Article pubs.acs.org/JPCC
SIMS of Delta Layers in Organic Materials: Amount of Substance, Secondary Ion Species, Matrix Effects, and Anomalous Structures in Argon Gas Cluster Depth Profiles M. P. Seah, R. Havelund,* and I. S. Gilmore Analytical Science Division, National Physical Laboratory, Teddington, Middlesex TW11 0LW, United Kingdom ABSTRACT: A study is reported of the quantification of the amount of matter by secondary ion mass spectrometry (SIMS) when depth profiling a nominally 3.1 nm thick delta layer of FMOC-L-pentafluorophenylalanine in Irganox 1010. The depth profiles are made using 5 keV Ar2300+ cluster ions with analysis by 25 keV Bi3+ ions. Data for 89 negative secondary ions shows profiles whose integrated intensities as a function of depth, even when normalized to the intensity for the pure material, still vary over a factor of 12. This variation mainly arises from matrix effects that are measured here using separate samples with mixed layers of three intermediate compositions of the two materials. Matrix enhancements or suppressions vary widely from secondary ion to secondary ion and are not related to the energy of the analyzing Bi3+ ion. Strong effects can cause the delta layer signal to show structure that may be misinterpreted. The compositional profile is established by using trial Dowsett profiles, representing the composition, which are then enhanced or reduced according to the measured matrix effect with the result then fitted to the normalized intensity data. By fitting each secondary ion profile separately, the computed amount of matter still varies weakly with the enhancement. This arises as a result of a longer wavelength roughening, equivalent to a rootmean-square value of about 2.5 nm, which causes, for example, the measured maximum intensity to be lower than the actual maximum intensity appropriate for the delta layer. The effective matrix enhancement is thus reduced by 20%. When this is included, it is found that the variation with enhancement disappears and the amount of matter is found to be equivalent to 3.22 ± 0.07 nm although the scatter from individual ions has a standard deviation of 0.45 nm. It is concluded that the matrix terms used are a good description of the phenomenon and that SIMS profiles may be made quantitative if suitable secondary ions are available and the matrix terms are measured. In the absence of measured matrix terms, the measured quantities are prone to large errors. concentration present.8 At dilute levels well below 1% this linearity was assumed, and excellent work is possible for homogeneous or delta layer dopants in semiconductors. In the profiling of organic materials using cluster ions, the matrix effect is associated with the analysis ion9 and both suppression and enhancement may occur.10,11 Deliberate enhancement, by altering the matrix, can improve the sensitivity.12,13 In 2015, Shard et al.14 measured the matrix effect in molecular mixtures of two molecules sputtered with argon cluster ions and analyzed with 25 keV Bi3+. These data were interpreted in terms of a simple model of charge transfer with two parameters α and β. The overall effect was summarized in a single parameter, Ξ, which described the total intensity across the phase diagram compared with the result in which there was no enhancement, i.e., at Ξ = 0 there is no enhancement and at Ξ =
1. INTRODUCTION The analysis of compositional depth profiles of organic materials by secondary ion mass spectrometry (SIMS) has now become relatively routine as a result of the use of argon gas cluster ions for the sputter removal of material.1,2 These cluster ions typically consist of 1000−5000 argon atoms and give a very gentle removal of surface layers that exposes a new surface in which the material chemistry is preserved with little degradation.1 The sputtering by the argon clusters permits depths of 1000 nm to be reached, also with little degradation to the depth resolution, which is typically around 10 nm even in the absence of sample rotation.3 This provides molecular depth profiling and three-dimensional imaging of a wide range of organic materials and biological specimens and has found applications, for example, in the analysis of organic electronic devices,4 drug delivery systems,5 and cells.6,7 The superb sensitivity and prospects for imaging in 3D are clear, but from the earliest work, it was recognized that the prospects for quantification were difficult. Matrix effects caused the ion signal not to be linearly proportional to the Published XXXX by the American Chemical Society
Received: August 26, 2016 Revised: September 30, 2016 Published: October 26, 2016 A
DOI: 10.1021/acs.jpcc.6b08646 J. Phys. Chem. C XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry C
Article
attributed this to increased annealing of the sputtered craters leading to reduced roughness at the higher temperatures. The crater bottom roughness is clearly important, but other effects must also be considered. Havelund et al. studied the depth dependence of the secondary ion generation and emission23 that leads to different apparent positions and widths of the delta layer profile for different characteristic secondary ions.24 Using a 50 keV Bi3+2 primary ion beam, the small CNO− fragment of the Irganox 3114 compound was observed 2.5 nm before the large characteristic Irganox 3114 fragment in the depth profile. This showed that the small fragment, on average, comes from deeper in the sputter crater. An even stronger effect on the delta layer profile, as we shall see in the present work, is that of the matrix effect. Ionization enhancement or suppression at interfaces have previously been reported to shift the apparent position of an interface more than 10 nm.15 Here, the thin layer studied is nominally a 3.1 nm layer of FMOC between thicker layers of Irganox 1010. Separate measurements of the matrix effect are required and are conducted on samples containing thicker, uniform, layers of mixtures of the two materials.
