2596
J. Phys. Chem. B 2008, 112, 2596-2605
Quantitative Molecular Depth Profiling of Organic Delta-Layers by C60 Ion Sputtering and SIMS† Alexander G. Shard,* Felicia M. Green, Paul J. Brewer, Martin P. Seah, and Ian S. Gilmore National Physical Laboratory, Teddington, Middlesex, TW11 0LW, UK ReceiVed: September 12, 2007; In Final Form: October 11, 2007
Alternating layers of two different organic materials, Irganox1010 and Irganox3114, have been created using vapor deposition. The layers of Irganox3114 were very thin (∼2.5 nm) in comparison to the layers of Irganox1010 (∼55 or ∼90 nm) to create an organic equivalent of the inorganic ‘delta-layers’ commonly employed as reference materials in dynamic secondary ion mass spectrometry. Both materials have identical sputtering yields, and we show that organic delta layers may be used to determine some of the important metrological parameters for cluster ion beam depth profiling. We demonstrate, using a C60 ion source, that the sputtering yield, S, diminishes with ion dose and that the depth resolution also degrades. By comparison with atomic force microscopy data for films of pure Irganox1010, we show that the degradation in depth resolution is caused by the development of topography. Secondary ion intensities are a well-behaved function of sputtering yield and may be employed to obtain useful analytical information. Fragments characteristic of highly damaged material have intensity proportional to S, and those fragments with minimal molecular rearrangment exhibit intensities proportional to S2. We demonstrate quantitative analysis of the amount of substance in buried layers of a few nanometer thickness with an accuracy of ∼10%. Organic delta layers are valuable reference materials for comparing the capabilities of different cluster ion sources and experimental arrangements for the depth profiling of organic materials.
Introduction The increased sputtering yield of cluster ions, in comparison to monatomic ions, has engendered the possibility of conducting sputter depth profiles for organic materials.1-21 This development is of enormous technological importance as few, if any, techniques are capable of producing three-dimensional chemical maps22,23 with molecular sensitivity and depth resolution approaching the nanometer level. For a number of organic systems it appears that the volume of material sputtered from the surface is sufficient to remove much of the damage caused by previous impacts and, after a transitory accumulation of chemical damage, a steady state can be achieved.16,24 In this state, characteristic signals from organic species are detectable by surface analytical techniques such as secondary ion mass spectrometry (SIMS) and X-ray photoelectron spectroscopy.25 Recently,24 we demonstrated that the volume of material sputtered (yield volume) by C60 ions has a nearly linear relationship to the energy of the impact, and that this volume is almost identical for two organic materials, poly(lactide) and Irganox1010. Reported data for poly(methyl methacrylate)26 and trehalose14,16 give similar yield volumes, at least for thin films, to poly(lactide) and Irganox1010. Delcorte and Garrison27 have graphically demonstrated the similarity in yield (measured in mass per incident ion) between these materials and molecular dynamic simulations of poly(ethylene) sputtering. These results are very encouraging as they indicate that a single, energy dependent, yield volume can be applied to a wide range of materials. However, more materials should be studied and the † Based on a paper presented at the 44th IUVSTA Workshop: Sputtering and Ion Emission by Cluster Ion Beams, UK, April 23-27, 2007. * Author to whom correspondence should be addressed. E-mail:
[email protected].
