Au sometimes brings about a significant increase in the secondary ion yields to allow ion image acquisition, it can also induce some changes in the sample surface; i.e., in principle, it is possible that the recorded distribution of the compounds on the sample surface can be quite different from the original distribution. Therefore, it is essential that the acquired images are interpreted with care. Further experiments are necessary in this regard.

CONCLUSION The deposition of Ag or Au on the surface of organic samples brings about a significant increase in positive secondary ion yields in TOF-S-SIMS measurements with both Ga+ and SF5+. In the case of the anionic A1 dye sample, metal treatment even leads to the detection of signals at higher m/z values, which were not recorded on untreated samples. The noble metals occur as nanometer-sized particles on the sample surface. Ion yields of metal-covered samples are lower in comparison with yields obtained on samples spin-casted on Ag and Au substrates, but outperform those produced on Si substrates. The best results are produced when a metal coating of 3-5 nm is applied to the sample surface. In general, Au causes a stronger enhancement effect than Ag. The intensity improvement appears to be very dependent on the type of ion, which causes differences between the relative signal intensities recorded on metallized and nonmetallized samples. In contrast to positive ions, negative ions do not seem to undergo any yield enhancement. Furthermore, the metal-coated samples tend to be unstable as a function of time. Therefore, it is difficult to predict the most suitable moment for measurement. Lack of reproducibility of recorded data may occur as a result of this time influence.

A major advantage of metal deposition is that it can be used on a more diverse set of samples, in particular, on real world samples. Its yield enhancement effect can even cause some ion signals that could not be detected without metallization to appear in the recorded spectra. Sometimes the secondary ion yields get high enough to acquire ion images. However, these images have to be interpreted with care, since the deposition of metals can, in principle, alter the sample. Further experimentation is needed in this regard. ACKNOWLEDGMENT L.A. was supported by the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWTFlanders). The authors thank Jan Vandecruys (Agfa, Belgium) for the acquisition of the SEM images.

Received for review June 18, 2004. Accepted September 3, 2004. AC049108D

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6785