1 the total intensity is twice that with no enhancement. For suppression, Ξ is negative but cannot be less than −1. An interlaboratory comparison15 showed that Ξ is consistent between instruments, as long as similar operating conditions are used, and so measurements of this may be useful for establishing widely applicable recommendations for measuring compositions. For the materials used in that interlaboratory comparison, the authors found that all study participants could have more accurately measured the layer compositions for 20:80 and 80:20 (fractions by volume) mixtures by correcting for the matrix effect. A special, and important, case is that of materials or devices consisting of thin layers of pure single materials where a measurement of the thicknesses of the layers provides, quantitatively, the amount of substance in the layers. In sputter depth profiling, the thickness of a layer will be reflected in the primary ion dose, i.e., number of ions per unit area, required to remove the layer, and in the integrated characteristic secondary ion signal intensity through that layer. In the present work, we study the quantification of the amount of substance in organic layers with thicknesses smaller than the depth resolution in SIMS. For these systems, the interfaces for the layers below and above the layer to be measured are not resolved in SIMS depth profiles and so quantification cannot be achieved from measuring the sputter ion dose between the interfaces and using a separate measurement of the sputtering yield. Instead, quantification must rely on an understanding and measurement of the matrix effect and effects that broaden and shift depth profiles. Carefully constructed organic delta layer reference materials have been instrumental for studying those effects. Shard et al.16 used a C60+ ion beam for sputter depth profiling a delta layer material consisting of 1 nm delta layers of Irganox 3114 separated by 50−100 nm layers of Irganox 1010. In the depth profiles, the delta layers were broadened to a width of more than 10 nm full width at half-maximum (fwhm) and were found to be described by Dowsett et al.’s analytical function17−20 for delta layer profiles originally intended to describe inorganic dopant delta layers obtained with monatomic sputter ion sources. This function is a convolution of two exponentials, one growing and one decaying with a Gaussian function, and is typically asymmetric with the decay length being longer than the growth length. The fwhm of the delta layer profiles is a commonly used definition of the depth resolution in sputter depth profiling. Importantly, Shard et al.16 showed that the surface height distribution, as measured by AFM, was described by the same Dowsett function indicating that the response function in organic depth profiling is largely a result of the crater bottom roughness. The effect of crater bottom roughness has also been observed for profiles obtained with argon cluster sputtering. Niehuis et al.21 measured the depth resolution for argon clusters of different E and n values and found that this degraded with increasing depth when sputtering with argon clusters below 5 keV. They found that the degrading depth resolution followed an increase in surface roughness and showed that the optimal depth resolution of 5 nm fwhm could be maintained to depths of more than 300 nm when rotating the sample during sputtering to reduce roughness formation. In a study of the effect of sample temperature in organic depth profiling, Seah et al.22 found that the apparent width of Irganox 3114 delta layers in a thin film of Irganox 1010 decreased with increasing sample temperature from 0 °C up to 45 °C and
2. EXPERIMENTAL SECTION Two types of sample were made using the same combination of materials; one of delta layers to establish the profiles and one of mixtures to establish the matrix effects. The delta layer samples were generated as described by Shard et al.24 and are detailed further in Havelund et al.23 Briefly, Irganox 1010 (pentaerythritol tetrakis(3-(3,5-ditert-butyl-4-hydroxyphenyl)propionate), C73H108O12, Mr = 1177.6) and fluorenylmethyloxycarbonyl-L-pentafluorophenylalanine (abbreviated here to FMOC, C24H16F5NO4, Mr = 477.4) from Sigma-Aldrich were each sublimed in a Qbox 450 (Mantis Deposition Ltd. Thame, UK) with relevant monitoring, shuttering, and sample rotation to create the delta layer structures as shown in Figure 1 of Shard et al.24 The evaporators were controlled by the outputs of stationary quartz crystal oscillators (QCOs) to deposit three layers of Irganox 1010 of 98.9 nm and then 2 final layers of 49.4 and 49.5 nm thickness. Each layer was separated by a 3.1 nm layer of FMOC. The QCOs were calibrated to relate their outputs to the thicknesses of each material deposited on the
Figure 1. Depth profiles with 5 keV Ar2300+ sputtering and analysis by 25 keV Bi3+ ions for the delta layer material using m/z 19.0 (red line) and 42.0 (black line) Da negative secondary ions characteristic of FMOC. The intensities are normalized by the intensities from pure FMOC, and the depth is calibrated to the known layer centroid positions to a standard deviation of 0.5 nm. B
DOI: 10.1021/acs.jpcc.6b08646 J. Phys. Chem. C XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry C
Article
The matrix parameters are determined as before25 using the profiles of the mixture sample with volume fractions of 0.0, 0.2, 0.5, 0.8, and 1.0 for the Irganox 1010 with FMOC. We fit the measured intensities in the equation describing the enhancement of the intensity of a secondary ion from A arising from a content of B, ϕB, as given by Shard et al.14 As before,25 we modify their equation to give
wafer substrates by ellipsometry using an M2000DI spectroscopic ellipsometer (Woollam, NE, USA). The mixture samples, however, were made as discussed by Seah et al.25 and shown in Figure 1 there. They were composed of 100 nm of pure layers, bracketing 100 nm layers of the mixtures. The pure layers were each prepared as single deposits of 100 nm, but the mixed layers were made using 50 cycles of small depositions of one material followed by deposition of the other; the pair of deposits being 2 nm thick. Since the molecular sizes are around 1 nm, this generates intimate mixing. Compositions of nominal 20, 50, and 80% by volume were prepared. These are called volume fractions and are prepared, for example, by mixing 20% by volume of Irganox 1010 and 80% by volume of FMOC with the total volume assumed unchanged as a result of mixing. The above multilayer samples were depth profiled by SIMS using 5 keV Ar2300+ gas cluster primary ions in an ION-TOF SIMS IV instrument (ION-TOF GmbH) with the incident ions at 45° to the surface normal and the cluster size distribution selected with a width of ∼30%. Negative secondary ions were measured using 15, 25, and 50 keV Bi3+(+) ions also at 45° incidence angle, but in an azimuth at 90° to the argon gas cluster sputtering beam. The sputtering beam was rastered, in interlaced mode, over an area of 500 μm by 500 μm, and the analysis was in a central zone of 200 μm by 200 μm. The relative Bi3+(+) dose was