effect of different experimental arrangements needs to be determined. For a number of organic materials, such as poly(methyl methacrylate) and aluminum tris(hydroxyquinolate) (Alq3), a true steady state is never achieved, and this is manifested, for example, by a gradual decline in characteristic secondary ion intensities and is possibly associated with both a reduction in sputtering yield24 and an increase in surface roughness.7 It is therefore important to identify and mitigate the causes of these long-term transitory effects if the analytical potential of molecular depth profiling is to be realized. One reported way to reduce these effects, for SF5+ cluster ions at least, is to cool the sample during depth profiling.7,9,10 In pursuance of the goal of improving comparability between instrumental arrangements, different ion sources and environmental conditions, and also to provide a greater understanding of sputtering yield changes and measuring depth resolution, we have developed a model organic material with multilayers of well-controlled thicknesses. By alternating thick layers of one material with thin layers of a second material, it is possible to construct analogues of the inorganic ‘delta-layers’ that have been usefully employed for many years as reference materials for depth profiling.28-32 We reported a prototype system in a recent communication24 with alternating thick layers of Irganox1010 and thin layers of Alq3. The disadvantage of this prototype was that Alq3 has a lower yield volume than Irganox1010, and consequently the remaining material accumulates damage more readily. This leads to uncertainty in determining the sputtering rate during the depth profile, particularly in the vicinity of the delta-layer. Here, we describe an organic delta-layered sample constructed from two materials, Irganox1010 and Irganox3114, which have the same yield volume. These delta-layers are employed to demonstrate that the sputtering yield is a well-
10.1021/jp077325n CCC: $40.75 Published 2008 by the American Chemical Society Published on Web 02/07/2008
Quantitative Molecular Depth Profiling
J. Phys. Chem. B, Vol. 112, No. 9, 2008 2597
Figure 1. Chemical structures of Irganox1010 and Irganox3114.
behaved function of the C60 ion energy and dose and, by comparing the SIMS response function with atomic force microscopy (AFM) data, that the depth resolution is dominated by the development of roughness. Experimental Section Irganox1010 and Irganox3114 (CIBA, Macclesfield, UK), with the chemical structures shown in Figure 1, were placed in separate crucibles within an Edwards AUTO306 vacuum coater. Particulates on the surface of silicon wafers were removed gently with tissue, and the wafers were cleaned by soaking overnight in isopropyl alcohol prior to being placed, face down, above the crucibles within the vacuum coater. Irganox layers were alternately evaporated onto the silicon wafer while monitoring relative thickness using an in situ quartz crystal microbalance (QCM). The total film thickness was measured ex situ using an M2000 spectroscopic ellipsometer (Woollam, NE), employing microfocus optics (short axis spot size 150 µm) to establish the uniformity of the coating thickness across the wafer. The depth of each layer was found by scaling the QCM measurements, assuming the QCM mass response to be linear over the range of thicknesses deposited and identical for the two different materials. In this manner, it was possible to obtain pinhole-free coatings with a variation in thickness of less than 5% across the ∼1 cm width of the wafer. The variation in thickness across the area analyzed (typically 100 µm × 100 µm) in depth profiling experiments is thus expected to be less than 0.05%. Analyses at different beam energies were taken within 4 mm of each other, leading to an error of ∼2% in the determination of relative sputtering yields arising from variations in layer thickness. Two sets of delta layer materials were prepared; one set with six delta layers (Irganox3114) separated by ∼90 nm layers of matrix (Irganox1010), referred to hereafter as ‘6 by
90’ and one with three delta layers separated by ∼55 nm layers, subsequently referred to as ‘3 by 55’. SIMS depth profiles were made using a TOFSIMS IV (IONTOF Gmbh, Mu¨nster, Germany) time-of-flight secondary ion mass spectrometer. The instrument was equipped with both a C60 ion source and a Bi ion source mounted orthogonally to each other and at 45° with respect to the sample surface. The time-of-flight mass analyzer was perpendicular to the sample surface. Depth profiles were carried out in the ‘interlaced’ mode, consisting of cycles of short pulses of 25 keV Bi3+ primary ions for SIMS analysis followed by longer periods of C60n+ ion sputtering, during which the analyzer extraction potential is switched off. The use of two ion beams is a convenient experimental arrangement because each beam can be optimized for either sputtering (high direct current, defocused) or analysis (pulsed current, focused), and the experiment is quicker than using a single ion source which must settle each time it is switched between modes. The time for each cycle (∼0.12 ms) determines the maximum detectable mass (∼1200 u) during the experiment. The C60 ion source has a maximum operating potential of 10 kV, and by selecting singly, doubly, and triply charged ions, this provides a range of ion energies up to 30 keV. Beam currents were measured using a Faraday cup on the sample holder both before and after each experiment. Beam currents of between 200 and 800 pA were employed and found to vary by less than 5% during the course of a typical experiment. In these experiments, we utilized 5 keV (C60+, 5 kV), 10 keV (C60+, 10 kV), 20 keV (C602+, 10 kV), and 30 keV (C603+, 10 kV) energies. During the profile, the C60 ion beam, which was defocused to have a spot size of ∼50 µm diameter, was rastered over a 400 µm × 400 µm area with a pixel spacing of ∼3 µm. The ion dose should therefore be constant within the central ∼150 µm × ∼150 µm area.
2598 J. Phys. Chem. B, Vol. 112, No. 9, 2008
Shard et al.
TABLE 1: Yield Volumes (nm3 per ion) and Unique Secondary Ions for the Materials Employed to Construct Delta-Layers (The Yield Parameters from eq 2 Used to Fit Delta-Layer Data Are Shown for Comparison) C60n+ energy
a
Irganox 101024
Irganox 3114 (this work)
5 keV
not measured
not measured
10 keV
82 (8
92 ( 9
20 keV
163 ( 16
186 ( 19
30 keV
254 ( 25
295 ( 30
unique secondary ions
C73H107O12- (1175 u; M1010 - H) C17H25O3- (277 u) C16H23O- (231 u)
C33H46N3O5- (564 u; M3114 - R) C18H24N3O4- (346 u) CNO- (42 u) CN- (26 u)
yield parameters from delta-layers (this work) S0 ) 36; S∞ ) 0a σDS ) 0.050 nm2 S0 ) 85; S∞ ) 0a σDS ) 0.118 nm2 S0 ) 174; S∞ ) 24 σDS ) 0.324 nm2 S0 ) 264; S∞ ) 80 σDS ) 0.618 nm2
Data insufficient to determine; value assumed for fitting purposes.
To avoid any potential problems arising from the overlap of the analyzed area with areas that have not received a constant C60 ion dose, only the central 100 µm × 100 µm area was analyzed with the Bi3+ primary ion beam. The Bi3+ primary ion beam had a spot size of 500 nm thickness. These films had been sputtered using a known dose of C60n+ ions, rastered over a 500 µm × 500 µm area. Micrographs were acquired using a scan range of 10 µm × 10 µm, and three micrographs from separate positions close to the center of each sputtered area. The micrographs were used to obtain height distribution functions (number of pixels at a given height, relative to the mean height) and root mean squared roughness (Rq) values. The virgin surface of Irganox1010 films is invariably found to be very smooth (Rq < 1 nm) relative to the C60n+ ion-irradiated surfaces. Results and Discussion The determination of energy dependent yield volumes for Irganox1010 has been described previously.24 Briefly, an homogeneous film is sputtered until the film is totally removed and the dose required (F) for the intensity of a characteristic secondary ion from the substrate to reach 50% of its maximal value is recorded. From this dose and knowledge of the film thickness (D), a yield volume can be directly calculated (S ) D/F), assuming that the yield does not change during the profile. A 65.4 nm film of Irganox3114 was used to establish that the yield volume for this compound was not significantly different to that of Irganox1010. Comparisons of yield volumes, along with a list of unique secondary ions that can be used to distinguish the two compounds, are provided in Table 1. The ions utilized in the majority of the paper are denoted as (M1010 - H)-, deprotonated Irganox1010 at 1175 u, and (M3114 - R)-,
the highest mass negative secondary ion observed from Irganox3114 at 564 u. As can be appreciated from the chemical structures, shown in Figure 1, there are numerous secondary fragment ions common to both compounds, and this is particularly problematical within the positive ion spectra. However, there are characteristic secondary fragment ions in the negative ion spectra, which arise as a consequence of the presence of nitrogen in Irganox 3114 and the higher yield of high mass negative secondary ions in comparison to high mass positive ions for these phenolic compounds. The observation of these high mass ions throughout the C60n+ profiles is an indication of the low level of accumulated damage. The (M1010 - H)intensity typically drops to ∼20% of its initial intensity within 20 nm of the surface and thereafter remains relatively constant. Considering that there are 191 atoms in the ion and a single damage event in a molecule will result in the incapacity to form this secondary ion, this represents an extremely low level of damage (
